Animal Science Reviews 2010

Animal Science Reviews 2010 Animal Science Reviews Animal Science Reviews provides scientists and students in the ﬁel...

Author: David Hemming

99 downloads 2331 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Animal Science Reviews 2010

Animal Science Reviews Animal Science Reviews provides scientists and students in the ﬁeld with timely analysis on key topics in current research. These chapters were originally published online in CAB Reviews. This volume makes available in printed form the reviews in animal science published during 2010.

Animal Science Reviews 2010

Edited by

David Hemming CAB International, UK

CABI is a trading name of CAB International

CABI Head Ofﬁce Nosworthy Way Wallingford Oxfordshire OX10 8DE UK Tel: +44 (0)1491 832111 Fax: +44 (0)1491 833508 E-mail: [email protected] Website: www.cabi.org

CABI North American Ofﬁce 875 Massachusetts Avenue 7th Floor Cambridge, MA02139 USA Tel: +1 617 395 4056 Fax: +1 617 354 6875 E-mail: [email protected]

g CAB International 2011. All rights reserved. No part of this publication may be reproduced in any form or by any means, electronically, mechanically, by photocopying, recording or otherwise, without the prior permission of the copyright owners.

ISBN: 978 1 84593 880 2 Commissioning editor: Nigel Farrar Production editor: Mike Bodinham Typeset by S R Nova Pvt Ltd, Bangalore, India Printed and bound in the UK by CPI Antony Rowe, Chippenham and Eastbourne

Contents 1. Protein nutrition and nitrogen balance in buffalo cows G. Campanile, G. Neglia, D. Vecchio, R. Di Palo, B. Gasparrini and L. Zicarelli

1

2. Expression QTL and their applications in genetic improvement of farm animals K. Wimmers, E. Murani and S. Ponsuksili

9

3. A brief history of mammalian coat colour genetics J. J. Lauvergne 4. Functional genomic approaches to understand the biological pathways underpinning intramuscular fat in beef L. Pannier, R. M. Hamill, A. M. Mullen and T. Sweeney

17

23

5. Assessing impacts of organic production on pork and beef quality Albert Sundrum

35

6. Whole genome marker-assisted selection Joel I. Weller

49

7. Quantitative trait mutations in cattle, sheep and pigs: a review Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens, and Nadine Buys

59

8. Technical, epidemiological and ﬁnancial implications of large-scale national vaccination campaigns to control HPAI H5N1 J. Hinrichs, J. Otte and J. Rushton 9. Accelerated lambing: Part 1. Housing for sheep G. L. Hunter 10. Accelerated lambing: Part 2. Increasing the frequency of pregnancy in sheep G. L. Hunter 11. Poultry sector development, highly pathogenic avian inﬂuenza and the smallholder production systems J. Rushton, R. E. Viscarra, N. Taylor, I. Hoffmann and K. Schwabenbauer 12. Fat and taste perception Philippe Besnard, Dany Gaillard, Patricia Passilly-Degrace, Ce´line Martin and Michael Chevrot 13. Microgoats in India and their role in future animal production in an era of climatic change Pramod Kumar Rout and M. C. Sharma

73 93 115

139 147

157

14. Assessing biosafety of GM plants containing lectins Morten Poulsen and Jan W. Pedersen

165

15. Ovine and caprine brucellosis (Brucella melitensis) Assadullah Samadi, M. M. K. Ababneh, N. D. Giadinis and S. Q. Laﬁ

171

16. Plants as reservoirs for human enteric pathogens Nicola J. Holden

185

17. Use of high-density marker genotyping for genetic improvement of livestock by genomic selection Jack C. M. Dekkers

197

18. Natural antimicrobials for food processing Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´n-Belloso

211

19. Palatability: principles, methodology and practice for farm animals J. Michael Forbes

229

20. Embryo cryopreservation in domestic mammalian livestock species C. R. Youngs, S. P. Leibo and R. A. Godke

245

21. Risk assessment of toxic contaminants in animal feed Alberto Mantovani and Chiara Frazzoli

257

Index

271

Animal Science Reviews 2010

Protein nutrition and nitrogen balance in buffalo cows G. Campanile*, G. Neglia, D. Vecchio, R. Di Palo, B. Gasparrini and L. Zicarelli Address: DISCIZIA, Faculty of Veterinary Medicine, ‘Federico II’ University, V.F. Delpino 1, Naples 80137, Italy. *Correspondence: G. Campanile. Fax: +39-081-292981. Email: [email protected] 21 September 2009 1 December 2009

Received: Accepted:

Abstract The buffalo species represents the mainstay of the rural economy of small farmers in many developing countries, and is also an important source of protein. Because of its area of origin, it shows some peculiar characteristics regarding protein requirements and nitrogen balance. Therefore, this review aims at clarifying some aspects of protein digestibility and, consequently, manure management in buffalo cows. Keywords: Buffalo, Protein, Nitrogen balance

Introduction Livestock products in tropical areas play a crucial role, which extends beyond their traditional supply of meat and milk. Unprecedented economic growth in developing countries, accompanied by the increases in income and purchasing power and by changes in food preferences, together with the growth of human population, has increased demands on the livestock sector. Livestock products, such as milk and meat, have undergone great modiﬁcation in response to these recent developments [1]. The buffalo is considered a triple-purpose animal, providing milk, meat and draught, though milk is the main product, followed by meat. However, it is worth pointing out that the swamp type of buffalo is used for work in Southeast Asia and China, and it often represents the mainstay of the rural economy of small farmers in many developing countries, as well as becoming an important source of meat and milk. When buffaloes are compared with other animals, they are more adaptable and show many homeo-kinetic mechanisms to maintain critical body functions. This condition, which is obtained at the expense of changes in other physiological functions, allows them to represent an economic and valuable species in rush areas of tropical countries [2]. The water buffalo is a domestic species raised in many different parts of the world. Its importance for the wealth and welfare of the human population has been recognized

g

in recent years. About 97% of the total world buffalo population is located in South and Southeast Asia, and about half of this population is bred in India [3]. In India, river buffaloes are an important source of milk, representing about 35% of India’s milch animals and supplying around 70% of the total milk production [3]. The nutritive forage characteristics of equatorial and sub-equatorial zones could be limiting factors for milk yield in buffaloes bred in these areas. Tropical forage shows low levels of growth (less than 20% annual growth), low digestibility (because of a high concentration of ligneous ﬁbre) and low crude protein (CP) concentration [4]. However, it is worth pointing out that buffalo breeding is usually preferred to that of dairy cattle in these zones, because of their superior quality of milk, better efﬁciency in utilization of nutrients from poorquality ﬁbrous tropical feeds and relatively better disease resistance and adaptability to tropical climates [5]. The dry matter (DM) intake in buffaloes is reduced when the proportion of indigestible ﬁbre increases [6, 7]. The forage protein content represents the main factor limiting productive capacity of buffalo cows. Low protein concentration in the feeds causes a considerable diminution of production and duration of lactation together with small variations in milk composition. This phenomenon is partially responsible for the low production (1713.83+576.26 litres), short lactation period and the chemical composition of buffalo milk in Brazil [8]. Consequently, protein nutrition in buffalo cows plays a

CAB International 2010

2

Animal Science Reviews 2010

fundamental role in milk yield and in reproductive activity. In the intensive breeding system, diets with higher protein content are always administered at start of lactation to promote milk yield and length of lactation. This condition could enhance nitrate pollution risk. Consequently, the optimization of protein requirements in intensive breeding favours the management of manure and risk of nitrate pollution. Hence, it is necessary to understand nitrogen metabolism to evaluate the nitrogen balance in buffaloes. The current review presents a summary of knowledge in nitrogen metabolism and nutrition in buffalo cows. The review also identiﬁes gaps in knowledge on protein nutrition and digestibility that could be considered important for the milk production and management of manure in the intensive breeding system.

Urea Metabolism The urea nitrogen concentration in plasma and milk in cattle is inﬂuenced by the amount of CP in the diet [9–11], as well as by degradable intake protein (DIP) and undegradable intake protein (UIP). Moreover, it is clearly correlated with dietary protein supply relative to requirements [11]. Roseler et al. [12] also showed that urea concentrations in blood and milk are inﬂuenced by changes in DIP and UIP and by increasing energy intake. The relation between the protein fraction of the diet and the protein/energy (P/E) ratio in cattle has previously been described [13, 14]. Blood urea concentration decreased in lactating dairy cows when an optimal level of ruminally fermentable carbohydrate was supplied to enhance the capture of DIP into microbial protein [15]. Excess degradable protein, responsible for increases in blood and milk urea (MU) levels, is shown to be poorly used in the production of milk proteins. In fact, Baker et al. [11] noted an increase in ‘true’ proteins when the DIP and UIP were consistent with production requirements. Azotaemia is closely correlated with MU concentration [11–13, 16]. This parameter can be used, since it facilitates monitoring for sampling feeding adequacy. This analysis cannot replace conventional feed analysis, but can be utilized as a valuable complement to it [13]. The change in MU content in response to an altered dietary composition is very rapid in cattle [14]. In buffalo cows, as in dairy cows, MU is positively related to blood urea (BU) [17]. The relationship between MU and BU in buffalo cows [8] is expressed by the following equation, which is similar to the results found by other authors in cattle [10–12]: BU(mmol/l)=1:675+0:817 MU(mmol/l); R2 =0:769: In buffaloes at the start of lactation, urea levels in milk and blood are conditioned by CP intake and days in milk, while protein degradability does not inﬂuence either parameter [18]. The relationship is described in the

g

following equation: MU(mmol/l)=712:475+0:008 CP intake (g) +0:029 days in milk; R2 =0:587 As speciﬁed above, nitrogen levels in blood are affected by days in milk and diet. The trend of nitrogen levels in blood of buffaloes fed a diet of low protein content is similar to that of DM intake throughout lactation. Low nitrogen levels in blood are responsible for higher growth hormone (GH) secretion in dairy cattle. GH favours fat mobilization, inhibits insulin and increases nitrogen availability from tissues, in order to optimize the P/E ratio for production [19]. When higher protein concentrations are utilized in the diet, which in our opinion are more suitable to guarantee protein requirements, in particular, at the start of lactation, feeding is characterized by low DM intake, and in high-producing buffaloes, the lowering in nitrogen levels does not occur [20]. On the contrary, in middle lactation [17] MU and BU are conditioned by the P/E ratio: MU(mmol/l)=75:371+41:60 CP/NSC753:58 UIP/NSC; R2 =0:906 where NSC is the non-structural carbohydrate. Carruthers et al. [21] found that in cows at late lactation, the P/E ratio increased microbial protein synthesis. Urea levels in blood and milk increase as a function of days open because of the higher DM intake and hence CP, as usually observed in the buffalo cow [22]. In buffaloes in mid-lactation, the urea concentration in blood and milk reached a higher level when a low-protein diet was administered for a fairly long period. In this species, a decline in circulating urea occurred after a sharp reduction [17]. We suppose that an increase of insulin levels reduced or blocked the normal amino acid breakdown, and therefore reduced the urea levels in the blood. Smith [23] reports that in ruminants from tropical areas, nitrogen deﬁciency decreases kidney clearance, increases ruminal return and decreases haematic levels of urea. As a result, the urea recycling in the digestive tract and the ruminal bacteria protein synthesis are improved [24]. The lower P/E ratio in buffalo milk compared with the bovine milk allows buffaloes to use forage characterized by low protein content [4, 5]. This may explain the adaptation of this species in South American countries where forage is characterized by a low P/E ratio [4]. In any case, we must underline that buffaloes react to protein deﬁciencies better than dairy cattle [25]. The utilization of diets characterized by a larger P/E ratio in buffalo species causes fewer deleterious effects than reported in dairy cattle. This is probably the result of higher ammonia utilization by rumen microbes or to a higher urea synthesis in the liver, which would reduce the increase of blood ammonia. In fact, buffaloes utilize diet nitrogen better than cattle in carbohydrate deﬁciency

CAB International 2010

G. Campanile, G. Neglia, D. Vecchio, R. Di Palo, B. Gasparrini and L. Zicarelli

[26], because the ruminal environment of the former species is more favourable for the growth of microbes that utilize non-protein nitrogen. Singh and Gupta [27] observed a higher level of volatile fatty acids and adenosine triphosphate (ATP) in the rumen of buffaloes that received a higher quantity of nitrogen through ammoniatreated straw. Hence, it is conﬁrmed that urea levels in blood and milk are positively correlated to CP/(NSC) ratio in buffalo species. In fact, the higher energy availability allows better ammonia utilization by microbial fauna for protein synthesis, causing low urea production [28]. Ammonia blood level is negatively affected by rumendegradable protein (RDP) intake and, as found in dairy cows, it is not conditioned by BU [29]. The authors attributed this result to the fact that urea peaks 4–8 h after feeding [29]. Ammonia values in buffalo cows are higher than those reported by Elrod and Butler [30] in dairy cows and McEvoy et al. [31] in sheep, and lower compared with the results of Jordan et al. [32] and GarciaBojalil [33] in cattle. We can conclude that when protein requirements are satisﬁed in both buffalo and cow, urea blood levels in buffalo are higher than those of dairy cattle, whose physiological values (28–32 mg) cannot be considered as a suitable reference. This has been conﬁrmed by the evidence that between October and February, buffaloes bred in Egypt, India and Iran receive a high quantity of Trifolium alexandrinum L., which is characterized by higher CP on DM than alfalfa (lucerne, >25%), without suffering any health problem such as alkalosis or laminitis.

Protein Requirements As mentioned above, protein requirements in buffalo species are lower than those reported for dairy cows if the same energy is produced by milk [4]. In a trial carried out on buffaloes fed with six different P/E ratio diets, Kurar and Mudgal [34] reported a maintenance requirement of digestible CP (DCP) of 3.2 g/kg W0.75/day and 166.34 g DCP/100 g of milk-produced proteins. Analogous values were previously observed in Indian Murrah buffaloes [35]. The results obtained in a study carried out to evaluate the effects of the P/E ratio on milk characteristics and on blood and MU levels showed 2.53 g of CP/kg W0.75 and 2.7 g of CP/g of protein produced with milk [17]. According to Rai and Aggarwal [36], a diet with a protein concentration of 11–14% should be administered to buffaloes during lactation. It has been noticed that the protein concentration used in Mediterranean buffaloes could be lower than 12% [37, 38]. The protein concentration depends on the genetic merit of the buffaloes and on the number of days in milk. As mentioned above, this changes throughout the year because of the concentration of calving both in seasonal herds and in the out-of-seasonal ones. Protein

g

3

concentrations 9% lower blood and MU levels and increase the milk freezing point, especially if combined to a high level of fermentable energy [17]. Therefore, it is important to evaluate the P/E ratio, to obtain a balanced diet. In buffaloes, this ratio may be higher, as the administration of high protein concentrations has less deleterious effects than in dairy cows. In our studies, we noticed that an increase of the protein concentration (up to 15% DM) at the beginning of lactation stimulates the productive activity of the buffaloes, without affecting the increase of fat mobilization. Nonesteriﬁed fatty acid (NEFA) values were lower than those registered in animals that ingested DM with lower CP levels. It is known that in cattle a higher protein concentration increases milk yield because the increase of absorbable proteins quickens fat mobilization [39, 40]. Post-ruminal casein infusion determines an increase of the palmitate percentage [41]. It seems that high protein concentrations stimulate GH [42, 43], which is responsible for the decreased response of the adipose tissue to insulin and for a greater sensitivity to the adrenergic stimulus at the end of gestation and at the beginning of lactation. These stimuli might be mediated by gastroenteric hormones that can modulate the sympathetic nervous system activity [44]. Conﬁrming this hypothesis, Cado´rniga and Lopez Diaz [45] noticed a greater fat mobilization in cattle that showed a higher Body Condition Score (BCS) and fed with diets containing a higher dose of rumen undegradable proteins (RUP); they also reported that their adipose tissues were more sensitive to epinephrine administrations. As mentioned above, dietary proteins and their digestibility in the rumen affect BU and MU concentration [9–11]. Overfeeding proteins has been associated, in sheep and cattle, with a decline in fertility in most [30–46], but not all [47] studies. Jordan et al. [32] found an increase in urea-N concentration in uterine secretions and plasma of cows fed 23% CP. Elrod and Butler [30] showed that an excess of either digestible (RDP) or indigestible (RUP) protein in the rumen increases BU and alters uterine pH to a similar degree, interfering with the normal inductive effects of progesterone on the microenvironment of the uterus, thereby providing sub-optimal conditions to support embryo development. Moreover, the elevation of ureic nitrogen in the local (reproductive) or systemic apparatus reduces the ovarian receptors’ luteinizing hormone (LH) linkage and, therefore, it would be responsible for the lower production of progesterone in the post-ovulation days and for the lower conception rate. It is known that this ovarian steroid is fundamental for embryo progression in the uterine tube, for the production of uterine milk that represents the only nourishment of the embryo before implantation and, ﬁnally, for pregnancy maintenance. A reduced hypophisary activity, characterized by a lower progesterone secretion in the ﬁrst phases of gestation, causes early embryo mortality. It seems that the unfavourable action of ammonia on reproductive activity is insulin mediated [48] through

CAB International 2010

4

Animal Science Reviews 2010

its known effects at the hypothalamic level. It has been hypothesized that ammonia acts on the central nervous system through the increase of glutamate and depletion of a-ketoglutarate, interrupting ATP synthesis [48]. A diet with 50% protein content higher than required led, in Mediterranean buffaloes, to an increase of heat phase followed by an appropriate luteinic phase and a prolongation of the inter-oestrus interval [49]. In this species, anoestrous is favoured by lack of CP and energy [50]. Analogous results have been noticed in beef cattle with a lack of energy and protein [51]. Protein and/or energetic underfeeding determines a decrease of ruminal microbial activity and, therefore, of the whole volatile fatty acid content. The organism responds through an increase in GH and a decrease in insulin, allowing the mobilization of the muscular proteins used for glucose synthesis and, therefore, for glycaemic homeostasis. Even in this case, the reduction of insulin activity could interfere with buffalo reproductive activity. Protein digestibility in the rumen does not inﬂuence reproductive activity in buffalo cows [18]. In a trial carried out on Italian Mediterranean buffalo cows [18], the urea values measured in vaginal mucus were similar [32] or higher than those reported, respectively, in dairy cow [52] and sheep [31]. By contrast, the ammonia levels in vaginal mucus were considerably lower than those found by Rumello et al. [52] in dairy cow and by McEvoy et al. [31] in sheep. High levels of ammonia in the uterus affect fertility in dairy cows because of the detrimental effect on embryo development [31]. It is possible that in buffalo, independently of the BU, a lower diffusion of ammonia occurs in the uterus, reducing the detrimental effect on reproductive efﬁciency. It is widely reported that the negative effect of a protein surplus on fertility, in cattle, may be the result of the RDP intake [29–31, 32]. The negative effect of RDP excess is not observed in buffalo cows [18] since this species uses nitrogen better than cattle, including when in NSC deﬁciency [26]. Energy deﬁciency decreases the ability of the liver to transform ammonia into urea. This results in an increase in blood ammonia levels and, by diffusion, levels of vaginal mucus ammonia. In addition, Schepers and Meijer [53] found a relationship between diet energy and MU. The high values of fertility registered in buffaloes fed with a high CP levels show that urea levels in the blood and in vaginal mucus do not have a negative effect on reproductive performance [18] in natural mating. This may well reﬂect the tropical (north of the Equator) origin of buffalo, which modiﬁes diet protein use according to season and feed availability. Indeed, in tropical regions, forage availability occurs only for a few months.

Nitrogen Balance The knowledge of nitrogen balance in buffalo species is fundamental in order to optimize the production and to

g

Table 1 Growth rate (g/d) and feed efﬁciency of growing Bhadawari calves fed on different energy diets [54] modiﬁed by Campanile et al. (unpublished data) Parameters Nitrogen balance Intake Faecal-N Digestible-N Urinary-N N balance % N absorption Growth Initial weight (kg) Final weight (kg) Growth rate (g/d) Growth rate for 1 gram of N1 Feed efﬁciency Feed conversion ratio CP conversion ratio

EL

EL1

EL2

SEM

41.4 18.5 22.9 8.4 14.4 34.8

38.2 16.9 21.2 10.9 10.3 27.0

39.2 17.8 21.4 6.0 15.4 39.2

2.71 1.52 1.47 0.83 1.16 2.14

75.9 143.7 368.5ab 25.5

74.9 129.8 298.4a 28.9

77.5 146.3 374.3b 24.3

7.62 12.9 – –

7.76 0.70

8.49 0.80

7.25 0.66

0.37 0.04

1

Growth rate (g/d): N balance. Superscripts within a row differ signiﬁcantly (P < 0.05); SEM, standard error of means.

a,b

perform adequate manure management. Few trials have been carried out in buffalo species compared with bovine. Some Indian authors [54] evaluated the nitrogen balance in female buffalo calves to be a mean between 76 and 140 kg of live body weight with three iso-nitrogenous concentrate mixtures (CM) of different energy density (CM1: 2.56, CM2: 2.20 and CM3: 3.04 M cal) to meet their energy and protein requirements in EL [55], EL1 (20% less metabolizable energy (ME) than EL) and EL2 (20% more ME than EL). Independently of the dietary energy levels, the animals intake a mean of 39.6 g/day of nitrogen and excreted 17.7 g/day as faeces and 8.4 g/die as urine (Table 1). Hence, all the groups were in positive N balance, although the urinary-N loss was higher in EL2 (10.9 g/day) than in other groups (Table 1). The authors concluded that the efﬁciency of protein utilization in buffaloes is reduced when the diet is characterized by low energy concentration [54]. However, if the data are analysed taking into account both the daily weight gain and N availability, it is clear that 1 g of retained N allows the growth of 25.5, 28.9 and 24.3 g for EL, EL1 and EL2, respectively. Hence, when high amounts of N are available, the efﬁciency of utilization is lower. In fact, in a recent trial performed on growing male Thai buffaloes [56], 1-year-old male calves were fed ad libitum with four diets, using four combinations of pineapple waste silage (P) and concentrate (C), in the proportions (on a DM basis) of 0.8 : 0.2 (P80:C20), 0.6 : 0.4 (P60:C40), 0.4 : 0.6 (P40:C60) and 0.2 : 0.8 (P20:C80). Also in this trial, all the groups were in positive N balance, although it has been observed that an increase in concentrate caused a signiﬁcant increase in N intake, N balance, microbial nitrogen [57] and urea-N in the

CAB International 2010

G. Campanile, G. Neglia, D. Vecchio, R. Di Palo, B. Gasparrini and L. Zicarelli

5

Table 2 Nitrogen (N) balance in adult buffalo and in growing buffalo (culling rate)

Live weight1 Days True days2 DM intake Milk yield per lactation Milk yield per year CP intake3 Retained CP Retained N Excreted CP Excreted N Volatilization nitrogen loss (28%) Stocked nitrogen

Milking period (kg)

Dry period (kg)

Buffaloes (kg)

670 270 224 3266 2200 1237 441 70 11.2 371 59 16.61 42.7

670 95 141 1269

670 365 365 4535

102 12 1.92 90 14 3.92 10

1237 543 82 13 461 73 21 53

Culling rate (kg) 300 365 2409 289 12 1.9 277 44 12.32 31.9

1

Live weight was calculated in adult buffalo considering a 20% incidence of animals at the ﬁrst lactation and in growing buffaloes the incidence of different categories. 2 True days were calculated by consideration of incidence of milking and dry period. 3 It has been calculated considering the diets utilized in different physiological phases.

urine [58]. The last-named ﬁnding is probably the result of an inefﬁcient utilization of protein in the rumen for energy and nitrogen excess, since in these conditions proteins break down into amino acids and undergo a deamination process, leading to urea formation that is excreted in the urine [59]. Furthermore, it is known that in diet characterized by nitrogen deﬁciencies, the buffalo reduces the kidney clearance and improves the rumen nitrogen utilization [17]. In any case, the proportion P40:C60 produced the best efﬁciency of urinary purine derivatives (PD) excretion (mmol) per kg of digestible organic matter intake [56]. The study of PD excretion rate has been used as an index to predict rumen microbial protein production in various ruminants. In fact, urinary allantoin and uric acid excretion rates reﬂect the magnitude of microbial protein ﬂow into the small intestine. Liang et al. [60] reported that PD excretion rates for buffaloes were 44–57% lower than for zebu cattle under similar feeding conditions. In a more recent trial [61], four male swamp buffaloes with an initial body weight of 244+19.8 kg were used to study the recovery rate of urinary PD after they were duodenally infused with incremental amounts of purine bases (PB). Similarly to cattle, in which values of 80–89% of total PD have been reported [60–62], the main PD was allantoin (84% of total PD), followed by uric acid and they were linearly correlated with PB input, whereas no relationships were found for other PD. Negligible amounts of hypoxanthine and xanthine were recorded in this study, conﬁrming the high activity of xanthine oxidase in tissue of liver, intestinal mucosa and blood of buffaloes [63]. The relationship between daily urinary PD (i.e. allantoin, uric acid) excretion (mmol) and duodenal PB infused (mmol per day) suggested that only 12% of supplied exogenous PB were excreted in the urine [61]. These results were in agreement with those previously reported in Murrah [63]

g

and Swamp [60] buffaloes. Furthermore, these studies were recently conﬁrmed by another trial [64] performed concurrently in adult Mediterranean dry buffaloes and adult dry Friesian cows. Also these authors reported a total PD urinary excretion of 11% compared with cattle (2.711 versus 23.707 mmol/l/day in buffalo and bovine, respectively). Interestingly, a higher urine amount was reported in buffalo compared with cattle during the experimental period (10.84 versus 8.54 l/day in buffalo and cattle, respectively). The low urinary PD excretion rate recorded in buffalo species is not easily explainable. Thanh and Ørskov [65] demonstrated that differences between buffalo and bovine arise only after rumen development, since no signiﬁcant variations in purine excretion were observed during weaning. However, when the study was performed on the same animals after weaning, buffaloes excreted around 38% of the total PD excreted by cattle [65]. Cutrignelli et al. [64] investigated the presence of allantoic acid, the ﬁnal product of purine degradation in several ﬁsh species, in the urine of buffalo, but it was not detected. The same authors also evaluated differences in the saliva content of PD, but the concentration was similar in both cattle and buffalo [66]. Other authors [60–63] suggested that this could be the result of higher recycling of plasma PD than in cattle. However, when the renal and non-renal losses of plasma PD in bovines and buffaloes were investigated by examining [8-14C] uric acid as tracer [67], PD lost via non-renal routes, as a measure of recycling, for buffaloes did not differ from that of cattle, although the recovery rate of plasma tracer PD for buffaloes was much higher than 12%. Another hypothesis considers the glomerular ﬁltration rate [65]. It was proposed that it was lower in buffaloes than in cattle, leaving more time in the blood for recycling to the rumen and metabolizing by bacteria the PD. This speculation is supported by a previous study

CAB International 2010

6

Animal Science Reviews 2010

[68], in which the glomerular ﬁltration rate was lower in buffalo than in bovine, but it is in contrast with other research [61] that demonstrated a higher glomerular ﬁltration rate in buffalo. Other hypotheses concern a higher permeability from the blood to the rumen in buffaloes than cattle [65] or a lower absorption rate at the small intestine [61]. In any case, further studies are needed to clarify this aspect. The knowledge of nitrogen balance is also fundamental in order to optimize manure management. Manure composition noticeably changes as a result of several factors, such as alimentation, management, manure storage and dilution with water, either rainwater or water for washing the milking room. Nitrogen balance in buffaloes has been considered similar to bovine and hence total nitrogen produced by an adult animal of 550 kg live weight is around 85 kg/year, taking into account the losses during the storage period. However, taking into account the productive characteristics, the duration of lactation and the dry period, this value cannot be utilized for an adult buffalo. In fact, in this species, nitrogen intake is lower than that reported for the dairy cow, because of the lower DM intake and the lower protein concentration of the diet [17–69]. If the nitrogen balance is calculated in buffalo species (Table 2), the mean quantity of nitrogen eliminated for a year for an adult buffalo and successively stocked is around 53 kg/year. This value is obtained by calculating the mean DM intake and its characteristics in terms of ingested and retained nitrogen in a buffalo that produces milk for 270 days and is in the dry period for 95 days. Furthermore, the value has been corrected considering the mean fertility of a buffalo herd (which is around 75%), which affects the length of lactation and consequently the quantities of ingested and retained nitrogen. These calculations are supported by some recent research carried out on buffalo manure, which demonstrates that nitrogen concentration is lower than that reported in bovine [70]. Furthermore, in some trials performed in buffaloes characterized by different days of lactation it has been observed that the buffalo excretes around 60% of nitrogen excreted by a bovine (Campanile et al., unpublished data). Growing buffaloes (culling rate) excrete on an average 32 kg of nitrogen throughout the year (Table 1). The incidence of growing subjects on the total number of animals present in the farm has to be considered, in order to determine the correct nitrogen balance. In fact, growing animals are around 50% of the total number of animals present in the farm and, hence, they weigh upon the nitrogen balance of the farm in this percentage. In conclusion, nitrogen excretion for buffalo bred in a farm is around 44.3 kg/year. Therefore, the manure characteristics deriving from a buffalo farm would need to be revaluated in order to manage an optimal nitrogen supplementation to the agronomic cultivation performed in the breeding zones of buffalo species.

g

Conclusions The peculiar productive characteristics of buffalo species, together with its low lactiferous habit and the length of the dry period, reduce nitrogen intake throughout the productive life. The satisfying of protein requirements depends on the tropical origin of these animals and the breeding management. In this review, the adaptation and the metabolic response of this species to diets characterized by different protein content have been demonstrated. Furthermore, the basis for the application of a new nitrogen balance calculation in buffalo species has been noted and it has been demonstrated that it is very different from that reported in bovine. The application of the value estimated for dairy cattle to buffalo would negatively affect buffalo breeders for both manure management and utilization for agronomic purposes.

References 1. Cruz LC. Trends in buffalo production in Asia. Italian Journal of Animal Science 2007;6(Suppl. 2):9–24. 2. Vale WG. Effects of environment on buffalo reproduction. Italian Journal of Animal Science 2007;6(Suppl. 2):130–42. 3. Zicarelli L. Buffalo milk production world-wide. In: Proceedings of the VI World Buffalo Congress, 20–23 May 2001, Maracaibo, Zulia, Venezuela, Volume I: 2001. p. 202–30. 4. Zicarelli L. Management under different environmental condition. Buffalo Journal 1994;2:17–38. 5. Paul SS, Mandal AB, Mandal GP, Kannan A, Pathak NN. Comparative dry matter intake and nutrient utilization efﬁciency in lactating cattle and buffaloes. Journal of Science Food and Agriculture 2003;83:258–67. 6. Campanile G, Di Palo R, De Filippo C, Zicarelli L. Tempi di ingestione e di ruminazione nella bufala in funzione della distanza dal parto. In: Proceedings XII Congresso Nazionale ASPA, 23–26 June 1997, Pisa, Italy; 1997. p. 211–12. 7. Campanile G, Di Palo R, De Filippo C, Zicarelli L. Dietary characteristics and feeding behaviour in buffalo cows. In: Proceeding of the 5th World Buffalo Congress, 6–10 October 1997, Caserta, Italy. Bubalus Bubalis S.r.l., Italy; 1997. p. 367–71. 8. Tonhati H, Lima ALF, Duarte JMC, Munoz MFC. Factors affecting milk yield and milk constituents in Brazilian buffaloes. Buffalo Newsletter 2004;20:15. 9. Carlsson J, Pehrson B. The inﬂuence of the dietary balance between energy and protein on milk urea concentration. Experimental trials assessed by two different protein evaluation systems. Acta Veterinaria Scandinavica 1994;35:193–205. 10. Gonda HL, Lindberg JE. Evaluation of dietary nitrogen utilization in dairy cows based on urea concentrations in blood, urine and milk, and on urinary concentration of purine derivatives. Acta Veterinaria Scandinavica Section A, Animal Science 1994;44:236–45. 11. Baker LD, Ferguson JD, Chalupa W. Responses in urea and true protein of milk to different protein feeding schemes for dairy cows. Journal of Dairy Science 1995;78:2424–34.

CAB International 2010

G. Campanile, G. Neglia, D. Vecchio, R. Di Palo, B. Gasparrini and L. Zicarelli 12. Roseler DK, Ferguson JD, Sniffen CJ, Herrema J. Dietary protein degradability effect on plasma and milk urea nitrogen and milk non-protein nitrogen in Holstein cows. Journal of Dairy Science 1993;76:525–34. 13. Oltner R, Wiktorsson H. Urea concentrations in milk and blood as inﬂuenced by feeding various amounts of protein and energy to dairy cows. Livestock Production Science 1983;10:457–67. 14. Oltner R, Emanuelson M, Wiktorsson H. Urea concentrations in milk in relation to milk yield, live weight, lactation number and amount and composition of feed given to dairy cows. Livestock Production Science 1985;12:47–57. 15. Hoover WH, Miller TK. Carbohydrate and protein consideration in ration formulation. In: Proceedings: Large Dairy Herd Manage Conference, Cornell Cooperative Extension, Cornell University, Ithaca, NY; 1990. p. 71.

7

(Bos indicus) on a feeding regimen deﬁcient in carbohydrates. Indian Journal of Veterinary Science 1969;38:333–6. 27. Singh SP, Gupta BN. Effect of ammonia treatment of wheat straw on ﬁbre digestion and volatile fatty acids production rates in the rumen of male Murrah buffaloes. Indian Journal of Animal Nutrition 1984;1:27–30. 28. Journet M, Verite R, Vignon B. L’azote non prote´ique du lait: facteurs de variation. Le Lait 1975;543–544:212–21. 29. Butler WR. Effect of protein nutrition on ovarian and uterine physiology in dairy cattle. Journal of Dairy Science 1998;81:2533–9. 30. Elrod CC, Butler WR. Reduction of fertility and alteration of uterine pH in heifers fed excess ruminally degradable protein. Journal of Animal Science 1993;71:694–701.

16. Gustafsson AH, Palmquist DL. Diurnal variation of rumen ammonia, serum urea, and milk urea in dairy cows at high and low yields. Journal of Dairy Science 1993;76:475–84.

31. McEvoy TG, Robinson JJ, Atken RP, Findlay PA, Robertson IS. Dietary excesses of urea inﬂuence the viability and metabolism of preimplantation sheep embryos and may affect foetal growth among survivors. Animal Reproduction Science 1997;47:71–90.

17. Campanile G, De Filippo C, Di Palo R, Taccone W, Zicarelli L. Inﬂuence of dietary protein on urea levels in blood and milk of buffalo cows. Livestock Production Science 1998;55:135–43.

32. Jordan ER, Chapman TE, Holtan DW, Swanson LV. Relationship of dietary crude protein to composition of uterine secretions and blood in high-producing postpartum dairy cows. Journal of Dairy Science 1983;66:1854–62.

18. Campanile G, Di Palo R, Infascelli F, Gasparrini B, Neglia G, Zicarelli F, et al. Inﬂuence of rumen protein degradability on productive and reproductive performance in buffalo cows. Reproduction, Nutrition, Development 2004;43:557–66.

33. Garcia-Bojalil CM, Staples CR, Thatcher WW, Drost M. Protein intake and development of ovarian follicles and embryos of superovulated non-lactating dairy cows. Journal of Dairy Science 1994;77:2537–48.

19. Ndibualonji BB, Dehareng D, Godeau JM. La sousalimentation e´nerge´tique et/ou azote´e chez la vache tarie. Effect sur les acides gras volatils du rumen, quelques me´tabolites et hormones du plasma et l’excre´tion urinaire de la 3-methylhistidine. Reproduction, Nutrition, Development 1995;35:137–54.

34. Kurar CK, Mudgal VD. Protein requirement of Murrah buffaloes in the early stage of lactation. Indian Journal of Dairy Science 1980;33:443–50.

20. Campanile G. Relationship between nutrition and reproduction in buffalo. In: Proceedings of the Third Course on Biotechnology of Reproduction in Buffaloes, 6–10 October 1997, Caserta, Italy. Bubalus bubalis S.rl, Italy; 1997. p. 217–35. 21. Carruthers VR, Neil PG, Dalley DE. Effect of altering the non-structural carbohydrate ratio in a pasture diet on milk production and ruminal metabolites in cows in early and late lactation. Animal Science 1997;64:393–402.

35. Sivaiah K, Mudgal VD. Protein requirement for maintenance and milk production in buffaloes (Bubalus bubalis) in ﬁrst lactation. Indian Journal of Animal Science 1985;53:236–42. 36. Rai SN, Aggarwal SK. Effect of substitution of green fodder with ammoniated straw on nutrient utilization and milk production in Murrah buffaloes. Buffalo Journal 1991;1:51–61. 37. Verna M, Bartocci S, Amici A, Agostini M. Effect of diets with different energetic concentration on the qualitative and quantitative yield in lactating buffaloes. In: Proceeding of the International Symposium on Prospect of Buffalo Production in the Mediterranean/Middle East, 9–12 November 1992, Cairo, Egypt; 1992. p. 258–61.

22. Campanile G, Di Palo R, Esposito L, Boni R, Di Meo C. Variazioni di alcune costanti ematiche in bufale in lattazione. In: Proceedings XI Congresso Nazionale ASPA, 19–22 June 1995, Grado (GO), Italy; 1995. p. 77–8.

38. Verna M, Bartocci S, Amici A, Agostini M. Effetto di diete diverse sulle prestazioni produttive di bufale in lattazione. Agricoltura e Ricerca 1994;153:73–8.

23. Smith RH. Reviews of the progress of dairy science. Dairy Research 1969;36:313–31.

39. Ørskov ER, Grubb DA, Kay RNB. Effect of postruminal glucose or protein supplementation on milk yield and composition on Friesian cows in early lactation and negative energy balance. British Journal of Nutrition 1977;38:397–405.

24. Houpt R. Transfer of urea and ammonia to the rumen. In: Phillipson AT, editor. Physiology of Digestion and Metabolism in the Ruminant. Oriel Press Limited, Royaume-Uni, New Castle Upon Tyne; 1970. p. 119–31. 25. Bertoni G, Amici A, Lombardelli R, Bartocci S. Variations of metabolic proﬁle and hormones in blood of buffaloes, cattle and sheep males fed the same diets. In: Proceeding of the International Symposium on Prospect of Buffalo Production in the Mediterranean/Middle East, 9–12 November 1992, Cairo, Egypt; 1993. p. 345–8. 26. Langer PN, Sidhu GS, Bhatia JS. A study of the microbial population in the rumen of buffalo (Bos bubalus) and zebu

g

40. Rulquin H. Effect sur la degestion et le metabolisme des vaches laitie`res d’infusions d’acides grass volatils dans le rumen et de caseinate dans le duode´num. II. Production et digestion. Reproduction, Nutrition, Development 1982;22: 905–21. 41. Ko¨nig BA, Oldham JD, Parker DS. The effect of abomasal infusion of casein of acetate, palmitate, and glucice kinetics in cows during early lactation. British Journal of Nutrition 1984;52:319–28. 42. Oldham JD, Hart IC, Bines JA. Effect of abomasal of casein, arginine, methionine or phenylalanina on growth hormone,

CAB International 2010

8

Animal Science Reviews 2010 insulin, prolactin, tyroxine and some metabolites in blood from lactating goats. The Proceedings of the Nutrition Society 1978;37:9.

43. Oldham JD, Hart IC, Bines JA. Formaldehyde-treated proteins for dairy cows. Effects on blood hormone concentration. British Journal of Nutrition 1982;48:543–7. 44. Choung J, Chamberlain DG. Protein nutrition of dairy cows receiving grass silage diets. The effects of postruminal supplements of proteins and aminoacids. Journal of Science, Food and Agriculture 1992;60:25–30. 45. Cado´rniga C, Lopez Diaz MC. Possible modulation of adipose tissue responsiveness to catecholamines by available dietary protein in dairy cows during early lactation. Reproduction, Nutrition, Development 1995;35:241–8. 46. Kaim M, Folman J, Neumark H, Kaufmann W. The effect of protein intake and lactation number on postpartum bodyweight loss and reproductive performance of dairy cows. Animal Production 1983;37:229–35. 47. Howard HJ, Aalseth AP, Adams GD, Bush LG, McNew RW, Dawson LJ. Inﬂuence of dietary protein on reproductive performance of dairy cows. Journal of Dairy Science 1987;70:1563–71.

57. Kozloski GV, Bonnecarrere LM, Cadorin Jr RL, Reffatti MV, Perez Neto D, Limal LD. Intake and digestion by lambs of dwarf elephant grass (Pennisetum purpureum Schm. cv. Mott) hay or hay supplemented with urea and different levels of cracked corn grain. Animal Feed Science and Technology 2006;125:111–22. 58. Kaur R, Nandra KS, Garcia SC, Fulkerson WJ, Horadagoda A. Efﬁciency of utilization of different diets with contrasting forages and concentrate when fed to sheep in a discontinuous feeding pattern. Livestock Science 2008;119:77–86. 59. Nolan JV. Nitrogen Kinetics. In: Forbes FM, France F, editors. Quantitative Aspects of Ruminant Digestion and Metabolism. CAB International, Wallingford, UK; 1993. p. 123–43. 60. Liang JB, Matsumoto M, Young BA. Purine derivative excretion and ruminal microbial yield in Malaysian cattle and swamp buffalo. Animal Feed Science and Technology 1994;47:189–99. 61. Pimpa O, Liang JB, Balcells J, Jelan ZA, Abdullah N. Urinary purine derivative excretion in Swamp buffaloes after duodenal purine base infusion. Animal Feed Science and Technology 2003;104:191–9.

48. Visek WJ. Ammonia: its effect on biological system, metabolic hormone, and reproduction. Journal of Dairy Science 1984;67:481–98.

62. Verbic J, Chen XB, MacLeod NA, Ørskov ER. Excretion of purine derivatives by ruminants: effect of microbial nucleic acid infusion on purine derivative excretion by steers. Journal of Agricultural Science, Cambridge 1990;114:243–8.

49. Campanile G, Di Palo R, Esposito L, Boni R, Zicarelli L. Inﬂuenza dei fabbisogni produttivi e del clima su alcune caratteristiche dell’estro nella bufala. In: Proceedings IX Congress Nazional ASPA, 3–7 June 1991, Rome, Italy; 1991. p. 180–200.

63. Chen XB, Samaraweera L, Kyle DJ, Ørskov ER. Urinary excretion of purine derivatives and tissue xanthine oxidase (EC 1.2.3.2) activity in buffaloes (Bubalis bubalis) with special reference to difference between buffaloes and Bos taurus cattle. British Journal of Nutrition 1996;75:397–407.

50. Salama AA, Ali HM, Shokry MM. The effect of feeding different energy levels on the reproductive performance of buffalo. In: Proceedings of the 2nd World Congress on Genetics Applied to Livestock Production, 4–8 October 1982, Madrid, Spain; 1982. p. 410–3.

64. Cutrignelli MI, Piccolo G, D’Urso S, Calabro` S, Bovera F, Tudisco R, et al. Urinary excretion of purine derivatives in dry buffalo and Friesian cows. Italian Journal of Animal Science 2007;6(Suppl. 2 – Part 1):563–6.

51. Randel RD. Nutrition and post-partum re-breeding in cattle. Journal of Animal Science 1990;68:853–62. 52. Rumello G, Bergero D, Tarantola M, Mimosi A, Ladetto G. Nutrizione azotata e metaboliti derivati nel latte e nel muco cervicale della bovina lattifera. Large Animal Review 1999;5:27–30. 53. Schepers AJ, Meijer RGM. Evaluation of the utilization of dietary nitrogen by dairy cows based on urea concentration in milk. Journal of Dairy Science 1996;81:579–84. 54. Singh S, Kundu SS, Kushwaha BP, Maity SB. Dietary energy levels response on nutrient utilization, nitrogen balance and growth in Bhadawari buffalo calves. Livestock Research for Rural Development 2009;21:http://www.lrrd.org/lrrd21/8/ bhad21125.htm. 55. NRC. Nutrients Requirements of Dairy Cattle. 7th ed. National Academy of Sciences, Washington, DC; 2001. 56. Jetana T, Suthikrai W, Usawang S, Vongpipatana C, Sophon S, Liang JB. The effects of concentrate added to pineapple (Ananas comosus Linn. Mer.) waste silage in differing ratios to form complete diets, on digestion, excretion of urinary purine derivatives and blood metabolites in growing, male, Thai swamp buffaloes. Tropical Animal Health and Production 2009;41:449–59.

g

65. Thanh VTK, Ørskov ER. Causes of differences in urinary excretion of purine derivatives in buffaloes and cattle. Animal Science 2006;82:355–8. 66. Cutrignelli MI. Escrezione di derivati purinici nella bufala [PhD Thesis]. Federico II University, Naples, Italy; 1997. 67. Pimpa O, Liang JB, Balcells J, Jelan ZA, Abdullah N. Recovery rate of plasma purine derivatives in the urine of swamp buffaloes (Bubalus bubalis) with special references to differences with zebu cattle. In: Proceedings of the Second Symposium on Sustainable Utilization of Agricultural Byproducts for Animal Production, 26–27 July 2001, Chulalongkorn University, Bangkok, Thailand; 2001. p. 112–5. 68. Norton BW, Moran JB, Nolan JV. Nitrogen metabolism in Brahman cross, buffaloes, Benteng and Shorthorn steers fed on low quality roughages. Australian Journal of Agricultural Research 1979;30:341–51. 69. Campanile G. Nutrition and milk production in dairy buffalo. In: Proceedings of the III Simposio bu`falos de las Ame´ricas and 2nd Buffalo Symposium of the Europe and Americas, 6–8 September 2006, Medelli`n, Columbia; 2006. p. 132–41. 70. St-Pierre NR, Thraen CS. Animal grouping strategies, sources of variation, and economic factors affecting nutrient balance on dairy farms. Journal of Animal Science 1999;77(Suppl 2):72–83.

CAB International 2010

Animal Science Reviews 2010

Expression QTL and their applications in genetic improvement of farm animals K. Wimmers*, E. Murani and S. Ponsuksili Address: Research Institute for the Biology of Farm Animals (FBN Dummerstorf), Wilhelm-Stahl-Allee 2, 18196 Dummerstorf, Germany. *Correspondence: K. Wimmers. Email: [email protected] Received: Accepted:

24 May 2009 6 November 2009

Abstract The integration of approaches based on genetic linkage maps and those based on function-driven holistic expression proﬁling, termed ‘genetical genomics’, has great potential for elucidating the expression and inheritance of complex traits. This review discusses the status and potential applications of ‘genetical genomics’ and how genomic/genetic, transcriptomic and phenotypic information can be used to promote the identiﬁcation of functional positional candidate genes for traits important in farm animal breeding schemes. Linkage and/or genome-wide association analyses for traits that are relevant in animal breeding provide information about the positional candidacy of a particular gene. Analysing the relationship of the expression levels of genes and these traits adds information about the functional role of a gene and gives insight into functional pathways involved in the conversion of the traits. Mapping of expression quantitative trait loci (eQTL) enables the display of regulatory networks and localization of genomic variations affecting the mRNA expression of a gene either within the genes itself (cis) or distant from the gene (trans). The key advantage of eQTL mapping is that it connects variation at the level of RNA expression to variation at the level of DNA. Only the latter provides versatile tools for breeding, whereas the ﬁrst reveals information on the biology of a trait and directs research to new candidate genes. Keywords: QTL, Linkage, Gene expression, Microarray, Genetical genomics Review Methodology: We searched the following databases: ISI Web of Knowledge, PubMed and Google Scholar (keyword search terms used: ‘genetical genomics’, ‘eQTL’). In addition, we used the references from the articles obtained by this method to check for additional relevant material.

Prerequisites and Rationale of ‘Genetical Genomics’ Researchers have used different strategies to detect genes controlling quantitative traits, quantitative trait loci (QTL); typically complex phenotypic characteristics (classical traits) are considered that vary in degree and can be attributed to effects of many genes (subsequently termed pheneQTL=pQTL). The most general approach to the identiﬁcation of genes controlling the ‘ﬁnal phenotype’ is the positional mapping by which the inheritance of a trait is compared with that of a large number of markers to locate the region(s) in which the gene(s) responsible for the trait lie. QTL analyses are in general hypothesis-free

g

approaches that allow researchers to address any variable heritable trait without knowledge of the physiology of that trait. Most of the successful approaches used to detect pQTL in pigs and chicken were based on genome scans performed in experimental F2 or backcross populations of divergent breeds, whereas in cattle mostly advantage was taken of the population structure by using existing paternal and grandpaternal pedigrees mainly within breeds. The ﬁrst pQTL detected in livestock by this approach were reported by Andersson and colleagues in 1994 [1]. Subsequently, linkage studies identiﬁed pQTL regions for various traits in livestock; the actual status of pQTL mapping in pig, cattle and chicken can be reviewed at public accessible internet databases (AnimalQTLdb;

CAB International 2009

10

Animal Science Reviews 2010

http://www.animalgenome.org/QTLdb/) [2]. QTL regions are generally large and contain many putative causal genes. Thus, pQTL analyses fail to identify anything more than large genomic segments. Subsequent experiments including ﬁne mapping and positional cloning have to be conducted. In fact, so far the underlying polymorphisms (quantitative trait nucleotides, QTNs) were identiﬁed by positional cloning and veriﬁed in livestock species for only a few pQTL: the Protein Kinase, AMP-activated, gamma 3 non-catalytic subunit (PRKAG3) locus and the Insulin-like growth factor 2 (IGF2) gene in swine [3–6], the DGAT1 (diacylglycerol O-acyltransferase homolog 1) gene in cattle [7] and the MSTN (myostatin; GDF-8, growth/ differentiation factor-8) (myostatin) gene in sheep [8]. That means that even after the successful identiﬁcation of genomic regions containing the pQTL that are responsible for the variation of ﬁnal phenotypes, high monetary and time expenditure are still necessary. The novel and powerful gene expression proﬁling techniques have brought a new level of hope to unravelling the secrets of complex classical traits with a polygenic inheritance. The introduction of microarray technology enabled expression proﬁling of thousands of genes simultaneously [9]. Functional genomics approaches can be used to generate information about gene function, as well as data on genetic interactions, not only among and between gene complexes but also in response to environmental stimuli. In order to successfully apply holistic expression proﬁling, an educated guess of the physiology of a trait is required in order to select the relevant tissues/cells at relevant developmental stages for the analysis. The outcome of bioinformatic analyses of microarray experiments are functional networks of genes reﬂecting interactions at the level of DNA, RNA or proteins. The results enable the derivation of new hypotheses on the nature of a trait. In livestock, expression microarrays were used regarding development and reproduction traits, growth, immune response and host–pathogen interaction [10–17]. One of the major challenges and opportunities is the combination of genetic linkage data of DNA and information of the abundance of RNAs and proteins to dissect complex classical traits or economic trait loci. These genomic data provide an understanding of metabolic, regulatory and developmental pathways. Jansen and Nap [18] proposed the merging of genomics and genetics into ‘genetical genomics’. This involves expression proﬁling and marker-based ﬁngerprinting of each individual of a segregating population, and exploits all the statistical tools used in the analysis of QTL. Doerge [19] proposed the use of current QTL methods and software to identify loci that affect the level of expression of a particular gene or a set of genes. That means that the quantitative data of gene expression analyses are handled as phenotypic data (here termed ‘expression phenotype’) and their inheritance is compared with that of large numbers of markers applying methods of QTL analyses. QTL analysis of expression

g

levels of a gene identiﬁes genomic regions that are likely to contain at least one causal gene with regulatory effect on the expression level, termed expression QTL (eQTL) [20–22]. Thus, the term ‘eQTL’ refers to the loci that control the abundance of particular transcripts, the expression phenotype [23]. However, in order to link the eQTL to the genetic background of a classical phenotypic trait of interest, it is necessary to establish a relationship between the variation of that classical phenotypic trait and its corresponding pQTL position on the one hand, and either the expression level of that particular transcript or the mapping position of the eQTL on the other hand.

Mode of Regulation of Expression and eQTL Mapping Combining microarray data with QTL linkage studies, by simply mapping genes that are differentially expressed depending on a classical trait, offers new options of understanding the biology at a global level and the genetic factors affecting the classical trait of interest, and it facilitates focusing on the most relevant candidate genes in each region shown to contain pQTL for that classical trait [20, 24, 25]. This projection of the position of traitassociated regulated genes onto QTL maps might be taken as ‘genetical genomics in the broader sense’. Differential expression of the gene in the pQTL region suggests a polymorphism in the gene itself (or close by) contributes to both variation of the transcript level and the classical trait of interest. These genes are functional positional candidate genes with cis-acting regulation. However, regulated genes depending on classical traits and mapping in the pQTL region might also be controlled by a locus mapping elsewhere, i.e. they might be under trans control. This implies that the genes are not causal or that they are not the only causal genes in the pQTL region. The differential expression of the genes might be an effect rather than a cause of variation. Mapping of eQTL, i.e. the identiﬁcation of regions harbouring loci that affect the expression of a gene by linkage analysis, allows us to get further insight into these possible scenarios. eQTL mapping enables the addressing of cis-acting control of the transcription of genes within the pQTL for a classical trait of interest but also the studying of gene interactions via trans-acting regulation, where either the regulated gene or the regulating gene (the eQTL) may be located in the pQTL. Polymorphisms associated with trait variation can either be in coding, non-coding or regulatory regions; functional polymorphisms will be likely to be located in coding or regulatory regions. Variants that affect gene expression (eQTL) may have a substantial impact on quantitative classical traits, whereas qualitative traits, including hereditary disease, may more often depend on structural variation [26, 27]. A literature survey revealed

CAB International 2009

K. Wimmers, E. Murani and S. Ponsuksili

that of 107 genes with experimentally veriﬁed functional cis-regulatory polymorphisms, 82% had proven effects on phenotypes [28]. In humans, more than 100 cis-regulatory polymorphisms were associated with phenotypic variation of classical traits [29]. Also, in livestock, current ﬁndings suggest high importance of cis-regulatory polymorphisms: two of the causal single nucleotide polymorphisms (SNPs) identiﬁed in livestock are regulatory SNPs, an SNP in the 0 3 untranslated region (UTR) of the ovine GDF8 gene, affecting a mRNA-binding site and thus modulating transcription and translation of GDF8 [8] and an SNP in intron 3 of the porcine IGF2 gene, affecting its transcription [6]. The fact that in most genome-wide association studies SNPs were detected in non-coding regions supports the view that polymorphisms in regulatory regions are important [30]. eQTL studies revealed that about one-third are cis-acting eQTL [31, 32]. Pomp et al. [32] hypothesize that the majority of genes affecting complex classical traits are acting in trans. The authors describe four modes of relationships between the dependent variation of gene expression and the independent position of the expression-modulating gene including (1) cisacting regulation; (2) regulation in trans mode represents cases where the expressed genes is regulated by an unlinked eQTL; but there are also cases where (3) a single eQTL regulates many unlinked expressed genes (pleiotropic effect of an eQTL, master transcript modulators); (4) expressed genes regulated by a common eQTL might be organized in closely linked groups: so-called ‘expression neighbourhoods’ [32, 33]. The identiﬁcation of the mode of action of an eQTL will provide clues as to the nature of the candidate gene underlying the trait of interest. The identiﬁcation of trans-acting eQTL provides genetical evidence for gene–gene interaction and epistasis effects on a trait of interest. The variation of complex classical traits is the result of many genes of largely unknown function combined with environmental inﬂuences. The extent to which epistasis is involved in regulating complex traits is not known, but estimates of statistical epistasis in mouse and chicken, i.e. via modelling of effects of interaction of QTL effects, indicated that up to 81% of phenotypic variation can be explained by epistatic QTL effects [34–38]. Mapping of eQTL detects interaction among DNA sequence variations that gives rise to a particular ﬁnal phenotype (genetical epistasis), and for genetic improvement, trans-acting eQTL direct researchers to the loci affecting the ﬁnal phenotype. A trans eQTL affecting the expression level of several genes (master controller) potentially affects several (correlated) traits; the eQTL has pleiotropic effects. This is relevant in attempts to select for trans-acting eQTL [39]. Genotypes environment interactions are responsible for variation in phenotypic differences among different genotypes, depending on environmental factors such as climate, feeding, housing and various treatments. The characterization of the relationship of expression level of a target

g

11

gene, ﬁnal phenotype and genotype, i.e. either cis- or transacting QTL, at various environments will enable selecting the right genotype for particular environments [40].

Feasibility of the eQTL Approach and Model Experiments It is evident that eQTL analyses are based on heritability of transcription. Expression studies in fact showed that inter-individual variability in gene expression is genetic, implying heritable variation in cis-acting sequences, transacting factors or both [41, 42]. Heritability is assessed by the resemblance among relatives and from parent– offspring regression between generations. The integration of QTL and expression analyses is based on the assumption that the functional polymorphism that affects the gene expression level also affects the classical trait in question. Such a ﬁnding will further increase the potential association of the gene with the trait in question. However, ﬁrst reports of eQTL studies did not primarily relate to any classical phenotypic trait but demonstrate the feasibility of this ‘genetical genomics’ approach. Brem et al. [43] demonstrated that among some 6000 genes analysed for 570 eQTL could be detected, 33% of which were cisacting. In human, lymphoblastoid cell lines obtained from 167 individuals of 15 Centre d’Etude du Polymorphisme Humaine (CEPH) families of 23 499 genes surveyed by microarray analysis, 2340 were examined to estimate eQTL based on the comparison of the expression level of an individual sample with the expression level obtained from a pool of all samples [44]. Monks et al. [44] found heritable variation in 762 transcripts, i.e. transcripts showing a variance of expression levels caused by polygenic effects as revealed by maximum-likelihood estimation. The median heritability for the 762 transcripts was 0.3. Linkage study was performed using 346 markers at 11 cM median marker intervals. Stringent signiﬁcance thresholds (pointwise signiﬁcance level of 0.000005) revealed 33 eQTL accounting for 50% of the variance of the expression phenotype, with 75% of the QTL having heritabilities of >0.76; [44]. Using human lymphoblastoid cell lines of members of 14 CEPH families (of which 7 were the same as used by Monks et al. [44]) for micorarray expression proﬁling of about 8500 genes, Morley et al. [45] found 142 eQTL in a genome scan performed with 2756 autosomal SNP markers (at P < 0.00000004). The authors report on 2 eQTL regions that control expression of several genes. About 50% of the genes controlled by these master regulators could be assigned to regulated clusters showing correlated expression exceeding the level expected by chance. This reﬂects a complex regulatory network where master transcriptional regulators control expression level of genes with similar expression patterns that might reside close together certain genomic regions. Interestingly, these kinds of hotspots were also detected in other species [21, 43].

CAB International 2009

12

Animal Science Reviews 2010

Monks et al. [44] also found six regions with statistically signiﬁcant enrichment of eQTL; however, the authors argue that because these hotspots correspond to linkage for highly variable and highly similar major histocompatibility complex (HLA) genes, this observation of hotspots might be an artefact of less reliable estimations of expression levels. Taking into account the particular phenotypic characteristics of the founders of the pedigrees used in eQTL studies enables us to relate expression levels, eQTL and phenotype. In mice, eQTL for obesity-related genes, diabetes, drug-abuse liability or allergic response were identiﬁed [21, 46–48]. The experiments demonstrate the feasibility of eQTL mapping: however, in particular, when comparing the two human studies by Monks et al. [44] and Morley et al. [45] it becomes obvious that because of differences in analytical and statistical methods, the outcome may differ in terms of number of eQTL, mode of eQTL and eQTL effects that are related to the power of the study [49]. Theoretical examinations of the eQTL experiments show up the limits of eQTL studies through experimental constraints (including speciﬁcity, sensitivity and efﬁciency of microarray experiments, dynamic range of detectable expression differences, size of intervals used to deﬁne cis eQTL) and insufﬁcient statistical power (mainly depending on sample number), given the complex nature of genetic regulation of transcription levels [23, 49]. The estimation of heritability of transcript variation is difﬁcult. Studies in model organisms and humans indicate that non-additive effects modulate transcript levels including epistatic and parent-of-origin effects [41]. At more stringent statistical thresholds, the proportion of cis-acting eQTL usually becomes higher in many studies, indicating a higher impact of cis eQTL that are consequently more readily detectable. This might be because effects of SNPs within the expressed genes themselves are easier to detect than second-order effects [23, 32, 41, 44]. However, a higher number of trans eQTL might be expected to be detected by chance than cis eQTL, because for the trans eQTL, the whole genome represents a target except one position, whereas for cis eQTL only exactly one position is the target.

Applications of the eQTL Approach The most obvious application of ‘genetical genomics’ is to scale down the number of candidate genes for a classical trait of interest. ‘Genetical genomics in the broader sense’ of mapping differentially expressed genes onto the pQTL map, as well as ‘genetical genomics in the narrow sense’ of mapping eQTL, enables a reduction in the number of positional candidate genes for a pQTL when seeking positional functional candidate genes. Arbilly et al. [25] reported that the integrated application of pQTL mapping and expression analysis allowed a narrowing down of a pQTL region for susceptibility to neuropathic pain of 12 Mb that contains some 150 genes by assigning

g

segregating SNPs to coding and regulatory regions of genes, addressing 34 polymorphic positional candidate genes, and ﬁnally prioritizing two positional functional candidate genes in mice. In chickens, this approach is promoted and has been exemplarily performed to identify genes related to Marek’s disease resistance [24]. Essentially, it has become common in expression analyses to refer to the position of the differentially expressed genes relative to existing pQTL for the trait of interest. With the aim of detecting genes affecting immune responsiveness De Koning et al. [50, 51] propose to extend this approach by obtaining microarray expression proﬁles of animals of a segregating F2 population after immunological stimulation and by genotyping the same animals in order to estimate eQTL and pQTL for immune-response-related traits. In order to increase the power and reduce the number of the analyses for eQTL analysis, individuals homozygous for the respective pQTL should be used, whereas for ﬁne mapping of the pQTL, recombinant individuals of the F2 should be selected. The use of animals that are homozygous for a putative pQTL for the estimation of eQTL increases the power to detect additive effects and the focus on just two contrasting groups and on predicted regions only allows simpler genetic tests and less multiple testing. The combination of the eQTL data and ﬁne mapping data will allow the accurate assigning of a candidate gene as a cis- or transregulated locus [51]. Fine mapping based on a subset of animals showing recombination between the initial identiﬁed ﬂanking markers of the pQTL enables the precise detection of cis-acting eQTL, whereas to detect of transacting eQTL, additional expression analysis using selected animals of the ﬁne mapping study may be required to resolve eQTL that correlate with the pQTL and those that are linked effects. Power analyses showed that the size of the resource populations required to detect eQTL depends on the number of QTL regions taken into account, the size of the effects of QTL to be detected, and the resolution of the genetic map. The authors state that the number of individuals that need to be genotyped using their proposed strategy is much lower than the size of the population [51]. Cardoso et al. [52] compared different strategies to select subsets of animals used for eQTL detection taking into account genetic and phenotypic dissimilarity, their combination and recombination events. Three samples sizes (80, 160 and 240 individuals of 510 F2 animals) were simulated. In terms of sensitivity and precision of estimates of QTL effects, selection of animals based on genetic dissimilarity and phenotypic extremes within genotype methods was preferential, in particular, when a low numbers of animals used [52]. There are numerous publications highlighting the possible merits of ‘genetical genomics’ applications in livestock breeding [32, 39, 40, 53, 54] and/or addressing issues of data evaluation and experimental design in silico [50, 51, 55–57]. However, there is a lack of experimental data on eQTL mapping obtained in non-model animal

CAB International 2009

K. Wimmers, E. Murani and S. Ponsuksili Genome: gene and marker genotypes

Transcriptome: expression levels

eQTL

cis

13

trans

Yes

No

Yes

Gene regulation not relevant to the classical trait

pQTL

Trait-depend. expression

Within pQTL

Neither gene nor Gene within pQTL eQTL within pQTL eQTL within PQTL

Outside pQTL

Likely a causal gene with Member of a relevant network a regulatory SNP

Member of a relevant network; link to a regulatory superior gene

The regulatory superior gene is a functional positional candidate gene

Final phenotype: records of traits

Figure 1. Outline of integration of data from eQTL and pQTL analyses and expression proﬁling towards identiﬁcation of candidate genes for complex classical traits facilitating the detection of markers for genetic improvement of livestock.

species. In lake whiteﬁsh, a valuable commercial freshwater ﬁsh, eQTL were mapped for white muscle and brain transcripts quantiﬁed in 66 and 57 individuals of the ﬁrst backcross generation of the F1 hybrid of dwarf and normal whiteﬁshdwarf whiteﬁsh [58, 59]. eQTL of genes that were differentially expressed between the founder lines points to genes involved in the divergence of these populations. Regions were identiﬁed that harbour genes controlling the expression of many loci of functional groups relevant to the phenotypic divergence of the dwarf and normal whiteﬁsh [58, 59]. The co-localization of eQTL and pQTL for growth enable researchers to address candidate genes for these classical traits, although positional data for the transcripts are lacking and thus a discrimination of cis- and trans-acting eQTL was not possible [59]. Recently, the ﬁrst study in livestock on eQTL mapping of trait-dependent expressed genes obtained from microarray experiments was reported in the pig [60]. The study aimed to ﬁnd relevant biological pathways and candidate genes for water-holding capacity of meat. Therefore, correlation between gene expression of 11 453 muscle expressed genes (out of 23 256 loci represented on the Affymetrix Porcine Genome Array) and drip loss phenotypic traits was calculated. A total of 1279 genes were signiﬁcantly associated with the trait (r=10.37–0.67l; P0.001; q0.004). Subsequently, eQTL were identiﬁed for transcripts that showed traitcorrelated expression, providing information about the

g

genomic location of putative regulatory loci relevant for the trait of interest. In total, 897 eQTL were identiﬁed with 104 eQTL mapping in pQTL target regions for water-holding capacity identiﬁed in the same population. Among the 104 eQTL co-localized with pQTL 96 represent trans-acting eQTL and 8 were cis-acting. Of the 793 eQTL mapping away from of any pQTL, 119 eQTL were trans-acting eQTL of 66 genes mapping to the pQTL target regions. These three groups of eQTL, in particular, provide additional insight into the control of the trait of interest: (1) The eight genes categorized as positional functional candidate genes, which map together with their corresponding eQTL in cis, are genes showing variation with impact on the trait of interest and its expression level, indicating that the nature of the variation is likely a polymorphism in regulatory regions of the gene. (2) The 96 eQTL detected within pQTL for which the functional candidate genes themselves are located elsewhere, i.e. trans control, are inﬂuenced by hierarchically superior genes located in the pQTL that actually represent candidate genes (positional candidate genes). (3) For 66 functional positional candidate genes under trans control, it can be speculated that the nature of variation affecting the ﬁnal phenotype of water-holding capacity is differential expression owing to polymorphisms in hierarchically superior genes and different responsiveness of the candidate genes to regulatory mechanisms. The study revealed a network of genes relevant to the trait of interest representing additive and pleiotropic as well as

CAB International 2009

14

Animal Science Reviews 2010

non-additive epistatic effects on the trait. cis-acting eQTL serves as an important new resource for the identiﬁcation of positional candidates in QTL studies.

4. Ciobanu D, Bastiaansen J, Malek M, Helm J, Woollard J, Plastow G, et al. Evidence for new alleles in the protein kinase adenosine monophosphate-activated gamma(3)-subunit gene associated with low glycogen content in pig skeletal muscle and improved meat quality. Genetics 2001;159:1151–62.

Summary and Implications

5. Georges M, Andersson L. Positional identiﬁcation of structural and regulatory quantitative trait nucleotides in domestic animal species. Cold Spring Harbor Symposia on Quantitative Biology 2003;68:179–87.

The beneﬁt of ‘genetical genomics’ approaches to identify functional positional candidate genes and to elucidate functional networks of genes relevant for complex classical traits has been demonstrated experimentally in humans, model organisms and livestock species. A key advantage of eQTL mapping is that it connects variation at the level of RNA to variation at the level of DNA. Variation at the transcript level, especially in holistic approaches, is conclusive towards identiﬁcation of relevant functional networks and for deriving hypothesis on the control of expression and inheritance of a trait of interest. However, measures at the level of RNA are less promising tools for breeding than genotyping at the DNA level. The application of expression microarray data by breeders including the estimation of breeding values for gene expression in ‘expression-assisted selection’ schemes [39, 40] is hampered by the costs and the high situation-speciﬁcity (tissue-, age-, sex-, housing-, feedingand treatment-depending) of expression analyses. Variation at the DNA level is easily detectable in any tissue at any developmental stage and under endogenous and exogenous conditions. Furthermore, variation at the DNA level is inherited. For genetic improvement, traitassociated markers or preferentially causal markers at the level of DNA are versatile tools. Integration of records of complex classical traits, genome scans and holistic transcriptome analyses for the identiﬁcation of pQTL and eQTL will largely contribute to the detection of such DNA markers for genetic improvement (Figure 1). Current applications of DNA marker genotype in best linear unbiased prediction (BLUP) animal models will be extended using information on pQTL and eQTL. The availability of SNP microarrays for livestock species for genome-wide association studies improves the mapping resolution for pQTL and eQTL, leading to more precise discrimination of cis and trans eQTL and further scaling down the length of lists of functional positional candidate genes.

References 1. Andersson L, Haley CS, Ellegren H, Knott SA, Johansson M, Andersson K, et al. Genetic mapping of quantitative trait loci for growth and fatness in pigs. Science 1994;263:1771–4. 2. Hu Z, Fritz ER, Reecy JM. AnimalQTLdb: a livestock QTL database tool set for positional QTL information mining and beyond. Nucleic Acids Research 2007;35:604–9. 3. Milan D, Jeon JT, Looft C, Amarger V, Robic A, Thelander M, et al. A mutation in PRKAG3 associated with excess glycogen content in pig skeletal muscle. Science 2000;288:1248–51.

g

6. Van Laere AS, Nguyen M, Braunschweig M, Nezer C, Collette C, Moreau L, et al. A regulatory mutation in IGF2 causes a major QTL effect on muscle growth in the pig. Nature 2003;425:832–6. 7. Grisart B, Coppieters W, Farnir F, Karim L, Ford C, Berzi P, et al. Positional candidate cloning of a QTL in dairy cattle: identiﬁcation of a missense mutation in the bovine DGAT1 gene with major effect on milk yield and composition. Genome Research 2002;12:222–31. 8. Clop A, Marcq F, Takeda H, Pirottin D, Tordoir X, Bibe´ B, et al. A mutation creating a potential illegitimate microRNA target site in the myostatin gene affects muscularity in sheep. Nature Genetics 2006;38:813–8. 9. Schena M, Shalon D, Davis RW, Brown PO. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 1995;270:467–70. 10. Bai Q, McGillivray C, da Costa N, Dornan S, Evans G, Stear MJ, et al. Development of a porcine skeletal muscle cDNA microarray: analysis of differential transcript expression in phenotypically distinct muscles. BMC Genomics 2003;4:8. 11. Moser RJ, Reverter A, Kerr CA, Beh KJ, Lehnert SA. A mixedmodel approach for the analysis of cDNA microarray gene expression data from extreme-performing pigs after infection with Actinobacillus pleuropneumoniae. Journal of Animal Science 2004;82:1261–71. 12. Ponsuksili S, Murani E, Schellander K, Schwerin M, Wimmers K. Identiﬁcation of functional candidate genes for body composition by expression analyses and evidencing impact by association analysis and mapping. Biochimica et Biophysica Acta – Gene Structure and Expression 2005;1730:31–40. 13. Ponsuksili S, Murani E, Walz C, Schwerin M, Wimmers K. Pre- and postnatal hepatic gene expression proﬁles of two pig breeds differing in body composition: insight into pathways of metabolic regulation. Physiological Genomics 2007;29:267–79. 14. Ponsuksili S, Murani E, Phatsara C, Jonas E, Walz C, Schwerin M, et al. Expression proﬁling of muscle reveals transcripts differentially expressed in muscle that affect water-holding capacity of pork. Journal of Agricultural and Food Chemistry 2008;56:10311–7. 15. Band MR, Olmstead C, Everts RE, Liu ZL, Lewin HA. A 3800 gene microarray for cattle functional genomics: comparison of gene expression in spleen, placenta, and brain. Animal Biotechnology 2002;13:163–72. 16. Suchyta SP, Sipkovsky S, Kruska R, Jeffers A, McNulty A, Coussens MJ, et al. Development and testing of a high-density cDNA microarray resource for cattle. Physiological Genomics 2003;15:158–64. 17. Cogburn LA, Wang X, Carre W, Rejto L, Porter TE, Aggrey SE, et al. Systems-wide chicken DNA microarrays, gene

CAB International 2009

K. Wimmers, E. Murani and S. Ponsuksili expression proﬁling, and discovery of functional genes. Poultry Science 2003;82:939–51. 18. Jansen RC, Nap JP. Genetical genomics: the added value from segregation. Trends Genet 2001;17:388–91. 19. Doerge RW. Mapping and analysis of quantitative trait loci in experimental populations. Nature Reviews Genetics 2002;3:43–52.

15

analysis reveals complex genetic determinants of circadian behavior in mice. Genome Research 2001;11:959–80. 37. Sugiyama F, Churchill GA, Higgins DC, Johns C, Makaritsis KP, Gavras H, et al. Concordance of murine quantitative trait loci for salt-induced hypertension with rat and human loci. Genomics 2001;71:70–7.

20. Wayne ML, McIntyre M. Combining mapping and arraying: An approach to candidate gene identiﬁcation. Proceeding of the National Academy of Sciences, USA 2002;99:14903–6.

38. Carlborg O, Kerje S, Schu¨tz K, Jacobsson L, Jensen P, Andersson L. A global search reveals epistatic interaction between QTL for early growth in the chicken. Genome Research 2003;13(3):413–21.

21. Schadt EE, Monks SA, Drake TA, Lusis AJ, Che N, Colinayo V, et al. Genetics of gene expression surveyed in maize, mouse and man. Nature 2003;422:297–301.

39. Walsh B, Henderson D. Microarrays and beyond: What potential do current and future genomics tools have for breeders? Journal of Animal Science 2004;82:292–9.

22. Tabakoff B, Bhave SV, Hoffman PL. Selective breeding, quantitative trait locus analysis, and gene arrays identify candidate genes for complex drug-related behaviors. Journal of Neuroscience 2003;23:4491–8.

40 Kadarmideen HN, von Rohr P, Janss LLG. From genetical genomics to systems genetics: potential applications in quantitative genomics and animal breeding. Mammalian Genome 2006;17:548–64.

23. Williams RBH, Chan EKF, Cowley MJ, Little PFR. The inﬂuence of genetic variation on gene expression. Genome Research 2007;17:1707–16.

41. Gibson G, Weir B. The quantitative genetics of transcription. Trends in Genetics 2005;21:616–23.

24. Liu H-C, Cheng HH, Tirunagaru V, Sofer L, Burnside J. A strategy to identify positional candidate genes conferring Marek’s disease resistance by integrating DNA microarrays and genetic mapping. Animal Genetics 2001;32:351–9. 25. Arbilly M, Pisante A, Devor M, Darvasi A. An integrative approach for the identiﬁcation of quantitative trait loci. Animal Genetics 2006;37(Suppl.1):7–9. 26. Farrall M. Quantitative genetic variation: a post-modern view. Human Molecular Genetics 2004;13(1):R1–7. 27. Yan H, Zhou W. Allelic variations in gene expression. Current Opinion in Oncology 2004;16:39–43. 28. Rockman MV, Wray GA. Abundant raw material for cisregulatory evolution in humans. Molecular Biology and Evolution 2002;19:1991–2004. 29. Wray GA. The evolutionary signiﬁcance of cis-regulatory mutations. Nature Reviews Genetics 2007;8:206–16. 30. McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JP, et al. Genome-wide association studies for complex traits: consensus, uncertainty and challenges. Nature Reviews Genetics 2008;9:356–69. 31. Doss S, Schadt EE, Drake TA, Aldons JL. Cis-acting expression quantitative trait loci in mice. Genome Research 2005;15:681–91. 32. Pomp D, Allan MF, Wesolowsky SR. Quantitative genomics: exploring the genetic architecture of complex trait predisposition. Journal of Animal Science 2004;82:300–12. 33. Oliver B, Parisi M, Clark D. Gene expression neighbourhoods. Journal of Biology 2002;1:4. 34. Brockmann GA, Kratzsch J, Haley CS, Renne U, Schwerin M, Karle S. Single QTL effects, epistasis, and pleiotropy account for two-thirds of the phenotypic F(2) variance of growth and obesity in DU6iDBA/2 mice. Genome Research 2000;10:1941–57. 35. Kim JH, Sen S, Avery CS, Simpson E, Chandler P, Nishina PM, et al. Genetic analysis of a new mouse model for non-insulin-dependent diabetes. Genomics 2001;74:273–86. 36. Shimomura K, Low-Zeddies SS, King DP, Steeves TD, Whiteley A, Kushla J, et al. Genome-wide epistatic interaction

g

42. Stamatoyannopoulos JA. The genomics of gene expression. Genomics 2004;84:449–57. 43. Brem RB, Yvert G, Clinton R, Kruglyak L. Genetic dissection of transcriptional regulation in budding yeast. Science 2002;296:752–5. 44. Monks SA, Leonardson A, Zhu H, Cundiff P, Pietrusiak P, Edwards S, et al. Genetic inheritance of gene expression in human cell lines. American Journal of Human Genetics 2004;75:1094–105. 45. Morley M, Molony CM, Weber TM, Devlin JL, Ewnes KG, Spielman RS, et al. Genetic analysis of genome-wide variation in human gene expression. Nature 2004;430:733–4. 46. Karp CL, Grupe A, Schadt E, Ewart SL, Keane-Moore M, Cuomo PJ, et al. Identiﬁcation of complement factor 5 as a susceptibility locus for experimental allergic asthma. Nature Immunology 2000;1:221–6. 47. Eaves IA, Wicker LS, Ghandour G, Lyons PA, Peterson LB, Todd JA, et al. Combining mouse congenic strains and microarray gene expression analyses to study a complex trait: the NOD model of type 1 diabetes. Genome Research 2002;12:232–43. 48. Palmer AA, Verbitsky M, Suresh R, Kamens HM, Reed CL, Li N, et al. Gene expression differences in mice divergently selected for methamphetamine sensitivity. Mammalian Genome 2005;16:291–305. 49. De Koning D-J, Haley C. Genetical genomics in human and model organisms. Trends in Genetics 2005;21:377–81. 50. De Koning D-J, Carlborg O¨, Haley CS. The genetic dissection of immune response using gene-expression studies and genome mapping. Veterinary Immunology and Immunopathology 2005;105:343–52. 51. De Koning D-J, Cabrera CP, Haley CS. Genetical genomics: combining gene expression with marker genotypes in poultry. Poultry Science 2007;86:1501–9. 52. Cardoso FF, Rosa GJ, Steibel JP, Ernst CW, Bates RO, Tempelman RJ. Selective transcriptional proﬁling and data analysis strategies for expression quantitative trait loci mapping in outbred F2 populations. Genetics 2008;180:1679–90.

CAB International 2009

16

Animal Science Reviews 2010

53. Kadarmideen HN, Reverter A. Combined genetic, genomic and transcriptomic methods in the analysis of animal traits. CAB Reviews: Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources 2007;2:16. 54. Wimmers K, Ngu NT, Murani E, Schellander K, Ponsuksili S. Linkage and expression analysis to elucidate the genetic background of muscle structure and meat quality in the pig. Archives of Animal Breeding 2006;49 (Special Issue):116–25. 55. Perez-Enciso M, Toro MA, Tenenhaus M, Gianoly D. Combining gene expression and molecular marker information for mapping complex trait genes: a simulation study. Genetics 2003;164:1597–606. 56. Perez-Enciso M. In silico study of transcriptome genetic variation in outbred populations. Genetics 2004;166:547–54.

g

57. Perez-Enciso M, Quevedo JR, Bahamonde A. Genetical genomics: use all data. BMC Genomics 2007;8:69. 58. Derome N, Bougas B, Rogers SM, Whiteley AR, Labbe A, Laroche J, et al. Pervasive sex-linked effects on transcription regulation as revealed by expression quantitative trait loci mapping in Lake Whiteﬁsh (Coregonus sp., Salmonidae). Genetics 2008;179:1903–17. 59. Whiteley AR, Derome N, Rogers SM, St-Cyr J, Laroche J, Labbe A, et al. The phenomics and expression quantitative trait locus mapping of brain transcriptomes regulating adaptive divergence in lake whiteﬁsh species pairs (Coregonus sp.). Genetics 2008;180:147–64. 60. Ponsuksili S, Jonas E, Murani E, Phatsara C, Srikanchai T, Walz C, et al. Trait correlated expression combined with expression QTL analysis reveals biological pathways and candidate genes affecting water holding capacity of muscle. BMC Genomics 2008;9:367.

CAB International 2009

Animal Science Reviews 2010

A brief history of mammalian coat colour genetics J.J. Lauvergne* Address: 147 C/3 avenue J.B. Cle´ment, 92 140, Clamart, France. *Correspondence: Email: [email protected] Received: Accepted:

19 August 2009 6 November 2009

Abstract The study of coat colour genes of mammals can be traced back to the rediscovery of Mendel’s laws in 1900. A phenotypic homology of coat colour loci based on a phenotypic similarity of allelic series has been demonstrated and is known as comparative coat colour genetics. This homology has been veriﬁed by molecular genetics. Loci controlling coat colour in mammals are numerous: more than 280 have been discovered in genetic analysis of the mouse, the leading laboratory species. Loci for which mutants have been kept in breeding species (mainly farm animals) are less numerous because only viable alleles have been saved. The colour loci are highly pleiotropic, being involved not only in coat colour pigmentation but also in many other biological processes. The comparative study of DNA chains of numerous variants in several species may help to follow evolution at the locus level in mammals. The colour mutants that segregate in those domesticated populations termed the primary populations may be used as markers to follow genetic dynamics after domestication. Keywords: Mammals, Coat colour genetics, History Review Methodology: The present review was initially prepared at the request of an informal committee which met at the Veterinary School of Alfort near Paris on 22–23 June 2009 devoted to re-launching in France a network on coat and plumage colour genetics.

The Visible Polymorphism in Domesticated Mammals

visible effects. The Mouse New Letter (MNL) list included the following:

During the Middle Ages, a polymorphism affecting visible traits of mammals was described in the miniatures of the Livre de Chasse from Gaston Phoebus [1] for the dog. Nineteenth-century authors such as Darwin [2, 3] and Isidore Geoffroy St Hilaire [4] then noticed the contrast between the wild state (in which this polymorphism was rather rare) and the domestic state, in which it was a general rule.

1. 2. 3. 4. 5. 6. 7.

The Mendelian Proﬁle of Visible Polymorphisms in Mammals After the rediscovery of Mendel’s laws in 1900, mutants inducing visible effects were sorted according to their

g

coat colour, skin ﬁbres or horn, skeleton, appendices, neurological and muscular, endocrinian defects (nanism, obesity), or reproductive organs.

The list can be used in other species such as sheep, as proposed by Dolling et al. [5]. In some mammalian species, such as the mouse, humans and cattle, it is possible to divide visible mutants into various categories. There is a worldwide network of laboratories studying the mouse – the best genetically known laboratory mammal. Thousands of animals are raised and all new mutants are scanned to be saved and studied. The genetic

CAB International 2009

18

Animal Science Reviews 2010

results were gathered in the MNL before being made available online by Mouse Genome Informatics (MGI), published by the Jackson Laboratory, Bar Harbor, Maine, USA. As far as man are concerned, every year, thousands of human births are surveyed for birth defects or congenital diseases in maternity hospitals. Several hospitals have sections devoted to hereditary traits whether pathologic or not. In the 1960s, the results of these genetic analyses were electronically compiled by McKusick [6, 7] in Mendelian Inheritance in Man (MIM). MIM is now available online under Online Mendelian Inheritance in Man (OMIM). For cattle (Bos taurus), a Mendelian Inheritance in Cattle (MIC) [8] has been published, covering the whole visible Mendelian inheritance in which ‘clinical, pathological and other visible hereditary traits’ could be scanned with the results of Artiﬁcial Insemination programmes [9]. From these explorations, the following is apparent: (1) Pleiotropy is a constant behaviour of mutants with visible effect and (2) mutants usually have a very low viability.

The Beginnings of Coat Colour Genetic Studies in Mammals Studies on the coat colour genetics of mammals were initiated in the mouse by Cue´not [10], in sheep by Davenport [11] and Wood [12], and in cattle by Barrington and Pearson [13] and others. At the end of the 1910s, Wright [14–23] reviewed coat colour in ten species: the mouse, rat, rabbit, guinea pig, cattle, horse, dog, pig, cat and humans. Later, Haldane [24] compared coat colour in rodents and carnivores, for which Mendelian studies were made easier by their high proliﬁcity, short generation intervals and low individual cost of maintenance. The same species were scanned by Little [25], followed by Searle [26], who has taken into account all the mammalian species then genetically analysed for their coat colour genetics (around 70).

The Laws of Comparative Coat Colour Genetics in Mammals Wright [14–23] postulated that genes were controlling the same enzyme from one species to another, and his 1917 and 1918 papers were obviously premonitory, but the notion of homology between mammalian species in its deﬁnitive shape must be assigned to Haldane [24], who compared phenotypes induced by allelic series in various species. These phenotypes were very similar: for example in the A (Agouti) series, there were alleles controlling similar patterns of distribution of phoeomelanic (red) and

g

eumelanic (black or brown) pigments in the coat, when the colour of eumelanin (black or brown) is under the control of alleles at Brown (B) locus. Such phenotypic and genotypic homology has been conﬁrmed by Little [25] and then by Searle [26]. The latter author moreover deduced rules for the genetics of mammalian coat colour. Therefore, before the true biochemical composition of hereditary particles (DNA chains) was known after Watson and Crick in the 1950s, followed by the deciphering of DNA chains in the 1970s, the comparison of coat colour genetics had already brought strong proof that, at least in mammals, evolution has used the same loci. Currently, the presence of the same loci in the genotype of three species as far from each other in the evolutionary tree as human, mouse and zebraﬁsh as compared by Oeting et al. [27] show that the lessons of the comparative coat colour genetics have a historical interest. From these pioneering times, the ﬁeld has inherited an alphabetical nomenclature (A, Agouti, B, Brown, C, Colour, D, Dilution, E, Extension, etc.) which, while it will probably vanish, may be kept for the variants that have not been biochemically identiﬁed.

Enumeration of Colour Loci in Mammals In the mouse, Silvers [28] enumerated more than 50 loci with 130 determinants (phenotypic alleles). Four years later, according to Bennett and Lamoreux [29], this number was 127, with more than 800 alleles, and 59 of these loci had been sequenced. Recently, Oetting et al. [27] gave a total number of 281 loci, of which 114 were cloned and 167 not yet cloned: twice the score of Bennett and Lamoreux! In farm animals, the number of coat colour loci is much lower, because only those which are viable have been kept and analysed. Therefore one may ﬁnd only 18 loci in the dog [30, 31], 14 in cattle [32] and 11 in sheep [33]. It also appears that there is a prevalence of some homologous loci (A, B, C, D, E, etc.) Even in the small number of farm animals having a catalogue of Mendelian loci such as sheep [34], cattle [35] or dog [30, 36], there is a prevalence of viable mutants belonging to the phenotypic category 1 (coat colour), over all the other phenotypic categories. The second phenotypic category to be represented is category 2 for skin ﬁbres and horn as observed in the dog: 52 alleles at 18 loci controlling coat colour have been described against only 16 alleles at 8 loci for the architecture of coat [30].

Present Studies The molecular genetic exploration of farm animals has followed the purely Mendelian studies reviewed by Searle

CAB International 2009

J.J. Lauvergne

[26]. In cattle, these studies go back to 1995 for cattle [37] and to 1999 for sheep [38–40]. The genes have received new names: MC1R for Extension, TRPY for Brown, etc.

An Explanation for the Uniformity of Coat in Wild Species, with Some Exceptions In wild species, where all animals are uniform, the disruptive role of coat colour for both prey species and for predators gives a strong selection pressure in favour of the best ﬁtted colour allele [41, 42]. Nevertheless, in several wild species, a uniformity of coat or shell colour is not observed. An example is given by the snail Cepaea nemoralis, studied by Lamotte [43]. Every allele controlling shell colour confers a disruptive advantage on a given coloured background and the colour of the background is of course variable. Therefore one of the reasons why we can observe several coat colour mutants in domestic mammals may result from the fact that in the domestic state the disruptive effect does not play a role in selection.

Biological and Evolutionary Interest of Coat Colour Genes The various levels of action of coat colour genes on coat colour have been summarized [26, 27]. They already affect a wide range of biological phenomena. The range of defects listed by the last author shows that the biological importance of colour genes is even much bigger. The interest of these colour genes in the study of evolution derives from an homology betwwen allelic series (Agouti, Brown and Extension etc.) in several species. Once the DNA chains will be deciphered the evolution between and intra species will be explored.

The Colour Gene to Follow Genetic Dynamics after Domestication Farm animal population in which several alleles for coat colour segregate were described in cattle, in England [44] and in Norway [45]; in sheep in Iceland [46] and in London cat [47]. In his World Dictionary of Breeds, Types and Varieties of Livestock, Mason [48] states that several breeds, types or varieties present a great diversity of external aspects, especially in their coat colour i.e. are multi-segregating for colour alleles. Such entities were then named ‘non-uniform populations or by a geographical term meaning cattle of such and such a place or breeds of such and such a place’, by Mason [48] and Lauvergne [49] proposed to use the shorter term

g

19

‘traditional population’. In turn, the term ‘traditional population’ has been replaced by ‘primary population’, because they are the ﬁrst type of populations to appear after domestication, Lauvergne [50]. According to the laws of population genetics established in the 1900s [51, 52], followed by Haldane [53–57] a stable gene frequency equilibrium is reached, provided that: coefﬁcients of selection do not disadvantage mutant carriers (homozygous or heterozygous), which is the case because under domestication the natural selection does not play a role; reproduction is made under random mating, which may still be veriﬁed in existing primary populations of goats in Cameroon (F. Meutchie`ye, personal communication, 2009) and Senegal (D. Bouchel, personal communication, 2009) or in Damara sheep of South Africa, (D. Du Toit, personal communication, 2009). In Europe and in other places, the primary populations have disappeared because breeders have been able to control reproduction more strictly, mating animals of the same phenotype with a choice that has mainly been of a given coat colour. This is the way modern standardized breeds have originated from primary populations. Since Mason [48] multi-segregating populations have been described in sheep in Iceland [58], in North Atlantic islands (Shetland, Orkney and Soay) [59] and in Corsica [60]. Several papers of the INRA 1986 Conference of Manosque [61] described also such primary sheep and goat populations. More recently in Africa studies in goat – in Algeria [62, 63] and in sub-Saharan countries such as Chad [64], Nigeria [65] and Cameroon [66] and Senegal (D. Bouchel, personal communication, 2009) show that primary populations are numerous in a very large zone. This could also be the case in Africa with sheep from the Damara population of Namibia and South Africa [67], for which great colour variability has been described [68, 69]. The great interest is that such primary populations are probably derived with minute variations (appraisal of new mutants) from the ﬁrst populations of the two species kept in Middle East during some centuries (or millenaries?) after domestication, 12 000 years ago [70].

Conclusion/Summary This very brief history of coat colour genetics in mammals stress the central biological role played by the genes belonging to this phenotypic category of loci as seen by the wide range of their pleiotropy.

CAB International 2009

20

Animal Science Reviews 2010

Moreover, their mutants are the most frequently observed category of viable mutants among the domestic mammals and may be used as markers.

13. Barrington A, Pearson K. On the inheritance of coat colour in cattle. Biometrika 1906;4:427–64. 14. Wright S. Color inheritance in mammals. II. The mouse. Journal of Heredity 1917;8:373–8. 15. Wright S. Color inheritance in mammals. III. The rat. Journal of Heredity 1917;8:426–30.

Acknowledgments

16. Wright S. Color inheritance in mammals. IV. The rabbit. Journal of Heredity 1917;8:473–5.

The subject matter of this review was proposed by Dr Genevie`ve Aubin-Houzelstein (Molecular and Cellular Genetics, UMR 955, INRA-ENVA, Maisons-Alfort, France) and Professor Ahmad Oulmouden (Animal Molecular Genetics UMR 1061, Limoges, France). Professor Carlo Renieri (Environmental Sciences Department, University of Camerino, Italy) has been kind enough to read the manuscript.

17. Wright S. Color inheritance in mammals. V. The guinea pig. Journal of Heredity 1917;8:476–80. 18. Wright S. Color inheritance in mammals. VI. Cattle. Journal of Heredity 1917;8:521–7. 19. Wright S. Color inheritance in mammals. VII. The horse. Journal of Heredity 1917;8:561–4. 20. Wright S. Color inheritance in mammals. IX. The dog. Journal of Heredity 1918;9:87–90. 21. Wright S. Color inheritance in mammals. VIII. Swine. Journal of Heredity 1918;9:33–8. 22. Wright S. Color inheritance in mammals. X. The cat. Journal of Heredity 1918;9:139–44.

References 1. Phoebus G. Le Livre de chasse (The Book of the Hunt). Manuscript, Ortez (France); 1369.

23. Wright S. Color inheritance in mammals. XI. Man. Journal of Heredity 1918;9:227–40.

2. Darwin C. 1. Variations under domestication. In: Darwin C, editor. On the Origins of Species by Means of Natural Selection. John Murray, London; 1859 (In Burrow JW, Darwin C. The Origin of Species. Penguin Books Ltd, Harmondsworth, Middlesex, UK; 1968. p. 71–100).

24. Haldane JBS. The comparative genetics of color in rodents and carnivora. Biological Review 1927;2:199–212.

3. Darwin, C. The Variation of Animals and Plants under Domestication. 2 volumes. John Murray, London; 1868.

26. Searle AG. Comparative Genetics of Coat Colour in Mammals. Logos Press/Academic Press, London/New York; 1968.

4. Geoffroy-Saint-Hilaire I. Taming and Domestication of Useful Animals. 4th ed. La Maison rustique, Paris; 1861.

27. Oetting WS, Montollu L, Bennett DC. Color genes (Mouse). European Society for Pigment Cell Research; 2009. Available from: URL: http://www.espr.org/micemut: 1–2.

5. Dolling CHS, Millar P, Denis B, Rae AL, Lauvergne JJ, Renieri C. Listing loci and alleles of sheep and goats for visible traits other than colour 1987. In: Lauvergne JJ, editor. Standardized Genetic Nomenclature for Sheep and Goats 1987, Loci for Visible Traits Other Than Colour and Blood and Milk Polymorphisms. Lavoisier, Paris; 1989. p. 27–32.

25. Little CC. Coat colour genes in rodents and carnivores. Quarterly Review of Biology 1956;33:103–37.

28. Silvers WK. The Coat Color of Mice. Springer-Verlag, Berlin; 1979. 29. Bennett D, Lamoreau ML. The colour loci of mice – a genetic Century. Pigment Cell Research 2003;16:333–44.

6. McKusick V. Foreword. In: McKusick V, editor. Mendelian Inheritance in Man. The John Hopkins Press, Baltimore, MD; 1966. p. vii–xi.

30. Sponenberg DP. Rothschild genetics of coat color and hair texture. In: Ruvinsky A, Sampson J, editors. The Genetics of the Dog. CABI Publishing, Wallingford, Oxon, UK; 2001. p. 61–85.

7. McKusick V, editor. Mendelian Inheritance in Man. The John Hopkins Press, Baltimore, MD; 1966.

31. Denis B. Dog Coat Colours. Aniwa SAS, Paris; 2008.

8. Millar P, Lauvergne JJ, Dolling CHS, editors. Mendelian Inheritance in Cattle 2000. Wageningen Pers, Wageningen, The Netherlands; 2000. 9. Huston K, Saperstein G, Steffen D, Millar P, Lauvergne JJ. Clinical, pathological and other visible traits loci except coat colour (Category 2). In: Millar P, Lauvergne JJ, Dolling CHS, editors. Mendelian Inheritance in Cattle. EAAP publication No. 101. Wageningen Pers, Wageningen, The Netherlands; 2000. p. 107–29. 10. Cue´not L. Mendel Law and Heredity of Pigmentation in Mouse. Archives de Zoologie Expe´rimentale et de Ge´ne´tique 3e`me se´rie 1902;10:27–30. 11. Davenport CB. The origin of black sheep in the ﬂock. Science NS 1905;22:674. 12. Wood. Note on the inheritance of horns and face colour in sheep. Journal of Agricultural Science (London) 1905;364–5.

g

32. Lauvergne JJ, Dolling CHS, Rae AL, Renieri C, Sponenberg DP, Denis B. 1. Coat colour loci (category 1). In: Millar P, Lauvergne JJ, Dolling CHS, editors. Mendelian Inheritance in Cattle. Wageningen Pers, Wageningen, The Netherlands; 2000. p. 37–105. 33. Sponenberg DP, Dolling CHS, Lundie RS, Rae AL, Renieri C, Lauvergne JJ. 1. Coat colour loci (category 1). In: Lauvergne JJ, Dolling CHS, Renieri C. Mendelian Inheritance in Sheep 1996 (MIS 96). COGOVICA/COGNOSAG and University of Camerino, Clamart/Camerino; 1996. pp. 15–57. 34. Lauvergne JJ, Dolling CHS, Renieri C, editors. Mendelian Inheritance in Sheep 1996 (MIS 96). COGOVICA/ COGNOSAG and University of Camerino, Clamart and Camerino; 1996. 35. Millar P, Lauvergne JJ, Dolling CHS. Mendelian inheritance in Cattle. EAAP publication No. 101. Wageningen Pers, Wageningen, The Netherlands; 2000.

CAB International 2009

J.J. Lauvergne 36. Nicholas FW. Genetics of morphological traits and inherited disorders. In: Ruvinsky A, Sampson J, editors. The Genetics of the Dog. CABI Publishing, Wallingford, Oxon, UK; 2001. p. 87–116. 37. Klungland H, Va˚ge D, Gomez-Raya L, Adalsteinsson S, Lien S. The role of melanocyte-stimulating hormone (MSH) receptor in bovine coat color determination. Mammalian Genome 1995;6:636–63. 38. Parsons YM, Fleet MR, Cooper DW. Short communication – isolation of the ovine Agouti coding sequence. Pigment Cell Research 1999;12:394–7. 39. Parsons YM, Fleet MR, Cooper DW. The Agouti gene: a positional candidate for recessive black in Australian Merino sheep. Australian Journal of Agricultural Research 1999;50:1099–3. 40. Va˚ge ID, Klungland H, Lu D, Cone RD. Molecular and pharmacological characterization of dominant black coat color in sheep. Mammalian Genome 1999;10:39–43. 41. Cot HB. Adaptative Coloration in Animals. Methuen, London; 1940. 42. Bourlie`re F. The Natural History of Mammals. Harrap, London; 1955. 43. Lamotte M. Recherches sur la structure ge´ne´tique des populations naturelles de Cepaea nemoralis (L.). Bulletin Biologique de France et de Belgique 1951;Suppl. XXXV: 1–239. 44. Wilson J. The colours of highland cattle. Scientiﬁc Proceedings of the Royal Dublin Society 1909;12:66–76. 45. Berge S. Inheritance of dun, brown and brindle colour in cattle. Heredity 1949;5:195–204. 46. Zophoniasson P. Nogle Bemaerkninger om enkelte Arvelighedsforhold hos de islandske Faar. Nordisk Jordbruksforskning 1934;16:217–3. 47. Searle AG. Gene frequencies in London cats. Journal of Genetics 1949;49:214–0. 48. Mason IL. A World Dictionary of Breeds, Types and Varieties of Livestock. Technical Communication No. 8 of the Commonwealth Bureau of Animal Breeding and Genetics, Edinburgh Commonwealth Agricultural Bureaux, Farnham House, Farnham Royal, Slough, Bucks., England; 1951. 49. Lauvergne JJ. Genetics in animal populations after domestication, the consequences for breed conservation. In: Proc. 2nd World Congress on Genetics applied to Livestock Production, Madrid, Volume 6; 1982. p. 77–87. 50. Lauvergne JJ. Breed development and breed differentiation. In: Proceedings CEC Workshop and Training Course, 7–9 December1992, Hannover. Commission of the European Communities, Agriculture, Brussels; 1993. p. 53–64. 51. Hardy GH. Mendelian proportion in a mixed population. Science 1908;28:49–50. 52. Weinberg W. U¨ber die Nachweis der Vererbung beim Menschen. Jahresb. Verein. Vaterl. Naturk. in Wu¨rttemberg 1908;64:368–82.

21

composition of Mendelian populations, and on natural selection. Proceedings of the Cambridge Philosophical Society – Biology 1924;1:158–63. 55. Haldane JBS. A mathematical theory of natural and artiﬁcial selection. III. Proceedings of the Cambridge Philosophical Society 1926;23:363–72. 56. Haldane JBS. A mathematical theory of natural and artiﬁcial selection. IV. Proceedings of the Cambridge Philosophical Society 1926;23:607–15. 57. Haldane JBS. A mathematical theory of natural and artiﬁcial selection. V. Selection and mutation. Proceedings of the Cambridge Philosophical Society 1927;23:838–44. 58. Adalsteinsson S. Colour inheritance in Icelandic sheep and relation between colour, fertility and fertilization. Journal of Agriculture Research Iceland 1970;2(1):3–135. 59. Ryder ML, Land RB. Coat colour inheritance in Soay, Orkney and Shetland sheep. Journal of Zoology London 1974;173:477–85. 60. Lauvergne JJ, Adalsteinsson S. Ge`nes pour la couleur de la toison de la brebis Corse. Annales de Ge´ne´tique et de Se´lection animale 1976;8:153–72. 61. Lauvergne JJ. Populations traditionnelles et premie`res races standardise´es d’Ovicaprinae dans le Bassin me´diterrane´en. Colloque INRA N 47 Gontard/Manosque France 30 juin–2 juillet 1986; 1988. 62. Khemici E, et al. Indice de primarite´ et diffe´renciation ge´ne´tique des populations caprines de la steppe (Arabia) et du de´sert (Mekatuatia) d’Alge´rie. Genetics Selection Evolution 1995;27:503–17. 63. Khemici E, et al. Etude des ressources ge´ne´tiques caprines de l’Alge´rie du nord a` l’aide des indices de primarite´. AGRI 1996;17:69–80. 64. Ozoje MO. Coat colour genes in West African dwarf sheep and goats: a theoretical appraisal. Proceedings of the Sixth World Congress on Genetics Applied to Livestock Production, Volume 26; 1998. p. 53–6. 65. Lauvergne JJ, Bourzat D, Souvenir Zaﬁndrajaona P, Zeuh V, Ngo Tama AC. Indices de primarite´ de che`vres au Nord Cameroun et au Tchad. Revue d’Elevage et de Me´decine Veterinaire des Pays Tropicaux 1993;46:651–65. 66. Meutchieye F. Caracte´risation morphome´trique de la che`vre locale des hautes terres de l’Ouest Cameroun. The`se de Master of Science en Biotechnologie et Productions animales, option Ame´lioration ge´ne´tique et syste`mes de production, Universite´ de Dschang (Cameroun), Faculte´ d’Agronomie et des Sciences agricoles, De´partement des Productions animales; 2008. 67. du Toit D, editor. The Damara of Southern Africa. Dawie Du Toit, PO Box 141, Prieska 8940, South Africa; 2007. 68. Lundie RS. Coat colour genetics of the Damara sheep. In: du Toit D, editor. The Damara of Southern Africa. Dawie Du Toit, PO Box 141, Prieska 8940, South Africa; 2007. p. 38–122.

53. Haldane JBS. A mathematical theory of natural and artiﬁcial selection. I. Transactions of the Cambridge Philosophical Society 1924;23:19–41.

69. Sponenberg DP. Colour genetics and Damara sheep. In: du Toit D, editor. The Damara of Southern Africa. Dawie Du Toit, PO Box 141, Prieska 8940, South Africa; 2007. p. 28–37.

54. Haldane JBS. A mathematical theory of natural and artiﬁcial selection. II. The inﬂuence of partial self-fertilization, inbreeding, associative mating and selective fertilization on the

70. Vigne JD. Les origines de la culture: les de´buts de l’e´levage. Collection Le Colle`ge de la Cite´, Le Pommier et Cite´ des Sciences et de l’Industrie, Paris; 2004.

g

CAB International 2009

This page intentionally left blank

Animal Science Reviews 2010

Functional genomic approaches to understand the biological pathways underpinning intramuscular fat in beef L. Pannier1,2, R. M. Hamill1, A. M. Mullen1 and T. Sweeney2* Address: 1 Teagasc, Ashtown Food Research Centre, Ashtown, Dublin 15, Ireland. 2 School of Agriculture, Food Science and Veterinary Medicine, University College Dublin, Belﬁeld, Dublin 4, Ireland. *Correspondence: Torres Sweeney. Fax: +35317166253. Email: [email protected] Received: Accepted:

4 August 2009 20 November 2009

Abstract Meat quality traits, such as intramuscular fat level, have a heritable component and are quantitative in nature, being inﬂuenced by many genes of variable individual effects. The identiﬁcation of the genetic variation associated with meat quality is an important step towards the improvement of meat quality in livestock selection programmes and meat management systems. Functional genomic approaches relate the static information of variation in genomic regions and individual genes with the dynamic interplay of gene expression pathways and biological processes. Therefore, through the application of functional genomic tools, researchers are gaining deeper insights into the biological pathways underpinning meat quality. Knowledge gained from these approaches can be beneﬁcial in deﬁning and optimizing management systems for quality, providing assurance of meat quality and in tailoring quality to suit market needs. Keywords: Functional genomics, Genetic markers, Bovine genome, Single nucleotide polymorphisms Review Methodology: This review paper was established by researching several databases such as PubMed, NCBI and Science Direct. Conference proceedings such as International Conference of Meat Science and Technology (ICoMST) and the European Association of Animal Production (EAAP) meetings, as well as CABI and Woodhead published books were used for information and references. In addition to this, advice and unpublished work from collaborating partners and colleagues was collected.

Intramuscular Fat (IMF) The concept of meat quality has been deﬁned in many ways, from safety through to sensory experience [1]. Palatability traits include juiciness, ﬂavour and tenderness. While consumers maintain an increasing desire for these traits, they are increasingly seeking lower fat/healthier options in food choices [2]. In beef, these consumer drivers are linked through the level of IMF [2], also subjectively assessed as marbling [3]. IMF has been shown to contribute 10–15% of the variance in palatability [4], which can be compromised if the fat content is less than 3%, but is enhanced as IMF levels increase from 3 to 6% [5]. Juiciness and ﬂavour tend to increase with increased IMF content. Fat can affect juiciness in a number of ways: (1) by enhancing the water holding capacity of meat, (2) by

g

lubricating muscle ﬁbres during cooking and (3) by stimulating the salivary glands during mastication [5]. Fat has also been identiﬁed as one of the major components of the meat ﬂavour lexicon [6]. When meat is cooked, the lipid components (principally triglycerides) melt and along with other meat constituents (amino acids and reducing sugars) undergo a series of thermally induced reactions, which form ﬂavour compounds. This is known as the Maillard reaction [7]. Fat is a precursor of a large number of ﬂavour compounds, such as aldehydes and ketones, and it contributes to undesired (e.g. rancidity) as well as desired tastes and aromas [7]. There is conﬂicting evidence on the precise nature of the relationship between fat content and tenderness. Some evidence is available that IMF increases meat tenderness [8], but other studies suggests that it has a negative effect [9, 10]. The relationship between those

CAB International 2009

24

Animal Science Reviews 2010

two aspects of quality is thought to be indirect [11] and an earlier study estimated that marbling explains just 5% of the variation in beef tenderness [12]. IMF clearly has an important inﬂuence on palatability given its speciﬁc contribution to juiciness, ﬂavour and possibly tenderness [12–14]. Furthermore, IMF is integrated in most industry grading systems and it is an important determinant of carcass value for certain markets such as Japan, where Japanese Black cattle are highly regarded for their higher marbling characteristics [15, 16]. Hence, in this review paper, we will discuss recent molecular research focused on increasing our understanding of the genetic component of IMF and review progress towards the identiﬁcation of biological pathways controlling IMF in the bovine. Genetic Control of IMF in the Bovine Genome Considerable progress has been achieved towards the goal of placing whole genome sequences and associated genomic tools into the public domain for high-priority domestic animal species. Today, the complete draft of the bovine genome has been released [17, 18] and is free to public access on different databases such as NCBI (www. ncbi.nih.gov/Genbank) and EMBL (www.ebi.ac.uk/embl/ index.html). The bovine genome contains approximately 2.87 billion base pairs and at least 22 000 protein-encoding genes [19]. Gene annotations are being continuously updated, e.g. by construction of a gene atlas of expressed transcript tags in up to 100 tissues [20]. Quantitative Trait Loci (QTLs) Most production and meat quality traits are complex traits, inﬂuenced by many genes with variable effects [21, 22]. While environmental factors such as nutrition have an inﬂuence, a signiﬁcant proportion of this variation is inﬂuenced by the genetic potential of the animal. Because of the consumer’s demand for high and consistent quality of beef, researchers are aiming to develop DNA markers of beef quality and to increase the understanding of muscle biology [23, 24]. Through mapping studies, genomic regions (QTLs) have been associated with meat quality [25]. Polymorphic markers are used to detect regions on a chromosome that have a substantial effect on quantitative traits and include restriction or ampliﬁed fragment length polymorphism (RFLP or AFLP) markers, microsatellites, single nucleotide polymorphisms (SNPs) and/or expressed sequence tags (ESTs) to map genes to large chromosomal regions which may contain hundreds of genes [24–26]. QTLs for marbling have been identiﬁed and are shown in Table 1 [25, 27, 28]. There are a number of drawbacks to the QTL mapping approach. Firstly, it can fail to detect trait loci with small effects and secondly large pedigrees with good phenotypic and genotypic data are required. Consequently, the

g

Table 1

QTLs for beef IMF/marbling

BTA

Literature

2 3 4 5 6 7 8 9 10 13 14 16 17 18 20 21 23 26 27 29

[29–31] [32–34] [35, 36] [29, 33] [36–38] [38, 39] [32] [33, 38, 40] [32, 33, 38] [35] [33, 41] [34] [42] [31] [38] [36, 38] [33] [31] [33] [42, 31]

BTA, Bos taurus autosome.

identiﬁcation of QTLs for many traits including meat quality has proven to be quite slow and in reality many studies lack power due to insufﬁcient sample sizes [25].

Polymorphisms in Candidate Genes The ultimate goal of functional genomics analysis is to identify the causative polymorphisms controlling variation in meat quality traits. Two different approaches have been used to identify candidate genes for further exploration. These are: (1) genes underlying particular QTL regions and (2) genes involved in biological pathways hypothesized to regulate a trait of interest. Sequencing phenotypically divergent individuals at candidate gene loci has led to the identiﬁcation of mutations, most of which are SNPs. These SNPs can inﬂuence the level of transcription of the gene or the amino acid structure of the transcribed protein. SNPs have been associated with numerous meatquality-related traits including tenderness [43, 44], waterholding capacity [45], colour [46], muscular hypertrophy [47] and growth rate [48]. SNPs in candidate genes associated with IMF/marbling scores and fat-related traits are presented in Table 2. With a view to uncovering the molecular basis of IMF deposition, research has focused on a number of candidate genes based on their major regulatory roles in pathways associated with triglyceride synthesis, adipose differentiation and fatty acid metabolism. Some of these are outlined below. The leptin gene, also known as the obese gene, encodes a 16 kDa peptide hormone, which is synthesized and secreted mainly from white adipocytes [72]. Leptin acts both centrally and peripherally and regulates adiposity, food intake, energy expenditure, body weight and overall

CAB International 2009

L. Pannier, R. M. Hamill, A. M. Mullen and T. Sweeney

25

Table 2 Overview of association studies of DNA markers for IMF, marbling score and fat related traits Marker

BTA

EDG1 RORC BM 1500 Leptin

3 3 4 4

CSSM34

5

ETH10

5

AFABP

14

BMS1747 CBFA2T1 CSSM66

14 14 14

DGAT1

14

DECR1 TG

14 14

FASN

19

GH1

19

FGF8 SCD

26 26

Association study

Literature

Association detected

Marbling breeding value in Japanese Black cattle Marbling scores in a Bos taurus population Marbling scores in purebred Angus, Charolais, Simmental and Hereford Increased fat deposition in purebred Angus, Charolais and Hereford IMF content in commercial cattle Fat and lean yield in commercial cattle IMF content in commercial cattle IMF and marbling score in cross-bred bull calves (HolsteinCharolais) Marbling scores in offspring of Angus and Shorthorn sires Marbling scores in offspring of Wagyu sires Marbling scores in offspring of Wagyu sires Marbling scores in offspring of Angus and Shorthorn sires Marbling score in a WagyuLimousin F2 population IMF content in commercial cattle Marbling scores in Angus and Shorthorn cattle Ultrasound marbling scores in Angus, Charolais and cross-bred cattle Marbling scores in offspring of Wagyu sires Marbling scores in offspring of Angus and Schorthorn sires Marbling scores in Angus and Shorthorn cattle IMF in German Holstein cattle Marbling score in Bos indicus cattle Backfat levels in B. taurus cattle IMF content in commercial cattle Milk fat content in dairy cattle Ultrasound backfat in Angus, Charolais and cross-bred cattle Marbling scores in offspring of Angus and Shorthorn sires Marbling scores in Angus and Shorthorn cattle IMF in German Holstein cattle Marbling score on a B. taurus population Backfat levels in B. taurus cattle Marbling score on a B. indicus population IMF content in commercial cattle Marbling score in purebred cattle Milk fat content in Holstein cattle Fatty acid composition of IMF in Japanese Black cattle Fatty acid composition of longissimus dorsi muscle in purebred Angus bulls Fatty acid composition of adipose tissue and milk fat in B. taurus cattle Marbling scores in Japanese Black cattle Marbling scores in Angus and Shorthorn cattle Carcass backfat in Angus, Charolais and cross-bred cattle MUFA content and melting point of IMF in Wagyu cattle

[15, 16] [49] [50] [51] [52] [52] [53] [54] [55] [55] [55] [55] [56] [57] [58] [59] [55] [55] [58] [60] [61] [41] [57] [62] [59] [55] [58] [60] [63] [41] [61] [57] [64] [65] [66] [67] [68] [69] [70] [59] [71]

Yes Yes Yes Yes No Yes No No Yes No Yes No Yes No No Yes Yes No No Yes No No No Yes Yes Yes Yes Yes No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes

BTA, Bos taurus autosome.

metabolism [73]. Because of leptin’s close location to microsatellite locus BM1500 [74], which was associated with fat measurements in beef bulls [50], the gene was hypothesized to inﬂuence fat-related traits in cattle [51, 54]. A number of other SNPs have been reported in the bovine leptin gene, both in the coding region [51, 54, 75] and in the promoter region [48, 76]. Numerous association studies have been carried out with these SNPs, not only for IMF or marbling levels (Table 2) but also for diverse meat quality traits. Polymorphisms have also been associated with fat deposition [51], feed intake [48, 54], fat and lean yield [52], quality and yield [77], milk

g

fat yield [76], milk yield [78] and average daily gain [48]. However, in some cases, where a different population was studied or different sample sizes used, these associations could not be replicated [53, 54, 77, 79, 80]. The thyroglobulin (TG) gene is a precursor molecule of the thyroid hormones that are known to affect adipocyte differentiation and lipid metabolism. The TG gene became a positional candidate gene for marbling after 0 the association of an SNP in the untranslated 5 region with marbling score, and is the source of a commercially available DNA test marker [81, 82]. However, conﬂicting results concerning its effect on marbling in beef cattle

CAB International 2009

26

Animal Science Reviews 2010

exist [60, 61, 63, 81, 83]. Recently, novel bovine SNPs 0 have been identiﬁed in the 3 ﬂanking region of the TG gene and an association was found between several of those SNPs and marbling score [64]. The diacylglycerol O-acyltransferase (DGAT1) gene codes for a microsomal enzyme that catalyses the ﬁnal step of the triglyceride synthesis pathway [84]. Studies have indicated that mice lacking both copies of DGAT1 are completely devoid of milk secretion, most likely because of deﬁcient triglyceride synthesis in the mammary gland [85]. Few SNPs were identiﬁed in the DGAT1 gene [62], with the most studied being a dinucleotide polymorphism in exon 8 which causes a lysine (K) to alanine (A) amino acid exchange at position 232 of the predicted protein [86]. Signiﬁcantly higher levels of IMF were identiﬁed in musculus semitendinosus in purebred German Holstein cattle for this polymorphism [60], but no signiﬁcant association with marbling score was observed in Bos indicus cattle [61]. The Adipocyte Fatty Acid Binding Protein (AFABP) gene codes for adipocyte fatty acid binding protein which is expressed almost exclusively in adipocytes [87] and is involved in the intracellular targeting of fatty acids [88]. The importance of AFABP for IMF deposition has been conﬁrmed by several independent studies [89, 90] and few SNPs have been identiﬁed which have a signiﬁcant association with marbling score and subcutaneous fat depth [56] and back fat traits [91]. Other candidate genes for IMF include: the sterol regulatory element binding protein-1 (SREBP-1), stearoylCoA desaturase (SCD) gene, fatty acid synthase (FASN) gene and the retinoid-related orphan receptor gamma (RORC) gene. They are considered important because of their related functions in fatty acid metabolism in cattle. SREBP-1 belongs to group of transcription factors that play a central role in energy homeostasis by promoting glycolysis, lipogenesis and adipogenesis [92, 93]. A deletion mutation in this gene contributed to a higher monounsaturated fatty acid (MUFA) proportion and lower melting point in IMF tissue in Japanese Black cattle [92]. The SCD gene located on BTA26 codes for SCD, which is the enzyme responsible for the conversion of saturated fatty acids into MUFA in mammalian adipocytes. It was proposed as a candidate gene for variation in fatty acid composition because of its role in fatty acid oxidation and several SNPs in the gene have been identiﬁed [71]. Associations between SNP genotypes and MUFA percentage and melting point of IMF were determined in Japanese Black cattle [71]. The FASN gene, mapped to BTA19 [68, 94], plays a role in de novo biosynthesis of long-chain fatty acids and lipogenesis in mammals [95]. Several SNPs on this gene have been identiﬁed by various independent studies and have been associated with milk fat content in Holstein–Friesian cattle [65] and with the fatty acid composition of IMF [66], longissimus dorsi muscle [67] and adipose fat and milk fat [68]. The RORC gene is a member of the steroid and thyroid hormone

g

receptor superfamily [96, 97] and binds retinoic acid as well as thyroid hormone. The gene is highly expressed in skeletal muscle [98] and associations between intronic and exonic alleles in the RORC gene and marbling score have been identiﬁed [49]. Furthermore, other genes which might affect fatness in cattle, are located in QTL regions on BTA3 [32–34, 99–101], these include the phosphatidylinositol 4-kinase (PIK4CB) [102], AMP-activated protein kinase b-2 subunit (PRKAB2) [103] and endothelial differentiation sphingolipid G-protein-coupled receptor 1 gene (EDG1) [15, 16]. Recently, an AFLP-based genome-wide scan has led to the direct identiﬁcation of a novel candidate gene associated with fat metabolism, i.e. the mitochondrial poly A polymerase gene (PAPD1), SNPs in which were subsequently found to contribute to extreme fat deposition in cattle [104, 105]. While progress has been made by using the candidate gene approach, the necessity of prior knowledge of the physiological and biological pathways of the candidate genes, the time in evaluating all genes potentially involved in a given trait, plus the potential contributions of yet uncharacterized genes and biological pathways have hindered progress.

DNA SNP Arrays, Genome-Wide Association Studies and Genomic Selection Today, progress in the bovine genome sequencing project [19] and associated SNPs discovery efforts [106, 107] has resulted in the identiﬁcation of hundreds of thousands of SNPs in all major breeds of cattle. Based on this research, a number of high-throughput SNP genotyping platforms have been developed including the Affymetrix GeneChip1 Bovine 10 and 25 K SNP chips featuring respectively 10 and 25 000 SNPs and the Illumina Inﬁnium1 Bovine SNP50 BeadChipä genotyping system, which permits interrogation of more than 54 000 SNPs evenly spaced throughout the genome. Larger chips are in development and to give an indication of the ultimate potential of bovine chips for ﬁne mapping, human chips containing more than 1 million SNPs (e.g. Human1M-Duo BeadChip, www.illumina.com) are currently available. Hence, higher-throughput platforms for the bovine are likely to be available in the near future. One of the major applications of these arrays is in whole-genome association (WGA) studies. Chips are currently being applied to identify genomic regions inﬂuential in relation to numerous economically important traits in livestock, including fertility, growth, milk production, reproduction, disease susceptibility and meat quality [108, 109]. Marker density is very important in determining the power to detect markers in linkage disequilibrium with causative SNPs and this depends on the extent of linkage disequilibrium in the genome. In cattle, linkage disequilibrium is quite extensive; however,

CAB International 2009

L. Pannier, R. M. Hamill, A. M. Mullen and T. Sweeney

achieving high power to identify speciﬁc variation inﬂuential on quantitative traits may require upwards of 100 000 SNPs [25]. A second major application of these markers lies in the enhancement of breeding programmes, via markerassisted or genomic selection [110, 111]. Genomic selection refers to the making of selection decisions based on genomic estimated breeding values (gEBV). These are calculated as the sum of the effects of dense genetic markers, or haplotypes of these markers, across the entire genome. This can potentially capture the effects of all QTLs that contribute to trait variation, even if their individual contribution is small [111, 112]. In order to calculate gEBV, a large reference population is ﬁrst genotyped and fully phenotyped and subsequently the genotype effects on the trait of interest are estimated. In subsequent generations, animals can be genotyped at birth to identify the chromosome regions they have inherited. Using this information, by summing the marker effects across the whole genome, gEBV can be estimated. Simulations have shown that EBV could be predicted with an accuracy of 0.85 from marker data alone, where markers are in high linkage disequilibrium with causative mutations [111]. Advantages of genomic selection include dramatic increases in the accuracy of EBV estimates, a more balanced approach to selection across traits, increased ability to select dams and decreases in generation interval and cost of progeny testing. Challenges include achieving sufﬁcient numbers of genotyped and phenotyped individuals, accounting for breed and family structure, the potential for loss of linkage disequilibrium between marker and trait over time and the development of appropriate statistical methods [110]. Genome-wide selection requires a large set of DNA markers, hence is becoming feasible with larger SNP chips [113]. While the approach needs to be robustly tested and validated in multiple populations, it is expected to increase the rate of genetic improvement two-fold per year in many livestock systems [112]. The increased application of high-density chips for livestock animals has the potential to revolutionize production methods in breeding, husbandry and nutrition, to improve beef quality and to produce differentiated products [23]. Genomic selection is already being applied to dairy cattle breeding programmes around the world [110, 113–116]. Meat quality traits such as IMF and tenderness have moderate or low heritabilities and can currently only be measured in a research setting. For these reasons, substantial genetic progress for important meat quality traits via WGA and genomic selection approaches is likely to be viable as long as appropriate methodologies are applied [25].

Gene Expression Expression arrays allow the detection of genes that are actively transcribing at a given time point. The

g

27

transcription rate of genes related to meat quality can be affected by endogenous factors such as cell type and stage of development as well as environmental factors such as nutrition, growth rate, handling and post-mortem changes [27, 117–119].

Gene Expression Techniques Traditional gene expression measurement techniques included northern blotting, slot and dot blotting and more recently, quantitative real-time PCR. These techniques have identiﬁed a number of individual genes associated with IMF levels in meat [90, 120, 121]. More recently, techniques have progressed to transcriptome-wide approaches, including suppressive subtractive hybridization (SSH) [122, 123], differential display [124, 125], serial analysis of gene expression (SAGE) [126], microarray hybridization [127–129] and next-generation cyclic array sequencing approaches which incorporate library construction, in vitro ampliﬁcation by e.g. emulsion PCR and iterative cycles of enzymatic synthesis and array imaging and analysis [130]. One of the widely applied technologies for exploring pathways underpinning meat quality and in particular, IMF, has been RNA expression microarray technology. This technology allows the simultaneous analysis of thousands of genes in a single experiment. This contrasts with ‘single gene’ studies where throughput is very limited. Two types of microarrays, spotted cDNA arrays and oligonucleotide arrays, are most commonly used for gene expression proﬁling. Commercial oligonucleotide arrays are now available for different domestic animal genomes (bovine, porcine, mouse, rat and chicken). For instance, the Affymetrix GeneChip1 Bovine Genome Array contains 24 027 probes sets which can be used to study gene expression of over 23 000 bovine transcripts. Given the complex nature of expression data, statistical reﬁnement through bioinformatics is crucial. Bioinformatics tools such as Genomatix Bibliosphere Pathway Edition (www. genomatix.de/) or PANTHER classiﬁcation system permit enhanced interpretation of the biological process and metabolic pathways involved. This further data mining can provide more information regarding the complex interplay of gene networks and their regulation by transcription factors [131].

Application of Microarrays to Enhance our Understanding of IMF Development Microarrays have been used to examine gene expression proﬁles in animals divergent in meat quality traits (Table 3). A cDNA microarray comprising 1915 bovine ESTs and 7291 anonymous cDNAs clones from bovine skeletal muscle and fat cDNA libraries [144] has been used in several breed and developmental stage

CAB International 2009

28

Animal Science Reviews 2010

Table 3 Bovine meat quality expression studies Array

Identiﬁcation of differentially expressed genes

Literature

Bovine skeletal muscle and fat cDNA array Bovine skeletal muscle and fat cDNA array Bovine skeletal muscle and fat cDNA array Bovine skeletal muscle and fat cDNA array Bovine skeletal muscle and fat cDNA array Bovine skeletal muscle, fat and liver cDNA array Bovine 0.5 K muscle cDNA array Bovine mixed tissue cDNA microarray Bovine mixed tissue cDNA microarray Bovine mixed tissue cDNA microarray Human and murine oligonucleotide array Human and murine oligonucleotide array

between Brahman steers fed with different diets

[132]

of Brahman and Brahman composite steers fed on different feeding regimes

[133]

between Japanese Black and Holstein cattle

[134]

between foetuses from WagyuHereford and PiedmonteseHereford crosses

[135]

between muscle biopsies from WagyuHereford and PiedmonteseHereford crosses taken at ﬁve different time points between different muscle samples in Korean cattle

[136]

between high IMF and low IMF longissimus samples of cross-bred cattle

[138]

between skeletal muscle samples of Polish Black and White bulls between 6 and 12 months of life between 12-month-old bulls of four cattle breeds

[139]

0

[137]

[140]

in Holstein–Friesian bulls carrying a polymorphism in the 5 -ﬂanking region of the mstn gene between normal and double-muscled cattle

[141]

between beef meat cuts differing in tenderness, juiciness and ﬂavour

[143]

comparative studies. When foetuses from both Wagyu Herford crosses were compared with foetuses from PiedmonteseHereford crosses, 82 differentially expressed genes were detected according to the breed of the foetus sire [135]. The most marked increase or decrease in gene expression was observed between day 195 of foetal development and birth [135]. The majority of changes in gene expression in foetal longisimus muscle comprised structural and metabolic components of extracellular matrix and muscle ﬁbres. Although IMF is a latedeveloping fat depot, this microarray study identiﬁed a number of genes associated with adipogenesis/lipogenesis (e.g. FABP4 and FABP5) to be already differentially expressed according to sire in early developmental stages and conﬁrms that the cellular development of adipocytes is ﬁxed relatively early in life. The same array was used to study and investigate differential gene expression in the longisimus muscle of Japanese Black (characterized with high IMF levels) and Holstein cattle at 11.5 months of age [134]. The results indicated that at 11 months of age, the genes associated with adipogenesis, mono-unsaturated fatty acid synthesis and fatty acid accumulation were highly expressed in Japanese Black cattle. A heterologous microarray (human and murine oligonucleotides sequences) of around 6000 genes expressed in muscle was used to identify genes that were differentially expressed between double-muscled (hypertrophy) and normal animals in the semitendinosus muscle [142]. Double-muscled animals are characterized by leaner carcasses with less fat. Several differentially expressed genes were identiﬁed (MEF2A, ID1, ZFX1B, MyoD1, RPL28 and

g

[142]

CoLIAI) which are involved in several physiological processes of muscle tissue development at 260 days of foetal development. The authors concluded that the biological trait of double-muscling is associated with speciﬁc gene proﬁles at the time of ﬁbre differentiation and/or specialization in late foetuses. For example, there was a decrease of collagen (COL1A1 and COL1A2) gene expression of late-stage bovine foetuses carrying the myosin mutation. Recently, a cDNA microarray was developed consisting of 448 genes and ESTs derived from an SSH library for IMF, tenderness and drip loss, plus 100 candidate genes from the literature. Thirty-eight genes were identiﬁed as being differentially expressed (e.g. ATP synthase F0 subunit 6, ACADVL, MRLC2 and Collagen type I) between two groups phenotypically divergent for IMF level [138].

Conclusions Major advances are occurring in functional genomics tools that will accelerate the acquisition of the pathways underpinning meat quality, in particular IMF, and it is clear that these will ultimately provide a mechanism to control quality for the consumer. A wealth of DNA SNP and RNA expression data are being generated experimentally through breed, nutritional status and developmental studies as well as studies focused on individual variation in meat quality. In order to develop a holistic systems biology approach to the understanding of IMF in beef, there is, however, an urgent need for the continuing development of the bioinformatical platforms to facilitate

CAB International 2009

L. Pannier, R. M. Hamill, A. M. Mullen and T. Sweeney

the effective integration of DNA SNP data and RNA expression data.

29

G-protein-coupled receptor 1 gene with marbling in Japanese Black beef cattle. Animal Genetics 2009;40:209–16. 16. Yamada T, Sasaki S, Sukegawa S, Miyake T, Fujita T, 0 Kose H, et al. Novel SNP in 5 ﬂanking region of EDG1 associated with marbling in Japanese Black beef cattle. Animal Science Journal 2009;80:486–9.

References 1. Hocquette JF, Gigli S. The challenge of quality. In: Hocquette JF, Gigli S, editors. Indicators of Milk and Beef Quality. EAAP publication No. 112. Wageningen Academic Publishers, The Netherlands; 2005. p. 13–22. 2. Harper GS, Pethick DW. Fatness in production animals: using genetic and environmental levers to meet consumer demand. Asia Paciﬁc Journal of Clinical Nutrition 2004;13:37. 3. Harper GS, Pethick DW. How might marbling begin? Australian Journal of Experimental Agriculture 2004;44:653–62. 4. Dikeman ME. Fat reduction in animals and the effect on palatability and consumer acceptance of meat products. In: Proceedings of the 40th Annual Reciprocal Meat Conference; Chicago, USA. National Livestock and Meat Board, Chicago, IL; 1987. p. 93–103. 5. Savell JW, Cross HR. The role of fat in the palatability of beef, pork, and lamb. In: Farrell DJ, editor. Designing Foods: Animal Product Options in the Marketplace. National Academy Press, Washington, DC; 1988. p. 345. 6. Johnsen PB, Civille GV. A standardized lexicon of meat WOF descriptors. Journal of Sensory Studies 1986;1:99–104. 7. Shahidi F. Lipid-derived ﬂavors in meat products. In: Kerry J, Ledward D, editors. Meat Processing: Improving Meat Quality. Woodhead Publishing Limited, Cambridge, UK; 2002. p. 105–21. 8. D’Souza DND, Mullan BP. The effect of genotype, sex and management strategy on the eating quality of pork. Meat Science 2002;60:95–101. 9. Romans JR, Tuma HJ, Tucker WL. Inﬂuence of carcass maturity and marbling on the physical and chemical characteristics of beef. Journal of Animal Science 1965;24:681–5. 10. Kipfmu¨ller H, Bodis K, Peschke W, Eichinger HM. Qualita¨t von Schweineﬂeisch. Erna¨hrung-Umschau 2000;47:416–22.

17. Liu Y, Qin X, Song XZH, Jiang H, Shen Y, Durbin KJ, et al. Bos taurus genome assembly. BMC Genomics 2009;10:180. 18. Zimin AV, Delcher AL, Florea L, Kelley DR, Schatz MC, Puiu D, et al. A whole-genome assembly of the domestic cow, Bos taurus. Genome Biology 2009;10(4):R42. 19. Elsik CG, Tellam RL, Worley KC, The Bovine Genome Sequencing and Analysis Consortium. The genome sequence of Taurine cattle: a window to ruminant biology and evolution. Science 2009;324:522–8. 20. Sonstegard TS, Keele JW, Harhay GP, Smith TPL, Matukumalli LK, Liu G, et al. Creation of a gene atlas in cattle using sequence-based transcriptional proﬁling. Journal of Animal Science 2007;85:415–415. 21. Williams JL. The use of marker assisted selection in animal breeding and biotechnology. Scientiﬁc and Technical Review, Ofﬁce International Des Epizooties 2005;24:379–91. 22. Andersson L, Georges M. Domestic animal genomics: deciphering the genetics of complex traits. Nature Reviews Genetics 2004;5:202–12. 23. Hocquette JF, Lehnert S, Barendse W, Cassar-Malek I, Picard B. Recent advances in cattle functional genomics and their application to beef quality. Animal 2007;1:159–73. 24. Gao Y, Zhang R, Hu X, Li N. Application of genomic technologies to the improvement of meat quality of farm animals. Meat Science 2007;77:36–45. 25. Allan MF, Smith TPL. Present and future applications of DNA technologies to improve beef production. Meat Science 2008;80:79–85. 26. Eggen A, Hocquette JF. Genomic approaches to economic trait loci and tissue expression proﬁling: application to muscle biochemistry and beef quality. Meat Science 2004;66:1–9.

11. Miller RK. Factors affecting the quality of raw meat. In: Kerry J, Kerry J, Ledward D, editors. Meat Processing: Improving Meat Quality. Woodhead Publishing Limited, Cambridge, UK; 2002. p. 27–63.

27. Ku¨hn C, Leveziel H, Renand G, Goldammer T, Schwerin M, Williams JL. Genetic markers for beef quality. In: Hocquette JF, Gigli S, editors. Indicators of Milk and Beef Quality. EAAP publication No. 112. Wageningen Academic Publishers, The Netherlands; 2005. p. 23–31.

12. Wheeler TL, Cundiff LV, Koch RM. Effect of marbling degree on beef palatability in Bos taurus and Bos indicus cattle. Journal of Animal Science 1994;72:3145–51.

28. Polineni P, Aragonda P, Xavier SR, Furuta R, Adelson DL. The bovine QTL viewer: a web accessible database of bovine quantitative trait loci. BMC Bioinformatics 2006;7:283.

13. Neeley TR, Lorenzen CL, Miller RK, Tatum JD, Wise JW, Taylor JF, et al. Beef customer satisfaction: role of cut, USDA quality grade, and city on in-home consumer ratings. Journal of Animal Science 1998;76:1027–32.

29. Stone RT, Keele JW, Shackelford SD, Kappes SM, Koohmaraie M. A primary screen of the bovine genome for quantitative trait loci affecting carcass and growth traits. Journal of Animal Science 1999;77:1379–84.

14. Thompson JM. The effects of marbling on ﬂavour and juiciness scores of cooked beef, after adjusting to a constant tenderness. Australian Journal of experimental Agriculture 2004;44:645–52.

30. Schimpf RJ, Winkelman-sim DC, Buchanan FC, Aalhus JL, Plante Y, Schmutz SM. QTL for marbling maps to cattle chromosome 2. In: Proceedings of the 27th International Conference on Animal Genetics, 22–26 July 2000, Minneapolis, USA; 2000. p. 48, abstract number B103.

15. Yamada T, Itoh M, Nishimura S, Taniguchi Y, Miyake T, Sasaki S, et al. Association of single nucleotide polymorphisms in the endothelial differentiation sphingolipid

g

31. MacNeil MD, Grosz MD. Genome-wide scans for QTL affecting carcass traits in Herefordcomposite double

CAB International 2009

30

Animal Science Reviews 2010 backcross populations. Journal of Animal Sciences 2002;80:2316–24.

32. Casas E, Stone RT, Keele JW, Shackleford SD, Kappes SM, Koohmaraie M. A comprehensive search for quantitative trait loci affecting growth and carcass composition of cattle segregating alternative forms of the myostatin gene. Journal of Animal Science 2001;79:854–60. 33. Casas E, Shackelford SD, Keele JW, Koohmaraie M, Smith TP, Stone RT. Detection of quantitative trait loci for growth and carcass composition in cattle. Journal of Animal Science 2003;81:2976–83. 34. Casas E, Keele JW, Shackelford SD, Koohmaraie M, Stone RT. Identiﬁcation of quantitative trait loci for growth and carcass composition in cattle. Animal Genetics 2004;35:2–6. 35. Mizoshita K, Mizoshita T, Watanabe T, Hayashi H, Kubota C, Yamakuchi H, et al. Quantitative trait loci analysis for growth and carcass traits in a half-sib family of purebred Japanese Black (Wagyu) cattle. Journal of Animal Sciences 2004;82(12):3415–20. 36. Mizoguchi Y, Watanabe T, Fujinaka K, Iwamoto E, Sugimoto Y. Mapping of quantitative trait loci for carcass traits in a Japanese black (Wagyu) cattle population. Animal Genetics 2006;37(1):51–4. 37. Kim JJ, Farnir F, Savell J, Taylor JF. Detection of quantitative trait loci for growth and beef carcass fatness traits in a cross between Bos taurus (Angus) and Bos indicus (Brahman) cattle. Journal of Animal Science 2003;81(8):1933–42. 38. Takasuga A, Watanabe T, Mizoguchi Y, Hirano T, Ihara N, Takano A, et al. Identiﬁcation of bovine QTL for growth and carcass traits in Japanese Black cattle by replication and identical-by-descent mapping. Mammalian Genome 2007;18(2):125–36. 39. Hirano T, Watanabe T, Inoue K, Sugimoto Y. Fine-mapping of a marbling trait to a 2.9-cM region on bovine chromosome 7 in Japanese black cattle. Animal Genetics 2008;39(1):79–83. 40. Imai K, Matsughige T, Watanabe T, Sugimoto Y, Ihara N. Mapping of a quantitative trait locus for beef marbling on bovine chromosome 9 in purebred Japanese black cattle. Animal Biotechnology 2007;18(2):75–80. 41. Moore SS, Li C, Basarab J, Snelling WM, Kneeland J, Murdoch B, et al. Fine mapping of quantitative trait loci and assessment of positional candidate genes for backfat on bovine chromosome 14 in a commercial line of Bos taurus. Journal of Animal Science 2003;81:1919–25. 42. Casas E, Shackleford SD, Keele JW, Stone RT, Kappes SM, Koohmaraie M. Quantitative trait loci affecting growth and carcass composition of cattle segregating alternate forms of myostatin. Journal of Animal Science 2000;78:560–9. 43. Page BT, Casas ER, Quaas L, Thallman RM, Wheeler TL, Shackelford SD, et al. Association of markers in the bovine CAPN1 gene with meat tenderness in large crossbred populations that sample inﬂuential industry sires. Journal of Animal Science 2004;82:3474–81. 44. Casas E, White SN, Wheeler TL, Shackelford SD, Koohmaraie M, Riley DG, et al. Effects of calpastatin and m-calpain markers in beef cattle on tenderness traits. Journal of Animal Science 2006;84:520–5. 45. Hamill RM, Reardon W, Aslan O, Sweeney T, Mullen AM. Association of single nucleotide polymorphisms in candidate

g

genes with meat quality in crossbred cattle. In: Book of Abstracts of the Third International Symposium on Animal Functional Genomics, 7–9 April 2008, Edinburgh, Scotland; 2008. p. 61. 46. Amills M, Vidal O, Varona L, Toma`s A, Gil M, Sa`nchez A, et al. Polymorphism of the pig 2,4-dienoyl CoA reductase 1 gene (DECR1) and its association with carcass and meat quality traits. Journal of Animal Science 2005;83:493–8. 47. Kambadur R, Sharma M, Smith TP, Bass JJ. Mutations in myostatin (GDF8) in double-muscled Belgian Blue and Piedmontese cattle. Genome Research 1997;7(9):910–5. 48. Nkrumah JD, Li C, Yu J, Hansen C, Keisler DH, Moore SS. Polymorphisms in the bovine leptin promoter associated with serum leptin concentration, growth, feed intake, feeding behaviour, and measures of carcass merit. Journal of Animal Science 2005;83:20–8. 49. Barendse W, Bunch RJ, Kijas JW, Thomas MB. The effect of genetic variation of the retinoic acid receptor-related orphan receptor C Gene on fatness in cattle. Genetics 2007;175:843–53. 50. Fitzsimmons CJ, Schmutz S, Bergen RD, McKinnon JJ. A potential association between the BM 1500 microsatellite and fat deposition in beef cattle. Mammalian Genome 1998;9:432–4. 51. Buchanan FC, Fitzsimmons CJ, Van Kessel AG, Thue TD, Winkelman-Sim DC, Schmutz SM. Association of a missense mutation in the bovine leptin gene with carcass fat content and leptin mRNA levels. Genetics Selection Evolution 2002;34:105–16. 52. Schenkel FS, Miller SP, Ye X, Moore SS, Nkrumah JD, Li C, et al. Association of single nucleotide polymorphisms in the leptin gene with carcass and meat quality traits of beef cattle. Journal of Animal Science 2005;83:2009–20. 53. Pannier L, Sweeney T, Hamill RM, Ipek F, Stapleton PC, Mullen AM. Lack of an association between single nucleotide polymorphisms in the bovine leptin gene and intramuscular fat in Bos taurus cattle. Meat Science 2009;81:731–7. 54. Lagonigro R, Wiener P, Pilla F, Woolliams JA, Williams JL. A new mutation in the coding region of the bovine leptin gene associated with feed intake. Animal Genetics 2003;34:371–4. 55. Barendse WJ. Assessing lipid metabolism. International Patent Publication WO 99/23248. 1999. 56. Michal JJ, Zhang ZW, Gaskins CT, Jiang Z. The bovine fatty acid binding protein 4 gene is signiﬁcantly associated with marbling and subcutaneous fat depth in WagyuLimousin F2 crosses. Animal Genetics 2006;37:400–2. 57. Pannier L, Stapleton P, Mullen AM, Hamill R, Sweeney T. Assessment of single nucleotide polymorphisms (SNPs) in genes on bovine chromosome 14 in association with intramuscular fat (IMF) in beef. In: Book of Abstracts of the 58th Annual Meeting of the European Association for Animal Production, 26–29 August 2007, Dublin, Ireland; 2007. p. 38. 58. Barendse W, Bunch R, Thomas M, Armitage S, Baud S, Donaldson N. The TG5 thyroglobulin gene test for a marbling quantitative trait loci evaluated in feedlot cattle. Australian Journal of Experimental Agriculture 2004;44:669–74.

CAB International 2009

L. Pannier, R. M. Hamill, A. M. Mullen and T. Sweeney 59. Marques E, Nkrumah JD, Sherman EL, Moore SS. Polymorphisms in positional candidate genes on BTA14 and BTA26 affect carcass quality in beef cattle. Journal of Animal Science 2009;87:2475–84. 60. Thaller G, Ku¨hn C, Winter A, Ewald G, Bellman O, Wegner J, et al. DGAT1, a new positional and functional candidate gene for intramuscular fat deposition in cattle. Animal Genetics 2003;34:354–7. 61. Casas E, White SN, Riley DG, Smith TPL, Brenneman RA, Olson TA, et al. Assessment of single nucleotide polymorphisms in genes residing on chromosome 14 and 29 for association with carcass composition traits in Bos indicus cattle. Journal of Animal Science 2005;83:13–19. 62. Winter A, Kra¨mer W, Werner FAO, Kollers S, Kata S, Durstewitz G, et al. Association of a lysine-232/alanine polymorphism in a bovine gene encoding acyl CoA:diacylglycerol acyltransferase (DGAT) with variation at a quantitative trait locus for milk fat content. Proceedings of the National Academy of Sciences, USA 2002;99:9300–5. 63. Rincker CB, Pyatt NA, Berger LL, Faulkner DB. Relationship among GeneSTAR marbling marker, intramuscular fat deposition, and expected progeny differences in early weaned Simmental steers. Journal of Animal Science 2006;84:686–93. 64. Gan QF, Zhang LP, Li JY, Hou GY, Li HD, Gao X, et al. Association analysis of thyroglobulin gene variants with carcass and meat quality traits in beef cattle. Journal of Applied Genetics 2008;49:251–5. 65. Roy R, Ordovas L, Zaragoza P, Romero A, Moreno C, Altarriba J, et al. Association of polymorphisms in the bovine FASN gene with milk-fat content. Animal Genetics 2006;37:215–8. 66. Abe T, Saburi J, Hasebe H, Nakagawa T, Misumi S, Nade T, et al. novel mutations of the FASN gene and their effect on fatty acid composition in Japanese Black beef. Biochemical Genetics 2009;47:397–411. 67. Zhang S, Knight TJ, Reecy JM, Beitz DC. DNA polymorphisms in bovine fatty acid synthase are associated with beef fatty acid composition. Animal Genetics 2008;39:62–70. 68. Morris CA, Cullen NG, Glass BC, Hyndman DL, Manley TR, Hickey SM, et al. Fatty acid synthase effects on bovine adipose fat and milk fat. Mammalian Genome 2007;18:64–74. 69. Chikuni K, Mitsuhashi T. Method of evaluating useful cattle. Patent Document CA 2441938. 2001. 70. Barendse W, Bunch RJ, Harrison BE, Thomas MB. The growth hormone GH1:c.457C>G mutation is associated with relative fat distribution in intra-muscular and rump fat in a large sample of Australian feedlot cattle. Animal Genetics 2006;37:211–4. 71. Taniguchi M, Utsugi T, Oyama K, Mannen H, Kobayashi M, Tanabe Y, et al. Genotype of stearoyl-coA desaturase is associated with fatty acid composition in Japanese Black cattle. Mammalian Genome 2004;15(2):142–8. 72. Zhang YY, Proenca R, Maffei M, Barone M, Leopold L, Friedman LM. Positional cloning of the mouse obese gene and its human homolog. Nature 1994;372:425–32. 73. Houseknecht KL, Baile CA, Matteri R, Spurlock ME. The biology of leptin: a review. Journal of Animal Science 1998;76:1405–20.

g

31

74. Stone RT, Kappes SM, Beattie C. Two polymorphic microsatellites within an 18 kb genomic clone containing the bovine ob gene. Animal Genetics 1996;27(2):64. 75. Haegeman A, Van Zeveren A, Peelman LJ. New mutation in the bovine leptin gene. Animal Genetics 2000;31:68–79. 76. Liefers SC, Veerkamp RF, te Pas MFW, Delavaud C, Chilliard Y, Platje M, et al. Leptin promoter mutations affect leptin levels and performance traits in dairy cows. Animal Genetics 2005;36:111–8. 77. Kononoff PJ, Deobald HM, Stewart EL, Laycock AD, Marquess FLS. The effect of a leptin single nucleotide polymorphism on quality grade, yield grade, and carcass weight of beef cattle. Journal of Animal Science 2005;83:927–32. 78. Madeja Z, Adamowicz T, Chmurzynska A, Jankowski T, Melonek J, Switonski M, et al. Short Communication: Effect of leptin gene polymorphisms on breeding value for milk production traits. Journal of Dairy Science 2004;87:3925–7. 79. Lusk JL. Association of single nucleotide polymorphisms in the leptin gene with body weight and backfat growth curve parameters for beef cattle. Journal of Animal Science 2007;85(8):1865–72. 80. Banos G, Woolliams JA, Woodward BW, Forbes AB, Coffey MP. Impact of single nucleotide polymorphisms in leptin, leptin receptor, growth hormone receptor, and diacylglycerol acyltransferase (DGAT1) gene loci on milk production, feed, and body energy traits of UK dairy cows. Journal of Dairy Science 2008;91:3190–200. 81. Barendse W, Bunch R, Thomas M, Armitage S, Baud S, Donaldson N. The TG5 thyroglobulin gene test for a marbling quantitative trait loci evaluated in feedlot cattle. Australian Journal of Experimental Agriculture 2004;44:669–74. 82. Barendse W, Bunch RJ, Harrison BE. The leptin C73T missense mutation is not associated with marbling and fatness traits in a large gene mapping experiment in Australian cattle. Animal Genetics 2005;36:86–8. 83. Anton I, Kova´cs K, Fe´su¨s L, Va´rhegyi J, Lehel L, Hajda Z, et al. Effect of DGAT1 and TG gene polymorphisms on intramuscular fat and on milk production traits in different cattle breeds in Hungary. Acta Veterinaria Hungarica 2008;56(2):181–6. 84. Cases S, Smith SJ, Zheng YW, Myers HM, Lear SR, Sande E, et al. Identiﬁcation of a gene encoding an acyl CoA: diacyl glycerol acyltransferase, a key enzyme in triacylglycerol synthesis. Proceedings of the National Academy of Sciences, USA 1998;95:13018–23. 85. Smith SJ, Cases S, Jensen DR, Chen HC, Sande E, Tow B, et al. Obesity resistance and multiple mechanisms of triglyceride synthesis in mice lacking DGAT. Nature Genetics 2000;25:87–90. 86. Grisart B, Coppieters W, Farnir F, Karim L, Ford C, Berzi P, et al. Positional candidate cloning of a QLT in dairy cattle: identiﬁcation of a missense mutation in the bovine DGAT gene with major effect on milk yield and composition. Genome research 2002;12:222–31. 87. Matarese V, Bernlohr DA. Puriﬁcation of murine adipocyte lipid-binding protein: characterisation as a fatty acid-and retinoic acid-binding protein. Journal of Biological Chemistry 1988;263:14544–51. 88. Gerbens F, Jansen A, Van Erp AJM, Harders F, Meuwissen THE, Rettenberger G, et al. The adipocyte fatty

CAB International 2009

32

Animal Science Reviews 2010 acid-binding protein locus: characterisation and association with intramuscular fat content in pigs. Mammalian Genome 1998;9:1022–6.

highly expressed in skeletal muscle. Journal of Biological Chemistry 1998;273:12443–50.

89. Wang YH, Byrne K, Reverter A, Harper G, Taniguchi M, McWilliam S, et al. Transcriptional proﬁling of skeletal muscle tissue from two breeds of cattle. Mammalian Genome 2005;15:201–10.

104. Xiao Q, Wu X, Michal J, Reeves J, Busboom J, Thorgaard G, et al. A novel nuclear-encoded mitochondrial poly(A) polymerase PAPD1 is a potential candidate gene for the extreme obesity related phenotypes in mammals. International Journal of Biological Sciences 2006;2:171–8.

90. Jurie C, Cassar-Malek I, Bonnet M, Leroux C, Bauchart D, Boulesteix P, et al. Adipocyte fatty acid-binding protein and mitochondrial enzyme activities in muscles as relevant indicators of marbling in cattle. Journal of Animal Science 2007;85(10):2660–9.

105. Michal JJ, Jiang Z, Xiao Q. Identifying animals having desirable intramuscular fat deposition by determining the presence of single nucleotide polymorphisms (SNPs) in a PAPD1 gene of the animal, which indicates intramuscular fat deposition. Patent WO2007134162-A2. 2008.

91. Cho S, Park TS, Yoon DH, Cheong HS, Namgoong S, Park BL, et al. Identiﬁcation of genetic polymorphisms in FABP3 and FABP4 and putative association with back fat thickness in Korean native cattle. BMB Reports 2008;41(1):29–34.

106. Gibbs RA, Taylor JF, Van Tassell CP, Barendse W, Eversole KA, Gill CA, The Bovine HapMap Consortium. Genome-wide survey of SNP variation uncovers the genetic structure of cattle breeds. Science 2009;324(5926):528–32.

92. Hoashi S, Ashida N, Ohsaki H, Utsugi T, Sasazaki S, Taniguchi M, et al. Genotype of bovine sterol regulatory element binding protein-1 (SREBP-1) is associated with fatty acid composition in Japanese Black cattle. Mammalian Genome 2007;18:880–6.

107. Van Tassell CP, Smith TP, Matukumalli LK, Taylor JF, Schnabel RD, Lawley CT, et al. SNP discovery and allele frequency estimation by deep sequencing of reduced representation libraries. Nature Methods 2008;5(3):247–52.

93. Shimano H. Sterol regulatory element-binding proteins (SREBPs), transcriptional regulators of lipid synthetic genes. Progress in Lipid Research 2001;40:439–52.

108. Feugang JM, Kaya A, Page GP, Chen L, Mehta T, Hirani K, et al. Two-stage genome-wide association study identiﬁes integrin beta 5 as having potential role in bull fertility. BMC Genomics 2009;10:10.

94. Roy R, Gautier M, Hayes H, Laurent P, Eggen A, Osta R, et al. Assignment of the fatty acid synthase (FASN) to bovine chromosome 19 (19q22) by in situ hybridisation and conﬁrmation by hybrid somatic cell mapping. Cytogenetics and Cell Genetics 2001;93:141–2.

109. Kim ES, Berger PJ, Kirkpatrick BW. Genome-wide scan for bovine twinning rate QTL using linkage disequilibrium. Animal Genetics 2009;40(3):300–7.

95. Wakil SJ, Stoops JK, Joshi VC. Fatty acid synthesis and its regulation. Annual Review of Biochemistry 1983;52:537–79.

110. Hayes BJ, Bowman PJ, Chamberlain AJ, Goddard ME. Invited review: Genomic selection in dairy cattle: progress and challenges. Journal of Dairy Science 2009;92(2):433–43.

96. Evans RM. The steroid and thyroid hormone receptor superfamily. Science 1988;240:889–95.

111. Meuwissen THE, Hayes BJ, Goddard ME. Prediction of total genetic value using genome-wide dense marker maps. Genetics 2001;157:1819–29.

97. Petkovich M, Brand NJ, Krust A, Chambon P. A human retinoic acid receptor which belongs to the family of nuclear receptors. Nature 1987;330:444–50.

112. Goddard ME, Hayes BJ. Mapping genes for complex traits in domestic animals and their use in breeding programmes. Nature Reviews Genetics 2009;10(6):381–91.

98. Hirose T, Smith RJ, Jetten AM. ROR-gamma, the 3rd member of ROR-RZR orphan receptor subfamily that is highly expressed in skeletal-muscle. Biochemical and Biophysical Research Communications 1994;205:1976–83.

113. Wiggans GR, Sonstegard TS, VanRaden PM, Matukumalli LK, Schnabel RD, Taylor JF, et al. Selection of single-nucleotide polymorphisms and quality of genotypes used in genomic evaluation of dairy cattle in the United States and Canada. Dairy Science 2009;92:3431–6.

99. Andersson-Eklund L, Danell B, Rendel J. Associations between blood-groups, blood protein polymorphisms and breeding values for production traits in Swedish red and white dairy bulls. Animal Genetics 1990;21:361–76.

114. VanRaden PM, Sullivan PG. National and international genomic evaluations for dairy cattle. Journal of Dairy Science 2009;92(1):175.

100. Barendse W. DNA markers for marbling. Patent Publication WO2004070055, 1–53. 2003. 101. Yamada T, Taniguchi Y, Nishimura S, Yoshioka S, Takasuga A, Sugimoto Y, et al. Radiation hybrid mapping of genes showing intramuscular fat deposition-associated expression changes in bovine musculus longissimus muscle. Animal Genetics 2006;37:184–5. 102. Nakanishi S, Catt KJ, Balla T. A wortmannin-sensitive phosphatidylinositol 4-kinase that regulates hormone-sensitive pools of inositolphospholipids. Proceedings of the National Academy of Sciences, USA 1995;92:5317–21. 103. Thornton C, Snowden MA, Carling D. Identiﬁcation of a novel AMP-activated protein kinase beta subunit isoform that is

g

115. Daetwyler HD, Schenkel FS, Sargolzaei M, Robinson JA. A genome scan to detect quantitative trait loci for economically important traits in Holstein cattle using two methods and a dense single nucleotide polymorphism map. Journal of Dairy Science 2008;91(8):3225–6. 116. Berry DP, Kearney JF, Harris BL. Genomic selection in Ireland. Interbull Bulletin 2009;39:29–34. 117. Davoli R, Zambonelli P, Bigi D, Fontanesi L, Russo V. Analysis of expressed sequence tags of porcine skeletal muscle. Gene 1999;233:181–8. 118. Hocquette JF. Where we are in genomics? Journal of Physiology and Pharmacology 2005;56(3):37–70. 119. Bahar B, Monaghan FJ, Moloney AP, Schmidt O, McHugh DE, Sweeney T. Long-term stability of RNA in post-mortem

CAB International 2009

L. Pannier, R. M. Hamill, A. M. Mullen and T. Sweeney bovine skeletal muscle, liver and subcutaneous adipose tissues. BMC Molecular Biology 2007;8:108.

33

tissue in Brahman steers during nutritional restriction. Journal of Animal Science 2005;83:1–2.

120. Bonnet M, Faulconnier Y, Leroux C, Jurie C, Cassar-Malek I, Bauchart D, et al. Glucose-6-phosphate dehydrogenase and leptin are related to marbling differences among Limousin and Angus or Japanese BlackAngus steers. Journal of Animal Science 2007;85:2882–94.

133. Reverter A, Byrne KA, Bruce HL, Wang YH, Dalrymple BP, Lehnert SA. A mixture model-based cluster analysis of cDNA microarray gene expression data on Brahman and Brahman composite steers fed high, medium and low quality diets. Journal of Animal Science 2003;81:1900–10.

121. Brandstetter AM, Sauerwein H, Veerkamp JH, Geay Y, Hocquette JF. Effects of muscle type, castration, age and growth rate on H-FABP expression in bovine skeletal muscle. Livestock Production Science 2002;75(2):199–208.

134. Wang YH, Byrne K, Reverter A, Harper G, Taniguchi M, McWilliam S, et al. Transcriptional proﬁling of skeletal muscle tissue from two breeds of cattle. Mammalian Genome 2005;15:201–10.

122. Wan J, Wright MB, Cai L, Flament A, Lindpaintner K. Efﬁcacy of SSH PCR in isolating differentially expressed genes. BMC Genomics 2002;3:12.

135. Lehnert S, Reverter A, Byrne KA, Wang YH, Nattrass GS, Hudson NJ, et al. Gene expression studies of developing bovine longissimus muscle from two different beef cattle breeds. BMC Development Biology 2007;7:95.

123. Mohan M, Hurst AG, Malayer JR. Global gene expression analysis comparing bovine blastocysts ﬂushed on day 7 or produced in vitro. Molecular Reproduction and Development 2004;68:288–98. 124. Liang P, Pardee AB. Differential display of eukaryotic mRNA by means of the polymerase chain reaction. Science 1992;257:967–71. 125. Davis Jr W, De Sousa PA, Schultz RM. Transient expression of translation initiation factor eIF-4C during the 2-cell stage of the preimplantation mouse embryo: identiﬁcation by mRNA differential display and the role of DNA replication in zygotic gene activation. Development Biology 1996;174:190–201. 126. Graff JC, Behnke M, Radke J, White M, Jutila MA. A comprehensive SAGE database for the analysis of {gamma} {delta} T cells. International Immunology 2006;18:613–26. 127. Schena M, Shalon D, Davis RW, Brown PO. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 1995;270:467–70. 128. Moody DE. Genomics techniques: an overview of methods for study of gene expression. Journal of Animal Science 2007;79:E128–35. 129. Rinaudo P, Schultz RM. Effects of embryo culture on global pattern of gene expression in preimplantation mouse embryos. Reproduction 2004;128:301–11. 130. Shendure J, Ji HL. Next-generation DNA sequencing. Nature Biotechnology 2008;26(10):1135–45. 131. Le Be´chec A, Zindy P, Sierocinski T, Petritis D, Bihoue´e A, Le Meur N, et al. M@IA: a modular open-source application for microarray workﬂow and integrative datamining. In Silico Biology 2008;8(1):63–9. 132. Byrne KA, Wang YH, Lehnert SA, Harper GS, McWilliam SM, Bruce HL, et al. Gene expression proﬁling of muscle

g

136. Wang YH, Bower NI, Reverter A, Tan SH, De Jager N, Wang R, et al. Gene expression patterns during intramuscular fat development in cattle. Journal of Animal Science 2009;87(1):119–30. 137. Yu SL, Chung HJ, Sang BC, Park CS, Lee JH, Yoon DH, et al. Identiﬁcation of differentially expressed genes in distinct skeletal muscles in cattle using cDNA microarray. Animal Biotechnology 2007;18(4):275–85. 138. Pannier L. A search for DNA polymorphisms and signature gene expression proﬁles associated with meat quality [PhD thesis]. University College Dublin; 2008. 139. Sadkowski T, Jank M, Oprzadek J, Motyl T. Age-dependent changes in bovine skeletal muscle transcriptomic proﬁle. Journal of Physiology and Pharmacology 2006;57(7):95–110. 140. Sadkowski T, Jank M, Zwierzchowski L, Oprzadek J, Motyl T. Transcriptomic index of skeletal muscle of beef breeds bulls. Journal of Physiology and Pharmacology 2009;60(1):15–28. 141. Sadkowski T, Jank M, Zwierzchowski L, Oprzadek J, Motyl T. Gene expression proﬁling in skeletal muscle of Holstein–Friesian bulls with single-nucleotide polymorphism 0 in the myostatin gene 5 -ﬂanking region. Journal of Applied Genetics 2008;49(3):237–50. 142. Cassar-Malek I, Passelaigue F, Bernard C, Leger J, Hocquette JF. Target genes of myostatin loss-of-function in muscle of late bovine fetuses. BMC Genomics 2007;8:63. 143. Bernard C, Cassar-Malek I, Le Cunff M, Dubroeucq H, Renand G, Hocquette JF. New indicators of beef sensory quality revealed by expression of speciﬁc genes. Journal of Agricultural and Food Chemistry 2007;55:5229–37. 144. Lehnert SA, Wang YH, Byrne KA. Development and application of a bovine cDNA microarray for expression proﬁling of muscle and adipose tissue. Australian Journal Experimental Agriculture 2004;44:1127–33.

CAB International 2009

This page intentionally left blank

Animal Science Reviews 2010

Assessing impacts of organic production on pork and beef quality Albert Sundrum* Address: Department of Animal Nutrition and Animal Health, University of Kassel, D-37213 Witzenhausen, Germany. *Correspondence: Email: [email protected] 21 November 2008 29 April 2009

Received: Accepted:

Abstract Organic livestock farming is based on a low-input production method, aiming to provide meat of a high product and process quality rather than maximizing production. Such high meat quality corresponds to the expectations of consumers, who are both seeking a premium product and who are willing to pay premium prices. This review focuses on the question of whether organic pork and beef production currently meet consumer demands, and it elaborates the potentials and limitations for producing high-quality meats under the current organic framework. Although deﬁned by speciﬁc and basic guidelines, organic livestock production is characterized by largely heterogeneous farming conditions that allow for huge differences in the availability of nutrient resources, the implementation of feeding regimes, the use of genotypes, etc., all of which variously impact meat production. Correspondingly, there is substantial variation in the quality of organic meat entering the marketplace. According to the available literature, the quality of organic pork and beef was inconsistent and often fell short of expectations, as it is often similar to the quality of conventionally produced meat. Obviously, organic guidelines play a minor role with respect to meat quality. On the other hand, there is evidence that less intensiﬁed production systems may present suitable preconditions that would yield meats of high eating quality assumed that the relevant pre- and post-farmgate factors are considered and balanced within quality assurance systems. However, the trade value of carcasses is determined primarily by lean meat and cut composition while qualitative traits of meat are of no relevance. Neither does continuous recording of qualitative traits take place nor is an outstanding meat quality rewarded by higher prices. Of the various parameters, the intramuscular fat is highly related to palatability traits in both pork and beef and could be used to distinguish between different levels of eating quality. Despite the fact that many consumers express their wish for high-quality meats, the payment and marketing systems counteract all efforts to follow consumer demands and fail to communicate adequately differentiated meat palatability. It is concluded that only a direct assessment of qualitative traits and a payment system that honours meat quality grades that are above average can contribute so as to improve the currently unsatisfying situation. This, however, requires a shift in the paradigm from a guideline-oriented to an output-oriented approach, and the implementation of a systematic approach for an effective and efﬁcient balancing of the multiple variables and complex interactions within each farming system. Keywords: Eating quality, Production method, Output orientation, System approach, Marketing

Introduction In the past, a lot of research effort has been devoted to meat quality. Nevertheless, problems with regard to the quality of meat are still evident in conventional production

g

and this explains why consumers are looking for alternatives [1]. The introduction of the ‘wholesomeness’ concept in meat production, most often represented by organic production, is mainly the result of a wish to reestablish a positive meat sector image, including meat

CAB International 2009

36

Animal Science Reviews 2010

quality, meat safety and animal welfare aspects [2, 3]. Consumers make a whole range of positive inferences from the label ‘organic’; referring not only to a better taste but also to the concern for the environment and health [4]. The holistic approach of organic farming is also in line with the growing wish of consumers for improved traceability, including the ever-rising consumer concern in relation to the introduction of genetically modiﬁed organisms in the feed and food production [5]. In Germany, guidelines are a characteristic feature of organic farming since 1954 and have been so since the country’s trademark legislation required clear criteria to identify organically produced goods [6]. Because the variety of production sites and the resulting product properties did not allow their identiﬁcation to be linked to products qualitatively in terms that could be described exactly and understood analytically, the production method itself became the identifying criterion. This fundamental principle of the original privately developed economic approach has been adopted in many countries by legislation to harmonize the rules of organic farming and to make all organic systems across countries subject to minimum standards. In Europe, the legal basis is provided by the revised Council Regulation No. 834/2007. One of the main objectives is to protect consumers from unjustiﬁed claims and to avoid unfair competition between those who label their products as being organic. The organic concept refers to the whole farm as the base of a comprehensive system where the production process is intended to ensure quality production rather than maximizing production, therewith meeting the demands of an increasing number of consumers, who are more and more critical of conventional production methods. Labelling meat as being ‘organic’ identiﬁes the products as deriving from a production method deﬁned by guidelines. So far, only a few studies have been conducted to assess the quality of pork and beef produced under these conditions. In this paper, some major ﬁndings, with a particular emphasis placed upon the eating quality of meat, are reviewed. The possibilities and limitations of both the production of high-quality meats and their assessment are outlined below.

Meat Quality, the Outcome of Complex Interactions Meat quality is multifaceted and comprises sensory, nutritive, hygienic or toxicological and technological factors [7]. Factors such as colour, ﬂavour, softness and tenderness are considered to be primary sensory factors. Nutritive factors in meat include the type and amount of protein, carbohydrate, lipid, vitamins and minerals. Hygiene and toxicological factors of beef and pork are outside the scope of this review and the reader is referred to corresponding reviews [8, 9]. Technological factors include the intrinsic properties of meat that govern its

g

suitability for manufacturing. Sensory factors are the major focus of this review, with passing reference being made to nutritive and technological factors. The intrinsic properties (i.e. those extant at the time of slaughter) of an animal’s muscles inﬂuence the eatingquality attributes of meat. Some of these intrinsic characteristics of muscle are changed little by processing and cooking, while others interact with pre-slaughter stress, post-slaughter processing conditions and method and time of storage to subsequently inﬂuence eating quality of meat [10, 11]. The conditions that apply in the 24–48 h immediately before and after slaughter are recognized as having the largest inﬂuence on pork and beef palatability [1, 12]. Thus, failure to control the pre- and postslaughter environment can have a particularly large impact on meat quality attributes. As there is no evidence for grave differences in post-farmgate factors such as preslaughter handling, transport, slaughter or chilling rate between organic and conventional production, the review is primarily focused on the effects of on-farm factors on meat traits. The signiﬁcant factors involved in generating high levels of meat quality have been in the focus of animal science for decades. However, most of the present knowledge is based on studies investigating the inﬂuence of a single or of very few factors pertaining to meat quality under ceteris paribus assumptions. Thus, there are limited data available concerning the understanding of how the various factors within the production chain interact in relation to the qualitative traits of meat [1, 13]. Under the current marketing conditions, carcasses with the leanest meat generate the highest proﬁt. At the same time it is the most efﬁcient strategy to reduce production costs as an increase in protein accretion goes along with a clear increase in daily weight gain and an improvement in feed conversion [14]. A lean carcass also provides the highest yield and proﬁt for the processing industry. In responding to market pressures and attracted by proﬁt taking, improved nutritional practices in combination with genetic selection have been implemented and have improved the leanness of pork and beef considerably within the recent decades [1]. However, striving for a high lean-meat percentage has provided several negative side effects. With an increase of lean meat and an extreme reduction of overlay fat, the risk increases that meat deﬁciencies will appear [15, 16]. In the carcasses of pigs, lean meat percentage and intramuscular fat (IMF) content in longissimus muscles are correlated negatively [17, 18]. Because of the substantial efforts to increase carcass lean content, an appropriate IMF content that would be associated with a high level of eating quality has almost been eliminated [19, 20]. Compared with highly fat-marbled pork chops, leaner chops are darker, have a more acceptable appearance and are more likely to be purchased [21]. However, they are also less tender, juicy, oily and ﬂavoursome than highly fatmarbled chops, indicating a disparity between purchase

CAB International 2009

Albert Sundrum

37

Table 1 Differences between organic and conventional livestock production in various aspects Council Regulation (European Commission) No 834/2007 on organic production Principles Management strategy Genotypes Housing conditions Nutritional potential Feeding regime Processing of products

Conventional

Quality production rather than maximizing production Maximizing efﬁciency within the whole farm system Breeding animals of organic origin; high variety of breeds between farms, often comparable to conventional genotypes but partly indigenous and robust breeds Special rules according to space allowance, lying and outdoor area as well as pasture; ban of tethering and fully slatted ﬂoors Limited availability of energy-rich feedstuffs for ruminants and high quality protein for monogastric animals; restrictions in quantity and quality of bought-in feedstuffs, ban of extracted meals and synthetic amino acids Prolonged weaning; restrictions in the use of concentrate for ruminants; pigs have to be provided with roughage Use of deﬁned products and substances in processing

intent based on visual evaluation, and quality or palatability attributes of the cooked product. While high glycogen levels at slaughter bears the risk of a low ultimate meat pH and consequently colour and drip loss, lower-than-normal muscle glycogen stores at the time of slaughter include the risk of dark cutting meat, because of a low lactic acid production after slaughter and a high ultimate meat pH [22, 23]. As ultimate pH is directly related to tenderness, a reduction in muscle glycogen level obtained through selection will indirectly increase tenderness in pork. On the other hand, glycogen showed a positive correlation to meat percentage and was negatively correlated to the feed conversion ratio (kg feed : kg gain) when studied in pure-breds [24]. Thus, selecting pigs for low glycogen levels may be in contradiction to the genetic improvement of performance traits. In beef production, fattening bulls are preferred in most European countries because of their higher economical efﬁciency. At the same time, bulls provide poorer preconditions for high eating quality compared with the meat of steers and heifers [25]. While the fattening of steers has nearly vanished in many countries, heifers are seldom used to produce quality meat. Based on these preconditions, beef production in Europe is more oriented to a quantitative rather than a qualitative production. Many beef-producing countries have attempted to utilize the potential of meaty, large and late-maturing beef breeds. On the other hand, beef cattle with a high lean-meat percentage such as White Blue Belgians showed a clearly lower IMF content compared with beef from dairy breeds [26]. Steers and heifers of small- and medium-sized breeds are well suited for extensive production, while bulls of larger breeds are not subjected to as much fattening, as a prolonged fattening period is expected to have a negative impact on the tenderness of meat [27]. This effect, however, has not been found in steers and heifers [28].

g

Minimizing production costs Maximizing productivity at the level of the farm animals High-yielding pure- or cross-bred, often bred for a speciﬁc production goal No prescriptions apart from the general regulations, primarily indoor housing No prescriptions apart from the general recommendations of good farming practice Good farming practice Good manufacturing practice

When offered a low feeding level, the IMF of heifers and steers was higher (4.8 and 3.4%, respectively) compared with a high feeding level (3.5 and 3.0%, respectively) [29]. At a high feeding level, the meat of the bulls contained less IMF (2.3%) than that of heifers and steers. Last but not least, while most of the consumers would prefer to receive meat that is tender rather than tough, it is preferable for certain manufactured products to rely on meat that does not disintegrate too easily [30]. If the fat content can vary within an established grade then this variance would not be helpful for processors wanting to produce a consistent and speciﬁed low-fat product.

Characteristics of Organic Livestock Production In recent years, organic agriculture has shown rapid growth and dynamic developments in many countries, but especially throughout Europe [31]. Based on worldwideapplied guidelines, organic livestock farming has set goals to establish environmentally friendly production methods, to sustain animals in good health, to achieve high animalwelfare standards and to produce products of high quality [32]. In the European countries, the EEC Regulation No. 834/2007 stipulates the maximum number of farm animals/ha in pasture and the maximum organic livestock numbers/m2 in covered and uncovered housing. Furthermore, these regulations include speciﬁcations for housing conditions, animal nutrition, care and breeding, disease prevention and veterinary treatment, while they create frameworks for organic livestock production and product labelling in all European countries on an equal and legal basis. Differences between organic and conventional conditions for livestock production are outlined in Table 1. For a producer or a farmer, an important principle is to rely mainly on the management of internal farm resources

CAB International 2009

38

Animal Science Reviews 2010

rather than on external input and, in relation to health management, to rely on prevention measures rather than on treatment. While in conventional production increasing demands for nutrients are covered mainly by broughtin feedstuffs, in organic production the possibilities for nutrient imports are limited. According to the new EU Regulation (EEC/834/2007), organic livestock farming intends to contribute to the equilibrium of agricultural production systems, to establish and maintain interdependence between soils, plants and animals; to establish land-related and rule out landless productions, and therewith support the development of a sustainable agriculture. The realization of this approach usually requires a complete re-organization of the farm, in which crop requirements have to be tailored to the concerns of animal husbandry and the extent and direction of animal husbandry adapted to home-grown feedstuffs [33]. Dealing with a limited availability of resources is therefore a main feature of organic production, while maximization of protein accretion is of less importance. However, organic agriculture is not organized uniformly, either with respect to the various objectives or in the degree of their implementation. There are different perspectives on organic agriculture with different understandings of what it is and what it develops. Not only farmers but also many different entities, such as policymakers, small agri-businesses, agri-food corporations, supermarkets and consumers, seek to inﬂuence the production process in accordance with their own goals and values. Alroe and Noel [34] identiﬁed three signiﬁcant perspectives on organic agriculture based on protest, meaning and market. These three perspectives on organic agriculture cannot be merged into one but can be helpful in understanding of its past and of its future development. Although organic meat production is clearly demarcated from conventional production by organic guidelines, framework conditions differ markedly between organic production systems [35]. The production conditions vary from outdoor production to indoor production and encompass huge differences between regions, not only in relation to environmental conditions and the genotype used but also especially in relation to the nutrient supply. While common ideas and guidelines exist only on a meta-level, the concrete implementation in farm practice results in a large variety of stocking densities, performance levels, nutrient ﬂows and efﬁciencies in the use of home-grown and bought-in feedstuffs. Thus, highly intensiﬁed conventional livestock systems and organic systems mark the cornerstones of opposite approaches, providing divergent implications for the understanding and the development of livestock systems. In farming practice, both approaches are often mixed. With regard to the different objectives, different priorities and different framework conditions, it has, however, to be taken into account that organic and conventional livestock productions belong to completely different farm systems. General conclusions derived from conventional

g

production systems are not directly compatible with organic livestock production and are therefore invalid.

Meat Quality in Organic Farming When organic meat is offered, those choosing the organic variety expect it to be of better quality across all quality dimensions, including taste and tenderness. The expectations focused on an alternative production method might be reasonable as the quality of conventional produced meat has been of concern for a long time. Several authors highlight that the limit of leanness that will still allow one to market palatable pork and beef has already been exceeded [36–38]. According to Andersen [19], the intensity of pork ﬂavour has decreased during recent years, probably as a result of the production of pork with a minimal content of IMF. Slaughtered pigs from intensiﬁed production units currently showed an average IMF content in the longissimus muscle around 1%, clearly below the threshold of 2.5%, having a resounding effect on palatability [39, 40]. Compared with conventional meat production, little information is available on the current relation of meat palatability attributes to technological quality for organic production. In a ﬁeld study, considerably higher variations in the lean meat content of carcasses have been found in organic pigs compared with conventionally produced pigs [41]. In a study by Jonsa¨ll et al. [42], loins from organically raised pigs were less juicy than those conventionally raised. Olsson et al. [43] found that the meat of a speciﬁc organic treatment was of poorer quality (higher drip loss and increased shear force values) compared with meat from the conventional treatment. A recently conducted ﬁeld study on organic farms in Germany and Austria re-vealed that the average IMF content in longissimus muscle ranged on a level of about 2.2%, at the same time showing a substantial variation within and between farms [44]. Different experiments were carried out to study the impacts of different organic feeding regimes on meatquality traits. Unbalanced organic diets enriched with grain legumes in the growing and ﬁnishing period led to an average IMF content of 2.8% in the longissimus muscle of pigs, compared with 1.2% in the conventional treatment [18, 45]. Also Millet et al. [46] found a higher IMF content associated with organic diets, although on a lower level. In addition, ultimate pH in ham and loin was lower and the meat was redder compared with the conventional treatment. In contrast, Millet et al. [47] did not ﬁnd an increased level of IMF when feeding diets consisting of organic feed ingredients. Also Hansen et al. [48] found no signiﬁcant difference in meat-quality characteristics when compared with results obtained in the conventional system. While the percentage of leanness increased in meat from organic pigs raised on 70% concentrate plus roughage compared with meat from pigs given 100%

CAB International 2009

Albert Sundrum

concentrate, restricted concentrate feeding gave rise to a decrease in tenderness compared with pork from pigs fed 100% concentrate. In a study by Partanen et al. [49], pigs fed diets with ingredients of organic origin had fatter carcasses, but the eating quality of organic pork did not differ from that of pork from pigs fed conventional diets. Heyer et al. [50] compared carcass and technological quality traits of indoor and outdoor raised pigs fed organic or conventional diets ad libitum or restrictively. Technologically based meat quality differed only slightly between indoor- and outdoor-raised pigs. The diet (organic/ conventional) for indoor pigs did not affect carcass and meat-quality traits. When comparing organic and conventional beef production systems in the USA (including ionophore implants), steers in organic ﬁnishing had signiﬁcantly lighter carcasses, smaller rib eyes and higher marbling than steers in conventional ﬁnishing [51]. In Europe, grazing in summer and feeding a minimum of 60% roughage during the whole year are inevitable components of an organic beef production system. According to Nielsen and Thamsborg [52], quality of organic beef in Denmark is affected by the production system used, especially differences in grazing and exercise. They assume that grazing and exercise may affect the eating quality by resulting in darker meat colour, lower risk of off-ﬂavour, yellow fat and a higher content of unsaturated fatty acids, including conjugated linoleic acid. Nevertheless, the authors conclude that the overall effect on sensory attributes may be of minor importance. According to Russo and Preziuso [53], carcasses from organically raised beef cattle in Italy were characterized by poor muscular development and low IMF content, while other organoleptic characteristics of the meat were not inﬂuenced by the organic rearing system. They concluded that the results were probably associated with the fact that native breeds were preferred, which are often characterized by a rather slow development, and the rations were mainly based on forage, with low energy yields from concentrates. Summing up, it can be assumed that the variations with respect to the eating quality of pork and beef in organic farming are caused by differences in the genotypes and the feeding regime used, and differences in the adaptation of the nutrient supply to the genetic protein and fat accretion of the animals in the various stages of life.

How to Improve Meat Quality In intensiﬁed livestock production systems, the procedures and measures used to achieve a certain product are highly standardized in order to obtain a more or less uniform and interchangeable product. This standardization includes the use of comparable genotypes and feeding strategies. In contrast, organic farmers are challenged by more heterogeneous preconditions.

g

39

Selection of Genotype and Sex Genetics are known to have a marked inﬂuence on both production and quality traits. The capacity for protein accretion is determined principally by the genotype. Meatrich genotypes achieved clearly higher protein accretion over the entire fattening period than those with an average muscle growth [54]. The superiority of animals with a high genetic capacity for protein accretion becomes manifest, however, only in combination with a corresponding protein and energy supply. In general, genotypes used in organic pig production do not differ markedly between organic and conventional production, except for some rare or robust breeds used on single farms [41]. Robust breeds, especially Iberian pigs, are known for their potential to emerge as high eating quality of pork. Muscles from pure-bred Iberian pigs contain signiﬁcantly higher amounts of IMF, haem pigments and iron than those from cross-bred pigs [55]. The use of other robust breeds, however, does not automatically provide clear advantages with respect to pork quality traits [17, 45]. Although an enormous amount of research has been directed at the pale, soft and exudative (PSE) problem in pig meat, and the genetic basis for an understanding of the impacts of the gene defects are known [56, 57], only a few countries have reacted accordingly [58]. In many countries, little progress to eliminate the gene defect or to minimize the problem is recorded [59]. In Germany, currently only few progeny-tested Pietrain boars have the homozygous positive genotype nn in relation to the malignant hyperthermia syndrome (MHS) genotype, but more than 50% of the boars are heterozygous gene carriers (Nn) [60]. The gene defect results in higher than normal muscle glycogen stores and in extended postmortem pH decline that leads to pork with a lower-thannormal ultimate meat pH, reduced water-holding capacities and greater cooking losses [61, 62]. If the main gene defects are excluded, genetics contributes less than 30% to the total variation of meat quality criteria [63]. Hence, most attributes referring to the eating quality of meat show a low heritability, except for that of IMF content [13]. Although the heritability of IMF and fat tissue is high, the genetic correlation between both is very low [64], which suggest that selection for high IMF in lean carcasses should be possible. A successful implementation of a breeding programme focusing on IMF content in pork has been reported from Switzerland [65]. Moreover, there has been considerable interest in the use of genotypes (especially Duroc) with high levels of IMF in an attempt to improve eating quality with variable results. While Edwards et al. [66] found no signiﬁcant advantage for the Duroc breed in eating quality, others highlighted the beneﬁt of this breed in several, but not all, quality traits [45, 67]. In the past, the favoured breeds with respect to beef quality were the small-framed and early maturing beef breeds such as Aberdeen Angus and Hereford, and the

CAB International 2009

40

Animal Science Reviews 2010

small-framed dairy breed Jersey [68]. In many countries, early maturing breeds have been nearly squeezed out by large-framed beef cattle. Meat of local rustic breeds that are still extant in some regions offers a certain attractiveness to consumers. In general, rustic breeds show the reddest meat, which is directly related to their higher content in muscular haem pigments [69]. In studies that compared sensory characteristics of beef between different rustic breeds in Spain and Germany, breeds were signiﬁcantly different from each other according to sensory attributes, indicating that some, although not all, rare breeds are providing high eating qualities [70, 71]. Furthermore, sensory analysis indicated that meat from the rustic breed may require a longer ageing period than meat from an improved cross-breed [72]. Data from Lively et al. [73] showed that dairy genotypes produce more tender meat than beef genotypes. Bulls selected for their high-phenotypic ultrasound IMF percentage and their fat-marbling score can be expected to produce steer progeny with signiﬁcantly higher amounts of marbling and quality [74, 75]. It also appears that marbling can be increased without a corresponding increase in external fat thickness. Meat from steers and heifers is expected to be tenderer compared with meat from bulls, which can be related to a higher IMF content and to a different rate of proteolysis post mortem [25, 76]. In studies of Maher et al. [77], bulls were more variable than steers for ultimate pH, Warner Bratzler shear force, ﬂavour and cooking loss. It has, however, to be taken into account that muscle type accounts for a greater proportion of the variation in the muscle characteristics such as ﬁbre crosssectional area, percentage of ﬁbre types, total and insoluble collagen and lipid concentration than breed and sex of the animals [78].

Feeding Regime Because of limitations in the availability of home-grown feedstuffs and restrictions concerning bought-in feedstuffs, the ﬁrst limitation in exploiting the genetic potential of growing pigs in organic farming is the supply of limited amino acids. As soybean meal from conventional processing and synthetic amino acids are banned, grain legumes represent the main protein source. Correspondingly, the performance capacity of pigs under organic conditions is diminished compared with conventional production. The preferred use of home-grown versus brought-in feedstuffs can cause a high variation in the nutrient content of feeding rations and this has consequences in carcass composition and meat quality. Possibilities and limitations of protein supply in organic pig production have been recently reviewed [79]. A number of feeding strategies have been proven to be potential control tools in the production of superior quality meat. According to Rosenvold et al. [80], the level

g

of macroglycogen before slaughter can be changed by dietary manipulation. A reduction in glycolytic potential results in an increase of ultimate pH and decreases the temperature 45 min post mortem, which will lead to reduced drip loss. According to Cisneros et al. [81], Sundrum et al. [18] and D’Souza et al. [82], the IMF content in the Musculus longissimus dorsi can be increased by feeding livestock diets low in essential amino acids without increasing fat area and back-fat thickness. Nutritional effects on IMF characteristics and content have shown to be clearly of higher importance than genetic effects [45, 83]. Under standardized conditions, a diet enriched with grain legumes in the growing and ﬁnishing periods led to an average IMF content of 2.8% in the longissimus muscle of pigs [45]. The results supported the hypothesis previously developed by Katsumata et al. [84] that a reduced supply of essential amino acids could provoke the de novo synthesis of fat in the muscle cell and therefore lead to a higher IMF content. Oksbjerg et al. [85] studied the inﬂuence of feeding strategies that combined outdoor and indoor rearing on pig meat quality. They conclude that ad libitum feeding in organic production systems gives superior meat quality compared with a restrictive feeding strategy. Variations in nutrient supply result in both varying growth processes within the different stages of life and varying ages at which each animal are slaughtered. These are expected to have an additional effect on IMF content [86]. For high variations in feed intake and growth rate, as well as in protein and fat accretion, differences in nutrient supply and digestibility [87], age [88], living conditions and environmental temperature [89], stocking rate and group size [90], rank order [91], feeding environment [92, 93] and generally stressful situations [94] must be considered. This makes a prediction of the speciﬁc requirements at the farm level very complex and difﬁcult. In organic farming, beef is primarily produced on a grass-based system consisting of a period of summer pasture feeding followed by an indoor, ﬁnishing period of grass silage feeding. Because of a lack of availability of energy-rich forage such as maize silage and restrictions in the use of concentrate, farmers are challenged to provide an appropriate energy supply. Tenderness in muscles with a high content of connective tissue decreases with age [27]. Prolonged grazing may decrease quality grade, either by impairing the ability of the animal to deposit IMF or by decreasing the time during which dietary energy supply is adequate for IMF deposition to occur [95]. In contrast, Scheeder et al. [28] found no relationship between an increased age of steers and heifers kept on grassland and the tenderness of their meat. The greatest potential for the manipulation of IMF accretion during fattening is via an increase in the net energy of feed rations, speciﬁcally by increasing the cereal grain content and the lipid content of the diet [75, 96]. In a study by Nuernberg et al. [97], done with bulls, the substitution of energy-dense concentrate ingredients with

CAB International 2009

Albert Sundrum

grass of lower energy concentration resulted in carcasses with lower IMF content and tougher meat when compared with concentrate-fed bulls. On the other hand, feeding grass had a positive effect on the fatty acid proﬁle of their meat. O’Sullivan et al. [98] found that beef from grass silage-fed animals had better overall quality in terms of colour, lipid oxidation and a-tocopherol levels than beef from maize silage fed animals. From an experiment conducted with varying proportions of autumn grass and grass silage, French et al. [99] concluded that the quality of meat from steers that consumed grass supplemented with low levels of concentrate produced the most tender and acceptable meat. However, further ageing eliminated all treatment effects on eating quality of beef. In a comparison between beef from Holstein Friesian bulls fed either ad libitum throughout the rearing period or subjected to a compensatory growth feeding strategy, the compensatory feeding was found to improve the tenderness and elevate the eating quality [100]. Beyond these establishments there are many uncertainties: experiments have examined the effects of diets with high protein, low protein, protected lipid, protected protein, added oil with and without calcium and vitamin A deﬁciency. None of these manipulations gave consistent improvement in marble score or IMF% [75]. Quality traits of pork and beef are characterized by a large variation between the animals of a herd and between the preconditions of farm systems while at the same time the degree and the source of the variation differs widely. Sources of variation in relation to meat quality attributes arise from interactions between genotype, age, sex, nutrient supply and their varying impacts on the dynamic nature of muscle and fat accretion, connective tissue composition and structure. To strive for a high eating quality requires high management skills; one must be capable enough to gain an overall picture of the complex interactions within a farming system, to reﬂect on the most relevant factors, to implement feedback mechanisms and guide the production process. Thus, it primarily depends on the management whether the potentials for a high level of eating quality are fully realized.

Barriers Scientiﬁc knowledge on how to produce pork and beef of a high intrinsic quality has markedly increased during the last few decades. However, it is obvious that knowledge alone is not sufﬁcient to change the current unsatisfactory situation, and that the production of a high eating quality is not solely the responsibility of the farmer. While the production processes at the farm level and the quality traits of meat products are characterized in the ﬁrst place by heterogeneity, branding programmes and marketing strategies often are reducing the heterogeneity of reality to very simple messages. Many brands put an emphasis only on some very speciﬁc traits while

g

41

neglecting or ignoring others [1]. The less that is known by the consumer about a product’s production processes and traits, the easier it is to construct an image, and this enables the consumer to project and connect individual and favourable ideas with the product while leaving plenty of room for speculation. On the other hand, a speciﬁc food item’s image and a consumer’s previous experience with its quality become more and more decisive in the purchase situation [101]. Based merely on single traits to transmit images to the consumer, food marketing ignores existing and production-related variability but nonetheless tries to sell a more or less uniform image of the product. The market is full of over-simpliﬁcations and inaccurate images and this makes it difﬁcult to employ more meaningful quality assessments into practical shopping/ consuming situations. Since the consumer is seldom able to evaluate whether the food product has delivered on promised process qualities, process-related qualities of food products are almost exclusively based on trust. Thus, food image and reputation, whether based on facts or anecdotal, are increasingly important in food marketing [1]. Currently, meat often has a negative image in the public both as a result of food scandals of past few years and the anecdotal statement that present-day meat is fatty and unhealthy from a nutritional viewpoint [102]. When purchasing meat products, price, appearance and place of purchase are all primary and interrelated meat-selection criteria [103]. Different consumers show different preferences and subjective perceptions in relation to different features and rank quality criteria differently. While some groups of consumers exhibit greatest preference for dark ﬁrm and dry (DFD) pig meat, others prefer normal appearing chops; a relevant number of consumers show a preference for PSE pork [104]. Reactions of consumers in four countries to the appearance and taste of pork with and without information concerning outdoor production were tested by Dransﬁeld et al. [105]. In all countries, consumers relied on colour of pork rather than marbling and drip loss to make their choice. Almost half of the British and Danish preferred the paler and the French preferred the darker pork. When information was provided in the form of labels, the vast majority of consumers preferred the pork labelled as originating from their own country as opposed to ‘imported’ and those labelled as pork from pigs ‘raised outside’ as opposed to ‘inside’. What consumers were willing to pay varied widely and the amount was higher for those consumers who found more of the characteristics they sought. A study about consumer visual preference and value for beef steaks showed that consumers who preferred low marbling seemed to desire lean products, and consumers who preferred high marbling seemed to desire products with high eating quality [106]. Boleman et al. [107] and Platter et al. [108] found that consumers could discriminate between steaks of different tenderness classes and that marbling scores and Warner–Bratzler shear

CAB International 2009

42

Animal Science Reviews 2010

forces inﬂuenced both the probability that consumers will purchase and the price they are willing to pay for strip loin steaks. Oliver et al. [109] concluded from their studies that the individual preferences could lead to a concept of market segmentation based on taste preferences. Although there are several attributes that are closely linked to the eating quality of meat, they are hardly assessed in the general production process. The marketing system does not reward high eating quality with a premium, but rather pays for quantity. Moreover, farmers are punished by deductions if they do not fulﬁl the quantitative demands of the meat industry. As long as qualitative traits are not directly assessed and not used to distinguish between different qualities, the use of the term ‘meat quality’ for marketing purpose is misleading and correspondingly leads to unfair competition between those producers who really strive for eating quality and those who, without proof, only claim to do so. Given a high variability in the preconditions of meat production, following the organic guidelines does not automatically lead to a higher eating quality of meat. From a superordinate perspective there is reason to conclude that organic livestock production is trapped between aroused consumer expectations and limited resources. Although there is evidence that organic production has the potential to offer a higher level of eating quality, there is concern that organic farming has manoeuvred itself into a blind alley because of a lack of both clear objectives and threshold values concerning acceptable eating qualities. Many consumers seem to be willing to pay more to purchase organic products. However, as long as relevant traits of meat quality are not routinely assessed and as long as the market does not distinguish between meat products of different quality there seems to be little chance to establish a speciﬁc target market based on quality products. To prevent the loss of credibility, organic farmers and retailers are obliged to take the burden of proof. Thus, there is a need for a shift in the paradigm from a guideline to an output-oriented approach [110]. Marbling respectively IMF content in both pork and beef can be easily used to differentiate between products of different eating quality because of the close relationship between IMF and various palatability traits. The grading systems for beef in USA, Australia and Japan provide examples of how grading relevant eating qualities and traits might work. Retailers of organic products so far have failed to provide a clear proﬁle of traits closely related to eating quality and have made no efforts to explain any differences to those consumers groups that are interested in what they eat. Thus, there is a need for more transparency and guidance in the confusing issue of eating quality both for the farmers and the consumers. Reliable monitoring systems for assessing meat quality are urgently required. No progress, however, can be expected with regard to eating quality if the farmer will not gain any beneﬁt and

g

proﬁt from the market. Retailers have to make sure that a high level of meat quality will be honoured by adequate prices to cover the additional efforts needed to ensure the production of high-quality meat. In the past, the marketing of organic products has done a lot to use the image of organic to increase turnover rates but has done nothing to ensure that organic products actually correspond to the image.

Scientiﬁc Challenges From a systemic point of view, meat quality is the outcome of a very complex process that emerges on the level of the single organism. Numerous factors such as genotype, age, sex, growth development, etc. interact within each organism and react with the various factors outside of the organism such as feeding, housing and management conditions. Accordingly, the large variation within a herd in relation to quality attributes of meat is the result of the potential of numerous interactions. There is a growing understanding within the scientiﬁc community that it is necessary to develop more comprehensive concepts that simultaneously consider a larger number of causal relationships. The isolated view under ceteris paribus assumptions appears to be guilty of selective attention and over-generalization, and is beginning to be replaced by a systemic approach [111]. This requires interdisciplinary collaboration, since this problem’s deﬁnition cuts across conventional commodity and disciplinary lines. The key feature of a systemic approach is that it captures the dynamics and interactions between the various elements within each system [112]. Thus, there is a need for a higher level of plausibility and coherence in relation to general statements and conclusions regarding marketing, labelling and branding. Furthermore, efﬁcient feedback mechanisms, such as monitoring concepts and quality assurance schemes have to be promoted to increase the predictability of meat quality. On the other hand, new hopes related to new methods in animal breeding should not be overestimated. The current progress in genomics will deﬁnitely increase the knowledge of physiology and genetics. These progressions are, however, not likely to improve meat quality automatically, especially when this approach does not match with the complexity of the meat growth processes within the farming system. The development of modelling systems which aim at predicting meat quality based upon data collected from an animal’s birth, through its slaughter, preparation and consumption, including all the details of its growth, may be helpful to better understand the sources of quality-related variance. On the other hand, the ordinal- and interval-scaled traits relevant for assessing the eating quality of meat aggravate any statistics, as operations such as multiplication and division cannot be carried out directly and coefﬁcient of variation cannot be deﬁned properly. Moreover, the partly antagonistic

CAB International 2009

Albert Sundrum

relationships between single traits and the bio-cybernetic self-regulation processes within both the organism and the farm system are opposed to the development of valid algorithms. The development of speciﬁc attributes of meat is inﬂuenced and regulated on different process levels (cell, organism, herd and farm system), which narrows the potential for valid conclusions drawn from results obtained on one process level to be transferred to other process levels. Following these considerations, there is limited potential to generalize results obtained under standardized conditions and to transfer these results into the heterogeneity of organic farm systems. Implications Currently, the quality of organic pork and beef is characterized by a substantial variation in different traits reﬂecting the various preconditions, objectives and implementations within the production chain. In general, organic can only partly set itself apart from the level of eating quality, or palatability, in intensiﬁed production units. Obviously, organic guidelines play a minor role with respect to meat quality. The various animal and farmrelated factors known to inﬂuence meat quality have to be balanced by management and the antagonistic relationships between some quantitative and qualitative traits must be taken into account. Within quality assurance systems, all aspects of meat production chain, including pre- and post-farmgate factors need to be optimized in order to minimize variation in eating quality. The trade value of carcasses is primarily determined by lean meat and the composition of its cuts while the production of meats with high levels of eating quality are not rewarded by the current market and are at risk of facing even lower market prices. Thus, the current marketing system is in contradiction to the production of pork and beef with high eating qualities. Only a direct assessment of qualitative traits and payment systems that honour qualities beyond average will contribute to improve this unsatisfying situation. This will, however, require a paradigm-shift from a guideline-oriented to an outputoriented approach. Additionally, as eating quality is the outcome of a very complex process, a systematic approach is needed to make the most effective and efﬁcient use of the multiple factors within each farming system and production chain.

References 1. Andersen HJ, Oksbjerg N, Therkildsen M. Potential quality control tools in the production of fresh pork, beef and lamb demanded by the European society. Livestock Production Science 2005;94:105–24. 2. Verbeke WA, Viaene J. Ethical challenges for livestock production: meeting consumer concerns about meat safety

g

43

and animal welfare. Journal of Agricultural and Environmental Ethics 2000;12:141–51. 3. Mayﬁeld LE, Bennet RM, Tranter RB, Wooldrige MJ. Consumption of welfare-friendly food products in Great Britain, Italy and Sweden, and how it may be inﬂuenced by consumer attitudes to, and behavior towards, animal welfare attributes. Journal of Sociology of Food and Agriculture 2007;15:59–73. 4. Bech-Larsen T, Grunert KG. Integrating the theory of planned behaviour with means-end chain theory – a study of possible improvements in predictive ability. In: Andersson P, editor. Proceedings of the 27th EMAC Conference, 20–23 May 1998, Stockholm, Sweden; 1998. p. 305–14. 5. Oughton E, Ritson C. Food consumers and organic agriculture. In: Cooper J, Niggli U, Leifert C, editors. Handbook of Organic Food Safety and Quality. Woodhead Publishing Limited, Cambridge, UK; 2007. p. 74–94. 6. Schaumann W. Der wissenschaftliche und praktische Entwicklungsweg des o¨kologischen Landbaus und seine Zukunftsperspektive. In: Schaumann W, Siebeneicher G, Lu¨nzer I, editors. Geschichte des o¨kologischen Landbaus, Volume 65. SO¨L-Sonderausgabe; 2002. p. 11–58. 7. Hofmann K. What is quality? Deﬁnition, measurement and evaluation of meat quality. Meat Focus International 1994;3:73–82. 8. Hocking AD, Arnold A, Jenson I, Newton K, Sutherland P. Foodborne Microorganisms of Public Health Signiﬁcance. Australian Institute of Food Science and Technology Inc., Sydney; 1997. 9. Nowak B, Sammet K, Klein G, Muefﬂing TV. Trends in the production and storage of fresh meat – the holistic approach to bacteriological meat quality. International Journal of Food Science and Technology 2005;41:303–10. 10. Oddy VH, Harper GS, Greenwood PL, McDonagh MB. Nutritional and developmental effects on the intrinsic properties of muscles as they related to the eating quality of beef. Australian Journal of Experimental Agriculture 2001;41:921–42. 11. Hambrecht E, Eissen JJ, de Klein WJH, Ducro BJ, Smits CHM, Verstegen MWA, et al. Rapid chilling cannot prevent inferior pork quality caused by high preslaughter stress. Journal of Animal Science 2004;82:551–6. 12. Ferguson DM, Bruce HL, Thompson JM, Egan AF, Perry D, Shorthose WR. Factors affecting beef palatability – farmgate to chilled carcass. Australian Journal of Experimental Agriculture 2001;41:879–91. 13. Rosenvold K, Andersen HJ. Factors of signiﬁcance for pork quality – a review. Meat Science 2003;64:219–37. 14. Susenbeth A. Anpassung der Aminosa¨urenversorgung an das Wachstumsvermo¨gen von Schweinen. Lohmann Information 2002;21–4. 15. Affentranger P, Gerwig C, Seewer GJ, Schwo¨rer D, Ku¨nzi N. Growth and carcass characteristics as well as meat and fat quality of three types of pigs under different feeding regimes. Livestock Production Science 1996;45:187–96. 16. Fischer K. Bedingungen fu¨r die Produktion von Schweineﬂeisch guter sensorischer und technologischer Qualita¨t. Mitteilungsblatt BAFF Kulmbach 2001;40:7–22.

CAB International 2009

44

Animal Science Reviews 2010

17. Baulain U, Ko¨hler P, Kallweit E, Brade W. Intramuscular fat content in some native German pig breeds. EAAP Publication 2000;100:181–4. 18. Sundrum A, Bu¨terfering L, Henning M, Hoppenbrock KH. Effects of on-farm diets for organic pig production on performance and carcass quality. Journal of Animal Science 2000;78:1199–205. 19. Andersen HJ. What is pork quality? EAAP Publication 2000;100:15–26. 20. Rosner F, Von Lengerken G, Maak S. The value of pig breeding herds in Germany and progress in improvement of meatiness and pork quality. Animal Science Papers and Reports 2003;21:153–61. 21. Brewer MS, Zhu LG, McKeith FK. Marbling effects on quality characteristics of pork loin chops: consumer purchase intent, visual and sensory characteristics. Meat Science 2001;59:153–63. 22. Warriss PD, Bevis EA, Ekins PJ. The relationship between glycogen stores and muscle ultimate pH in commercially slaughtered pigs. British Veterinary Journal 1989;145:378–83. 23. Wulf DM, Emnett RS, Leheska JM, Moeller SJ. Relationships among glycolytic potential, dark cutting (dark, ﬁrm, and dry) beef, and cooked beef palatability. Journal of Animal Science 2002;80:1895–903. 24. Oksbjerg N, Henckel P, Andersen S, Pedersen B, Nielsen B. Genetic variation of in vivo muscle glycerol, glycogen, and pigment in Danish purebred pigs. Acta Agriculture Scandinavica, Section A, Animal Science 2005;54:187–92. 25. Temisan V, Augustini C. Qualita¨tsrindﬂeisch – Deﬁnition, Standardisierung, Wege zur Erzeugung. Fleischwirtschaft 1989;69:31–7. 26. Nuernberg K, Wegner J, Ender K. Factors inﬂuencing fat composition in muscle and adipose tissue of farm animals. Livestock Production Science 1998;56:145–56. 27. Purchas RW, Burnham DI, Morris ST. Effect of growth potential path on tenderness of beef longissimus muscle form bulls and steers. Journal of Animal Science 2002;80:3211–21. 28. Scheeder MR, Meili E, Bezencon M, Spring J, Kreuzer M. Beziehungen zwischen dem Alter von Rindern und Ochsen aus einem Weidemastverfahren und Merkmalen der Fleischtextur (Zartheit). Schriftenreihe Institut fu¨r Nutztierwissenschaften, ETH Zu¨rich, 2007;29:140–3. 29. Frickh JJ, Steinwidder A, Baumung R. Einﬂuss von Rationsgestaltung, Geschlecht und Mastendgewicht auf die Fleischqualita¨t von Flechvieh-Tieren. Zu¨chtungskunde 2003;75:16–30. 30. Fisher A. Beef carcass classiﬁcation in the EU: an historical perspective. EAAP Publication 2007;123:19–30. 31. Willer H, Yusseﬁ M. The World of Organic Agriculture 2007: Statistics and Emerging Trends. 7th ed. International Federation of Organic Agriculture Movements Publication (IFOAM), 2007. Available from: URL: http://www.soel.de/ oekolandbau/weltweit.html (accessed 15 November 2008). 32. IFOAM (International Federation of Organic Agricultural Movement). Principles of Organic Agriculture. Bonn, Germany; 2006. 33. Sundrum A. Organic livestock farming – a critical review. Livestock Production Science 2001;67:207–15.

g

34. Alroe HF, Noel E. What makes organic agriculture move – protest, meaning or market? A polyocular approach to the dynamics and governance of organic agriculture. International Journal of Agricultural Resources, Governance and Ecology 2008;7:5–22. 35. Vaarst M, Padel S, Arsenos G, Sundrum A, Kuzniar A, Walkenhorst M, et al. Challenges for animal health and welfare in the implementation of the EU legislation on organic livestock production: analysis of questionnaire survey among SAFO participants. In: Rymer C, Vaarst M, Padel S, editors. Future Perspectives for Animal Health on Organic Farms: Main Findings, Conclusions and Recommendations from the SAFO Network. 5th SAFO-Workshop, 30–31 May 2006, Odense, Denmark; 2006. p. 43–74. 36. NCBA (National Cattlemens’s Beef Association). National Beef Instrument Assessment Plan II. NCBA, Englewood, CO; 2002. 37. Ender K, Fiedler I, Dietl G. Muskelstrukturmerkmale fu¨r die Selektion beim Schwein. Archiv Tierzucht Dummerstorf 2003;46:72–7. 38. Murray A, Pommier S, Robertson W, Johns M. Effect of genotype on marbling content. Advances in Pork Production 2004;15, Abstract 17. 39. Mo¨rlein D. Bestimmung des intramuskula¨ren Fettgehaltes (IMF) im M. longissimus von Schweinen mittels Ultraschallspektralanalyse (dissertation). Halle University; 2005. 40. Fischer K, Lindner JP, Judas M, Ho¨reth R. Schlachtko¨rperzusammensetzung und Gewebebeschaffenheit von schweren Schweinen – II. Mitteilung: Merkmale der Fleisch- und Fettqualita¨t. Archiv Tierzucht 2006;49:279–92. 41. Sundrum A, Ebke M. Problems and challenges with the certiﬁcation of organic pigs. In: Hovi M, Sundrum A, Padel S, editors. Proceedings of the 2nd SAFO-Workshop, 25–27 March 2004, Kassel, Germany; 2004. p. 193–8. 42. Jonsa¨ll A, Johansson L, Lundstro¨m K, Andersson KH, Nilsen AN, Risvik E. Effects of genotype and rearing system on sensory characteristics and preference for pork. Food Quality and Preference 2002;13:73–80. 43. Olsson V, Andersson K, Hansson I, Lundstro¨m K. Differences in meat quality between organically and conventionally produced pigs. Meat Science 2003;64:287–97. 44. Abel S, Weissensteiner R, Marien C, Zollitsch W, Sundrum A. Effect of a feeding strategy to increase intramuscular fat content of pork under the conditions of organic farming. In: Proceedings of the Second Scientiﬁc Conference of the International Society of Organic Agriculture Research (ISOFAR), 18–20 June 2008, Modena, Italy, 2008. p. 198–201. 45. Sundrum A, Aragon A, Schulze-Langenhorst C, Bu¨tfering L, Henning M, Stalljohann G. Implications of genotype and amino acid sup-ply on pork quality. In: Proceedings of the European Joint Congress Organic Farming and European Rural Development, 30–31 May 2006, Odense, Denmark; 2006. p. 542–3. 46. Millet S, Hesta M, Seynaeve M, Ongenae E, De Smet S, Debraekeleer J, et al. Performance, meat and carcass traits of fattening pigs with organic versus conventional housing and nutrition. Livestock Production Science 2004;87:109–19.

CAB International 2009

Albert Sundrum 47. Millet S, Ongenae E, Hesta M, Seynaeve M, De Smet S, Janssens GP. The feeding of ad libitum dietary protein to organic growing-ﬁnishing pigs. Veterinary Journal 2006;171:483–90. 48. Hansen LL, Claudi-Magnussen C, Jensen SK, Andersen HJ. Effect of organic pig production systems on performance and meat quality. Meat Science 2006;74:605–15. 49. Partanen K, Siljander-Rasi H, Alaviuhkola T. Feeding weaned piglets and growing-ﬁnishing pigs with diets based on mainly home-grown organic feedstuffs. Agricultural and Food Science 2006;15:89–105. 50. Heyer A, Andersson KH, Lundstro¨m K. Performance, carcass and technological meat quality of pigs in indoor and outdoor production system.Acta Agriculturae Scandinavica, Section A Animal Science 2006;56:55–64. 51. Woodward BW, Fernandez MI. Comparison of conventional and organic beef production systems II. Carcass characteristics. Livestock Production Science 1999;61:225–31. 52. Nielsen BK, Thamsborg SM. Welfare, health and product quality in organic beef production: a Danish perspective. Livestock Production Science 2005;94:41–50. 53. Russo C, Preziuso G. Carcass and meat quality of organic beef: a brief review. Animal Breeding Abstracts 2005;73:11N–14N. 54. Susenbeth A, Dickel T, Diekenhorst A, Hohler D. The effect of energy intake, genotype and body weight on protein retention in pigs when dietary lysine is the ﬁrst-limiting factor. Journal of Animal Science 1999;77:2985–9. 55. Ventanas S, Ventanas J, Jurado A, Este´vez M. Quality traits in muscle biceps femoris and back-fat from purebred Iberian and reciprocal IberianDuroc crossbred pigs. Meat Science 2006;73:651–9. 56. Fujii J, Otsu K, Zorzato F, De Leon S, Khanna VK, Weiler JE, et al. Identiﬁcation of a mutation in the porcine ryanodine receptor associated with malignant hyperthermia. Science 1991;253:448–51. 57. Monin G, Larzul C, Le Roy P, Culioli J, Mourot J, RoussetAkrim S, et al. Effects of the halothane genotype and slaughter weight on texture of pork. Journal Animal Science 1999;77:408–15. 58. Oksbjerg N. Effects of manipulating muscle glycogen on meat quality in pigs. In: Garnsworthy PC, Wiseman J, editors. Recent Advances in Animal Nutrition 2006. Nottingham University Press, Nottingham, UK; 2007. p. 229–41. 59. Cassens RG. Historical perspectives and current aspects of pork meat quality in the USA. Food Chemistry 2000;69:357–63. 60. Stamer E, Brade W, Kalm E. Genetic trends in different MHS genotypes of Pietrain boars in Lower Saxony. Zu¨chtungskunde 2008;80:279–90. 61. Le Roy P, Moreno C, Elsen JM, Caritez JC, Billon Y, Lagant H, et al. Interactive effects of the HAL and RN major genes on carcass quality traits in pigs: preliminary results. EAAP Publication 2000;100:139–42. 62. Jonsa¨ll A, Johansson L, Lundstro¨m K. Sensory quality and cooking loss of ham muscle (M. biceps femoris) from pigs reared indoors and outdoors. Meat Science 2001;57:245–50.

g

45

63. De Vries AG, Faucitano L, Sosnicki A, Plastow GS. Inﬂuence of genetics on pork quality. EAAP Publication 2000;100:27–35. 64. Wood JD. Consequences for meat quality of reducing carcass fatness. In: Wood JD, Fisher AV, editors. Reducing Fat in Meat Animals. Elsevier Applied Science, London; 1990. p. 344–97. 65. Schwo¨rer D, Hofer A, Lorenz D, Rebsamen A. Selection progress of intramuscular fat in Swiss pig production. EAAP Publication 2000;100:69–72. 66. Edwards SA, Wood JD, Moncrieff CB, Porter SJ. Comparison of the Duroc and Large White as terminal sire breeds and their effect on pigmeat quality. Animal Production 1992;54:289–97. 67. Ellis M, Webb AJ, Avery PI, Brown I. The inﬂuence of terminal sire genotype, sex, slaughter weight, feeding regime and slaughter-house on growth performance and carcass and meat quality in pigs and on the organoleptic properties of fresh pork. Animal Science 1996;62:521–30. 68. Campion DR, Crouse JD, Dikeman ME. Predictive value of UDSA beef quality grade factors for cooked and meat palatability. Journal of Food Science 1975;40:1225–8. 69. San˜udo C, Albertı´ P, Campo MM, Olleta JL, Panea B. Instrumental quality of beef meat from seven Spanish breeds. Archives of Zootechnology 1998;48:397–402. 70. Augustini C, Pirchner F, Eichinger H, Reinsch N, Ko¨gel J. Fleisch-Leistung der gefa¨hrdeten bayrischen Rinderrassen. 2. Mitteilung: Fleischqualita¨t. Zu¨chtungskunde 1998;70:328–37. 71. Serra X, Guerrero L, Gua`rdia MD, Gil M, San˜udo C, Panea B, et al. Eating quality of young bulls from three Spanish beef breed-production systems and its relationships with chemical and instrumental meat quality. Meat Science 2008;79:98–104. 72. Vieira C, Garcı´ a-Cacha´n MD, Recio MD, Domı´ nguez M, San˜udo C. Effect of ageing time on beef quality of rustic type and rusticCharolais crossbreed cattle slaughtered at the same ﬁnishing grade. Spanish Journal of Agricultural Research 2006;4:225–34. 73. Lively FO, Moss BW, Keady TW, Patterson DC, Gordan AG. The effect of genotype and carcass hanging method on eating quality. In: Proceedings of the 51st International Congress of Meat Science and Technology, 7–12 August 2005, Baltimore, MD; 2005. p. 1808–14. 74. Sapp RL, Bertrand JK, Pringle TD, Wilson DE. Effects of selection for ultrasound intramuscular fat percentage in Angus bulls on carcass traits of progeny. Journal of Animal Science 2002;80:2017–22. 75. Bindon BM. A review of genetic and non-genetic opportunities for manipulation of marbling. Australian Journal of Experimental Agriculture 2004;44:687–96. 76. Morgan JB, Wheeler TI, Koohmaraie M, Savell JW, Crouse JD. Meat tenderness and the calpain system in longissimus muscle of young bulls and steers. Journal of Animal Science 1993;71:1471–6. 77. Maher SC, Mullen AM, Keane MG, Buckley DJ, Kerry JP, Moloney AP. Variation in the eating quality of M. longissimus dorsi from Holstein-Friesian bulls and steers of New Zealand and European/American descent, and Belgian BlueHolstein-Friesians, slaughtered at two weights. Livestock Production Science 2004;90:271–7.

CAB International 2009

46

Animal Science Reviews 2010

78. Schreurs NM, Garcia F, Jurie C, Agabriel J, Micol D, Bauchart D, et al. Meta-analysis of the effect of animal maturity on muscle characteristics in different muscles, breeds, and sexes of cattle. Journal of Animal Science 2008;86:2872–87. 79. Sundrum A, Schneider K, Richter U. Possibilities and limitations of protein supply in organic poultry and pig production. In: Proceedings of the European Joint Congress Organic Farming and European Rural Development, 30–31 May 2006, Odense, Denmark, 2006. p. 528–9. 80. Rosenvold K, Petersen JS, Laerke HN, Jensen SK, Therkildsen M, Karlsson AH, et al. Muscle glycogen stores and meat quality as affected by strategic ﬁnishing feeding of slaughter pigs. Journal of Animal Science 2001;79:382–91. 81. Cisneros F, Ellis M, Baker DH, Easter RA, McKeith FK. The inﬂuence of short-term feeding of amino acid-deﬁcient diets and high dietary leucine levels on the intramuscular fat content of pig muscle. Journal of Animal Science 1996;63:517–22. 82. D’Souza DN, Pethick DW, Dunshear FR, Mullan BP. Nutritional manipulation increases intramuscular fat levels in the longissimus muscle of ﬁnisher pigs. Australian Journal of Agricultural Research 2003;54:745–9. 83. Cameron ND, Enser M, Nute GR, Whittington FM, Penman JC, Fisken AC, et al. Genotype with nutrition interaction on fatty acid composition of intramuscular fat and the relationship with ﬂavour of pig meat. Meat Science 2000;55:187–95. 84. Katsumata M, Kobayashi S, Matsumoto M, Tsuneishi E, Kayi Y. Reduced intake of dietary lysine promotes accumulation of intramuscular fat in the M. longissimus dorsi muscle of ﬁnishing gilts. Animal Science Japan 2005;76:237–44. 85. Oksbjerg N, Strudsholm K, Lindahl G, Hermansen John E. Meat quality of fully or partly outdoor reared pigs in organic production. Acta Agriculturae Scandinavica, Section A Animal Science 2005;55:106–12. 86. Candek-Potokar M, Zlender B, Lefaucheur L, Bonneau M. Effects of age and/or weight at slaughter on longissimus dorsi muscle: biochemical traits and sensory quality in pigs. Meat Science 1998;48:287–300. 87. Schinckel AP, Smith JW, Tokach MD, Dritz SS, Einstein M, Nelssen JL, et al. Two on-farm data collection method to determine dynamics of swine compositional growth and estimates of dietary lysine requirements. Journal of Animal Science 2002;80:1419–32. 88. Lebret B, Juin H, Noblet J, Bonneau M. The effects of two methods of increasing age at slaughter on carcass and muscle traits and meat sensory quality in pigs. Animal Science 2001;72:87–94. 89. Le Bellego L, Van Milgen J, Noblet J. Effect of high ambient temperature on protein and lipid deposition and energy utilization in growing pigs. Animal Science 2002;75:85–96. 90. Hyun M, Ellis M. Effect of group size and feeder type on growth performance and feeding patterns in growing pigs. Journal of Animal Science 2001;79:803–10. 91. Scho¨nfelder A. The effect of rank order a feed intake and growth of fattening pigs. Deutsche Tiera¨rztliche Wochenschrift 2005;112:215–8.

g

92. Botermans JA, Svendsen J. Effect of feeding environment on performance, injuries and behaviour in growing-ﬁnishing pigs: group-based studies. Acta Agriculture Scandinavica Section A, Animal Science 2000;50:237–49. 93. Wolter BF, Ellis M, Curtis SE, Parr EN, Webel DM. Effect of feeder-through space and variation in body weight within a pen of pigs on performance in a wean-to-ﬁnish production system. Journal of Animal Science 2002;80:2241–6. 94. Whittemore CT, Tullis JB, Emmans GC. Protein growth in pigs. Animal Production 1988;46:437–45. 95. Sainz RD, Vernazza Paganini RF. Effects of different grazing and feeding periods on performance and carcass traits of beef steers. Journal of Animal Science 2004;82:292–7. 96. Pethick DW, Harper GS, Oddy VH. Growth, development and nutritional manipulation of marbling in cattle: a review. Australian Journal of Experimental Agriculture 2004;44:705–15. 97. Nuernberg K, Dannenberger D, Nuernberg G, Ender K, Voigt J, Scollan ND, et al. Effect of a grass-based and a concentrate feeding system on meat quality characteristics and fatty acid composition of longissimus muscle in different cattle breeds. Livestock Production Science 2005;94:137–47. 98. O’Sullivan A, O’Sullivan KO, Galvin K, Moloney AP, Troy DJ, Kerry JP. Grass silage versus maize silage effects on retail packaged beef quality. Journal of Animal Science 2002;80:1556–63. 99. French P, O’Riordana EG, Monahan FJ, Caffrey PJ, Vidal M, Mooney MT, et al. Meat quality of steers ﬁnished on autumn grass, grass silage or concentrate-based diets. Meat Science 2000;56:173–80. 100. Hansen S, Therkildsen M, Byrne DV. Effects of a compensatory growth strategy on sensory and physical properties of meat from young bulls. Meat Science 2006;74:628–43. 101. Grunert KG, Bredahl L, Brunso K. Consumer perception of meat quality and implications for product development in the meat sector – a review. Meat Science 2004;66:227–59. 102. Kubberod E, Ueland O, Tronstad A, Risvik E. Attitudes towards meat and meat-eating among adolescents in Norway: a qualitative study. Appetite 2002;38:53–62. 103. McEachern MG, Schro¨der MJ. The role of livestock production ethics in consumer values towards meat. Journal of Agricultural and Environmental Ethics 2002;15:221–37. 104. Jeremiah LE. Consumer responses to pork loin chops with different degrees of muscle quality in two western Canadian cities. Canadian Journal of Animal Science 1994;74:425–32. 105. Dransﬁeld E, Ngapo TM, Nielsen NA, Bredahl L, Sjo¨den PO, Magnusson M, et al. Consumer choice and suggested price for pork as inﬂuenced by its appearance, taste and information concerning country of origin and organic pig production. Meat Science 2005;69:61–70. 106. Killinger KM, Calkins CR, Umberger WJ, Feuz DM, Eskridge KM. Consumer visual preference and value for beef steaks differing in marbling level and color. Journal of Animal Science 2004;82:3288–93. 107. Boleman SJ, Boleman SL, Miller RK, Taylor JF, Cross HR, Wheeler TL, et al. Consumer evaluation of beef of known categories of tenderness. Journal of Animal Science 1997;75:1521–4.

CAB International 2009

Albert Sundrum 108. Platter WJ, Tatum JD, Belk KE, Koontz SR, Chapman PL, Smith GC. Effects of marbling and shear force on consumers’ willingness to pay for beef strip loin steaks. Journal of Animal Science 2005;83:890–9. 109. Oliver MA, Nute GR, Font i Furnols M, San Julian R, Campo MM, Sanudo C, et al. Eating quality of beef, from different production systems, assessed by German, Spanish and British consumers. Meat Science 2006;74:435–43. 110. Sundrum A. From a standard to an output oriented approach in organic live-stock farming. In: Proceedings of the European Joint Congress Organic Farming and European

g

47

Rural Development, 30–31 May 2006, Odense, Denmark; 2006. p. 508–9. 111. German Research Foundation (DFG). Future Perspectives of Agricultural Science and Research. Wiley-VCH Publisher, Bonn, Germany; 2005. 112. Sundrum A. Achievements of research in the ﬁeld of livestock systems and quality production. In: Rosati A, Tewolde A, Mosconi C, editors. Animal Production and Animal Science Worldwide. WAAP Book of the Year 2006. Wageningen Academic Publishers, Wageningen, The Netherlands; 2007. p. 95–106.

CAB International 2009

This page intentionally left blank

Animal Science Reviews 2010

Whole genome marker-assisted selection Joel I. Weller* Address: Institute of Animal Sciences, A.R.O., The Volcani Center, PO Box 6, Bet Dagan 50250, Israel. *Correspondence: Email: [email protected] 24 November 2009 25 January 2010

Received: Accepted:

Abstract Most applications of genomic selection (GS) have so far been in animals, especially dairy cattle, although theoretical studies have also been conducted for maize. For the last 50 years, commercial breeding programmes for dairy cattle have been based on the progeny test scheme, which results in an average generation interval of approximately 7 years for the sire to dam path. It should be possible to double rates of genetic gain by the application of GS, if generation intervals are reduced to close to the biological minimum. During the last decade, methods were developed for highthroughput genotyping of thousands of single nucleotide polymorphisms per individuals. The number of potential polymorphic markers per species was increased from about 1000 to tens of thousands, and costs per individual marker genotype were reduced from several dollars to less than one cent. The application of marker-assisted selection based on genome-wide association studies requires solutions to new statistical problems. Speciﬁcally how should information from pedigree, phenotypic records and genotypes be combined to optimally rank candidates for selection? Various linear and Bayesian methods have been proposed and tested to compute genomic estimated breeding values. Linear models require signiﬁcantly less computing time, and perform nearly as well as Bayesian methodologies. Methods are generally evaluated by comparing genomic evaluations based only on pedigree and genotype to genetic evaluations based on daughter records of the same bulls. Genomic selection programmes for dairy cattle have been implemented in the USA, Canada, Australia, New Zealand and the Netherlands. With declining genotyping costs it becomes economically viable to genotype more individuals, including candidate bull dams. More emphasis can be placed on low heritability traits, such as fertility and disease traits, and it will be easier to control the increase in inbreeding in commercial animal populations. Keywords: Genomic selection, linkage disequilibrium mapping, marker-assisted selection, genomic estimated breeding values, single nucleotide polymorphism Review Methodology: ISI Web of Knowledge and PubMed were searched using the key words ‘genomic selection’.

Introduction During the 1950s and 1960s, advance breeding programmes were developed for most economically important plant and animal species. These programmes were based on the ‘inﬁnitesimal model’ of quantitative genetic inheritance [1], that is, quantitative traits are controlled by a very large number of genetic and environmental factors, and the effect of any speciﬁc factor is very small. The inﬁnitesimal model and the statistical techniques developed made virtually no use of the principles of Mendelian genetics.

g

In 1923, Sax [2] ﬁrst demonstrated with beans that individual genes affecting quantitative traits, denoted ‘quantitative trait loci’ (QTL), could be detected with the aid of morphological genetic markers. In 1961, NeimannSørensen and Robertson [3] used blood groups as genetic markers to detect QTL in dairy cattle. Although a number of papers were written during the next two decades on QTL detection, the application of QTL to breeding was limited by the lack of segregating genetic markers in commercial populations. With the advent of molecularlevel genetic markers in the 1980s and microsatellites or

CAB International 2010

50

Animal Science Reviews 2010

‘simple sequence repeats’ (SSRs) in the 1990s [4], markerassisted selection (MAS) in both plants and animals became possible; although costs of deﬁning individual genotypes were still relatively expensive. During the last decade, methods were developed for high-throughput genotyping of thousands of single nucleotide polymorphisms (SNPs) per individual [5]. The number of potential polymorphic markers per species was increased from about 1000 to tens of thousands, and costs per individual marker genotype were reduced from several dollars to less than one cent. In this review, I will describe in detail the application of genomic selection (GS) to cattle, because most of the theoretical literature has dealt with this species. I will then brieﬂy describe methodologies for GS in plants, and ﬁnally consider directions for future research.

50 candidate bulls 30 000 cows 5000 daughters

Records on daughters

100 elite cows

3 foreign bulls

200 elite cows

4 local bulls

120 000 cows 20 bulls

5 bulls are selected 55 bulls are culled

Figure 1 A traditional progeny-test breeding scheme for a dairy cattle population of 120 000 cows

DNA-level Genetic Markers, SSRs versus SNPs Traditional Dairy Cattle Breeding Schemes Commercial dairy cattle breeding has been based on the constraints that nearly all economic traits are expressed only in females, while female fertility is extremely limited. Heritabilites of all important economic traits are low to moderate. Therefore, in order to utilize the nearly unlimited fertility of males, most advanced breeding programmes are based on the ‘progeny test scheme’. An example is given in Figure 1. A cohort of male calves is selected based on pedigree, and each of these calves sires between 50 and 100 milking cows. On the basis of the records of their daughters and other female relatives, about 10% of these calves are returned to general service and the rest are slaughtered. This scheme allows for high selection intensities along the sire-to-sire, sire-to-dam and dam-to-sire genetic paths, but results in generation intervals much greater than required by biological constraints. Although bull calves reach sexual maturity at the age of 1 year, milk production records on their ﬁrst crop of daughters are only available when the bull calves are 5 years old. Rates of genetic gain could be doubled if the genetic paths generation intervals were reduced to near the biological limits [6, 7]. All else remaining equal, reducing the mean generation interval on the sire-to-dam path from 7 to 3 years will increase the annual rate of genetic gain by 20%. The rate of genetic gain per generation is a function of the selection intensity and the accuracy of the estimated breeding values (EBV). ‘Accuracy’ is deﬁned as the correlation between the genetic evaluation and the actual genetic value, while ‘reliability’ is deﬁned as the square of this correlation. Based on pedigree only, that is, the EBV of the calf’s sire and dam, the accuracy of the calf’s genetic evaluation for milk production traits with heritabilities close to 0.3 will be at most 0.6, as compared with an accuracy of 0.9 after completion of the progeny test.

g

SSRs consist of tandem repeats of a sequence of 1–4 DNA base pairs, the most common being ‘TG’. SSRs are polymorphic in the number of repeats of the core sequence. In the early 1990s, SSRs became the marker of choice, chieﬂy because of their high polymorphic information content and relative ease of genotyping [8]. Thousands of SSR sequences are scattered throughout the genomes of all advance organisms. Genetic maps of at least several hundred SSRs were developed for nearly all the important agricultural species (e.g. http://www.thearkdb.org/). Unlike SSRs, SNPs are nearly always biallelic. SNPs occur at a frequency of approximately 0.3–1 SNP/kb throughout the human genome [5], and apparently at equal frequencies in other mammalian species. Advantages of SNPs are summarized by Werner et al. [9]. Genotyping error rates tend to be lower for SNPs [10, 11], larger numbers of markers can be run jointly and genotype determination is completely automatic, eliminating what is generally the largest cost element of genotyping [11, 12]. In January 2008, Illumina announced the release of the Inﬁnium1 BovineSNP50 BeadChip, which includes 54 001 SNPs approximately evenly spaced across the entire bovine genome (http://www.illumina.com/products/bovine_snp50_wholegenome_genotyping_kits.ilmn). Similar bead chips have been developed for the other major commercial animal species, and SNP chips with >1 million SNPs are available for humans.

QTL Detection and MAS Prior to SNPs Detection of QTL requires generation of linkage disequilibrium (LD) between the genetic markers and the QTL. In plants, this is generally accomplished by crossing inbred lines to produce either F2 or backcross populations [13]. These options are not available for farm animals in which inbred lines are not available and breeding is conducted within breeds. Until 2006, the most appropriate experimental designs for QTL detection and

CAB International 2010

Joel I. Weller

analysis in cattle were the daughter and granddaughter designs [14]. These designs have been applied to nearly all advanced dairy cattle populations [13]. Numerous QTL have been identiﬁed in farm animals, and this information has been summarized for cattle at two web sites: http:// www.vetsci.usyd.edu.au/reprogen/QTL_Map/ and http:// www.animalgenome.org/QTLdb/cattle.html. Limited use was made of this information in commercial animal breeding programmes. Unlike plant breeding, animal breeding is generally conducted within a single breed, or within a small number of breeds. For a QTL to be useful in an MAS animal breeding programme it must fulﬁl the following criteria: 1. Given the huge multiple comparison problem, normal signiﬁcance levels of 0.01 or even 0.001 are meaningless. Conﬁdence that a segregating QTL has in fact been detected is only obtained if the observation is repeated in several independent studies [15]. 2. The effect should be sufﬁciently large so that the conﬁdence interval for QTL location is small enough so that a haplotype including the QTL can be determined and followed through pedigrees. This problem is alleviated if the causative quantitative trait nucleotide has been identiﬁed, but this has been accomplished for dairy cattle in only two cases [16, 17]. 3. The net effect of the QTL on the selection index must be signiﬁcant. For example, in the polymorphism in the gene DGAT1 the allele that increases fat production decreases protein production [17], so the net effect on most current selection indices is close to zero. 4. Scope for selection is possible only if the frequency of the economically positive allele is relatively low.

51

criterion of the false discovery rate, many more statistically signiﬁcant SNP effects have been found, as compared with linkage-based genome scans [24–26]. The apparent reasons are as follows: 1. LD effects can be tested by a simple linear model of number of ‘+’ alleles on the bulls’ EBV or daughter yield deviation (DYD) [26]. This test is inherently more powerful than the test of signiﬁcance for linkage that must be conducted within families. 2. Genome coverage with tens of thousands of SNPs is more complete. 3. Unlike granddaughter designs, which can only utilize bulls from large half-sib families, all sires with evaluation can be included in LD analyses. Thus, effective sample sizes are greater. The results of Cole et al. [25] and VanRaden et al. [26] for the US Holstein population conﬁrm that at least with respect to the QTN that have been detected, results of GWAS do correspond to the results obtained previously by granddaughter designs. The largest effects found were for protein concentration on Bos taurus chromosome (BTA) 6 and fat concentration on BTA14. The effects on BTA6 ﬂanked the ABCG2 gene, which has been shown to have a major effect for this trait [16], and the effects on BTA14 ﬂanked the DGAT1 gene [17], which has a major effect on fat concentration with lesser effects on milk and fat yield. Both effects were ﬁrst discovered by daughter and granddaughter designs.

Estimation of QTL Effects from Genome Scans

No QTL detected so far in dairy cattle meet all four of these criteria. Nevertheless, MAS breeding programmes based on microsatellites were started in Germany and France in 2001, both based on following a rather limited number of chromosomal regions [18, 19].

The application of MAS based on GWAS requires solutions to new statistical problems. Speciﬁcally, how should information from pedigree, phenotypic records and genotypes be combined to optimally rank candidates for selection? Goddard and Hayes [27] proposed that GS could be divided into three steps:

Genome Scans, Within Family Linkage versus Population-wide LD

1. Use the markers to deduce the genotype of each animal at each QTL. 2. Estimate the effects of each QTL genotype on the trait. 3. Sum all the QTL effects for selection candidates to obtain their genomic estimated breeding values (GEBV).

Meuwissen and Goddard [20] proposed that conﬁdence intervals for QTL location could be dramatically reduced by the application of population-wide LD mapping. Unlike analysis of genetic linkage within families that extends over tens of cM, population-wide LD extends in dairy cattle at most only over individual cMs [21]. The application of LD mapping therefore requires much denser genetic maps than required for the detection of QTL via linkage, but does not require a speciﬁc family structure [22]. Since the ﬁrst reports of genome-wide association studies (GWAS) based on dense arrays of SNPs in 2006 [23], very few speciﬁc results of effects detected and their locations have been published. All studies agree that based on the

g

Estimation of QTL effects by LD from genome scans is problematic, ﬁrstly because only a very small fraction of the population will be genotyped. Furthermore, these will generally be males without records on the traits of interest. In addition, there are at least ﬁve potential sources of bias. Firstly, estimated effects will be underestimated because LD between any speciﬁc marker and the QTL will be incomplete. Thus only part of the QTL effect will be detected. Secondly, if the effects of the SNP

CAB International 2010

52

Animal Science Reviews 2010

are analysed on sires, then the dependent variable analysed will generally be the sires’ EBV or DYD. EBV for commercial dairy cattle populations are generally computed by an individual animal model, based on the principles of best linear unbiased prediction (BLUP) [13, 28]. All animals with records and all known ancestors are included in the analysis, and the heritability is assumed known. The additive genetic effect is assumed to be random, and ‘nuisance’ effects, such as the cows’ herd–year– seasons are assumed to be ﬁxed. The genetic variance is computed as the numerator relationship matrix times a scalar equal to the genetic variance among unrelated animals. EBV computed by the animal model are regressed in proportion to the amount of information available for each animal. Thus, if QTL effects are estimated by analysis of EBV, the effects will be underestimated [29]. The problem is somewhat alleviated if DYD, which are unregressed means of daughter records corrected for ﬁxed effects, are analysed instead of EBV. Unlike EBV, the variances of DYD decrease with an increase in the number of daughters. Thus, weighting DYD by the bulls’ reliabilities in an analysis of marker effects is in accordance with the generalized least-squares principle that records with greater variance should be given smaller weights. However, estimates of QTL effects based on analysis of DYD will still be biased [30]. In addition, there are three sources of upward bias. In a simple regression model, animals that are related will tend to have a higher probability to inherit the same marker alleles identical by descent. Since these animals also have a common polygenic variance, the estimated effect will also include a polygenic component because of relationships [31]. Various studies have proposed that this problem could be solved by inclusion of the inverse of relationship matrix in the analysis (e.g. [32]). However, if EBV or DYD are analysed, and the QTL effect is assumed to be a random variable, the distributional properties of the model are problematic. If the residual variance in this model represents deviation of the ‘record’ from the animal’s true additive variance, a DYD or EBV based on thousands of daughters should then have a residual variance approaching zero. The second source of upward bias is because of the fact that only a small fraction of the markers will be linked to segregating QTL. Beavis [33] ﬁrst noted that the estimates for effects deemed ‘signiﬁcant’ will be biased as only the largest effects will be reported and retained for further analysis. Bias will be greater for small effects for the following reason. Although large QTL effects will be deemed signiﬁcant in any case, small or marginal effects will be denoted ‘signiﬁcant’ only if the estimate is larger than the actual effect [33, 34]. Since in nearly all QTL analyses only the absolute value of the effect is considered, leastsquares estimates will be inﬂated because of non-zero residuals. Assuming that the residual and the actual QTL effects are uncorrelated, the variance of the least-squares estimates will be equal to the sum of the residual and true

g

QTL effect variance. Thus, positive values for QTL effects will be obtained even in the absence of a segregating QTL. The third source of upward bias results from the fact that when a LOD score or regression effect is maximized over many point-wise tests in interval mapping, the locusspeciﬁc effect-size estimate is also maximized, and this will tend to be greater than the actual QTL effect [35]. This will also be the case for the LD mapping [20, 36].

Studies on the Distribution of QTL Effects Several studies have proposed that the second source of upward bias could be alleviated by the application of Bayesian estimation methods (e.g. [37]). Bayesian estimation differs from maximum likelihood estimation in that in Bayesian estimation the likelihood function is multiplied by the ‘prior distribution’ of the parameters. This generally results in ‘shrinkage’ of the parameter estimates, but requires assumptions about the nature of distribution of QTL effects. If many QTL are analysed jointly, it should be possible to estimate both the QTL effects and the parameters of the distribution of QTL effects [38]. The extent of shrinkage will vary inversely with the sample size. Hayes et al. [39] estimated that the number of detectable QTL affecting milk production is on the order of 150, based on a whole genome scan with 10 000 SNPs, while Chamberlain et al. [40] estimated the total number of QTL at 30 by analysis of a daughter design. These studies demonstrate that most of the additive genetic variance can be explained by QTL that can be detected by SNP genome scans, provided that the number of animals analysed and the SNP densities are sufﬁcient. A number of studies have considered the question of the appropriate distribution for QTL effects, and in nearly all cases a single-sided distribution was assumed; that is, the QTL effect was assumed to vary from zero to inﬁnity. Hayes and Goddard [37] estimated the distribution of QTL effects for cattle and swine by combining results from several studies. They assumed a gamma distribution for the QTL effects. The dairy cattle analysis was based on the QTL estimates from three granddaughter design analyses, considering only ‘signiﬁcant’ effects. Thus, a truncated gamma distribution was assumed. Weller et al. [38] estimated the parameters of the distribution of QTL effects for nine economic traits in dairy cattle from a daughter design analysis of the Israeli Holstein population including 490 marker-by-sire contrasts [41]. A separate gamma distribution was derived for each trait. The estimates derived for the individual QTL effects using the gamma distributions for each trait were regressed relative to the least-squares estimates, but the regression factor decreased as a function of the least-squares estimate. On simulated data, the mean of least-squares estimates for effects with nominal 1% signiﬁcance was more than twice the simulated values, while the mean of the maximum

CAB International 2010

Joel I. Weller

likelihood estimates was slightly lower than the mean of the simulated values. The coefﬁcient of determination for the Bayesian estimates was ﬁve-fold the corresponding value for the least-squares estimates.

Appropriate Criteria for Evaluation of GEBV Evaluation of methodology to compute GEBV are generally based on analysis of a population of bulls with EBV derived from progeny tests. The population is then divided into two sets. In the ‘training set’ consisting of older bulls, all information including genotypes, genetic relationships and daughter records are used to estimate the marker effects. In the ‘validation set’ consisting of younger bulls, GEBV are computed based only on the marker genotype effects derived from the training set and pedigree. The GEBV of the validation set of bulls are then compared with the EBV of these bulls based on their daughter records and relationships [26, 42]. Most studies that have compared GEBV with EBV based on daughter records have done so on the basis of coefﬁcients of determination between the two evaluations [42]. Although this criterion is important, a second criterion is the bias of GEBV. That is, GEBV are unbiased if the regression of true breeding values on GEBV is not signiﬁcantly different from unity and the y-intercept is not signiﬁcantly different from zero [43]. If the regression is less than unity, then the bulls with the highest GEBV will be inﬂated relative to the true genetic values of these bulls.

Implementation of Methodology to Compute GEBV for Dairy Cattle ‘Interbull’, a sub-committee of the International Committee for Animal Recording, is responsible for promoting the development and execution of international genetic evaluations for cattle. Nine countries (Australia, Canada, France, Germany, Ireland, Israel, New Zealand, the Netherlands and the USA), responded to the Interbull survey question: ‘Which methodology is being used to estimate SNP effects [24]?’ Several different methods are being implemented. Four countries have adopted Bayesian methods, described brieﬂy previously. ‘Bayes-A’ and ‘Bayes-B’ methodology differ in that in Bayes-A all marker included in the analysis are assumed to have a non-zero effect on the trait analysed, while in Bayes-B methodology it is assumed that most markers have no effect [44]. Bayes-B methods are clearly closer to reality, but require signiﬁcantly greater computing time. Other methodologies will now be discussed in detail. VanRaden [32] proposed analysis of DYD as the dependent variable with all SNPs included as random effects. Genotypes for 38 416 informative markers of the 54 001 included on the BeadChip and the August, 2003,

g

53

genetic evaluations for 3576 Holstein bulls born before 1999 were used to predict the January, 2008, daughter deviations for 1759 bulls born from 1999 through 2002. GEBV were computed using linear and non-linear genomic models. For linear predictions, the traditional additive genetic relationship matrix was replaced by a genomic relationship matrix, which is equivalent to assigning equal genetic variance to all markers. Final GEBV combined three terms by selection index: 1. Direct genomic prediction. 2. Parent averages computed from the set of genotyped ancestors using traditional relationships. 3. Published parent averages or pedigree indexes, constructed as 0.5(sire EBV)+0.25(maternal grandsire EBV) + 0.25(birth year mean EBV). Combined predictions were more accurate than ofﬁcial parent averages for all 27 traits analysed [26]. Reliabilities were 0.02 to 0.38 higher than non-linear genomic predictions included as compared with parent averages alone. Misztal et al. [43] found that regressions of EBV based on progeny tests on GEBV derived by this method were less than unity for the trait ‘ﬁnal score’ (a conformation trait). Regressions for other traits have not been published. Variations of this method have also been applied to other national dairy cattle populations, e.g. as described by Liu et al. [45]. GEBV have been published in the US since April 2008. Nearly all of the top US bulls are currently young bulls with GEBV, but without progeny tests (http:// aipl.arsusda.gov/dynamic/sortnew/current/OHOnm.html). The same data were also analysed by Bayes-A and Bayes-B procedures [32]. In the Bayes-A analysis, the prior distribution was a simple, heavy-tailed distribution generated from a normal variable divided by 1.25abs(s–2), where s is the number of standard deviations from the mean and 1.25 determines departure from normality. In the Bayes-B analysis, only 700 markers of 50 000 included in the analysis were assumed to have non-zero effects. Gains in realized reliabilities were minimal, as compared with the linear model [26]. Similar results were found for Australian dairy cattle [42]. This is not surprising considering that very large samples of bulls were analysed. Therefore ‘shrinkage’ of estimates by Bayesian methodology should be minimal. Advantages of the multistage system for genomic evaluation include no change to the regular evaluations and simple steps for predicting genomic values for young genotyped animals. Disadvantages include weighting parameters, such as variance components [46] or selection index coefﬁcients [26], loss of information and biased evaluations [43]. Furthermore, the extension to alternative analysis models, such as multi-trait evaluations or test-day models, is not obvious. Finally, tracing back anomalies in a two- or three-step procedure might become very complicated. As for the loss of information, several problems exist in the use of DYD for bulls or

CAB International 2010

54

Animal Science Reviews 2010

‘yield deviations’ for cows [28]. These problems are weights (caused by different amount of information in the original data set), bias (caused by selection, for example), accuracy (for animals in small herds) and colinearity (for example, the yield deviations of two cows in the same herd). Finally, the expectation of Mendelian sampling in selected animals is not zero [47]. Other studies proposed methods for deriving GEBV from direct analysis of the complete population, even though only a small fraction is actually genotyped. Legarra et al. [48] assumed that SNPs effects are random, with conditioning of the genetic value of ungenotyped animals on the genetic value of genotyped animals via the selection index (e.g. pedigree information), and then used the genomic relationship matrix for the latter. This results in a joint distribution of genotyped and ungenotyped genetic values, with a pedigree–genomic relationship matrix H. In this matrix, genomic information is transmitted to the covariances among all ungenotyped individuals. Matrix H is suitable for iteration on data algorithms that multiply a vector times a matrix, such as preconditioned conjugated gradients [49]. This method was applied to 10 466 066 US Holsteins records for the ﬁnal score [43]. GEBV were computed based on 6508 bulls genotyped for the Illumina BovineSNP50 BeadChip and records up to 2004. GEBV were compared with those obtained by the multi-step method [32] on the same data. Comparisons were based on regressions of 2009 EBV of bulls without daughter records prior to 2005 on GEBV, and coefﬁcients of determination. This approach includes a parameter, l, that represents the fraction of the additive variance explained by the genomic information. By estimating l, the goodness of ‘genomic’ ﬁt can be determined without creating the training and validation populations. Subsequently, comparisons of different models are simpliﬁed. With ‘optimal’ scaling, this method was more accurate and less biased than the multi-step method [43]. Another alternative that has been proposed is to use genotypes to construct a realized relationship matrix between individuals, as opposed to the average relationship matrix generally included in animal model evaluations. In the realized relationship, matrix elements are the realized proportion of the genome that is identical by descent between pairs of individuals, based on shared marker haplotypes. Hayes et al. [50] demonstrated that by replacing the average relationship matrix derived from pedigree with the realized relationship matrix, the accuracy of breeding values can be substantially increased, especially for individuals with no phenotype of their own. They also demonstrated that this method of predicting breeding values is exactly equivalent to the GS methodology where the effects of QTL contributing to variation in the trait are assumed to be normally distributed. The accuracy of breeding values predicted using the realized relationship matrix can be deterministically predicted for known family relationships, for example half-sibs.

g

Visscher [51] compared the theoretical results with empirical data from 4401 pseudo-independent full-sib pairs from Australian families. The siblings and their parents were genotyped for microsatellite markers, and additive and dominance coefﬁcients of relationships were estimated at each cM and genome-wide using exact multipoint probability calculations. The mean and S.D. were 0. 498 and 0.036 for the additive coefﬁcients and 0.248 and 0.040 for the dominance coefﬁcients. Both the mean and variance were close to expectation. The extreme values for the additive coefﬁcients were 0.37 and 0.63, so that at the low range, some full-sibs are between average half-sibs and average full-sibs, and at the high range, the sibs are between average full-sibs and monozygotic twins.

Calculation of GEBV Based on Selection of Markers Similar to Bayes-B methodology, several studies have proposed a priori selection of relatively small sets of markers for computation of GEBV, e.g. Aulchenko et al. [52]. Hayes et al. [42] proposed a two-step procedure in which the effect of each SNP is ﬁrst evaluated in a simple linear model. All signiﬁcant SNPs in the linear model are then analysed jointly with SNP effects assumed to be random. This model requires an estimate of the variance of the SNP effects. To estimate this variance, Hayes et al. [42] assumed that the SNPs included in the analysis explain all the additive genetic variance among the bulls, and are approximately equal. In this case, the additive genetic variance, s a 2, can be approximately modelled as follows: s 2a =2g2

m X

pi (17pi ),

i

where g is the additive effect of each SNP, m is the number of SNPs included in the analysis and pi is the frequency of the less frequent allele for the ith SNP. The variance of the SNP effects, s g2, is then assumed to be equal to g2, and is computed as follows: s 2g =s 2a =2

m X

pi (17pi ):

i

Approximately 4000 SNPs were included in joint analysis. Thus, the effect of each individual SNP was assumed to be quite small. Weigel et al. [53] compared GEBV computed with all valid SNP markers to GEBV computed with subsets of markers. A total of 3305 US Holstein bulls born from 1952 to 1998 were used as the training set, and 1398 bulls born from 1999 to 2002 were used as the testing set. The coefﬁcient of determination (R2) from regressing the April 2008 progeny test EBV of bulls in the testing set on their

CAB International 2010

Joel I. Weller 2

August 2003 GEBV was 0.375. The R values for subsets of the 300, 500, 750, 1000, 1250, 1500 and 2000 SNP with largest effects were 0.184, 0.236, 0.279, 0.289, 0.307, 0.313 and 0.322, respectively. Solberg et al. [54] proposed the selection of markers based on either partial least-square regression or principal component regression. With 8080 total markers, the correlation between estimated and true breeding values reached a maximum when 350 principal components were ﬁtted. Both methods gave a lower accuracy and greater bias compared with a Bayes-B analysis including all markers, but computing time was 65 times greater for the all-marker analyses. These results indicate that a lowdensity assay comprising selected SNP may be a costeffective alternative for selection decisions.

Explaining the Genetic Variation A further question of interest is how efﬁcient are current SNP chips in explaining the genetic variation in a given trait. Although this question has not been dealt with in detail in farm animals, it has been considered for humans, especially with respect to height and several high-heritability disease traits [55]. Three groups of researchers analysed the genomes of huge populations for genetic variants associated with the height, which has a heritability of 80%, in the ‘broad sense’. (‘Narrow-sense’ heritability only includes ‘additive’ genetic variance that responds to selection, while broad-sense heritability includes all genetic variance.) More than 40 effects were veriﬁed by detection in independent analyses, but these altogether accounted for little more than 5% of the variance. Similar results were found for autism and schizophrenia, even though both traits have very high heritability. Various explanations have been advanced to explain these disappointing results. Firstly, sample sizes of tens of thousands of individuals are required to obtain the statistical power to detect QTL explaining < 0.5% of the total variance [56]. For example, with a sample size of 10 000, the statistical power to detect a variant that explains 0.2% of the variance is only 29%. Therefore, the chance that two independent studies of this size will both detect this variant is only 0.292=0.08. A signiﬁcant fraction of the genetic variation may be the result of rare alleles, which are very difﬁcult to detect [55]. Also genes affecting only the non-additive genetic variance will not be detected by most models used for analysis of SNP effects. Finally, a large fraction of the genetic variation in quantitative traits may be the result of polymorphisms in DNA other than SNPs, for example copy number variation [55].

Genomic Selection in Plants Until the development of high-throughput genotyping methodology for SNPs, the application of MAS in plants

g

55

was clearly ahead of application in animals. Now the situation has been reversed. Most applications of GS have so far been in animals [57]. The costs of producing and scoring each individual plant are generally much lower than for animals. Thus plant MAS programmes have generally been based on generating populations speciﬁcally for MAS, while animal programmes have been based nearly entirely on analysis of animals from the commercial population. For a simulated breeding programme of maize based on doubled haploids generated from inbred lines, the response of genome-wide selection was 18–43% larger than for marker-assisted recurrent selection [58]. Efﬁciency of GS with low-cost genotyping in a composite line from a cross between inbred lines was evaluated for a trait with heritability of 0.10 or 0.25 using a lowdensity marker map [59]. GS was the sum of estimates of effects of all marker intervals across the genome, ﬁtted either as ﬁxed or random effects. Reponses to selection over 10 generations, starting from the F2, were compared with standard BLUP selection. Both genomic strategies outperformed BLUP selection, especially in initial generations. Random GS outperformed ﬁxed GS in early generations and performed slightly better than ﬁxed GS in later generations. Random GS gave higher genetic gain when the number of marker intervals was greater, whereas ﬁxed GS gave higher genetic gain when the number of marker intervals was low.

Conclusions and Future Prospects As demonstrated by ‘The Pirate Bay’ (http://thepiratebay. org/), the technology always wins. This will also be the case for genetic markers. Genotyping costs will continue to decline. Human GWAS will now be based on a new family of genotyping chips designed to simultaneously assay 4 million sites of variation. In the near future, new SNP chips in cattle will be developed with much higher density than the BovineSNP50 BeadChip (D. Schnabel, personal communication). Complete Genomics (Mountain View, CA, USA) plans to offer complete human genome sequence per individual for US$5000 next year [60]. Increasing SNP density will increase the level of LD between SNP genotypes and the QTLs. LD is generally measured by the parameter r2 [61]. Simulation studies demonstrate that the accuracy of GEBV increases dramatically with an increase in average r2 between adjacent markers [62]. However, the development of a new higher-density chip would probably require reanalysis of all bulls with evaluations. This, in fact, may be advantageous in that LD relationships will tend to decay over time. Thus, it will probably be necessary to reanalyse populations after several generations in any event. The number of young bulls that could be progeny tested was limited by the cost of keeping each bull for 5 years, and the potential pool of cows that could be

CAB International 2010

56

Animal Science Reviews 2010

mated to these bulls. Even at $400 per individual, genotyping costs are much less that the cost of producing a progeny-tested bull. Recently, genotyping costs for the BovineSNP50 BeadChip have declined to $175 per individual (D. Pomp, personal communication). With the reduction of genotyping costs it becomes possible to increase the number of bull candidates. Also as genotyping costs decrease, genotyping of potential bull dams becomes economically viable [63]. Since bull dams generally have reliabilities of only 0.3 to 0.4 for milk production traits, this could also make a signiﬁcant increase in rates of genetic gain. For secondary traits with lower heritability, such as fertility, the potential gain will be even greater [42]. Finally with the information on the genotypes of individual animals, it becomes easier to control inbreeding. Current methods of genetic analysis include the complete relationship matrix. For young bulls without records, until now, this has been the only source of information. Therefore, bull candidates are generally all sons of the sires and dams with the highest evaluations. This tends to increase inbreeding in the population. With genomic data as a new source of information, it should be possible to increase the pool of potential parents of candidate bulls [42]. So far, no studies have dealt with economic optimization of breeding programmes under the new realities.

Acknowledgements This research was supported by grants from the Israel milk marketing board and the European Sixth Research and Technological Development Framework Programme, Proposal No. 016250-2 SABRE. I thank M. Ron and M. Shani for their critical reading and suggestions.

References 1. Fisher RA. The correlation between relatives on the supposition of Mendelian inheritance. Transactions of the Royal Society of Edinburgh 1918;52:399–433. 2. Sax K. The association of size differences with seed-coat pattern and pigmentation in Phaseolus vulgaris. Genetics 1923;8:552–60. 3. Neimann-Sørensen A, Robertson A. The association between blood groups and several production characters in three Danish cattle breeds. Acta Agriculturae Scandinavica Section A – Animal Science 1961;11:163–96. 4. Litt M, Luty JA. A hypervariable microsatellite revealed by in vitro ampliﬁcation of a dinucleotide repeat within the cardiac muscle actin gene. American Journal of Human Genetics 1989;44:397. 5. Marth G, Yeh R, Minton M, Donaldson R, Li Q, Duan SG, et al. Single-nucleotide polymorphisms in the public domain: how useful are they? Nature Genetics 2001;27:371–2. 6. Spelman RJ, Garrick DJ, van Arendonk JAM. Utilization of genetic variation by marker assisted selection in commercial

g

dairy cattle populations. Livestock Production Science 1999;59:51–60. 7. Schaeffer LR. Strategy for applying genome-wide selection in dairy cattle. Journal of Animal Breeding and Genetics 2006;123:218–23. 8. Glowatzki-Mullis ML, Gaillard C, Wigger G, Fries R. Microsatellite based parentage control in cattle. Animal Genetics 1995;27:7–12. 9. Werner FAO, Durstewitz G, Habermann FA, Thaller G, Kramer W, Kollers S, et al. Detection and characterization of SNPs useful for identity control and parentage testing in major European dairy breeds. Animal Genetics 2004;35:44–9. 10. Bonin A, Bellemain E, Eidesen PB, Pompanon F, Brochmann C, Taberlet P. How to track and assess genotyping errors in population genetics studies. Molecular Ecology 2004;13:3271–73. 11. Kennedy GC, Matsuzaki H, Dong SL, Liu WM, Huang J, Liu GY, et al. Large-scale genotyping of complex DNA. Nature Biotechnology 2003;21:1233–7. 12. Anderson EC, Graza JC. The power of single nucleotide polymorphisms for large scale parentage inference. Genetics 2006;172:2567–82. 13. Weller JI. Quantitative Trait Loci Analysis in Animals, 2nd edn. CABI Publishing, London, UK; 2009. 14. Weller JI, Kashi Y, Soller M. Power of ‘daughter’ and ‘granddaughter’ designs for genetic mapping of quantitative traits in dairy cattle using genetic markers. Journal of Dairy Science 1990;73:2525–37. 15. Lander ES, Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting reporting linkage results. Nature Genetics 1995;11:241–7. 16. Cohen-Zinder M, Seroussi E, Larkin DM, Loor JJ, Everts-van der Wind A, Lee JH, et al. Identiﬁcation of a missense mutation in the bovine ABCG2 gene with a major effect on the QTL on chromosome 6 affecting milk yield and composition in Holstein cattle. Genome Research 2005;15:936–44. 17. Grisart B, Coppieters W, Farnir F, Karim L, Ford C, Berzi P, et al. Positional candidate cloning of a QTL in dairy cattle: identiﬁcation of a missense mutation in the bovine DGAT1gene with major effect on milk yield and composition. Genome Research 2002;12:222–31. 18. Bennewitz J, Reinsch N, Reinhardt F, Liu Z, Kalm E. Top down preselection using marker assisted estimates of breeding values in dairy cattle. Journal of Animal Breeding and Genetics 2004;121:307–18. 19. Boichard D, Fritz S, Rossignol MN, Guillaume F, Colleau JJ, Druet T. Implementation of marker-assisted selection: practical lessons from dairy cattle. In: Proceeding of the 8th World Congress of Genetics Applied to Livestock Production Belo Horizonte (MG), Brazil; 2006. CD Communication No. 22-11. 20. Meuwissen THE, Goddard ME. Fine mapping of quantitative trait loci using linkage disequilibria with closely linked marker loci. Genetics 2000;155:421–30. 21. Sargolzaei M, Schenkel FS, Jansen GB, Schaeffer LR. Extent of linkage disequilibrium in Holstein cattle in North America. Journal of Dairy Science 2008;91:2106–17.

CAB International 2010

Joel I. Weller 22. Grapes L, Dekkers JC, Rothschild MF, Fernando RL. Comparing linkage disequilibrium-based methods for ﬁne mapping quantitative trait loci. Genetics 2004;166:1561–70. 23. Goddard ME, Hayes B, McPartlan H, Chamberlain AJ. Can the same genetic markers be used in multiple breeds? In: Proceeding of the 8th World Congress of Genetics Applied to Livestock Production, Belo Horizonte (MG), Brazil; 2006. CD Communication No. 22-14. 24. Loberg A, Du¨rr J. Interbull survey on the use of genomic information. In: Interbull Bulletin No. 39. Proceedings of the Interbull International, Uppsala, Sweden; 2009. p. 3–14. 25. Cole JB, VanRaden PM, O’Connell JR, Van Tassell CP, Sonstegard TS, et al. Distribution and location of genetic effects for dairy traits. Journal of Dairy Science 2009;92: 2931–46. 26. VanRaden PM, Van Tassell CP, Wiggans GR, Sonstegard TS, Schnabel RD, Taylor JF, et al. Invited review: Reliability of genomic predictions for North American Holstein bulls. Journal of Dairy Science 2009;92:16–24.

57

39. Hayes BJ, Chamberlain A, Goddard ME. Use of linkage markers in linkage disequilibrium with QTL in breeding programs. In: Proceeding of the 8th World Congress of Genetics Applied to Livestock Production, Belo Horizonte (MG), Brazil; 2006. CD Communication No. 30-06. 40. Chamberlain AJ, McPartlan HC, Goddard ME. The number of loci that affect milk production traits in dairy cattle. Genetics 2007;177:1117–23. 41. Ron M, Feldmesser E, Golik M, Tager-Cohen I, Kliger D, Reiss V, et al. A complete genome scan of the Israeli Holstein population for quantitative trait loci by a daughter design. Journal of Dairy Science 2004;87:476–90. 42. Hayes BJ, Bowman PJ, Chamberlain AJ, Goddard ME. Invited review: Genomic selection in dairy cattle progress and challenges. Journal of Dairy Science 2009; 92:433–43.

27. Goddard ME, Hayes BJ. Genomic selection. Journal of Animal Breeding and Genetics 2007;124:323–30.

43. Aguilar I, Misztal I, Johnson LD, Legarra A, Tsuruta S, Lawlor TJ. A uniﬁed approach to utilize phenotypic, full pedigree and genomic information for genetic evaluation of Holstein ﬁnal score. Journal of Dairy Science 2010;93:743–52.

28. VanRaden PM, Wiggans GR. Derivation, calculation and use of national animal model information. Journal of Dairy Science 1991;74:2737–46.

44. Meuwissen THE, Hayes BJ, Goddard ME. Prediction of total genetic value using genome-wide dense marker maps. Genetics 2001;157:1819–29.

29. Thomsen H, Reinsch N, Xu N, Looft C, Grupe S, Kuhn C, et al. Comparison of estimated breeding values, daughter yield deviations and deregressed proofs within a whole genome scan for QTL. Journal of Animal Breeding and Genetics 2001;118:357–70.

45. Liu Z, Seefried F, Reinhardt F, Reents R. Dairy cattle genetic evaluation using genomic information. In: Interbull Bulletin No. 39. Proceedings of the Interbull International, Uppsala, Sweden; 2009. p. 23–8.

30. Israel C, Weller JI. Estimation of candidate gene effects in dairy cattle populations. Journal of Dairy Science 1998;81:1653–62. 31. Calus MPL, Veerkamp RF. Accuracy of breeding values when using and ignoring the polygenic effect in genomic breeding value estimation with a marker density of one SNP per cM. Journal of Animal Breeding and Genetics 2007;124:362–8. 32. VanRaden PM. Efﬁcient methods to compute genomic predictions. Journal of Dairy Science 2008;91:4414–23. 33. Beavis WD. The power and deceit of QTL experiments: lessons for comparative QTL studies. In: Annual Corn Sorghum Research Conference, Volume 49. Washington, DC; 1994. p. 252–68. 34. Georges M, Nielsen D, Mackinnon M, Mishra A, Okimoto R, Pasquino AT, et al. Mapping quantitative trait loci controlling milk production in dairy cattle by exploiting progeny testing. Genetics 1995;139:907–20. 35. Goring HH, Terwilliger JD, Blangero J. Large upward bias in estimation of locus-speciﬁc effects from genomewide scans. American Journal of Human Genetics 2001;69:1357–69. 36. Meuwissen THE, Goddard ME. Mapping multiple QTL using linkage disequilibrium and linkage analysis information and multitrait data. Genetics, Selection, Evolution 2004;36:271–79. 37. Hayes B, Goddard ME. The distribution of the effects of genes affecting quantitative traits in livestock. Genetics, Selection, Evolution 2001;33:209–29. 38. Weller JI, Shlezinger M, Ron M. Correcting for bias in estimation of quantitative trait loci effects. Genetics, Selection, Evolution 2005;37:501–22.

g

46. Guillaume F, Fritz S, Boichard D, Druet T. Short communication: correlations of marker-assisted breeding values with progeny-test breeding values for eight hundred ninety-nine French Holstein bulls. Journal of Dairy Science 2008;91:2520–22. 47. Party C, Ducrocq V. Bias due to genomic selection. In: Interbull Bulletin No. 39. Proceedings of the Interbull International, Uppsala, Sweden; 2009. p. 77–82. 48. Legarra A, Aguilar I, Misztal I. A relationship matrix including full pedigree and genomic information. Journal of Dairy Science 2009;92:4656–66. 49. Misztal I, Legarra A, Aguilar I. Computing procedures for genetic evaluation including phenotypic, full pedigree, and genomic information. Journal of Dairy Science 2009;92: 4648–55. 50. Hayes BJ, Visscher PM, Goddard ME. Increased accuracy of artiﬁcial selection by using the realized relationship matrix. Genetic Research, Cambridge 2009;91:47–60. 51. Visscher PM. Whole genome approaches to quantitative genetics. Genetica 2009;136:351–8. 52. Aulchenko YS, de Koning DJ, Haley C. Genomewide rapid association using mixed model and regression: a fast and simple method for genomewide predigree-based quantitative trait loci association analysis. Genetics 2007;177:577–85. 53. Weigel KA, de los Campos G, Gonza´lez-Recio O, Naya H, Wu XL, Long N, et al. Predictive ability of direct genomic values for lifetime net merit of Holstein sires using selected subsets of single nucleotide polymorphism markers. Journal of Dairy Science 2009;92:5248–57.

CAB International 2010

58

Animal Science Reviews 2010

54. Solberg TR, Sonesson AK, Woolliams J, Meuwissen THE. Reducing dimensionality for prediction of genome-wide breeding values. Genetics, Selection, Evolution 2009;41:29.

59. Piyasatian N, Fernando RL, Dekkers JC. Genomic selection for marker-assisted improvement in line crosses. Theoretical and Applied Genetics 2007;115:665–74.

55. Maher B. The case of the missing heritability. Nature 2008;456:18–21.

60. Hayden EC. Genomics shifts focus to rare diseases. Nature 2009;461:458–9.

56. Visscher PM. Sizing up human height variation. Nature Genetics 2008;40:489–90.

61. Hill WG. Estimation of effective population size from data on linkage disequilibrium. Genetic Research 1981;38:209–16.

57. Piepho HP. Ridge regression and extensions of genomewide selection in maize. Crop Science 2009;49: 1165–76.

62. Calus MP, Meuwissen THE, de Roos AP, Veerkamp RF. Accuracy of genomic selection suing different methods to deﬁne haplotypes. Genetics 2008;178:553–61.

58. Bernardo R, Jinming Y. Prospects for genomewide selection for quantitative traits in maize. Crop Science 2007;47: 1082–90.

63. Schaeffer LR. Strategy for applying genome-wide selection in dairy cattle. Journal of Animal Breeding and Genetics 2006;123:218–23.

g

CAB International 2010

Animal Science Reviews 2010

Quantitative trait mutations in cattle, sheep and pigs: a review Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens and Nadine Buys* Address: Laboratory of Livestock Physiology, Immunology and Genetics, Department of Biosystems, KULeuven, Kasteelpark Arenberg 30, 3001 Leuven, Belgium. *Correspondence: Nadine Buys. Fax: +32(0)16/32 19 94. Email: [email protected] 19 October 2009 4 February 2010

Received: Accepted:

Abstract Most economically important traits in farm animals, such as daily gain, muscularity, meat quality, milk production, reproduction and others, are complex multifactorial traits that are controlled by an unknown number of genes combined with environmental factors. Since these traits show a continuous distribution rather than discrete values as monogenic or qualitative traits do, they are called quantitative traits and a polymorphism that affects such a quantitative trait is called a quantitative trait mutation (QTM). During the last few decades, several QTMs for different economically important quantitative traits in farm animals, such as muscularity, meat quality and milk production, were discovered. Also, for various congenital disorders, causal polymorphisms could be found or are under investigation. However, for other quantitative traits, no such QTMs have been revealed so far. An explanation for this discrepancy lies in the extent of the effect of the different polymorphisms that underlie a certain quantitative trait since this effect can vary from very small to quite large (up to 25–30% of the total phenotypic variance of a particular trait). Although new, emerging cost-effective genome-wide single-nucleotide polymorphism (SNP) genotyping techniques have become available, only QTMs that explain a large portion of the phenotypic variation are worth unravelling, especially in the light of genome-wide selection. In this review, the research for QTMs is positioned against these contemporary techniques. Moreover, an overview is given of, past and present, research efforts in identifying QTMs in farm animals and the incorporation of these polymorphisms in modern animal breeding. Keywords: Farm animals, Quantitative trait mutations, Candidate gene approach, Linkage analysis, Linkage disequilibrium, Whole genome association analysis, Genomic selection Review Methodology: We searched the following databases: Entrez Pubmed, AnimalQTLdb and Web of Science (Keyword search terms used: genomic selection, quantitative trait nucleotides in farm animals, causative mutations, animal breeding, etc.). In addition, we used the references from the articles obtained by this method to check for additional relevant material.

Introduction Since animal breeding is especially focused on economically important traits, farm animals are of particular interest in the search for genes that control growth, muscularity, meat quality, behaviour, reproduction and other traits [1]. Moreover, since productivity is dependent on good health, farm animals are also very important in the research for disease-related genes [2]. Some of the researched traits have a simple monogenic basis: the so-called qualitative traits. However, this is the

g

case for a minority of the traits. Most economically important traits, such as muscularity and reproduction, are complex multifactorial traits that are controlled by an unknown number of genes – a few with a major effect and a lot with a minor effect – in combination with several environmental factors. Since these traits show a continuous distribution rather than discrete values as qualitative traits do, they are called quantitative traits. Therefore, the region on a chromosome that bears one or more genes that affect a quantitative trait is called a quantitative trait locus (QTL) [1, 3]. Until recently,

CAB International 2009

60

Animal Science Reviews 2010

nucleotide variation underlying a QTL was called a quantitative trait nucleotide (QTN) [4]. In 2009, however, Leif Andersson introduced the expression ‘quantitative trait mutation’ (QTM), since structural mutations such as copy number variation, insertions, deletions and inversions can also be causative mutations for some QTLs [5]. The identiﬁcation of genes for complex traits would greatly enhance our understanding of quantitative traits, but there would also be a great practical beneﬁt to livestock breeding. Traditionally, the genetics of complex traits in domestic animal species has been studied without studying the genes underlying these traits. Selection was based on estimated breeding values calculated from phenotypic records and pedigrees, and on knowledge of the heritability of a trait [6]. Nowadays, QTL information is used in breeding programmes to improve a range of traits, using marker-assisted selection (MAS) with markers linked to the QTM, markers in linkage disequilibrium (LD) with the QTM or, ultimately, the QTM itself [2, 7]. The research for markers applicable in MAS in farm animals has taken a giant leap forward during the last two decades. In this review, it is explained why this research was given such a priority and why farm animals, in particular, are such prized subjects for this research. Also, an overview is given of the different research methods used in the search for markers, and more speciﬁcally QTM, during the last few decades. The use of QTM and MAS is positioned against a background of the most contemporary techniques using single-nucleotide polymorphism (SNP) chips. Finally, an overview is given of known QTMs involved in production and reproduction traits and in congenital disorders in farm animals.

The Genome of Farm Animals: A Wealth of Functional Mutations The magnitude of variation in domestic animals was created during the 10 000-year-old process of domestication and selective breeding. This process can be seen as an extensive screening for genetic variants with a phenotypic effect and without a deleterious effect on ﬁtness [5]. Also, this process led to a rather remarkable LD structure in domestic animals. This structure is best characterized in dogs, but is comparable in other farm animals [8–11]. Before the creation of different breeds, the LD extended over chromosome segments of 10 kb, with 3–5 haplotypes per LD block. These haplotypes were responsible for the major part of the genetic variation. About 200 years ago, the creation of the modern breeds began. They were created as genetically isolated, closed populations and, in almost all cases, experienced episodes of severe bottlenecks and long-time decreases of the effective population size. From a chromosomal point of view, the creation of breeds can be seen as the selection of a limited number of haplotypes from a large pool of ancestral chromosomes. Nowadays, the LD within

g

different dog breeds still extends over a very long chromosomal segment, since not enough time has passed for recombination to have mixed the different haplotypes. Therefore, within breeds, the long-range haplotypes still span 0.5–1.0 Mb with a small number of haplotypes (3–5) that account for the majority of the observed variation. Within the different long-range haplotypes, however, independent of the breed, the same short-range haplotypes can be found. This short-range haplotype corresponds to one of the haplotype variants that were present in the original population, before breed creation [9]. It is this remarkable genome structure that makes domestic animals such appreciated models for the exploration of genotype–phenotype relationships of quantitative traits. One of the major drawbacks of positional cloning of genes with an inﬂuence on quantitative or complex traits is their multifactorial character. This makes it very hard to detect the effects of the different QTLs and therefore, large groups with thousands of individuals would be needed in QTL mapping experiments. However, in domestic animals, the localization of genes underlying complex traits is simpliﬁed because of the recent creation of breeds and the particular LD structure that arose from it [1, 3, 7]. The complexity of many economically valuable traits has decreased because of the ancestral effects that accompanied the creation of different breeds. This is because, in domestic animals, the effective population size is very small. For cattle and pigs, the effective population size of most breeds is about 100 animals [7, 9]. Because of this, the number of animals necessary to localize genes that underlie a certain quantitative trait is reduced drastically. For disease alleles causing a simple Mendelian dominant trait, the probability of detecting the locus is over 99% given 100 affected and 100 unaffected animals [9]. For recessive Mendelian traits, detecting the locus is already possible with smaller number of animals. Karlsson et al. [12] successfully mapped recessive Mendelian trait loci with 10 affected and 10 unaffected dogs [12]. Also, in cattle, recessive disorders were characterized based on 12–30 animals [13]. For mapping risk factors that lead to a 2–5 higher incidence of a polygenic disorder, it was estimated that 200 animals should be sufﬁcient [9].

QTM Mapping Throughout (Recent) History To date, thousands of QTLs for numerous traits have been reported (e.g. see http://www.animalgenome.org/ QTLdb/). However, the identiﬁcation of the mutations that underlie these QTLs, the so-called QTMs, has been hard [5].

The Candidate Gene Approach Over the past 20 years, two approaches have been used to discover genes and polymorphisms underlying the

CAB International 2009

Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens and Nadine Buys

phenotypic variation in complex traits. The ﬁrst approach is the so-called candidate gene approach, which is based on the identiﬁcation of segregating marker alleles in a candidate gene associated with the trait of interest [6]. The approach can be justiﬁed when candidate genes fulﬁl one or more of the four criteria. The ﬁrst two criteria involve the function of the candidate gene: the gene either has a known physiological role in the trait of interest or the gene affects the trait of interest based on studies of knockouts, mutations or transgenics in other species. The next two criteria involve time and place of expression of the candidate gene: the gene is preferentially expressed in organs related to the quantitative trait and the gene is preferentially expressed during developmental stages related to the phenotype [14, 15]. A problem with the candidate gene approach is that a lot of genes fulﬁl at least one criterion and, until now, none of the genes in which a QTM was discovered meets all four of them [15]. On the other hand, the advantage of the candidate gene approach is that it can be very powerful and can detect loci even with small effects, provided that the candidate gene harbours a true causative mutation [1]. Examples of QTM that were found using the candidate gene approach include the QTM in Ryanodine receptor 1 (RYR1) [16] and Myostatin (MSTN) in cattle [17–19] and dogs [20],

Using Linkage and LD The second approach used to detect QTM in domestic animals in the past is used to map the genes that affect the trait of interest to a chromosomal location using genetic markers, also called positional cloning. Until recently, a whole genome scan was performed using linkage and low-density microsatellite panels [21], and the progress in mapping causal genes for complex traits has been slow because of the techniques used [6]. Linkage analysis uses recombination events in the pedigree and traces chromosome segments to a common ancestor [6]. This approach, however, gives a poor resolution and conﬁdence intervals for the detected QTL are typically 10–20 Mb wide and may contain hundreds of genes. Therefore, this method should be followed by LD ﬁne mapping of the most prominent QTL via high density, de novo developed, SNP maps [5, 7]. LD mapping depends on chromosome segments inherited before the recorded pedigree, since it is the inheritance of identical chromosome segments by multiple descendents from a common ancestor that causes LD. The use of LD allows one to do precise mapping, because LD decays as the distance between the marker and the QTL increases [6]. In cattle, an LD measure of 50% was noticed for marker pairs separated by less than 5 cM, but a rapid decay to 16% was witnessed, reaching a plateau slightly below 14% for more distant markers [22]. Identiﬁcation and production of dense SNP marker panels to reduce the conﬁdence

g

61

intervals of the most eminent QTL was expensive and time-consuming, until the arrival of genome-wide SNP panels [6]. QTMs detected using linkage and LD scanning include the mutations in insulin-like growth factor 2 (IGF2) [23] and in protein kinase AMP-activated gamma 3 non-catalytic subunit (PRKAG3) in pigs [24, 25], in MSTN in mice and sheep [26, 27], in acylCoA:diacylglycerol acyltransferase (DGAT1) in cattle [28], the callipyge (CLPG) gene [29, 30] and the Booroola fecundity (Fec) gene in sheep [31].

Whole-genome Association Analysis (WGAA) Nowadays, high-density genome-wide SNP maps have been developed for almost all domestic animals and, at the moment, it is possible to genotype 20–60 000 SNP markers using random SNP chips, available for cattle, horse, sheep, pig, dog and chicken. Using the SNP chips (instead of applying the traditional separate mapping and ﬁne-mapping strategy), it is possible to merge two steps into a single combined linkage and LD analysis or direct WGAA [7, 32]. The basic design of a WGAA study is that a certain trait is recorded in a group of animals and a genome-wide panel of SNP markers is assayed to detect signiﬁcant associations between the trait of interest and any of the markers. The validated markers of the WGAA study can subsequently be used in MAS in domestic animals, in two different ways [6]. A ﬁrst method for implementing validated WGAA markers into MAS is to make direct use of markers that are in LD with the QTM. To avoid bias, the effect associated with each allele of the signiﬁcant marker or markers is estimated in a population that is independent from the one in which the signiﬁcant markers were found. Breeding values for selection candidates can then be estimated combining pedigree, markers and phenotype information [6]. This type of MAS has already been applied, in practice, but it has some major disadvantages. Firstly, alleles at linked markers cannot be used to predict the phenotype until the association between alleles at the marker and alleles at the trait gene is known (called ‘phase’). This phase, however, is only valid within the tested population and may change in subsequent generations through recombination [2]. Therefore, for every new population, a new association analysis has to be done. Secondly, this kind of MAS is mainly a tool for traits that are easy to measure and are therefore recorded routinely. However, many traits that have an important impact on proﬁtability and welfare, such as feed conversion efﬁciency, fertility and disease resistance, are difﬁcult to measure. Obtaining the necessary phenotype information to carry out WGAA would be very expensive [2]. A possible solution for these disadvantages and a second method for implementing WGAA markers into MAS is to identify the functional mutations or QTMs underlying the traits of interest and use these as markers

CAB International 2009

62

Animal Science Reviews 2010

for MAS. Since it concerns causal mutations, it is possible to predict the effect of a particular allele in all animals in a population, without having to determine the phase [2]. However, the identiﬁcation of QTMs also requires considerable investment. Therefore, the search for QTMs is only economically feasible if it concerns mutations in genes that explain a large portion of the phenotypic variance of a speciﬁc trait, or so-called major genes. Research for QTMs using WGAA has been successful for genes causing simple Mendelian traits [12, 13]. For the unravelling of quantitative traits, however, this approach has not yet proven to be successful.

QTM in Production and Reproduction Traits in Farm Animals As mentioned in the previous paragraph, during the last few decades, several QTMs for different economically important quantitative traits related to production and reproduction in farm animals were discovered using both the candidate gene approach and combined linkage and LD analysis. In this paragraph, an overview of these QTMs in different farm animal species is given.

Production Traits RYR1 The ﬁrst QTN involved in muscularity in pigs was detected in the RYR1 gene, the gene encoding the predominant Ca2+ release channel in the sarcoplasmic reticulum of skeletal muscle. A substitution at position 1843 in the porcine RYR1 gene (Ryr1 g.1843C>T), leading to an alteration from an arginine residue to a cysteine residue, was associated with higher lean meat content and muscularity in different pig breeds, including the heavily muscled Pie´train breed. However, the mutation was also associated with malignant hyperthermia, which is a disorder that can be triggered by inhalational anaesthetics such as halothane or by skeletal muscle relaxants such as succinylcholine. In pigs, homozygous for the defect, malignant hyperthermia can also be triggered by stress and this is referred to as the porcine stress syndrome (PSS) in pigs [16]. PSS is a disorder witnessed typically in Pie´train, Poland, China and strains of Landrace pigs, and is triggered by stimuli as transportation, increase in environmental temperature, physical exercise, restraint or general handling. The PSS is characterized by muscle tremor, increased respiratory rate, systemic acidosis and a rise in body temperature. Typically, animals that die in this state, either naturally or by slaughtering, exhibit an immediate rigor mortis, leading to the so-called pale, soft and exudative (PSE) meat [33]. This post-mortem manifestation of the PSS leads to major economical losses in the pig breeding industry and is an important downside to the selection for the RYR1 g.1843C>T polymorphism [16].

g

PRKAG3 The rendement Napole or RN7 phenotype is common in Hampshire pigs and is characterized by a 70% increase in glycogen content in skeletal muscle and large effects on meat characteristics. Meat from RN7 pigs has a low ultimate pH, a reduced water-holding capacity, and gives a reduced yield of cured cooked ham [24, 25]. However, animals bearing the mutation also show higher growth rates and meat content in the carcass [34]. In 1996, the RN gene was mapped to porcine chromosome 15 [35]. A candidate gene for this QTL was the muscle-speciﬁc PRKAG3, which encodes the gamma 3 isoform of AMPactivated protein kinase (AMPK). This enzyme can inactivate glycogen synthase, the key regulatory enzyme of glycogen synthesis by phosphorylation [24]. A nonconservative, causative arginine to glutamine substitution (R200Q) was found in the most conserved region of the gene denoted as RN7. It concerns an activating mutation that increases glucose uptake but it may also have other effects like fatty acid oxidation or synthesis. The mutation was supported by very strong genetic evidence [24, 25]. A second mutation (V199I) was found by Lindahl et al. [36, 37], denoted as rn*. This gives three alleles as followed: 199V-200R for the wild-type rn+, 199V-200Q for RN7 and 199I-200R for rn* [36]. The RN7 allele was dominant over rn* and rn+ concerning ultimate pH, water-holding capacity and cooking loss but with regard to residual glycogen content RN7 was only dominant in entire male pigs and not in female pigs, where rn* had a glycogen-lowering effect [36]. Both mutations also have an inﬂuence on the colour of pork meat [37]. IGF2 In 1998, a paternally expressed QTL with large effects on muscularity, heart weight and fat deposition was described at the distal tip of porcine chromosome 2 [38–40]. Based on the conserved synteny between the QTL on Sus scrofa chromosome 2 (SSC2) and human chromosome 11p, IGF2 was suggested as a candidate gene for this QTL in an intercross between European wild boar and Large White domestic pigs and another intercross between Large White and Pie´train domestic pigs [39, 40]. The IGF2 protein plays a key role in skeletal muscle growth, since it has an effect on carbohydrate and fat metabolism, protein turnover, growth and differentiation in skeletal muscle [41, 42]. The effects of the identiﬁed QTL on muscle mass and fat deposition are of the same magnitude as those reported for the RYR1 locus, although apparently without the associated negative effects on meat quality. It is estimated that both loci jointly explain about 50% of the phenotypic difference for muscularity and leanness between Large White and Pie´train [40]. In 2003, a QTM was discovered in the porcine IGF2 gene. A substitution in an evolutionary conserved CpG island in intron 3 of the IGF2 gene (IGF2 intron3-g.G>A) leads to an abrogation of the interaction of a regulatory

CAB International 2009

Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens and Nadine Buys

63

locus in the IGF2 gene and a repressor protein. This subsequently leads to a threefold increase in IGF2 mRNA levels and a subsequent increase in muscle growth and heart size and a decrease in fat deposition [23]. Oczkowicz et al. [43] noticed signiﬁcant effects of this mutation on body composition and growth performance and no signiﬁcant effect on meat quality in Polish pig breeds [43]. In 2004, also the existence of porcine IGF2 antisense transcript (IGF2-AS) was reported. Its expression was also shown to be imprinted, paternally expressed and affected by the QTM [44].

genes DLK1 and PEG11, and it is the overexpression of DLK1 that causes the muscular hypertrophy. In animals that are homozygous for the CLPG mutation, both the maternally and the paternally expressed genes are overexpressed, but since this leads to a post-transcriptional trans-inhibition of the paternally expressed proteinencoding genes by the maternally expressed miRNA genes, no phenotypic difference can be seen in these animals [7].

CLPG In 1993, a mutation causing muscular hypertrophy, associated with leanness and improved feed efﬁciency was identiﬁed in the Dorset sheep breed. Animals bearing the mutation show little external fat and unusual muscling in the hind quarters. Therefore, the locus in which this mutation occurred was called the callipyge (CLPG) locus, from the Greek calli- which means beautiful and -pyge buttocks [45, 46]. The CLPG locus was mapped to a 400 kb chromosome fragment of ovine chromosome 18 and displays an unusual parent-of-origin effect referred to as polar overdominance: only heterozygous individuals having inherited the CLPG mutation from their sire exhibit the muscular hypertrophy [29, 30]. In 2002, an A to G mutation causing the CLPG phenotype was located in a dodecamer region that was highly conserved in sheep, cattle, human and mouse genome sequences. Although the mutation was located outside the boundaries of any previously identiﬁed transcript, it was proven to be located within an (unknown) expressed transcript and shown to alter a serine codon into a proline residue [47, 48]. Within the CLPG region, different genes that are preferentially expressed in skeletal muscle and subject to parental imprinting in this tissue were studied. The results showed that the CLPG mutation did not alter the imprinting status of the genes, but it enhanced their expression levels in cis, while maintaining their exclusive expression from either the paternal or maternal allele. This indicated that the CLPG mutation affects a long-range control element and either boosts an enhancer or inhibits a silencer regulatory element [30, 48]. As only the heterozygous individuals that inherited the CLPG mutation from their father (+mat/CLPGpat) exhibited the CLPG phenotype, it was hypothesized that the mutation involves a physiological interaction in trans between a maternally expressed repressor and its paternally expressed growthpromoting target [48]. From this hypothesis, the actual working mechanism could be deduced. Animals that inherit the CLPG mutation from the mother only overexpress the maternally expressed long non-coding RNA genes hosting tandem clusters of C/D small nucleolar RNAs and microRNAs present in the CLPG locus. Animals that inherit the CLPG mutation from the father only overexpress the paternally expressed protein-encoding

Naturally occurring mutations of the MSTN gene were found in several species, including cattle, mice, humans, sheep and dogs [17–20, 26, 27, 49–51]. The ﬁrst species in which a mutation of the MSTN gene was found was cattle. In total, 12 different mutations were found in the MSTN gene of the so-called double-muscled cattle, 8 of them were located in the coding sequence and 4 were located in the intronic sequences of the gene [17–19, 49–51]. Of the 8 polymorphisms that were located in the coding region of the MSTN gene, 6 mutations led to a disruption of the function of the protein, one was a conservative mutation and one was a silent mutation. Of the 11 different double-muscled cattle breeds, the analysed double-muscled animals were either homozygous for one of the mutations predicted to disrupt the function of the MSTN protein or compound heterozygous for two distinct of these mutations. In two of the double-muscled breeds studied, namely Limousin and Blonde d’Aquitaine, there was no clear evidence for the role of the MSTN loss-of-function mutations in double muscling in these breeds. This suggests that either MSTN is not responsible for the muscular hyperthrophy in these breeds or there are additional mutations in bovine MSTN that are not located in the coding sequence of the gene [49–51]. The hypermuscular compact (Cmpt) mouse was the ﬁrst animal species other than cattle that was found to have a naturally occurring functional mutation in the MSTN gene. It concerns a 12-bp deletion in the propeptide region of the MSTN gene and so the apparent loss of function of the MSTN protein is not because of disruption of the growth factor domain as in cattle. But the proregion may play a role in the proper folding, efﬁcient secretion or regulation of the target of the mature C-terminal portion of the MSTN gene. Since the deleted region is conserved in all known vertebrate MSTN genes, it is possible that one or more of these functions are affected, which may decrease the amount or the availability of the MSTN protein or confuse the proper targeting of the MSTN protein [26]. Recently, a mutation in the canine MSTN gene was found. A 2 bp deletion in the third exon of the MSTN gene in some individuals of the whippet dog breed led to a double-muscled phenotype known as the ‘bully’ whippet. This mutation caused the formation of a premature stop

g

MSTN

CAB International 2009

64

Animal Science Reviews 2010

codon at amino acid 313 of the canine MSTN, and subsequently led to a truncated protein [20]. In sheep, a deletion in the coding region of the MSTN gene of Norwegian White Sheep was found to disrupt the reading frame from amino acid position 320, ending in a premature stop codon in amino acid position 359 (c.960delG). As the bioactive part of MSTN is known to constitute the C-terminal domain, this mutation is most likely to generate a completely non-functional protein [52]. In addition to the c.960delG mutation, also a mutation was found in the non-coding region of the MSTN gene of the exceptionally muscled Texel breed. The possibility that a functional mutation could occur in the non-coding region of the MSTN gene was already postu0 0 lated by Grobet et al. [49]. In the 3 -UTR (3 -untranslated region) of the ovine MSTN gene, a guanine to adenine 0 transition (c.2630G>A) creates one of the 3 -UTR octamer motifs (ACATTCCA) that correspond to miRNA targets. This octamer is a binding site for mir1 and mir206, microRNAs that are highly expressed in skeletal muscle, and binding of these microRNAs causes translational repression of the MSTN gene and double-muscling in Texel sheep [27]. It was shown that ‘double-muscled’ sheep are either homozygous for the c.960del G or c.2360G>A mutation or compound heterozygous for both mutations. When comparing the phenotypes of sheep bearing different mutations, it could be seen that the most extreme phenotypes regarding carcass conformation and fatness were found in the animals that were homozygous for the c.960delG mutation, followed by the compound heterozygous animals and then the homozygous c.2360G>A group. An explanation for this difference can be found in the type of mutation. The deletion in the coding sequence of the MSTN gene causes a premature stop codon that leads to a non-functional protein. The c.2360G>A mutation causes a translational inhibition, which decreases the amount of circulating MSTN to approximately one-third compared with wildtype animals [52]. Also in humans, a functional mutation in the non-coding MSTN gene was found. It concerns a guanine to adenine transition at position +5 in intron 1 of the human MSTN gene. This is a common location of splicing and in vitro studies show that approximately 69% of the MSTN gene bearing the transition is misspliced, leading to a severely truncated protein, as also seen in cattle [53]. DGAT1 In 1998, a QTL with a major effect on milk fat content and other milk characteristics was mapped onto chromosome 14 in cattle [54]. A very strong positional candidate gene within the QTL conﬁdence interval was DGAT1, a gene that encodes diacylglycerol-O-acyltransferase, a microsomal enzyme that catalyses the ﬁnal step of triglyceride synthesis [28, 55]. In exon 8 of the DGAT1 gene, an ApA to GpC dinucleotide substitution was found that caused an amino acid substitution of a conserved positively

g

charged lysine residue into a neutral, hydrophobic alanine residue (K232A) [28]. This mutation leads to an increased activity of the enzyme and a subsequent higher production of triglycerides [56]. In 2004, an additional polymorphism was found at position 1465 in the promoter gene of the DGAT1 gene that was associated with milk fat percentage variation within the German Holstein population. Alleles of a variable number of tandem repeats (VNTR) polymorphism were found to be associated with milk fat content in animals that were homozygous for the lysine to alanine substitution. Since a potential transcription factor binding site is present within the VNTR, it also might concern a causal mutation [57]. ATP-binding cassette sub-family G member 2 (ABCG2) In 2005, another QTM for cattle milk production was proposed on chromosome 6, affecting milk fat and protein concentration [58]. Besides being a strong positional candidate gene, ABCG2 was also an interesting functional candidate, since ABCG2 is strongly expressed in the mammary gland of mice, cows and humans and it is responsible for the active secretion of clinically and toxicologically important substrates into mouse milk [59]. A single nucleotide change from A to C encoded for a tyrosine residue instead of a serine residue (Y581S) in the ABCG2 transporter. Animals with the tyrosine residue showed a decrease in milk yield and a change in milk composition with a higher fat percentage and a higher protein percentage compared with those with the serine residue [58, 60].

Reproduction Traits The B. fecundity gene and other genes involved in proliﬁcacy in sheep At a workshop on the Booroola Merino at Armidale, Piper and Bindon [61] were the ﬁrst to report the B. fecundity gene (FecB) as the ﬁrst major gene causing an increase in the ovulation rate in sheep. The effect of FecB is additive for ovulation rate (increasing by about 1.6 corpora lutea per cycle for each copy) and the FecB mutation increases litter size by one to two extra lambs with each copy [31]. The QTM underlying FecB was found to be a G to A transition at position 746 of the coding sequence of the Bone Morphogenetic Protein Receptor-type 1B (BMPR-1B) gene, located on sheep chromosome 6. This polymorphism leads to an alteration from a glutamine residue to an arginine residue at position 249 of the BMPR-1B protein, which is a member of the transforming growth factor-b (TGF-b) receptor family [31, 62–64]. Other mutations increasing the ovulation rate in sheep were found in two members of the TGF-b family: the Bone Morphogenetic Protein 15 (BMP15) and the Growth and Differentiation Factor-9 (GDF9) [65–67]. In the BMP15 gene on the X chromosome (FecX), not less than 5 independent mutations have been found that

CAB International 2009

Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens and Nadine Buys

all result in the same phenotype: one copy of the mutation gives an increase in ovulation rate, whereas two copies make the ewe sterile with small underdeveloped ovaries containing follicles with not more than one layer of granulose cells. The mutations in BMP15 are called FecXI, FecXH, FecXG, FecXB and FecXL. FecXI or the Inverdale mutation is a T to A transition at nucleotide position 896 in the cDNA, resulting in a substitution of valine with aspartic acid at amino acid 31 of the mature protein (V31D). FecXH is called the Hanna mutation. This is a transition of C to T residue at nucleotide position 871 and leads to a stop codon at position 23 of the mature protein (Q23stop). FecXG (Galway), FecXB (Belclare) and FecXL (Lacaune) are C to T, G to T and G to A transitions at nucleotides 718, 1100 and 1196 of the cDNA, respectively. They are responsible for a stop codon at position 239 of the unprocessed protein (T239stop), a substitution of a serine with an isoleucine at position 99 (S99I) and a substitution of a cysteine with a tyrosine at position 53 (C53Y) of the mature protein, respectively [64]. The mutation in the GDF-9 gene on chromosome 5 is an A to T transition at position 1184 of the cDNA is a substitution of a serine with a phenylalanine (S77F). It is called the High Fertility mutation (FecGH). Homozygous FecGH/FecGH ewes are (like the homozygous FecX mutant carriers) infertile but, in contrast to the FecX phenotype, their ovarian follicles develop to an abnormal type 5 early antral stage [68].

QTMs in Congenital Disorders in Farm Animals In addition to the search for genes involved in the ‘classical’ (re)production traits, the search for diseaserelated genes in farm animals becomes increasingly important. Several QTMs involved in the susceptibility to various (congenital) disorders in dogs, cattle and sheep are known or are under investigation. The advent of the SNP chip has speeded up the progress in this area. Examples are the discoveries of the genetic components of congenital muscular dystonia (CMD) types 1 and 2, foetal ichytiosis and complex vertebral malformation (CVM) in cattle [13, 69]. It has to be noted, however, that it concerns simple monogenic disorders and not the more complicated polygenic disorders. However, it is expected that risk factors that lead to a 2–5 higher incidence of a polygenic disorder could also be mapped using the SNP chip, screening as few as 200 animals [9].

Transmissible Spongiform Encephalopathy (TSE) TSEs are transmissible prion diseases, caused by abnormally folded prion proteins (PrPs) that induce the conversion of the non-infectious, cellular form of the host PrP into the abnormal, infectious form. This group of

g

65

diseases includes scrapie in sheep and goats, bovine spongiform encephalopathy (BSE) in cattle, chronic wasting disease in deer and elk, feline spongiform encephalopathy in cats, transmissible mink encephalopathy and variant Creutzfeldt–Jacob disease (CJD) in humans. All TSE diseases are dependent on the expression of the host PrP protein and substitutions in this gene are found to have direct consequences for TSE pathogenesis, TSE susceptibility and the host species barrier of transmissibility [70–72]. Scrapie is an invariably fatal, neurodegenerative disease in sheep and goats [73]. However, since scrapie outbreaks in goats in many countries are very rare, the study of scrapie in goats is not a priority. Therefore, in what follows, only the progressions in the research for scrapie in sheep will be discussed. Clinical observations of classical scrapie are increased anxiety, teeth grinding, pruritus, a crouched or wide-based stance, ataxia, hyperthermia, tremors and a loss of weight. Neuropathologically, the disease is characterized by the presence of vacuoles in neurons and neuropil, together with PrP proteins in the brain [74]. The host PrP protein is known to be a major component of scrapie and polymorphisms appear throughout the entire PrP gene. However, only three mutations in the part of the sequence that encodes for the highly structured C-terminal domain have been found to be associated unambiguously with TSE susceptibility and incubation period: an alanine to valine substitution at amino acid (AA) 136 (A136V), an arginine to histidine substitution at AA 154 (R154H) and a glutamine to arginine/histidine substitution at AA 171 (Q171R/H) [70, 75, 76]. Between these polymorphisms, a complex relationship exists. From a simplistic perspective, V136, R154 Q171 and H171 were linked to susceptibility to scrapie, while A136, H154 and R171 were linked to resistance. However, of the 12 possible combinations of these polymorphisms, only ﬁve occur with any frequency: ARR (A at codon 136, R at codon 154 and R at codon 171), ARQ, AHQ, ARH and VRQ [77, 78]. More recently, also the AHR and VRR alleles were reported, however, no information on the susceptibility to scrapie is available [79, 80]. When examining these alleles, it becomes apparent that a high degree of susceptibility is associated with the VRQ allele. Furthermore, the resistance conferred by the ARR allele is also very clear [70, 78]. Other alleles/genotypes that play a role in the susceptibility of scrapie include the ARQ/ARQ genotype, the ARQ and ARH alleles and the ARH/VRQ genotype. However, these effects are either largely dependent on sheep breed and scrapie strain or are rather subtle [70]. In addition to the susceptibility for scrapie, also the incubation period for scrapie is genetically determined [81]. Between the susceptibility of a genotype and the age at death, a clear negative relationship exists, such that sheep of the genotypes with the greatest scrapie risk (VRQ/VRQ and ARQ/ VRQ) die, on average, the youngest. However, exceptions do exist [70].

CAB International 2009

66

Animal Science Reviews 2010

Atypical scrapie is quite different from the classical form of the disease. Sheep with this unusual type of scrapie displayed predominantly ataxia without pruritus. Neuropathological ﬁndings of vacuolation and PrP proteins were restricted to cerebellar and cerebral cortices and not present so much in the brainstem at the level of the obex [74]. The three most common atypical scrapie PrP alleles are ARR, AHQ and AF141RQ. The latter is ARQ containing a phenylalanine residue at amino acid position 141 instead of a leucine [82]. BSE is another example of a TSE and is a fatal, neurodegenerative disease in cattle that causes a spongy degeneration in the brain and spinal cord. BSE has a long incubation period of about 4 years and mainly affects adult animals at a peak age onset of 4–5 years (http://fsrio. nal.usda.gov/document_fsheet.php?product_id=169). In cattle, two forms of BSE exist. Classical BSE is associated with ingestion of BSE-contaminated feed and is therefore greatly dependent on environmental factors. H- and L-type BSE, collectively known as atypical BSE, differs from classical BSE by displaying a different disease phenotype and not being linked to the consumption of contaminated feed. Atypical BSE thus is much more independent of environmental factors than classical BSE [72, 83]. As for scrapie in sheep, also in cattle, polymorphisms in the PrP gene, associated with the susceptibility for BSE were found. In cattle with classical BSE, a 23-bp indel poly0 morphism was found in the 5 -ﬂanking sequence of the bovine PrP gene and a 12-bp indel polymorphism in intron 1. These two indel polymorphisms, which contain binding sites for the RP58 and SP1 transcription factors, were found to modulate the expression of the PrP gene in vivo, and can therefore play a role in the susceptibility for BSE in cattle. For both mutations, the deletion alleles confer a higher risk of developing BSE in comparison with the insertion alleles; however, whether the major effect is associated with the 23-bp indel or 12-bp indel differs between studies [71, 84, 85]. In cattle with H-type atypical BSE, a polymorphism at codon 211 (E211K) of the bovine PrP gene was found. This mutation resulted in the substitution of a glutamic acid residue into a lysine residue and is highly analogous to a human polymorphism described as the most common cause of inherited CJD in humans (E200K). The presence of the lysine residue at this codon in humans resulted in complete penetrance of CJD, but the functional signiﬁcance of the mutation is still unknown. However, taken together, these ﬁndings suggest that the bovine PrP E211K mutation is most likely to cause BSE in animals carrying the mutation [72, 85].

Bovine and Canine Leukocyte Adhesion Deﬁciency (LAD) Bovine LAD (BLAD) in dairy cattle is a lethal autosomal recessive congenital disease characterized by recurrent bacterial infections, delayed wound healing and stunted

g

growth, and is also associated with persistent marked neutrophilia [86]. BLAD is characterized by lack of expression of adhesion molecules of the CD11/CD18 family, which play an essential role for host defence on the leukocyte surface. Therefore, the gene encoding bovine CD18 was sequenced as a candidate gene for BLAD in Holstein cattle. An A to G point mutation was found at position 383 of the CD18 gene, which caused an aspartic acid to glycine substitution at amino acid 128 in the glycoprotein. This mutation occurs near the centre of 26 consecutive amino acids that are identical in normal bovine, human and murine CD18 and lies within a large extracellular region that is highly conserved across integrin b subunits [86, 87]. Restriction analysis of PCR ampliﬁed DNA containing position 383 of the CD18 gene allows discrimination between normal, carrier and affected animals [87]. Also in dogs, and more speciﬁcally in Irish Setters, leukocyte adhesion deﬁciency was found (canine LAD (CLAD)) and the clinical manifestations are very similar to those of BLAD. Also, similarly to cattle, a mutation in the CD18 gene (G to C transversion) is responsible for this neurodegenerative disease in dogs. The G to C mutation leads to a replacement of a conserved cysteine residue to a serine residue at codon 36 of the CD18 protein. This mutation leads to the inability of CD18 to form heterodimers with the human CD11 subunit, as normal canine CD18 does [88].

CMD Types 1 and 2 CMD types 1 and 2 are important recessive genetic defects that occur in Belgian Blue cattle. Calves affected with CMD1 show impaired swallowing, fatigue upon stimulation or exercise, and muscle myotonia, resulting in an inability to ﬂex limbs and injurious falling. CMD1 calves usually die within a few weeks as a result of respiratory complications. CMD2 calves suffer severe episodes of myoclonus upon acoustic or tactile stimulation and typically die within a few hours after birth. For both these defects, functional mutations were found using the recent technique of the genome-wide SNP analysis. Using this genome-wide SNP scan on DNA samples from affected individuals and controls, a 2.12 Mb segment encompassing the CMD1 locus was found. Within this interval, a missense mutation leading to a R559C substitution in exon 14 of ATP2A1 was found. It is known that Arg559 in bovine ATP2A1 is located within the nucleotidebinding domain and is a highly conserved, functionally signiﬁcant residue that directly interacts with the b-phosphate moiety of ATP. Substitution of this residue leads to impaired ATP binding and is therefore very likely to be the causative mutation underlying CMD1. For CMD2, a 3.61 Mb segment containing the CMD2 locus was found on chromosome 29, including a strong candidate gene: SLC6A5. In exon 4 of this gene, a missense

CAB International 2009

Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens and Nadine Buys

mutation was found, resulting in a L270P substitution in the third membrane-spanning domain of GlyT2. This Leu270 is not predicted to be involved in the binding of glycine or Na+, but was shown to be conserved at this position in all vertebrates studied. This indicates that the residue has an important structural or functional role. Indeed, it was found that the mutation disrupted the presynaptic uptake of glycine by GlyT2 and was causative for CMD2 in cattle [13].

Crooked Tail Syndrome (CTS) CTS is another genetic defect in Belgian Blue cattle. Typical characteristics of animals with CTS are a crooked tail and shortened head, growth retardation, extreme muscularity and spastic paresia. The disease is not lethal but can cause substantial economical losses due to growth retardation and treatment [13]. Carrier animals have an enhanced muscular development, and this heterozygote advantage is the cause for a selective sweep, which explains the fact that the incidence of CTS has risen very suddenly in Belgian Blue cattle and that 25% of the animals now appear to be CTS carrier [89]. Using a genome-wide SNP scan on DNA samples from affected individuals and controls, a 2.42 Mb segment on bovine chromosome 19 encompassing the mutation causing CTS was found [13]. The mutation is a 2 bp deletion in the open reading frame of the Mannose Receptor C type 2 (MRC2), creating a frameshift which is the cause for a premature stop. Consequently, the 180 kDa endocytic transmembrane glycoprotein can no longer be formed in affected animals [89].

Foetal Ichytiosis Severe forms of ichytiosis fetalis (IF) have been reported in Chianina cattle. Calves affected with ichtyosis fetalis show cutaneous lesions, with deep ﬁssures separating hyperkeratotic skin plaques and eversion of mucocutaneous junctions, comparable to harlequin ichtyosis in humans. Using a genome-wide SNP scan on DNA samples from affected individuals and controls, an 11.78 Mb segment encompassing the IF locus was found. In this interval, a strong candidate gene was found: namely the ABCA12 gene. In exon 39 of this gene, a missense A5804G mutation was found, resulting in an H1935R substitution in the fourth extracellular loop. It was found that this mutation was causal for the disease [13].

CVM CVM is a recessively inherited disorder with onset during foetal development, leading to frequent abortions of foetuses or perinatal death. Animals affected with CVM

g

67

show misshapen and fused vertebrae around the cervicothoracic junction. It was found that CVM is caused by a missense mutation in the Golgi-resident nucleotidesugar transporter (UDP-N-acetylglucosamine transporter) encoded by positional candidate gene SLC35A3. It concerns a G to T transversion that replaces valine with phenylalanine at position 180. This mutation was functionally tested by yeast complementation assay in the Kluyveromyces lactis mutant mnn2–2, which is deﬁcient in transport of N-acetylglucosamine in Golgi vesicles. This assay showed a partial or complete loss of function of the mutant protein. From these results, it could be concluded that indeed the missense polymorphism in SLC35A3 was the causative mutation for CVM. It also excluded the possibility that a mutation in another gene, tightly linked to SLC35A3 is causing the defect [69].

Factor XI (FXI) A deﬁciency in blood coagulation FXI has been found in Holstein cattle. The mutation responsible for this deﬁciency is a 76 bp insertion in exon 12 of the FXI gene. This polymorphism leads to a stop codon, which results in a mature FXI protein lacking the functional protease domain encoded by exons 13, 14 and 15 [90]. The disorder seems to have an impact on reproductive traits and udder health in cattle [91].

Genomic Selection Traditional selection has been successful in improving a large number of traits for a long time. However, it requires widespread, reasonably accurate and preferably early-in-life recording of traits. Functional traits only partially fulﬁl these requirements [92]. Also, estimated breeding values should be able to predict more accurately by implementing information on DNA polymorphisms underlying traits of interest. Research towards MAS has been active and extensive but implementation of results has been limited and the increase in genetic gain is mostly rather disappointing [93]. Therefore, quite recently, in addition to research for markers for MAS, genomic selection has become an area of intense research. Genomic selection could be deﬁned as MAS on a genomewide scale. Different to the MAS previously described, where selection was for only a few causal or linked markers, genomic selection is the simultaneous selection for many (tens or hundreds of thousands of) markers, which cover the entire genome in a dense manner. This way, all genes are expected to be in LD with at least some of the markers so that potentially all the genetic variance is explained by the markers [6, 94, 95]. In 2001, when Meuwissen et al. introduced this technique, major limitations to implementation were the large number of markers required and the cost of genotyping these

CAB International 2009

68

Animal Science Reviews 2010

markers. Nowadays, however, with the development of the SNP chips for almost all domestic animals and the developments in SNP genotyping technology, these limitations have been overcome. Simulation results and limited experimental results suggest that breeding values can very accurately be predicted using genetic markers alone. However, more validation is required, especially in samples of populations that are different from the original population wherein the effect of the markers was estimated. Nevertheless, many livestock breeding companies are implementing or are planning to implement genomic selection in their breeding programmes [6].

Conclusion/Summary In conclusion, although numerous QTLs have been found for quantitative traits in different species of domestic animals, the number of causative mutations (QTMs) identiﬁed is very limited. This is due to the fact that most QTLs have a mild phenotypic effect, so the mutations that underlie them are difﬁcult to distinguish from neutral polymorphisms [3]. The QTMs, such as the one of IGF2, which have a major inﬂuence and are responsible for quite a large amount of phenotypic variance, are therefore the ones ﬁrst found. Another factor that complicates the ﬁnding of the QTL mutations is the fact that a good proportion of these are located in regulatory regions. It becomes even more complex to unravel the underlying mutations when epistasis or epigenetic inheritance is involved in the quantitative genetic variation [3]. Both candidate gene approaches and whole genome linkage analysis followed by LD mapping have contributed to the characterization of QTMs. Whole genome scans using the SNP panels recently developed for different species already lead to the identiﬁcation of mutations involved in monogenic traits. Their usefulness in detection of QTMs remains to be elucidated, although it is expected that these could also be mapped. Although the new trend in livestock breeding or genomic selection does not require the knowledge of causative mutations, QTMs are the only type of mutations that do not require validation in each new population. Hence, besides its scientiﬁc value, or its contribution to increasing the knowledge on the regulatory mechanisms underlying phenotypic traits, the identiﬁcation of QTM in domestic animals will remain important for breeding purposes.

3. Andersson L, Georges M. Domestic-animal genomics: deciphering the genetics of complex traits. Nature Reviews Genetics 2001;5:202–12. 4. Mackay TFC. Quantitative trait loci in Drosophila. Nature Reviews Genetics 2001;2:11–21. 5. Andersson L. Genome-wide association analysis in domestic animals: a powerful approach for genetic dissection of trait loci. Genetica 2009;136:341–9. 6. Goddard ME, Hayes BJ. Mapping genes for complex traits in domestic animals and their use in breeding programmes. Nature Reviews Genetics 2009;10:381–91. 7. Georges M. Mapping, ﬁne mapping and molecular dissection of quantitative trait loci in domestic animals. Annual Review of Genomics and Human Genetics 2007;8:131–62. 8. Harmegnies N, Farnir F, Davin F, Buys N, Georges M, Coppieters W. Measuring the extent of linkage disequilibrium in commercial pig populations. Animal Genetics 2006;37:225– 31. 9. Lindblad-Toh K, Wade CM, Mikkelsen TS, Karlsson EK, Jaffe DB, Kamal M, et al. Genome sequence, comparative analysis and haplotype structure of the domestic dog. Nature 2005;438:803–19. 10. De Roos APW, Hayes BJ, Spelman R, Goddard ME. Linkage disequilibrium and persistence of phase in Holstein Freisian, Jersey and Angus cattle. Genetics 2008;179:1503–12. 11. MacEachern S, Hayes BJ, McEwan J, Goddard ME. An examination of positive selection and changing effective population size in Angus and Holstein cattle populations (Bos taurus) using a high density SNP genotyping platform and the contribution of ancient polymorphism to genomic diversity in domestic cattle. BMC Genomics 2009;10:181. 12. Karlsson EK, Baranowska I, Wade CM, Salmon Hillbertz NHC, Zody MC, Anderson N, et al. Efﬁcient mapping of Mendelian traits in dogs through genome-wide association. Nature Genetics 2007;39:1321–8. 13. Charlier C, Coppieters W, Rollin F, Desmecht D, Agerholm JS, Cambisano N, et al. Highly effective SNP-based association mapping and management of recessive defects in livestock. Nature Genetics 2008;40:449–54. 14. Yu TP, Tuggle CK, Schmitz CB, Rothschild MF. Association of PIT1 polymorphisms with growth and carcass traits in pigs. Journal of Animal Science 1995;73:1282–8. 15. Ron M, Weller JI. From QTL to QTN identiﬁcation in livestock – winning by points rather than by knock-out: a review. Animal Genetics 2007;38:429–39. 16. Fujii J, Otsu K, Zorzato F, de Leon S, Khanna VK, Weiler SE, et al. Identiﬁcation of a mutation in porcine ryanodine receptor associated with malignant hyperthermia. Science 1991;253:448–51. 17. Grobet L, Martin LHR, Poncelet D, Pirottin D, Brouwers B, Riquet J, et al. A deletion in the bovine myostatin gene causes the double-muscled phenotype in cattle. Nature Genetics 1997;17:71–4.

References 1. Andersson L. Genetic dissection of phenotypic diversity in farm animals. Nature Reviews Genetics 2001;2:130–8.

18. Kambadur R, Sharma M, Smith TPL, Bass JJ. Mutations in myostatin (GDF8) in double-muscled Belgian Blue and Piedmontese cattle. Genome Research 1997;7:910–5.

2. Williams JL. The use of marker-assisted selection in animal breeding and biotechnology. Revue Scientiﬁque et Technique (International Ofﬁce of Epizootics) 2005;24:379–91.

19. McPherron AC, Lee SJ. Double muscling in cattle due to mutations in the myostatin gene. Proceedings of the National Academy of Sciences of the USA 1997;94:12457–61.

g

CAB International 2009

Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens and Nadine Buys 20. Mosher DS, Quignon P, Bustamante CD, Sutter NB, Mellersh CS, Parker HG, et al. A mutation in the myostatin gene increases muscle mass and enhances racing performance in heterozygous dogs. PLoS Genetics 2007;3:e79. 21. Georges M, Nielsen D, Mackinnon M, Mishra A, Okimoto R, Pasquino AT, et al. Mapping quantitative trait loci controlling milk production in dairy cattle by exploiting progeny testing. Genetics 1995;139:907–20. 22. Farnir F, Coppieters W, Arranz J, Berzi P, Cambisano N, Grisart B, et al. Extensive genome-wide linkage disequilibrium in cattle. Genome Research 2000;10:220–7. 23. Van Laere A, Nguyen M, Braunschweig M, Nezer C, Collette C, Moreau L, et al. A regulatory mutation in IGF2 causes a major QTL effect on muscle growth in the pig. Nature 2003;425:832–6. 24. Milan D, Jeon JT, Looft C, Amarger V, Robic A, Thelander M, et al. A mutation in PRKAG3 associated with excess glycogen content in pig skeletal muscle. Science 2000;288:1248–51. 25. Andersson L. Identiﬁcation and characterization of AMPKg 3 mutations in the pig. Biochemical Society Transactions 2003;31:232–5. 26. Szabo´ G, Dallmann G, Mu¨ller G, Patthy L, Soller M, Varga L. A deletion in the myostatin gene causes the compact (Cmpt) hypermuscular mutation in mice. Mammalian Genome 1998;9:671–2.

69

and technological meat quality in Hampshire-sired pigs. Livestock Production Science 1997;47:221–9. 35. Milan D, Woloszyn N, Yerle M, Le Roy P, Bonnet M, Riquet J, et al. Accurate mapping of the “acid meat” RN gene on genetic and physical maps of pig chromosome 15. Mammalian Genome 1996;7:47–51. 36. Lindahl G, Enfa¨lt A, Von Seth G, Josell A, Hedebro-Velander I, Andersen HJ, et al. A second mutant allele (V199I) at the PRKAG (RN) locus – I. Effect on technological meat quality of pork loin. Meat Science 2004;66:609–19. 37. Lindahl G, Enfa¨lt A, Von Seth G, Joseli A, Hedebro-Velander I, Andersen HJ, et al. A second mutant allele (V199I) at the PRKAG (RN) locus – II. Effect on colour characteristics of pork loin. Meat Science 2004;66:609–19. 38. Andersson-Eklund L, Marklund L, Lundstro¨m K, Haley CS, Andersson K, Hansson I, et al. Mapping quantitative trait loci for carcass and meat quality traits in a wild boarLarge White intercross. Journal of Animal Science 1998;76:694–700. 39. Jeon JT, Carlborg O¨, To¨rnsten A, Giuffra E, Amarger V, Chardon P, et al. A paternally expressed QTL affecting skeletal and cardiac muscle mass in pigs maps to the IGF2 locus. Nature Genetics 1999;21:157–8. 40. Nezer C, Moreau L, Brouwers B, Coppieters W, Detilleux J, Hanset R, et al. An imprinted QTL with major effect on muscle mass and fat deposition maps to the IGF2 locus in pigs. Nature Genetics 1999;21:155–6.

27. Clop A, Marcq F, Takeda H, Pirottin D, Tordoir X, Bibe´ B, et al. A mutation creating a potential illegitimate microRNA target site in the myostatin gene affects muscularity in sheep. Nature Genetics 2006;38:813–8.

41. Florini JR, Magri KA, Ewton DZ, James PL, Grindstaff K, Rotwein PS. ‘Spontaneous’ differentiation of skeletal myoblasts is dependent upon autocrine secretion of insulin-like growth factor-II. Journal of Biological Chemistry 1991;266:15917–23.

28. Grisart B, Coppieters W, Farnir F, Karim L, Ford C, Berzi P, et al. Positional candidate cloning of a QTL in dairy cattle: identiﬁcation of a missense mutation in the bovine DGAT1 gene with major effect on milk yield and composition. Genome Research 2001;12:222–31.

42. Oksbjerg N, Gondret F, Vestergaard M. Basic principles of muscle development and growth in meat-producing mammals as affected by the insulin-like growth factor (IGF) system. Domestic Animal Endocrinology 2004;27:219–40.

29. Cockett NE, Jackson SP, Shay TL, Nielsen D, Moore SS, Steele MR, et al. Chromosomal localisation of the callipyge gene in sheep (Ovis aries) using bovine DNA markers. Proceedings of the National Academy of Sciences of the USA 1994;91:3019–23. 30. Charlier C, Segers K, Karim L, Shay T, Gyapay G, Cockett N, et al. The callipyge mutation enhances the expression of coregulated imprinted genes in cis without affecting their imprinting status. Nature Genetics 2001;27:367–9. 31. Wilson T, Wu X, Juengel JL, Ross IK, Lumsden JM, Lord EA, et al. Highly proliﬁc Booroola sheep have a mutation in the intracellular kinase domain of bone morphogenetic protein IB receptor (ALK-6) that is expressed in both oocytes and granulose cells. Biology of Reproduction 2001;64:1225–35. 32. Meuwissen THE, Goddard ME. Mapping multiple QTL using linkage disequilibrium and linkage analysis information and multitrait data. Genetics Selection Evolution 2004;36:261–79.

43. Oczkowicz M, Tyra M, Walinowicz K, Ro´zycki M, Rejduch B. Known mutation (A3072G) in intron 3 of the IGF2 gene is associated with growth and carcass composition in Polish pig breeds. Journal of Applied Genetics 2009;50:257–9. 44. Braunschweig MH, Van Laere AS, Buys N, Andersson L, Andersson G. IGF2 antisense transcript expression in porcine postnatal muscle is affected by a quantitative trait nucleotide in intron 3. Genomics 2004;84:1021–9. 45. Jackson SP, Miller MF and Green RD. The effect of a muscle hypertrophy gene on muscle weights of ram lambs. Proceedings, Western Section, American Society of Animal Science 1993;44:196–8. 46. Jackson SP, Miller MF, Green RD, Brdecko KS. Carcass characteristics of Rambouillet ram lambs with genetic muscle hypertrophy. Proceedings, Western Section, American Society of Animal Science 1993;44:167–9.

33. Lister D. The physiology and biochemistry of the porcine stress syndrome. In: Tarrant PV, Eikelboom G, Monin G, editors. Evaluation and Control of Meat Quality in Pigs. Martinus Nijfhoff Publishers, Boston, MA; 1987. p. 3–16.

47. Freking BA, Murphy SK, Wylie AA, Rhodes SJ, Keele JW, Leymaster KA, et al. Identiﬁcation of the single base change causing the callipyge muscle hyperthrophy phenotype, the only known example of polar overdominance in mammals. Genome Research 2002;12:1496–506.

34. Enfalt A, Lundstro¨m K, Hansson I, Johansen S, Nystro¨m P. Comparison of non-carriers and heterozygous carriers of the RN7 allele for carcass composition, muscle distribution

48. Georges M, Charlier C, Cockett N. The callipyge locus: evidence for the trans interaction of reciprocally imprinted genes. Trends in Genetics 2003;19:248–52.

g

CAB International 2009

70

Animal Science Reviews 2010

49. Grobet L, Poncelet D, Royo LJ, Brouwers B, Pirottin, Michaux C, et al. Molecular deﬁnition of an allelic series of mutations disrupting the myostatin function and causing double-muscling in cattle. Mammalian Genome 1998;9:210–3.

62. Souza CJH, MacDougall C, Campbell BK, McNeilly AS, Baird DT. The Booroola (FecB) phenotype is associated with a mutation in the bone morphogenetic receptor type 1 B (BMPR1B) gene. Journal of Endocrinology 2001;169(2):R1–6.

50. Karim L, Coppieters W, Grobet L, Valentini A, Georges M. Convenient genotyping of six myostatin mutations causing double-muscling in cattle using a multiplex oligonucleotide ligation assay. Animal genetics 2000;31:396–9.

63. Mulsant P, Lecerf F, Fabre S, Schibler L, Monget P, Lanneluc I, et al. Mutation in bone morphogenetic protein receptor-IB is associated with increased ovulation rate in Booroola Me´rino ewes. Proceedings of the National Academy of Sciences of the USA 2001;98(9):5104–9.

51. Marchitelli C, Savarese MC, Crisa` A, Nardone A, Marsan PA, Valentini A. Double muscling in Marchigiana beef breed is caused by a stop codon in the third exon of myostatin gene. Mammalian Genome 2003;14:392–5.

64. Fabre S, Pierre A, Mulsant P, Bodin L, Di Pasquale E, Persani L, et al. Regulation of ovulation rate in mammals: contribution of sheep genetic models. Reproductive Biology and Endocrinology 2006;4:20.

52. Boman IA, Klemetsdal G, Blichfeldt T, Nafstad O, Va˚ge DI. A frameshift mutation in the coding region of the myostatin gene (MSTN) affects carcass conformation and fatness in Norwegian White Sheep (Ovis aries). Animal genetics 2009;40:418–22.

65. Galloway SM, McNatty KP, Cambridge LM, Laitinen MPE, Juengel JL, Jokiranta S, et al. Mutations in an oocyte-derived growth factor gene (BMP15) cause increased ovulation rate and infertility in a dosage-sensitive manner. Nature Genetics 2000;25(3):279–83.

53. Schuelke M, Wagner KR, Stolz LE, Hu¨bner C, Riedel T, Ko¨men W, et al. Myostatin mutation associated with gross muscle hyperthrophy in a child. New England Journal of Medicine 2004;350:2682–8.

66. Hanrahan JP, Gregan SM, Mulsant P, Mullen M, Davis GH, Powell R, et al. Mutations in the genes for oocyte-derived growth factors GDF9 and BMP15 are associated with both increased ovulation rate and sterility in Cambridge and Belclare sheep. Biology of Reproduction 2004;70(4):900–9.

54. Coppieters W, Riquet J, Arranz JJ, Berzi P, Cambisano N, Grisart B, et al. A QTL with major effect on milk yield and composition maps to bovine chromosome 14. Mammalian Genome 1998;9:540–4. 55. Winter A, Kra¨mer W, Werner FAO, Kollers S, Kata S, Durstewitz G, et al. Association of a lysine-232/alanine polymorphism in a bovine gene encoding acylCoA:diacylglycerol acyltransferase (DGAT1) with variation at a quantitativ trait locus for milk fat content. Proceedings of the National of Academy of Sciences of the USA 2002;99:9300–5. 56. Grisart B, Farnir F, Karim L, Cambisano N, Kim JJ, Kvasz A, et al. Genetic and functional conﬁrmation of the causality of the DGAT1 K232A quantitative trait nucleotide in affecting milk yield and composition. Proceedings of the National Academy of Sciences of the USA 2004;101:2398–403.

67. Davis GH. Major genes affecting ovulation rate in sheep. Genetic Selection Evolution 2005;37(Suppl. 1):S11–23. 68. McNatty KP, Galloway SM, Wilson T, Smith P, Hudson NL, O’Connell A, et al. Physiological effects of major genes affecting ovulation rate in sheep. Genetics Selection Evolution 2005;37(Suppl. 1):S25–38. 69. Thomsen B, Horn P, Panitz F, Bendixen E, Petersen AH, Holm LE, et al. A missense mutation in the bovine SLC35A3 gene, encoding a UDP-N-acetylglucosamine transporter, causes complex vertebral malformation. Genome Research 2006;16:97–105. 70. Baylis M, Goldmann W. The genetics of scrapie in sheep and goats. Current Molecular Medicine 2004;4:385–96.

57. Ku¨hn C, Thaller G, Winter A, Bininda-Emonds ORP, Kaupe B, Erhardt G, et al. Evidence for multiple alleles at the DGAT1 locus better explains a quantitative trait locus with major effect on milk fat content in cattle. Genetics 2004;167:1873–81.

71. Sander P, Hamann H, Dro¨gemu¨ller C, Kashkevich K, Schiebel K, Leeb T. Bovine prion protein gene (PRNP) promotor polymorphisms modulate PRNP expression and may be responsible for differences in bovine spongiform encephalopathy susceptibility. Journal of Biological Chemistry 2005;280:37408–14.

58. Cohen-Zinder M, Seroussi E, Larkin DM, Loor JL, Everts-Van der Wind A, Lee J, et al. Identiﬁcation of a missense mutation in the bovine ABCG2 gene with a major effect on the QTL on chromosome 6 affecting milk yield and composition in Holstein cattle. Genome Research 2005;15:936–44.

72. Nicholson EM, Brunelle BW, Richt JA, Kehrli ME, Greenlee JJ. Identiﬁcation of a heritable polymorphism in bovine PRNP associated with genetic transmissible spongiform encephalopathy: evidence of heritable BSE. PLoS One 2008;3:e2912.

59. Jonker JW, Merino G, Musters S, Van Herwaarden AE, Bolscher E, Wagenaar E, et al. The breast cancer resistance protein BCRP (ABCG2) concentrates drugs and carcinogenic xenotoxins into milk. Nature Medicine 2004;11:127–9. 60. Olsen HG, Nilsen H, Hayes B, Berg RR, Svendsen M, Lien S, et al. Genetic support for a quantitative trait nucleotide in the ABCG2 gene affecting milk composition of dairy cattle. BMC Genetics 2007;8:32. 61. Piper LR, Bindon BM. The Booroola Merino and the performance of medium non-peppin crosses at Armidale. In Piper LR, Bindon BM, Nethery RD, editors. The Booroola Merino, Proceedings of a Workshop, Armidale, 24–25 August 1980, CSIRO; 1982, pp. 9–19.

g

73. Prusiner SB. Research on scrapie. Lancet 1982;2:494–5. 74. Konold T, Davis A, Bone G, Bracegirdle J, Everitt S, Chaplin M, et al. Clinical ﬁndings in two cases of atypical scrapie in sheep: a case report. BMC Veterinary Research 2007;3:2. 75. Goldmann W, Hunter N, Foster JD, Salbaum JM, Beyreuther K, Hope J. Two alleles of a neural protein gene linked to scrapie in sheep. Proceedings of the National Academy of Sciences of the USA 1990;87:2476–80. 76. Goldmann W, Hunter N, Benson G, Foster JD, Hope J. Different scrapie-associated ﬁbril proteins (PrP) are encoded by lines of sheep selected for different alleles of the Sip gene. Journal of General Virology 1991;72:2411–7.

CAB International 2009

Anneleen Stinckens, Martine Schroyen, Liesbet Peeters, Steven Janssens and Nadine Buys 77. Belt PBGM, Muileman IH, Schreuder BEC, Bos-De Ruijter JB, Gielkens ALJ, Smits MA. Identiﬁcation of ﬁve allelic variants of the sheep PrP gene and their association with natural scrapie. Journal of General Virology 1995;76:509–17. 78. Ikeda T, Horiuchi M, Ishiguro N, Muramatsu Y, Kai-Uwe GD, Shinagawa M. Amino acid polymorphisms of PrP with reference to onset of scrapie in Suffolk and Corriedale sheep in Japan. Journal of General Virology 1995;76:2577–81. 79. Kutzer T, Pfeiffer I, Brenig B. Identiﬁcation of new allelic variants in the ovine prion protein (PrP) gene. Journal of Animal Breeding and Genetics 2002;119:201–8. 80. Tkacikova L, Hanusovska E, Novak M, Arvayova M, Mikula I. The PrP genotype of sheep of the improved Valachian breed. Folia Microbiologica 2003;48:269–76. 81. Houston EF, Halliday SI, Jeffrey M, Goldmann W, Hunter N. New Zealand sheep with scrapie-susceptible PrP genotypes succumb to experimental challenge with a sheep-passaged scrapie isolate (SSBP/1). Journal of General Virology 2002;83:1247–50.

71

87. Shuster DE, Kehrli ME, Ackermann MR, Gilbert RO. Identiﬁcation and prevalence of a genetic defect that causes leukocyte adhesion deﬁciency in Holstein cattle. Proceedings of the National Academy of Sciences of the USA 1992;89:9225–9. 88. Kijas JMH, Bauer TR, Ga¨fvert S, Marklund S, Trowald-Wigh G, Johannisson A, et al. A missense mutation in the b-2 integrin gene (ITGB2) causes canine leukocyte adhesion deﬁciency. Genomics 1999;61:101–7. 89. Fasquelle C, Sartelet A, Li W, Dive M, Tamma N, et al. Balancing selection of a frame-shift mutation in the MRC2 gene accounts for the outbreak of the crooked tail syndrome in Belgian Blue Cattle. PLos Genet 2009;5(9):e1000666. doi:10.1371/journal.pgen.1000666. 90. Marron BM, Robinson JL, Gentry PA, Beever E. Identiﬁcation of a mutation associated with factor XI deﬁciency in Holstein cattle. Animal Genetics 2004;35:454–6.

82. Juling K, Schwarzenbacher H, Williams JL, Fries R. A major genetic component of BSE susceptibility. BMC Biology 2006;4:33.

91. Gurgul A, Rubis D, Slota E. Identiﬁcation of carriers of the mutation causing coagulation factor XI deﬁciency in Polish Holstein–Friesian cattle. Journal of Applied Genetics 2009;50:149–52.

83. Saunders GC, Cawthraw S, Mountjov SJ, Hope J, Windl Q. PrP genotypes of atypical scrapie cases in Great Britain. Journal of General Virology 2006;87:3141–9.

92. Meuwissen T. Genomic selection: marker assisted selection on a genome wide scale. Journal of Animal Breeding and Genetics 2007;124:321–2.

84. Haase B, Doherr MG, Seuberlich T, Dro¨gemuller C, Dolf G, Nicken P, et al. PRNP promoter polymorphisms are associated with BSE susceptibility in Swiss and German cattle. BMC Genetics 2007;8:15.

93. Dekkers JC. Commercial application of marker- and gene-assisted selection in livestock: strategies and lessons. Journal of Animal Science 2004;82:E313–28.

85. Richt JA, Hall SM. BSE case associated with prion protein gene mutation. PloS Pathogens 2008;4:e1000156.

94. Meuwissen THE, Hayes BJ, Goddard ME. Prediction of total genetic value using genome-wide dense marker maps. Genetics 2001;157:1819–29.

86. Nagahata H. Bovine leukocyte adhesion deﬁciency (BLAD): a review. Journal of Veterinary Medical Science 2004;66:1475–82.

g

95. Goddard ME, Hayes BJ. Genomic selection. Journal of Animal Breeding and Genetics 2007;124:323–30.

CAB International 2009

This page intentionally left blank

Animal Science Reviews 2010

Technical, epidemiological and ﬁnancial implications of large-scale national vaccination campaigns to control HPAI H5N1 J. Hinrichs1*, J. Otte1 and J. Rushton2 Address: 1 FAO, Viale delle Terme di Caracalla, 00153 Rome, Italy. North Mymms, Hatﬁeld, Hertfordshire, AL9 7TA, UK.

2

Royal Veterinary College (RVC), Hawkshead Lane,

*Correspondence: J. Hinrichs. Email: [email protected] 30 September 2009 15 January 2010

Received: Accepted:

Abstract This paper reviews the literature on highly pathogenic avian inﬂuenza (HPAI) vaccination of poultry and the resulting technical, epidemiological and ﬁnancial implications for the control of HPAI H5N1 through national vaccination campaigns. Large-scale HPAI vaccination programmes have been implemented in a number of Southeast Asian countries and in Egypt, requiring substantial ﬁnancial and human resources. HPAI remains endemic in these countries and it thus appears warranted to comprehensively assess all major aspects of large-scale vaccination campaigns for the control of HPAI. The paper reviews existing data and literature on HPAI vaccines, HPAI epidemics and large-scale HPAI H5N1 vaccination programmes and draws on calculations carried out by the authors to assess and evaluate the potential contribution and the ﬁnancial implications of mass vaccination in the control of HPAI H5N1 in domestic poultry in developing countries. We conclude that the high and recurrent costs, technical difﬁculties and epidemiological drawbacks of large-scale, open-ended, blanket vaccination programmes in national efforts to control HPAI call for careful targeting of vaccination in national control strategies, which ‘intelligently’ combine available disease control measures. Keywords: Vaccination, Epidemiology, Economics, Highly pathogenic avian inﬂuenza (HPAI), Poultry production systems, Chicken, Ducks

Introduction Since its emergence in 1996 in China, highly pathogenic avian inﬂuenza (HPAI) H5N1 virus has infected 61 countries, been associated with more than 260 human fatalities, and resulted in disease mortality and culling of several hundred million domestic birds. In most of the affected countries, the H5N1 virus could be eliminated through swift and determined interventions of national animal health systems. In some countries, however, the virus appears to have become endemic in speciﬁc eco- and production systems, leading to resurgence of infection in poultry and humans, the moment control efforts are relaxed. The major countries in which HPAI H5N1 virus can currently be considered endemic comprise China, Egypt, Indonesia and Vietnam, all of which have

g

included vaccination as a part of their national control strategy. Under laboratory conditions, vaccination trials against HPAI have shown that a variety of commercially available vaccines protect against clinical signs and reduce virus shedding in the case of contact with ﬁeld virus but do not prevent infection in all vaccinated birds [1, 2]. However, prior to the massive epidemics of H5N1 in Southeast Asia, only very few attempts to control HPAI outbreaks in domestic poultry populations by vaccination have been reported. Pre-H5N1 experience in the use of vaccination in large-scale HPAI control programmes was gained in Mexico (H5N1, 1994), Italy (H7N1, 2000) and Pakistan (H7N3, 2003) [3]. In the course of the current H5N1 avian pandemic, several large-scale HPAI vaccination campaigns have been implemented by national animal

CAB International 2010

74

Animal Science Reviews 2010

health authorities in China, Hong Kong, Vietnam, Indonesia and Egypt. Experience has shown that despite the theoretical potential of vaccination to control HPAI epidemics [3, 4], this potential is in practice not fully realized in largescale vaccination efforts due to the numerous constraints to delivering and administering vaccine in large and heterogeneous poultry populations [3]. Large-scale vaccination campaigns require substantial ﬁnancial and human resources and therefore are unlikely to be sustainable over long time periods. It thus appears warranted to assess the technical, epidemiological as well as the ﬁnancial implications of large-scale vaccination campaigns for HPAI control, both from a theoretical as well as from a practical perspective. This paper reviews existing data and literature on HPAI vaccines, HPAI epidemics and large-scale HPAI H5N1 vaccination programmes and draws on calculations carried out to assess and evaluate the potential contribution and the ﬁnancial implications of mass vaccination in the control of HPAI H5N1 in domestic poultry in developing countries. The paper starts by reviewing the literature on various characteristics of commercially available HPAI vaccines that determine their utility as a tool for the control of HPAI in different poultry species. The following section presents the three HPAI vaccination strategies proposed by the Ofﬁce International des Epizooties (OIE) and provides an overview of estimates of immunization rates that would be necessary to suppress H5N1 virus transmission to a level where infection dies out. Estimates of the maximum vaccination coverage that would be achievable in different poultry production systems through mass vaccination campaigns carried out under ideal conditions are presented in ‘Practical Challenges of Large-Scale HPAI Vaccination Programmes’. The following section summarizes common challenges for the implementation of large-scale vaccination programmes and experiences with the implementation of such programmes are compiled in ‘Experiences from Large-Scale HPAI H5N1 Vaccination Programmes’. ‘Costs and Incentives for HPAI Vaccination’ provides assessments of the costs and returns of vaccination from a production systems perspective as well as considering the broader public aspects of embarking on vaccination as a part of a national HPAI control programme. Finally, the last section discusses the advantages and drawbacks of large-scale vaccination programmes and their potential contribution to HPAI control.

Characteristics of Commercial HPAI Vaccines Route and Schedule of Vaccine Administration All commercially available HPAI vaccines require administration to poultry via injection with a syringe. Vaccines for administration via other routes, such as aerosols, drinking water or injection during egg incubation are not

g

currently available [5]. Generally, a two-dose vaccination schedule is required to achieve satisfactory protection against HPAI in any poultry species. Application of the booster vaccination is often inconvenient and may not be ensured in all production systems for a number of reasons. Also, in many commercial production systems it is most efﬁcient to vaccinate birds at one day of age. Trovac, a recombinant live H5N1 vaccine (A/Turkey/Ireland/83 recombinant fowlpox vector) is the only commercially available vaccine purposely developed for a one-shot at day-old age vaccination schedule.

Vaccine Efﬁcacy From a technical and biological perspective, for individual birds, vaccination efﬁcacy can be measured using the following three parameters: (i) degree of protection from infection when exposed to a given amount of infectious virus, (ii) the degree of reduction of morbidity and mortality given infection occurred and (iii) the level of reduction of virus excretion by infected poultry. The efﬁcacy of most commercially available vaccines has been determined in studies with chickens or turkeys, since, from an economic perspective, globally they represent the most important poultry species. However, in many developing countries, particularly in Southeast Asia, other species such as ducks, muscovy ducks and quails also represent signiﬁcant parts of the poultry sector. Thus, vaccination strategies for poultry populations with signiﬁcant shares of species other than chicken and turkeys require vaccines with proven efﬁcacy in these species to contain viral transmission. Peyre et al. [2] have compiled a list of commercially available H5 and H7 vaccines with information on reductions in mortality and viral shedding based on experimental results from challenge trials under laboratory and controlled ﬁeld conditions. Although almost all commercially available vaccines provide some level of protection from infection with virulent virus and signiﬁcantly reduce morbidity and mortality in infected chicken, no HPAI vaccine has so far proved to satisfactorily perform on all three of the above parameters [3, 6]. Challenge trial results on the efﬁcacy of Trovac are conﬂicting. Inui [7] conducted challenge trials with Trovac and observed 90–100% mortality 3, 4 and 6 weeks postvaccination in chickens when challenged with several clades of the Asian lineage HPAI H5N1 virus (clades 1(08), 2.3.4 (08), 1(04)). In contrast, Swayne et al. [8], in a challenge trial with H5N2 virus isolated in Mexico, observed 90–100% protection against morbidity and mortality, and signiﬁcant reductions in virus shedding in chickens vaccinated at one day of age. Bublot et al. [9] challenged Trovac-vaccinated chicken with HPAI H5N1 A/chicken/Vietnam/0008/2004 and observed full protection against morbidity and mortality. This diversity of challenge trial results with different HPAI H5N1 ﬁeld

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton

strains for Trovac indicates a limited breadth of protection of Trovac vaccine, and most likely of other HPAI vaccines as well, against ﬁeld strains that change over time. Unpublished results of challenge trials carried out by the Friedrich–Loefﬂer Institute in 2008 showed that vaccinated layer chicken reared under commercial conditions did not attain sterile immunity and that infection with HPAI virus may facilitate bacterial infections which can then dominate the clinical disease picture [10]. In a vaccination trial with goose parents and a low pathogenic avian inﬂuenza virus strain vaccine (A/duck/ Potsdam/1402/86 (H5N2)), Rudolf et al. [11] found that vaccinated geese became infected and transmitted challenge virus (A/Cygnuscygnus/Germany/R65/06 (H5N1)) to non-vaccinated geese. The contribution of ducks to the maintenance and spread of HPAI H5N1 virus has been shown in several studies [12–14]. Several vaccines with proven efﬁcacy in domestic ducks and geese are commercially available. A H5N3 reversegenetics vaccine has been shown by Webster et al. [15] to control clinical signs and virus shedding in Peking ducks challenged with a duck-lethal H5N1 virus after the application of two doses of vaccine. A H5N1 inactivated vaccine was used by Beato et al. [16] to vaccinate Peking ducks at 1 and 30 days of age. The vaccinated ducks were subsequently challenged with an Asian-lineage H5N1 virus and neither clinical signs nor virus shedding were detected. After one shot of a recombinant H3N3 or H5N1 vaccine administered to 2-week-old SPF Peking ducks and a challenge with duck lethal HPAI Dk/Laos/25/06 (clade 2.3.4), no clinical signs or virus shedding were detectable [17]. A speciﬁc challenge in developing vaccines for ducks and monitoring the effectiveness of duck vaccination campaigns is the lack of correlation between the serum antibody levels of vaccinated ducks and their degree of immunity. Kim et al. [17] found complete protection in vaccinated ducks in the absence of detectable antibody responses. This may indicate that cell-mediated immune response is important in protecting ducks from HPAI and would constitute a major gap in the current capacity of assessing HPAI vaccine efﬁcacy in this species, which constitutes a major part of the poultry population in a number of developing countries.

Onset and Duration of Immunity The time lag between vaccination and protective immunity and the respective duration of protection depends on the vaccine used, timing of vaccination, number of doses given, species, immunologic condition of the birds and the challenge virus. The literature is dominated by studies with SPF birds reared under laboratory conditions, the results of which cannot safely be extrapolated to ﬁeld conditions, and in which protection is often assessed by serology, the assumption being that birds with a speciﬁed

g

75

antibody level (e.g. a haemagglutination inhibition (HI) titre P16) are protected. However, under conditions of antigenic variability and diversity of the HPAI viruses circulating in the ﬁeld, a given titre found under laboratory conditions cannot unequivocally be interpreted as protective against ﬁeld virus challenge, and the observed variation in the level and length of protection also makes generalizations on vaccine efﬁcacy problematic. Despite these limitations, some examples of published literature on vaccine efﬁcacy and the duration of protection under ﬁeld and laboratory conditions are summarized in the following. Field studies with a killed oil-adjuvanted H5N2 (A/chicken/Mexico/232/94/CPA) vaccine in Hong Kong showed effective protection by 13–18 days postvaccination [18, 19]. Serologic response monitoring after vaccination of 3 ﬂocks of white leghorn layer chicken in commercial farms in California, USA with two doses of killed LPAI H6N2 (CK/CA/0379/02) vaccine at 7.5–9 and 11–13 weeks of age showed signiﬁcant differences in the onset and duration of immunity. In one ﬂock, 72% (15/21) of the tested chicken were sero-positive, based on agar gel immunodiffusion tests, within 21 days after the ﬁrst vaccination and 100% were sero-positive 4 days after the second vaccination, but by 20.5 weeks after the second vaccination, all chicken were sero-negative again. In two other ﬂocks, situated on a different farm, serologic response monitoring only began 13 weeks after the ﬁrst vaccination. Nearly 100% of the tested 32 chicken were sero-positive from 13–19 weeks post vaccination. Sixtynine weeks after the second vaccination, the proportion of seropositive chicken had declined to 22.9% [20]. Under laboratory conditions, vaccination experiments with 6-week-old white leghorn SPF chicken, using a single vaccination with H7N1 (A/Chicken/Italy/99) and H7N3 (A/Chicken/Pakistan/95) vaccines resulted in HI titres P16 11 days post vaccination. In challenge experiments with a H7N7 virus (A/Chicken/Netherlands/621557/03) 2 weeks post-vaccination, virus could not be recovered from tracheal and cloacal swabs and chicken did not infect unvaccinated chicken that were in close contact [21]. Qiao et al. [22] vaccinated 4-week-old white leghorn SPF chicken once with a vaccine derived from A/Harbin/ Re-1/2003 (Re-1) and conducted challenge trials with A/goose/Guangdong/1/96 (GSGD/96). Two weeks postvaccination and 3 days post challenge, shedding of 0.9 log10 EID50 was detected in one chicken out of a group of eight. Five days after challenge, no virus shedding was detectable in groups of chicken that had been vaccinated 2 or 43 weeks before the challenge trial. An ex-ante assessment of control strategies in the case of H5N1 outbreaks in the UK, based on a mathematical simulation model, assumes a period of 21 days for vaccination to become effective [23]. During the HPAI outbreaks in the Netherlands in 2003, van Boven et al. [24] based their advice on the use of vaccination on the assumption that it takes 2–4 weeks until vaccination offers

CAB International 2010

76

Animal Science Reviews 2010

Table 1 Vaccination strategies, objectives, time frame, critical success factors and alternative/complementary control measures Vaccination strategy

Objective

Time frame

Critical factor

Alternatives

Preventive

Protect individual/speciﬁc ﬂocks/birds

Emergency

Curtail potential of an acute epidemic after virus introduction Reduction of mortality/ production losses in endemic situations; in longer term, may facilitate eradication of HPAI virus presence in domestic poultry Reduction of human health risk

Variable, depending on risk of exposure to infectious virus Short-term

Accuracy of the exposure risk assessment Time to achieve immunity

Medium- to long-term

Effective immunization coverage (reduction of between-ﬂock Rn < 1)

Improve biosecurity, limit contact to secure sources Movement control and pre-emptive depopulation Passive and active surveillance with rapid stamping out

Routine

Reduction of viral shedding by infected birds

effective protection and therefore concluded that ring vaccination in a radius < 50 km would have little effect in reducing the size of an outbreak. No conclusions were provided for vaccination in a radius >50 km.

Vaccine Cost and Storage Requirements Legok and Harbin Werke vaccines purchased in large quantities for use in vaccination programmes in Indonesia and Vietnam cost about US$0.02–0.03 per dose. Swayne and Kapczinski [5] provide a wholesale price for HPAI vaccines of US$0.05–0.15 per dose. Avian inﬂuenza vaccines have to be stored within a temperature range of 2–8 C [25] and cold storage and a cold chain is required to maintain the efﬁcacy of all commercially available vaccines [26].

Antigenic Drift and Long-term Vaccine Efﬁcacy Avian inﬂuenza viruses vary antigenically and evolve rapidly, which poses a major challenge for the use of vaccines as an effective and sustainable HPAI control measure. H5N1 virus isolates from human cases in Vietnam show evidence of antigenic drift [27]. Although several studies demonstrated cross-protection for HPAI viruses with regard to morbidity and mortality, a correlation between virus shedding and antigenic differences of vaccine and ﬁeld strains was shown by Lee et al. [28] for the Mexican lineage H5N2 virus and by Swayne and Suarez [29] with nine different H5 HPAI viruses representing 87.3–100% deduced amino acid identity in the HA1 between the vaccine and challenge virus. During and after the extensive use of about 2 billion doses of H5N2 vaccine in commercial poultry farms in Mexico, molecular

g

Behaviour change reducing human exposure

drifts with a yearly trend have been shown [28, 30]. Antigenic drift of avian inﬂuenza viruses was observed in the USA after vaccination programmes for LPAI in commercial poultry [31]. H5N1 HPAI outbreaks in Hong Kong in December 2008 were speculated to be the result of the vaccine used being ineffective as a result of antigenic shift of circulating ﬁeld virus. Variant ﬁeld strains that escaped the protection by the used vaccines emerged in Shanxi China during 2006, in Egypt in late 2006 and in Indonesia early 2007 [5].

HPAI Vaccination ‘Strategies’ and Effective Immunization Rates OIE [32] lists three HPAI vaccination ‘strategies’ with distinct objectives: (i) preventive vaccination, (ii) emergency vaccination and (iii) routine vaccination. A summary of the objective, time frame and critical success factors of these HPAI vaccination strategies is given in Table 1.

Preventive Vaccination Preventive vaccination is proposed as an option to prevent the infection of poultry ﬂocks in a country or region that is free of disease but at ‘high’ risk of virus introduction and in which early detection and elimination of infection may not be feasible or realistic. Incorporation of DIVA1 is recommended as part of such a strategy. For example, in the Netherlands, vaccination of free-range

1

DIVA: differentiating infected from vaccinated animals, i.e. vaccinated birds can be distinguished from (vaccinated and subsequently) infected birds.

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton

laying hens and hobby poultry with inactivated H5N9 vaccine was permitted as an alternative risk reduction measure to indoor housing in 2006. In Hong Kong, a killed oil-adjuvanted H5N2 vaccine is used in broiler chicken farms since the HPAI outbreak in 2002 to reduce the likelihood of outbreaks, if introduction of infection were to occur from mainland China [33].

Emergency Vaccination This vaccination strategy is considered an option for the control of HPAI introduced into the national ﬂock when the epidemiological situation suggests an immediate and high risk of massive and rapid spread of infection, which cannot be contained by culling and movement restrictions. Emergency vaccination includes ‘ring’ vaccination of ﬂocks located within a pre-deﬁned (but not further speciﬁed) radius around detected outbreaks to create a ‘buffer zone’. This strategy was applied in northern Pakistan within a 3 km ring after H7N3 outbreaks in 2003 when layer and breeder ﬂocks were vaccinated [33].

Routine Vaccination Routine vaccination is listed as an appropriate measure in ‘countries and regions where the disease is endemic and where the classical control cannot be effectively implemented to eliminate the virus’. It can achieve a reduction in poultry mortality and in the longer term decrease the prevalence of infection to a level where surveillance and stamping out could be applied cost-effectively. Eradication of HPAI virus is not stated in the OIE [32] document as an objective that is achievable solely through routine vaccination.2 In addition the contribution of vaccination to reducing the risk of human cases via reducing the virus load is mentioned in OIE [32] as a potential result of any vaccination strategy. All three of the above ‘strategies’ can be applied in different ‘tactics’, i.e. either in a ‘mass or a targeted manner’, targeting speciﬁc sub-populations (age, species, location and production systems) and/or times within production cycles.

Effective Immunization Rates Whichever the applied vaccination strategy, its effectiveness depends on the proportion of poultry which are rendered immune and will therefore not signiﬁcantly

2 ‘Routine vaccination’ was successfully used for the eradication of other trans-boundary diseases such as Rinderpest in Africa [34] and FMD in parts of South-America [35].

g

77

contribute to virus transmission in the case of exposure. As poultry populations are segregated into ﬂocks (or other types of management units), both within-ﬂock and between-ﬂock transmission need to be sufﬁciently contained to avoid sustained virus transmission and potential development of endemicity. A number of studies estimated the minimum withinﬂock immunization rate required to stop virus transmission within a poultry ﬂock. Depending on the assumptions made, the calculated within-ﬂock immunization rate required to avoid disease spread range from 50 to 90%: Tiensin et al. [36] analysed within-ﬂock HPAI transmission data from the 2004 epidemic in Thailand and conclude that 80% of birds need to be immunized to avoid major within-ﬂock disease spread. Bouma et al. [37] estimated that a within-ﬂock immunization rate of 60–80% was necessary to avoid major outbreaks, which they deﬁned as more than 50 infected birds. Mathematical modelling conducted by Savill et al. [38] indicated a required 90% within-ﬂock coverage to reduce the outbreak probability by 50%. According to Lesnoff et al. [39] a ﬂock immunity rate of 50–67% is necessary to completely interrupt within-backyard-ﬂock virus transmission. A factor of high importance for disease control is the contact rate between ﬂocks and the level of risk of each contact to transmit infection [40]. The objective of emergency and routine vaccination would be to interrupt the infection chain between ﬂocks/farms, i.e. achieve a between-ﬂock/farm-to-farm ‘reproductive number’ (Rn) that is below unity. The required proportion of ﬂocks that would have to be immunized within an affected region can be derived from the infection dynamics in that region in the absence of control measures. For HPAI outbreaks in the Netherlands, Canada and Italy, Garske et al. [41] estimated mean farm-to-farm reproductive numbers prior to the introduction of control measures to range from 1.1 to 2.4. For Vietnam and Romania, countries with less industrialized poultry sectors, Rn was estimated to have been in the order of 2–3 prior to the introduction of control measures [42, 43]. Under the ideal condition of full protection of vaccinated ﬂocks, immunization of between half and two-thirds of ﬂocks would be necessary to stop sustained between-ﬂock transmission under the situation prevailing in Vietnam (based on the fraction of 1–1/R0 [44]). Under conditions where only partial immunity of vaccinated ﬂocks is achieved, a higher proportion of immunized ﬂocks would be required to interrupt disease transmission to the extent needed to control an outbreak.

Maximum Vaccination Coverage Achievable with Vaccination Campaigns As pointed out by Alders et al. [45] production system characteristics need to be considered to estimate the potential vaccination coverage and whether this coverage

CAB International 2010

78

Animal Science Reviews 2010

is capable of preventing, or at least signiﬁcantly reducing, virus transmission. Important characteristics are the average lifespan of poultry, origin of replacement birds (home-bred versus bought), synchronization of ﬂock age and the disease status and immuno-competency of ﬂocks in the respective production system. Furthermore, poultry owners’ incentives to vaccinate against HPAI strongly inﬂuence potential vaccination coverage. Although incentives are generally similar for owners of ﬂocks of the same type of production system, they may still vary depending on their ﬂocks point in the production cycle. It would, for example, be very unlikely that a broiler farmer would allow vaccination of a ﬂock close to the point of sale because of fear that the vaccination will affect the birds and potentially kill a proportion.

Poultry Production Systems Within the poultry sector, four main production systems can be identiﬁed: (i) breeder ﬂocks, (ii) layer ﬂocks (iii), broiler ﬂocks and (iv) backyard multi-purpose ﬂocks. This production system classiﬁcation roughly applies to both chickens and ducks [46]. Sub-systems, such as short- and long-lived broiler production and free-range grazing duck production systems prevail in some countries. The major poultry production systems are brieﬂy described in the following to provide an overview of the main characteristics relevant for the effectiveness of HPAI vaccination campaigns. Breeder ﬂocks (grandparent and parent) are kept in closed houses and cages and the birds are bought from specialized poultry genetic supply companies. Several batches of birds of differing ages are required to meet the continuous demand for day-old chicks (DOCs). The average lifespan of breeder chicken for the production of DOCs varies between 63 and 65 weeks. The production period of breeder duck ﬂocks varies between 52 and 104 weeks. Breeder ducks are kept in houses with outdoor access [47]. Layer chickens are typically kept in cages in open or closed housing without outdoor access. Hens start laying eggs at 23–25 weeks of age and are kept for 63–74 weeks. Continuous egg supply requires a layer ﬂock with chickens of several age groups. Either DOCs or pullets (at 16–25 weeks of age) are bought to replace spent hens, which are either sold for immediate slaughter or for fattening. In Asia, layer ducks are kept between 1 and 3 years and for a varying proportion of this time ﬂocks are freeranging in rice ﬁelds to use left over rice, weeds and snails as a feed resource. Semi-conﬁned ﬁshponds with temporary or permanent shelters are used to house the ducks when they are not left to range in rice ﬁelds [47]. Two broiler chicken production systems need to be distinguished, namely long- and short-ﬁnish systems. Short-ﬁnish, ‘industrial’, broilers are usually kept indoors

g

on the ground, whereas long-ﬁnish, cross-bred broilers are kept under semi-conﬁned conditions to utilize some feed resources from the natural environment. The attention to hygiene and animal health management is normally lower than in breeder and layer ﬂocks. Chicken broilers are kept for a relatively short lifespan. ‘Industrial’ broilers are kept in batches of the same age for 5–7 weeks, while ‘cross-bred’ broilers are kept for 9–26 weeks and achieve a premium price in local markets. Similar to broiler chicken production systems, the production cycle length of broiler ducks varies depending on the breed and feeding system used. Scavenging broiler ducks are reared under similar conditions to layer ducks and are usually sold for slaughter after about 80 days. Conﬁned and intensively fed ducks, usually of speciﬁc meat type breeds, are ﬁnished in about 60 days [47–49]. Mixed backyard poultry production systems are characterized by scavenging indigenous birds that consume leftover feed and produce their own replacement chicks with very little cash investment by the owner [50]. The systems typically have high mortality rates in the early stages of life because of predation, poor diets and diseases [51–54]. Given the high mortality rates, particularly in young birds, a relatively high population turnover is common, in which birds get replaced by newly hatched chicks. This high population turnover signiﬁcantly limits the duration of ﬂock immunity that can be maintained with vaccination campaigns. The potential coverage of HPAI or Newcastle disease vaccination in backyard poultry with uncontrolled but more or less continuous replacement dynamics has been modelled by several authors (e.g. [39, 55, 56]). For nonbackyard poultry production systems, in which birds are periodically replaced by controlled reproduction, and which in most countries constitute a signiﬁcant share of the standing poultry population, no estimates of the maximum vaccination coverage achievable through mass vaccination were available in the literature. We therefore had to make our own estimates, which are based on the simplifying assumption that the birds within a speciﬁc ﬂock and production system are of the same age (i.e. all-in/ all-out ﬂock management) and that the age of all ﬂocks in a region is uniformly distributed (i.e. no seasonal production). The maximum ﬂock vaccination coverage with a vaccination campaign, in which vaccination teams visit each farm once only (or twice in the case of booster application) is then theoretically given by the time a particular ﬂock is eligible for vaccination within its production cycle (including the idle time between batches). The estimated maximum achievable vaccination coverage applies to the area that can be vaccinated within a day. For areas that require more than one day for vaccination teams to visit all ﬂocks, the theoretical maximum vaccination coverage is lower as a result of ﬂock turnover. The results of the calculations are displayed in Table 2.

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton

79

Table 2 Estimated ﬂock vaccination coverage and time until ﬂocks are fully susceptible again after an HPAI vaccination campaign

Layer chicken; vaccination during laying period Layer chicken; no vaccination during laying period Broiler chicken; industrial Broiler chicken; cross-bred Broiler ducks; intensive conﬁned Broiler ducks; scavenging

Time Length of Idle time eligible production between for cycle batches 1st shot (days) (days) (days)

Time eligible for 2nd shot (days)

Proportion of ﬂocks Proportion eligible for of ﬂocks 2nd shot eligible for 14 days 1st shot (%) later (%)

Time until all ﬂocks are 100% nai¨ve again after 1st shot (days)1

Time until all ﬂocks are 100% nai¨ve again after 2nd shot (days)1

490

0

476

462

97

94

180

180

490

0

140

126

29

26

147

133

32

14

18

4

39

9

25

11

120

14

106

92

79

69

113

99

60

21

46

32

57

40

53

39

80

–

66

52

83

65

73

59

Source: authors’ calculations. 1 Protection assumed to be lost 180 days after vaccination because of waning immunity.

Maximum Vaccination Coverage of Breeder and Layer Flocks

breeder and layer ﬂocks usually do not receive a booster vaccination [57].

Chicken breeder and layer ﬂocks In theory, relatively high vaccination coverage could be achieved for chicken layer ﬂocks with a production cycle length of 490 days, assuming owners would allow vaccination during the egg-laying period (Table 2). Since layer ﬂocks usually comprise of birds of several age groups (are not kept under all in and all out management), the coverage of 94% for a single-shot campaign and 92% for a double-shot campaign is a reﬂection of the share of birds that could be vaccinated at any point in time within a ﬂock of heterogeneous age composition. It should, however, be recognized that in most layer ﬂocks, birds are vaccinated before point of lay and owners are reluctant to revaccinate during egg-laying periods. If vaccination during the laying period is not accepted, within-ﬂock vaccination coverage would not exceed 26% in a double-shot vaccination campaign.

Scavenging layer duck ﬂocks The age distribution and the number of layer duck ﬂocks are related to the rice harvest seasons. Vaccination coverage similar to those in layer chicken could be achieved with well-timed vaccination campaigns that take into account the seasonality of rice harvest and laying period. However, the poor accessibility of scavenging duck ﬂocks in rice paddies makes the administration of vaccine, especially the second booster shot, very difﬁcult and experience from China shows that scavenging duck

g

Maximum Vaccination Coverage of Broiler Flocks Short-lived industrial chicken broiler ﬂocks During the average 32-day production cycle of an industrial batch of broiler chicken, vaccine can be administered over a period of 18 days, since vaccination should only be applied at a minimum age of 7 days and requires at least 7 days to confer protection, i.e. needs to be applied more than 7 days prior to slaughter to have any effect. Given an assumed average idle time of 14 days for cleaning and disinfection between batches, a signiﬁcant proportion (18/46) of broiler houses in a country or region will either be empty or not be populated by birds eligible for vaccination on any given day. Hence, the maximum achievable vaccination coverage with a single injection in a one-day visit vaccination campaign is 39% of all industrial broiler ﬂocks within a country or region. As described in ‘Characteristics of Commercial HPAI Vaccines’, two injections are required to achieve full protection in domestic poultry. Therefore, if each broiler ﬂock needs to be given a booster shot 14 days after the initial vaccination, only the ﬂocks that received their ﬁrst injection during the ﬁrst 4 days of the 18-day window will be eligible for the booster shot 14 days later. Although some previously non-eligible farms will have become available for the ﬁrst vaccination in the 14-day interval, only 9% of all broiler ﬂocks would have received the two injections required to achieve full protection. All industrial broiler ﬂocks would

CAB International 2010

80

Animal Science Reviews 2010

be nai¨ve again within 25 days after a single injection vaccination campaign, since vaccinated birds will have been slaughtered and replaced.

campaign, the population immunity rate is estimated to have dropped to 25%.

Intensively raised duck broiler ﬂocks The relatively short production cycle length of about 60 days for duck broilers in intensive closed systems results in a low maximum coverage of 40% for a two-injection at 14-day interval HPAI vaccination campaign. However, experience from China shows that ﬂock owners usually do not vaccinate these birds given their short life span [57].

Maximum Vaccination Coverage of National Poultry Populations

Long-lived cross-bred chicken broiler ﬂocks The maximum achievable vaccination coverage for crossbred broiler ﬂocks would be 79 and 69% for a one- or two-injection campaign, respectively. The higher coverage compared with industrial broiler ﬂocks results from the considerably longer production cycle. All cross-bred broiler ﬂocks would be nai¨ve again within 113 days after a single injection vaccination campaign. Scavenging duck broiler ﬂocks Seasonal peaks of available rice in harvested paddies are utilized by raising broiler ducks around the time of rice harvests. Therefore the timing of vaccination campaigns should take into account this seasonality. The maximum achievable vaccination coverage by a vaccination campaign will depend on the age distribution of duck broiler ﬂocks within a region. However, under the assumption that the start of broiler ﬂock raising in a region is uniformly distributed over a time period of at least 80 days, 65% of all ﬂocks can be vaccinated in a two-shot campaign. It should be noted that the delivery of a second shot is a challenge due to the difﬁcult accessibility of scavenging ducks in rice paddies and will require signiﬁcantly higher vaccinator time inputs for travelling and catching ducks.

Immunization Rates in Extensively Reared, Mixed Backyard Flocks Based on a spreadsheet poultry population model, Taylor [55] estimates that a maximum immunization rate of 52% of all backyard birds can be achieved with a two-shot vaccination campaign under the assumption that 80% of all chicken older than 4 weeks are caught for an initial vaccination and a booster shot 14 days later. Immunization coverage would fall to 19% within 17 weeks, because of replacement. These results assume a vaccine efﬁcacy of 80% and that 50% of the eggs are used for human consumption and the remaining 50% are hatched. Also under a two-shot vaccination scenario, vaccinating all poultry P14 days and assuming 80% vaccine efﬁcacy, Lesnoff et al. [39] estimate a maximum achievable population immunity rate of 55%. Seventeen weeks after the vaccination

g

The maximum vaccination coverage of national poultry populations achievable through vaccination campaigns is determined by the speciﬁc mix of ﬂock types in the national poultry industry and the time required to carry out a campaign. Given that in most countries, broilers are the most common type of poultry followed by backyard birds, while layer and breeder ﬂocks are comparatively rare, immunization rates needed to break infections chains will be difﬁcult to achieve unless vaccination campaigns are complemented by restocking bans.

Practical Challenges of Large-Scale HPAI Vaccination Programmes There are three main steps in the process of implementing vaccination programmes: (a) planning, which includes estimation of vaccine needs, identiﬁcation of vaccine source, storage, delivery, as well as procedures and schedules of administration, and respective budget allocation, (b) information campaigns and monitoring of vaccine quality, delivery and administration, and (c) programme evaluation that covers technical and cost elements and would lead to a revision of the programme if necessary. For this purpose, long(er)-term performance indicators are required. Depending on whether the public veterinary staff are responsible for the implementation of vaccination campaigns or whether their role is restricted to monitoring and ensuring the efﬁcacy of the campaign, public agents are involved in all or a part of the following tasks: selection of vaccines, monitoring the production or importation of vaccine, organizing the timely distribution of vaccines, and monitoring of ﬁeld virus strains and the efﬁcacy of the vaccine(s) employed. It needs to be emphasized that due to scarce veterinary staff and animal health funds, the inevitable effect of large-scale vaccination campaigns is a detraction of public animal health services from other disease control activities, both with respect to HPAI and other diseases (e.g. FMD outbreaks increased in Vietnam during and after HPAI vaccination campaigns). It has been emphasized by several authors that the achievable protection through vaccine use in the ﬁeld is unlikely to reach the potential shown under experimental studies with SPF chicken [3, 29]. Since commercially available HPAI vaccines are not thermostable, maintaining the cold chain from production to administration is crucial for obtaining high levels of immunization in vaccinated birds in the ﬁeld. This represents a signiﬁcant challenge in many developing countries with high daytime temperatures

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton

and shortages of cold storage capacity. A list of factors which compromise the effectiveness of avian inﬂuenza vaccination campaigns are provided by Wooldridge [58] and Alders et al. [45]: Inappropriately matched ﬁeld and vaccine strains. Vaccines that are inappropriate for the species or age group vaccinated. Spoiled (ineffective) vaccine. Inadequate hygiene during vaccine administration to birds (clean needle and disinfectants). Vaccine administration to inappropriate tissues of the birds. Infection of birds with immunosuppressive disease agents. Assessments of rinderpest [34] and classical swine fever [59] vaccination campaigns have identiﬁed similar risk factors to those mentioned above for avian inﬂuenza vaccination campaigns. The results of an evaluation of the cold chain for Newcastle disease vaccination in Mozambique have been documented by Chicamisse et al. [60]. All transport and storage systems exceeded the recommended temperatures. Some reasons for nonoptimal storage temperature were: storing too much vaccine in the refrigerator, too high temperatures in vaccine storage rooms, and old refrigerators. In addition to the above listed practical and technical issues, the perception of poultry owners and vaccinators with respect to the beneﬁts of vaccination is a crucial factor for vaccination to be effective. If owners and vaccinators are not convinced of the need for vaccination they are unlikely to be cooperative, which increases the risk of not achieving immunization targets. Information and communication campaigns are therefore necessary to ensure the preparedness and optimal cooperation of all stakeholders. The target audience comprises of a wide range of stakeholders with different information requirements, e.g. senior government decision-makers, veterinary staff, local authorities, poultry owners and poultry traders [45]. Properly trained and motivated (paid) vaccinators and cooperative poultry keepers are the key to achieve high vaccination coverage and immunization levels. Wooldridge [58] describes a formal risk assessment for vaccination campaigns which serves for identifying different risk pathways and risk reduction measures. Identiﬁcation of the necessary data to quantify the risk linked to each pathway facilitates the planning and monitoring of vaccination campaigns. An effective tracing and monitoring scheme for the vaccines used, storage locations and temperature, vaccinator, poultry ﬂocks and post-vaccination protection rate is necessary to accurately determine the cause of poor performance and implement the appropriate corrective measures. While such a standardized approach to vaccination campaign monitoring is seductive in its apparent accuracy it appears

g

81

to lack the necessary appreciation of the importance of the people involved in a vaccination campaign and their incentives to collaborate. The assumption is that their interest and willingness is a given, something that is rarely if ever the case. Given the changing nature of HPAI virus strains circulating in the ﬁeld, the need for post-vaccination monitoring of vaccine efﬁcacy is stressed in the HPAI literature. Virus isolation and sequencing should be an essential part of any vaccination strategy to detect potential genetic shifts and to monitor vaccine efﬁcacy. The possibility that the efﬁcacy of the vaccine(s) used in the programme decreases over time needs to be considered in the planning process because the frequency with which a new representative master-seed for the production of adapted vaccines needs to be found directly inﬂuences the proﬁtability of commercial vaccine production. Producers need a certain level of security about the potential scale and duration of market demand for their vaccine to embark in vaccine development and production. The recommended number of blood samples to test for sero-conversion and swab samples to test for virus presence and potentially isolate ﬁeld virus represents a major challenge for the capacity of most laboratories in developing countries.

Experiences from Large-Scale HPAI H5N1 Vaccination Programmes China Since the introduction of vaccination in China in 2004, the national vaccination policy has evolved. Until the end of 2005, all waterfowl and terrestrial poultry near waterways and wetlands were vaccinated [33]. Eight billion birds (60% of China’s domestic poultry population) were vaccinated between the onset of vaccination in 2004 and November 2005 [61]. In 2006, compulsory vaccination of all poultry was mandated and 8.2 billion head of poultry were vaccinated from January to September 2006 [62]. A vaccination coverage of 20–50% was achieved in mainly backyard poultry [2]. Since then, a twice-yearly vaccination campaign is estimated to result in the annual administration of 11 billion doses of vaccine [63]. Relatively low vaccination rates have been observed in waterfowl. The required booster vaccination for layer and breeder ducks is not actually administered in most ﬂocks and broiler ducks are usually not vaccinated because of their short life span [57]. Testing of 1113 chicken sera from Guangdong and Guiyang Provinces collected at markets in 2005 and 2006 revealed that only 180 (16%) were positive against Ck/HK/YU22/02 (H5N1) and that 55 of the positive sera had low or no neutralizing antibodies against the predominant FJ-like sub-lineage [64].

CAB International 2010

82

Animal Science Reviews 2010

Hong Kong A killed oil-adjuvanted H5N2 (A/chicken/Mexico/232/94/ CPA) vaccine was used in Hong Kong after HPAI outbreaks in February–April 2002. HPAI vaccination was introduced after a 12-month vaccination ﬁeld trial in commercial broiler chicken (yellow meat) ﬂocks to evaluate its effectiveness. All broiler chicken between 8 and 55 days of age were vaccinated followed by a booster 4 weeks later. In 75.8% (188/248 for a total of 1.35 million broilers) of the batches vaccinated, results were considered successful (P70% of chickens per batch had a HI titre of P16 and a geometric mean batch titre P20, see results in Table 4). Some HPAI outbreaks occurred on farms shortly after vaccination. In one farm, where chicken developed clinical HPAI 9 days after vaccination, vaccination had not yet induced a protective immune response. In a few other farms that became infected it was shown that vaccination provided protection from disease and reduced virus excretion by 13–18 days postvaccination [18]. Compulsory vaccination was introduced for local chicken farms in June 2003 and by the end of 2003 all chicken from Hong Kong entering the live poultry were vaccinated using the killed H5N2 vaccine. The farm biosecurity enhancements introduced between 1998 and 2002 were not sufﬁciently effective to prevent the spill over of circulating virus from markets to farms. The contact rates between broiler chicken farms and live poultry markets were relatively high as a result of rearing of several batches of different ages on the same farm. HPAI vaccination was introduced as an obligatory condition for poultry farms to have market access, which required substantial enforcement and laboratory capacity (Sims, personal communication). In addition to vaccination, bio-security measures were further enhanced on farms and live poultry markets [33]. No outbreaks were detected until 2008. Surveillance results showed that virus circulation in live poultry markets had ceased for the 5-year period after the introduction of vaccination [65]. The re-emergence of an HPAI outbreak on a vaccinated chicken farm in Hong Kong in December 2008 led to speculation3 about vaccine efﬁcacy.

Egypt Voluntary vaccination with registered vaccines was permitted by the government in 2006. Over 750 million birds were vaccinated with H5N2 and 150 million with H5N1 in commercial farms [66]. For backyard or household village

3 The China Post cited Guan Yi from University of Hong Kong on 12 December 2008, based on a press release on 9 December 2008 by Dr York Chow, Secretary for Food and Health Bureau, Hong Kong.

g

poultry, vaccine was provided by the government since mid-2007 [67]. Seventy million birds were vaccinated with H5N1 in backyards and 11 million DOCs were vaccinated with H5N2.

Indonesia By March 2008, 20 vaccines were registered for use in poultry in Indonesia. Vaccine seed strains include a highly pathogenic local strain, A/chicken/Legok/2003 (H5N1), A/ckichen/Mexico/94 (H5N2), A/turkey/ Wisconsin/69 (H5N9) and A/turkey/England/73 (H5N2). The widely used Legoc/03 strain is not fully protective against all circulating viruses. The OFFLU network is therefore collecting HPAI virus isolates and screening these to identify a new vaccine seed strain, which affords better protection against circulating HPAI viruses [68]. A compulsory vaccination policy for all poultry was adopted by the Government of Indonesia in June 2004. Poultry classiﬁed as sector 4 (backyard) and sector 3 (small-scale commercial) up to maximum ﬂock size of 5000 birds were to be vaccinated free of charge by government services. Problems were reported with respect to vaccine availability and conﬁdence of ﬂock owners in the beneﬁts of vaccination [69]. Vaccination coverage assessed by the Ministry of Agriculture rarely exceeded 30% of the poultry population. Post-vaccination monitoring of HI-titres ranged from 11% with protective titres (HI-titre P16) in native chicken populations in Bali to 78% protected in industrial breed chicken populations in Medan [69]. The commercial poultry sector implemented vaccination programmes using a variety of protocols mainly in chicken layer ﬂocks (including grandparent and parent ﬂocks) until point of lay. Short-lived broilers were not vaccinated except in areas considered to be at high risk and was based on a half dose shot at 7 days of age [69]. No information was available on vaccination in long-lived broilers and male layers (90–120 day ﬁnish), which make up a substantial proportion of the poultry raised for meat. A vaccination pilot study in semi-intensive and native chicken layer farms coupled with the use of sentinel chicken was conducted in Sukabumi to assess the efﬁcacy of locally produced homologous vaccines. Farmers were reluctant to have sentinel birds in their systems. After receiving two doses of vaccine, 75% of the semi-intensive layer chicken hens developed HI-tires P32, while only 15–40% of the native layer chicken hens in 4 vaccinated ﬂocks showed a titre P16. Blood samples from sentinel birds remained sero-negative for HI [69]. Vaccination of backyard poultry was conducted under a World Bank and USAID funded operational research project. A total of 2.9 million birds in 425 villages located in 32 sub-districts in the provinces West Java, Yogyakarta and Central Java received two doses of vaccine. It took 3 weeks each to administer the ﬁrst and second shots of

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton

83

Table 3 Estimated vaccination coverage in Vietnam, 2005 and 2006 HPAI vaccination campaigns

Year

Campaign

Shot

Chicken population vaccinated (%)

2005 2005 2006 2006 2006 2006

1st 1st 1st 1st 2nd 2nd

1st 2nd 1st 2nd 1st 2nd

66 58 51 35 47 22

Duck population vaccinated (%) 79 73 58 22 72 26

Source: Taylor [78], based on poultry population census data.

Legok 2003 H5N1 vaccine and required 64 community vaccinator coordinators and 1088 community vaccinators. The campaign was estimated to have covered only 32% of the 9.0 million poultry reported by ofﬁcial livestock statistics to be in the sub-districts where vaccination was implemented [70]. About 18 000 blood samples from vaccinated backyard chicken and ducks were taken for sero-monitoring, of which 33.1% showed a protective HI titre P16 [71], Table 4). The participatory disease surveillance and response (PDSR) project also vaccinated poultry. The PDSR teams visited 9326 villages, detected HPAI infection in 494 of them and vaccinated poultry ﬂocks in 56 (11.3%) of the infected villages [67].

Pakistan Strategic vaccination and increased biosecurity was used in 1995 to control outbreaks of H7N3 HPAI. The outbreak was controlled within 4 months. However, H7N3 HPAI virus re-emerged in November 2003. The HPAI virus originated from an LPAI outbreak in April 2003. Breeder and layer ﬂocks were vaccinated prophylactically and no H7N3 virus has been isolated since January 2005 [72]. The vaccination strategy in the area affected by HPAI H5N1 in 2007 has been to vaccinate all ﬂocks within 3 km of the outbreak with a water-based vaccine. Ten days after the ﬁrst vaccination a second vaccination with an oil-based vaccine was given. Breeder and layer ﬂocks were given up to 4 shots per bird. Infected ﬂocks were destroyed by immediate burial. Mild clinical signs with some mortality have been observed in vaccinated ﬂocks [33].

Vietnam A large-scale vaccination programme of domestic poultry against HPAI based on twice-yearly vaccination campaigns has been conducted since mid-2005 by public veterinary services. This vaccination programme was combined with

g

supplementary age-based vaccination of larger commercial poultry ﬂocks between the main campaigns [73]. The vaccination policy, in terms of the eligible poultry population and cost-recovery schemes, has evolved over time. Currently, all domestic chicken and ducks older than 7 and 15 days, respectively, are eligible for vaccination. However, broiler chickens and ducks kept for less than 70 days until slaughter and for broiler chicken to be slaughtered within 15 days are exempted. Vaccine delivery has required immense labour inputs from public veterinary services and contracted commune animal health workers. According to a survey conducted by Cristalli in 2006, about 25 000 commune animal health workers were involved in administering vaccine during the 2005 and 2006 campaigns [33, 73, 74]. Farm labour inputs to these campaigns have not been estimated. Initially, ﬂocks with less than 2000 birds were vaccinated free of charge. This ﬂock size cut-off point for free vaccination has been reduced to 500 birds for the 2009 campaigns. Each campaign, comprising of two injections within a 4-week interval, required about 60 days to be completed. In 2005, 158 million chicken and 72 million ducks were vaccinated, while for 2006, the respective numbers were 177 million and 28 million [75]. In 2007, 164 million poultry were vaccinated during the ﬁrst campaign, comprising of 87 million chicken and 74 million ducks plus 42 million doses of vaccine administered by private livestock ﬁrms [73, 75, 76]. Vaccination costs have been estimated at US$20 million annually [77]. In 2008, the use of 500 million doses of HPAI vaccine was planned. Vaccination coverage estimations based on the number of vaccinated domestic poultry in relation to available domestic poultry census have been conducted by Taylor [78] (Table 3) and To et al. [79]. The data show a decreasing trend in vaccination coverage, which could be related to increasing vaccination fatigue of vaccinators and reduced vaccination acceptance of poultry owners. Post-vaccination monitoring results with regard to the prevalence of protective antibody levels in vaccinated poultry are provided in Table 4. The proportion of tested poultry with protective antibody levels varied between 33 and 72%. No outbreaks of avian inﬂuenza were reported in Vietnam between December 2005 and December 2006.

CAB International 2010

84

Animal Science Reviews 2010

Table 4 Prevalence of protective HPAI antibody levels in poultry after vaccination campaigns Country

Species

‘Protected’

Comment(s)

Source

China China

n.s. Chicken

69% (n=n.s.) 16% (n=1113)

In 2004 Sera collected from November 2005 to April 2006 in Guangdong and Guiyang Provinces; HI titre P20

EFSA [33] Smith et al. [64]

Egypt Hong Kong

n.s. Broiler chicken

Indonesia

Chicken and ducks

25.6% (n=160) 75.8% batches (n=248 batches) 33.1% (n18 000)

Vietnam

n.s.

53.8% (n=364)

Vietnam

n.s.

44.9% (n=203)

Vietnam

n.s.

56.2% (n=269)

Vietnam

n.s.

33.3% (n=43)

Vietnam

n.s.

72.1% (n=1263)

Vietnam

Ducks

55% (n=182)

Vietnam

Chicken Ducks

40% (n=30) 63% (n=302)

Chicken

37% (n=57)

1 month after second vaccination, HI titre P16 1–2 months after campaign ended, H5N1 HI titre P16 1–2 weeks after end of 1st round of 2005 campaign 3–4 weeks after end of 1st round of 2005 campaign 1–2 weeks after end of 1st round of 2006 campaign 3–4 weeks after end of 1st round of 2006 campaign 1 month after end of 1st round of 2007 campaign 4–6 weeks after end of the 1st round of the 2007 campaign >12 weeks after end of the 1st round of the 2007 campaign

CIRAD [80] Ellis et al. [18] McLaws [71] Nguyen Van Long [81] Nguyen Van Long [81] Nguyen Van Long [81] Nguyen Van Long [81] Nguyen Van Long [81] Henning et al. [82] Henning et al. [82]

n.s.: not speciﬁed.

Table 5 Mass vaccination campaign costs per vaccination (USD/100 birds) Layer

Broiler

Backyard

Cost component

Indonesia

Vietnam

Indonesia

Vietnam

Indonesia

Vietnam

Vaccine Vaccinator (labour, equipment) Storage and distribution Post-vaccination monitoring Planning and communication Total

3.18 0.84 0.04 0.02 0.03 4.12

4.06 0.36 0.01 0.02 0.31 4.75

3.18 0.84 0.04 0.02 0.03 4.12

2.03 0.36 0.01 0.02 0.31 2.72

3.18 4.88 3.27 0.02 0.03 11.38

2.03 1.36 0.01 0.02 0.31 3.72

(77%) (21%) (1%) (0%) (1%) (100%)

(85%) (8%) (0%) (0%) (6%) (100%)

(77%) (21%) (1%) (0%) (1%) (100%)

(75%) (13%) (0%) (1%) (11%) (100%)

(28%) (43%) (29%) (0%) (0%) (100%)

(55%) (36%) (0%) (0%) (8%) (100%)

Source: authors’ calculations.

However, assessment of the contribution of vaccination to the absence of reported outbreaks is challenging, since vaccination has not been implemented as an isolated control measure. Other factors, such as improved hygiene practices together with a signiﬁcant reduction in the susceptible poultry population subsequent to culling and market reactions in 2004, probably also have contributed to the reduction of outbreaks. However, in 2005, at the time the decision to embark on mass vaccination campaigns was taken, a reduction of the human infection risk rather than disease eradication was the main goal in an emergency situation with several human cases [83]. Table 4 provides a compilation of post-vaccination sero-monitoring results obtained in different countries which have embarked on large-scale HPAI vaccination programmes. In individual birds, the prevalence of antibody levels regarded as protective ranges from 16 to 72.1%.

g

Costs and Incentives for HPAI Vaccination Costs of Large-scale Vaccination Campaigns in Indonesia and Vietnam Essential cost components of any mass vaccination programme comprise of (i) planning, monitoring and communication, (ii) vaccinator labour and equipment, (iii) storage and distribution of vaccine and equipment, (iv) post-vaccination sero-monitoring and (v) vaccine. The recurrent and investment costs differ between countries because of different economic and (veterinary) infrastructure conditions and because of differences in the structure of the poultry industry. Unfortunately, comprehensive cost estimates for vaccination campaigns are only available for Indonesia and Vietnam. The vaccination costs presented in Table 5 are derived from ex-ante assessments for a government-run mass

CAB International 2010

0.09

0.28

40

120

Source: Poultry production survey data from May 2008 in Vietnam [85]; international egg production costs comparison data from Thailand, Vietnam and Indonesia for June 2006 [86]; backyard production parameters and costs from Otte [50]; Aviagen1 and Hy-Line1 poultry management guides; own calculations for selected examples of existing production systems. Prod. cycle= production cycle. 1 ‘Standard ﬂock’ of 16 birds including 5 chicks, 6 growers, 4 hens, 1 cock.

0.02 0.71 0.46 0.68 0.39 0.15 0.03 0.22 0.08 2.17 70.13 0.63 0.08 70.12 0.02 0.12 0.06 70.03 0.15 70.03 0.35 0.92 0.09 0.09

0.80 3.14 0.80 0.74 0.37 4.86 4.86 0.41 0.26 19.55 0.37 2.26 0.37 0.37 0.31 3.84 3.84 0.25 0.09 14.38

5 5 5 5 10 12 12 20 46

16.0 45.0 61.0 4.6 8.6 39.0 39.0 21.0 3.0

25.0 54.0 79.0 6.7 25.7 41.0 41.0 23.0 5.0

10.0 43.0 53.0 5.0 4.4

0.25 0.23 0.23 0.44 0.44

0.46 0.46 0.46 0.62 0.62

2.26 3.86 3.86 3.09 3.55 4.29 4.29 3.84 0.25 38.20

3.14 4.89 4.89 5.71 5.68 5.44 5.44 4.86 0.37 48.70

320 320

0.15 0.11

0.94 0.48 70.25 160 30.00 10.00

5

63.0

65.0

52.0

0.23

0.46

3.86

4.89

0.22

44.97 13.26 6.00 18.00 160 4.89 3.86 0.46 0.23 52.0 65.0 63.0 74.00 38.00

DOC to grandparent (meat purpose) layer spent hen DOC to parent (meat purpose) layer spent hen DOC to pullet Pullet to spent hen DOC to spent hen DOC to broiler (industrial) DOC to broiler (cross-bred) Backyard grower to spent hen Backyard grower to cock Backyard chick to grower Backyard egg to chick Backyard ﬂock1

Max. Min. Production system

5

Max. Min. Eggs (no.) Min. Max. Max. Min.

Feed costs Feed (USD/kg) (kg/prod. cycle) Min. Max. Prod. cycle Mortality length (weeks) (%/prod. cycle) Min. Max. Bird input (USD)

Table 6 Chicken values and gross margins in different production systems in Southeast Asia

Bird output (USD)

Egg price (USD)

Gross margin (USD/bird & week)

J. Hinrichs, J. Otte and J. Rushton

g

85

vaccination campaign in Vietnam and a planned mass vaccination campaign in Western Java, Indonesia. The total costs per vaccination vary between US$0.03 in broiler ﬂocks and US$0.12 in backyard ﬂocks in Indonesia. HPAI vaccination costs per bird differ between production systems because of varying accessibility and ﬂocksize-related differences in achievable vaccinations per vaccinator and day4. Predominantly scavenging extensive backyard ﬂocks in remote areas demand signiﬁcantly higher labour input for vaccination than conﬁned chicken broiler production systems. The relatively high storage and distribution costs for the vaccination of backyard poultry in Indonesia result from necessary investments in motorbikes to supply vaccinators with vaccine.

Private Incentives for HPAI Vaccination If poultry owners see an economic advantage in vaccinating their ﬂock against HPAI compared with (or in addition to) applying other control measures, they are more likely to comply with a compulsory national vaccination strategy and a higher level of vaccination coverage can be achieved [84]. Whether poultry owners regard vaccination as a ﬁnancially worthwhile risk reduction measure not only depends on the infection risk and cost of vaccination but also on the overall proﬁtability of their respective poultry enterprise. Indicative values for production inputs, prices of inputs and outputs, performance indicators and production margins for the various production systems in Southeast Asia are provided in Table 6. As market prices for feed and poultry products have been highly volatile and differ between countries in the region, the presented values and prices reﬂect the price range since 2006. The gross margins for the production systems are calculated per production cycle and only account for feed costs and the costs of replacement birds. From a ﬂock owner’s perspective vaccination represents a protective measure against the economic impact of a HPAI outbreak. The decision on whether to contract the ‘vaccination insurance policy’ depends on the vaccination costs (‘insurance premium’), which also include production impacts of vaccination such as decreased egg laying, the expected economic loss in the case of an outbreak, and the perceived probability of an outbreak. The ratio of vaccination costs to outbreak losses (‘breakeven outbreak risk per year’ in Table 7), adjusted for an average 80% vaccine efﬁcacy, indicates the probability of ﬂock infection at which expenditure on vaccination would be proﬁtable for a risk-neutral ﬂock owner. The expected absolute loss in the case of an outbreak

4

One hundred and ﬁfteen (300) birds per day and vaccinator assumed in backyard systems and 500 (500) birds per day and vaccinator assumed in broiler and layer systems in Indonesia (Vietnam).

CAB International 2010

86

Animal Science Reviews 2010

Table 7 Financial breakeven outbreak risks for HPAI vaccination

Production system DOC to grand parent (meat purpose) layer spent hen DOC to parent (meat purpose) layer spent hen DOC to pullet Pullet to spent hen DOC to spent hen DOC to broiler (cross-bred) DOC to broiler (industrial) Backyard grower to spent hen Backyard grower to cock Backyard chick to grower Backyard egg to chick Backyard ﬂock

Vaccinations required per production cycle

Vaccination costs per production cycle (US$ cents)

Vaccination costs per year (US$ cents)

HPAI outbreak loss (US$)1

Breakeven outbreak risk per year (%)2

Breakeven outbreak frequency in years2

min.

max.

min.

max.

min.

min.

min.

max.

4

16

19

13

15

55

182

289

1141

4

16

19

13

15

1

6

3

20

5

39

3 1 4 2

12 4 16 5

14 5 19 8

24 4 11 10

41 5 16 40

1 3 2 2

2 5 4 4

18 1 3 3

84 3 9 28

1 39 11 4

6 104 32 35

2

5

8

33

65

1

6

7

75

1

14

1

4

11

5

15

2

3

2

8

13

56

1

4

11

5

15

2

3

2

9

11

48

1

4

11

8

28

3

3

3

14

7

31

2

7 72

23 219

77 453

394 2191

0.02 23

1 33

190 17

29 987 120

0 1

1 6

max.

0.1

max. 0.3

Source: authors’ calculations. 1 A potential outbreak is assumed to occur in the middle of the production cycle which would lead to the loss of 50% of the bird value at the production cycle end and 4 weeks of lost gross margin due to subsequent downtime for cleaning and disinfection. 2 Based on 80% vaccine efﬁcacy.

varies during the production cycle. The applied absolute losses for the calculations of the break even risks in Table 7 are based on the assumption of a 50% loss of the maximum bird value and a 4-week gross margin loss associated with production downtime subsequent to an outbreak. In case outbreaks occur when birds are of lower value than assumed in Table 7 or a salvage value can be derived from selling sick birds, the calculated breakeven risks are underestimated. The breakeven risks presented in Table 7 are underestimates due to several other reasons. In all production systems, the ‘background’ mortality requires vaccinating more birds than will eventually be produced. Vaccination can depress production, leading to the foregoing of revenue. In Vietnam, layer chicken reportedly showed a decrease in egg production of 5% over 3 weeks following vaccination [73]. This cost component has not been accounted for in the estimation of vaccination costs for layer ﬂocks. The vaccination campaign costs in Table 5 were used to calculate vaccination costs per production cycle in Table 7, while an age-based vaccination scheme to achieve the maximum immunization coverage of the birds during the production cycle is assumed. For shortproduction-cycle broiler ﬂocks, in particular, this is not achievable through mass vaccination campaigns. A more continuous vaccination delivery system would be required, for which the costs are likely to be higher than

g

the cost estimates from vaccination campaigns used to calculate the ﬁgures in Table 7. Grandparent and layer-chicken production systems have a relatively low estimated breakeven outbreak risk. Compared with vaccination in other production systems, vaccination in these production systems would be most cost-effective and ﬂock owners have a relatively high private incentive to adopt vaccination and HPAI vaccination is commonly used by parent stock keepers in Indonesia [69]. However, even for relatively valuable laying hens with an estimated potential HPAI outbreak loss of US$3–5 per bird, the breakeven risk of 1–3% would not economically justify the use of vaccination for a risk-neutral poultry keeper under infection risk conditions similar to the peak HPAI H5N1 incidence in 2004 in Thailand and Vietnam, which was around 0.2% [87]. Nevertheless, HPAI vaccination is reportedly widely used by layer ﬂock owners in Vietnam and Indonesia (Sims, personal communication), which indicates that these poultry producers are risk-averse and regard vaccination as an important component of their strategy to reduce the likelihood of the high economic losses an HPAI outbreak would cause. Although the value of broiler ﬂocks increases over their relatively short life span, broiler producers have a relatively small ﬁnancial incentive to vaccinate against HPAI with a vaccine that becomes fully effective only

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton

about 13 to 21 days after the ﬁrst of two injections, since they only keep chicken for a short time thereafter. The ﬁnancial incentive to vaccinate duck broilers, despite their longer production cycle is also low as they normally only show mild clinical signs of disease when infected with HPAI [14]. Vaccination of an average backyard chicken ﬂock would only be ‘proﬁtable’ for a risk-neutral ﬂock owner, if the annual risk of HPAI infection were higher than 17% and resulted in a loss of US$1–3 per bird. Since such a high infection risk is very unlikely, the average beneﬁt of freeof-charge vaccination for backyard chicken ﬂock owners would be marginal. Notwithstanding the relatively high gross margins of backyard poultry production, poultry keepers have shown little motivation to vaccinate against HPAI or other more prevalent diseases such as Newcastle disease. A survey on the willingness to pay for HPAI vaccination in Vietnam showed that only 32% out of the surveyed 62 poultry owners used Newcastle vaccine at an average price of US$0.026 per chicken [85]. If other vaccines to prevent more ﬁnancially relevant diseases were delivered in addition to HPAI vaccination and if HPAI-vaccinated ﬂocks would be exempted from culling, subsidized HPAI vaccination might be a pro-poor disease control intervention. The breakeven outbreak frequency is given by the reciprocal of the breakeven outbreak risk per year (per production cycle) and indicates the time period in years (in production cycles) within which the use of vaccination is expected to result in the same costs as the disease itself. The breakeven outbreak frequency for an industrial broiler ﬂock owner would be one outbreak every 10 to 82 production cycles, depending on the assumed vaccination costs and losses. However, outbreak risk and absolute loss of an outbreak are not equally distributed over the time period of a production cycle. The economic value at risk is correlated with the feed investments in growing broilers. Broilers within a batch usually do not grow homogenously and are therefore not ﬁnished at exactly same time, which leads to the sale of smaller batches over several days. The interaction with traders is likely to increase the infection risk of the remaining birds towards the end of a production cycle. Protection through vaccination for this speciﬁc time point, however, needs to be started early in the production cycle. If market access of poultry producers is made conditional on the proven use of vaccination, as is the case in Hong Kong and Ho Chi Minh City, Vietnam, vaccination is likely to be used by producers supplying these markets. However, this only holds if the capacity to control market access and the diagnostic capacity to test for antibodies is sufﬁciently high. The higher the difference between the calculated breakeven outbreak risk and the actual outbreak risk, the higher is the incentive for ﬂock owners to ‘bypass’ market access restrictions. Flock owners may consider other available protection measures such as improvements of production hygiene

g

87

and investments in cleaning and disinfection equipment. The calculated vaccination costs to achieve the maximum achievable protection through vaccination could be considered as a benchmark for the maximum expenditure on other protection measures. For a ﬂock of 1000 industrial broilers the annual vaccination costs vary between US$325 and US$651. Annual vaccination costs would rise to US$106–157 for an integrated layer production system with a ﬂock size of 1000 birds. Detailed assessments of the speciﬁc risk factors for the entry of HPAI virus into these production systems would be essential to estimate the potential feasibility, costs and effectiveness of achieving a higher disease protection level for the respective ﬂocks. Nevertheless, simple improvements, such as cleaning and disinfection of equipment, cages, and work clothes are likely to cost less than the above estimated costs for vaccination. It is recognized that the ease of applying such measures will differ between systems and the quality of housing. For example, layer units with multiage ﬂocks may not be in a position to regularly disinfect units and may have difﬁculty in cleaning egg trays, whereas an all-in/all-out broiler system with concrete ﬂooring may be in a better position to apply cleaning and disinfection measures. Such measures require signiﬁcant investments in training and then need to be followed with management so that they are continuously applied. Production hygiene improvements would also likely have additional beneﬁts from reduced mortality and morbidity and subsequently increased productivity.

Public ‘Returns’ to Vaccination Positive externalities from vaccinating poultry ﬂocks result from the reduced probability of secondary outbreaks and the public health beneﬁts from reduced exposure to HPAI virus, neither of which is taken into account in the estimated ﬁnancial incentives for vaccination presented in Table 7. The main beneﬁt of vaccination for the community of poultry producers stems from the expectation that vaccinated ﬂocks will act as ‘dead ends’ of infection chains and thereby also indirectly ‘protect’ non-vaccinated ﬂocks. This indirect beneﬁt is deemed to outweigh the ‘direct’ beneﬁts of vaccination, but, given the uncertainty surrounding prevented infections/outbreaks, the inclusion of prevented outbreaks in public cost-beneﬁt analyses is problematic. For Vietnam, Soares Magalhaes et al. [88] estimated that vaccination of all commercial smallholder farms would more than halve ‘secondary’ outbreaks by reducing Rn from 2.24 in the case of culling of infected premises and pre-emptive cull in a 3-km zone to 1.05 in the case of the same measures complemented by vaccination. Additional immunization of 25% of backyard ﬂocks was estimated to further reduce secondary outbreaks from 1.05 to 0.23 per infected premises. These major reductions in secondary outbreaks would result in much

CAB International 2010

88

Animal Science Reviews 2010

higher public cost-beneﬁt ratios than those that can be derived from Table 7 for individual producers, the order of magnitude depending on the generations of cases avoided. Assuming three successive generations of cases resulting from one HPAI H5N1 outbreak, vaccination of commercial smallholder farms (in addition to standard culling of infected and surrounding premises) would theoretically lead to a 90% reduction of the outbreak size, i.e. a ten-fold increase in cost-effectiveness. Disincentives are needed to counter the ‘free-rider’ problem that results from the above positive externalities that reduce overall disease risk, from which farmers who do not vaccinate beneﬁt without incurring respective costs. To reduce the public health risk, it would be essential to limit the contact of humans with infected poultry. Broilers represent the largest share of poultry that is marketed through live poultry markets and in contact with a magnitude of consumers in many developing countries. Effective immunization of broilers would reduce the exposure of live-bird market customers to HPAI virus. However, high immunization coverage of marketed broilers is not likely to be achieved because of low economic incentive for ﬂock owners to vaccinate their birds. Subsidized vaccination in these ﬂocks may be justiﬁed for public health reasons, but even if the vaccine delivery was entirely free, owners’ willingness to participate may be affected by their perception of risk and the impact of a vaccination campaign on production in terms of potential losses of birds and their condition. In addition, these costs need to be assessed against market hygiene interventions. The required vaccination costs to supply a medium size live bird market with a daily trade volume of 1000 broilers would amount to US$1151–1707 per month. Similarly to the situation on broiler farms, a detailed assessment of the costs, effectiveness and feasibility of other market hygiene improvements and behaviour changes need to be considered in order to choose the most cost-effective risk reduction strategy.

Discussion and Conclusions The available literature on ﬁeld vaccination experiments with commercially available vaccine indicates that antibody titres considered as protective can develop within 13 days after the ﬁrst vaccination. However, with the exception of Trovac, two injections at two-week intervals are required to achieve full protection and one of the few long-term serologic response studies indicates that immunity is lost in most chicken 20.5 weeks after vaccination. In general, vaccinated birds have been shown to shed less amounts of virus than unvaccinated controls at speciﬁc times post challenge. Thus, most commercial vaccines have the potential to reduce the level of circulating virus in infected chicken populations. However, a crucial factor for achieving signiﬁcant reductions in circulating virus in poultry ﬂocks are sufﬁciently high vacci-

g

nation coverage levels (50–90% immunization of at least 50% of all ﬂocks at risk of infection) with a vaccine that protects against the circulating virus(es). Both theoretical considerations as well as ﬁeld observations show that such high immunization rates are difﬁcult to attain in large poultry populations through vaccination campaigns and that they are even more difﬁcult to maintain over a longer time period given the high population turnover in short-lived commercial broiler5 and mixed-age backyard poultry ﬂocks. There are also problems of maintaining immunity levels in long-lived commercial layer and parent ﬂocks as the currently available vaccines do not lead to lifetime immunity. The short- to medium-term gains in reducing the virus load with vaccination are not likely to result in a cost-effective long-term control approach, if no additional measures are in place, because infection chains are unlikely to be totally interrupted and virus will not be eliminated from the entire poultry population. A major drawback of vaccination is that the probability of detecting outbreaks may decrease because of a lack or reduction of clinical signs, which could lead to the silent spread of virus [38]. Incentives for disease reporting are relatively low and masking disease signs through vaccination further depresses an already low level of reporting. For northern Vietnam, Walker et al. [42] estimated a 45% effective vaccination coverage achieved by mass vaccination campaigns, not only leading to a greatly reduced transmission of virus between communes but also to an increase in the commune-level infectious period because of outbreaks remaining unreported for a longer duration. The same authors estimated that, had detection levels been maintained at pre-vaccination levels, around twothirds of outbreaks which occurred in the 2007 wave in northern Vietnam would have been prevented. This highlights the fact that, regardless of the underlying reasons for less rapid reporting of outbreaks, in order to translate the reductions in disease transmission following vaccination into greater gains in disease control, more effective reporting and surveillance strategies are required. Another drawback of the extensive use of vaccination is the increased likelihood of genetic drift as seen in Mexico [28, 30] and the USA [31]. Therefore close virus monitoring of circulating ﬁeld strains, continuous vaccine testing via challenge trials, and subsequent development of new vaccines that protect from infection with evolving ﬁeld strains are an inevitable component of any longerterm routine vaccination programme. This requires considerable ﬁnancial resources and supporting activities have to be based on surveillance systems that have a high

5

Theoretically, close to 50% of all vaccinated broilers have been replaced by non-vaccinated birds in the 60 days required by the Vietnamese animal health system to conduct one national vaccination campaign.

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton

probability of detecting circulating HPAI viruses even in the absence of signiﬁcant clinical disease. It also requires the sharing of isolates with laboratories capable of assessing the suitability of the vaccines used. At present, these signiﬁcant ‘collateral’ investments to vaccination are rarely found in countries with problems of HPAI endemicity. In Indonesia, a donor-funded OIE/FAO network of expertise on Avian Inﬂuenza (OFFLU) is monitoring avian inﬂuenza virus variants and the efﬁcacy of vaccines used in commercial poultry production systems [89]. Short-lived broilers, mainly chicken but also ducks, constitute a relatively large share of the standing poultry population of most countries, which, given their rapid turnover, provide a constant and ample supply of susceptible avian hosts. Campaign-based vaccination programmes can only achieve a very low coverage in these systems, particularly if two injections are required to achieve immunity. An age-based vaccination schedule for broilers would be an option to achieve higher vaccination coverage and its maintenance over time, but the logistical requirements for age-based vaccine delivery and associated costs differ signiﬁcantly from those of vaccination campaigns. The private incentives for owners of broiler ﬂocks to regularly vaccinate replacements are low and even if owners do vaccinate, broiler ﬂocks will remain at least partially susceptible for 2 to 3 weeks, i.e. most of their life span (unless Trovac is used and protects against circulating virus strains). Broilers thus represent the ‘Achilles’ heel’ of any HPAI control strategy that relies, at least to some extent, on the use of vaccination. Although vaccination of more valuable breeder and layer ﬂocks is generally more ‘proﬁtable’ from the ﬂock owners’ perspective, the incentives to vaccinate are not constant over the production cycle and immunity of birds might have waned towards the end of their productive life. Also, as breeder and layer ﬂocks have relatively high contact rates with other ﬂocks and as HPAI vaccination is frequently used in these production systems, postponed detection of infection because of potential masking of symptoms by vaccination may undermine the success of a vaccination strategy in breeder and layer systems. Upgrading of bio-security is likely to be safer and more costeffective in these production systems than vaccination. From a public health and national health security perspective, the reduction of human cases of avian inﬂuenza as a means of reducing the risk of a national panic and global pandemic is most important. Human cases of avian inﬂuenza receive high media attention and the political pressure to act is high. The impact of poultry vaccination on human health risk is controversially debated in the scientiﬁc community. Human cases of H5N1 infections in China in January 2009 raised concerns about the role of vaccination in increasing the virulence of HPAI virus and masking its symptoms in poultry [63]. Hygiene practices and awareness of risk factors for poultry to human transmission are possibly as important for preventing human infections as reducing virus shedding by vaccination.

g

89

The cost-effectiveness of national vaccination efforts need to be weighed against those of alternative measures to reduce disease spread in the national ﬂock. In Vietnam for example, the culling strategy employed during the ﬁrst wave of outbreaks led to the destruction of about 44 million birds (20% of the standing poultry population) and caused major losses to poultry owners and costs to the government. However, even this extensive depopulation of poultry ﬂocks was not sufﬁcient to break the chain of infection in all locations [90]. As a consequence, the government decided to use vaccination as an additional control measure, which, in combination with a modiﬁed culling policy, reduced the number of culled poultry, but added substantial vaccination costs. On the other hand, Thailand managed to very signiﬁcantly reduce or even eliminate the circulation of H5N1 virus in its domestic poultry population within 2 years without resorting to vaccination, largely through intensive active and passive surveillance combined with, progressively restricted, culling in the case of outbreaks. The high and recurrent costs, technical difﬁculties, and epidemiological drawbacks of large-scale, open-ended blanket vaccination programmes in national efforts to control HPAI call for careful targeting of vaccination in national control strategies, which ‘intelligently’ combine available disease control measures. In principle, vaccination can be targeted spatially, temporally, and/or by production system to maximize its impact and costeffectiveness. Effective targeting, however, requires sound risk assessments, for which data and expertise are often lacking. Strengthening of the epidemiological capacity of national animal health systems would thus be a major prerequisite for large-scale use of vaccination in the control of HPAI.

Acknowledgements We thank all those who provided the sources on which we drew and are grateful for the constructive suggestions made by colleagues, in particular Professor Dr Dirk Pfeiffer at Royal Veterinary College and Leslie Sims, and reviewers in the course of the gestation of this paper. We also express our gratitude to the UK Department for International Development (DFID) for funding this work through the HPAI risk reduction project (GCP/UK/804/ INT).

References 1. Swayne DE, Kapczinski D. Vaccines, vaccination and immunology for avian inﬂuenza viruses in poultry. In: Avian Inﬂuenza. Blackwell Publishers, Ames, IA; 2008. p. 407–451. 2. Peyre M, Fusheng G, Desvaux S, Roger F. Avian inﬂuenza vaccines: a practical review in relation to their application in the

CAB International 2010

90

Animal Science Reviews 2010 ﬁeld with a focus on the Asian experience. Epidemiology and Infection 2009;137:1–21.

vaccine suppresses shedding and prevents viral meat colonization in commercial (Pekin) ducks challenged with HPAI H5N1. Vaccine 2007;25(20):4064–72.

3. van den Berg T, Lambrecht B, Marche´ S, Steensels M, Van Borm S, Bublot M. Inﬂuenza vaccines and vaccination strategies in birds. Comparative Immunology, Microbiology and Infectious Diseases 2007;31(2-3):121–65. doi:10.1016/ j.cimid.2007.07.004.

17. Kim JK, Seiler P, Forrest HL, Khalenkov AM, Franks J, Kumar M, et al. Pathogenicity and vaccine efﬁcacy of different clades of Asian H5N1 avian inﬂuenza A viruses in domestic ducks. Journal of Virology 2008;82(22):11374–82.

4. Swayne DE, Garcia M, Beck JR, Kinney N, Suarez DL. Protection against diverse highly pathogenic avian inﬂuenza viruses in chickens immunized with a recombinant fowl pox vaccine containing an H5 avian inﬂuenza hemagglutinin gene insert. Vaccine 2000;18:1088–95.

18. Ellis TM, Sims LD, Wong HKH, Bissett LA, Dyrting KC, Chow KW, et al. Evaluation of vaccination to support control of H5N1 avian inﬂuenza in Hong Kong. In: Schrijver RS, Koch G, editors. Wageningen UR Frontis Series: Avian Inﬂuenza, Prevention and Control. Springer, Dordrecht, London; 2005.

5. Swayne DE, Kapczinski D. Strategies and challenges for eliciting immunity against avian inﬂuenza virus in birds. Immunological Reviews 2008;225:314–31. 6. Swayne DE. Principles for vaccine protection in chickens and domestic waterfowl against avian inﬂuenza: emphasis on Asian H5N1 high pathogenicity avian inﬂuenza. Annals of the New York Academy of Sciences 2006;1081:174–81. 7. Inui K. Vaccination against HPAI H5N1 in Vietnam. Unpublished presentation at OFFLU Technical Activity on Vaccination in Southeast Asia. 1st meeting, Jakarta, 14 November 2008. 8. Swayne DE, Beck JR, Mickle TR. Efﬁcacy of recombinant fowl poxvirus vaccine in protecting chickens against a highly pathogenic Mexican-origin H5N2 avian inﬂuenza virus. Avian Diseases 1997;41(4):910–22. 9. Bublot M, Pritchard N, Cruz JS, Mickle TR, Selleck P, Swayne DE. Efﬁcacy of a fowlpox-vectored avian inﬂuenza H5 vaccine against Asian H5N1 highly pathogenic avian inﬂuenza virus challenge. Avian Diseases 2007;51(1 Suppl.):498–500. 10. Rudolf M, Poeppel M, Blohm U, Teifke J, Beer M, Harder TC. Use and efﬁcacy of inactivated AIV-H5 vaccines in the ﬁeld (industrialized holdings). Unpublished presentation at OFFLU Technical Activity on Vaccination in Southeast Asia. 1st meeting, Jakarta, 14 November 2008.

19. Ellis T, Leung C, Chow M, Bissett L, Wong W, Guan Y, et al. Vaccination of chickens against H5N1 avian inﬂuenza in the face of an outbreak interrupts virus transmission. Avian Pathology 2004;33(4):405–12. 20. Cardona CJ, Charlton BR, Woolcock PR. Persistence of immunity in commercial egg-laying hens following vaccination with a killed H6N2 avian inﬂuenza vaccine. Avian Diseases 2006;50(3):374–9. 21. van der Goot JA, Koch G, de Jong MCM, van Boven M. Quantiﬁcation of the effect of vaccination on transmission of avian inﬂuenza (H7N7) in chickens. Proceedings of the National Academy of Sciences, USA 2005;102(50):18141–6. 22. Qiao C, Tian G, Jiang Y, Li Y, Shi J, Yu K, et al. Vaccine developed for H5 highly pathogenic avian inﬂuenza in China. Annals of the New York Academy of Sciences 2006;1081:182–92. 23. Truscott J, Garske T, Chis-Ster I, Guitian J, Pfeiffer D, Snow L, et al. Control of a highly pathogenic H5N1 avian inﬂuenza outbreak in the GB poultry ﬂock. Proceedings of the Biological Sciences 2007;274(1623):2287–95.

11. Rudolf M, Poeppel M, Froehlich A, Mettenleiter T, Beer M, Harder T. Efﬁcacy of a commercial inactivated H5 inﬂuenza vaccine against highly pathogenic avian inﬂuenza H5N1 in waterfowl evaluated under ﬁeld conditions. Revue Scientiﬁque et Technique (International Ofﬁce Epizootics) 2009;28(1): 275–91.

24. van Boven M, Boender G, Elbers A, et al. Epidemiologische consequenties van vaccinatie. Rapport voor het Ministerie van Landbouw, Natuur en Voedselkwaliteit. cited in English in: Vaccination of poultry against avian inﬂuenza: epidemiological rules of thumb and experimental quantiﬁcation of the effectiveness of vaccination. In: Proceedings of the Frontis Workshop on Avian Inﬂuenza: Prevention and Control Wageningen, The Netherlands, 13–15 October 2003. p. 85–92.

12. Chen H, Deng G, Li Z, Tian G, Li Y, Jiao P, Zhang L, Liu Z, Webster RG, Yu K. The evolution of H5N1 inﬂuenza viruses in ducks in southern China. Proceedings of the National Academy of Sciences, USA 2004;101:10452–7.

25. Amorij J-P, Huckriede A, Wilschut J, Frijlink HW, Hinrichs WLJ. Development of stable inﬂuenza vaccine powder formulations: challenges and possibilities. Expert Review. Pharmaceutical Research 2008;25(6):2008.

13. Gilbert MP, Chaitaweesub T, Parakamawongsa S, Premashthira T, Tiensin W, Kalpravidh H, et al. Free-grazing ducks and highly pathogenic avian inﬂuenza, Thailand. Emerging Infectious Diseases 2006;12:227–34. 14. Hulse-Post DJ, Sturm-Ramirez KM, Humberd J, Seiler P, Govorkova EA, Krauss S, et al. Role of domestic ducks in the propagation and biological evolution of highly pathogenic H5N1 inﬂuenza viruses in Asia. Proceedings of the National Academy of Sciences, USA 2005;102:10682–7. 15. Webster RG, Webby RJ, Hoffmann E, Rodenberg J, Kumar M, Chu H, et al. The immunogenicity and efﬁcacy against H5N1 challenge of reverse genetics-derived H5N3 inﬂuenza vaccine in ducks and chickens. Virology 2006;351(2):303–11. 16. Beato MS, Toffan A, De Nardi R, Cristalli A, Terregino C, Cattoli G, Capua I. A conventional, inactivated oil emulsion

g

26. CAST. Avian Inﬂuenza Vaccines: Focusing on H5N1 High Pathogenic Avian Inﬂuenza (HPAI). Special Publication No. 26. Council for Agricultural Science and Technology (CAST). Ames, IA; 2007. ISBN 1-887383-28-X. 2007. 27. WHO. Evolution of H5N1 avian inﬂuenza viruses in Asia. The World Health Organization (WHO) Global Inﬂuenza Program Surveillance Network. Emerging Infectious Diseases 2005;11(10):1515–21. 28. Lee CW, Senne DA, Suarez DL. Effect of vaccine use in the evolution of mexican lineage H5N2 avian inﬂuenza virus. Journal of Virology 2004;78:8372–8381. 29. Swayne DE, Suarez DL. Highly pathogenic avian inﬂuenza. Revue Scientiﬁque et Technique (International Ofﬁce Epizootics) 2000;19(2):463–82.

CAB International 2010

J. Hinrichs, J. Otte and J. Rushton 30. Escorcia M, Va´zquez L, Me´ndez ST, Rodri´guez-Ropo´n A, Lucio E, Nava GM. Avian inﬂuenza: genetic evolution under vaccination pressure. Virology Journal 2008;5:15. 31. Suarez DL, Lee CW, Swayne DE. Avian inﬂuenza vaccination in North America: strategies and difﬁculties. Developmental Biology (Basel) 2006;124:117–24. 32. OIE. Avian Inﬂuenza Vaccination. OIE Information Document, Verona Recommendations; 2007. Available from: URL: http:// www.oie.int/eng/info_ev/Other%20Files/A_Guidelines%20on %20AI%20vaccination.pdf 33. European Food Safety Authority (EFSA). Scientiﬁc opinion on ‘vaccination against H5 and H7 subtypes in domestic poultry and captive birds’. EFSA-Q-2006-30. The EFSA Journal 2007;489. 34. Normile D. Rinderpest; driven to extinction. Science 2008;319(5870):1606–9.

91

systems. Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources. In press. 47. Desvaux S, Ton VD. A General Review and a Description of the Poultry Production in Vietnam. PRISE, Agricultural Publishing House; 2008. Available from: URL: http// www.prise-pcp.org. 48. Seng S. Gender and socio-economic impacts of HPAI and its control. Rural livelihoods and bio-security of smallholder poultry producers and poultry value chain in Cambodia. Case study in 36 villages in 4 provinces. Consultancy report by Cambodian Center for Study and Development in Agriculture (CEDAC) for FAO. July 2007. 49. Songserm T, Jam-on R, Sae-Heng N, Meemak N, Hulse-Post DJ, Sturm-Ramirez KM, et al. Domestic ducks and H5N1 inﬂuenza epidemic, Thailand. Emerging Infectious Diseases 2006;12:575–81.

35. Melo EC, Lo´pez A. Control of foot and mouth disease: the experience of the Americas. Revue Scientiﬁque et Technique (International Ofﬁce Epizootics) 2002;21(3):695–8.

50. Otte J. The hen which lays the golden eggs – or why backyard poultry are so popular. FAO PPLPI: 2006. Available from: URL: http://www.fao.org/ag/AGAinfo/programmes/en/pplpi/ docarc/feature01_backyardpoultry.pdf

36. Tiensin T, Nielen M, Vernooij H, Songserm T, Kalpravidh W, Chotiprasatintara S, et al. Transmission of the highly pathogenic avian inﬂuenza virus H5N1 within ﬂocks during the 2004 epidemic in Thailand. Journal of Infectious Diseases 2007;196:1679–84.

51. Dessie T. Studies on village poultry production systems in the central highlands of Ethiopia [MSc Thesis]. Department of Animal Nutrition and Mangement, Swedish University of Agricultural Sciences, Uppsala, Sweden; 1996.

37. Bouma A, Claassen I, Natih K, Klinkenberg D, Donnelly CA, et al. Estimation of transmission parameters of H5N1 avian inﬂuenza virus in chickens. PLoS Pathogens 2009;5(1):e1000281. doi: 10.1371/journal.ppat.1000281.

52. Gunaratne SP, Chandrasiri ADN, Mangalika Hemalatha WAP, Roberts JA. Feed resource base for scavenging village chickens in Sri Lanka. Tropical Animal Health and Production 1993;25:249–257.

38. Savill N, Rose S, Keeling M, Woolhouse M. Silent spread of H5N1 in vaccinated poultry. Nature 2006;442:757.

53. Rushton J, Ngongi SN. Poultry, women and development: old ideas, new applications and the need for more research. World Animal Review 1998;91(2):43–9.

39. Lesnoff M, Peyre M, Duarte PC, Mariner JC. A simple model for simulating immunity rate dynamics in a tropical free range poultry population after avian inﬂuenza vaccination. Epidemiology and Infection 2009;137:1405–13. 40. Beach RH, Poulos C, Pattanayak SK. Farm economics of bird ﬂu. Canadian Journal of Agricultural Economics 2007;55:471–83. 41. Garske T, Clarke P, Ghani AC. The transmissibility of highly pathogenic inﬂuenza in commercial poultry in industrialized countries. PLoS ONE 2007;2(4):e349. doi:10.1371/journal. pone.0000348 42. Walker P, Cauchemez S, Metras R, Dung DH, Pfeiffer D, Ghani A. Modelling the temporal and spatial dynamics of the spread of HPAI H5N1 in Northern Viet Nam. HPAI Research Brief, No. 19. 2009. Available from: URL: http://www. hpai-research.net 43. Ward MP, Maftei D, Apostu C, Suru A. Estimation of the basic reproductive number (R0) for epidemic, highly pathogenic avian inﬂuenza subtype H5N1 spread. Epidemiology and Infection 2009;137:219–26. 44. Anderson RM, May RM. Infectious Diseases of Humans: Dynamics and Control. Oxford University Press, New York; 1992. 45. Alders RG, Bagnol B, Young MP, Ahlers C, Brum E, Rushton J. Challenges and constraints to vaccination in developing countries. Developmental Biology (Basel) 2007;130:73–82. 46. Rushton J, Viscarra RE, Taylor N, Hoffman I, Schwabenbauer K. Poultry sector development, highly pathogenic avian inﬂuenza and the smallholder production

g

54. Spradbrow PB. Newcastle disease in village chickens. Poultry Science Review 1993;5:57–96. 55. Taylor N. HPAI vaccination strategies in Indonesia – modelling of alternative strategies. Consultancy Report for FAO, Rome. July 2008. 56. Udo HMJ, Asgedom AH, Viets TC. Modelling the impact of interventions on the dynamics in village poultry systems. Agricultural Systems 2006;88:255–69. 57. Chen H. Avian inﬂuenza vaccination: the experience in China. Revue Scientiﬁque et Technique (International Ofﬁce Epizootics) 2009;28(1):267–74. 58. Wooldridge M. Risk modelling for vaccination: a risk assessment perspective. Developmental Biology (Basel) 2007;130:87–97. 59. Taylor N, McLeod A, Thuy NT, Stone M, Binh VT, Lan LTK, et al. Examining the options for a livestock disease-free zone in the Red River Delta of Vietnam. Strengthening of Veterinary Services in Vietnam (SVSV). ALA/96/20. Supported by the European Commission; 2003. 60. Chicamisse M, Harun M, Alders R, Young M. Evaluation of the cold chain encountered by ‘Wet’ I-2 Newcastle Disease Vaccine from the Vaccine Production Department to the Village Chicken in Mozambique. Paper presented at the International Conference on ‘Opportunities for village chickens to assist with poverty alleviation with special emphasis on the sustainable control of Newcastle disease’ hosted by the AusAID Southern Africa Newcastle Disease Control Project in collaboration with the FAO, Dar es Salaam, Tanzania, 5–7 October 2005.

CAB International 2010

92

Animal Science Reviews 2010

61. Cyranoski D. China steps up drive to vaccinate all domestic birds. Nature 2005;438:406. 62. MoA China. Poultry avian inﬂuenza vaccination in China. Ministry of Agriculture, PR China (MoA China); 2006. Available from: URL: http://www.agri.gov.cn/ztzl/gdztzl/ P020061023368529330005.pdf 63. FAO. Avian Inﬂuenza Disease Emergency (AIDE) news. Situation Update 57, 15 January 2009. 64. Smith GJD, Fan XH, Wang J, Li KS, Qin K, Zhang JX, et al. Emergence and predominance of an H5N1 inﬂuenza variant in China. Proceedings of the National Academy of Sciences, USA 2006;103(45):16936–41. 65. OIE. Animal Health Information, WAHID Interface. 2009. Available from: URL: http://www.oie.int/wahis/public.php (accessed June 2009). 66. Samaha H. Highly pathogenic avian inﬂuenza in Egypt. Presentation at the HPAI technical meeting at FAO Rome, 27–29 June 2007 by Professor Dr Hamed Samaha (CVO, GOVS); 2007. 67. FAO. Aide News No. 56. November 2008. 68. FAO. OFFLU FAO interim recommendations on poultry vaccination against HPAI in Indonesia – on behalf of OFFLU-March 2008. Outcomes of the 4th OFFLU technical review meeting on vaccination in Indonesia 12–13 November 2008; 2008.

78. Taylor N. An assessment of post-vaccination sero-monitoring and surveillance activities, and the data generated, following HPAI vaccination in Viet Nam (2005–2006). Technical report (1) for FAO. January 2007. 79. To TL, Bui QA, Dau NH, Hoang VN, Van DK, Taylor N, et al. Control of avian inﬂuenza: a vaccination approach in Viet Nam. Developmental Biology (Basel) 2007;130:159–60. 80. CIRAD. EPIAAF survey (EPidemiology of Avian Inﬂuenza in AFrica). Unpublished presentation at FAO; 2008. 81. Nguyen VL. Post-vaccination surveillance and monitoring for AI virus circulation in Vietnam. Epidemiology Division, Department of Animal Health. Presentation at the International Avian Inﬂuenza Research Workshop in Hanoi 16–18 June 2008. 82. Henning J, Wibawa H, Purnomo P, Henning K, Ha T, Long NT, et al. Maintenance and Transmission of HPAI virus in backyard and small-scale commercial poultry production systems in South-East Asia. Unpublished data; 2008. 83. MARD. Integrated National Operational Program for Avian and Human Inﬂuenza (OPI) 2006–2010. Ministry of Agriculture and Rural Development (MARD), Socialist Republic of Vietnam, May 2006. 84. McLeod A, Rushton J. Economics of animal vaccination. Revue Scientiﬁque et Technique (International Ofﬁce Epizootics) 2007;26(2):313–26.

69. Sawitri Siregar E, Darminto, Weaver J, Bouma A. The vaccination programme in Indonesia. Developmental Biology (Basel) 2007;130:151–158.

85. FAO. Data from a HPAI vaccination willingness to pay survey commissioned by FAO Viet Nam and conducted by Abt Associates Hanoi; 2008.

70. FAO. Vaccination and surveillance costs evaluation for operational research project in Indonesia. Draft Consultant Report by Mohammed Iqbal Rafani to FAO; 2009.

86. World Poultry. Egg costs remain complicated. World Poultry 2007;23(1):29.

71. McLaws M. Operational Research in Indonesia for more Effective Control of HPAI – Seromonitoring. Presentation at ILRI closing workshop Bandung, 1 December 2009.

87. Otte J, Hinrichs J, Rushton J, Roland-Holst D, Zilberman D. Impacts of avian inﬂuenza virus on animal production in developing countries. CAB Reviews: Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources 2008;3(080):1–18.

72. Naeem K, Siddique N, Ayaz M, Jalalee MA. Avian inﬂuenza in Pakistan: Outbreaks of low- and high-pathogenicity avian inﬂuenza in Pakistan during 2003–2006. Avian Diseases 2007;51:189–93. 73. Cristalli A. The Vietnamese vaccination campaign against H5N1 HPAI virus subtype. Internal Consultant Report for FAO; 2006. 74. MARD. Decision No: 3541 by Ministry for Agricultural and Rural Development, Viet Nam (MARD), November 2007. 75. Taylor N, Sims L. Avian inﬂuenza (AI) vaccination strategy – future directions. Unpublished consultancy report for FAO Viet Nam; 2007. 76. FAO. Aide News No. 50. February 2008. 77. McLeod A, Rushton J, Riviere-Cinnamond A, Brandenburg B, Hinrichs J, Loth L. Economic issues in vaccination against HPAI in developing countries. Developmental Biology (Basel) 2007;130:63–72.

g

88. Soares Magalhaes R, Pfeiffer D, Wieland B, Dung D, Otte J. Commune-level simulation model of HPAI H5N1 poultry infection and control in Viet Nam. FAO-PPLPI Research Report 06-07; 2006. 15pp. Available from: URL: http:// www.fao.org/ag/againfo/programmes/en/pplpi/docarc/ rep-hpai_modelupdate.pdf 89. Domenech J, Dauphin G, Rushton J, McGrane J, Lubroth J, Tripodi A, et al. Experiences with vaccination in countries endemically infected with highly pathogenic avian inﬂuenza: the Food and Agriculture Organization perspective. Revue Scientiﬁque et Technique (International Ofﬁce Epizootics) 2009;28(1):293–305. 90. Tuan NA. Avian inﬂuenza, poultry culling and support policy of the Viet Nam Government. In: McLeod A, Dolberg F, editors. Future of Poultry Farmers in Viet Nam after HPAI. FAO and MARD Workshop Held at Horison Hotel, Hanoi, 8–9 March 2007. 99 pp.

CAB International 2010

Animal Science Reviews 2010

Accelerated lambing: Part 1. Housing for sheep G. L. Hunter* Address: Previously of the Department of Agricultural Technical Services, Stellenbosch, South Africa *Correspondence: Email: [email protected] Received: Accepted:

31 July 2009 3 December 2009

Abstract This presentation has been written for those who may be considering practical intensiﬁcation of sheep production and need information about its implications, potential and hazards, as well as for those with interests in this ﬁeld of research. The limited literature on sheep housing and the management of a housed ﬂock are considered, together with the importance of proliﬁcacy and the suitability of a number of South African breeds to accelerated lambing. The productivity of a housed Ile de France ﬂock in which three lambings a year was an objective, and difﬁculties met during 14 years of the ﬂock’s existence, are presented. ‘Effective teaser rams’ (ETRs) that both induce oestrus and increase lambing rate by as much as 20% have been noted and it is suggested that these as well as vasectomized goats (Effective teaser goats (ETGs)) may proﬁtably be used for these purposes.

Housing for Sheep (for Frequent Lambing or for Other Reasons) Plans for Sheep Housing 1. Prior to the 1970s, photographic illustrations or designs of successful and extensive housing for experimental sheep were published only very occasionally (e.g. [1]) and although commercial housing systems were already in use in some areas (particularly in the USA), published details for the housing of commercial ﬂocks (as opposed to animals used for research) prior to 1964 are surprisingly scarce, even for very cold climates. Fell [2] included photographs of such a construction and (in the same journal) Williams [3] offered useful insights into the management of housed sheep, presumably based on considerable practical experience, but no details of the housing are recorded in the paper. The use of housing has sometimes been indicated in research papers, but no details provided. Vega [4], for example, recorded results from an attempt to achieve two lambings per annum in Aragon sheep (in Spain): two groups of 50 and 46 ewes were kept indoors with rams, for two years (except for the period November 1963 to March 1964 for one of the groups), but nothing was noted in the abstract regarding

g

the housing. A review of ‘housing and its effects on farm animals’ by Baxter [5] recorded some very early attempts to eliminate the use of ‘litter’ (bedding) in animal houses (but does not mention sheep), the development of systems aimed at increasing the efﬁciency of space utilization, and also considered problems associated with increased stocking density. 2. A CAB Abstracts literature search of the period from 1973 to 2008, using ‘Sheep’ and ‘Housing’ as key words, produced (inter alia) the following 15 titles (presented in chronological order), with abstracts that are either summarized below or included verbatim: to begin with, an apparently dogmatic conclusion by Hume [6] stated that ‘although there had been a revival of interest in sheep housing in Scotland . . ., it is shown that there is no possibility of an improvement in proﬁtability resulting from the housing of sheep’ (from the CABI abstract – the original paper has not been seen, so the author’s justiﬁcation for this statement is not known. It may be speculated, perhaps, that something like the following should have been added to the abstract: ‘unless the feed can be home produced and/or if there is also an adequate increase in the productivity of and income from the ﬂock’ – see paras 25 and 40 for further discussion). Loboda’s paper [7] deals with a potentially useful topic

CAB International 2010

94

Animal Science Reviews 2010

(‘production methods used’), but unfortunately only the title and no abstract was presented in the CABI list. (The reference is listed for those who are able to access the paper and can read Russian.) Robertson [8] considered, in the light of new techniques that had become available, that temporary housing of ewes on low-ground farms (presumably in Britain, although this was not stated in the abstract) had become justiﬁable. For example, the control of lambing by synchronizing oestrus can reduce the accommodation required – batches of lambs born 7–10 days apart can use (in succession) the same housing. Floor space, troughs, ﬂooring, ventilation and layout are also discussed – adapted from CABI Abstract. This method for part-time housing during lambing (and perhaps at other stages during production) bears consideration, as a ﬁrst stage in the development of a permanent housing scheme. 3. de Montis [9] – The suitability of earth, concrete, wood or metal for ﬂoors (including slatted) is discussed and ‘the concrete ﬂoor with slatted elements was favoured’. In a second paper by the same author, given at the same seminar in Aberdeen, Scotland, de Montis [10] was concerned mainly with the important topics of building materials, techniques and costs, rather than with designs. Linklater and Watson [11], in a useful paper, illustrate and brieﬂy discuss three designs for sheep buildings, including the associated handling facilities and the importance of ventilation, ﬂooring and bedding as well as the foot ailments that occurred, feed troughs and the supply of water. Other topics discussed included numbers per pen, conditions at lambing, the need for shearing housed sheep and a ‘Reading List’ is appended. 4. Westendorp [12] – Housing types described include stall with litter, the same but with an attached uncovered area, and stall with slatted ﬂoor. Different feeding and drinking systems are assessed and other equipment and aids described, including one for hoof treatment. Burgkart [13] – Considers stall hygiene, climate and ﬂoorings, with a preference for ‘perforated’ over slatted ﬂoors, and recommends the space ‘requirement for feeding, equipment and exercise areas’ (CABI abstract). Hujnak [14] – An adaptable, easily erected and cost-effective, prefabricated sheep housing system is described, consisting of steel frame supports, wood panels and corrugated roof elements, the wall and roof panels allow both side and roof ventilation, and various designs are presented diagrammatically. 5. Moreau, [15] – ‘The use (as sheep housing) of plastic tunnels, of the type originally designed for market gardening, is discussed’. They are ‘easy to construct and remove, durable, resistant to weather, cheap and environmentally acceptable’ (although, as pointed out by E.M. Hughes (personal communication, 2009), their suitability should be queried, since the animals need shade and ventilation rather than the light and heat provided by plastic tunnels). Plans for the on-farm construction of housing from conventional materials, are also proposed’ (CABI abstract). Szalay [16]: ‘A number of sheep housing barns are described. Most are constructed of wood or a

g

combination of a steel frame and wooden cladding, often with a thermal insulation layer. In some, concrete ﬂoors are covered by suspended wooden slats for added insulation and easier cleaning. The barns are (large, being) designed to house around 500 ewes each; construction parameters of the individual types of barn are given together with some details of their interior arrangement’ (CABI abstract). Dornic [17]: ‘An assessment of construction trends in large capacity animal housing over the past 40 yr. is made. Schematic diagrams of typical contemporary cattle, sheep and pig housing units are presented, together with a summary of optimum temperature and humidity regimes in east Slovakia’s extreme climatic conditions. Mistakes frequently occurring in the choice of construction materials and overall design are discussed’ (CABI abstract). 6. Slade [18] – ‘The justiﬁcation for housing sheep in the winter in the United Kingdom (UK) is examined. The basic requirements of sheep housing in terms of location, ventilation, space requirements, manure handling, water supply, and safety and access are outlined. Feeding arrangements for housed sheep, and housing structures and ﬂoors are discussed. The use of housing to improve the success of lambing, and the effects of shearing on winterhoused sheep are considered. Finally, the disease risks of intensive housing are touched upon’ – CABI abstract. Berge [19] concluded that (in cold climates) sheep sheds need no insulation, but also discussed are manure storage and handling, types of ﬂooring, feeding and watering systems, space requirements, group sizes, ventilation and lighting, lambing season requirements, and design examples are given. Finally, Romero [20] described in Spanish (with special reference to conditions in Spain) different types of housing suitable for sheep under extensive or intensive management, taking account of ﬂock size, whether for milk or meat production, reproductive cycle, geographical location, climate and the labour available. 7. A small number of additional papers listed in the CABI literature search were concerned with the ventilation of, or hygiene in, rather than with the design of buildings for sheep. 8. A paper not identiﬁed by the CABI literature search, by Bøe et al. [21], in which the housing of sheep in Norway was reported, commences with the observation that ‘Norwegian winters can often be cold, . . . with much snow. Consequently, sheep, like all other production animals, are kept indoors during the winter period, usually from October to May (depending on location within Norway)’. This paper is concerned with the effect of insulation of the building (when ‘the lowest average monthly temperature was 713 C) on food intake, performance and body temperature of sheep over a four-year period, and it was concluded ‘that uninsulated buildings are well-suited for sheep’. 9. To the above, the following notes are added (after 14 years experience with a housed ﬂock in South Africa, and for the beneﬁt of any who may, for similar conditions, be seeking practical guidance on the construction of

CAB International 2010

G. L. Hunter

low-cost, but effective sheep housing). To begin with, for the breeding ﬂock, two simple structures were erected, each about 3010 m, consisting of a frame of creosotetreated wooden poles that were concreted vertically into the ground to support a roof of corrugated iron sheets, with sufﬁcient slope for disposal of rain water. This was only about 2.5 m above ground level; the buildings were clad on three sides with similar sheets and proved to be suitable for sheep housing in the mild winters and even during the hot summers of the Mediterranean climate of the South Western Cape Province. It should be noted, however, that on the eastern (long) side of the ﬁrst building, existing tall trees provided good shade during the mornings, and fast growing, American Ash trees were planted that soon provided good summer shade on the west sides of the buildings. Later, however, when a third building was required for the housing of weaned lambs, steel supports (instead of treated poles) were used to support a roof that was some 3–4 m above the ground; this greatly improved the ventilation from the open side of the building. The fourth (long) side of all three buildings, adjacent to the gravelled feeding passage that ran the length of the structures along their eastern side, was enclosed with vertical, spaced wooden slats that extended from ﬂoor to roof, which also improved the ventilation. In addition, in the ﬁrst two buildings that were erected (in which the roof sheets were only about 2.5 m above ground level) a heat-reﬂecting insulating material (‘sisalation’) was attached beneath the purloins that supported the iron roof. This was unnecessary in the third building, because of the roof’s greater height above ground. 10. Inside each building, wooden hurdles divided the area into three or more separate pens for groups, so that each pen could be provided with feeding troughs on two sides; this enabled 30 or more animals in a pen to feed simultaneously. There must be adequate head space at the feed troughs to allow all animals simultaneously to compete for feed (and allowance must be made for sufﬁcient space when the animals are carrying long ﬂeeces). Two metres of head space at the troughs was adequate for ﬁve ewes with long ﬂeeces. These considerations are regarded as more important than the recommendation of a minimum ﬂoor area per animal. But, as a guide, a pen of 68 m can hold about 30–35 ewes and their lambs, with sufﬁcient head space for all if feed troughs are provided on two sides of the pen. Furthermore, as already indicated, it is recommended that the interior of the building be designed so that fodder can be provided from feeding passages, rather than to have to enter the pen to do so, as Fell’s illustrations suggest [2]. 11. In addition, inside one of the buildings, 12 small pens were constructed for use by individual ewes and their new-born lambs. Each pen, of course, had its own water trough, with a ﬂoatation valve for controlling the supply of drinking water. Fluorescent lights were installed in the buildings for adults, for convenience when night shepherding was necessary during lambing, and could have

g

95

been used (by means of time switches) to enable ‘day length’ to be altered as a ‘treatment’ to promote frequent lambing (see discussion in part 2, paras 62–65); this was not, however, employed for this ﬂock. 12. The crowded and sometimes rather dusty conditions in sheep houses (especially those with bedding material on an earth ﬂoor) are not ideal for top-quality wool production. The ﬂoor of an enclosed sheep house may be slatted (at a price, and as long as adequate access for the easy removal of accumulated manure is provided). An earth or solid ﬂoor worked well in the three buildings described above, as no animals from other ﬂocks (which might have been carriers of foot rot) were introduced, apart from one ram and a nanny goat. Floors do not need to be impervious, but should be well drained (and remain dry) and such a ﬂoor should be covered with adequate amounts of bedding, such as straw or other suitable material. Our ‘deep litter’ system was the basis for the production of excellent ‘weed-seed-free kraal manure’, which was popular with gardeners, although its removal and bagging was labour-intensive and would have been difﬁcult to mechanize. 13. In a warm climate, it cannot be over-emphasized that it is especially necessary for the roof of the sheep building to be at an adequate height above the ﬂoor, to ensure that there is sufﬁcient ventilation from the open side(s) of the building, and the planting of suitable shade trees adjacent to the building should be implemented if possible. 14. Facilities outside of an enclosed building, adjacent to (and part of) the facility (but unroofed) included ‘outside runs’ (and access to an adjacent dipping tank) that contained a suitably designed restraining or sorting facility (a crush or race: one per building) of adequate dimensions for the number of animals being housed; these are required for sorting groups, as well as for handling and treatment of the animals, such as inoculations, weighing, and feet trimming (the animals’ hooves will need attention from time to time, whatever type of ﬂoor is provided, even if the daily programme includes the use of an exercise regime – see para. 22). These ‘outside runs’ are also useful when the ﬂock is subjected to teasing by vasectomized rams to identify ewes in oestrus and which are to be separated for mating by a breeding ram, but their use may be restricted to ﬁne weather.

General Management of Housed Flock 15. Some advantages of housing a ﬂock. Housing may, of course, be temporary. The use of both controlled lighting (if required) and teaser rams to stimulate the onset of ovarian activity (both still to be discussed) are likely to be particularly effective in the more conﬁned space in buildings – provision of effective controlled lighting is hardly practicable in large grazing paddocks – and ‘hand’ (i.e. supervised) mating of the ewes may be tightly

CAB International 2010

96

Animal Science Reviews 2010

controlled (by using teaser rams to identify those in oestrus, followed by hand-mating by the stud ram). Recording and use of the mating dates will enable appropriate regrouping of ewes so that feeding levels may be adjusted according to the nutritional requirements of late pregnancy. Following the mating period, it can be helpful (as an alternative to pregnancy diagnosis) to continue the teasing beyond the intended mating period for about an additional 21=2 –3 weeks (16–19 days being the normal range of oestrous cycle lengths in many breeds – see [22], for example), in order to identify mated ewes that remain not pregnant – they usually continue to cycle after the intended breeding period has ended, or alternatively, non-pregnant ewes can be identiﬁed using ultrasound (or other) techniques. Recent references produced by a CAB Abstracts search for ‘pregnancy diagnosis in sheep’, all in English, include Johns [23], Gracia et al. [24], Stevens [25], Karen et al. [26], Verberckmoes et al. [27], Yotov [28] and Barr [29], and non-pregnant ewes should be transferred to the next group that are to be mated. This also enables the manager to ensure that these ‘skips’ will have another mating opportunity sooner than they would have in a conventional, ‘one-opportunity-a-year’ mating system, which of course helps to increase in the ﬂock the numbers of pregnancies and lambs born per annum. 16. Expected lambing dates of the pregnant ewes can be estimated (about 146 days after their mating dates, depending on breed – but see para. 27) to facilitate good feeding, management and shepherding practices especially before, during and after lambing. These include planning and execution of the necessary annual inoculation programme (para. 19), being on hand to assist lamb births as necessary, and to ensure that mothering (bonding of the ewe and her newly born lambs) is timeously achieved thereafter (temporary provision of single pens for ewes and their newborns constructed in the building in which lambing takes place, is recommended – para. 11); as well as lamb marking soon after birth (using numbered ear tags, tattoos and/or notches, or other suitable methods). Prompt marking and bonding ensure the correct identiﬁcation of the lambs born to each ewe and help to avoid the consequences of ‘lamb stealing’, which may occur in crowded conditions while the lambs are still young. Tail docking and castration (if necessary) are best performed early. 17. ‘Creeps’ for the supplementary feeding of suckling lambs should be erected inside the ewe pens. If feeding levels of both the lactating ewes and of the lambs (in their ‘creep’) are adequate, weaning of housed lambs can take place by six weeks of age. This enables the ewes to be remated promptly – limiting the length of lactation to less than the more usual 3 months is fundamental to a successful ‘frequent lambing programme’, but accommodation of the weaned young in a separate building is probably essential. There is also merit in separating (in different pens) lactating ewes with single lambs from ewes with multiples, which do better in smaller groups. This also

g

makes it possible to supply the ewes rearing multiples, as well as their lambs (in the creep), with more nutritious rations to reduce the weight differences between singles and multiples at weaning. 18. Instead of artiﬁcially rearing (bottle feeding?) one or more of the lambs born as triplets, or those born to good breeding ewes that have lost some udder function through mastitis, lambs can be fostered to other lactating females, including a ‘nanny goat’ (of a good milking strain such as a Saanen). This process is probably easier to manage if the animals are housed. A goat with sufﬁcient milk may foster two or three lambs in addition to her own progeny; they can be surprisingly amenable to the practice. A Saanen was acquired for this purpose for use in the Ile de France ﬂock to be described from para. 44; this had an unexpected consequence, to be described in para. 46. The mating of the goat should, of course, be ‘timed’ to ensure that she will be in milk when her services are required. (The possibility of teasing the breeding ﬂock using male goats as teasers will be suggested in para. 53 – the ‘nanny’ can also be a producer of such teasers.) 19. Diseases and Pests in Sheep Houses. There is at least one major beneﬁt which arises from the housing of sheep. Internal parasites in the animals are seldom if ever a problem, so that even occasional dosing was unnecessary. On the other hand, two problems arose in my own sheep houses which were unpleasant, but ultimately controlled, namely a severe ﬂea infestation and an invasion by rats! The former were eventually eliminated by scattering malathion powder onto the litter in the pens, and the rats succumbed to poisoning. Furthermore, there was no escaping a carefully planned inoculation programme, which is not only essential [against diseases such as (in no particular order) enterotoxaemia, pasteurella, tetanus, Brucellosis, Chlamydia, lamb dysentery, Staphylococcus aureus, Corynebacterium (now Arcanobacterium) pyogenes, and blue tongue (if in Africa)] but, in an accelerated lambing system, can be something of a challenge to implement, with ewes at different stages of production at any one time and with lambs being born during three different periods each year – each group of mated ewes and each group of contemporary lambs required its own inoculation schedule or timetable. Coccidiosis can be a problem in housed lambs; but was not, in my experience, presumably because conditions in the pens remained dry. Lacasta et al. [30] have reported that ‘Pneumonia is responsible for important economic losses in the sheep industry in Spain and many other sheep-rearing countries, being the main cause of lamb death’, while losses also resulted from condemnations in abattoirs and from treatment costs. Building design and wind direction (presumably when lambs were young and vulnerable) were implicated as causes of pneumonia. 20. Another hazard that should be noted is chronic copper poisoning, especially in early weaned, housed, fattening lambs [31, 32]. The former authors noted that ‘stress in adult sheep can precipitate the condition if

CAB International 2010

G. L. Hunter

copper levels are high’ and recommended that not more than 5–10 p.p.m. should be included in the diet of housed sheep. 21. Mastitis was mentioned in para. 18. Lactating ewes in houses may be at greater risk of mastitis than when grazing; the shepherd should be very vigilant for indications of possible udder infections. An opportunity for udder inspections at feeding time, when all heads are in the trough, should not be missed, and signs of hunger especially among the younger suckling lambs should not be ignored. Prompt treatment is essential; keep supplies of medication in stock for quick action. The responsible organism should, however, be identiﬁed (at the local Veterinary Diagnostic Laboratory) for effective treatment. 22. Exercising of Housed Sheep. Experience has shown that sheep kept mainly for meat production can be permanently housed and appear very amenable to full-time housing. The following paragraph summarizes experience gained and observations made during the permanent housing of a small breeding ﬂock of ewes and their lambs, for which a daily exercise routine was provided. This involved taking each housed group daily on a prepared (fenced) route round part of the perimeter of a sloping, 6 ha property, a distance of perhaps 1 km, of which about half was uphill and fairly steep, rising some 50 m in the ﬁrst 0.5 km. The daily run for a group took about 10–15 min, so the speed averaged about 4–5 km per hour, possibly faster when the border collie was the ‘chaser’ of a group. Although no studies of the beneﬁts of regular exercise to permanently housed sheep are known, exercise per se has been shown to beneﬁt the breeding ability of rams [33] and the concept probably ‘stands to reason’ (considering the current attitude of many people to their own beneﬁts from exercise), but its value for a housed, breeding ﬂock still needs testing. 23. Selection of Young Stock. With adequate provision of suitable rations, most of the lambs (of both sexes) can become productive members of the breeding ﬂock (or be sold pregnant) from about 10–12 months of age, since puberty occurs well before this age [34, 35]. The ultimate objective of the system to be described from para. 32 was the production by each ewe of twins every eight months. The progeny of breeding ewes that achieved this level of productivity were (in the Ile de France ﬂock) candidates for selection into the breeding ﬂock [36]. It would be useful to know whether there is a genetic correlation between proliﬁcacy and frequent lambing. Turner [37] has described ‘experiments in which selection (was successfully) practised for incidence of multiple births . . . for rams based on ranking coefﬁcients for their dams . . . using mean numbers of lambs born to ewes of each age in these groups. Ewes (were) rejected if they or their dams failed to . . . produce . . . one multiple birth in their ﬁrst three lambings, while (in the control group) rams (were) rejected if their dams produced any multiple births’. The cost of running a control group in order to assess progress in selection for proliﬁcacy would be high and

g

97

possibly prohibitive in some commercial situations, but without one, evidence of progress in the selection for proliﬁcacy will be circumstantial at best (see paras 28 and 55). 24. Shepherding. The ﬂock manager or shepherd of a housed ﬂock develops a great sense of ‘control’ over his or her charges; this is an assumed ‘service’ to any ﬂock, whether housed or not, and is usually not an activity that can be withheld without some dire consequences. One comes across the occasional ﬂock owner who claims that he prefers that his ewes do not produce twins (usually because multiple births are associated with greater lamb losses or underfed lambs); this is a very short-sighted view, as greater lamb production is basic to proﬁtability in most ﬂocks, whether housed or grazed: multiple births should rather be encouraged; ewes with multiple lambs should be fed accordingly and the services of a competent, well-trained shepherd provided. The need for good, conscientious shepherding is paramount for virtually any system of sheep production and there should be no excuses for not providing it; this service to the ﬂock is particularly vital if maximum productivity is to be realized (both Gunn [38] and Giles [39] refer to the important role of the shepherd), and selection of a suitable candidate as well as adequate training are essential. In fact, the availability of good, responsible shepherding can be regarded as vital, especially during lambing, to ensure that the ﬂock’s welfare is not neglected; Phillips’ comprehensive review of animal welfare considerations [40] does not refer to this need, presumably because shepherding is a function of management, rather than of the breeding programme. On the other hand, in total contrast to this view, a ‘News and Notes’ item in ABA, vol. 43, abstract 4952, records a reference (in New Zealand Journal of Agriculture, April 1975, p. 16) to ‘Successful lambing without shepherding’, where ‘a lambing rate of 103% was achieved in a ﬂock of 2500 New Zealand Romney ewes that was left unattended at lambing time; in the previous six seasons, with normal shepherding, it had averaged 102%. One of the reasons for withdrawing shepherding on this 1400-acre property of large, steep paddocks was the possibility that disturbance in routine handling causes more losses than shepherding prevents’. In contrast to this attitude, it could be argued that the removal of ‘shepherding services’ during lambing could result in unnecessary distress among ewes that need assistance during lambing. The implication of the use of the word ‘successful’ seems to have been that the lambing rate was as ‘good’ as it had been for six seasons, but is it acceptable ‘ethically’ to withhold ‘midwifery services’, and should this be considered by Jarvis et al. [41] to be another example of practices that should be ﬁrmly discouraged? 25. Cost of Housing. Finally, it should be added that the costs of adequate housing, ‘proper ‘in-house’ feeding’ and good shepherding can be high, and should be carefully estimated before launching into such a programme. Intensive production may currently be economically

CAB International 2010

98

Animal Science Reviews 2010

justiﬁed if the sale of breeding stock is a primary objective; the proﬁtability of intensiﬁcation may be questionable if the objective is to be limited to the sale of slaughter stock (see para. 2). Tempest [42] has produced a useful summary, not only of the management practices needed in a ﬂock that lambs frequently but also, brieﬂy, of the economic performance of such a ﬂock. ‘The ﬂock (in the study) produced extremely high margins relative to annual lambing lowland ﬂocks recorded by the (British) Meat and Livestock Commission [43]. The purchased forage costs included hay and an allowance for grazing leys and sugar beet tops. Thus high stocking rates can be practised and high gross margins per hectare achieved, but it is emphasized that purchasing forage is purchasing hectares’. Similar objective economic assessments are required for other situations elsewhere. It can be expected, of course, that harvesting a crop to be fed indoors will increase costs, compared with the situation where the ﬂock can graze it. 26. Nutrition of Housed Sheep. This topic has been left to the ‘end’, not because it is unimportant – to the contrary – but it is outside the intended scope of this review. Attention is drawn only to three important papers on this topic, the ﬁrst two by J.J. Robinson [44, 45], which appear to be comprehensive, although, being now more than 20 years old, new developments are likely to have been published subsequently. This paper includes the headings ‘Recommended nutrient allowances’, ‘Nutritional principles’, ‘Diets and feeding regimes for continuously housed sheep’, ‘Nutritional problems speciﬁc to housed sheep’, and 29 useful references. Robinson made what seems to be a helpful suggestion that a ‘mild degree of under-nutrition (allowing ewes steadily to lose up to 5% of their body weight . . . during the second and third month of pregnancy) enhances placental growth in the fourth and ﬁfth month of pregnancy, this being the period when the foetus achieves over 80% of its growth’ and would thus be of special beneﬁt to ewes carrying litters of lambs. One of the referees of this review kindly pointed out that ‘there is indeed substantial later literature on the adverse effect of excessive nutrient intake on placental weight’, and gave Wallace et al. [46] as an example. And Ørskov’s report should also be noted [47], viz. ‘Early weaning and fattening of lambs’. His discussion emphasizes feed processing, protein requirements and manipulation of body composition.

Proliﬁcacy in the Ewe and Choice of Breed for Accelerated Lambing 27. In [32], Robinson and Ørskov suggested that increasing the production by the ewe to four lambs per year (note: not per pregnancy) ‘could be considered a realistic practical target’ and would achieve 80% of the theoretical improvement (as calculated by others to whom they refer), ‘in which . . . carcass production per

g

unit of food energy input has been estimated for a range of values for each component’; these ‘calculations . . . involve extrapolation of existing information to levels of production not yet attained in practice . . .’. Other than the Finnish Landrace (Finnsheep), which, in the early 1970s, was not available to breeders in South Africa (as import permits were not issued by the state’s veterinary authorities) there were few breeds with proliﬁcacy at such a level. Owen [48] noted that, ‘under commercial conditions, the optimum level of reproduction is limited to about 23=4 lambs at any lambing and a mean ﬂock lambing interval of about 8 months’. He added that ‘breeding frequency is also limited in many circumstances by the length of post-partum anoestrus, the period following lambing (when) the ewe cannot be successfully remated’. In addition, incidentally, he also observed that the ‘Finnish breed (with its unusually high proliﬁcacy) has a very short gestation length of little more than 140 days’ – gestation in most mutton breeds averages about 146 days, but depending on the number of lambs being carried by the ewe, usually varies in my experience between 144 and 150 days, with multiple births being associated with shorter pregnancies. (Owen also added a summary of how ‘proliﬁcacy can be seen as the end result of the progressive erosion of potential litter size from the number of ova shed by the ewe at oestrus . . . The ﬁrst hurdle is fertilization, the second . . . during the hazardous free living and early implantation stage, the third, more recently suspected, may occur after implantation but still early in pregnancy, the fourth in late pregnancy, manifested in still births and abortions. Parturition itself is a ﬁfth traumatic hurdle which brings down many lambs, followed closely by the sixth stage of establishing a good mother-offspring relationship, essential for the young lamb. The seventh and ﬁnal hurdle comes in the growing period where some proportion of lambs may still succumb’.) 28. Since Donald and Read [49] drew attention to the potential of the Finnsheep in a cross-breeding programme for the production of highly proliﬁc females, its use has become widely accepted and the breed is currently used in a number of countries for cross-breeding to produce a proliﬁc cross-bred ewe that is expected to carry ‘litters of lambs’ (i.e. at least triplets at each lambing) to be reared for slaughter or for further breeding. However, it seems to be a matter of opinion whether ‘litters of lambs’ are desirable, since few ewes are capable of rearing successfully more than two lambs at a time: and if they are required to do so, the shepherding would have to be outstanding. Is it not preferable or at least the alternative may be considered that ewes may be placed on accelerated lambing, with the aim of producing and rearing two of ‘Owen’s 23=4 lambs’ every eight months (see previous para.), thus coming close to meeting the requirements suggested at the start of para. 27)? Is this not an alternative to producing annually a crop of lambs for artiﬁcial rearing’? Is there not, in other words, at least some merit in

CAB International 2010

G. L. Hunter

ewes lambing at eight-month intervals, producing two lambs per pregnancy for rearing by their own mothers, rather than by cross-breeding with Finnsheep and being faced with the artiﬁcial rearing of lambs? Donald and Read [49] have suggested that as heritable differences between breeds ‘seem to be small’ (about 0.15 at one lambing), the breeding of proliﬁc cross-bred ewes to produce lambs for slaughter, may be a quicker path to follow than attempting to improve proliﬁcacy by selection within breeds. This view is now widely accepted. Hogue et al. [50] reported two schemes for rearing lambs from large litters (unfortunately, the lambs born per ewe are not given in the abstract). First, the ewes were allowed to (rear as many lambs as possible) and second, the lambs were weaned at birth, presumably reared artiﬁcially and fed on dry diets from about 10 days of age. The feeding of the ewes is described and they weighed the same at breeding, on 1 day after lambing and at 30 days after weaning. Results for ten ewes which reared all lambs born showed that the additional energy required to rear 3–4 lambs, instead of 2–3, may be similar to the additional 13–15% increase recommended by (American) NRC (standards) for ewes suckling twins compared with a single lamb. In the second scheme, the early weaning trial, with 24 ewes, Suffolk Finnsheep lambs in four groups were reared (1) as singles on the ewe to eight weeks, (2) reared from birth to 21 days on liquid milk replacer and then weaned to a dry diet, (3) reared from birth to ten days on milk replacer, drenched with ‘liquid starting feed’ and weaned to dry diet, or (4) with similar treatment except that the drench was of rumen contents from a wether on the starting feed. All lambs grew well and, of those weaned early, those given rumen contents gained most, 0.6 lb (32.4 g?) daily from 20 to 100 days of age. So these authors have suggested some interesting possibilities for rearing many lambs. However, I am inclined to suggest instead that there is an alternative to consider: viz. accelerated lambing, with a slightly lower litter size, which has the substantial practical advantage that all lambs can be reared by their mothers. The 1967 suggestion of Donald and Read should also perhaps be ‘revisited’, in view of Turner’s conclusions [37] from her work with the Merino that selection for improved proliﬁcacy may be achievable reasonably quickly after all (see para. 23). Ultimately, each producer has the choice: either to produce ‘litters for artiﬁcial rearing’, or to increase lambing frequency. Ultimately the intention of this review is to demonstrate how the alternative of increasing lambing frequency may be implemented. The rearing of twins by each ewe at eightmonth intervals may in fact be achievable more quickly than the development of proliﬁc strains by crossing with the Finnsheep: simply read para. 60, and start using teaser goats! Each producer may decide for themselves which approach to adopt: both are entirely feasible. The choice will in the end be based on the ﬂock owner’s preferences, either to increase productivity by means of more lambs per pregnancy or by increasing the frequency of lambing.

g

99

Each approach has its advantages and its disadvantages and its own set of problems for solving. The approach is entirely in the hands of the ﬂock owner or its manager, who should decide at an early stage which production system is preferable: more frequent pregnancies or more lambs per pregnancy (see paras 27–29). 29. Donald and Read reported that of the ﬁve Finnish ram lambs acquired by the Animal Breeding Research Organisation in Edinburgh, Scotland to start this project in 1962, one had been born in a litter of seven and another in a litter of six and that, after four years the little group of imported Finnsheep had produced ‘average litter sizes (at birth) of 2.0, 3.0 and 3.4 for ewes aged one, two and three or more years respectively’. Survival rates and birth and weaning weights are tabulated and losses seemed disappointingly high. The sheep were housed from October when cold weather would be starting, were mated and lambed indoors, and put out to graze in May (late spring in Scotland). At weaning (age not mentioned, but perhaps at about 3 months), lambs reared singly by one-year-old and two-year-old ewes weighed 51 and 62 lb (23 and 28 kg), respectively, and lambs in litters of two or more weighed 43 and 64 lb (19.5 and 29.1 kg), respectively, while older ewes weaned lambs averaging 49 lb (22.3 kg) – it is not clear how many lambs were reared by the ewes themselves – this may or may not, however, present a serious problem regarding the use of the breed. Since Finnsheep are to be crossed in a breeding programme with other breeds, presumably fewer lambs would be expected per pregnancy (from the cross-bred ewes) and it is possible that many if not most of the lambs can be reared by the cross-bred mothers. In a commercial situation, few ewes will successfully rear more than two lambs at a time, unless the shepherding is outstanding, so if the mean litter size does exceed two, artiﬁcial rearing of additional lambs may have to become ‘standard practice’. Walton and Robertson [51] reported from Canada that of Finnsheep ewes mated twice a year in autumn and spring, a third of the ewes conceived at each of the ﬁve breeding periods and 72.2% conceived at least four times. Mean lambing rate over the ﬁve consecutive breeding periods was, however, only 1.77 lambs per ewe exposed, equivalent to a mean lambing rate of 3.54 lambs per ewe exposed for a 12-month period. An important additional reason for not using the Finnsheep is that the cross-bred as well as the pure-bred Finn have some drawbacks as producers of top-quality lamb carcases [52] – the lambs tend to fatten early at relatively low carcase weights [53, 54], so carcase composition is less than ideal and carcase classiﬁcations are unlikely (in South Africa) to be graded as ‘super’. Is there not more merit in the development of recognized South African mutton breeds already with long breeding seasons (see paras 30–31) and potential for improved proliﬁcacy and which respond well in early summer to the introduction of teaser rams? Such breeds already exist in South Africa. These could meet the requirements of Robinson and Ørskov ([32] – para. 27), as many of them

CAB International 2010

100

Animal Science Reviews 2010

are already able to lamb every eight months. This lambing rate would be realistic in practice as the ewes can be expected (with good shepherding) to rear their lambs themselves. The successful development during the 1930s in South Africa (outlined above in paras 30–31) when new breeds with long breeding seasons were produced, was a ﬁrst step in the process; whether the additional development of a new, proliﬁc mutton breed from crosses with the Finnsheep is now also necessary may be considered, but existing breeds, already capable of mating from November to June (early summer to mid-winter), could also be selected for improved proliﬁcacy. Both options are possible. As will be apparent, a ﬂock can lamb at eightmonth intervals as quickly as the necessary management can be introduced. CABI has published an important review by Fahmy [55] on proliﬁc sheep, in which the main research workers in the ﬁeld of sheep proliﬁcacy have made important contributions including the role of breeds other than the Finn. The resources are available to maximize productivity in the sheep, the world-wide potential is phenomenal and the resources virtually limitless, sufﬁcient to meet different criteria and environmental constraints. I trust this review will restore some perspective to the potential of the ewe’s productivity.

South African Breeds suitable for Accelerated Lambing in Sheep Houses or on Grazing 30. Although the Merino ewe produces most of the sheep meat consumed in South Africa, it is not a ‘mutton breed’, and breeds (and cross-breds) that are better suited to lamb and mutton production have ﬂeeces that do not meet the high standards for wool set by the Merino. It must be appreciated that crossing the Merino to facilitate selection for proliﬁcacy would be detrimental to the quality of the wool and is unlikely to be contemplated in countries where the Merino wool industry is important – see reference to Turner in para. 23. So, during the 1930s in South Africa, other breeds were developed as specialist lamb and mutton producers, particularly from crosses with the Dorset Horn, reputed in its country of origin, the UK, also to have long breeding seasons (derived, it is said, from a belief that the Dorset may have some ‘Merino blood’, even though there are those in the UK who maintain that the Merino’s reputation in this regard is not justiﬁed – [56], for example). This longer breeding season was rightly thought to be an important trait for the new composite breeds in South Africa, as at that time, summer lambs were prone to internal parasites that were then more difﬁcult to control than they are today, so early lambs were preferred. This required mothers that were able to show oestrus well before midsummer and to lamb in autumn, which was achieved. Consequently this trait (which is also vital in a frequent- or accelerated-lambing ﬂock) is already a feature of some of our breeds.

g

31. The Dorper breed was one of the important and successful composites developed in South Africa during the early 1930s from crosses between the Dorset and the Black Headed Persian, for use in the warmer, drier areas such as the Karoo and the Eastern Cape Province [57], and another, the Dormer, was developed from about 1940 at the Elsenburg College of Agriculture, in the winter rainfall region near Stellenbosch in the Western Cape Province, from crosses between the Dorset and the then-named German (now South African) Mutton Merino [58]. (The purebred Dorset had been found to be susceptible in the Western Cape Province, to infestations by lungworm (Muellerius sp.), but the cross-bred was apparently less susceptible – [59].) It can be expected that both of these breeds (as well as the South African Mutton Merino itself, and yet another South African (Merino) derivative of merit, the Do¨hne Merino), if similarly managed, are already quite capable of producing lambs three times in two years, and in many ﬂocks also with a good twinning rate, although further improvement in this regard remains desirable in some; Basson et al. [60] demonstrated 40 years ago that these breeds are able to do so if adequately fed. This was conﬁrmed for the Do¨hne Merino in a short communication by Karberg and Fourie [61], as well as by Schoeman [62]. Following these successes, the further development of these breeds, already with long breeding seasons, but also with improved proliﬁcacy, may be achievable – with the Australian selection experiments (reviewed by Turner [37]) as examples (see para. 23). The assessment by Schoeman et al. [63] of Finn composites and their comparison with the Dorper, may be unnecessary and the alternative, selecting for proliﬁc strains by introducing highly proliﬁc breeds, should be reconsidered, at least in South Africa. More frequent production of twins may be preferable to the production of ‘litters of lambs’ requiring artiﬁcial rearing. The South African breeds mentioned above, and others, such as the Ile de France, already have the potential to produce twins or more at eight-month intervals.

Experience with ‘Accelerated Lambing’ in a Housed Flock 32. After reviewing this topic some 40 years ago [64, 65], and from research experience with teaser rams, as listed in the bibliography, I was convinced that with the ‘right’ breed of ewe, and the use of some of the techniques discussed in the original review (in particular, weaning the lambs by six weeks of age and using teaser rams to ‘stimulate’ the recommencement of oestrous cycles in the ewes), it seemed entirely feasible that many ewes could be induced to remate within a few weeks after lambing, instead of only annually. An opportunity to put this to the test in a privately owned ﬂock, was presented in about 1970 and the experience and results thus gained will be described brieﬂy in the following paragraphs to give

CAB International 2010

G. L. Hunter

some practical perspective to the review that is to follow in Part 2. 33. Selection of Breed for Housed ﬂock. The Ile de France breed was selected for this project. The objective of the originator of the Ile de France, a veterinarian named Auguste Yvart, was said to be ‘to produce a Dishley in the ﬂeece of a Merino’, at a time when the price of wool was experiencing one of its periodic downturns. In 1824, he crossed two breeds, the ‘Dishley Leicester’ and the Spanish Merino, a ﬂock of which had been established at Rambouillet, near Paris, France, a present from the Spanish monarch. 34. It is possible, however, that a more important contribution to the new cross-bred than its wool quality (and one that may not even have been appreciated at the time), could have been the Merino’s extended breeding season (in spite of dissenting views such as that already noted in para. 30), especially as the breed in France today is required to produce lamb (aged about 3–4 months) for both the ‘Christmas’ and ‘Easter markets’; implying that the ewes are capable of lambing in autumn as well as spring. Moreover, the Ile de France breeders had (by the 1970s) developed a sophisticated breeding programme to promote the traits desirable in the marketing of slaughter lambs, having established the testing facility at Verdilly, although the breed’s proliﬁcacy was not then considered to be exceptional. 35. The Dishley Leicester had been developed by Robert Bakewell, a renowned eighteenth century English breeder, who was reputed to use selection methods that included ‘progeny tests’ to identify the animals that improved the traits in which he was interested. A biography of Bakewell was published [66] and reviewed in Animal Breeding Abstracts [67], and includes material that recorded his breeding and selection methods. The reviewer noted, ‘Bakewell’s success can . . . be attributed . . . to methods . . . which have been used by successful breeders since his day: (1) Selection based on progeny tests and governed by exceptional skill in the assessment of merit. (2) Readiness to breed from closely related animals if other superior stock could not be procured’. It is regrettable that the Biography is apparently not still available. Attempts to access more information from the Leicester Sheep Breeders’ Society have been unsuccessful. 36. A small ﬂock of 30 young, mated Ile de France ewes and two rams were selected in France and imported to the Cape in 1974 (to be managed by the author) with ‘frequent lambing’ as a major objective, in addition to the sale of breeding stock to generate income. To keep inbreeding at bay (see ABA review by Lamberson and Thomas [68]), the imported females were allotted to three unrelated groups of ewes, each with an unrelated ram, a third ram being selected from those sired in France and raised on the little ‘farm’ of 6.3 ha., referred to earlier (from para. 9). Thereafter, all females born in the ﬂock remained for breeding in the group in which they were born, while rams selected for breeding were used only in

g

101

the ‘next’ group on unrelated females. Towards the end of the ﬂock’s existence it became necessary to purchase an unrelated ram for use in one breeding group. His dam, for some years, produced only twins. 37. Being in a district with a Mediterranean climate and winter rainfall (Stellenbosch, in the Western Cape Province, some 60 km from Cape Town), annual rye grass/ clover pastures were established to provide some grazing during the cool winters and in spring, the annual germination and growth of which was determined by the rainfall in autumn and winter. Only limited use of this grazing could be made when it was available, as the number of animals in the ﬂock was soon substantially greater than could be supported by the grazing, so most of the stock remained conﬁned to the buildings, apart from their daily exercise (para. 22). The size of the property was altogether too small for the production of sufﬁcient fodder for the ﬂock, and most had to be purchased (this being the reason for housing the ﬂock) together with wheat straw for use as bedding in the pens. Ultimately, the cost of purchased feed (as anticipated) became the ‘Achilles’ heel’ of the project; the breed became well established, the sale of surplus breeding stock became more competitive and prices started to drop. 38. It may also be mentioned that the ultimate, longterm plan on the small property, had been to produce hydroponically, during the hot, dry summers, a suitable fodder crop such as oats or barley (see [69], for recommended procedures), using water which was to be channelled from an existing state storage dam, but the frequent lambing project had to be terminated (after 14 years) before the water became available. 39. As mentioned in para. 27, the Finnsheep was not available at that time in South Africa, but a number of South African mutton breeds were likely also to be suited to ‘frequent lambing’, although none are as proliﬁc as the Finnsheep and its crosses. The choice of the Ile de France for the planned ‘frequent lambing ﬂock’ was made at least partly because of the expectation that the ultimate sale of surplus breeding stock to other breeders (registered with the Breed Society) would be easier because of less competition from already well-established breeders: the Ile de France was not, at that time, a well-established breed in South Africa, but was believed (because of its origins) to have the potential to lamb in autumn (because of its Merino parent and because of the continued interest in this trait by current breeders in France), and would display good mutton characters (because of the merits and contribution of the Dishley Leicester (paras 33–35). 40. The opinion expressed by Hume ([6], see paras 2 and 25) remains very relevant to the housing of sheep kept exclusively for meat production. It was probably valid in many areas apart from Scotland and was certainly valid in the 1970s in South Africa. It may not, however, be applicable to the economics of rearing stock in houses unless surplus animals can also be sold for breeding purposes for higher prices than if sold only as slaughter stock.

CAB International 2010

102

Animal Science Reviews 2010

Intensiﬁcation of sheep production by ‘frequent lambing’ is likely to become important in the future as a means of increasing the productivity of sheep – huge potential in this regard has been demonstrated by Robinson and Ørskov ([32] – see para. 27), although it should be borne in mind that their use of artiﬁcial lighting to extend the breeding season is feasible only with housed ewes – in most large ﬂocks frequent lambing will, in practice, depend to a large degree on the successful stimulation of breeding activity by the use of teaser rams while the ﬂock is grazing, rather than by changing the photoperiod with artiﬁcial light, with or preferably without the use of progestagen-impregnated sponges and other hormones. It is one thing to use hormone and light treatments in experimental ﬂocks or even in small-scale farm ﬂocks, but under the truly extensive conditions in regions of Australia, South America or Southern Africa, the scale of the operation may differ substantially from those elsewhere. At present in South Africa, the cost of purchased feed for housed animals is certainly still too high for such a project to be economically viable, unless the surplus stock can be sold as breeding animals rather than for slaughter (and if they are cross-breds, this may be difﬁcult), or if housed animals can be fed only home-grown, conserved fodder. In areas with severe climates, housing for sheep may be a necessary provision, but even then the feed required during housing should, purely for economic reasons, be home-grown and conserved for use indoors. Furthermore, the progesterone-impregnated sponges and gonadotrophic and other hormones sometimes used to synchronize breeding cycles or to promote multiple births, should also be costed, and accurate prices of these drugs must be included in any economic assessment of the treatments. Research papers commonly acknowledge the donation by the manufacturers of such items, but the price that would be charged for their commercial use is not widely known. 41. Adaptation and Productivity of the Imported Animals. High productivity was anticipated from the small number of imported Ile de France animals and their descendants, although the proliﬁcacy of the chosen breed was somewhat limited (see para. 55). Marshall [70] ﬁrst recorded that sheep transferred from one hemisphere to another altered their breeding pattern to conform to the changed photoperiod. Marshall based his report to the Royal Society on information supplied by Dr L.L. Roux in South Africa, following the importation to the Cape Province (latitude not stated) of a ﬂock of 21 pregnant Southdown ewes (age was not recorded either, but they would have been young). These ewes lambed in 1933 during January and were remated in May 1933 (late autumn in South Africa) ‘so that (according to Marshall) the changeover to conform to the sexual season of the southern hemisphere was at the beginning very rapid’. Para. 42 and the Appendix Table record the ﬁrst lambings and remating of the 30 ewes imported from France in 1974. The actual mating dates of the Southdowns were not supplied by

g

Marshall, so it is not possible now to consider whether perhaps their mating in South Africa was the result of a ram effect, rather than of an adaptation to the changed photoperiod, to which attention was ﬁrst drawn by Underwood et al. only ten years later in 1944 [71]. In 1934, the Southdown ewes, according to Marshall, ‘came on heat in April, and in 1935 and 1936 they did so in March (autumn) which is nearly the reverse of the normal tupping (mating) time in England. It is thus seen that in just over two years they became almost if not quite adjusted to the South African seasons’. 42. Marshall’s further comments are relevant to thoughts to be considered in Part 2 (paras 63–65), where it will be suggested that the sheep’s responses to changes in the photoperiod may be more complex than so far appreciated. Marshall went on (in his report to the Royal Society) to describe two cases of Southdowns imported to the Argentine with consequences that were similar to the importations into South Africa. In addition, he mentioned the consequences of importing pregnant ewes into Australia, and referred to what he termed ‘the exceptional breeds, Dorset Horns and Merinos which can have lambs twice a year’. He added that ‘in Scotland the tupping time is later (Oct. to Dec.), especially in the Highlands, and speaking generally is more limited in duration the further north we go’. Adaptation to the changed photoperiod appeared to present no problem in the Ile de France animals that were shipped to their new home at 34 S, from their French homes at latitude 49 N; 17 of 30 imported ewes were pregnant on arrival; the rest conceived to fertile matings within a matter of months, following teasing by vasectomized rams. Travelling on the deck and being quarantined for 30 days at the docks before ﬁnal transport to the farm was also a period of adjustment to the change in photoperiod. A comparison of the effect of change in photoperiod on ewes from the far north of Scotland, with similar ewes from further south would be of interest (see para. 64). An important publication should be noted at this stage, viz. the proceedings of a workshop or symposium held in Edinburgh, Scotland and edited by Land and Robinson [72], on the genetics of reproduction in sheep. The authors present several reasons for the extensive study of the topic, including the fact that the recent discovery of the ‘single gene inheritance of the high litter size of the Booroola Merino in Australia, had opened the door to considerable exploitation by both breeders and scientists’. 43. The subsequent selection into the Ile de France ﬂock of young breeding stock (of both sexes) was based mainly on the lamb production of their dams, particularly during their ﬁrst two lambings, following ﬁrst mating by 12 months of age, remating at about 18 months, and having successfully reared lambs for 40 days – see Turner ([37] – para. 23) for a description of two Australian experiments on selection for twinning rate. By late 1976, there were 56 Ile de France breeding ewes in the three breeding groups, and stock of both sexes for sale – so at about this

CAB International 2010

G. L. Hunter

stage a third building had to be erected for young stock (para. 9). 44. The Appendix Table, described in paras 44–55, summarizes the performance of the Ile de France ﬂock – rather tedious ‘circumstantial evidence’ (for which I apologize – skip perhaps to para. 50?) of the level of ﬂock performance achieved in a frequent lambing system, in which teaser rams were with the ewes to be mated for two weeks prior to each mating period. The number of ewes that mated (column 3 of the table) is expressed as a percentage of those with a mating opportunity (after the second slash in column 1). In column 9, the lambing rate for each year is calculated and presented, and it will be seen that this was greater than 100% during the ﬁrst ﬁve years. Apart from 1974 (see para. 47), this was achieved as a direct result of the preliminary teasing periods – ewes that lamb once a year can never exceed 100% lambing rate, because there are typically a few ewes in the ﬂock that do not lamb, lambing rate is usually less than 100%. In column 1 of the Table, the number of ewes available for each mating period is calculated. The 1979 lambing rate (column 9) fell below 100% because of the sale of 11 pregnant ewes during the year; if these had lambed before being sold, the lambing rate for the year would have been 102%! In 1980, the ﬂock’s lambing rate fell to 75%, signiﬁcantly below 100% for the ﬁrst time, because of the failure of the November–December mating of 1979, when only two of 66 ewes were mated, one successfully. 45. Without a control group of unmated ewes, in which the occurrence of oestrus can be determined without introducing rams, and the introduction of rams is the only means of determining ‘standing oestrus’ and, furthermore, because the presence of the teasers actually causes the onset of oestrus, one has a real dilemma: vaginal smears have been used to determine ovarian cycles [73–75], but Ducker and Boyd [76] found them unreliable even for the detecting of ovulation – only 35% giving a correct diagnosis, so, clearly, their use for the determination of standing oestrus, would be of little value. 46. Because of the fact that in the ewe standing oestrus can be determined only by her response to the mating behaviour of a ram, it was something of a surprise to hear, early one morning, the very vocal sign (as it turned out) of standing oestrus in the Saanen goat (mentioned in para. 18). She was taken for mating to a nearby ‘cheese-fromgoat’s-milk-enterprise’, where another surprise was in store, the witnessing (for the ﬁrst time) of the strong sex drive of a male Saanen. This is not intended to be facetious, but this experience provided ‘food for thought’ when considering the events to be described below (especially paras 49–53), namely the failure of the teaser rams to repeat the successes of 1975–1978. 47. The Table records in column 1, for each year, the numbers of females available for breeding, and, in columns 3, 5 and 6, the numbers that mated and lambed and of lambs born and weaned. In the ﬁrst line of the Table, it

g

103

will be seen that, prior to shipping, an unknown number of the selected young ewes was mated in France during the autumn months of 1973, when they were 10–14 months of age, and also that 17 lambed during January–March 1974. A further six were mated in South Africa during February 1974 and lambed during June–July; 23 were remated during May–June, and 22 of those (more than half the ﬂock) lambed during October–November 1974; thus 45 lambings were recorded during 1974 by the 30 ewes in the ﬂock; this accounts for the high total lambing rate (150%) recorded in the last column – seemingly an auspicious start for the project, but the ﬁrst mating in France had taken place during the French autumn of 1973, so the ﬁrst lambing of 1974 was very early in the year, and, of course, this schedule could not subsequently be repeated. Furthermore, the 22 ewes that lambed during October– November 1974 had little prospect of being remated at the next mating period that took place during the same months in which they lambed – and while they were rearing lambs, so only four ewes mated during October– November 1974. During the next 12 months, the number of ewes in the ﬂock increased brieﬂy to 39 (column 1), was soon reduced to 38, and 39 lambings were recorded during 1975, so the lambing percentage for the second year (calculated as shown in column 8) was 103%. During October–November 1974 (column 2), there were only eight ewes to be mated (because 22 had lambed during the mating period – see column 1) and columns 3 and 5 show that four were served, but only three lambed during March–April 1975. 48. For the November–December mating in 1975, an asterisk in column 1 is used for the ﬁrst time. As the Table’s footnote explains, in all cases thus marked, the number of ewes mated (41, as shown in column 3) was greater than the number expected to mate (ewes in ﬂock less the number that lambed at the previous lambing), because some of the most recently lambed ewes were also mated. In this instance, 41 mated, although (44 less 12) were ‘due’ to be mated. The 68 lambings in 1976 were achieved by only 44 ewes in the ﬂock, giving a total lambing rate for 1976 of 154.5% (column 8). 49. During the November–December mating period in 1978, when there were 100 ewes in the ﬂock (column 1), the Table shows that 35 of them had recently lambed (see previous line, October–November, columns 4 and 5). As most of these would still be lactating during the mating period, only 65 were expected to mate (column 1); in the event, 43 were mated (column 3) which was 66.2% of those available for mating (the last value in column 1). From these matings, 35 lambed during April–May 1979 (column 5 again). Thus, from the values recorded in column 1, the numbers of non-pregnant ewes for each mating period may be calculated. 50. The main feature to be observed in the ﬂock records is the presumed effect of the presence of teaser rams for 14 days before each mating period, especially during the early summer (November–December

CAB International 2010

104

Animal Science Reviews 2010

mating) – ‘presumed’ as there was no ‘unteased’ group to serve as a ‘control’. On the other hand, the effect of the preliminary period of teasing could be measured by the overall lambing rate per ewe in the ﬂock during the course of each year, or the percentage of the ewes in the ﬂock that lambed in each calendar year, shown in column 9 of the Table. Flocks that lamb only once per year obviously cannot exceed 100%. But in 1979, for the ﬁrst time in six years, the total lambing rate fell below 100%, reﬂecting the sale of 11 pregnant ewes. Reverting to the mating period already presented, from 15 November to 24 December 1975 (a period of 40 days) that had followed 14 days of ‘teasing’, the Table shows that 41 of the total ﬂock of 44 ewes were mated, as a result of which 34 ewes lambed during April–May 1976 and 47 lambs were born. During about the same period in the following year (from 12 November to 17 December 1976), only 26 of the ewes mated, perhaps because 24 of them had lambed during the October–December period of 1976, compared with only 12 during October 1975. 51. It is important to consider two more years of farm records. From 14 November to 12 December 1977, the table shows that 44 ewes mated and 40 subsequently lambed; that during the next year, 1978, the ﬂock had increased to 100 ewes (the sale of ewes had until then been disappointingly slow); and the table also shows that during November and December 1978, 43 ewes were mated, with 35 of these lambing during April–May 1979. There was thus some small variation in the actual dates of mating in similarly designated periods that are not indicated in the table, but this is not regarded as of signiﬁcance. However, during the November–December mating of 1979 (from 12 November to 21 December), after some sales, and having had what were regarded during the previous four years as adequate responses to the use of teasers at that time of year, it was decided to treat the Ile de France ﬂock with progestagen-impregnated sponges. Now (almost 30 years later) the reasons for this decision can only be guessed – perhaps it was because there had been some encouraging results in experiments in which ewes had been synchronized by teasing plus vaginal sponge treatment [77–81]. But in this ﬂock, the treatment was a complete failure, with one ewe showing oestrus before treatment started, and another showing oestrus after treatment (both were mated, but the latter did not conceive). 52. The success of the early summer mating is vital in a frequent lambing programme, as the ewes that lamb as a result of this mating have an excellent chance of mating again during the subsequent autumn and early winter, when the days are becoming shorter, this latter period being regarded as the ‘natural’ breeding season. So, in the following year, during November–December 1980, as the table shows, the previously apparently successful use of teaser rams for two weeks before mating was resumed. This time, however, the outcome was disastrous, as not a single ewe showed oestrus during the mating period.

g

However, there was also the fact (noted in the record book) that the group of teaser rams (by then increased by new vasectomies) was noticeably reluctant to ‘work’: in the 1980 ﬂock mating records, it was noted that prior to the ‘disastrous’ mating, for reasons that were not apparent, on introduction into the ewe pens, the teaser rams showed comparatively little interest in the ewes. Was the serious change of routine during the previous year, a possible cause of this change in behaviour, was there a reason for the ewes to be unattractive, or were the rams uncharacteristically lazy? 53. This ‘development’ needs consideration. There was little that could be done at the time to ‘activate’ (i.e. to improve or increase the mating efforts or sexual drive of) the teasers. It was considered that the use of more than one teaser per group to provide competition for the rams was self-defeating – competition between the rams might divert their attention altogether from the ewes. The rams were fed before being introduced to a pen of ewes, to prevent them from being distracted by hunger from paying attention to their task, and the teasers in each group of ewes were changed twice in 24 h. Fresh drinking water was always available. Some of the teasers were no longer youthful, but younger, newly vasectomized rams (as mentioned in the previous paragraph) was available to augment the group. In the end, it was conceded that, on this occasion, the teaser rams were simply ineffective – and the Table conﬁrms that in November–December 1980, the system that hitherto had appeared to be effective, had failed. Furthermore, it failed again in November– December 1982 and in 1984, presumably for similar reasons: the apparent sloth of the teaser rams. The comparatively sensational activity of the Saanen male (para. 46) was remembered, and the question asked – would there be merit in future, in using teaser goats to stimulate oestrous activity in the Ile de France ewes? On the other hand, Gordon ([82], on p. 148, quoting MacDonald [83]) recorded that in New Zealand, ‘ram introductions . . . particularly in the North Island (why there especially?), showed that there was no response among ewes to the ram effect in some seasons, yet a satisfactory response in others’. And on the following page, Gordon quotes Scott and Johnstone [84], who also recorded ‘variations between years in the ram effect when Coopworth or Dorset rams were introduced to seasonally anoestrous Coopworth ewes; these authors concluded that the variation in expression of the ram effect is probably governed by the depth of anoestrus of the ewes at the time of their exposure to rams’. Perhaps Scott and Johnstone should also have considered whether their teasers were sufﬁciently active and effective in all years, or whether variation in ‘depth of anoestrus’ is another challenging topic for investigation. 54. A consequence of these failures is reﬂected in the right-hand column of the Table, where the percentage of the breeding ewes that lambed in a calendar year is shown. In a ﬂock that lambs only once per annum, if, say

CAB International 2010

G. L. Hunter

95% of the ewes lamb, this could be regarded as normal and acceptable – inevitably each year, a few ewes will ‘skip’ lambing; the percentage of skips usually varies from year to year, but the numbers that lamb can never be more than 100% of the ﬂock. In the Ile de France ﬂock this percentage was less than 100% for the ﬁrst time in the sixth year of the project, 1979, when the ﬂock had been reduced by sales after the ﬁrst mating of that year, and it was even lower in 1980 (due to the failure of the mating in November–December 1979 – para. 51). In 1985 (because of the further failure of the mating in November– December 1984 – para. 52), the lambings per ewe were only 102%, because, although conception rate during the two matings was good, ﬂock size had again been reduced and there were only 73 lambings altogether, as well as small reductions in ﬂock size. 55. Proliﬁcacy in the growing Ile de France ﬂock was ‘promoted’ by placing emphasis on this trait when selecting rams as well as young ewes (see para. 23). This appeared to result in the required gradual increase, during the 14 years of the ﬂock’s existence, in the number of lambs born per lambing. In column 8 of the Appendix Table the weighted mean number of lambs born per lambing has been calculated for each year. These means improved during the ﬂock’s existence, from less than 1.2 lambs per ewe in 1974, when the foundation ewes were still young, to more than 1.6 lambs per ewe in 1983 and 1984. There was, however, an increase in age of the ewes, and no unselected control group with which to compare, so this trend should be viewed with caution, but it seemed at the time to be encouraging. More proliﬁc ewes would perhaps have performed better, and if the effect of teaser rams had been greater and more consistent, pregnancy rates could also have been higher.

Conclusions and Recommendations 56. A number of breeds in South Africa have been noted, in addition to the Ile de France, which display ‘long breeding seasons’ (para. 31); the incidence of oestrus in these ﬂocks increases gradually from November, when day length is still increasing and is promoted especially by the introduction of sexually active rams. If these are breeding rams (not teasers), it is advisable (in November, to achieve adequate conception rates) to allow mating to continue for eight weeks or longer (instead of only 5–6 weeks, which is usually sufﬁcient later in the breeding season), or alternatively to introduce active teaser rams for two weeks immediately prior to the introduction of the breeding rams, to stimulate or promote regular breeding cycles [85, 86]. Individual ewes of these breeds usually start their breeding cycles rather gradually and young ewes, in particular, often have short heat periods of lower intensity, so that their conception rates, especially at this time of year, are generally lower than those of more mature ewes. This can be signiﬁcantly improved by

g

105

running them in smaller groups from which mature ewes are excluded and using more than the usual (1 : 50) ram : ewe ratio – a ratio of 1 ram to 40 or less, or ‘hand mating’ may be used, although E.M. Hughes (personal communication, 2009) suggested that the latter practice (especially if used with hormonally synchronized oestrus) might mask cases of ‘sub-fertility’, which would be undesirable in any ﬂock; and that such young ewes should rather be identiﬁed and eliminated from the ﬂock. 57. Lamb production in ‘Cape ﬂocks’ (i.e. of the breeds noted in para. 31) may be increased by as much as 20% by mating later in the breeding season, during February– March (the southern hemisphere’s late summer and autumn), either instead of, or as well as during November–December. Following weaning (by six weeks) of lambs born during April–May and the use of teaser rams, a third mating period within 12 months can be scheduled in June. This programme is, in fact, the kernel of a frequent lambing system in the southern hemisphere. All these matings can be carried out in grazing paddocks – the use of buildings and artiﬁcial lighting (the latter to be discussed in Part 2) are unnecessary for frequent lambing in the Western Cape – on the other hand, if lambs are to be weaned by six weeks, they should be creep-fed before weaning and probably housed after weaning to ensure proper feeding. There may, however, be other reasons for housing a breeding ﬂock, and the early part of this review has summarized the needs and possibilities. With the lambs of each crop weaned on to suitable, economic rations by 6 weeks of age, by scheduling all three matings to take place between November and June (or early in July at the latest), and by allowing, as before, for some reluctance among the young ewes to be rebred so promptly, up to 50% of the ﬂock can be expected to lamb in each period (350=150 lambings per 100 ewes per annum), producing as many as 135 lambs per 100 lambings, depending on ewe proliﬁcacy. If this schedule is used each year, there can be three 5-week matings and lambings per annum, rather than two lambings in one year and one the next. This is ‘theoretically’ achievable in a frequent lambing programme. It should, however, be appreciated that the provision of adequate fodder for pregnant and lactating ewes during three different periods of the year might be possible only with irrigation during the dry summers – unless the spring growth can be harvested and conserved for use during the summer months. In the Mediterranean climate of the Western Cape, this is possibly a greater challenge than the induction of frequent lambing by the ewes. With three lambings each year, the facilities for the indoor rearing of early weaned lambs can be about half the size of those required for lambing three times in two years, and the facilities remain in use for 6–8 months instead of only 2–3 months per annum. The marketing of the lamb crop is thus spread over a longer period, and ewes which ‘skip’ do not have to wait 8 months to be remated, but only about 3–4 months: they join the group for the next scheduled mating period (see

CAB International 2010

106

Animal Science Reviews 2010

para. 15). In the southern hemisphere, as shown in the Appendix Table, the sequence of annual mating periods in the Ile de France ﬂock was February/March (late summer), May/June (early winter), and November/December (early summer), but the exact dates were not ‘set in stone’; they may vary somewhat (as long as mating in October by breeding rams is avoided – teasers may, however, be joined from mid-October). The sequence of the annual lambing periods was therefore April/May (autumn), June/ July (winter) and October/November (early summer) of each year. The schedules of two frequent lambing systems will be shown diagrammatically in Table 1 (in part 2), for the FLASH and STAR systems (see para. 62, and also para. 81 in part 2); some ﬂexibility in the mating period is acceptable in the FLASH system – an advantage in comparison with STAR. 58. For calculating lambings per ewe in the ﬂock, much depends on when and how ﬂock size and structure for the year are deﬁned; assessment of the performance of the Ile de France ﬂock was complicated by the sales of breeding ewes, as the consequent ﬂuctuations from year to year in the age structure of the ewes in the ﬂock affected the lambing opportunities – the prospect of lambing twice in a year was greatest for adult ewes that mated following teasing during early summer. At the bottom of the Appendix Table, 14-year totals are given and may be expressed as annual means, as follows: the number of ewes mated averaged 78 per annum, 90% of mated ewes lambed, lambs born averaged 151% of lambings, 86% of lambs were weaned and the mean lambing rate per ewe was 119.45%, including the years when both teasers and summer matings were ineffective (paras 48–52). This is a far cry from the potential of 150 lambings per 100 ewes suggested in para. 57. On the other hand, by using ‘really’ active ETRs that not only stimulate the onset of oestrus but also increase lamb production by as much as 20% (see end of para. 60), much potential remains for improving ﬂock performance. 59. Objective, long-term assessments are required of the effects of ETRs on mating behaviour in ewes, and the role of teasers should not be ignored in research on, or in practical attempts at, accelerated lambing. In spite of the apparent failures of the teaser rams in four years of the Ile de France project (1979, 1980, 1982 and 1984), to induce oestrus during the November–December matings, they seemed to have been successful in eight years (1975, 1976, 1977, 1978, 1981, 1983, 1985 and 1986), when the total lambing rate (column 8 in the table) exceeded 100%, although there was substantial variation between years. 60. There is a need to test the suggestion made in para. 53 that male goats might be very effective teasers (ETGs) in a frequent lambing programme for sheep. The males should, however, be vasectomized, as there is evidence that hybrid embryos can survive, though often not to term [87]. On the other hand, McGovern [88] records that on the Island of Malta hybrids not only exist, but are common enough to be named ‘shoats’, while in Chile they are

g

known as ‘chabins’ and in Jamaica as ‘sheep-royal’. It should also be noted, in a report by Sambraus [89], that virtually all female sheep (and goats), if given the choice, ‘approach the male of their own species (when in oestrus) much more frequently than the other male’. However, ‘because of their more intense sexual behaviour, male goats (bucks) were more attractive to females than rams of the other species’. There is also a need to consider factors affecting libido in rams, especially teasers. An example of research into the behaviour of rams was provided in an annual report of the New Zealand Ministry of Agriculture and Fisheries [90], in which, at the Ruakura Research Institute, the libido of ram lambs was tested and shown to be inﬂuenced by mode of rearing. Isolation of rams after weaning was found by Roux and Barnard [91, 92] to be detrimental to normal sexual behaviour; although Bryant [93] found no differences in the development of sexual behaviour when male lambs were reared in isolation or in a group. After Alison [94] had concluded that most rams are capable of mating with 200 ewes or more in a 17-day mating period, Alison and Davis [95] observed that although ‘ewe fertility was high when three breeding rams were joined with 100 ewes, increasing the number of ewes to 300 per three (rams) resulted in fewer ewes mated in the ﬁrst 17 days of mating, fewer rams mating each ewe, and an increase in barrenness’. Other possibilities that could be considered for the determination of oestrus in ewes are the use of testosterone-treated wethers [23, 96], or even of stilboestrol-treated ewes [96, 97]. The use of such ‘teasers’ might even be a method for the experimental determination of standing oestrus without causing a ‘ram effect’ or inﬂuencing the length of oestrus and time of ovulation, as was demonstrated by Parsons and Hunter [98] and Parsons et al. [99]. Furthermore, there is an important ability, reported by King and Coetzer1 [100], of ‘highly sexed’ Ronderib Afrikaner rams (that were compared with Merino rams), if allowed to mate throughout the oestrous periods, as they were reportedly able to do, of increasing the ovulation rate in ewes, resulting in a substantial 21% increase in lambing rate (over ﬁve years). This may be expected also of ETGs, but needs still to be tested. As already mentioned, Roux and Barnard have shown that the method of rearing young Karakul rams can inﬂuence their subsequent libido, so circumstances during rearing potential teaser rams are important. 61. The availability of fodder, as well as the costs of production and income generated should be considered in farm (and research) projects on frequent lambing (para. 25, and see [101] for an analysis of economic response to unit change in litter size). The proposed

1

See this and other publications on this topic in the Website of Grootfontein Agricultural College, Middelburg, Cape: http://gadi.agric.za.

CAB International 2010

G. L. Hunter

system should be subject to rigorous and ongoing economic scrutiny and include the capital costs of buildings, which will be essential if artiﬁcial lighting is to be used, as well as the costs of shepherding and management (see para. 78 in part 2). Furthermore, the rearing and sale of ‘hoggets’ as opposed to lamb may be encouraged by growing consumer resistance to the increasing price of lamb (already evident in South Africa, where Hirzel [102], over 35 years ago, suggested that hogget should be marketed instead of lamb – the term ‘hogget’ for a young animal, 10–14 months of age, is used in the UK and in New Zealand for a ‘yearling’ slaughter or breeding animal that is no longer a lamb, but not yet an adult. The term is not widely known in South Africa, but may come into more general use if the needs arise to increase the age at slaughter of young stock, as well as to decrease the age at which female sheep are ﬁrst bred.) – hoggets could be brought to slaughter more slowly and hopefully more cheaply than lamb, but must be marketed before the eruption of permanent incisors, to avoid discrimination during carcase grading. With the prospect of rearing young slaughter stock in future with less grain feeding [103], the marketing of hoggets for slaughter in place of lamb may perhaps be expected to become widespread practice, with the price of meat from lambs becoming increasingly prohibitive for many consumers. In her ABA review, Turner [37] noted that Australian ‘lamb must come from sheep which have not cut a permanent incisor tooth and so covers animals up to about 15 months of

g

107

age. Animals aimed speciﬁcally for the prime lamb market, however, are usually slaughtered at 4–6 months of age’. She did not, 30 years ago, mention the possibility of the marketing of hoggets. On the other hand, Barber [104] reported that the demand for lamb in France remained high although a decline in sheep numbers was apparent. Inability to ﬁnd more recent information on this topic is regretted. 62. The system demonstrated in the housed Ile de France ﬂock to promote frequent lambing on the small property near Stellenbosch in the South Western Cape Province, with its emphasis on the use of teaser rams to stimulate oestrus, together with the improved fecundity expected to be achieved by ‘really active teasers’, as demonstrated by King and his co-workers, will, in Part 2 of this review, be referred to as the ‘FLASH’ (Frequent LAmbing System StellenboscH, shortened from FLASSH!). The use of ETRs and/or ETGs to stimulate oestrus, and also during mating to promote proliﬁcacy in the ewes, is expected to be of primary importance (at latitudes up to about 40 N or S) in the management of a ﬂock that is to lamb frequently. For not only was the Ile de France ﬂock housed in the Stellenbosch district of the Western Cape, where the suggestion originated that the use of really active teaser goats (ETGs) should be investigated,(as an alternative to the use of ‘ordinary’ teaser sheep-rams), but King’s research into the effectiveness of Ronderib Afrikaner rams (para. 58) earned him his Ph.D. at the University of Stellenbosch.

CAB International 2010

108

APPENDIX 1

g

CAB International 2010

Ewes in ﬂock >12 m/no. excluding those that lambed last time*/% that mated

Mating period

30/?/? 30/13/46.1% 30/24/95.8% 30/8/50% 39/36/69.4% 38/14/100% 44/32*/128.1% 44/10/100% 44/34/88.2% 56/32/81.3% 56/33*/182.6% 66/32/87.5% 63/40*/110.% 68/28/100% 83/59/76.3% 100/65/66.2% 100/65/63.1% 89/54/38.9% 66/45/4.4% 65/64/53.1% 49/16/81.3% 46/35/0% 46/46/95.7% 45/4*/325% 52/40/52.5% 52/31/90.3% 52/24/145.8% 63/31/3.2% 63/62/75.8% 54/7/51.9% 84/31/44.1% 84/59/71.2% 69/29/82.8% 83/60/0% 73/73/78.1.1% 67/12*/150% 56/38/23.7 % 56/47/97.9% 55/12*/208.3% 77/56/66.1% 9/46.4/100% 78/47/44.7%% 91/10/

1974 In France In South Africa February May–June October–November 1975 February–March May–June November–December 1976 February–March May–July November–December 1977 February–March May–June November–December 1978 January–February May–June November–December 1979 January–March May–June November–December 1980 January–February June November–December 1981 January–February June November–December 1982 January-February June–July November–December 1983 January–February June–July November–December 1984 January–February June–July November–December 1985 February–March May–June November–December 1986 February–March May–June November–December 1987 February–March May–June November–December 14-year totals: 1091

Ewes mated ? 6 23 4 25 14 41 10 30 26 42 28 44 28 45 43 41 21 2 34 13 0 44 13 21 28 35 1 47 28 26 42 24 0 57 18 9 46 25 37 32 21

Period of lambing

Ewes lambed/% of ﬂock

Lambs born/ weaned

1974 January–March 17/30 20/18 June–July 6/20 6/6 October–November 22/73 26/25 1975 March–April 3/30 4/4 July–August 24/39 36/34 October 12/38 19/14 1976 April–May 34/77.3 47/43 July 10/22.7 15/14 October–December 24/54.5 33/28 1977 April 23/41.1 36/30 July–August 34/60.1 52/43 October–November 23 38/35 1978 April–May 40 53/45 July 24 35/32 October–November 35 46/39 1979 April–May 35 53/46 June–July 35 47/44 October–November 21 31/27 1980 May 1 1/1 June–July 33 55/53 October–November 11 14/13 1981 – – June–July 41 7/60 October–November 12 18/15 1982 April–May 21 31/29 June–July 28 38/34 November–December 31 47/45 1983 April 1 2/2 June–July 47 85/71 October–November 25 36/32 1984 April–May 25 46/39 June–July 40 68/53 October–December 23 34/26 No ewes in oestrus 1985 July–August 55 83/79 September–November 18 24/23 1986 April–May 9 17/16 June–July 43 70/64 October–November 21 28/26 1987 April–May 37 52/50 June–July 31 49/36 September–November 19 28/27 Oestrous cycles synchronized, ewes inseminated artiﬁcially 985

1490/1294

Lambs born per lambing 1.18 1.00 1.18 1.33 1.5 1.58 1.38 1.50 1.38 1.57 1.53 1.65 1.33 1.46 1.31 1.51 1.42 1.48 1.00 1.67 1.18 – 1.63 1.50 1.48 1.36 1.52 – 1.80 1.44 1.84 1.70 1.48

Weighted mean lambing (% p.a.) )

1974–1987 total lambing rate per ewe (% per annum)

115.6

17/30+6/30+22/30=56.7+20+73.3=150.0%

151.2

3/30+24/39+12/38=10+61.5+31.6=103.1%

139.8

34/44+10/44+24/44=77.3+22.7+54.5=154.5%

157.6

23/56+34/56+23/56=41.1+60.7+41.1=121.0%

135.4

40/63+24/68+35/83=63.5+35.3+42.2=141.0%

146.8

35/100+35/100+21/89=35+35+23.6=93.6%

153.5

1/66+33/65+11/49=1.5+50.8+22.4=74.7%

160.1

0+41/46+12/45=89.1+26.7=115.8%

114.5

21/52+28/52+31/52=40.4+53.8+59.6=153.8%

167.9

1/63+47/63+25/54=1.6+74.6+46.3=115.9%

168.2

25/84+40/84+23/69=29.8+47.6+33.3=104.8%

) ) ) ) ) ) ) ) ) ) )

1.51 141.1 55/73+18/67=75.3+26.9=102.2% 1.11 ) 1.89 1.63 157.6 9/56+43/56+21/55=16.1+76.8+37.5=130.4% 1.33 ) 1.41 1.58 147.0 37/77+31/69+19/78=48.1+44.9+24.4=111.5% 1.45 and ﬂock was sold early in 1988 Mean: 119.45%

*The number of ewes available for mating (the number between slashes in column 1) was usually the number of ewes in the ﬂock (the ﬁrst number in column 1) less the number that lambed during the previous lambing (column 5). The *in column 1 in the body of the Table, indicates that, to the contrary, at this mating, the mated ewes included some that had lambed during the previous lambing.

Animal Science Reviews 2010

Appendix Table Growth of Ile de France ﬂock and summary of breeding records over 14 years, 1974–1987 (see text paras 42–53)

G. L. Hunter

109

Baraboo lle de France Stud

Figure 2. Sons of a Baraboo ram sold to a breeder in the Southern Cape. Figure 1.

The imported ewes in their prime.

Figure 4.

Figure 3. Champion Baraboo lamb carcasses at the Queenstown Show.

g

CAB International 2010

10-month-old Baraboo ewes ready for mating.

110

Animal Science Reviews 2010

References (1) ABA=Animal Breeding Abstracts The paragraph(s) in which references are quoted are recorded after each {in parentheses} 1. Hunter GL. Effects of season and mating on oestrus and fertility in the ewe. Proceedings of the South African Society for Animal Production 1964;3:196 [ABA abstr.] {1} 2. Fell HR. Housing the ewe ﬂock. Journal of the Royal Agricultural Society 1968;128:38–45 [ABA abstr.] {1, 10} 3. Williams S. Indoor sheep – their behaviour, comfort and management. Journal of the Royal Agricultural Society 1968;128:46–52 [ABA abstr.] {1} 4. Vega VY. Possibilities of permanent indoor housing as a method of managing Aragon sheep. I and II. Avances en Alimentacion y Mejora Animal 1967;8:765–85; 889–894 [Sp with Eng. Summ.] [ABA (1969) 37, abst. 1601] {1} 5. Baxter SH. Intensive housing and its effects on farm animals. South African Journal of Animal Science 1974;4:259–64 [ABA 43, abstr. 2727] {1}

18. Slade CFR, Stubbings L. Sheep housing. In: Charles DR, editor. Livestock Housing. CAB International, Wallingford, UK; 1994. p. 359–78 {6} 19. Berge E. Housing of sheep in cold climate. Livestock Production Science 1997;49:139–49 [ABA abstr.] {6} 20. Romero G. Bases for the design of sheep housing. Albeitar 2003;65:10–13 [ABA abstr.] {6} 21. Bøe K, Nedkvitne JJ, Austbø D. The effect of different housing systems and feeding regimes on the performance and rectal temperature of sheep. Animal Production 1991;53:331–7 [ABA abstr.] {8} 22. Hunter GL. Some effects of plane of nutrition on the occurrence of oestrus in Merino ewes. In: Proceedings of the Fourth International Congress on Animal Reproduction, The Hague, Netherlands; 1961. p. 197–201 [ABA abstr.] {15} 23. Johns MA. Estimation of the week of conception in Merino ewes using real-time ultrasonic imaging. Australian Journal of Experimental Agriculture 1993;33:839–41 [ABA abstr.] {15, 60}

6. Hume A. Low cost sheep housing. Farm Business Outlook 1975;3:15 [ABA abstr.] {2, 40}

24. Garcia A, Neary MK, Kelly GR, Pierson RA. Accuracy of ultrasonography in early pregnancy diagnosis in the ewe. Theriogenology 1993;39:847–61 [ABA abstr.] {15}

7. Loboda KI. Production methods used during the permanent housing of sheep indoors at an experiment farm. Belorusskii Nauchno-Issledovatel’skii Institut Zhivotnovodstva 1976;17:99–108 [ABA abstr.] {2}

25. Stevens DR. Ewe nutrition: decisions to be made with scanning information. Proceedings: New Zealand Society of Animal Production 1999;59:93–4 [ABA abstr.] {15}

8. Robertson AM. Sheep housing. Farm Building Progress 1978;51:1–4 [ABA abstr.] {2} 9. de Montis S. Flooring for sheep housing in Sardinia. In: Seminar of the Commission Internationale du Genie Rural, Modelling, Design and Evaluation of Agricultural Buildings, Scottish Farm Buildings Investigation Unit, Aberdeen, Scotland; 1981. p. 237–43 [ABA abstr.] {3} 10. de Montis S. Performances from conventional and new building materials and components for sheep housing in Sardinia. In: Seminar of the Commission Internationale du Genie Rural, Modelling, Design and Evaluation of Agricultural Buildings, Scottish Farm Buildings Investigation Unit, Aberdeen, Scotland; 1981. p. 245–51 [ABA abstr.] {3} 11. Linklater KA, Watson GAL. Sheep housing and health. Veterinary Record 1983;113:560–4 [ABA abstr.] {3} 12. Westendorp TJ, Elfrink BJ. ‘Do it yourself’ sheep housing. Publikatie, Instituut voor Mechanisatie, Arbeid en Gebouwen 1984;192:61pp. [ABA abstr.] {4} 13. Burgkart M. Sheep housing on farms. In: Moderne Haltungssysteme und Tiergesundheit. FreisingWeihenstephan, GFR; 1985 [ABA abstr.] {4}

27. Verberckmoes S, Vandaele L, de Cat S, El-Amiri B, Sulon J, Duchateau L, et al. A new test for early pregnancy diagnosis in sheep: determination of ovine pregnancy associated glycol-protein (ovpag) concentration by means of a homologous radioimmuno-assay. Vlaams Diergeneeskundig Tijdschrift 2004;73:119–27 [ABA abstr.] {15} 28. Yotov S. Diagnostics of early pregnancy in Stara Zagora dairy sheep breed. Bulgarian Journal of Veterinary Medicine 2005;8:41–5 [ABA abstr.] {15} 29. Barr F. Getting the best results from ultrasonography. In Practice 2007;29:520–5 [ABA abstr.] {15} 30. Lacasta D, Ferrer LM, Ramos JJ, Gonzales JM, de las Heras M. Inﬂuence of climatic factors on the development of pneumonia in lambs. Small Ruminant Research 2008;80:28–32 [ABA abstr.] {19} 31. Belonje P, Hunter G, Heine E. Copper toxicity in housed sheep. Veterinary Clinician 1971;5:14–15. {20}

14. Hujnak J. Novel design for cattle and sheep housing. Mechanizace Zemedelstvi 1986;36:402–8 [ABA abstr.] {4} 15. Moreau V. Light weight sheep housing. Bulletin Technique du Machinisme et de l’Equipement Agricoles 1987;14:51–5 [ABA abstr.] {5} 16. Szalay J. Universal wooden housing for sheep. Mechanizace Zemedelstvi 1987;37:310–3 [ABA abstr.] {5} 17. Dornic V. The development of cattle, pig and sheep housing in east Slovakia, its problems and solutions. Mechanizace Zemedelstvi 1989;39:121–4 [ABA abstr.] {5}

g

26. Karen A, Szabados K, Reiczigel J, Beckers JF, Szenci O. Accuracy of transrectal ultrasonography for determination of pregnancy in sheep: effect of fasting and handling of the animals. Theriogenology 2004;61:1291–8 [ABA abstr.] {15}

32. Robinson JJ, Ørskov ER. An integrated approach to improving the biological efﬁciency of sheep meat production. World Review of Animal Production 1975;11:63–75 [ABA abstr.] {27, 29, 40} 33. Combrink GC, Schoeman SJ. The inﬂuence of exercising rams on the lambing performance of a Merino ﬂock. South African Journal of Animal Science 1993;23:24 [ABA abstr.] {22} 34. Joubert DM. Puberty in female farm animals. ABA Review 1963;31:295–306 {23}

CAB International 2010

G. L. Hunter 35. Dy´rmundsson O´R. Puberty and early reproductive performance in sheep. I. Ewe lambs. ABA 1973;41:239–88 {23} 36. Van der Westhuysen JM. The relationship of birth status and early reproductive performance with lifetime reproductive performance in Merino ewes. South African Journal of Animal Science 1973;3:29 [ABA abstr.] {23} 37. Turner H. Newton, Australian sheep breeding research. ABA Review 1977;45:9–31 {23, 28, 30, 43, 61} 38. Gunn RC. A note on difﬁcult birth in Scottish hill ﬂocks. Animal Production 1968;10:213–5 [ABA 36, abstr. 2716] {24} 39. Giles JR. A comparison of two lambing management systems. Proceedings of the Australian Society of Animal Production 1968;7:235–8 [ABA 37, abstr. 2565] {24} 40. Phillips CJC. Animal welfare considerations in future breeding programmes for farm animals. ABA Review 1997;65:645–54 {24} 41. Jarvis S, Day JEL, Reed B. British Society of Animal Science Ethical Guidelines for research in animal science. In: Proceedings of the Britain Society of Animal Science, 4–6 April. British Society of Animal Science, York; 2005. p. 247–53. 42. Tempest WM. Management of the frequent lambing ﬂock. In: Haresign W, editor. Sheep Production. Butterworths, London; 1983. Chapter 24 [ABA Review?] {25} 43. Meat and Livestock Commission. Feeding the Ewe. Meat and Livestock Commission Meat Improvement Services, Bletchley, UK; 1980. {25} 44. Robinson JJ. Some aspects of ewe nutrition. Veterinary Records 1973;92:602–6 [ABA abstr.] {26} 45. Robinson JJ. Nutrition of housed sheep. In: Marai IFM, Owen JB, editors. New Techniques in Sheep Production. London, Butterworths; 1987. Chapter 16 [ABA Book Reviews, p.] {26} 46. Wallace JM, Da Silva P, Aitken RP, Cruickshank MA. Maternal endocrine status in relation to pregnancy outcome in rapidly growing adolescent sheep. Journal of Endocrinology 1997;155:359–68 [ABA abstr.] {26} 47. Ørskov ER. Nutrition of housed sheep. In: Marai IFM, Owen JB, editors. New Techniques in Sheep Production. Butterworths, London; 1987. Chapter 17 [ABA, Book Reviews, p.] {26} 48. Owen JB. Sheep Production. Baillie´re Tindall, London; 1976. [ABA, Book Reviews, p.] {27} 49. Donald HP, Read JL. The performance of Finnish Landrace sheep in Britain. Animal Production 1967;9:471–6 [ABA 36 abstr. 390] {28, 29} 50. Hogue DF, Magee BH, Travis HF. Feed ewes to produce more than two lambs, or wean lambs to dry diets at 10 days of age? In: Proceedings of the Cornell Nutrition Conference (October/November); 1978. p. 141–3 {28} 51. Walton P, Robertson HA. Reproductive performance of Finnish Landrace ewes mated twice yearly. Canadian Journal of Animal Science 1974;54:35–40 [ABA abstr.] {29} 52. Jenkins TG. Post weaning performance and carcase characteristics of crossbred ewe lambs produced in accelerated or annual lambing systems. Journal of Animal Science 1986;63:1063–71 [ABA abstr.] {29}

g

111

53. Gregory KE. Livestock production systems for increased yield on resources. South African Journal of Animal Science 1972;2:139 [ABA abstr ] {29} 54. Greef JC. Evaluation of the Finnish LandraceMerino and Merino as dam lines in crosses with ﬁve sire lines: Slaughter and carcase traits of ram lambs. South African Journal of Animal Science 1992;22:21 [ABA abstr.] {29} 55. Fahmy MH, editor. Proliﬁc Sheep. CABI, Wallingford, UK; 1996 [ABA Review] {29} 56. Tempest WM, Boaz TG. The seasonality of reproductive performance of Merino sheep in Britain. Animal Production 1973;17:33–41 [ABA, 41, abst. 5330] {30} 57. Joubert DM. Indigenous South African sheep and goats: their origin and development. Tropical Science 1969;11:185–95 [ABA 39, abstr. 473] {31} 58. Van Wyk JB, Fair MD, Cloete SWP. Revised models and genetic parameter estimates for production and reproduction traits in the Elsenburg Dormer sheep stud. South African Journal of Animal Science 2003;33:213 [ABA abstr.] {31} 59. Van Rensburg PJJ. Slaughter lamb production in the Western Cape Province. In: Handbook for Farmers in South Africa, Volume 3, Government Printer, Pretoria; 1957. p. 209–14 {31} 60. Basson WD, Van Niekerk BDH, Mulder AM, Cloete JG. The productive and reproductive potential of three sheep breeds mated at 8-monthly intervals under intensive feeding conditions. Proceedings of the South African Society for Animal Production 1969;8:149–54 [ABA 38, abstr. 2640] {31} 61. Karberg FOC, Fourie AJ. (The reproductive potential of a Do¨hne Merino ﬂock in an 8-month breeding cycle. South African Journal of Animal Science 1982;12:306 [ABA abstr.] {31} 62. Schoeman SJ. Production parameters for Do¨hne Merino sheep under an accelerated, intensive lambing system. South African Journal of Animal Science 1990;20:174 [ABA abstr.] {31} 63. Schoeman SJ, de Wet R, Van der Merwe CA. Assessment of the reproductive and growth performance of two sheep composites, developed from the Finnish landrace, compared to the Dorper. South African Journal of Animal Science 1993;23:207 [ABA abstr.] {31} 64. Hunter GL. Increasing the frequency of pregnancy in sheep. 1. Some factors affecting rebreeding during the post-partum period. ABA 1968;36:347–78 {32} 65. Hunter GL. Increasing the frequency of pregnancy in sheep. 2. Artiﬁcial control of rebreeding, and problems of conception and maintenance of pregnancy during the post-partum period. ABA 1968;36:533–53. {32} 66. Pawson HC. Robert Bakewell: Pioneer Livestock Breeder. With foreword by Sir James A. Scott Watson. Crossby Lockwood and Son Ltd, London; 1957 [ABA 26, Book Reviews, p. 221] {35} 67. White RG 1958; Pawson, H.C. (1957): Robert Bakewell: Pioneer Livestock Breeder. ABA 26, Book Reviews, p. 221 {35} 68. Lamberson WR, Thomas DL. Effects of inbreeding in sheep: a review. ABA 1984;52:287–97 {36} 69. Harris D. Hydroponic fodder production. In: Hydroponics. Purnell and Sons, Cape Town; 1977. Chapter 14. {38}

CAB International 2010

112

Animal Science Reviews 2010

70. Marshall FHA. On the change over in the oestrous cycle in animals after transference across the equator, with further observations on the incidence of the breeding seasons and the factors controlling sexual periodicity. Proceedings of the Royal Society B 1937;122:413–28 [ABA abstr.] {41, 42} 71. Underwood EJ, Shier FL, Davenport N. Studies in sheep husbandry in Western Australia. V. The breeding season of Merino, crossbred and British breed ewes in the agricultural districts. Journal of the Department of Agriculture, West Australia 1944;21:135–43 [ABA abstr.] {41} 72. Land RB, Robinson DW. Genetics of Reproduction in Sheep. Butterworths, London; 1985. [ABA Review] {42} 73. Marshall FHA. The oestrous cycle and the formation of the corpus luteum in the sheep. Philosophical Transactions of the Royal Society 1904;196:47 {45} 74. Cole HH, Miller RF. The vaginal smear of the ewe. Proceedings of the Society for Experimental Biology 1931;28:841 [ABA abstr.] {45}

South African Society for Animal Production; 1965;4:136–9 [ABA abstr.] {56} 86. Lyle AD, Hunter GL. Effect on occurrence of oestrus in the ewe during spring of six and 14 day teasing periods prior to mating. Proceedings of the South African Society for Animal Production 1965;4:140–2 [ABA abstr.] {56} 87. Hancock JL, McGovern PT, Stamp JT. Failure of gestation of goatsheep hybrids in goats and sheep. Journal of Reproduction and Fertility, Supplement 1968;3:29–36 [ABA abstr.] {60} 88. McGovern PT. Goat and sheep hybrids. Animal Breeding Abstracts 1969;37(1):1–11 [ABA Review, 37, 1–11] {60} 89. Sambraus HH. On the selection of a mating partner in oestrous sheep and goats. Schweizer Archiv fu¨r Tierheilkunde 1974;116:339–46 [ABA 43, abstr. 4067] {60} 90. New Zealand, Ministry of Agriculture and Fisheries. Annual Report of Research Division; 1973. p. 1971–72 [ABA 42, abstr. 3015] {60}

75. Robinson TJ, Moore NW. The interaction of oestrogen and progesterone on the vaginal cycle of the ewe. Journal of Endocrinology 1956;14:97 [ABA abstr.] {45} 76. Ducker MJ, Boyd JS. An evaluation of the vaginal smear technique for detecting the occurrence of ovulation in the ewe. Journal of Reproduction and Fertility 1974;41:249–51 [ABA 43, abstr.1766] {45} 77. Hunter GL. The role of the ram when synchronising the mating of sheep with progestagens. Proceedings of the South African Society for Animal Production 1969;8:143 [ABA abstr.] {51} 78. Hunter GL, Belonje PC, Van Niekerk CH. Effects of season, suckling and teasing on post-partum interval to ovulation in ewes. Proceedings of the South African Society for Animal Production 1970;9:179 [ABA abstr.] {51} 79. Hunter GL, Belonje PC, Van Niekerk CH. Synchronised mating and lambing in Spring-bred Merino sheep: the use of progestagen-impregnated intravaginal sponges and teaser rams. Agroanimalia 1971;3:133–40 [ABA 40, abstr. 3271] {51} 80. Van der Westhuysen JM, Van Niekerk CH, Hunter GL. Time of application and possible application of artiﬁcial insemination in sheep on a time basis after the use of progestagen sponges. Proceedings of the South African Society for Animal Production 1970;9:183 [ABA abstr.] {51} 81. Van der Westhuysen JM, Van Niekerk CH, Hunter GL. Duration of oestrus and time of ovulation in sheep: effect of synchronisation, season and ram. Agroanimalia 1970;2:131–7 [ABA 40, abstr. 601] {51} 82. Gordon I. Controlling oestrus and ovulation. In: Reproductive Technologies in Farm Animals. CABI Publishing, Wallingford, UK; 2004. p. 140–63 [ABA Review] {53} 83. McDonald MF. Factors associated with onset of the breeding season in sheep. Sheep Farming Annual Massey University; 1971. p. 23–30 [ABA abstr.] {53}

91. Roux PJ, Barnard JP. The effect of heterosexual contact on libido and mating dexterity in Karakul rams. South African Journal of Animal Science 1974;4:171–4 [ABA 43, abstr. 3478] {60} 92. Roux PJ, Barnard JP. The effect of heterosexual contact on libido and mating dexterity in Karakul rams. Yearbook, Karakul Breeders’ Society of South Africa 1975;17:99–107 [ABA 44, abstr.1715] {60} 93. Bryant MJ. A note on the effect of rearing experience upon the development of sexual behaviour in ram lambs. Animal Production 1975;21:97–9 [ABA 43, abstr. 4582] {60} 94. Alison AJ. Optimum ram/ewe ratios. In: Proceedings of the Ruakura Farmers’ Conference, 1975. vol. 27, p. 8–13 [ABA 44 abstr. 4266] {60} 95. Alison AJ, Davis GH. Studies on mating behaviour and fertility of Merino ewes. 1. Effects of number of ewes joined per ram, age of ewe and paddock size. New Zealand Journal of Experimental Agriculture 1976;4:259–67 [ABA 45, abstr. 1382] {60} 96. Signoret J-P, Fulkerson WJ, Lindsay DR. Effectiveness of testosterone-treated wethers and ewes as teasers. Applied Animal Ethology 1982;9:37–45 [ABA abstr.] {60} 97. Lishman AW, Hunter GL. Induction of masculine sexual behaviour in the ewe using stilboestrol and PMS. Proceedings of the South African Society for Animal Production 1965;4:142–5 [ABA abstr.] {60} 98. Parsons SD, Hunter GL. Effect of the ram on duration of oestrus in the ewe. Journal of Reproduction and Fertility 1967;14:61–70 [ABA abstr.] {60} 99. Parsons SD, Hunter GL, Rayner AA. Use of probit analysis in a study of the effect of the ram on time of ovulation in the ewe. Journal of Reproduction and Fertility 1967;14:71–80 [ABA abstr.] {60}

84. Scott IC, Johnstone PD. Variations between years in the ram effect when Coopworth or Poll Dorset rams are introduced to seasonally anovular Coopworth ewes. New Zealand Journal of Agricultural Research 1994;37:187–93 [ABA abstr.] {53}

100. King PR, Coetzer WA. Effect of Ronderib Afrikaner rams on the plasma LH concentration in Merino ewes during the oestrous cycle. In: Proceedings of the 34th Congress, South African Society for Animal Science; 1995. Available from: URL: http://gadi.agric.za [ABA abstr.] {60}

85. Parsons SD, Hunter GL. Practices controlling fertility and ﬂock improvement in sheep farming. Proceedings of the

101. Nitter G. Economic response to increasing genetic potential for reproductive performance. In: Marai IFM, Owen, JB,

g

CAB International 2010

G. L. Hunter editors. New Techniques in Sheep Production. Butterworths, London; 1987. Chapter 23. [ABA Book Reviews, p.] {61} 102. Hirzel R. Thoughts on meat production. South African Journal of Animal Science 1972;2:125 [ABA abstr.] {61} 103. Mc Clymont GL. Animal production in a grain hungry world – or competition between man and animals n a resource limited

g

113

world. South African Journal of Animal Science 1976;6:129–37 [ABA abstr.] {61} 104. Barber J. The French sheep industry from farm to table. In: New Developments in Sheep Production. BSAP Occasional Publication No. 14, BSAP, Edinburgh, UK; 1990. p. 39–43 [ABA abstr.] {61}

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Accelerated lambing. Part 2. Increasing the frequency of pregnancy in sheep G.L. Hunter* Address: Formerly of the Department of Agricultural Technical Services, Stellenbosch, South Africa. *Correspondence: Email: [email protected] Received: Accepted:

31 July 2009 3 December 2009

Abstract This presentation has been written for those who may be considering practical intensiﬁcation of sheep production and need information about its implications, potential and hazards, as well as for those with interests in this ﬁeld of research. The literature on factors affecting accelerated lambing is reviewed, including seasonal breeding and the photoperiod, the involuting uterus, effects of lactation and the age of lambs at weaning, some effects of nutrition and of the introduction of rams. Systems developed in the USA, Scotland, South Africa and elsewhere are described and evaluated. Finally, the importance for ﬂock managers of the advice of those with access to developments in this area is illustrated. An index is provided.

Seasonal Breeding and the Photoperiod 63.1 A brief survey by Yeates [1] describes how the role of the photoperiod began to be appreciated by biologists only in the 1920s, and he records some early reviews of the ‘extensive literature’ that developed. According to Yeates (and as mentioned in Part 1 – see para. 41), it was Marshall [2] who ﬁrst ‘observed that when sheep are transported across the equator . . . they ﬁnally alter by six months the time of the calendar year at which they breed’ and it was Yeates’s own [3, 4] work that ‘supplied ﬁnal proof that light plays a major role in regulating the reproductive pattern of sheep’. Since the publication of the two-part review of frequent lambing [5, 6], in which the effect of light on reproduction in the sheep was brieﬂy considered, important observations by Goot [7] and by Ducker and Bowman [8–10] showed that the use of artiﬁcial light to create ‘long days’, followed by suitable reductions in their length, made it unnecessary (in studies of the effects of day length on the breeding season) to use the blacked-out pens of previous researchers [4] with their associated need for adequate ventilation. Hunter

Please note: Reprints are not available from the author, but CABI (Commonwealth Agricultural Bureau International) has digitized all ABA Review papers and they are available electronically as pdf ﬁles. 1

The paragraph numbers for this article continue from Part 1.

g

and van Aarde [11] subsequently also made use of similar reductions in an artiﬁcial day length during spring and early summer, when the natural photoperiod was increasing, and concluded ‘that in most periods of the year, many (South African Mutton Merino) ewes can be remated within 2–3 months of lambing and should thus be able to lamb three times in two years’. Furthermore, they concluded that most mature ewes of this breed are capable of an early return to oestrus, even after lambing in July (when natural day length at the Cape of Good Hope has started to increase), after ﬁnding that the ewes showed oestrus soon after their lambs were weaned. 64. It is obvious but nonetheless noteworthy that since day length changes vary according to the distance from the equator, in, say, Aberdeen, Scotland (at latitude 57 ), the role of the photoperiod in a frequent lambing programme may be found to be more important than at lower latitudes, such as near Cape Town (+34 ) or even Auckland (+37 ), where seasonal day length changes are substantially less than in the far north of Scotland (see para. 42), or for that matter in the south of New Zealand’s South Island at 46 . It also seems possible that sheep born and reared in such relatively ‘extreme’ conditions in Scotland may be more sensitive to and affected by changes in photoperiod than those reared at lower latitudes. Shelton et al. [12] have concluded, for instance, after transferring Rambouillet ewes reciprocally between Texas (about latitude 30 ) and the ‘Northwest’

CAB International 2010

116

Animal Science Reviews 2010

(Utah?, about latitude 40 ), that sheep at the ‘more southerly locations should adapt better to out-of-season or accelerated lambing’ and that ‘the number of hours of darkness appeared to be the primary factor controlling oestrus and ovulation rate’. Perhaps, at lower latitudes, other effects, such as that caused by ram introduction, may be a primary factor? It would be interesting to compare the adaptation to a change in photoperiod in ewes from the north of Scotland with animals from lower latitudes: a more extreme exchange than that of Shelton et al. [12]. Gordon [13] has suggested that melatonin secretion by the ewe is transmitted to the foetus and that this inﬂuences the lamb’s subsequent response to the photoperiod. If correct, this could have wide implications. It might, for example, explain the observations of Shelton et al. [12], mentioned above, and also those of Robinson et al. [14], who reported (according to the ABA abstract – the original paper is not available in South Africa) that ‘although day length was carefully controlled, there was an innate natural seasonal effect on the onset of behavioural oestrus after pessary withdrawal’. 65. Some of the extraneous factors that were reported to inﬂuence the occurrence of ovulation and oestrus in the postpartum ewe were noted in the previous review [5]. These included seasonal factors, particularly changes in the photoperiod. Subsequent observations by Goot [7] and by Ducker and Bowman [8] have shown that the use of artiﬁcial light to create ‘long days’, followed by suitable reductions in their length, made it unnecessary (in studies of the effects of day length on the breeding season) to use the blacked-out pens of previous researchers such as Yeates [4]. Goot, incidentally, also suggested that ‘in sheep with a long sexual season, genetic heterozygosity may be responsible for regulating the breeding season by allowing other environmental stimuli, besides light, to trigger-off sexual activity’ – such as the ram effect? (my addition). 66. A Factorial Experiment: Whichever factor may be ‘primary’ in an area, since my earlier review, many observations and a few reviews on these topics have been published. The following is an attempt to summarize them under appropriate headings and in chronological order. But, to begin with, rather than quoting this reference under three different headings, the effects of three factors, namely season of lambing, lactation and teasing by vasectomized rams, were studied simultaneously by Hunter et al. [15] in Border LeicesterMutton Merino cross-bred ewes, whose breeding season in the southern hemisphere was restricted and conﬁned approximately to the period from February through June. The experiment was designed as a 322 factorial. The ewes lambed (in the South Western Cape) in August–September, October–November or in March (the latter group having been induced to mate ‘out of season’). The lambs of half of each group were removed from their mothers 24 h after birth; the remaining ewes suckled their lambs for 40 days. Thirty days after lambing, half of each group was exposed

g

to teaser rams, while the others remained isolated from rams until their ﬁrst ovulation (determined at laparotomy) had occurred. The season of lambing was clearly the greatest inﬂuence on the interval between lambing and ﬁrst ovulation. After the ﬁrst two lambing periods, most ewes remained anoestrous until the start of the subsequent breeding season. A high percentage of the ewes ovulated in November/December, apparently as a result of the presence of rams. Suckling had no signiﬁcant effects, however, on the postpartum periods, and there was no evidence of a ‘lactational anoestrus’ in any of the groups, including that which lambed in the breeding season (March). Subsequently, Pope et al. [16] also reported an effect of season (spring versus autumn) and lactation (suckling for 40 days versus non suckling), by monitoring plasma concentrations of progesterone in Polypay, Dorset, St Croix and Targhee ewes, to assess days from lambing to ﬁrst normal ovulation, concluding that ‘the response of different breeds to various components of postpartum fertility varies with season and management of the ﬂock’. Fahmy [17] published a report on work in Canada on the use of cross-breeding with the Finnsheep in accelerated lambing, and Fahmy and Lavallee [18] evaluated the productivity of Dorset, and Polypay ewes and crosses between the two breeds that lambed three times in two years (3/2) or ﬁve times in three years (5/3). Production details are given (fertility, lambs born and weaned, lambings per year), and it was concluded that ‘unless fertility in the 5/3 system is improved, the 5/3 system may not be any more advantageous than the 3/2 system’. To exploit the ‘ram effect’ established by King and his collaborators, a system of hand mating is essential and ewes need to be exposed to the ‘really effective sheep or goat rams (teasers, of course – see pars. 58 and 60)’.

The Role of the Involuting Uterus 67. A detailed review of this topic by Hunter [6] concluded that if ewes are to be remated not less than 2–3 months after lambing, the involuting uterus itself was unlikely to be a limiting factor. On the other hand, if the period between pregnancies is to be reduced to only a month, retained blood and cellular debris, rather than a lack of involution of the uterus itself, may be a problem. Van Wyk et al. [19] examined the effect of exogenous hormones and season of lambing on the ewe’s uterine involution, and Schirar et al. [20] have shown that hysterectomy performed within two days of parturition ‘resulted in all cases within two weeks in the formation and development of persistent (more than 65 days) and actively progesterone-secreting corpora lutea immediately after the ﬁrst postpartum luteinizing hormone (LH) surge’. It was concluded that ‘the length of the ovarian cycles after parturition depends mainly on uterine inﬂuences’ and it was also suggested that ‘in lactating

CAB International 2010

G.L. Hunter

ewes, hyperprolactinaemia could lead to a more long lasting progesterone secretion by antagonizing the deleterious effect of the uterus on progesterone secretion’. See also the observations of Mallampati et al. [21] noted at the end of para. 68. Hayder and Ali [22] used 74 Farafra ewes ‘to estimate the effects of lambing season, parity, litter size, ewe body weight, milk yield and their interactions, on the time taken to complete uterine involution and onset of postpartum luteal activity’, using transrectal ultrasonography twice a week. Only the season of lambing affected involution, the process being quicker following lambing in February, compared with lambing in June, with a high milk production delaying the onset of luteal function.

Effects of Lactation and Age of Lamb at Weaning 68. Not surprisingly, in the title of the next paper in the ‘project’ introduced in para. 66, Hunter [23] asked the question ‘Is there a lactation anoestrus in the sheep?’ In the experiment now reported, 20 S.A. Mutton Merino ewes (whose breeding season in the Cape begins in October/November) and 20 Mutton MerinoBorder Leicester ewes (whose breeding season begins in February) lambed during the second half of October (early summer in this area). Half of each group suckled single lambs for six weeks and half for 13–15 weeks. Quoting from the summary, ‘The post-partum interval to ﬁrst oestrus was determined using teaser rams. The crossbred ewes showed oestrus at the usual time at the beginning of the breeding season, and there was no difference in mean postpartum interval resulting from differences in length of lactation. The Mutton Merino ewes which suckled for six weeks showed ﬁrst oestrus in 87.7+21.7 days, vs. 93.1+19.6 days in ewes that suckled for 13–15 weeks. This difference was not signiﬁcant, but 80% of the ewes that suckled for the shorter period showed oestrus within 100 days of lambing, vs. 40% of those suckling for 13–15 weeks. The remaining ewes in the latter group were in oestrus within one cycle length of weaning’. It was concluded that in ewes that lactate for a maximum of about 3 months and which are adequately fed, lactation anoestrus is of little practical signiﬁcance; the season of lambing in relation to the breeding season appeared to be of far greater signiﬁcance. Land [24] also observed the incidence of oestrus in Finnsheep, Dorsets and crosses between the two breeds. Mallampati et al. [21] bred ten Targee ewes each month for a year. On lambing, the ewes were either (1) suckled for 42 days, (2) suckled until killed at 21 days, (3) had lambs removed 24 h after birth, or (4) had lambs removed after 24 h and were killed at 21 days. From parturition, vasectomized rams with markers were with the ewes until ﬁrst oestrus or until killing. Suckling had a pronounced effect on the length of postpartum oestrus, being signiﬁcantly longer in the 42-day versus the 1-day ewes, but there were no seasonal differences in the suckling effect. Non-suckling

g

117

ewes had higher concentrations of follicle stimulating hormone (FSH), but luteinizing hormone (LH) concentrations were similar, and there was no seasonal variation in either hormone. Follicle diameters were greater in non-suckling than in suckling ewes and varied according to season; but there were no observable effects of suckling on uterine involution. 69. Pelletier and Thimonier [25] concluded that although LH release was lower in lactating than in nonlactating ewes, this does not appear to limit the associated ovulation, but it may be related to the reduced fertility of lactating females. Hunter and van Aarde [26] concluded from their observations that ‘if lactating and non-lactating Mutton Merino ewes are fed to meet their respective nutritional requirements, the length of their postpartum anoestrous periods will not differ’. Shevah et al. [27], however (working in Scotland, but not in the extreme north, with ewes that had lambed in December), reported that eight of 17 FinnDorset cross-bred ewes that suckled for one day, showed oestrus between 2 and 20 days after lambing, compared with only one of 33 lactating ewes, and concluded from the latter (in which there was a delay in the onset of oestrus as well as in the release of LH, compared with those that did not lactate) that lactation affected LH release, and furthermore that their failure to observe differences caused by nutritional status between lactating and non-lactating ewes was because of ‘a direct effect of lactation’. Peclaris [28] recorded the effect of suppression of prolactin with bromocryptine treatment for 15 or 16 days on the interval to oestrus and on fertility in Karagounico ewes (a Greek dairy breed) during seasonal anoestrus. ‘The results suggest that suppression of prolactin during lactation improved fertility’. 70. Fuentes [29] introduced his/her paper by noting that in the ewe there are two factors that determine reproductive behaviour: ‘photoperiod and lactation. During the immediate postpartum period the ewe is not sexually active, and when lactation has ended the photoperiod exerts its inﬂuence and the ewe enters a period of anoestrus’. He/she continued with a short review of ‘several methods (that) have been tried to increase the ewe’s productivity by inducing oestrous behaviour during anoestrus’; these included changes to the photoperiod, early weaning, and treatment with gonadotrophic and other hormones, including melatonin. Fuentes concluded that his/her treatments induced oestrus in the ewe ‘during anoestrus and as early as three weeks after parturition’. Also in 1989, Schirar et al. [30] recorded the LH surge in lactating and non-lactating ewes of the Pre´alpes de Sud breed that had lambed in autumn, and also the ‘postpartum interval to ﬁrst luteal phase’, which was similar in the two groups. The interval to ﬁrst oestrus was, however, shorter in the non-suckling ewes, but this was ‘followed by short luteal phases in 60% of non-suckling ewes (but) in only 7% of suckling ewes’. They concluded that a primary consequence of suckling is the regulation of the ‘conditions of resumption of cyclic ovarian activity

CAB International 2010

118

Animal Science Reviews 2010

after parturition’. In the same year, 1989, Newton and Edgerton recorded the effects on LH secretion in postpartum ewes of season (January – winter, March – spring and June – summer) and lactation (the latter by testing for LH at 5, 20 or 40 days postpartum) [31]. ‘Mean basal concentrations of LH for ewes on days 5, 20 or 40 postpartum ranged from 1.6 to 4.6 ng/ml and did not differ. Mean concentrations of LH during the post-GnRH (gonadotrophin releasing hormone) sampling interval were greater for ewes bled on days 20 or 40 postpartum than for ewes bled on day 5 or for unbred control ewes. Weaning on day 37 depressed GnRH-induced LH secretion on day 40 postpartum’, and there was no difference in basal or in GnRH-induced LH secretion between days 20 and 40 postpartum in January or March, while in June, ewes had lower LH secretions on day 20 than on day 40. Other details are also presented and it was concluded that ‘seasonal modiﬁcations of the releasable pool of LH may mask or modify the effect of the post partum interval upon this response’. 71. Subsequently, in 1990, Peters and Lamming published a lengthy review of lactational anoestrus in farm animals, but did not, however, include any of the references listed in paras 63–67 (which did not ‘support’ their conclusion), not even those mentioned in paras 65 and 66 (which were published in the UK, so would have been easily accessible to them – unlike the others, perhaps?), but concluded instead that ‘a period of anoestrus after parturition occurs in all the farm species’, and proceeded to advance an hypothesis ‘to explain the return of ovarian cyclicity’. Reviewers should not be so blatantly selective in their choice of what to ignore! Peters and Lamming [32] also noted that at that time there was ‘little information on the effect of the male in postpartum reproductive activity in the sheep and cow’, in spite of the evidence already published for the sheep ([5, 33] – see para. 74). This view was soon further superseded for the sheep, when Wright et al. [34] demonstrated ‘that reduced nutrient status of postpartum ewes can (a) delay the onset of ovarian cyclicity postpartum; (b) inhibit the occurrence of ram-induced oestrus (but not ovulation); and (c) hasten the onset of cessation of oestrous cycles’. Schoeman [35] evaluated the reproduction and other parameters in a Do¨hne Merino ﬂock in an accelerated lambing system, but observed a lower productivity compared with studies elsewhere, which required further investigation – being a Merino derivative, it is unlikely that the breeding season length of the Do¨hne is more restricted than hitherto believed. 72. Meanwhile, Fogarty et al. [36] also used teaser rams before an early summer mating and concluded that ‘high pregnancy and lambing rates can be achieved from natural joining in the spring following late winter lambings. On an 8-monthly lambing regime, ewes have sufﬁcient time postpartum to return to normal cyclic activity if they are within their breeding season. This highlights the necessity of using breeds with an extended breeding season’.

g

Furthermore they continued that, with ‘management practices such as the use of the ram effect (reviewed by [37]), weaning lambs prior to rejoining, and ensuring adequate nutrition of ewes are expected to contribute to higher performance’, and ﬁnally concluded that ‘In these circumstances there seems little advantage in weaning lambs earlier than 12 weeks of age’.

Effect of Nutrition 73. Vosloo et al. [38] subjected three groups of maiden, Dorper ewes (see para. 18 for information about this breed), during late pregnancy and during about 100 days of lactation, to a high, medium or low feeding level. The ewes lambed in mid winter (June–July), after being mated in January–February, when day length was starting to shorten. After weaning, the ewes were laparotomized to determine date of ﬁrst ovulation and placed on ad libitum feeding in the presence of raddled teaser rams. The highest feeding level signiﬁcantly reduced the postpartum interval to ovulation, but the difference in this interval between the medium and low feeding levels was not signiﬁcant. Variation in the weight of the ewes at weaning accounted for twice the variation in the interval to ovulation as accounted for by the variation in postpartum body weight alone, which was comparable to the observations of Hunter and Lishman [39] working with ewes of a different breed and in a different season (spring, when the ewes were about to start their new breeding season). They reported that 75% of spring lambing Mutton Merino ewes (in a 22 factorial experiment) were stimulated to ovulate by the introduction of teaser rams. See para. 30, and the conclusion of Wright et al. [34], regarding a nutritional effect on ovarian cyclicity, already noted in para. 71.

Effect of the Ram 74. Hunter [40] presented evidence that, for best results when synchronizing oestrus using progesterone, the presence of a ram during treatment as well as for the determination of oestrus after the end of treatment, was to be recommended. Chesworth and Tait [41] noted ‘a large and sudden increase in the levels of LH in ewes exposed to rams’, and also a small, insigniﬁcant effect on the starting date of the breeding season, as well as on the number of lambs born. Also in 1974, Louw et al. reported the inﬂuence on the lambing pattern of spring-mated Corriedale ewes following the introduction of vasectomized rams [42]. Signoret [43] conﬁrmed this observation and also established that it was the smell of the rams that was responsible for the effect. In a review of the effects on the responses of the ewe to ram introduction, Martin et al. [33] concluded that ‘the ram effect operates through neural circuitry that is independent of the circuitry

CAB International 2010

G.L. Hunter

controlled by photoperiod and negative feedback’ and, furthermore, that ‘the response is not dependent upon the presence of oestradiol, but is simply more easily detected, because the initial pulse frequency is far lower, so the change in frequency induced by the ram effect is greater than in the absence of steroid’. In addition, according to these authors, the ram effect is ‘not irreversible’ and should be regarded as ‘a temporary blockade or bypassing of the anoestrous condition’, and that ‘the blockade of ovulation by ram withdrawal [44, 45]’ suggests that ‘the continuous presence of the ram is important for the release of the LH surge’. Lishman et al. [46] had reported that the levels of LH in the plasma of ewes that failed to exhibit oestrus during lactation, as well as in ewes isolated from rams, were similar to the intraoestrus levels of ewes that were cycling. This observation, it was suggested, supported his earlier observation [47] that ewes isolated from rams tend to become anoestrous. Fukui et al. [48] reported results from two experiments in which treated ewes were given progesterone or a synthetic GnRh, or saline, and introduced to both teasers or (two weeks later) to breeding rams. Results are recorded in the abstract, but no conclusions. Finally, Wright et al. [34], working in the south of Australia (about latitude 36–38 S), demonstrated that reduced nutrient status of postpartum ewes can (a) delay the onset of ovarian cyclicity post partum; (b) inhibit the occurrence of ram-induced oestrus (but not ovulation) and (c) hasten the onset of cessation of oestrous cycles. 75. Gordon [49], in his comprehensive chapter on ‘Advancing the Breeding Season’, reviewed work on the manipulation of the light environment and the use of the ‘ram effect’, and commented that ‘ram sexual drive may be a further factor’ to consider in studies of the ram effect (see paras 51 and 54); he also pointed out that (at least in the Irish work) care was taken to ensure that, prior to introducing rams in order to produce ‘an effect’, the rams are ‘well and securely separated from the ewes’ prior to joining. In point of fact, this may have been a serious deﬁciency in the housing system already described in Part I for the Ile de France project, as, when not in use, the rams were housed in a pen at one end of the same building that contained the ewes. It could be argued that to achieve a satisfactory ram effect, when not in use, the rams should have been housed much further away on a different part of the farm, but this would have been difﬁcult on the small property. In the Ile de France ﬂock (see paras 48–52), this could, however, explain some of the variation in response, as well as the lack of response to teaser introduction observed in some years (see also the discussion in paras 95–97).

Reports from Virginia on Accelerated Lambing 76. A report received too late for inclusion in the 1968 review, inadvertently sent to me by surface mail by

g

119

Copenhaver and Carter [50], described the effect (on the rebreeding of 26 Hampshire or Suffolk cross-bred ewes) of weaning ﬁve successive crops of lambs at approximately one month of age. Mean lambing intervals of 203, 205, 249 and 194 days were recorded. Of 169 lambs born, 151 were marketed and weight gains and feed consumption data are presented. Then, in a second similar study, ‘on a scale approaching that of a fair-sized farm unit’, production comparisons were made between 321 ewes in the early weaning – multiple lambing group and 74 ewes in the conventional lambing group. In the two methods, lambs born per ewe per year (the ewes were not particularly proliﬁc) were 1.64 and 1.46, birth weights of the ‘multiple lambing group’ were 10.5 versus 9.8 lb, and daily gain to market and other lamb growth data are presented. Of 525 lambs born in the frequent lambing group, 82.7% were marketed, compared with 86% of the once-a-year lambing group. The authors concluded that the two studies compared well with each other, and list problems that arose, as well as their recommendations for early weaning/frequent lambing systems. Two facts in this report were surprising. First, the breeding of the cross-bred ewes used (both the Hampshire and the Suffolk are known in South Africa to have restricted breeding seasons, and in neither is proliﬁcacy particularly noteworthy, in comparison with the Finnsheep and its crosses; but the other parent of the ewes is not stated – see para. 27 for brief discussion on the importance of proliﬁcacy). Second, there was no mention at any time, of the use of teaser rams. However, three subsequent reports from the same institution were published by Copenhaver and Carter [51], in which 113 DorsetHampshire-sired lambs of 113 ewes were weaned at an average age of 38 days and weighing 32 lb (14.5 kg) and detailed results are given of this and also of a second ﬁve-year experiment involving a comparison of the performance of SuffolkRambouillet ewes that were either housed indoors or kept at pasture. Carter and Copenhaver [52] then compared Dorset, Rambouillet and cross-breds in a ﬁve-year study, apparently resolutely avoiding the Finnsheep cross-breds until Dr Carter’s death, after which Notter and Copenhaver [53] compared cross-bred Finnsheep ewes that lambed at an improved proliﬁcacy rate, with the Hampshire and Suffolk crossbreds. Over ﬁve years, 1=2 - and 1=4 -bred Finnsheep– Rambouillet and 1=2 -bred Suffolk cross-breds that lambed three times in two years, were compared. Conception rates at mating in August (late summer – 90%), November (early winter – 79%) and April (spring – 53%) differed, but breed groups differed only in April, when the 1=2 -bred Finnsheep produced 52% more lambs per litter and 38% more kg of lamb than the SuffolkRambouillet crossbreds, the unexpected conclusion being that Finnsheep cross-breds can be used in some but not all seasons to increase substantially the rate of lamb production in commercial ﬂocks. This should probably be conﬁrmed – see para. 27.

CAB International 2010

120

Animal Science Reviews 2010

Systems Developed for Accelerated Lambing 77. The Rowett Project: In a review, Blaxter [54] described how intensive methods of nutrition and husbandry in sheep production were ‘being developed (at the Rowett Research Institute in Aberdeen, Scotland) and accepted by the farming community (my italics), largely as a result of economic pressures; (these methods are) based on alterations of the breeding cycle, (principally) by control of day length’. Robinson, with co-authors at the Institute, started at about this time [55–58], for example to publish a series of reports on the Institute’s ‘frequent lambing project’ (my term, not theirs, as far as I know). The commencing motivation, that the ‘estimated ‘biological ceiling’ of ﬁve lambs per ewe per pregnancy, and a potential mean lambing interval of 6 months’ was said to be ‘much further from (being achieved) than any other form of domestic livestock production’ has already been mentioned in para. 27, and the group set out ‘to draw attention to the independent nature of the factors controlling two of the main determinants of ewe productivity, namely litter size and frequency of breeding, and to show how . . . it is possible to intensify production’ by either or both. The animals in this project were housed in two separate areas, where day length can be controlled artiﬁcially. Each area provides individual pens for 48 ewes, which are the ‘unselected’ (does this mean ‘randomly selected’?) F1 progeny of Finnish Landrace rams and Polled Dorset ewes. The treatments adopted included ‘photostimulation, the hormonal control of oestrus and the abrupt weaning of the lambs onto solid feed at 1 month of age’. The ingenious light pattern used is illustrated [57], and it is evident that this is intended to emulate the natural pattern, but it changes at nearly twice the speed of the natural photoperiod: ‘Two months after mating the ewes are subjected daily to 18 h. of artiﬁcial light for a period of 1 month. Thereafter the light period is reduced by 31=2 min. per day’. Thus ‘at parturition ewes are always on an August day length (i.e. for Aberdeen, Scotland, at 57 N) and at mating are on an October day length’, so that four cycles of treatment are completed in 27 months. Robinson [57] has pointed out ‘that ewes which have repeatedly failed to breed or which have had a particularly difﬁcult lambing in one pregnancy, leaving them in an unﬁt condition for rebreeding, have not been discarded, (but) are included in the results’. There is little doubt, however, that such ewes would be culled under traditional commercial production systems. Robinson [59] reviewed the literature on the incidence of breeding activity in ewes maintained under natural day length conditions at different latitudes, and pointed out ‘that between latitudes 35 N and 35 S, most breeds are intermittently polyoestrus, and at the extremes of these latitudes some breeds will come into season as day length increases, and artiﬁcial photoperiods involving an abrupt increase to 18–22 h of light daily for 1 month, followed by an abrupt reduction to 8–10 h or to natural day length, will stimulate

g

a fertile oestrus’ (see paras 27, 40). Robinson’s abstract continues: ‘Improvements in lamb production in breeding systems that rely solely on artiﬁcial photostimulation have seldom exceeded 20%, and breed differences in the suitability of ewes for such systems have been reported’. Furthermore, ‘there are conﬂicting reports on the effects of nutrition on breeding frequency, some of which can be resolved on the basis that reductions in body weight of up to 15% have little detrimental effect in the short term, but will reduce long-term production’ (see para. 73). 78. Before proceeding further to review the series of papers that have been published on this project, a plea must again be made for a consideration of all the costs that will arise if and when the system that evolves is ever to be widely ‘commercialized’. As has been mentioned already (para. 25), if the object is to market only lambs for slaughter, housing and the use of purchased feeds in an intensive system of production may not be economically viable everywhere. It is also noteworthy that in the ‘Rowett Project’ the ewes and their progeny are F1 crossbreds, so it is questionable whether any sales of such ‘breeding stock’ may be contemplated. It is obvious that there is little point in developing a production system that is uneconomical in practice. On the other hand, a change of habit on the part of consumers to an appreciation of the good quality of cheaper meat from hoggets (in place of young lamb) that have grown more slowly and more cheaply than the early weaned, well-fed lamb, may be a development that should be encouraged (see para. 61). However, a CAB Abstracts search produced a reference to Haresign [60] that includes an ‘abstract’ that is in reality a listing of the contents of the publication, which consists of 29 reviews of various topics on sheep production. No. 24 [61] is one of seven listed under reproduction that is especially relevant to this review, the purpose thereof being to ‘show how these techniques can be synthesised into a frequent lambing system of production, to detail the management factors to which particular attention must be made and to assess the performance of such a system’. His summary concluded with a short presentation of its ﬁnancial performance, which showed ‘that the ﬂock has produced extremely high margins relative to annual lambing lowland ﬂocks recorded by the Meat and Livestock Commission. The purchased forage costs included purchased hay and an allowance for grazing cow lays and sugar beet tops. Thus high stocking rates can be practiced and high gross margins/hectare achieved, but it is emphasised that purchasing forage is purchasing hectares’ (my italics) – the presumed basis of Blaxter’s claim underlined in para. 77. The same author [62] had, in fact, written previously on this topic, but this paper is not available in South Africa. 79. In addition, in the Rowett project, the oestrous periods of all ewes are synchronized using progesteroneimpregnated pessaries. Thus, a tight breeding schedule of 205 days (6.8 months) is possible between one mating and the next. With the ewes in individual pens, the teaser ram

CAB International 2010

G.L. Hunter

can have little if any place in this programme at least until its role as a stimulant of oestrus can be shown to be more effective and predictable (in Scotland) than experience to date has suggested. The Rowett ‘project managers’ are certainly aware of the potential role of the teaser elsewhere, as the observations (of [63]) that ‘when some British breeds are given continuous access to the ram, individuals can successfully breed well outside the recognised breeding season’ are quoted by Robinson [57], although they may be interested to learn of the differences in effectiveness between breeds of rams (see paras 51 and 59). It may be necessary in Aberdeen to take more serious notice of the ram effect when the costs of commercializing the present Rowett regime are considered; of these, Robinson [57] is clearly aware, when he wrote that ‘its adoption on a commercial scale may be prohibitive on economic grounds’. Is this comment compatible with the claim by Blaxter [54] that was underlined in para. 77? Furthermore, to be critical, it is not apparent what was gained by using ‘unselected’ F1 ewes in the ﬁrst place, and by not culling those that do not ‘perform’ for any reason: in the hands of ‘traditional’ producers, some selection at least of the more proliﬁc F1 ewes would be likely and any that become unﬁt will soon receive short shrift. And how is the Rowett project to proceed, when all the original ewes have ‘shot their bolt’? Is there to be no attempt to select progeny from the better performers? This will surely occur on commercial farms. 80. But to continue with the reports that have been published on the ‘Rowett project’, Robinson and rskov [58] reviewed and discussed in detail the requirements for and costs of feeding both ewes and lambs. Their considerations could be extended to include the possibility of increasing lambing frequency to four times in two years (perhaps by allowing their ewes only one mating opportunity per cycle) ‘as favoured by the French researchers’, referring to a personal communication to these authors from J. Thimonier, and as applied also in the Ile de France ﬂock presented in Part 1 (see end of para. 15) – so that those ewes which do not conceive to the ﬁrst oestrus after lambing simply join the next group that are to be mated, rather than extend the length of the present cycle). Furthermore, the possibility, in a commercial situation in which ewes are not individually penned, of using effective teaser goats instead of progestagen pessaries, should also be considered, especially if it can be conﬁrmed that these also increase the litter size, as achieved by the Ronderib Afrikaner teaser rams of King [64] and of King and Coetzer [65]. (Their ‘super’ rams are not widely available outside of South Africa, but there should be little difﬁculty in acquiring ‘super’ male goats.) Incidentally, there appears to be a small but important error in Fig. 1 of Robinson’s [57] paper, where the length of the ﬁrst of the four artiﬁcial lighting cycles used in the intensive unit is shown as about 7 months, whereas the other three appear to be about or perhaps slightly less than 6 months. Commercial producers may become

g

121

enthusiastic adopters of the ‘Rowett System’ if each of the cycles can be limited to 182 days, so that four cycles can be completed in two years and then repeated, and if cull ewes may be replaced as necessary. If Blaxter’s claim (para. 76) was justiﬁed in 1973, some ‘ﬁne tuning’ of the system may result in even more interest being shown by commercial producers than was then apparent to him. 81. Second, the STAR System: Hogue’s [66] disappointing presentation of Frequent Lambing Systems (the ﬁrst that I have traced, although Hogue claims that by this date he had been for some 15 years involved with STAR) includes no references and no information about the management of the animals on the systems he presents; which made it difﬁcult for others to use them, let alone credit the originator with any innovation. Two systems are brieﬂy described by Hogue et al. [66–67], both originating at Cornell University in the USA. To quote from Hogue’s summary, the CAMAL (Cornell Alternate Month Accelerated Lambing) system ‘allows ewes the maximum opportunity to lamb at intervals of less than 12 months, but is more difﬁcult to manage than the STAR system’, while ‘the STAR system is based on the fact that a calendar year (365 days) contains exactly ﬁve 73 day periods, each of which is . . . half of a sheep’s pregnancy (146 days – conﬁrmed in para. 21). By expressing the calendar in circular fashion with each 73-day period marked off and connecting alternate points, a perfect star is formed’, illustrated by Hogue on p. 61. ‘Ewes are managed to breed at every third point (of the star), thus allowing 73 days for lambing, lactation and weaning and 146 days for pregnancy. Optimally, ewes can lamb ﬁve times in three years. The system,’ Hogue concludes, ‘ﬁts into sheep biology and the calendar year, and because lambing and breeding dates are exactly coincident, management of the sheep is easier than most systems. Hogue ‘considers the STAR system (to be) the frequent lambing sheep production system of choice’. How ewes are managed is not explained, but it is possible that further critical details have been given in other original publications that I have not found: for instance, at what age are the lambs weaned, or is artiﬁcial lighting used (as in the Rowett system), or perhaps the use of teaser rams may be part of management; if so, until more details are available it is impossible to ‘credit’ Hogue’s farm manager with any innovation, apart from the reasoning behind the design of the star! The STAR system may have management practices in common with other frequent lambing systems; so it is regrettable that Hogue was somewhat economical with important details. The problem was eventually solved (para. 83), but not until 19 years later, in 2006! 82. The report of Snyder and Milligan [68] ‘examines and compares economic factors for three levels of lambing performance (on) the STAR system with the results that the same operator could expect under (on) a traditional lambing programme in the USA’. In this paper, it is claimed that the STAR system ‘offers signiﬁcant potential gains in the number of lambs raised per

CAB International 2010

122

Animal Science Reviews 2010

ewe per year . . . the result of improved breeding and feeding programs and (consequently) in increased (out-ofseason) lambing and productivity for the ewe ﬂock. The study also (compares) economic factors for three levels of lambing performance (on) the STAR system with the results the same operator could expect (on, from, with, but not ‘under’ – see ﬁnal comment, para. 103) a traditional annual lambing programme in the USA . . . The analysis is based on a 300 ewe ﬂock (on) the STAR system and a 150 ewe ﬂock (on) the annual system’. The results indicate that the STAR system enables a higher net cash income and return than the annual system (see para. 59). 83. Matthews [69], presumably a veterinarian ‘gets onto the accelerated lambing act’ when he listed a number of ways in which he saw a roll for himself. Magee [70] described the ‘STAR Accelerated Lambing System’ and considers ‘the major limitations of the system under commercial conditions’; a pity that this publication is not available in South Africa. Lewis et al. [71] evaluated the effects on fertility of ewe age, season of mating and the interval from lambing to mating, using the breeding records of 1084 Dorset ewes in the STAR accelerated lambing system, and produced a helpful summary that was a step in the right direction: Flock fertility ‘changed in a cyclic and predictable manner . . . (and) matings that occurred within the typical breeding season (Aug, Oct. and Jan.) were more fertile than matings in Mar. and Jun. Ewe fertility was affected by ewe age and the interval from lambing to mating. Except for matings in Jun., fertility at the ﬁrst postpartum mating increased with increasing ewe age’. Conclusions are also drawn regarding ewe fertility at various stages and ‘a matrix . . . was calculated . . . to describe the combined effect of season, ewe age and the interval from lambing to mating’. 84. Eventually (and at last), Smith [72] took the long awaited large step, and described for the ﬁrst time the 1981 development and management of the STAR accelerated lambing system, by the manager of the Cornell University sheep ﬂock. The system is ‘based on the concept that there are ﬁve sheep half-gestations in a year’. According to the abstract, the system ‘limits (unspeciﬁed) risks associated with having the entire ﬂock pregnant or in the barn at one time, and favours coccidiosis control by the small spread in lamb age. However, continual presence of susceptible sheep perpetuates contagious ecthyma and ampliﬁes worm burdens on pasture. Pregnancy toxaemia threatens early lambing proliﬁc ewes, paratuberculosis is easily spread if thin ewes are housed with the lambing ﬂock, and newly weaned lambs are exposed to diseases carried by unthrifty lambs from the preceding lambing period if management is not excellent’. Management of the system is described brieﬂy as follows: ‘The 365-day year is divided into ﬁve seasons; each is 73 days long ([71] – para. 81). Breeding and lambing occur concurrently at the beginning of each season. After an up-to-1-month-long breeding period (to permit each ewe two opportunities to

g

conceive – although normal cycle lengths can be as long as 18–19 days – para. 15) the rams are removed from the ﬂock and gestation proceeds in the pregnant ewes. At the beginning of the next season (i.e. the next point of the star), the rams are again introduced for an additional breeding period and removed on schedule. At the beginning of the third cycle, the ewes that had conceived during the ﬁrst breeding period are identiﬁed based on palpation of the developing udder and separated off to a lambing area. Those not close to lambing (which included ewes half way through gestation and non-pregnant ewes) spend another breeding period with the ram. The end of lambing coincides with the end of the breeding period, and the ewes continue to lactate until weaning occurs over a two-day period at the end of the 73-day cycle. The lambs are, on average, 55 days old at weaning [71] and are creep-fed. The ewes are given water, but not feed, in order to facilitate the drying off of the lactating ewes’. Following weaning, the ewes are returned to pasture with the breeding rams, many cycle and conceive in the month after weaning, resulting in about a 7.2-month lambing interval. An ewe that lambs ﬁve times at intervals of 7.2 months is designated a STAR ewe, while All-STAR ewes produce twins or better at each of the ﬁve lambings [73]. Altogether a most helpful summary of what is ﬁnally beginning to sound like a good system, but there is more: Smith continues by pointing out that since the ‘farm is located at a north temperate latitude of 42 30 min, not all the breeding periods fall within the traditional autumn (decreasing photoperiod) breeding season. No hormones are used at any time in the Cornell ﬂock, neither is supplemental light used to control the breeding season. Instead, genetic selection has been used to develop a ﬂock that will breed year round. (The) Polled Dorset breed (chosen for mothering ability and a relatively long breeding season) and Finnsheep (for proliﬁcacy) were crossed to produce the FinnDorset ewe, on which the STAR system is based. Rams for breeding are selected from twin and triplet litters of All-STAR ewes; these are ‘aseasonal’ ewes that have conceived in each of the breeding periods’; and some attention has been given to scrotal dimensions in the selection of rams. Smith continues with a helpful discussion of the ‘economic advantages of the system’, of ‘veterinary and animal health effects’ and ﬁnally of ‘potential mistakes when using the STAR system’. Anyone interested in using STAR or a modiﬁcation of the system, is strongly advised to obtain a copy of this ‘four pager’. It would be interesting to see the consequences of using teaser goats for two weeks before each mating period. The success of the system may be as much the result of the productivity of the two breeds as the discipline of management imposed by the design of the STAR. A summary of the STAR system is included in Table 1. It may be of interest to know whether breeding stock from this proliﬁc ﬂock (after some years of selection for the ability to lamb frequently) is available for sale to local commercial producers, and also whether

CAB International 2010

G.L. Hunter

123

Table 1 Calendar for STAR and FLASH systems (years 1 and 2 only, for southern hemisphere – see para. 84). From year 3, both systems will include additional lambings during January and March (STAR) and during April/May (FLASH).

STAR System

FLASH System Year 1

30 d mating

1

30 d mating

2

January February

Tease6 38 d mating

March April

30 d mating & lambing

3

May Tease6 38 d mating

June July 30 d mating & lambing

4

Lambing7

August September

30 d mating & lambing5

October Tease6 38 d mating

November

Lambing7

December Year 2 30 d mating & lambing1

January February

30 d mating & lambing

2

30 d mating & lambing

3

Tease6 38 d mating

March April

Lambing7

May Tease6 38 d mating

June July 30 d mating & lambing

4

30 d mating & lambing

5

Lambing7

August September October November

7

Lambing

Tease6 38 d mating

December 1

Beginning 1 January. Beginning 15 March. 3 Beginning 27 May. 4 Beginning 8 August. 5 Beginning 20 October. 6 For 14 days. 7 Lambs to be weaned by 6 weeks of age. 2

embryos from this ﬂock might be exported to other countries. 85. The following reports on accelerated lambing were three that were published since Hogue’s 1987 [66] ‘release’ – there is no mention of the use of ‘STAR’ or of teaser rams; all are from a project in the Virgin Islands – paras 90 and 94 contain ‘post 1987’ references to STAR. (1) Dodson and Godfrey [74] reported that Dorpers (para. 31) are being crossed with St Croix White (DPRX), the local hair sheep. Ewe production was evaluated in an extensive management and accelerated lambing system and compared with equally small ﬂocks of Barbados Black Belly (BB) and St Croix White (STX) ewes. Breeding

g

occurred for 35 days in October and June and ovulation rates were determined by laparoscopy 7–9 days after mating; lambs were born in March and November and 24-h milk yield of the ewes was determined at intervals after lambing. Lambs were weaned at 63 days in May and June, respectively. Results are presented and show ‘that it is possible to incorporate DRPX ewes into an accelerated lambing system’. (2) Godfrey et al. [75] evaluated two small ﬂocks of BB and STX ewes, both on Panicum maximum rotationally grazed. ‘Single sire breeding took place during 35-day periods in Feb., Jun. and Oct . . . so that each ﬂock produced three lamb crops every two years. Lambs were . . . weaned at 63 days . . . Replacement

CAB International 2010

124

Animal Science Reviews 2010

animals were selected for multiple births, breed characteristics and adjusted weaning weight using an index’. Productivity of the ewes (including proliﬁcacy, which improved to 2.01 in ten years) was rather better than that of the Ile de France ﬂock reported in the Appendix table (Part 1), especially during October, and survival rate to weaning improved to 91.3%. It was concluded that ‘productivity can be . . . enhanced (by) using accelerated lambing’. (3) Godfrey et al. [76] evaluated production traits of DPRX ewes in an accelerated lambing system over a period of 2 years, comparing them with an established STX ﬂock of 40 ewes (four lamb crops, 163 births). The ewes lambed in March and November (2004), July (2005) and March (2006). Productivity ﬁgures for the breeds conﬁrm that the Dorper cross-breds are able to produce well under tropical conditions. 86. The FLASH System (see para. 61). The Ile de France frequent lambing system described from para. 31 needs to be summarized at this point (see Table 1). Breeds or cross-bred ewes with potential for good proliﬁcacy should be selected from those with long breeding seasons (para. 38). The objective of FLASH is to arrange for each ewe three lambing opportunities in two years by means of three ﬂock matings annually. Housing of the ﬂock is not essential, but conserved fodder will be required for lactating ewes during dry periods, so production costs must always be considered (see paras 29, 39, 55 and 59). Three ﬁve-week mating periods are scheduled annually, each preceded by teasing for 14 days, using ‘really active’ vasectomized goats or sheep, the main consideration being the ‘intensity’ of their sexual behaviour – see para. 58. The ﬁrst mating is in March (early autumn in the southern hemisphere), the second in June (winter) and the third in November (early summer). Lambing therefore takes place in July/August (late winter), October/ November (early summer) and March/April (autumn). For northern hemisphere ﬂocks, mating and lambing will be similarly scheduled, but in appropriate months. Breeding animals of both sexes should be selected on the proliﬁcacy of their mothers (para. 55). Neither artiﬁcial light nor hormone treatment of the ewes is required to promote accelerated lambing in the FLASH system. Lambs are weaned by 42 days. Ile de France and Dorper embryos are already being exported from South Africa to Australia, where another period of low prices for Merino wool is currently being experienced (see para.33).

Other Publications on Accelerated Lambing – in chronological order 87. The First from Eyal et al. [77], who reported the change (in a ﬂock of Awassi and Awassi cross-bred dairy ewes in Israel) from lambing annually to frequent lambing, ‘which included four 70-day mating periods per year with 20-day intervals between them’. Their report was, however, concerned with the milk yield of the animals, rather

g

than with their reproduction, and does not mention the dates of the mating periods. (2) Espego Diaz [78] reported that three lambings in two years were compared with ‘continuous’ (all year round) lambing, for Aragon sheep in north eastern Spain, for other breeds in the central region, and for Merinos in the south and west of that country. The system was used for ﬁve years in a ﬂock of 672 Aragon ewes, resulting in an annual production of 1.4 lambs weaned per annum. The conception rate of ewes mated in November was 80%, versus 60% for ewes mated in March (3) In 1976, Jennings and Lawlor produced a short report on ewes that had lambed in December in Ireland (breed not stated in the abstract), and were treated in February–March with a 60 mg MAP sponge and 500 i.u. PMS [79]. Lambs were removed at 48 h or at 4 or 8 weeks or ‘suckled continuously’. The percentage of ewes mating and lambing after treatment is given for the four treatment groups, but there was apparently no follow-up in the subsequent season. (4) Also in 1976, Flanagan and Quirke presented data on Galway and 1=4 Finn–3=4 Galway ewes that lambed in April, December and August during a two-year period [80]. The numbers are given in the abstract. These show that although many of the ewes mated, conception rates were lower than in ﬂocks mated once a year. (5) In 1978, Eyal et al. reported the production achieved by dairy ewes in Israel [81]. Awassi ewes and cross-breds were mated four times per year for three years, during summer (June–July), autumn (September–October), winter (December– January) and spring (March–April). Treatment with a 15-day progesterone course and Pregnant Mare Serum Gonadotrophin (PMSG) improved the incidence of oestrus as well as the lambing rates of the cross-breds, but not of the Awassi ewes. ‘Ewes which had lambed twice and averaged 1.5 lactations . . . produced 30% more milk per year than those which lambed only once’, and further details are recorded. 88. (6) Two reports on the productivity of proliﬁc Romanov cross-breds (as alternatives to Finnsheep) (a) by Marzin et al. [82] and (b) by Marzin and Brelurut [83], summarized the performance of (a) two ﬂocks of Limousin and Romanov cross-bred ewes at one location and of Arles Merino and Romanov cross-breds that lambed every two years at different locations, and (b) of 4656 matings during a 4-year period in a Limousin ﬂock and a ﬂock of RomanovLimousin F1 ewes. In (a), oestrus was synchronized using sponges and PMSG and the ewes were inseminated two days after the PMSG injection. In four years, the number of matings recorded for the four groups was 2 836, 1 820, 880 and 641 respectively, and the number of lambs weaned per ewe per year was 1.93, 2.66, 1.24 and 1.92, respectively. For Limousin and Romanov cross-breds, productivity was signiﬁcantly greater after the October matings than after mating in January and June, and ‘lamb production remained high up to the 9th mating’. In (b), both ﬂocks lambed three times in two years, and details of their performance are

CAB International 2010

G.L. Hunter

presented, concluding with ‘the number of lambs weaned per 100 ewes mated was highest in both genotypes for ewes lambing at 2–4 years (about 130 for purebreds and 165 for cross-breds); it fell to about 65 at ten years of age for purebreds, and to about 120 at 8–9 years for crossbreds’. (7) A report published in 1980 by Dzakuma et al. [84] described how ﬁve breed combinations aged 7–8 years, some including Finnsheep blood, performed in two cycles of accelerated lambing, totalling six mating periods in two years. The lambing rate in both years averaged 76%. Mean levels of production are given for the ﬁve groups. (8) Cognie et al. [85] concluded that postpartum ovarian activity is related more to the presence of the lamb than to the amount of milk produced by the ewe that the relationship between ewe and lamb and/or a high level of prolactin ‘retards the establishment of positive feedback to oestrogen at 3–4 weeks postpartum in the breeding season, and suppresses it in the non-breeding season’ that a ‘lack of response to gonadotrophins is the main cause of ovarian inactivity during postpartum anoestrus’ and that ‘seasonal anoestrus is related more to levels of gonadotrophin, particularly FSH, than to the increase in prolactin that occurs with longer days, but high levels of prolactin are often associated with a high incidence of short cycles’. (9) Jankowski [86] reviewed the experimental work in Poland on frequent lambing and (10) Eyal [87] reported to the 32nd EAAP meeting on results of both research and commercial use of accelerated lambing in dairy sheep. (11) Dyrmundsson [88] reported to the same meeting of the EAAP on frequent lambing in Icelandic sheep. (12) Valls Ortiz [89, 90] described ‘two management systems commonly used in Spain to increase lambing frequency; these are keeping rams permanently with the ewe ﬂock, and arranging mating periods so that ewes lamb at intervals of 8 months. In the Churro breed, the average number of lambings per ewe per year has increased in recent years from 1.1 to 1.6. Factors inﬂuencing lambing frequency, e.g. oestrous activity in spring, feeding, lactation and the presence of rams in the ewe ﬂock are discussed. Data obtained on 16 commercial farms, where Aragon ewes were mated at different intervals, are tabulated. Ewes mated every four months (four systems, depending on the months rams are ﬁrst joined with ewes) averaged 1.58–2.00 lambs per ewe per year. The advantage in lamb production obtained by using Romanov rams on Aragon ewes was about 40%. ‘More frequent lambing than three times in two years does not seem practicable’. (13) Also in 1981, Notter (in Virginia – see para. 76) examined repeatabilities of conception rates (CR) and litter sizes for 131 Finnsheep cross-bred ewes in an accelerated lambing system in which the ewes lambed three times every two years, in September (autumn), January (winter) and April (spring). From 717 lambings ‘repeatability estimates . . . ranged from 0.076+0.022 to 0.127+0.024 [91]. Within season repeatabilities for litter size were signiﬁcant only for the Sep. lambings (0.188+0.046) and the use of conception

g

125

rate for the autumn lambings as a culling criterion would have maximised the performance of the remaining selected ewes in all seasons. Overall, ewe effects on litter size were signiﬁcant, and the repeatability of litter size was 0.156+0.32 over all seasons’. Other interactions and correlations are also presented. Finally, it was concluded that ‘direct selection for litter size within each season would have nearly maximized the performance of (the) remaining ewes’. 89. (14) Dzakuma ([92], in a dissertation abstract) and Dzakuma et al. [93] published in a journal article, were obviously reports from the same project. The productivity, or fertility and proliﬁcacy data are presented for different (2- and 3-breed) combinations of the Dorset, Rambouillet and Finnsheep breeds in an accelerated lambing system, in which the 4–5-year-old ewes went through two cycles of lambing three times in two years. In the journal article, it was reported that ‘reproductive performance averaged over the two cycles showed reduced fertility after May–June matings (spring–early summer 47.8%), compared with matings in January– February (late winter–early spring, 91.8%) and in September–October (autumn, 90.6%)’. Fertility levels varied also between purebred and cross-bred rams, as a result of which, in the thesis, various interactions and the fertility of different breed combinations were presented. More data are to be found in the two abstracts, and see also no. 7 in para. 88. (15) Hulet [94] reported the effects of early weaning, lactation and ‘day-of-year lambing on the ability of Polypay ewes (a ‘cocktail’ now recognized in the USA as a commercial breed), developed from Polled DorsetTargheeFinnsheepRambouillet breeds, to rebreed following winter and summer lambings. Ewes lambing in winter did not mate successfully while lactating, but weaning the winter-born lambs at 31 days resulted in more ewes rebreeding (35.7%) and producing summer lambs than when lambs were weaned at 41 days postpartum (23.6%). Furthermore, ‘ewes that lambed during the early part of the winter lambing period had an advantage over ewes lambing later’ – a higher percentage of the former lambed in the following summer. Further details are available in the summary. (16) A short abstract is listed next from the 1981–82 Annual report of the South African Department of Agriculture, on accelerated lambing in the Do¨hne Merino breed – which is probably the same project on which Schoeman’s [35] paper reported in rather more detail (see para. 71). Lamb production by the ewes was reported to be 33% higher than that attained with lambing once a year. (17) Amir et al. [95] reported a study in which Finnsheep (FL)German Mutton Merino and FLAwassi ewes were studied after lambing in autumn (October) and winter. After 70% had ovulated without showing heat, 90% of all ewes showed ﬁrst postpartum oestrus within 60 days of lambing in October, and it was ‘suggested that (selection of) breeding ewes with ovarian activity after oestrus induction during anoestrous periods might increase the

CAB International 2010

126

Animal Science Reviews 2010

reproductive performance of the ﬂock in a frequent lambing system’. 90. (18) Also in 1984, Goot et al. [96], in a ﬁnal report to the United States–Israel Bi-national Agricultural Research and Development Fund, on a project in which the performance was analysed of Finnsheep cross-breds either in an experimental or in two commercial ﬂocks . The wording of the abstract is not clear, but it seems that both ﬂocks lambed three or four times per two years. Data on growth rates and mortality, rather than on reproduction of the ewes are summarized. (19) Hackett and Wolynetz [97] reported that 330 ewe lambs in two ﬂocks maintained indoors from ‘three synthetic crossbred strains being developed at the Animal Research Centre in Ottawa’ (no further details of origin are given in the abstract), were ﬁrst bred (without preliminary hormone treatment) at 6–7 months. The ﬂocks ‘were bred alternately in Jan (winter), May (spring) or Sep (early autumn) to establish an 8-month breeding cycle. The rams (at least ten months old) were maintained at alternating day lengths in one of two lighting regimes’; 15% of the ewes lambed’, producing an average of 1.6 lambs, and it appeared that ‘ewe lambs mated to rams maintained at a constant day length (of 10 h) had lower fertility than lambs mated to rams maintained at alternating day lengths (four months of 10 h light and four months of 18 h light. Although fertility level was low, the data also suggest that the breeding of ewes aged 6–7 months is feasible’. (20) Iniguez et al. [98] evaluated two frequent lambing systems, ‘Morelam (using Morelam sheep and the Camal system developed at Cornell University, but not described here) or using Dorset ewes, for age at ﬁrst lambing, lambing interval and probability of conception. The Morelam ewes were continuously exposed to rams over the year, while the Dorsets were exposed every other month . . . Lambing intervals averaged 293 and 303 days in the two types of ewe, and longer lambing intervals resulted when ewe lambs were mated before they were 12 months of age and when the previous lambing occurred in winter . . . The overall probability of conception for the Morelam system (0.16) was relatively higher than for the Camal system (0.14) and the numbers of lambings per ewe per year were 1.28 and 1.21, respectively. Estimates of heritability of age at ﬁrst lambing, lambing interval and conception probability were 0.31, 0.06 and 0.30, respectively. (21) Veress et al. [99] presented data from large sheep farms to show that the Hungarian Merino is capable of lambing every 7–9 months at an average of 1.2–1.4 lambings per year, and ‘suggested that combined selection for accelerated breeding and twinning should be undertaken for this breed’. (22) From Mexico, Urrutia et al. [100] reported that during a three-year period, matings were carried out at eight-month intervals in a ﬂock of 180 Rambouillet ewes and results include the fact that overall the number of lambs born per year averaged 1.49. 91. (23) Gabin˜a [101] presented repeatability and heritability estimates of reproductive traits in four Aragon

g

ﬂocks that were lambing frequently, and concluded, after combining estimates into a selection index, that ‘litter size was the most important trait, accounting for 74–96% of the variation in the economic value of genetic progress’. (24) Aboul-Naga et al. [102] studied the performance of Finnsheep (FL) and Finn cross-bred ewes over 18 successive mating periods. The ‘crossbreds retained the high proliﬁcacy of FL ewes, but low conception rates and high lamb mortality markedly reduced number of lambs weaned per ewe joined. Conception rates and number of lambs born per ewe lambing were signiﬁcantly higher in F1 and backcross ewes than in purebreds . . . Ewes performed better when mated in Sep than in Jan or May . . . and all breed (groups) showed marked seasonal variation in proliﬁcacy’. And, in 1991, Aboul-Naga reported on the reproductive performance of two Egyptian sheep breeds in accelerated lambing systems. (25) Lahlou-Kassi et al. [103], (26) Bradford et al. [104] and (27) Berger et al [105], are three papers from a project at the Agronomy and Veterinary Institute, in Rabat, Morocco (in cooperation with researchers at the University of California, Davis), where ‘sheep are an important part of the agricultural economy, with an estimated 14 million animals’. Reproduction traits were recorded in Sardi and D’Man sheep (Moroccan breeds) and their crosses on an accelerated lambing programme. The traits included fertility, litter size, postpartum anoestrus, puberty, ovulation rate, embryo survival, and lamb mortality, as well as growth and production per ewe; both breeds were found to perform well in the system used. ‘It was concluded that the D’Man breed has excellent potential for increasing total lamb production when crossed with less proliﬁc breeds’. (28) Mendel et al. [106] compared year-round lamb production of two breeds ‘known for their extended breeding season’, the MerinoLandschaf and the Tyrol Mountain sheep (TM) and certain crosses, which showed that crosses with the TM were superior in all aspects except birth weight. Other differences ‘were minor’. The animals were ‘kept under intensive management conditions’, details of these are not provided, but other observations are contained in the abstract. (29) Veress et al. [99] recorded, from 1978 to 1987, accelerated lambing in a ﬂock of Hungarian Merinos. Flock statistics include the annual lamb crop (1.53 per ewe) and lamb survival to weaning (91%). 92. (30) Aboul-Naga [107] and Aboul-Naga et al. [108] reported very favourably on the breeding performance of substantial numbers of Rahmani and Ossimi fat-tailed Egyptian sheep in ﬁve ﬂocks that were bred every eight months over 18 successive mating seasons. 5066 and 3740 records on 1127 Rahmani and 830 Ossimi ewes were obtained; 35-day mating periods were recorded in September, May and January. Results’ conﬁrm the good breeding activity of both breeds on accelerated lambing. This system has produced 41.9 and 54.7% more lambs per ewe mated, compared with once-a-year matings. Both breeds performed better in September than in May,

CAB International 2010

G.L. Hunter

with the poorest performance after the January matings. Further details are provided in the abstract. (31) Two reports by Schoeman [109] and another by Schoeman and Burger [110] describe the production of three South African breeds, the Do¨hne Merino (DM), the S.A. Mutton Merino and the Dorper, in ‘an intensive accelerated lambing system’, concluding that ‘the Dorpers were the most efﬁcient breed for reproduction and growth, but overall efﬁciency was highest in the DM ewes when wool production was included’ – see para. 31. (32) Wheaten et al. [111] treated Finnish Landrace cross-bred and Columbia ewes with ‘type G controlled internal drug release devices’ (CIDR-G pessaries) and rams were joined with the ﬂock at pessary removal, in a complex trial, which the authors found difﬁcult in their abstract to describe so that I could follow it! They concluded, however, ‘that CIDR-G and/or sudden exposure to rams reduces lambing interval’. (33) Mavrogenis and Chimonides [112] managed a ﬂock of 145 Chios ewes (‘where’ is not stated in the abstract, the breed is not known to me and no address is given for the authors – perhaps in Greece, as the ewes were milked following weaning of the lambs); the ewes lambed three times in two years (the 3/2 group), while 119 similar ewes was managed to lamb once a year as a control (1/1) group. The ewes on both systems were kept indoors, but had some access to grazing – the ﬂock seems to have been managed similarly to the Ile de France ﬂock described in part 1 (paras 44–55), with the lambs being weaned at 42+3 days (except for those born in November and December, which were rebred in April) but the ewes were also milked following weaning, production being ‘computed from monthly test-day records following weaning and reﬂects commercial yield’. The ewes in both groups were hand-mated, after teasing twice daily, but there is no suggestion in the abstract that the ewes were teased for a period before mating commenced. It was concluded that ‘ewe fertility in the accelerated breeding system (3/2 group – 60%), although lower than that of the (1/1) control group, probably reﬂects seasonal fertility . . . and total milk production per ewe, although somewhat higher in the 3/2 system, was not signiﬁcantly different from the control ewes. It was concluded that ‘the use of proliﬁc breeds appears to be advantageous for accelerated breeding systems – this, of course, agrees with the views of Donald and Read [113]), although there was no suggestion in this abstract that crossing with the Finnsheep was being considered, so perhaps these authors may also consider the alternative suggested by the later work of Turner [114], viz. to select for improved proliﬁcacy (para. 23). (34) Diaz Infante et al. [115] described the reproductive performance of a ﬂock of +90 Rambouillet ewes in Mexico, which was ‘mated over a period of six years in January to March, May, July, September or November (sic) with an average lambing interval of 10 months . . . The annual lamb production of ewes lambing at 10-month intervals was 20% greater than that of ewes with a lambing interval of 12 months, (and)

g

127

it was concluded that a 10-month lambing interval had no adverse effects on ewe productivity’. (35) Schoeman and Botha [116] examined the reproductive traits of 917 ewes in an accelerated lambing trial ‘to examine whether reproductive traits could be improved by melatonin treatment’. Because melatonin treatment should be administered only to non-lactating ewes near the end of the breeding season, the need for accurate pregnancy diagnoses in commercial ﬂocks complicates the management and its use was therefore not recommended. 93. (36) Gordon [49], in the sixth of his 15 chapters, reviewed ‘More frequent lambings in sheep’; it should certainly be studied by those interested in this topic. (37) Bittante et al. [117] in a complex experiment, mated Finnsheep (FL) ewes to rams of an Alpine breed, the Lamon, the reciprocal cross, and also included purebred Lamon ewes (if I followed the abstract correctly). Their purpose was to evaluate the direct effect of breed and maternal breed, as well as individual and maternal heterosis, on litter traits and ewe productivity. ‘Most differences between FL and Lamon were in maternal genetic effects, whereas differences in the additive genetic effect were not signiﬁcant for most traits . . . Even though FL maternal effect decreased average lamb weight at birth (by 2.5 kg) and at weaning by 1.6 kg, the greater proliﬁcacy of FL-derived ewes resulted in a 6.7 kg heavier litter per ewe mated’ – I suspect ‘at weaning’ should be added to the end of that sentence. Other parameters are included in the abstract, and the conclusion was drawn that ‘the highest improvement in lamb production from crossing of Lamon sheep with FL was obtained by using FL sheep as the maternal line’. (38) Another example of the potential of the Romanov breed (see no. 6 in para. 86), together with Salz, Aragonese and F1 ewes is provided by Maria and Ascaso [118]. Litter size, lamb mortality at birth and lambing intervals were examined. 25 960 records were collected from 1986 to 1992; these included data from 5372 ewes that lambed three times every two years. The main factors, genotype, parity and lambing season and ﬁrst-order interactions were examined; ‘all the main factors had signiﬁcant effects on the three traits studied’. Lambing intervals of the four breed groups were 280, 255, 257 and 228 days, respectively, with that of the F1 ewes being signiﬁcantly shorter and of the Romanovs being signiﬁcantly longer than the others. Litter size increased with increasing parity and was signiﬁcantly larger in winter and spring, while lamb mortality and lambing interval decreased with increasing parity. The latter trait was longer in spring and summer than during autumn and winter. It was ‘concluded that Saltz sheep have good potential for production systems requiring improved proliﬁcacy or accelerated lambing’. (39) Kusakari and Ohara [119] reported results from feeding melatonin to Suffolk ewes on an accelerated lambing system on the northern Japanese island of Hokkaido ‘to determine the effect on increasing lamb production of

CAB International 2010

128

Animal Science Reviews 2010

accelerated lambing caused by melatonin feeding in the spring. ‘Forty-six Suffolk ewes . . . that were expected to give birth in Feb. were divided into two equal groups. (The animals on) the accelerated lambing system were fed melatonin for 60–90 days starting in late Mar, followed by ram introduction in May’ and were rebred after treatment. The ewes on the natural lambing system were isolated from rams until the autumn and so lambed only once in the year. Eighteen of 23 ewes (78%) ‘successfully produced three sets of lambs’ in two years, giving a ‘proliﬁcacy’ per year of 224% versus 159% for natural lambing. It was concluded that ‘the accelerated lambing system with melatonin feeding once every two years could result in greater lamb production than the natural lambing system’. (40) Urrutia et al. [120], in Mexico, stallfed 129 ‘Merino Rambouillet’ ewes, recording 349 matings of ewes, which were bred at eight-month intervals in summer (June – in-season), late winter (February – out-ofseason) or autumn (October – in season). Lambing rates from these matings did not differ (98.5, 88.9 and 96.5%, respectively), but litter size at birth (1.39 and 1.41 after the June and October matings) differed from the mean of 1.14 lambs born after the February mating). Lamb survival was similarly affected by season of mating (88.2 and 83.8 following breeding in June and October versus 74.1 after breeding in February). Weaning rate was also affected by breeding date, fewer lambs were weaned (84 per 100 ewes) from the February bred ewes, compared with 100 lambs per 100 ewes from those bred in October and 121 lambs per 100 ewes from the ewes bred in June. Both proliﬁcacy and the weaning weight of the lambs improved as ewe age increased; accelerated lambing increased lamb production by 37% at birth and by 28% at weaning. (41) Fisher [121] modelled three economic production systems over a seven-year period, for sheep that lambed either in spring, winter or in an accelerated lambing system, and determined production costs for each system using linear programming over seven years. The results showed that over the seven-year period, one ewe contributed $357 with spring lambing from a total of 9.80 lambs, $517 in a winter lambing system from 12.25 lambs, and $755 in an accelerated lambing system from 17.64 lambs. In addition, it was concluded that ‘although preference had shifted to larger lambs in the past ﬁve years, Christmas and Easter are by far the most important and preferred target markets for any sized lamb’. (42) Notter [122] reviewed and discussed ‘the potential for genetic improvement of reproduction in sheep in both annual autumn and accelerated lambing systems in the USA’ and found ‘genetic differences in seasonal breeding patterns . . . among and within sheep breeds . . . when coupled with careful management of breeding rams and the use of the ram effect, resulting in fertility levels of 50–70% in most seasons. Selection within breeds may further reduce seasonality, but requires carefully designed evaluation programmes to identify superior individuals’. (43) Goulet and Castonguay [123] synchronized (using progestagen

g

sponges) 79 primiparous cross-bred Lacaune ewes, for mating at 75 or 90 days postpartum, and concluded that their productivity improved with the increased interval after lambing. 94. (44) Keskin et al [124] offered no data, but pointed out the merits of increasing the numbers of lambs produced per ewe per year, by the use of accelerated lambing systems and brieﬂy describe the CAMAL and STAR systems developed at Cornell University. (45) Tosh et al. [125]: Data from 682 ewes and 58 sires over seven years produced 5043 observations or records for analysis. These were used to study fertility in four seasons of the year; although these are not actually deﬁned in the abstract. Assuming they are deﬁned in the presentation, it will presumably be possible, and easier, to summarize the ﬁndings from that source. The authors summarize their ﬁndings as ‘Seasons, year and age of ewe . . . inﬂuenced fertility, and least squares means for the four seasons are presented’ and they conclude that ‘results suggest (that there is) potential for expanding the breeding season by selecting ewes that cycle early in the year’. Brevity may be the soul of wit, but is not appreciated if overdone in an abstract. (46) Banos et al. [126] estimated and presented genetic parameters for fertility and proliﬁcacy in ewes of two age groups managed in a frequent lambing system and also included estimates of the interval between lambings. (47) Susˇic [127] aimed to describe procedures by which frequency of lambing, as well as the total annual production of lambs, can be increased. (48) Susˇic and Sˇtokovic [128] presented the results of a survey of data on the breeding season in breeds in Croatia (that reach high fertility by means of a short inter lambing period as well as large numbers of lambs per litter, with special attention being paid to the Finnsheep and Romanov breeds’ and their efﬁciency in various exploitative systems with frequent lambing. The need for ‘exploring the breeding season of sheep bred in Croatia and the possibility of extending that season by selection, is emphasised’. (49) Koyuncu [129] determined ‘the effects of . . . lambing [three times in two years, in Jul (summer), Mar (early spring) and Nov (late autumn/early winter)] in Kivircik ewes in Pakistan, and presents data on lambing rates, litter size at birth, and lamb survival at weaning, and considered that the breed had ‘good potential for production systems requiring improved proliﬁcacy or accelerated lambing management’. (50) Menegatos et al. [130], in a small trial using only 15 crossbred ewes (KaragounikaMytilene, both are presumed to be breeds known in Greece – apologies for my ignorance!), studied the productivity of the ewes in two lambing seasons, ‘as part of an accelerated lambing system of four lambings in three years . . . Oestrus synchronisation was induced by the ‘ram effect’ ’ and lactation length was limited to eight weeks. A substantial amount of data is presented and it was concluded that for the two seasons selected, ‘increased ewe ovarian activity, as estimated by serum progesterone levels reﬂects the stage of the breeding season of the ewes.

CAB International 2010

G.L. Hunter

Thyroid function showed the typical seasonal pattern, irrespective of the ewe’s productive stage’, and also that ‘summer and spring lambing seasons as a part of an accelerated lambing system of four lambings in three years could result in increased ewe productivity (in terms of proliﬁcacy, lamb birth weight and weaning weight) – being greater for the spring lambing season’. And (51) is a report by El-Saied et al. [131], who used data from 6318 lactations for 1391 Spanish Churra ewes in ten ﬂocks, that ‘were used to study parameters of lifetime traits in annual and accelerated lambing systems’. The conclusion was that ‘management to ensure high fertility, especially among young ewes, would be of more importance than removing non-lambing ewes’. 95. (52) In another reference from Eastern Europe, this time from the Lithuanian Veterinary Academy, Zapasnikiene [132] analysed data collected over seven years from 1999, in various ﬂocks of the local coarse wooled (LCW) and Lithuanian Blackface (LB) sheep. The seasons of lambing differed between the two breeds, with 55.4% of the LCW lambing during spring and less than 1% during winter, while 46.7% of the LBs lambed during winter. The differences in lambing rate between the LCW in the different seasons were also substantial; those lambing in spring were 30% more fertile (proliﬁc) than those lambing in summer and autumn, while the ‘fertility’ of the LB ewes was almost the same in different seasons (1.4 lambs per ewe). The progeny of LCW ewes were heaviest in summer, while the LB lambs born in autumn were heaviest. ‘More frequent lambing had (a) negative effect on the litter size of both breeds and (on) the progeny weight of the LCW sheep. However, the lambs of LB sheep with the lambing frequency of eight months were heavier and gained better until weaning than the lambs of ewes with the 12-month lambing frequency’. The abstract appears to have omitted much that would be of interest to those interested in frequent lambing, but this work should be considered together with the ﬁnal report to be listed in this paragraph, viz. (53) Zapasnikiene and Nainiene [133]: a ﬂock of the endangered LCW breed was established in 1995; currently (2008) the ﬂock numbers about 100 and has been ‘evaluated for its biological and farming qualities’. An analysis was performed in 2008 of the ﬂock’s ‘exclusive traits’; including ‘aseasonal’ heat and the traits were evaluated for the period 1995–2007. In 1995–96, the ﬂock lambed only once, but in the period 1997 to 2001 the ewes lambed every six to eight months. Two lambs were born per ewe during both periods, but the clarity of the abstract deteriorated at this point and became incomprehensible to me – a pity that some language assistance was not sought for the presentation of what seems to be an interesting project! 96. Interesting and important conclusions have been drawn by Stewart et al. [134–136], and also in a personal communication to me in 2009), following her observations of the suckling behaviour of calves and of the diurnal pattern of secretion of melatonin. She observed (between

g

129

August and October), in suckling beef cows at Cedara in Natal (latitude +29–30 ), that an increase in melatonin secretion by the pineal gland began at sunset and continued after dark, and so increased in concentration during the hours of darkness to a peak at about midnight. Furthermore, because the hormone inhibits the onset of oestrus, perhaps by interacting with endorphins and encephalins, she concluded that suckling late at night interfered with the mother’s resumption of oestrus and delayed rebreeding. Eventually, the cow suckled her calf sufﬁciently early in the evening not to cause this hormonal interference, and breeding cycles were resumed. Clearly there is much to be inferred from her observations concerning the timing of the weaning of suckling lambs from ewes that are to be rebred promptly in a frequent lambing programme. For example, it may prove to be highly advantageous, for two weeks before a mating period is due to begin and when teaser rams are about to be introduced, temporally each evening to remove suckling lambs from their mothers for the night. They may be returned to the mothers next morning. This temporary separation, of course, prevents late night suckling which should encourage the resumption of oestrus, ovulation and fertility. In support of this suggestion, Gordon and Stiegmann [137] have pointed out that ‘circumstantial evidence is accumulating from various human and nonhuman primate studies to suggest that night-time suckling may be more important (for the lactational inhibition of breeding cycles) than daytime suckling’ and undertook to determine ‘the natural suckling pattern of ewes and their lambs throughout day and night once a week for the ﬁrst seven weeks of lactation’. Observations were made of the suckling behaviour of Merino ewes with single lambs born in May (late autumn) in Australia, in a 3030 m. grassy paddock, using an image intensiﬁer at night. ‘Suckling frequencies fell from 36+5.5 at 1–2 weeks of age to 14.3+2.78 suckles per 24 h at 6–7 weeks of age’; the duration of suckling bouts declined with age of lamb; and suckling frequencies per hour were greater during daylight (P < 0.05). It was suggested that ‘the declining frequency and duration of suckling bouts with age probably accounts for the waning inﬂuence of lactation on the reproductive state of the ewe, and that the maintenance of a critical ‘interbout’ interval may ﬁrst break down at night. It was a pity that the authors did not publish the times during the night at which suckling took place and whether this changed with increasing age of the lamb, as with the observations of Stewart et al. to which reference was made at the beginning of this paragraph. And, strangely enough, Gordon and Steigmann in the end did not record, or at least did not publish their records of the resumption of oestrous behaviour in the ewes, even though this appeared to be the reason for initiating the study in the ﬁrst place – but suggest, nevertheless, ‘that the decline in suckling frequency and duration with age of the lamb provides the mechanical basis for the waning inhibitory effects of lactation on reproduction in the ewe’.

CAB International 2010

130

Animal Science Reviews 2010

The temporary weaning for the night, suggested above, would certainly result in much noise, and perhaps some temporary inhibition of breeding cycles in the mothers, due to stress – there is still much to learn in this area. 97. Subsequently, Nowers et al. [138] studied the effects of melatonin implants, ﬂushing and teasing on reproduction of spring-mated Do¨hne Merino ewes, and reported that, following melatonin treatment, there was ‘a signiﬁcant reduction in the spread of dates of ﬁrst oestrus during the mating period following melatonin treatment’, and that the treatment also advanced the date of the period of ﬁrst oestrus. The authors suggest that ‘the ability to advance the mean date of ﬁrst oestrus and to condense the mating period round this date would be a useful attribute – in some circumstances . . . a less tight synchronisation, as in the melatonin-treated ewes in this trial, may be preferable. The authors also claim that it would be more economical to treat with melatonin (@R7 – < $1 per treatment) than with the combined ﬂushing and teasing treatment – a pity, perhaps that the costs of these two were combined? 98. Haresign [139], however, offered the following explanation in his introduction, that ‘the principle roˆle of melatonin is to advance the period of sensitivity to the ‘ram effect’ and (uses) ram introduction to promote a greater degree of synchrony in mating’ (my italics). Can he and Stewart both be correct? Both sound plausible. It is apparent that we have much to learn about melatonin and its important role. Haresign went on to report on two trials designed to investigate the ability of melatonin implants to induce early breeding in lowland sheep ﬂocks in the UK, and concluded ‘from the mating patterns observed (that his) suggested optimum melatonin implantation dates for Suffolk crossbreds in his experiment (would be from) mid-May to mid-June (early summer) and, for his Mule ewes, from mid-May to early July (a slightly longer period lasting past mid-summer’s day). Haresign [139] added that ‘Melatonin treatment also successfully increased mean litter size, but (that) the magnitude of this varied across ﬂocks. Useful diagrams illustrate the cumulative percentage of ewes mated following ram introduction (these are very similar to the trend previously reported in South Africa by Parson and Hunter [140]). Haresign concluded that ‘appropriately timed treatment of lowland ewes (with melatonin) can be used to promote early breeding with an additional lift in litter size’. In a second paper, Haresign [141] reported that ‘increasing the interval from treatment to ram introduction from four to six weeks, was associated with a signiﬁcant, progressive reduction in the time from ram introduction to mating, as well as a reduction in the spread of mating across the group’. Furthermore, ‘melatonin caused an increase in ovulation rate (from 0.44 to 0.48 ovulations per ewe, although the increase in litter size (from 0.19 to 0.36) was more variable’. 99. Bettencourt et al. [142] suggested a hypothesis that ‘pregnancy success could be improved in early postpartum

g

ewes by prolonging the lifespan of the corpus luteum via active immunisation against prostaglandin F2a’. The authors point out that ‘theoretically, ewes could lamb twice a year if they conceived within 35–40 days post partum’. The treatments administered in a series of experiments are described and the authors conclude from the results that ‘premature luteal regression was not the reason for failure of occurrence of pregnancy’. They acknowledge that the results were unexpected, and concluded also that ‘Immunisation against the prostaglandin was effective in blocking ovulation, but not in inhibiting oestrous behaviour or the formation of persistent luteal tissue. Treatment of immunised ewes with exogenous prostaglandin restored the ability of ewes to ovulate, providing further evidence for the involvement of prostaglandin in ovulation. The ability of oestradiol to induce luteolysis in immunised ewes was associated with the presence of uterine ﬂuid.’ 100. Notter and Cockett [143] developed a population of sheep with an extended breeding season ‘through selection for fertility in spring matings, and (the ﬂock) provides opportunities for further study of candidate genes inﬂuencing seasonal breeding. In particular, the Melatonin receptor 1a gene is polymorphic in many sheep breeds and appears to inﬂuence a number of seasonal reproductive responses . . . Mutations in these clock genes have been identiﬁed and shown to inﬂuence circadian periodicities and reproductive patterns in golden hamster and mouse. In sheep, expression of clock genes in the suprachaismatic nucleus and pars tuberalis (PT) suggests that ‘calendar’ cells in the ovine PT play a role in maintaining circannual rhythms. Thus the various clock genes represent potentially important candidate genes that may be involved in control of seasonal breeding’.

Final Conclusions and Recommendations 101. Accelerated Lambing of ewes with long breeding seasons (three lambings in two years) is entirely feasible, but lambing more frequently (twice a year) is considered to be less achievable at present. It is clear that whichever system is favoured, there are some basic requirements to be met. The ewes should be proliﬁc and should be encouraged in every possible way to conceive promptly so that they breed regularly at +8-month intervals. Replacements for the ﬂock (both sexes) should be selected to promote these traits. There are a number of breeds that have been shown to have these abilities; cross-breds incorporating the genes of the Finnsheep may not be essential, in spite of the reasons that led to the acquisition by ABRO of Finnsheep in the ﬁrst place ([113] – see para. 28); Turner’s review (para. 23) suggests that the topic should be ‘revisited’: Perhaps this trait can after all be improved at a sufﬁciently fast rate in purebred ﬂocks that do not contain Finnsheep genes – this needs reassessment. If this path is to be followed, there is likely to be a need to

CAB International 2010

G.L. Hunter

run a control ﬂock to be certain that real genetic progress is being achieved; this should be seriously considered also in commercial undertakings. The carcase traits of the ewes that are to lamb frequently are less important, at least to begin with, since the lambs to be marketed will be sired by rams that are bred for the purpose – the top ewes in the ﬂock will be selected and bred primarily to lamb frequently, such as the ‘ALL STAR’ ewes in the Cornell system. Less productive ewes in the ﬂock will be the producers of lambs for the market and, once identiﬁed, should be mated only by a specialized mutton breed. The FLASH system (para. 61) from Stellenbosch has promise, perhaps particularly under grazing conditions, assuming the need for ‘super teasers’ can be met. The calendars of STAR and FLASH are compared in Table 1. Neither system requires changes in photoperiod nor treatment with hormones; housing of the ﬂock is not necessary for frequent lambing, but excellent management and shepherding are important in both systems. 102. The two systems differ with regard to the need for control by the ﬂock manager. With respect to the frequency of mating and lambing – is it really necessary to conform to the pattern dictated by the star diagram? Mating periods every two months seem to be excessive – is it just that some ﬂock managers need and welcome the discipline imposed by the Star diagram? On the other hand, judgement should probably be made on the performance achieved by the ﬂocks, rather than on the demands that are made on managers. Reports of actual achievements in STAR ﬂocks are not common, but see para. 85. Smith [72] has reported on the seasonal effects, stating that ‘because there are ﬁve evenly spaced breeding periods a year . . . not all (of them) fall within the . . . (autumnal) decreasing photoperiod’, so that ‘genetic selection has been used to develop a ﬂock that will breed year-round’. This may be because the system was conceived by placing insufﬁcient emphasis on the importance of using ewes with a ‘long breeding season’ or of the factors (such as the Ram Effect) that promote this trait, in spite of the fact that the Dorset was selected to meet this need. Was the Dorset’s reputation sufﬁcient? Perhaps individuals that were crossed with the Finn to establish the basic stock, were insufﬁciently proven in this respect? As Smith put it, the polled Dorset was selected and used because of its reputation for both ‘mothering ability and a relatively long breeding season’. Was this done because of the breed’s reputation only, or were the foundation ewes actually performance tested in some way? – I have found no suggestion of this – and Hogue [66] makes no mention of this need in his chapter on Frequent Lambing Systems, of which I have already been critical. If this important trait was not evaluated, but assumed to be present in the foundation stock at the start of the project, it may now be necessary to select young stock from the progeny of ewes that have actually performed as required. This approach with regard to such a vitally important trait was risky, in my opinion; and the next question should be, ‘Should

g

131

there be a ‘control’ ﬂock against which selection progress in this trait can be measured?’ 103. The same questions should, of course, be asked of a ﬂock manager who is using FLASH, as I was with the Ile de France ﬂock. It can be interesting, but unhelpful to acquire knowledge and perspective during or after an experience – I certainly wish I knew, while managing the Ile de France ﬂock, what I have learnt during the preparation of this review. The performance of the ﬂock left much to be desired (due to ignorant management, rather than the inability of the breed) and could have been greatly improved had I kept up with developments, particularly about the selection and use of teaser rams, or had actually used goats as teasers. (The possible importance of this suggestion has grown for me since the ﬂock was sold – I learnt of King’s work only recently – see para. 80.) Regarding the knowledge required by producers, 14 references are listed (duplicated) in both parts of this review. Of the 236 listed altogether, 73 (33.8%) had been published by 1974, the year in which the Ile de France ﬂock was founded, and a further 24 (10%) of those listed were published by 1980, before the ﬂock was sold, but ‘farmers’ did not and still do not have access to the University Library! This only serves to emphasize the need for researchers to undertake extension and advisory work amongst farmers, and for farmers to ask questions and seek answers at every opportunity, to keep up with developments in research. 104. I cannot claim that every relevant item in the literature published since my 1968 reviews, has been noted in this one. Future research should prioritize the role of melatonin; although many studies of melatonin have been found via CABI listings, its role during the ewe’s puerperium has not yet, to my knowledge, received sufﬁcient attention. Gordon’s [13, 49] suggestion (already mentioned, par. 64) that ‘melatonin secretion by the ewe may ‘prime’ the unborn lamb to be responsive to the light patterns to which the ewe has been exposed’, may also prove rewarding. More recently, Mateescu et al. [144] have compared the association between a melatonin receptor gene in Dorset and cross-bred ewes that expressed different reproductive performances (intervals to ﬁrst lambing and between ﬁrst and second lambings); this research will no doubt continue and be of interest in frequent lambing projects. Research institutions will presumably continue to breed ﬂocks that lamb frequently – but, with guidance from researchers on speciﬁc aspects of the enterprise, especially the fundamental physiology, genetics and sound economics, it is important that commercial undertakings are also encouraged to do so. There is still much to be learned about the effects of the photoperiod in the breeding of sheep, particularly the effects of high latitude on the ewes’ responses – with controlled and innovative lighting facilities, the effects of moving ewes (or other animals) from one latitude to another could be investigated without them actually leaving the property. Marshall’s [2] observations (paras 40 and 41) should be

CAB International 2010

132

Animal Science Reviews 2010

conﬁrmed, with recognition of the effects of ram introduction on the reproductive performance of transported ewes. Finally, a mild protest against the growing use of the phrase ‘under (instead of ‘on’) accelerated lambing’ – many of the culprits are listed below!

Acknowledgements I wish to dedicate this review and record my gratitude to Mr Neil MacGillivray (1920–2008), of Gartmore in the Howick district of KwaZulu-Natal, for his generous support, interest and encouragement at all times, both prior to but especially during the ‘Ile de France period’. I must also express my appreciation to Professor C.W. Cruywagen, Head of the Department of Animal Science, University of Stellenbosch, for enabling me to use the facilities of the University Library, as well as for his helpful suggestions. I am much indebted to Mr Pieter Du Plessis, Faculty Librarian, Agricultural Sciences, for his efﬁcient assistance in meeting my requests, as well as to his predecessors Mrs Carine Tymbios and Mrs Hanlie Strydom. I particularly wish to acknowledge the practical and constructive suggestions at an early stage of the draft, of Mr E.M. Hughes, retired farmer and Dorper breeder, and I greatly valued the assessment and suggestions of Professor Geoff Antrobus of the Department of Economics and Economic History, Rhodes University, Grahamstown, in the Eastern Cape Province. The meticulous checking at a late stage and the suggestions of Professor P.C. Belonje were much appreciated. Dr Gretha Snyman of the Agricultural Research Institute, Middelburg, Cape, S. Africa, was most helpful in putting me in touch with the publications from that institution on the use and effects of Ronderib Afrikaner rams. I may add that the production of this document on a computer, instead of the three copies produced laboriously on a small portable typewriter, which was required 40 years ago for the previous review, was a welcome development, but the practical advice and guidance of Mrs Karen Burns in subduing the apparent vagaries of Windows Vista, was willingly provided and greatly appreciated, as was the comprehensive abstracting service of the CABI – efﬁcient and satisfactory in every respect, not to mention the advice and necessary encouragement given during the preparation of this document, by ‘Mr Editor’ of Animal Breeding Abstracts, Dr M. Djuric.

2. Marshall FHA. On the changeover in the oestrous cycle in animals after transference across the equator, with further observations on the incidence of the breeding seasons and the factors controlling periodicity. Proceedings of the Royal Society B 1937;122:413–28 [ABA abstr.] {63} 3. Yeates NTM. Inﬂuence of variation in length of day upon the breeding season in sheep. Nature (London) 1947;160:429 [ABA abstr.] {63} 4. Yeates NTM. The breeding season of the sheep with particular reference to its modiﬁcation by artiﬁcial means using light. Journal of Agricultural Science, Cambridge 1949;39:1 [ABA] {63, 65} 5. Hunter GL. Increasing the frequency of pregnancy in sheep. 1. Some factors affecting rebreeding during the post-partum period. ABA 1968;36:347–78 {62, 64, 71} 6. Hunter GL. Increasing the frequency of pregnancy in sheep. 2. Artiﬁcial control of rebreeding, and problems of conception and maintenance of pregnancy during the post-partum period. ABA 1968;36:533–53 {62, 66} 7. Goot H. Effect of light on spring breeding of Mutton Merino ewes. Journal of Agricultural Science, Cambridge 1969;73:177–80 [ABA] {62, 64} 8. Ducker MJ, Bowman JC. Photoperiodism in the ewe. 5. An attempt to induce sheep of three breeds to lamb every eight months by artiﬁcial daylength changes in a non-light-proofed building. Animal Production 1972;14:323–34 [ABA 40, abstr.3262] {62, 64} 9. Ducker MJ, Bowman JC. Effect of artiﬁcial daylight changes on the reproductive rate of sheep. In: Marai IFM, Owen JB, editors. Nutrition of Housed Sheep. Chapter 16 in New Techniques in Sheep Production. Butterworths, London; 1974. [ABA Book Reviews, p.] {62} 10. Ducker MJ, Bowman JC. Effect of artiﬁcial daylight changes on the reproductive rate in sheep. Veterinary Records 1974;95:96–8 [ABA 42, abstr. 4900] {62} 11. Hunter GL, van Aarde IMR. Inﬂuence of age of ewe and photoperiod on the intervals between parturition and ﬁrst oestrus in lactating and non-lactating ewes at different nutritional levels. Journal of Reproduction and Fertility 1975;42:205–12 [ABA 43, abstr. 2367] {62} 12. Shelton M, Hulet CV, Gallagher JR. Inﬂuence of season, location and source of ewe on estrus and ovulation rate of Rambouillet ewes. In: Sheep and Angora Goat, Wool and Mohair Research Report. Texas Agriculture Experiment Station, College Station, TX; 1973. p. 26–28 [ABA, 44, abstr. 3775] {64} 13. Gordon I. More frequent lambings in sheep. In: Controlled Reproduction in Sheep and Goats. CABI Publishing, Wallingford, Oxon, UK; 1997. Chapter 5 [ABA abstr.] {64, 104} 14. Robinson JJ, Frazer C, Mc Hattie I. The use of progestagens and photoperiodism in improving the reproductive rate of the ewe. Annales de Biologie Animale, Biochemie, Biophysique 1975;15:345–52 [ABA 44, abstr. 1714] {64}

References (2) ABA=Animal Breeding Abstracts The paragraph(s) in which references are quoted are recorded after each {in parentheses} 1. Yeates NTM. Daylight changes. In: Hammond J, editor. Progress in the Physiology of Farm Animals, Volume 1. Butterworths Scientiﬁc Publications, London; 1954. Chapter 8 {63}

g

15. Hunter GL, Belonje PC, van Niekerk CH. Effects of season, suckling and teasing on post-partum interval to ovulation in ewes. Proceedings of the South African Society for Animal Production 1970;9:179–81 [ABA 39, abstr. 4779] {65} 16. Pope WF, McClure KE, Hogue DE, Day ML. Effect of season and lactation on postpartum fertility of Polypay, Dorset,

CAB International 2010

G.L. Hunter St Croix and Targhee ewes. Journal of Animal Science 1989;67:1167–74 [ABA abstr.] {66} 17. Fahmy MH. The accumulative effect of Finnsheep breeding in crossbreeding schemes: Ewe productivity under an accelerated lambing system. Canadian Journal of Animal Science 1990;70:967–71 [ABA abstr.] {65}

133

31. Newton GR, Edgerton LA. Effects of season and lactation on luteinizing hormone secretion in postpartumewes. Theriogenology 1989;31:885–94 [ABA abstr.] {70} 32. Peters AR, Lamming GE. The lactation anestrus in farm animals. Oxford Reviews of Reproductive Biology 1990;12:245–88 [ABA abstr.] {71}

18. Fahmy MH, Lavallee D. Productivity of Polypay, Dorset and PolypayDorset Ewes under two accelerated breeding systems. Small Ruminant Research 1990;3:269–281 [ABA abstr.] {66}

33. Martin GB, Oldham CM, Cognie´ Y, Pearce DT. The physiological responses of anovulatory ewes to the introduction of rams – a review. Livestock Production Science 1986;15:219–47 [ABA abstr.] {71, 74}

19. Van Wyk LC, Van Niekerk CH, Hunter GL. Inﬂuence of exogenous hormones and season of lambing on uterine involution in the sheep. Agroanimalia 1972;4:77–82 [ABA 42, abstr. 3227] {67}

34. Wright PJ, Geytenbeek PE, Clarke IJ. The inﬂuence of nutrient status of post-partum ewes on ovarian cyclicity and on the oestrous and ovulatory responses to ram introduction. Animal Reproduction Science 1990;23:293–303 [ABA abstr.] {71, 73, 74}

20. Schirar A, Meusnier C, Paly J, Levasseur MC, Martinet J. Resumption of ovarian activity in post-partum ewes: role of the uterus. Animal Reproduction Science 1989;19:79–89 [ABA abstr.] {67} 21. Mallampati RS, Pope AL, Casida LE. Effect of suckling on post-partum anestrus in ewes lambing in different seasons of the year. Journal of Animal Science 1971;32:673–677 [ABA 39 abstr. 4785] {67, 68} 22. Hayder M, Ali A. Factors affecting the postpartum uterine involution and luteal function of sheep in the subtropics. Small Ruminant Research 2008;79:174–8 [ABA abstr.] {67} 23. Hunter GL. Is there a lactation anoestrus in the sheep? South African Journal of Animal Science 1971;1:55–7 [ABA 40, abstr. 1963] {67}

35. Schoeman SJ. Production parameters for Do¨hne Merino sheep under an accelerated, intensive lambing system. South African Journal of Animal Science 1990;20:174–9 [ABA abstr.] {71, 90} 36. Fogarty NM, Hall DG, Dawe ST, Atkinson W, Allan C. Management of highly fecund ewe types and their lambs for 8-monthly lambing. 1. Effect of lamb weaning age on ewe reproductive activity in spring. Australian Journal of Experimental Agriculture 1992;32:421–8 [ABA abstr.] {72} 37. Pearce DT, Oldham CM. The ‘Ram Effect’, its mechanisms and application to the management of sheep. In: Lindsay DR, Pearce DT, editors. Reproduction in Sheep. Australian Wool Corporation, Canberra; 1984. p. 26–34 [ABA abstr.] {72}

24. Land RB. The incidence of oestrus during lactation in Finnish Landrace, Dorset Horn and Finn–Dorset sheep. Journal of Reproduction and Fertility 1971;24:345–52 [ABA 39, abstr. 3464] {68}

38. Vosloo LP, Hunter GL, Carstens JdeW. Inﬂuence of level of nutrition during gestation and lactation on post-partum interval to ovulation and rebreeding of ewes. Proceedings of the South African Society for Animal Production 1969; 8:145–6 [ABA 38 abstr. 2677] {73}

25. Pelletier J, Thimonier J. Comparison of the induced preovulatory LH discharge in lactating and dry sheep during seasonal anoestrus. Journal of Reproduction and Fertility 1973;33:310 [ABA 42, abstr. 207] {69}

39. Hunter GL, Lishman AW. Effect of the ram early in the breeding season on the incidence of ovulation and oestrus in sheep. Proceedings of the South African Society for Animal Production 1967;6:199–201. [ABA abstr.] {72}

26. Hunter GL, van Aarde IMR. Inﬂuence of season of lambing on post-partum intervals to ovulation and oestrus in lactating and dry ewes at different nutritional levels. Journal of Reproduction and Fertility 1973;32:1–8 [ABA 41, abstr. 1677] {69}

40. Hunter GL. The role of the ram when synchronising the mating of sheep with progestagens. Proceedings of the South African Society for Animal Production 1969;8:143 [ABA abstr.] {74}

27. Shevah Y, Black WJM, Carr WR, Land RB. The effect of lactation on the resumption of reproductive activity and the pre-ovulatory release of LH in Finn x Dorset ewes. Journal of Reproduction and Fertility 1974;38:369–78 [ABA 42, abstr. 5353] {69} 28. Peclaris GM. Effect of suppression of prolactin on reproductive performance during the postpartum period and seasonal anoestrus in a dairy ewe breed. Theriogenology 1988;29:1317–26 [ABA abstr.] {69} 29. Fuentes VO. Effect of naloxone, nalbuphine, progesterone and pregnant mare’s serum gonadotrophin on the sexual behaviour of ewes. Veterinary Records 1989;24:274–76 [ABA abstr.] {70} 30. Schirar A, Cognie Y, Louault F, Poulin N, Levasseur MC, Martinet J. Resumption of oestrous behaviour and cyclic ovarian activity in suckling and non-suckling ewes. Journal of Reproduction and Fertility 1989;87:789–94 [ABA abstr.] {70}

g

41. Chesworth JM, Tait A. A note on the effect of the presence of rams upon the amount of luteinizing hormone in the blood of ewes. Animal Production 1974;19:107–10 [ABA 42, abstr. 4356] {74} 42. Louw BP, Marx FE, Yeates CD. The inﬂuence of vasectomised rams on the lambing pattern of spring-mated Corriedale ewes. South African Journal of Animal Science 1974;4:167 [ABA 43, abstr. 3473] {74} 43. Signoret JP. [Effect of the presence of the ram on reproductive mechanisms in the ewe.] In 1e`res journe´es de la recherche´ ovine et caprine, 2–3 et 4 de´cembre 1975. Tome II: Espe`ce ovine; 1975. p. 303–313 [ABA 44 abstr. 3776] {74} 44. Signoret JP, Fulkerson WJ, Lindsay DR. Effectiveness of testosterone-treated wethers and ewes as teasers. Applied Animal Ethology 1982;9:37–45 [ABA abstr.] {74} 45. Oldham CM, Pearce DT. Mechanism of the ram effect. Proceedings of the Australian Society of Reproductive Biology 1983;15:72 [ABA abstr.] {74}

CAB International 2010

134

Animal Science Reviews 2010

46. Lishman AW, Stielau WJ, Dreosti IE, Stewart AM, Botha WA. Plasma luteinising hormone levels in ewes failing to exhibit oestrus during lactation and in ewes isolated from rams. South African Journal of Animal Science 1974;4:45–9 [ABA 42, abstr. 4907] {74} 47. Lishman AW. The seasonal pattern of oestrus among ewes as affected by isolation from and joining with rams. Agroanimalia 1969;1:95 [ABA 39, abstr. 2013] {74} 48. Fukui Y, Akaike M, Ise K, Kobayashi K, Hitoshi O. Effect of progesterone pre-treatment methods associated with ‘ram effect’ on estrus induction and lambing rate in seasonally anoestrous ewes. Japanese Journal of Animal Reproduction 1988;34:204–208 [ABA abstr.] {74} 49. Gordon I. Advancing the breeding season. In: Controlled Reproduction in Sheep and Goats. CABI Publishing, Wallingford, Oxon, UK; 1997. Chapter 6. [ABA abstr.] {75, 93} 50. Copenhaver JS, Carter RC. Early weaning and multiple lambing. In Livestock Research Report 1965–66 Virginia Agriculture Experiment Station; 1966. p. 54–7. Very early weaning (conﬁnement rearing of lambs), rebreeding ewes for multiple lambing. In 1967–68 Livestock Research Report. Virginia Polytechnic Institute, No. 126. p. 61–9 [ABA 37, abst. 455] {76} [Google] 51. Copenhaver JS, Carter RC. Accelerated lamb production. In Livestock Research Report, Virginia Agricultural and Experimental Station; 1972–73. p. 41–44 [ABA abstr.] {76} 52. Carter RC, Copenhaver JS. Performance of ewe breeds and crosses under accelerated lambing. In Livestock Research Report 1973;1972–73:94–6 [ABA abstr.] {75} 53. Notter DR, Copenhaver JS. Performance of Finnish Landrace crossbred ewes under accelerated lambing. 1. Fertility, proliﬁcacy and ewe productivity. 2. Lamb growth and survival. Journal of Animal Science 1980;51:1033–50 [ABA abstr.] {76} 54. Blaxter KL. The nutrition of ruminant animals in relation to intensive methods of agriculture. Proceedings of the Royal Society London, Series B, Biological Science 1973;183: 321–36 [ABA, abstr.] {77, 78, 79, 80} 55. Robinson JJ, Frazer C, Gill JC. Preliminary observations on the performance of Finnish LandraceDorset Horn ewes in an intensive system. In: Proceedings of the 54th Meeting of the British Society of Animal Production; 1972. p. 132 (Abstr.) [ABA abstr.] {77} 56. Robinson JJ. Some aspects of ewe nutrition. Veterinary Records 1973;92:602–6 [ABA, abstr.] {77} 57. Robinson JJ. Intensifying ewe productivity. Proceedings of the British Society of Animal Production 1974;3:31–40 [ABA 43, abstr. 2955] {77, 79} 58. Robinson JJ, rskov ER. An integrated approach to improving the biological efﬁciency of sheep meat production. World Review of Animal Production 1975;11:63–75 [ABA abstr] {73, 80}

61. Tempest WM. Management of the frequent lambing ﬂock. In: Haresign, W, editor. Sheep Production. London, Butterworths; 1983. Chapter 24. [ABA Review] {78} 62. Tempest WM. The development and performance of a frequent lambing system in the U.K. In 32nd Annual Meeting of the European Association for Animal Production; 1981. p. 13 [ABA abstr.] {78} 63. Lees JL. The reproductive pattern and performance of sheep. Outlook on Agriculture 1969;6:82–8 [ABA abstr.] {79} 64. King PR. Stimulation of sexual activity in Merino rams. Karoo Agrie 1994;6:1–2 [ABA abstr.] {80} 65. King PR, Coetzer WA. Effect of Ronderib Afrikaner rams on the plasma LH concentration in Merino ewes during the oestrous cycle. In Proceedings of the 34th Congress, South African Society of Animal Society; 1995. Available from: URL: http://gadi.agric.za (publications)] [ABA abstr.] {80} 66. Hogue DE. Frequent lambing systems. In: Marai IFM, Owen JB, editors. New Techniques in Sheep Production. Butterworths, London; 1987. p. 57–63 [ABA review] {81, 102} 67. Hogue DE, Magee BH, Travis HF. Feed ewes to produce more than two lambs, or wean lambs to dry diets at 10 days of age? Proceedings of the Cornell Nutrition Conference 1978; October/November:141–3 [ABA abstr.] {81, 102} 68. Snyder DP, Milligan RA. A comparative economic analysis of the STAR accelerated and annual lambing systems. A.E. Research Cornell University 87-12; 1987. 33 pp [ABA abstr.] {82} 69. Matthews D. The role of private practitioners in accelerated lambing. Veterinary Clinics of North America, Food Animal Practice 1990;6:585–95 [ABA abstr.] {83} 70. Magee B. STAR accelerated lambing system. In: Symposium on Diseases, Small Ruminant, American Association of Small Ruminant Practice, Oregon; 1990. p. 47–56 [ABA abstr.] {83} 71. Lewis RM, Notter DR, Hogue DE, Magee DH. Ewe fertility in the STAR accelerated lambing system. Journal of Animal Science 1996;74:1511–22 [ABA abstr.] {82, 83} 72. Smith MC. Veterinary experiences with the Cornell STAR system of accelerated lambing. Small Ruminant Research 2006;62:125–8 (www.elsevier.com/locate/smallrumres) [ABA abstr.] {84, 102} 73. Thonney ML. STAR Management. 2005. Available from: URL: http://www.sheep.cornell.edu/sheepmanagement/ breeding/star/index.html (Quoted by Smith (2006) {84} 74. Dodson RE, Godfrey RW. The use of Dorper crossbred ewes in an accelerated lambing and extensive management system in the tropics. Journal of Animal Science 2005;83:343 [ABA abstr.] {85} 75. Godfrey RW, Collins JR, Hensley EL, Buroker HA, Bultman JK, Weis AJ. Hair sheep performance in an accelerated lambing and extensive management system in the tropics: a ten year summary. Journal of Animal Science 2005;83:13 [ABA abstr.] {93}

59. Robinson JJ. Photoperiodic and nutritional inﬂuences on the reproductive performance of ewes in accelerated lambing systems. In: 32nd Annual Meeting of the European Association for Animal Production; 1981. p. 10 [ABA abstr.] {77}

76. Godfrey RW, Dodson RE, Vinson MC, Driscoll RC. Production traits of Dorper crossbred ewes in an accelerated lambing system in the tropics. Journal of Animal Science 2007;85:39. Suppl. 2 [ABA abstr.] {93}

60. Haresign W, editor. Sheep Production – Nottingham Easter School Proceedings 35. Butterworths, London; 1983. [ABA Review] {77}

77. Eyal E, Folman Y, Morag M. Lamb production in frequently lambing dairy sheep. World Review of Animal Production 1973;9:64–69 [ABA abstr.] {87}

g

CAB International 2010

G.L. Hunter 78. Espejo Diaz. Planning reproduction in sheep ﬂocks. II. Use of the system of three lambings in two years. ITEA, Revista de la Asociatio´n Interprofesional para dos Desarollo Agrario 1975;6:111–25 [ABA 45, abstr. 262] {87} 79. Jennings J, Lawlor MJ. Towards increased lambing frequency. Farm and Food Research 1976;7:56–7 [ABA 44, abstr. 5734] {87} 80. Flanagan S, Quirke JF. Accelerating the frequency of lambing. Research Report, Agricultural Institute, Dublin; 1976. p. 162–3 {87} 81. Eyal E, Lawi A, Folman Y, Morag M. Lamb and milk production of a ﬂock of dairy ewes under an accelerated breeding regime. Journal of Agricultural Science, Cambridge 1978;91:69–79 [ABA abstr.] {87} 82. Marzin J, Proud’hon M, Brelurut A, Angevain J, Reboul G. Performance of Local Ewes and their Crosses with the Romanov under an Accelerated Lambing System. INRA, ITOVIC, Paris; 1979. p. 349–66 [ABA abstr.] {88} 83. Marzin J, Brelurut A. Comparative performance of Limousin and RomanovLimousin ewes under relatively intensive husbandry. Bulletin Technique, Centre de Recherches Zootechniques et Veterinaire de Theix, Vannes, France, 1979;37:15–23 [ABA abstr.] {88} 84. Dzakuma JM, Whiteman JV, Fields JE, Spencer R. Lambing performance of crossbred ewes of Finnsheep, Dorset and Rambouillet breeding under two cycles of an accelerated lambing program. Animal Science Research Report Oklahoma State University (M.P. 107); 1980. p. 14–18 [ABA abstr.] {88} 85. Cognie Y, Gayerie F, Oldham CM, Poulin N, Mauleon P. Frequent lambing; underlying physiology. In: 32nd Meeting, EAAP (III-1); 1981. 18 pp [ABA abstr.] {88} 86. Jankowski S. A review of experiments on frequent lambing in Poland. In: Proceedings of the 32nd Annual Meeting, EAAP; 1981. 9 pp [ABA abstr.] {88} 87. Eyal E. Accelerated lambing frequency in dairy sheep under experimental and commercial conditions. Proceedings of the 32nd Annual Meeting, EAAP; 1981. 9 pp [ABA abstr.] {88} 88. Dyrmundsson OR. Frequent lambing in Icelandic sheep. In: 32nd Annual Meeting of the European Association for Animal Production; 1981. p. 6 [ABA abstr.] {88} 89. Valls Ortiz M. Frequent lambing of sheep ﬂocks in Spain. Productivity and management consequences. In: 32nd Annual Meeting, EAAP. (III-4); 1981. 8 pp [ABA abstr.] {88} 90. Valls Ortiz M. Frequent lambing of sheep ﬂocks in Spain. Productivity and management consequences. Livestock Production Science 1983;10:49–58 [ABA abstr.] {88} 91. Notter DR. Repeatability of conception rate and litter size for ewes in an accelerated lambing system. Journal of Animal Science 1981;52:643–50 [ABA abstr.] {88} 92. Dzakuma JM, Stritzke DJ, Whiteman JV. Fertility and proliﬁcacy of crossbred ewes under two cycles of accelerated lambing. Journal of Animal Science 1982;54:213–20 [ABA abstr.] {89} 93. Dzakuma JM. Productivity of crossbred ewes under accelerated lambing and accuracy of estimating lifetime productivity. Dissertation Abstract International 1981;B42:3–4 [ABA abstr.] {89}

g

135

94. Hulet CV. Effect of time of early weaning and time of lambing on accelerated lambing in Polypay sheep. Theriogenology 1983;20:141–8 [ABA abstr.] {89} 95. Amir D, Rosenberg M, Schindler H. Oestrus and ovarian activities of Finn-cross ewes during the post-partum and the seasonal anoestrous periods. Journal of Agricultural Science, Cambridge 1984;103:155–60 [ABA abstr.] {89} 96. Goot H, Eyal E, Foote WC, Matthews DH. Lamb performance and wool production in Finn-cross sheep in accelerated lambing programs. Final Report to U.S.–Israel Agricultural Research Development Fund 1984. p. 23–36 [ABA abstr.] {90} 97. Hackett AJ, Wolynetz MS. Fertility of ewe lambs maintained indoors year-round on an accelerated breeding program. Journal of Animal Science 1984;59:1129–34 [ABA abstr.] {90} 98. Iniguez LC, Quaas RL, van Vleck LD. Lambing performance of Morelam and Dorset ewes under accelerated lambing systems. Journal of Animal Science 1986;63:1769–78 [ABA abstr.] {90} 99. Veress L, Ve´gh J, Turai I, Tarno´czi T, Ecsedi F. Some conclusions concerning the large scale accelerated lambing of Hungarian Merino ewes. Acta Agronomica Hungarica 1988;37:111–21 [ABA abstr.] {91} 100. Urrutia MJ, Martinez RF, Sa´nchez GF, Pijoan AP. A programme of accelerated lambing using Rambouillet ewes on the high plateau of Mexico. In: 1er Congreso Nacional de Produccion Ovina, Mexico, AMTEO; 1988. p. 144–6 [ABA abstr.] {90} 101. Gabin˜a D. Improvement of the reproductive performance of Rasa Aragonesa ﬂocks in frequent lambing systems. I. Effects of management system, age of ewe and season. II. Repeatability and heritability of sexual precocity, fertility and litter size. Selection strategies. Livestock Production Science 1989;22:69–85, 87–98 [ABA abstr.] {97} 102. Aboul-Naga AM, Aboul-Ela MB, Mansour H, Gabr M. Reproductive performance of Finn sheep and crosses with subtropical breeds under accelerated lambing. Small Ruminant Research 1989;2:143–50 [ABA abstr.] {92} 103. Lahlou-Kassi A, Berger GF, Bradford GE, Boukhliq R, Tibary A, Derqaoui L, et al. Performance of D’Man and Sardi sheep on accelerated lambing. I. Fertility, litter size, postpartum anoestrus and puberty. Small Ruminant Research 1989;2:225–39 [ABA abstr.] {91} 104. Bradford GE, Lahlou-Kassi A, Berger YM, Boujenane I, Derqaoui L. Performance of D’Man and Sardi sheep on accelerated lambing. II. Ovulation rate and embryo survival. Small Ruminant Research 1989;2:241–52 [ABA abstr.] {91} 105. Berger YM, Bradford GE, Essaadi A, Bourﬁa M, Lahlou-Kassi A. Performanceof D’Man and Sardi sheep on accelerated lambing III lamb mortality, growth and production per ewe. Small Ruminant Research 1989;2:307–21 [ABA abstr.] {91} 106. Mendel C, Scholaut W, Pirchner F. Performance of Merinolandschaf and Bergschaf under an accelerated lambing system. Livestock Production Science 1989;21: 131–41 [ABA abstr.] {91} 107. Aboul-Naga AM. Breeding activity of two subtropical Egyptian sheep breeds underaccelerated lambing system. Small Ruminant Research 1991;4:227–83 [ABA abstr.] {92} 108. Aboul-Naga AM, Mansour H, Aboul-Ela H, Almahdy H. Breeding activity of two subtropical sheep breeds under

CAB International 2010

136

Animal Science Reviews 2010 accelerated lambing system. Small Ruminant Research 1991;4:285–92 [ABA abstr.] {92}

109. Schoeman SJ. Productivity of purebred Do¨hne Merino, S.A. Mutton Merino and Dorper sheep under an intensive accelerated lambing system. In: Proceedings of the 4th World Congress on Genetics and Applied Livestock Production, Edinburgh XV; 1990. p. 373–6 [ABA abstr.] {92} 110. Schoeman SJ, Burger R. Performance of Dorper sheep under an accelerated lambing system. Small Ruminant Research 1992;9:265–81 [ABA abstr.] {92} 111. Wheaten J, Windels HF, Johnston LJ. Accelerated lambing using exogenous progesterone and the ram effect. Journal of Animal Science 1992;70:2628–35 [ABA abstr.] {92} 112. Mavrogenis AP, Chimonides I. Reproductive and production efﬁciency of Chios ewes under an accelerated breeding system. Small Ruminant Research 1992;7:353–60 [ABA abstr.] {92} 113. Donald HP, Read JL. The performance of Finnish Landrace sheep in Britain. Animal Production 1967;9:471–6 [ABA 36, abstr. 390] {92, 101} 114. Turner H. Newton. Australian sheep breeding research. ABA Review 1977;45:9–31 {92, 101} 115. Diaz Infante CM, Urrutia Morales J, Ochoa Cordero MA. Reproductive performance of Rambouillet ewes subjected to an accelerated lambing system. Revista Latinoamericana de Pequen˜os Rumiantes 1995;1:211–9 [ABA abstr.] {92} 116. Schoeman SJ, Botha MA. The effect of melatonin treatment on reproductive efﬁciency in an accelerated lambing system. Journal of South African Veterinary Association 1995;66:230–4 [ABA abstr.] {92} 117. Bittante G, Gallo L, Carnier P, Cassandro M, Mantovani R, Pastore E. Effects on fertility and litter traits under accelerated lambing scheme in crossbreeding between Finnsheep and an Alpine sheep breed. Small Ruminant Research 1997;23:43–50 [ABA abstr.] {93} 118. Maria GA, Ascaso MS. Litter size, lambing interval and lamb mortality of Salz, Rasa Aragonesa, Romanov and F1 ewes on accelerated lambing management. Small Ruminant Research 1999;32:167–72 [ABA abstr.] {93} 119. Kusakari N, Ohara M. Effect of accelerated lambing system with melatonin feeding on reproductive performance for two years in Suffolk sheep raised in Hokkaido. Journal of Reproduction and Development 1999;45:283–8 [ABA abstr.] {93}

124. Keskin M, Bicer O, Gu¨l S. Accelerated Lambing Systems. Mustafa Kemal Universetisi, Ziraat Facultesi, Antakya-Hatay; 2002. p. 89–94 [ABA abstr.] {90, 94} 125. Tosh JJ, Wilton JW, Kennedy D. Heritability of fertility in four seasons for ewes under accelerated lambing. In: Proceedings of the 7th World Congress on Genetics Applied to Livestock Production; 2002. 8 pp. [ABA abstr.] {94} 126. Banos G, Lewis RM, Notter DR, Hogue DE. Genetic proﬁle of fertility and proliﬁcacy in maiden and mature ewes managed in a frequent lambing system. In: Proceedings of the 7th World Congress on Genetics Applied to Livestock Production; 2002. 4 p. (sic) [ABA, abstr.] {94} 127. Susˇic V. Frequency of sheep lambing. Veterinarska Stanica 2003;34:157–63 [ABA abstr.] {94} 128. Susˇic V, Sˇtokovic I. Duration of breeding season and inter lambing period in highly fertile sheep breeds. Veterinarska Stanica 2004;35:43–9 [ABA abstr.] {94} 129. Koyuncu M. Reproductive performance of Kivircik ewes on accelerated lambing management. Pakistan Journal of Biological Sciences 2005;8:1499–502 [ABA abstr.] {94} 130. Menegatos J, Goulas C, Kalogiannis C. The productivity, ovarian and thyroid activity of ewes in an accelerated lambing system in Greece. Small Ruminant Research 2006;65:209– 16 [ABA abstr.] {94} 131. El-Saied UM, de la Fuente LF, San Primitovo F. Lifetime traits comparison between annual and accelerated lambing systems for dairy ewes. Livestock Science 2006;101:180–90 [ABA abstr.] {94} 132. Zapasnikiene B. The effect of lambing season and frequency on local ewe reproduction and progeny weight. Baisogala: Institute of Animal Science of Lithuanian Veterinary Academy 2007;49:24–32 [ABA abstr.] {95} 133. Zapasnikiene B, Nainiene R. Analysis of the Exclusive Traits of Lithuanian Local Coarse Wooled Sheep. Volume 51. Institute of Animal Science of Lithuanian Veterinary Academy, Baisogala; 2008. p. 50–7 [ABA abstr.] {95} 134. Stewart IB, Louw BP, Lishman AW. Suckling behaviour and fertility of beef cows on pasture 1. Suckling behaviour. South African Journal of Animal Science 1993;23:176–80 [ABA abstr.] {96} 135. Stewart Iona B, Louw BP, Lishman AW. Suckling behaviour and fertility of beef cows on pasture 2. Inﬂuence of twelvehour calf separation on interval to ﬁrst oestrus after onset of mating period. South African Journal of Animal Science 1993;23:180–2 [ABA abstr.] {96}

120. Urrutia J, Ochoa M, Meza HCA, Mancilla C. Reproductive performance of Merino Rambouillet ewes under three cycles of accelerated lambing.Wool Technology and Sheep Breeding 2001;49:193–201 [ABA abstr.] {93}

136. Stewart Iona B, Louw BP, Lishman AW, Stewart PG. Latenight suckling inhibits onset of postpartum oestrus activity in beef cows. South African Journal of Animal Science 1995;25:26–9 [ABA abstr.] {96}

121. Fisher JW. An economic comparison of production systems for sheep. Canadian Journal of Agricultural Economics 2001;49:327–36 [ABA abstr.] {93}

137. Gordon K, Stiegmann M. Suckling behaviour of ewes in early lactation. Physiology and Behavior 1991;50:1079–81 [ABA abstr.] {96}

122. Notter DR. Opportunities to reduce seasonality of breeding in sheep by selection. Sheep and Goat Research Journal 2001;17:20–32 [ABA abstr.] {93}

138. Nowers CB, Coetzer WA, Morgenthal JC. Effect of melatonin implants, ﬂushing and teasing on the reproductive performance of spring-mated Do¨hne Merino ewes. South African Journal of Animal Science 1994;24:22–6 [ABA abstr.] {97}

123. Goulet F, Castonguay FW. Inﬂuence of lambing-torebreeding interval on ewe reproductive performance in the anestrous season. Canadian Journal of Animal Science 2002;82:453–456 [ABA abstr.] {93}

g

139. Haresign W. The effect of implantation of lowland ewes with melatonin on the time of mating and reproductive

CAB International 2010

G.L. Hunter performance. Animal Production 1992;54:31–9 [ABA abstr.] {98} 140. Parsons SD, Hunter GL. Practises controlling fertility and ﬂock improvement in sheep farming. Proceedings of the South African Society for Animal Production 1965;00:136–9 [ABA abstr.] {98} 141. Haresign W. Responses of ewes to melatonin implants: Importance of the interval between treatment and ram introduction on the synchrony of mating, and effects on ovulation rate. Animal Production 1992;54:41–5 [ABA abstr.] {98}

g

137

142. Bettencourt CMV, Moffatt RJ, Keisler DH. Active immunisation of ewes against prostaglandin F2alpha to control ovarian function. Journal of Reproduction and Fertility 1993;97:123–31 [ABA abstr.] {99} 143. Notter DR, Cockett Noelle E. Opportunities for detection and use of QTL inﬂuencing seasonal reproduction in sheep: a review. Genetics, Selection, Evolution 2005;37 (Suppl. 1):S39–S53 [ABA abstr.] {100} 144. Mateescu R, Lunsford A, Thonney M. Association between melatonin receptor 1A gene polymorphism and reproductive performance in Dorset ewes. Journal of Animal Science 2009;87:2485–2488 [ABA abstr.] {104}

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Poultry sector development, highly pathogenic avian inﬂuenza and the smallholder production systems J. Rushton1, R.E. Viscarra2, N. Taylor3, I. Hoffmann4 and K. Schwabenbauer5 Address: 1 RVC, Hawkshead Lane, Hatﬁeld, Herts AL9 7TA, UK.2 CEVEP, Casilla 10474, La Paz, Bolivia.3 VEERU, University of Reading, Earley Gate, Whiteknights Road, Reading RG6 6AT, UK.4 FAO (AGAH) Viale delle Terme di Caracalla, Rome, 00153, Italy.5 FAO (AGAP) Viale delle Terme di Caracalla, Rome, 00153, Italy. Correspondence: Email: [email protected] 30 September 2009 27 February 2010

Received: Accepted:

Abstract Poultry sectors are the fastest-changing component of the general livestock sector. These changes are typiﬁed by tremendous growth in poultry populations in industrial production systems. These industrial systems generally separate feed mills, breeding, production and processing into different operations. In a competitive market environment, such operations may be physically separated, but are closely integrated by strong private regulations. In less competitive market environments, encouraged by protection policies, industrial poultry chains are loosely integrated and are weakly regulated by governments. Large breeding ﬂocks continue to exist, but are very weakly linked to low investment fattening and layer production units1. Investment in processing is low and live bird markets are commonly used to market products. Loosely integrated industrial chains create enormous risks for the spread of contagious diseases such as highly pathogenic avian inﬂuenza (HPAI), and affect the smallholder poultry producers who live and work alongside the industrial units and traditionally use live-bird markets for the sale of their own products and purchase of poultry. The review will discuss how changes in sector policy could improve industrial poultry sector performance and in turn reduce risks to the smallholder producers. Keywords: Poultry, Production Systems, Highly Pathogenic Avian Inﬂuenza.

Introduction During the 1990s, there was a growing realization that the livestock sector was in a process of change. FAO pioneered early classiﬁcation systems to detail how the livestock production units were developing and where they were concentrating [1, 2]. Building on these approaches, it was recognized that a ‘livestock revolution’ was ongoing, responding to the growing demands of urban populations in developing countries [3]. It was

1 The authors recognize that most countries have integrated chains, but in badly affected countries these chains produce and process a minority of birds.

g

documented at an early stage that much of the growth in the livestock sector was coming from the intensive monogastric systems and to some extent from a growth in milk production. For many reasons, these dramatic changes in livestock production were celebrated, although some concerns were raised about poorer livestock producers being left behind [4–7] and issues on the potential negative impacts on the environment have been well investigated [4, 8]. What were less well anticipated were the growing problems with the control of transboundary animal diseases and more speciﬁcally the emergence and resurgence of dangerous zoonotic diseases [9]. The most notable of these is highly pathogenic avian inﬂuenza H5N1 (HPAI H5N1), which was ﬁrst reported to have caused serious problems in domestic poultry and also human mortalities in 1997 in Hong Kong [10, 11]. The disease

CAB International 2010

140

Animal Science Reviews 2010

was quickly controlled, largely by massive culling2, and marketing systems were modiﬁed to reduce risk, but it reappeared in 2001 and 2002, indicating that a reservoir of the disease continued to be maintained [12]. These relatively small-scale outbreaks were overshadowed by events of 2003/04 when HPAI H5N1 spread quickly across Southeast Asia [12]. At ﬁrst, it appeared that the disease would be contained in this region, but 2005 saw outbreaks ﬁrst in Russia, then in Turkey, that were successfully controlled. However, since early 2006, the disease has become established outside of Southeast Asia in Egypt and Nigeria and recently in Myanmar and Bangladesh, with outbreaks also in neighbouring India. Sporadic cases have also occurred in Europe. The persistence and the continuing spread of HPAI H5N1 has created global concerns of a human ﬂu pandemic, as a result of the potential for AI viruses to mutate, allowing rapid human to human transmission, and people’s awareness of ‘bird ﬂu’ across the world is very high. The spectre of a pandemic has stimulated one of the largest concerted animal disease control efforts that the world has seen. The original control measures have focused on the use of culling as a means to stop the disease spreading, which is the ﬁrst time that developing countries have consistently applied such measures over extended periods of time. In some countries, there has also been the widespread adoption of vaccination to control HPAI, with Vietnam often cited as being the country where this measure has been implemented most successfully. In the background and growing more important has been the need to upgrade biosecurity measures in the poultry sector, often linked with the need to restructure the sector. A majority of these culling and vaccination campaign measures are based on experiences gained with the global rinderpest eradication3, and poultry disease control measures applied in the industrialized countries. A number of issues have been strongly debated about the disease. For example, the initial hypothesis on disease spread centred on wild birds being an important element [13]. It seems that wild birds play an important role at certain times in some regions, but trade as a transmission channel cannot be dismissed as insigniﬁcant. Trade appears to have played a role in some regions for international spread, and has almost certainly been the most important mechanism for national and local spread of the disease. There have also been emotive debates and some very harsh decisions with regard to who is to blame for

2 Culling is a disease control intervention, which, in its most effective application, removes sick and at risk birds through slaughter, disposal of the carcass and disinfection of the property the birds were held on. No birds should be kept on the property for a period after the culling. 3 While rinderpest was largely controlled in Europe with culling and movement control, the eradication in other regions of the world has been based on vaccination and strong monitoring of vaccination campaigns.

g

the entry and maintenance of HPAI in countries. Convenient scapegoats have been the backyard producers, who, with their low-input low-output systems, are perceived as being non-biosecure. It has been argued that these systems are riskier than the large industrial-scale units, but data collected on disease outbreaks indicate that the quantitative justiﬁcation for this view is limited [14]. Savill et al. [15] have also raised the possibility that HPAI may be maintained unnoticed in poorly vaccinated ﬂocks. What recurs in many places is the idea that a technical, biological solution can be found, and that the poultry sector can be conveniently divided into different components or segments [16]. By focusing measures, often Draconian, on particular components of the poultry sector, but in particular the backyard producers, it is assumed that the disease will disappear. However, after nearly 5 years of HPAI control efforts, it is becoming increasingly clear that classical public-sector-led control measures, which have been successfully used in developed countries over the last decade, are not working well in developing countries. The focus of this review will be to examine where the smallholder poultry ﬁt within the poultry sectors of the countries worst affected by HPAI H5N1. In the process it will attempt to answer three questions: Are our current methods of describing the poultry sector adequate for the planning of HPAI control? Are the control measures so far applied appropriate for the poultry sectors where the disease is now endemic? Can HPAI H5N1 be effectively controlled and ultimately eradicated from domestic poultry ﬂocks? In attempting to answer these questions, the authors will make suggestions on how to view the poultry sector to understand linkages between its different components. It is argued that with such a perspective and underlying information, disease control intervention planning can be improved, and differentiated control policies can be better designed. The overall aim is, of course, the control and future elimination of HPAI H5N1 from domestic poultry.

Methods The article is a review of the available information on poultry sector development and HPAI, which are subjects that have been studied in isolation but have generally not been brought together. For example, early work on Newcastle disease only focused on backyard systems and did not consider the more general poultry sector [17, 18]. An early HPAI impact assessment for Southeast Asia was also weak on the poultry sector [19]. More recently, work in Egypt has included analysis of the poultry sector [20], assessment of compensation mechanisms [21, 22] and the livelihood and market shock impacts of the

CAB International 2010

J. Rushton, R.E. Viscarra, N. Taylor, I. Hoffmann and K. Schwabenbauer

disease [23, 24], but this has not been combined with HPAI epidemiological data. The global experiences of the authors with strong ﬁeldlevel investigations4 provides a basis for the development of conceptual thinking on poultry sector development and the associated problems faced with new poultry sectors in the control of HPAI H5N1.

Discussion Problems with Existing Classiﬁcation Systems In 2004, FAO proposed a classiﬁcation system for the poultry sector based on perceived biosecurity levels of different production units, Sector 1 with the highest biosecurity level and Sector 4 the lowest. The classiﬁcation system has been interpreted in different ways by different people and organizations, but in general people would state that chicken breeding ﬂock farms are in Sectors 1 and 2, integrated broiler and layer farms may be sector 2, the rustic broiler production and layer units are in sector 3 and the backyard5 smallholder systems are sector 4. Ducks have largely been ignored by many within this classiﬁcation. The classiﬁcation system has universal appeal and suits many perceptions of biosecurity, which often equates to investments in physical barriers. However, the different interpretations of the classiﬁcation system indicate that it requires some modiﬁcation (see [19] for a fuller discussion for Southeast Asia). The following sections will attempt to answer whether the classiﬁcation system helps to understand the general development of the poultry sectors where there are problems with HPAI control. It will also suggest what type of systems analysis may be required to assist in HPAI control.

Industrial Chicken Systems The industrial chicken production systems, layers and broilers, are the fastest-growing component of the poultry sector [25]. In general, broilers in industrial production systems, which include birds in both high- and low-investment production units, make up half or more of the poultry population6. These industrial production

4 Two authors are involved in a Wellcome Trust project on HPAI in Vietnam Project Number 079282/Z/06/Z, one author manages a global project for HPAI and backyard producers (see Schwabenabauer et al., this edition) and the other author coordinates FAO’s socio-economics work on HPAI. 5

Backyard is used in its widest sense and covers the rooftop systems found in Egypt. 6

Pym et al. [26] state that only 10% of standing population are broilers, but it is believed that their deﬁnition of industrial is limited to highinvestment units. A mixture of ofﬁcial livestock population data and

g

141

systems have a pyramid shape in terms of population, i.e. the great-grandparent stock are few in number, there are more grandparent stock, yet more parent stock and many production birds either broiler fatteners or layers [27]. These populations are physically separated, in many cases by large distances, but are linked into ‘contact networks’ through movements of fertile eggs and at a local level through feed supply7. Elmsley [27] provides a very good analysis of the international nature of the poultry breeding business, with a strong analysis of the formal routes of transfer. In a fully integrated system, the feed, breeding and production units are under the management and control of one ﬁrm: therefore the contact network is entirely within the control of that ﬁrm. However, in the countries currently badly affected by HPAI, the integration of industrial chicken production systems is weak or absent, breeding farms produce fertile eggs or day-old chicks that are then sold to producers to fatten or produce eggs. The formal contract between companies with parent stock ﬂocks, some of whom also have grandparent stock farms, and the production units are limited and generally weak and various levels of brokers, traders and transporters are involved in the contact network. In these countries, there are examples of breeding farms, with high levels of physical biosecurity, being infected during different points of the HPAI epidemic (personal communication with people working in Egypt and Indonesia). The potential impact of outbreaks of a disease such as HPAI in such important points in the industrial chicken systems is enormous as they are linked to so many other farms. To all intents and purposes, the industrial chicken production units in these countries are part of one system or network. Therefore, the lowest level of biosecurity in any one unit will ultimately affect the biosecurity status of all other units. This recognizes that while birds and products probably move only in one direction, the ﬂow of people, transport and physical infrastructure to carry eggs and birds can be in both directions. Therefore, fomite spread is of importance in both directions in the poultry value chains. The FAO sector classiﬁcation system does not recognize these linkages and in fact gives an impression that the breeding sectors are separate and distinct from the layer and fattening units. The sector classiﬁcation system also does not provide any understanding of the importance of private regulation and enforcement across the industrial chicken production system. When private sector regulation and enforcement are strong, biosecurity levels are tightened through the application of disease prevention measures such as

estimates of poultry sector growth from Egypt, Indonesia, Thailand, Bangladesh all indicate high proportions of the standing bird populations are in industrial broiler units. 7

There are data to suggest that outbreaks in an African country in the industrial chicken systems were through feed lorries.

CAB International 2010

142

Animal Science Reviews 2010

vaccination and the general safety of the system is increased. It is argued that tightly integrated poultry chains are usually associated with, and fostered by, strong private regulation and enforcement, with governments providing overall guidance and a legal framework for operations. The key question for governments is how to create a policy environment that can encourage private industrial chicken sector actors to adopt tighter private regulatory systems. It is suggested that the most effective, though also potentially the harshest, mechanism is to open these production systems to international trade, where competition against the most competitive poultry sectors in the world will lead to greater exposure to improved management methods, private regulatory procedures and increased productivity. It is recognized that such sector liberalization could well have very negative impacts, and probably a much sounder strategy is to encourage large industrial producers to compete in world markets through the export of their products. The best example of the positive beneﬁts of such entry into world trade is the reaction of Thai exporting poultry companies after HPAI. They have tightened their industrial chicken production systems, eliminating contract growers from their systems and fully integrating the growing units. The exporting countries have also rapidly changed from chilled to cooked products to maintain export markets, while this shift is about their market, the integration of the chain is more about risk perception and desire to protect valuable breeding ﬂocks. Those who worry that more strongly integrated industrial poultry chains will lead to the loss of employment and livelihood potential in small-scale fattening and layer units need to reﬂect on the fact that the largest beneﬁciary from improved efﬁciency and productivity from this component of the poultry sector will be the urban-based consumer. Therefore there are strong social welfare gains from improved efﬁciency. By retaining relatively inefﬁcient, and in some cases risky, poultry production units within poultry value chains, the welfare of a large group of consumers is negatively affected, as they have to pay more than is necessary for poultry products. Historically, these small-scale poultry industrial units did not exist in the countries currently badly affected by HPAI, and in developed countries they disappeared as economies of scale made small-scale production economically unviable. In recognition that a change in structure of the industrial chicken production systems will lead to a loss of employment in rural areas, it is recommended that the State ﬁrst make estimates of the numbers of people involved and contemplates alternative business or employment opportunities.

of scavenge base feed resources and household scraps [28–30]. The ﬂock size may vary with seasonal availability of feed resources, and hence there may also be seasonality in the supply of birds and eggs. The ability of these small units to respond to seasonal demand is potentially ﬂexible, and at certain times of year they may become part of the bigger poultry ﬂock that supplies products to urban markets. Such backyard systems are distinct from the small-scale rustic industrial units for layer and broiler production, which rely on purchased feed inputs, and market a high proportion of their eggs and birds outside the household. These backyard systems may be involved in transmission cycles of HPAI H5N1 if they have signiﬁcant mixture of poultry species, and have direct or indirect links with medium- and small-scale broiler fattening or layer units. The latter is the case for Egypt, Bangladesh, Vietnam and Indonesia, and the linkages are generally indirect via traders and feed suppliers. The backyard systems in some villages exist in close proximity to the small-scale industrial chicken production systems, creating risks for these units as they are constantly exposed to external chicken production systems through the transportation of feed, birds or eggs and of course the constant movement of people involved in the more general poultry sector. Data from Vietnam would suggest that not all villages have these small-scale rustic industrial chicken units, so not all villages are exposed to similar levels of risk (Taylor, unpublished data from Wellcome Trust project). Similarly, Geerlings et al. [24] report that not all Egyptian villages interviewed had small-scale rustic industrial units. What can be stated with clarity is that backyard production systems, once infected, have a limited ability to infect other ﬂocks, because of the number of birds in the ﬂocks and their tendency not to be linked into wideranging trade8. They may also be able to eliminate risks of further spread, but they probably lack information and resources to apply control and containment measures with rigour. It is argued that the State needs to investigate this further to ensure that poorer families are not negatively affected by HPAI and are given the appropriate means to deal with this disease. There is a strong need for additional socio-economics data on the poultry sector in the different countries to make sure that the policies to be applied are proportionate to the risk of spreading the disease as well as for human health.

Duck Production Systems The organization of the duck production systems in countries affected is less well documented than the

Backyard Production Systems With regard to the backyard systems, these are deﬁned as units with small ﬂocks whose size relates to the availability

g

8

See Leibler et al. [31] for the HPAI risks of different production systems.

CAB International 2010

J. Rushton, R.E. Viscarra, N. Taylor, I. Hoffmann and K. Schwabenbauer

chicken production systems. In the countries with continuing H5N1 problems there are signiﬁcant duck populations, and in some countries the expansion of duck populations in the last 20 years has been very quick [32]. In other countries, the ofﬁcial estimates of duck populations are well below what ﬁeld estimates indicate. For example, Egypt ofﬁcially reported a duck population of 9 million, while estimates from ﬁeld data indicate that the population is 40–60 million [33]. Similarly to chickens, ducks are raised in different production systems. There are industrial-style systems with well-organized breeding sectors and associated fattening and layer units. In some countries, the companies and people involved in these industrial duck systems have signiﬁcant political lobbying power, similar to the large-scale industrial chicken producers, who have political lobbying power in all badly affected countries9. In contrast, the small-scale and backyard chicken and duck farmers do not have political lobbying power. A large number of ducks are found in smallholder ﬂocks, often mixed with chickens and other poultry species. A further duck category is also found in a number of countries with production systems that are large-scale, in terms of ﬂock size, non-industrial, rustic and free-ranging. These are often associated with rice paddy production and mainly found in Asia. It has become increasingly obvious that the understanding of the duck component is critical, in some cases not well-enough studied10 and on others not well recognized for means of designing and implementing effective disease control interventions (Slingenbergh, personal communication, 2008).

Poultry Marketing and Processing In countries where HPAI has become endemic, there is minimal investment in poultry slaughterhouses, processing units and retailing systems. The main method of marketing birds is through live-bird markets. Most of these markets do not specialize in birds from particular production systems or particular species. Therefore, there is mixing of birds and people from different production systems within these markets. Often, smaller markets and big consumers such as restaurants are supplied via large wholesale live-bird markets, which can maintain large transient populations of birds. These markets are particularly identiﬁed as being the key in the spread of HPAI [35]. They also complete a part of the jigsaw of the poultry sectors in the badly affected countries, in that they

9

Countries that have lost a signiﬁcant proportion of their poultry ﬂock through disease and control measures and are suspected to be endemically infected as a result of continued reports of H5N1. 10 There are occasional exceptions, such as the study by Gilbert et al. [34].

g

143

link the industrial poultry production systems with the backyard systems. As mentioned above, it is possible that these linkages are seasonal and dependent particularly on the ability of smallholder producers to produce live birds and poultry products at different times of year. At certain times of the year, it is likely that the poultry sector is ‘one system or network’ with many connections through marketing of birds, movement of people and farm-level inputs. The concept that different production systems and different species may be linked into the same system or network is not captured in the sector classiﬁcation.

Avian Inﬂuenza Control Measures Capua and Marangon [36] and Sims [11] have reviewed avian inﬂuenza control methods. Their reviews have taken a very technical, biological approach, and the aim of the following discussion is to put the control measures they have reviewed into a socio-economic, political and institutional context. All countries so far affected by HPAI H5N1 have begun with the use of culling as the principal means to control disease. The initial culling policy has been based on ring culling around an outbreak, usually in large areas of 1–10 km radius. After a period of time, the culling policy has been modiﬁed, usually limiting its use to infected ﬂocks and dangerous contacts. There are several reasons for such modiﬁcations: people in non-infected areas resist the culling of apparently healthy birds even if compensation is being offered, the human and logistical resources required for ring culling where there are many outbreaks at one time often quickly outstrip available resources and the cost to the government in terms of compensation when using ring culling can quickly run into millions of dollars. Overall, culling supported with compensation is still an important tool in HPAI control, but it needs reﬁnement in order to ﬁt the epidemiological and socio-economic reality in which it is applied, and to ensure future cooperation in disease control measures. Vaccination has been well covered from both a scientiﬁc [37] and practical viewpoint [11, 38]. Capua and Marangon [37] indicate some of the difﬁculties that could be encountered with vaccination campaigns without appropriate monitoring systems and Alders et al. [38] detail the difﬁculties of implementing vaccination under different production systems. Sims [11] states that while vaccination is a useful tool to reduce virus load in the environment it will not, on its own, eliminate the virus from domestic poultry. However, strategic use of vaccination in poultry sectors that contain differing production systems and species has not been well discussed. The authors argue that the industrial chicken and duck production systems and in particular the breeding units, both grandparent and parent ﬂocks, and hatcheries, within these components need to use every measure to reduce the risks of becoming infected by HPAI H5N1. The reasons as explained above are to do with the potential

CAB International 2010

144

Animal Science Reviews 2010

serious impact of infection in these units on neighbouring and connected poultry producers. However, given the large scale of these units, the State should probably limit their involvement to monitoring circulating viruses, licensing the vaccines used, assessing the efﬁcacy of available vaccines and monitoring vaccination coverage. For the backyard systems, the currently available vaccines bring into question the ability of countries badly affected to vaccinate effectively in either campaigns or as a scheduled approach11. These producers could well receive the beneﬁt of day-old chicks being vaccinated12 at the hatchery level and ring vaccination could be used around conﬁrmed outbreaks. A critical problem where logistical and human resources are scarce is balancing the resources needed for a large-scale vaccination campaign for backyard sector with the need to strengthen disease detection, investigation and response work [40].

Conclusion The conclusion from the discussion is that while the sector classiﬁcation [16] was a good start when little was known about the poultry sectors with HPAI problems, it requires modiﬁcation in order to understand how the poultry industry operates and to reﬁne the planning of disease prevention and control. More information is required on the important linkages between inputs, production units, marketing and processing systems. An improved method of looking at the poultry sector is to describe the pyramid shape of the populations, examination of the linkages between production units and the investigation of the seasonality of the linkages. Of greatest urgency is a better understanding of how the duck production and marketing systems ﬁt within the general poultry sector. In addition, the organizations or people who are involved in driving the poultry sector need to be identiﬁed and their motivations understood. With such information local, national and international risk assessments can be made with greater certainty, allowing the identiﬁcation of risk points along chains, and within units. The people or organizations involved in these risky points can then be contacted to reach agreements on how risks can be reduced. The underlying aspect of this process of analysis is that many poultry units are linked through physical proximity, movement of inputs, people, transport or through poultry and poultry products. The current structure of poultry sectors in badly HPAI-affected countries indicates that at certain times of the year the poultry sector becomes a single system or network,

11

Birds vaccinated at a particular point in the production or age cycle.

12 A vaccine has been developed for application to day-old chicks but there are conﬂicting results on its efﬁcacy of protection (see [39]).

g

linking naive populations with endemically infected populations and creating seasonal epidemics. Government action for such poultry sectors that have become badly affected by HPAI should search for policy measures that facilitate the sustainable development of the poultry sector. In the worst-affected countries, it is unlikely that the State has the capacity to become the sole regulatory and enforcement agent for the poultry sector. In fact it should be recognized that in the majority of countries, poultry sectors are private-sector-led and -regulated. Therefore the State needs to develop a strategy with a legal and institutional framework that facilitates and encourages the industrial chicken and duck production systems to adopt tighter private regulatory and enforcement frameworks, improve biosecurity measures, motivate the adoption of disease control tools and improve general productivity. The State should focus its role to regulating and monitoring critical aspects of the poultry sector in order that the private-sector functions in eliminating food-borne disease risks [41]. The State also has a very strong role in ensuring that if contagious disease enters the production and processing systems then bio-containment measures are quickly applied and enforced to ensure that other poultry producers are not affected and that consumers experience limited impacts as a result of supply constraints. To further encourage the private sector it is recommended that the State provides information on poultry health management schemes, and where appropriate provide support to upgrade key poultry sector infrastructure, including subsidies where global public goods may be created. Allowing the private industrial sector to deal with most of the upgrading of the industrial chicken and duck systems will leave the State to play a stronger and more effective role in providing services and information to backyard producers. The type of policies and actions adopted for this component of the poultry sector need to take advantage of lessons learned from other rural development initiatives that have involved poor livestock producers. It is recommended that this process builds on the pioneering work of the Australians on Newcastle disease during the 1980s and 1990s [17, 42–46], some of the understanding of how to work with isolated rural communities [47, 48] and the more recent work of the German-funded AHBL project [49]. All of these studies look to work with communities in the development of general poultry health management rather than a strong focus on a particular disease. The authors believe that HPAI H5N1 can be controlled and ultimately eliminated from domestic poultry in the countries affected. This can be achieved by a better understanding of how the poultry sector operates (with a greater emphasis on ducks), who is involved and what are their incentives to contributing to disease control measures. The evidence so far compiled would indicate that there is a need for differentiated policies that address the needs of the industrial poultry production systems

CAB International 2010

J. Rushton, R.E. Viscarra, N. Taylor, I. Hoffmann and K. Schwabenbauer

and the small-scale backyard and rooftop systems. The former requires a greater understanding of how to create a favourable policy environment that encourages stronger private regulation, the latter more grassroots work in information provision and resources when producers are affected. In general, there is a need for a people-centred approach with strong technical leadership that promotes proportionate disease control measures.

145

with John Hopkins, School of Public Health, USA, University of California, USA and RVC, London, UK; 2007. 15. Savill NJ, Rose SG, Keeling MJ, Woolhouse MEJ. Silent spread of H5N1 in vaccinated poultry. Nature 2006;442:757. 16. FAO. FAO Recommendations on the Prevention, Control and Eradication of Highly Pathogenic Avian Inﬂuenza in Asia. FAO Position Paper, September 2004. FAO, Rome Italy; 2004. 59p. 17. Spradbrow PB. Newcastle disease in village chickens. Poultry Science Review 1993;5:57–96.

References 1. Sere C, Steinfeld H. World Livestock Production Systems: Current Status, Issues and Trends. Animal Production and Health Paper No. 127. FAO, Roma, Italia; 1996. 89p. 2. Steinfeld H, Ma¨ki-Hokkonen J. A classiﬁcation of livestock production systems. World Animal Review 1995;84/85:83–94. 3. Delgado C, Rosegrant M, Steinfeld H, Ehui S, Courbois C. Livestock to 2020. The Next Food Revolution. Food, Agriculture and the Environment Discussion Paper 28. IFPRI, Washington DC; 1999. 72p. 4. de Haan C, Schillhorn van Veen T, Brandenburg B, Gauthier J, Le Gall F, Mearns R, et al. Livestock Development: Implications for Rural Poverty, The Environment and Global Food Security. The World Bank, Washington DC; 2001. 96p. 5. Heffernan C. Livestock and the poor: issues in poverty focused development. In: Owen E, Smith T, Steele MA, Anderson S, Duncan AJ, Herrero M, et al., editors. Responding to the Livestock Revolution. The Role of Globalisation and Implications for Poverty Alleviation. BSAS Publication No. 33, Nottingham, UK; 2002. p. 229–45. 6. Owen E, Best J, Devendra C, Ku-Vera J, Mtenga L, Richards W, et al. Introduction - the need to change the ‘mind-set’. In: Owen E, Kitalyi A, Jayasuriya N, Smith T, editors. Livestock and Wealth Creation. Improving the Husbandry of Animals Kept by Resource-Poor People in Developing Countries. Nottingham University Press, Nottingham, UK; 2005. p. 1–11. 7. FAO. Livestock Policies and Poverty Reduction in Africa, Asia and Latin America. Pro-Poor Livestock Policy Initiative Policy Brief. FAO, Roma, Italia; 2005. 8. Steinfeld H, Gerber P, Wassenaar T, Castel V, Rosales M, de Haan C. Livestock’s Long Shadow Environmental Issues and Options. FAO, Rome, Italy; 2006. 390p. 9. Greger M. The human/animal interface: emergence and resurgence of zoonotic infectious diseases. Reviews in Microbiology 2007;33:243–99. 10. Capua I, Alexander D. The challenge of avian inﬂuenza to the veterinary community. Avian Pathology 2006;35(3):189–205. 11. Sims L. Lessons learned from Asian H5N1 outbreak control. Avian Diseases 2007;51:174–81. 12. Capua I, Alexander D. Avian inﬂuenza recent developments. Avian Pathology 2004;33(4):393–404. 13. Morris R. Highly pathogenic avian inﬂuenza. Keynote Presentation at the International Society for Veterinary Epidemiology and Economics, 6–11 August 2006. 14. Otte MJ, Roland-Holst D, Pfeiffer D, Soares-Magalhaes R, Rushton J, Graham J, et al. Industrial livestock production and global health risks. Research Report, FAO-PPLPI, Rome, Italy

g

18. Rushton J, Ngongi SN. Poultry, women and development: old ideas, new applications and the need for more research. World Animal Review 1998;91(2):43–9. 19. Rushton J, Viscarra RE, Guerne Bleiche E, Mcleod A. Impact of avian inﬂuenza outbreaks in the poultry sectors of ﬁve South East Asian countries (Cambodia, Indonesia, Lao PDR, Thailand, Viet Nam) outbreak costs, responses and potential long term control. World Journal of Poultry Science 2005;61(3):491–514. 20. El Naggar A, Ibrahim A. Case study of the Egyptian Poultry Sector. In: Thieme O, Pilling D, editors. Poultry in the 21st Century: avian inﬂuenza and beyond. Proceedings of the International Poultry Conference held in Bangkok, Thailand, 5–7 November 2007. FAO Animal Production and Health Proceedings, No. 9. Rome, 2008. p. 31. 21. Shalaby M. Rapid Assessment of Highly Pathogenic Avian Inﬂuenza Producers for reimbursement of Industrial Poultry Producers after the HPAI outbreak Egypt. FAO Consultant Report. FAO, Rome, Italy; 2006. 37p. 22. Ghonem M. Rapid Assessment of Highly Pathogenic Avian Inﬂuenza Procedures for Reimbursement of Backyard and Industrial Poultry Producers after the HPAI Outbreak. FAO Consultant Report. FAO, Rome, Italy; 2006. 42p. 23. Ibrahim A, Albrechtsen L, Upton M, Morgan N, Rushton J. Market Impacts of HPAI Outbreaks: A Rapid Appraisal Process Egypt. FAO, Rome, Italy; 2007. 34p. 24. Geerlings E, with Albrechtsen L, Rushton J. Highly pathogenic avian inﬂuenza: a rapid assessment of the socio-economic impact on vulnerable households in Egypt. Report for FAO/ WFP. FAO, Rome, Italy; 2007. 78p. 25. Rushton J, Viscarra RE, Taylor N. Is highly pathogenic avian inﬂuenza a symptom of rapid poultry sector development? Why understanding value chains will be a step towards disease control. Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources, CABI; (forthcoming). 26. Pym RAE, GuerneBleich E, Hoffmann I. The relative contribution of indigenous chicken breeds to poultry meat and egg production and consumption in the developing countries of Africa and Asia. In: XII European Poultry Conference, Verona, 10–14 September 2006 [CD Proceedings]. 27. Elmsley. Risk of AI transmission through breeding stock. In: Proceedings of the Symposium for Markets and Trade Dimensions of Avian Inﬂuenza, FAO, Rome, Italy, 13 November 2006. 28. Gunaratnes SP, Chandrasiri ADN, Mangalika Hemalatha WAP, Roberts JA. Feed resource base for scavenging village chickens in Sri Lanka. Tropical Animal Health Production 1993;25:249–57.

CAB International 2010

146

Animal Science Reviews 2010

29. Roberts JA. The scavenging feed resource base in assessments of productivity of scavenging village chickens. In: Spradbrow P, editor. Newcastle Disease in Village Chickens. Control with Thermostable Oral Vaccines. Proceedings of an International Workshop held in Kuala Lumpur, Malaysia, 6–10 October 1991. ACIAR, Canberra, Australia; 1992. p. 29–32. 30. Dessie T. Studies on village poultry production systems in the central highlands of Ethiopia [MSc Thesis]. Department of Animal Nutrition and Mangement, Swedish University of Agricultural Sciences, Uppsala, Sweden; 1996. 31. Leibler JH, Otte J, Roland-Holst D, Pfeiffer DU, Soares Magalhaes R, Rushton J, et al. Industrial food animal production and global health risks: exploring the ecosystems and economics of avian inﬂuenza. EcoHealth 2009;6(1): 58–70. 32. FAO (2007) FAOSTAT. Available from: URL: http:// faostat.fao.org (accessed 23 November 2007). 33. Hogerwerf L, Siddig A. Ducks and HPAI H5N1 in the Nile Delta, Egypt. FAO, Rome, Italy; 2007. 47p. 34. Gilbert M, Xiao X, Chaitaweesub P, Kalpravidh W, Premashthira S, Boles S, et al. Avian inﬂuenza, rice production and domestic duck production in Thailand. Agriculture, Ecosystems and Environment 2007;119:409–15. 35. Van Kerkhove MD, Vong S, Guitian J, Holl D, Mangtani P, San S, et al. Poultry movement networks in Cambodia: implications for surveillance and control of highly pathogenic avian inﬂuenza (HPAI/H5N1) Vaccine; 2009. p. 6345–52. 36. Capua I, Marangon S. Control of avian inﬂuenza in poultry. Emerging Infectious Diseases 2006;12(9):1319–24. 37. Capua I, Marangon S. Vaccination for avian inﬂuenza in Asia. Vaccine 2004;22:4137–8. 38. Alders R, Bagnol B, Young M, Ahlers C, Brum E, Rushton J. Challenges and constraints to vaccination in developing countries. Development of Biologicals 2008;130:73–82. 39. Hinrichs J, Otte MJ, Rushton J. Technical, epidemiological and ﬁnancial implications of large-scale national vaccination campaigns to control HPAI H5N1. CAB Reviews, Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources 2010;5(021):20 pp. 40. Rushton J, Upton M, Ayala G, Velasco R. Balancing active and passive policies for the prevention of transboundary diseases.

g

In: Proceedings of the International Society for Veterinary Epidemiology and Economics, 6–11 August 2006. 41. Schwabenbauer K, Rushton J. Veterinary services for poultry production. In: Thieme O, Pilling D, editors. Poultry in the 21st Century: avian inﬂuenza and beyond. Proceedings of the International Poultry Conference held in Bangkok, Thailand, 5–7 November 2007. FAO Animal Production and Health Proceedings, No. 9. Rome, 2008. p. 54. 42. Copland JW (editor). Newcastle Disease in Poultry. A New Food Pellet Vaccine. Monograph No. 5. ACIAR, Canberra, Australia; 1987. 43. Spradbrow PB (editor). Newcastle disease in village chickens. Control with thermostable oral vaccines. In: Proceedings No. 39, ACIAR, Canberra, Australia; 1992. p. 181. 44. Alders RG. Facilitating women’s participation in village poultry projects: experiences in Mozambique and Zambia. In Proceedings of the XXth World’s Poultry Congress, 2–5 September 1996, New Delhi, India; 1996. Vol. III, p. 441–7. 45. Alders RG. Extension methodologies for village poultry production. In: ACIAR/ADRI Workshop on Newcastle Disease Vaccines for Village Chickens, Animal Disease Research Institute, Dar es Salaam, 9–13 December 1996. 46. Alders RG, Spradbrow PB. Controlling Newcastle Disease in Village Chickens. A Field Manual. ACIAR Monograph No. 82. ACIAR, Canberra, Australia; 2001. p. 112. 47. Rushton J, Viscarra RE. Meeting the needs of poor, marginalised livestock keeping families – Concerted Action on Livestock and Livelihoods (CALL) project in Boluvia. In: Rowlinson P, Wachirapakorn C, Pakdee P, Wanapat M, editors. Proceedings of Integrating Livestock–Crop Systems to Meet the Challenges of Globalisation. Vol. 2. AHAT/BSAS International Conference 14–18 November 2005, Khon Kaen, AHAT, Bangkok, Thailand; 2005. p. T40. 48. Viscarra RE, Chipana O, Rushton J. Identifying and meeting the veterinary education needs of poor isolated communities in Bolivia. In: Proceedings of the International Society for Veterinary Epidemiology and Economics, 6–11 August 2006. 49. Schwabenbauer K, Besbes B, De Haan N, Thieme O, Rushton J. Animal Disease Control – A Multidisciplinary Task. In: Proceedings of the International Society for Veterinary Epidemiology and Economics, 10–14 August 2009. CD ROM.

CAB International 2010

Animal Science Reviews 2010

Fat and taste perception Philippe Besnard*, Dany Gaillard, Patricia Passilly-Degrace, Ce´line Martin and Michael Chevrot Address: Physiologie de la Nutrition, UMR INSERM U886/Universite´ de Bourgogne, AgroSup Dijon, 1, Esplanade erasme, F-21000 Dijon, France. *Correspondence: Philippe Besnard. Tel/Fax. [+33] (0)380 396 691. Email: [email protected] 14 September 2009 15 January 2010

Received: Accepted:

Abstract An obesity epidemic is spreading rapidly across the world, especially in children. It constitutes a major risk in serious diet-related chronic diseases, thus decreasing life expectancy. The origin of the obesity epidemic is clearly dependent on multiple parameters. Increased consumption of more energy-dense, nutrient-poor foods might be a contributory factor to the epidemic. Behavioural experiments in rats and mice show that both have a spontaneous preference for fatty foods. This lipid attraction is so strong that a mouse with free access to a source of oil in addition to standard laboratory chow becomes rapidly obese. In humans, studies also report that obese subjects have a greater preference than lean subjects for lipids. This last observation suggests that an inappropriate lipid perception might inﬂuence obesity risk by impacting feeding behaviour. Why we like fatty foods is not yet fully understood. Recent data demonstrate that low quantities of long-chain fatty acids can be speciﬁcally detected in the oral cavity by humans and laboratory rodents. Interestingly, lipid sensors have been found in rodent taste buds, suggesting that gustation can also play a role in the orosensory perception of lipids in combination with textural and olfactory cues. The sense of taste informs the organism about the quality of ingested food. It encompasses ﬁve sub-modalities: allowing the perception of sweet, salt, sour, bitter and umami stimuli. The discovery of lipid sensors raises the possibility for a sixth taste modality (‘fatty’) directed to dietary fat detection. This mini-review highlights recent ﬁndings in this new ﬁeld of investigation in both rodents and humans. Keywords: Dietary lipids, Lipid sensors, CD36, GPR120, Taste buds, Gustation, Fatty taste, Feeding behaviour, Obesity

Introduction In recent years, obesity has reached epidemic proportions across the world, with more than 1 billion adults overweight and at least 300 million of them clinically obese (body mass index 30). This phenomenon affects both adults and children [1]. Recent data suggest that the increase in prevalence of obesity is associated with a decrease in life expectancy in children [2]. Indeed, overweight and obesity constitute a major risk for serious diet-related chronic diseases, including type 2 diabetes, cardiovascular diseases, hypertension and certain forms of cancer. Genes are known to play an important role in the susceptibility to weight gain. Nevertheless, the profound changes in our way of life over recent decades, which

g

have led to a progressive decline in the physical activity associated with an overconsumption of diets high in energy, mainly explain this phenomenon. An abundance of food has obvious consequences: it promotes our speciﬁc appetites. Lipids account for about 40% of the calories ingested in Western countries, whereas nutritional recommendations are 5–10% lower. The mechanism responsible for this fat preference is a complex phenomenon that is not fully understood. It results from integration of multiple signals triggered by the chemoreception of lipids along the digestive tract (oral cavity, stomach and intestine) and by post-absorptive lipid-sensing by various organs including adipose tissue, liver and brain. In this system, detection of lipids in the oral cavity appears to be essential for determining the

CAB International 2010

148

Animal Science Reviews 2010

Dietary lipids = TG

A

Foliates

Taste buds

TG + lipase = LCFA

Circumvallates

von Ebner’s glands = Lingual lipase

Fungiforms

B Accumbens nucleus Hypothalamus

NS T

Digestive tract X

IX

... ... ... ... ... ...

VI I

... ... ... Figure 1 Anatomy and physiology of the gustatory system. (a) Localization of gustatory papillae in the mouse tongue and speciﬁc anatomic features of circumvallate papillae. (b) Main nervous circuitry involved in the sense of taste

choice to consume fat-rich foods preferentially. Orosensory perception of lipids was long thought to involve only textural and olfactory cues. Recent ﬁndings challenge this limited viewpoint, strongly suggesting that the sense of taste also plays a signiﬁcant role in dietary lipid perception and might therefore be involved in the preference for fatty foods and, hence, might contribute to lipidinduced chronic diseases. This minireview analyses recent data relating to the molecular mechanisms and physiological consequences of orosensory lipid perception.

There is a Spontaneous Preference for Lipids Compelling evidence strongly suggests that laboratory rodents have an orosensory system devoted to lipid detection. The two-bottle preference test is a classical method for studying the feeding behaviour of animals in a free-choice situation. This simple paradigm shows clearly that rats [3] and mice [4] display a strong preference for lipid-rich solutions. Although dietary lipids are mainly composed of triglycerides (TG), long-chain fatty acids (LCFA; number of carbons is >16) seem to be responsible

g

for oral lipid perception. In a free-choice situation, rats have a weaker preference for TG and medium-chain fatty acids (number of carbons ranges from 8 to 14) than for LCFA [3, 5]. This chemical selectivity is very tight, as LCFA derivatives, such as methyl-LCFA, are not recognized [3]. The ability of rodents to detect LCFA speciﬁcally has also been conﬁrmed with the conditioned test aversion (CTA) paradigm in which a naive animal learns to avoid a newly encountered tastant after suffering adverse post-ingestive effects triggered by an intraperitoneal injection of LiCl. Both rats and mice can be conditioned to avoid speciﬁc LCFA [6, 7], with a submicromolar detection threshold [6, 8]. Lingual lipase, which is responsible for an efﬁcient release of LCFA from TG in rodents, seems to play a signiﬁcant role in oral fat perception. Indeed, its pharmacological inhibition leads to a dramatic decrease in lipid preference in the mouse [9]. This may explain why mineral oil, which is not digestible, is not as attractive as vegetable oil in a free-choice situation [9]. Interestingly, lingual lipase is known to be released directly into the clefts of foliate and circumvallate papillae by von Ebner’s glands in rodents (Figure 1a). This anatomical feature appears to be ideal for the efﬁcient

CAB International 2010

Philippe Besnard, Dany Gaillard, Patricia Passilly-Degrace, Ce´line Martin and Michael Chevrot

hydrolysis of TG and generates high LCFA levels close to the taste buds, facilitating their subsequent detection by the taste receptor cells (TRC). Far fewer studies have been carried out on humans, but psychophysical studies carried out by Richard Mattes and co-workers strongly suggest that taste plays a role in fat perception in humans [10]. Healthy adult subjects can speciﬁcally detect the presence of saturated and unsaturated LCFA in the oral cavity. The detection threshold for LCFA is much lower (0.028%, w/v, on average) [11] than that for TG (5.6 and 17.3%, w/v, on average, in young and old subjects, respectively) [12]. Lipid-rich foods may contain up to 0.5% free fatty acids [10], so TG hydrolysis by lingual lipase does not seem to be required for orosensory fat detection. This point is important, given continuing debate concerning whether humans have an efﬁcient lingual lipase [13, 14].

Why Might We Think that Sense of Taste Plays a Role in the Fat Preference? However, the interpretation of these observations remains complex because the spontaneous food attraction is known to result from the integration of olfactory, somesthesic, gustatory and post-ingestive signals. Thus, to explore whether gustation plays a role in fat preference, the relative importance of each of these parameters for the spontaneous lipid attraction has been systematically explored in rodents. Firstly, spontaneous fat attraction is maintained in anosmic rats and mice, in which olfaction is blocked by chemical or surgical (i.e. section of olfactory nerve) means, demonstrating that smell does not play a signiﬁcant role in this behaviour [5, 15, 16]. Secondly, the two-bottle preference test clearly reveals that mice prefer vegetable oils to texturing agents, such as xanthan gum or mineral oil, suggesting that texture is not a major cue in orosensory fat perception [4]. Thirdly, fat preference is not abolished in very-short-term experiments (0.5–5 min) designed to minimize post-ingestive effects [3, 6, 17]. The persistence of an attraction for lipids in rats and mice in which textural, olfactory and post-ingestive cues have been simultaneously minimized strongly suggests that the sense of taste plays a signiﬁcant role in the spontaneous fat preference in rodents [5, 15]. In humans, the fact that the orosensory perception of LCFA is conserved in subjects in which olfactory and somatosensory (post-ingestive) cues are minimized also suggests the involvement of gustation in this phenomenon [11].

Anatomy and Physiology of the Gustatory System The sense of taste is ensured by TRC, which are clustered in specialized onion-shaped structures: the taste buds. They are found at high density on the tongue and at low

g

149

density in the soft palate, larynx, pharynx and the upper part of the oesophagus. In the lingual epithelium, taste buds are located in three types of gustatory papillae with different spatial distributions (Figure 1a). Most are the fungiform papillae, which cover the front two-thirds of the tongue. These papillae are mushroom-shaped and have a small number (1–3) of taste buds on their apical surface. The circumvallate and foliate papillae are located on the central and lateral regions, respectively, of the posterior third of the tongue. The circumvallate papillae consist of a circular depression, the walls of which contain a few hundred taste buds. Humans have about ten circumvallate papillae, whereas rodents have only one, in a central position. Foliate papillae are located at the posterior lateral edge of the tongue and contain hundreds of taste buds. Chemosensitive proteins (ion channels, metabotropic and ionotropic receptors) located on the apical side of the TRC are responsible for taste reception. Interactions between a tastant and its speciﬁc detection system lead to changes in intracellular free calcium concentration ([Ca2+]i), resulting in depolarization of TRC. This change leads to neurotransmitter release, which generates, in turn, the ﬁring of gustatory afferent nerve ﬁbres [18, 19]. TRC from fungiform papillae and some of the anterior foliate papillae establish synaptic contacts with the chorda tympani nerve ( i.e. VII pair of cranial nerves), whereas the posterior parts of the foliate and circumvallate papillae are innervated by the glossopharyngeal nerve (i.e. IX pairs of cranial nerves). Vagus ﬁbres from the X pairs of cranial nerves innervate taste buds found outside of the oral cavity. Afferent ﬁbres carry taste information from the tongue to the brain via the nucleus of the solitary tract (NST) in the brainstem, which constitutes the ﬁrst synaptic relay in the nervous gustatory pathway. NST is connected to different brain areas involved in the feeding behaviour, reward and memorization, as well as to the digestive tract through efferent ﬁbres of vagus nerves. This last circuitry is responsible for the cephalic phase of digestion induced by taste sensory input and leading to early digestive secretions, preparing the body to the food arrival (Figure 1b).

Lingual CD36 Displays the Properties of a Gustatory Lipid Sensor Since chemosensitive proteins located at the apical side of TRC are responsible for taste reception, gustatory perception of lipids must also require the existence of speciﬁc receptors displaying a high afﬁnity for LCFA in TRC. A plausible candidate for this function has recently been identiﬁed in rodent taste buds: the glycoprotein CD36 [20]. CD36 is a plasma membrane glycoprotein expressed in a wide variety of tissues. It is a multifunctional protein belonging to the family of class-B scavenger receptors. It increases the uptake of LCFA by cardiomyocytes and

CAB International 2010

150

Animal Science Reviews 2010

LCFA CD36

A

B

Gustducin

α β

Src Glycosylation site Hydrophobic pocket

? Ca2+

?

SOC IP3 Ca2+

Na + Plasma

TRPM5

membrane

Cytoplasm SCrc-tTeKr

D epolarization

« Fatty » taste signal Neurotransmitters release Figure 2 Predictive structure of CD36 and the putative CD36-mediated fat taste transduction pathway. (a) The glycoprotein CD36 displays a receptor-like structure, with a large extracellular hydrophobic pocket located between two short cytoplasmic tails. The existence of a physical interaction between the intracellular C-terminal tail of CD36 and Src proteintyrosine kinases can result in the formation of a functional complex which appears to be especially adequate to transfer an exogenous lipid signal to TRC. (b) In CD36-positive TRC immunomagnetically isolated from mouse circumvallate papillae, Src-protein tyrosine kinases (Src-PTK) Yes and Fyn are activated when long-chain fatty acids (LCFA) bind to CD36, leading to a rise in [Ca2+]i secondary to the opening of SOC channels. LCFA–CD36 interaction also increases IP3 expression level, suggesting that mobilization of endoplasmic reticulum Ca2+ could also contribute to the increase in [Ca2+]i. Like sweet taste transduction, [Ca2+]i rise can activate the TRPM5 channels, leading to a cellular depolarization at the origin of the neurotransmitters release

adipocytes [21, 22] and that of oxidized LDL by macrophages [23], modiﬁes platelet aggregation by binding to thrombospondin and collagen [24], facilitates the phagocytosis of apoptotic cells by macrophages [25] and increases the cyto-adhesion of erythrocytes infected with Plasmodium falciparum [26]. CD36 is also a receptor-like protein that binds saturated and unsaturated LCFA with an afﬁnity in the nanomolar range [27]. Interestingly, this plasma membrane lipid-binding protein displays the structural and functional features required for being a taste-based lipid receptor. Firstly, CD36 appears to be restricted to the gustatory epithelium in the mouse and rat tongues [20, 28]. Immunohistochemical staining has

g

shown that the CD36 protein is present mostly at the apical side of some of the TRC lining the taste pores in the mouse [20]. This distribution of a protein with a very high afﬁnity for LCFA is particularly suitable for the generation of a lipid signal by taste buds. Indeed, CD36-positive TRC is directly exposed to a microclimate potentially rich in LCFA, because of the local release of lingual lipase in the clefts of circumvallate papillae (Figure 1a). Secondly, a role for CD36 as a lipid sensor is also supported by the predicted structure of this protein (Figure 2a). This plasma membrane protein has a hairpin structure, with a large extracellular hydrophobic pocket located between two short cytoplasmic tails [29]. The existence of a

CAB International 2010

Philippe Besnard, Dany Gaillard, Patricia Passilly-Degrace, Ce´line Martin and Michael Chevrot

physical interaction between the intracellular C-terminal tail of CD36 and Src protein-tyrosine kinases [30] results in the formation of a functional complex, allowing the transfer of an exogenous lipid signal into the TRC (Figure 2a). Thirdly, CD36 gene invalidation abolishes both fat preference [20, 31] and the cephalic phase of digestion triggered by an oral LCFA deposition [20]. Taken together, these ﬁndings strongly suggest that CD36 is likely to be a lipid receptor involved in the orosensory perception of dietary lipids in rodents.

Evidence for the Involvement of the Gustatory Pathway in CD36-Mediated Fat Perception The key question at this stage concerned the possible mediation of oral lipid detection by the gustatory pathway. Studies of the lipid transduction signal in TRC and of the afferent nerve route used to transfer the fat signal to the central nervous system were required to address this question.

Mechanisms of Fat Signal Transduction The tastant-induced release of neurotransmitters towards afferent nerve ﬁbres leads to the orosensory perception of sapid molecules. As mentioned above, this event is known to be mediated by changes in [Ca2+]i in TRC [18]. If lingual CD36 acts as a lipid receptor, LCFA binding to CD36 may also affect [Ca2+]i. We tested this hypothesis by determining [Ca2+]i in CD36-positive TRC isolated from mouse circumvallate papillae by afﬁnity puriﬁcation with magnetic beads [7]. Saturated and unsaturated LCFA triggered a rapid and robust increase in [Ca2+]i in CD36-positive cells. This effect was strictly CD36 dependent, as it was not observed in CD36-negative cells. Moreover, addition of the speciﬁc CD36-binding inhibitor sulfo-N-succinimidyl oleic acid ester (SSO) [32] to the culture medium completely abolished the linoleic acidmediated rise in [Ca2+]i in CD36-positive cells. These data provided the ﬁrst demonstration that LCFA increases [Ca2+]i in the taste bud cells and that this event is CD36 dependent. The reception of tastes other than salty and sour requires the heterotrimeric gustducin complex [33]. CD36-expressing TRC also contain the a-subunit of gustducin [7, 20], but this G-protein complex is not involved in fat preference. Indeed, attraction for LCFAenriched solutions is maintained in a-gustducin-null mice [34]. This result was expected, because CD36 does not belong to the G-protein-coupled receptor family, unlike the T1R and T2R receptors responsible for sweet, umami and bitter taste detection. Since interactions between CD36 and Src protein-tyrosine kinases Fyn, Lyn and Yes have been reported in other cell types, an alternative mechanism for the transduction of lipid signal was

g

151

possible. Using CD36-positive TRC isolated from mouse circumvallate papillae, we have found that linoleic acid (C18 : 2, n76) induced the phosphorylation of Src protein-tyrosine kinases Fyn and Yes [35]. It has also been shown that this event was responsible for the rapid and huge rise in the [Ca2+]i, which is known to be a necessary step for the neurotransmitter release by TRC. Transient receptor potential protein-5 (TRPM-5) is known to play an important role in taste transduction. This Na+/K+ channel responds to rapid changes in [Ca2+]i by inducing transient membrane depolarization leading to neuromediator release. Interestingly, it has been reported that in TRPM57/7 mice, the attraction for LCFA is weaker than that in controls [34]. Taken together, these ﬁndings provide the ﬁrst evidence that LCFA affect the function of mouse TRC via the activation of a signalling cascade dependent on CD36 (Figure 2b).

Gustatory Nerves Convey the Fat Signal The potential involvement of gustatory nerves in the LCFA-mediated fat preference has recently been explored in rodents by studying the impact of bilateral transection of the gustatory nerves: chorda tympani (CTX) and/or glossopharyngeal (GLX) nerves. In rats, CTX decreases fat preference in a free-choice situation [36, 37]. Consistent with these data, CTX rats display much weaker conditioned aversion to linoleic acid than control rats [38]. Paradoxically, the impact of total denervation of the peripheral gustatory nerves (CTX+GLX) was not investigated in these rat studies, although such an exploration would be required for a full demonstration. In mice, the lack of functional peripheral gustatory nerves (CTX+GLX animals) fully abolishes both spontaneous linoleic acid preference and conditioned aversion [7]. These ﬁndings demonstrate that afferent gustatory nerve ﬁbres play a crucial role in the orosensory perception of LCFA in rodents. Although CD36 in TRC probably acts as a lipid receptor in these mammals, the link between lingual CD36 and transfer of a fat signal by gustatory nerves remains to be demonstrated by using a direct electophysiological recording in CD36+/+ and CD367/7 mice subjected to an oral lipid load.

Oral Lipid Deposition and Brain Activation Lipid signals therefore appear to be transmitted by a peripheral nerve route known to be involved in the transfer of gustatory information to the brain. The NST in the brainstem is the ﬁrst synaptic relay in the nervous gustatory cascade. Immunohistochemical detection of Fos, the protein encoded by the immediate early gene c-fos, has been successfully used to identify populations of neurons activated by LCFA in the NST. In mice, oral linoleic acid deposition triggers the activation of NST

CAB International 2010

152

Animal Science Reviews 2010

NAc HT NST 5

Dietary lipids

Lipid preference

Saliva

TG

Gustatory nerves (VII and IX)

CD36

Taste receptor cell

6

Lingual lipase LCFA

Cephalic phase of digestion

3

1

Digestive secretions

TG [Ca2+]i

Hormones, fatty acids…

4

Adipose tissue, liver…

LCFA

2

Intestinal lumen

Enterocyte

Figure 3 Working model for the gustatory perception of LCFA in the mouse. ˚ LCFA released from triglycerides (TG) by lingual lipase bind to CD36, which acts as a gustatory lipid receptor in taste receptor cells. ¸ The recognition of LCFA by CD36 induces an increase in [Ca2+]i, an event known to generate the release of neurotransmitters by TRC. Lipid taste signal is then transmitted by the gustatory nerves, the chorda tympani nerve (VII) and the glossopharyngeal nerve (IX), to the gustatory area in the nucleus of the solitary tract (NST) in the brainstem (˝). The projections of the NST to central nuclei involved in eating behaviour and to peripheral tissues, including the digestive tract, could account for the CD36-mediated attraction for lipids (˛) and cephalic phase of digestion (ˇ) reported in mice subjected to an oral lipid stimulation. HT, hypothalamus; NAc, nucleus accumbens

areas known to receive CT and GL afferent ﬁbres. This activation appears to be CD36 dependent, as it is not observed in CD36-null mice subjected to oral stimulation with linoleic acid [7]. The known involvement of the lateral hypothalamus and nucleus accumbens in food intake and reward, respectively, and in the reception of synaptic inputs from the NST [38] might account for the spontaneous preference for LCFA-rich food observed in mice. The digestive projections of the NST [38] might also contribute to a lipid-mediated reﬂex, controlling pancreato-biliary secretions directly and/or indirectly, through the production of intestinal hormones. Axons from the mandibulary branch of the trigeminal nerve innervating the anterior tongue project into taste areas of the NST [39], but mechanical or textural stimulation is unlikely. Indeed, water alone or mixed with xanthan gum to mimic the texture of lipids does not affect neuronal activity in the NST [7]. Thus, the fat signal triggered by the interaction of LCFA

g

with CD36 in the oral cavity is transmitted through the NST, the central and peripheral projections of which affect the activity of nervous nuclei known to be involved in feeding behaviour and the cephalic phase of digestion. A working model for the gustatory perception of LCFA in the mouse is shown in Figure 3. Other Lingual Lipid-Sensor Candidates Two other candidates for the chemioreception of dietary lipids have been identiﬁed in rodent tongues: the delayedrectifying potassium channel Kv1.5 and the G-proteincoupled receptor GPR120. Delayed-Rectifying Potassium Channels In TRC, various voltage-activated ion channels (K+, Na+ and Ca2+) contribute to the release of neurotransmitters

CAB International 2010

Philippe Besnard, Dany Gaillard, Patricia Passilly-Degrace, Ce´line Martin and Michael Chevrot

after chemical stimulation [19, 20]. LCFA are known to regulate ion channels in various cell types. Gilbertson and his colleagues from the University of Utah (USA) therefore used patch-clamp recording to explore the putative effect of free fatty acids (FFA) on membrane potential in TRC isolated from rat fungiform papillae. They reported that FFA are able to inhibit the delayed rectifying K+ (DRK) channels, which are known to be involved in the transduction pathways for various taste stimuli [40]. This action is direct and strictly mediated by polyunsaturated fatty acids (PUFA). PUFA inhibition is effective only if these molecules are applied extracellularly, as in the physiological context. This suggests that the responsiveness to PUFA of taste cells may contribute to fat preference, thereby indirectly affecting body mass. To explore this hypothesis, the effect of PUFA on DRK currents was explored by patch-clamp recording in isolated fungiform TRC from obesity-resistant (S5B/P1) and obesity-prone (Osborne–Mendel) rats. Unexpectedly, PUFA-mediated depolarization was greater in TRC from obesity-resistant rats, which are known to prefer carbohydrates, than in those from obesity-prone animals, which prefer fats [41]. This strain-speciﬁc response was attributed to a difference in the pattern of expression of DRK channel isotypes in TRC, with obesity-resistant rats having more K+ channels responsive to PUFA than obesity-prone animals [41]. Various DRK channels are found in rat fungiform papillae [42], but the shaker Kv1.5 channel, speciﬁcally inhibited by PUFA in cardiac cells [43], has been shown to be strongly expressed in TRC from obesity-resistant S5B/ P1 rats [42]. The mechanism by which PUFA inhibit Kv1.5 channels in taste bud cells remains unknown. However, a direct effect is likely, since a physical interaction between PUFA and the extracellular domain of the Kv1.5 protein has already been reported in cardiomyocytes [43]. The Kv1.5 channel may therefore be considered as an ionotropic receptor in TRC. Taken together, these data suggest that the control of Kv1.5 channels in TRC by PUFA does not explain the spontaneous fat preference observed in rodents.

G-Protein-Coupled Receptors (GPCR): GPR120 GPR120 belongs to the GPCR family [44, 45]. It is abundantly expressed in enteroendocrine cells, particularly in the distal part of the small intestine (ileum) and the colon, in both mice and humans. In these cells, GPR120 functions as a receptor for unsaturated LCFA, leading to the secretion of incretins such as glucagon-like peptide-1 [46] and cholecystokinin [47]. Preliminary studies have shown that GPR120 is also expressed in TRC in rats [48] and mice [49]. To date, the physiological role of this receptor in the taste buds remains unknown. GPR120 is also found in the mouse enteroendocrine STC-1 cell line. Interestingly, STC1 cells have several genotypic and phenotypic features in common with TRC.

g

153

For instance, they also express the genes encoding receptors for sweet taste, T1R2 and T1R3 [50], and for bitter taste, T2R [51], and the ﬁve basic taste stimuli induce an increase in [Ca2+]i [52]. In vitro studies in STC-1 cells have shown that unsaturated LCFA–GPR120 interaction also leads to an increase in [Ca2+]i [47]. This result is reminiscent of the tastant-mediated increase in [Ca2+]i in TRC. GPR120 is a receptor for unsaturated fatty acids [47] and its involvement in the orosensory perception of dietary lipids is thus plausible. However, further studies including the effect of GPR120 gene manipulation (invalidation or overexpression) on fat preference are required to validate this hypothesis.

Conclusion and Perspectives Dietary lipid perception clearly depends on multiple factors. Recent data show that a gustatory cue devoted to fat perception operates in mice and likely in humans, in parallel with texture and olfaction. LCFA play the role of tastants that provide the stimulus for fat perception. In mice, CD36 acts as a gustatory lipid receptor, enabling the organism to obtain sufﬁcient energy by selecting and promoting the digestion of lipids (Figure 2). This system, which might be considered as a sixth taste modality, would clearly be advantageous in times of food scarcity. Indeed, fat-rich foods are an important source of energy, contain essential fatty acids and carry lipid-soluble vitamins (A, D, E and K) with many important, fundamental biological functions. Conversely, this ‘fatty’ taste might contribute to obesity risk during periods of food abundance. Attraction for lipids is so strong that mice given free access to oil as an optional diet rapidly become obese [53]. In humans, there is a positive correlation between liking fat and body mass index [54]. However, the origin of this correlation remains unclear. It has also been reported that obese individuals like fat to a greater extent than lean subjects do [55, 56], suggesting that obese people live in a different orosensory world. Several questions are raised by these ﬁndings. What is the physiological signiﬁcance of the presence of several putative lipid sensors (CD36, Kv1.5 channels and GPR120) in rodent taste buds? How can we account for the multifunctionality of CD36? What are the molecular mechanisms responsible for the chemodetection of LCFA in the oral cavity in humans? Are they related to what is found in rodents? The identiﬁcation of relevant, noninvasive biological markers in humans may make it possible to answer this last question in the near future.

Acknowledgements This work was supported by the French National Research Agency (ANR) through the Research Program in Nutrition, grants from INRA/INSERM (SensoFAT) and

CAB International 2010

154

Animal Science Reviews 2010

from the Burgundy Council. D.G. and M.C. are supported by a fellowship from the French Ministry of Research and Technology and C.M. by the ANR.

17. Smith JC, Fisher EM, Maleszewski V, McClain B. Orosensory factors in the ingestion of corn oil/sucrose mixtures by the rat. Physiology and Behavior 2000;69:135–46. 18. Gilbertson TA, Boughter Jr JD. Taste transduction: appetizing times in gustation. Neuroreport 2003;14:905–11. 19. Sugita M. Taste perception and coding in the periphery. Cellular and Molecular Life Sciences 2006;63:2000–15.

References 1. Malecka-Tendera E, Mazur A. Childhood obesity: a pandemic of the twenty-ﬁrst century. International Journal of Obesity (London) 2006;30:S1–3. 2. Olshansky SJ, Passaro DJ, Hershow RC, Layden J, Carnes BA, Brody J, et al. A potential decline in life expectancy in the United States in the 21st century. New England Journal of Medicine 2005;352:1138–45. 3. Tsuruta M, Kawada T, Fukuwatari T, Fushiki T. The orosensory recognition of long-chain fatty acids in rats. Physiology and Behavior 1999;66:285–8. 4. Takeda M, Imaizumi M, Fushiki T. Preference for vegetable oils in the two-bottle choice test in mice. Life Sciences 2000;67:197–204. 5. Fukuwatari T, Shibata K, Iguchi K, Saeki T, Iwata A, Tani K, et al. Role of gustation in the recognition of oleate and triolein in anosmic rats. Physiology and Behavior 2003;78:579–83. 6. McCormack DN, Clyburn VL, Pittman DW. Detection of free fatty acids following a conditioned taste aversion in rats. Physiology and Behavior 2006;87:582–94. 7. Gaillard D, Laugerette F, Darcel N, El-Yassimi A, PassillyDegrace P, Hichami A, et al. The gustatory pathway is involved in CD36-mediated orosensory perception of long-chain fatty acids in the mouse. FASEB Journal 2008;22:1458–68. 8. Yoneda T, Saitou K, Mizushige T, Matsumura S, Manabe Y, Tsuzuki S, et al. The palatability of corn oil and linoleic acid to mice as measured by short-term two-bottle choice and licking tests. Physiology and Behavior 2007;91:304–9. 9. Kawai T, Fushiki T. Importance of lipolysis in oral cavity for orosensory detection of fat. American Journal of Physiology 2003;285:R447–54. 10. Mattes RD. Fat taste and lipid metabolism in humans. Physiology and Behavior 2005;86:691–7. 11. Chale-Rush A, Burgess JR, Mattes RD. Evidence for human orosensory (taste?) sensitivity to free fatty acids. Chemical Senses 2007;32:423–31. 12. Schiffman SS, Graham BG, Sattely-Miller EA, Warwick ZS. Orosensory perception of dietary fat. Current Directions in Psychological Science 1998;7:137–43. 13. Moreau H, Laugier R, Gargouri Y, Ferrato F, Verger R. Human preduodenal lipase is entirely of gastric fundic origin. Gastroenterology 1988;95:1221–6. 14. Hamosh M. Lingual and gastric lipases. Nutrition 1990;6:421–8. 15. Takeda M, Sawano S, Imaizumi M, Fushiki T. Preference for corn oil in olfactory-blocked mice in the conditioned place preference test and the two-bottle choice test. Life Science 2001;69:847–54. 16. Saitou K, Yoneda T, Mizushige T, Asano H, Okamura M, Matsumura S, et al. Contribution of gustation to the palatability of linoleic acid. Physiology and Behavior 2009;96:142–8.

g

20. Laugerette F, Passilly-Degrace P, Patris B, Niot I, Febbraio M, Montmayeur JP, et al. CD36 involvement in orosensory detection of dietary lipids, spontaneous fat preference, and digestive secretions. Journal of Clinical Investigation 2005;115:3177–84. 21. Coburn CT, Knapp Jr FF, Febbraio M, Beets AL, Silverstein RL, Abumrad NA. Defective uptake and utilization of long chain fatty acids in muscle and adipose tissues of CD36 knockout mice. Journal of Biological Chemistry 2000;275:32523–9. 22. Hajri T, Ibrahimi A, Coburn CT, Knapp Jr FF, Kurtz T, Pravenec M, et al. Defective fatty acid uptake in the spontaneously hypertensive rat is a primary determinant of altered glucose metabolism, hyperinsulinemia, and myocardial hypertrophy. Journal of Biological Chemistry 2001;276:23661–6. 23. Febbraio M, Hajjar DP, Silverstein RL. CD36: a class B scavenger receptor involved in angiogenesis, atherosclerosis, inﬂammation, and lipid metabolism. Journal of Clinical Investigation 2001;108:785–91. 24. Chen H, Herndon ME, Lawler J. The cell biology of thrombospondin-1. Matrix Biology 2000;19:597–614. 25. Ren Y, Silverstein RL, Allen J, Savill J. CD36 gene transfer confers capacity for phagocytosis of cells undergoing apoptosis. Journal of Experimental Medicine 1995;181:1857– 62. 26. Oquendo P, Hundt E, Lawler J, Seed B. CD36 directly mediates cytoadherence of Plasmodium falciparum parasitized erythrocytes. Cell 1989;58:95–101. 27. Baillie AG, Coburn CT, Abumrad NA. Reversible binding of long-chain fatty acids to puriﬁed FAT, the adipose CD36 homolog. Journal of Membrane Biology 1996;153:75–81. 28. Fukuwatari T, Kawada T, Tsuruta M, Hiraoka T, Iwanaga T, Sugimoto E, et al. Expression of the putative membrane fatty acid transporter (FAT) in taste buds of the circumvallate papillae in rats. FEBS Letters 1997;414:461–4. 29. Rac ME, Safranow K, Poncyljusz W. Molecular basis of human CD36 gene mutations. Molecular Medicine 2007;13:288–96. 30. Huang MM, Bolen JB, Barnwell JW, Shattil SJ, Brugge JS. Membrane glycoprotein IV (CD36) is physically associated with the Fyn, Lyn, and Yes protein-tyrosine kinases in human platelets. Proceedings of the National Academy of Sciences, USA 1991;88:7844–8. 31. Sclafani A, Ackroff K, Abumrad NA. CD36 gene deletion reduces fat preference and intake but not post-oral fat conditioning in mice. American Journal of Physiology 2007;293:R1823–32. 32. Harmon CM, Luce P, Beth AH, Abumrad NA. Labeling of adipocyte membranes by sulfo-N-succinimidyl derivatives of long-chain fatty acids: inhibition of fatty acid transport. Journal of Membrane Biology 1991;121:261–8.

CAB International 2010

Philippe Besnard, Dany Gaillard, Patricia Passilly-Degrace, Ce´line Martin and Michael Chevrot 33. Wong GT, Gannon KS, Margolskee RF. Transduction of bitter and sweet taste by gustducin. Nature 1996;381:796–800. 34. Sclafani A, Zukerman S, Glendinning JI, Margolskee RF. Fat and carbohydrate preferences in mice: the contribution of alpha-gustducin and Trpm5 taste-signaling proteins. American Journal of Physiology 2007;293:R1504–13. 35. El-Yassimi A, Hichami A, Besnard P, Khan NA. Linoleic acid induces calcium signaling, SRC-kinase phosphorylation and neurotransmitters release in mouse CD36-positive gustatory cells. Journal of Biological Chemistry 2008;283:12949–59. 36. Stratford JM, Curtis KS, Contreras RJ. Chorda tympani nerve transection alters linoleic acid taste discrimination by male and female rats. Physiology and Behavior 2006;89:311–9. 37. Pittman D, Crawley ME, Corbin CH, Smith KR. Chorda tympani nerve transection impairs the gustatory detection of free fatty acids in male and female rats. Brain Research 2007;1151:74–3. 38. Berthoud HR. Multiple neural systems controlling food intake and body weight. Neuroscience Biobehavior Review 2002;26:393–428. 39. Hamilton RB, Norgren R. Central projections of gustatory nerves in the rat. Journal of Comparative Neurology 1984;222:560–77. 40. Gilbertson TA, Fontenot DT, Liu L, Zhang H, Monroe WT. Fatty acid modulation of K+ channels in taste receptor cells: gustatory cues for dietary fat. American Journal of Physiology 1997;272:C1203–10. 41. Gilbertson TA, Liu L, Kim I, Burks CA, Hansen DR. Fatty acid responses in taste cells from obesity-prone and -resistant rats. Physiological Behavior 2005;86:681–90. 42. Liu L, Hansen DR, Kim I, Gilbertson TA. Expression and characterization of delayed rectifying K+ channels in anterior rat taste buds. American Journal of Physiology 2005;289:C868–80. 43. Honore E, Barhanin J, Attali B, Lesage F, Lazdunski M. External blockade of the major cardiac delayed-rectiﬁer K+ channel (Kv1.5) by polyunsaturated fatty acids. Proceedings of the National Academy of Sciences, USA 1994;91:1937–41. 44. Fredriksson R, Hoglund PJ, Gloriam DE, Lagerstrom MC, Schioth HB. Seven evolutionarily conserved human rhodopsin G protein-coupled receptors lacking close relatives. FEBS Letters 2003;554:381–8.

g

155

45. Rayasam GV, Tulasi VK, Davis JA, Bansal VS. Fatty acid receptors as new therapeutic targets for diabetes. Expert Opinions on Therapeutic Targets 2007;11:661–71. 46. Hirasawa A, Tsumaya K, Awaji T, Katsuma S, Adachi T, Yamada M, et al. Free fatty acids regulate gut incretin glucagon-like peptide-1 secretion through GPR120. Nature Medicine 2005;11:90–4. 47. Tanaka T, Katsuma S, Adachi T, Koshimizu TA, Hirasawa A, Tsujimoto G. Free fatty acids induce cholecystokinin secretion through GPR120. Naunyn Schmiedebergs Archives of Pharmacology 2008;377:523–7. 48. Matsumura S, Mizushige T, Yoneda T, Iwanaga T, Tsuzuki S, Inoue K, et al. GPR expression in the rat taste bud relating to fatty acid sensing. Biomedical Research 2007;28:49–5. 49. Matsumura S, Eguchi A, Mizushige T, Kitabayashi N, Tsuzuki S, Inoue K, et al. Colocalization of GPR120 with phospholipase-C beta2 and alpha-gustducin in the taste bud cells in mice. Neuroscience Letters 2009;450:186–90. 50. Dyer J, Salmon KS, Zibrik L, Shirazi-Beechey SP. Expression of sweet taste receptors of the T1R family in the intestinal tract and enteroendocrine cells. Biochemical Society Transactions 2005;33:302–5. 51. Wu SV, Rozengurt N, Yang M, Young SH, Sinnett-Smith J, Rozengurt E. Expression of bitter taste receptors of the T2R family in the gastrointestinal tract and enteroendocrine STC-1 cells. Proceedings of the National Academy of Sciences, USA 2002;99:2392–7. 52. Saitoh O, Hirano A, Nishimura Y. Intestinal STC-1 cells respond to ﬁve basic taste stimuli. Neuroreport 2007;18:1991–5. 53. Takeda M, Imaizumi M, Sawano S, Manabe Y, Fushiki T. Long-term optional ingestion of corn oil induces excessive caloric intake and obesity in mice. Nutrition 2001;17:117–20. 54. Barthoshuk LM, Duffy VB, Hayes JE, Moskowitz HR, Snyder DJ. Psychophysics of sweet and fat perception in obesity: problems, solutions and new perspectives. Philosophical Transactions of the Royal Society B 2006;361:1137–48. 55. Drewnowski A, Brunzell JD, Sande K, Iverius PH, Greenwood MR. Sweet tooth reconsidered: taste responsiveness in human obesity. Physiology and Behavior 1985;35:617–22. 56. Mela DJ. Sensory assessment of fat content in ﬂuid dairy products. Appetite 1988;10:37–44.

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Microgoats in India and their role in future animal production in an era of climatic change Pramod Kumar Rout* and M.C. Sharma Address: Central Institute for Research on Goats (CIRG), Makhdoom, Farah, Mathura, Uttar Pradesh, 281122, India. *Correspondence: Dr Pramod Kumar Rout. Email: [email protected] and [email protected] 22 May 2008 22 October 2009

Received: Accepted:

Abstract The major objective of this review is to analyse how microlivestock can help us to reduce poverty through sustainable development and enhance food production to meet the demands of the growing population. The role of microgoats has been examined on the basis of sustainability indicators in relation to climatic variation. The sustainability of microgoats has been analyzed on the basis of ecological adaptability, economically viability and energy efﬁciency in particular production environments and their suitability in diverse environmental conditions. The present review also highlights the importance of microgoats for future agriculture. Microgoats play multifaceted roles such as providing nutrition, fulﬁlling family needs, maintaining soil fertility and playing a vital role during disasters in disadvantaged areas. Microgoats are distributed in areas that show extreme climatic variation and seem to be efﬁcient in food production under particular situations. Livestock production is increasingly going to face challenges of heat stress, water scarcity, coastal ﬂooding, disease onset and food security. Therefore it is necessary to evaluate the different genotypes present in speciﬁc locations to establish which of them will be suitable in future for national and global applications. It is necessary to analyse suitable livestock components for sustainable agriculture and analyse closely the consequence of climate change as it will cause problems for the environment and subsequently disturb the peace, harmony and security of the whole world. Keywords: Micro livestock, Micro goats, Sustainable production, Climate change, Adaptability Review Methodology: The speciﬁc subject-related search was carried out in CAB Abstracts, Google Search, PubMed, Agricola, FAO website, Science Direct and some speciﬁc journal websites, Intergovernmental Panel on Climate Change (IPCC) and other websites. We have also searched the different reports from institutes and government departments to check for relevant data in this area.

Introduction Livestock production is an important component of agriculture as it provides 30% of the total protein requirement for humans and the demand for livestock products is increasing all over the world. Livestock play an important role in the Indian economy as they are the major source of food production as well as employment generation and contribute signiﬁcantly to the gross domestic product. As a result of an expanding economy and higher population growth, there is strong demand for livestock products in India and it has also been predicted that the demand for livestock products will continue over the coming

g

decades [1]. The livestock sector in India has to expand to meet the growing population’s requirements by adapting to the environmental changes. The nature of livestock production in the Indian subcontinent varies with environment and culture. India is the highest milk-producing as well as milk-consuming country in the world [2]. India contributed about 15% to the total world milk output in 2004, compared with less than 7% in 1980 [3]. However, individual milk yield has not increased in comparison to other developed countries. The meat consumption in India remains the lowest in the world [4]. In India, the consumer prefers goat or sheep meat over buffalo meat. The most widely consumed meats in India are goat

CAB International 2010

158

Animal Science Reviews 2010

and poultry [5]. In India, the importance of livestock production goes beyond its food production function. Livestock are used for draught power, ﬁbre and fur production, organic manure, dung as fuel and other byproducts such as leather, bone and horn [6]. In India, almost everyone depends on livestock for their livelihood. Farmers will beneﬁt a lot from the rapidly growing livestock market and the opportunity for global trade through retail. Therefore the current situation presents a unique opportunity to reduce poverty through livestock production and marketing. Simultaneously, the present situation is really complex with respect to the following factors: (i) food production is showing almost no growth and there may be a deﬁciency in food grain for human consumption; (ii) the possibility of cultivating biofuel is adding additional problems to this situation; and (iii) the grazing area is diminishing because of increasing cropping area to accommodate a large and growing population. At the same time, water supply is reducing and creating problems for management of the situation and climate change in future will bring harsher environments. Therefore the present situation demands the analysis of adaptability of different livestock species in varied production environments and the formulation of breeding strategies to enhance adaptive attributes (adaptive genetics) and disease attributes (disease genetics) for sustainable livestock production.

Projected Changes Associated with Climate Change and their Effect on Agriculture in India Agriculture is the main source of livelihood for more than 70% of the population in India. Climate change is one of the most serious threats mankind will face in the near future. Those who will suffer most from the impacts of climate change will be the poorest people living in Africa and Asia. Climate change will have effects on various aspects of agriculture ranging from water stress to coastal ﬂooding. Climate change will cause variation in rainfall patterns and shifting of temperature zones, which will have signiﬁcant negative effects on agriculture. The situation will need more attention because of changes in precipitation patterns and increasing salinity of ground water because of the increase in sea level. The most signiﬁcant impact of climate change is expected to be in respect of availability of water. The melting of glaciers will lead to decreased river ﬂow. The Intergovernmental Panel on Climate Change (IPCC) estimate shows that 500 million people of South Asia will be affected by reduced river ﬂow. The rise in sea level will lead to coastal ﬂooding and it has been determined by the IPCC that the mega deltas of India will be vulnerable [7, 8]. The threat of coastal ﬂooding will have serious impacts on crop and livestock production. Moreover, the increase in frequency of drought and ﬂood will hamper the prospects of agricultural production. The devastating effect of climate

g

change will be mainly on agriculture. In the Central and South Asia region, yield could decrease by up to 50% by 2050 [7]. The Peterson Institute for International Economics predicts that agricultural production in developing countries may fall by 10–25% and, if global warming progress is unabated, ‘India’s agriculture capacity would fall as much as 40%’ [9]. India is a country of South Asia and will face challenges of rising sea level, water stress and heat waves. The ambient temperature in India is projected to increase by 0.10–0.30 C in Kharif (summer) and 0.30–0.70 C in Rabi (winter) by 2010 and to 0.40–2.00 C in Kharif and 1.10–4.50 C in Rabi by 2070 [10]. Mean rainfall is projected not to change by 2010 but may increase by 10% during Rabi by 2070. At the same time, there is an increased possibility of climate extremes, such as the timing of onset of monsoon and intensities and frequencies of droughts and ﬂoods. Agricultural productivity can be affected by climate change because of changes in temperature, precipitation and/or CO2 levels and indirectly through changes in soil and infestation by pests, insects, diseases or weeds. The climate of India varies from tropical monsoon to temperate dry in the Himalayan region. Livestock production is the world’s largest user of land either through grazing or through consumption of food grains and fodder. The livestock sector will be affected by climate change; however, the livestock production has to be more productive to meet the demand of growing population. Livestock production in India will be affected by the following factors as a result of the changing environment: a more warm, humid and disease-prone environment; shrinking of grazing land and water resources; the production system will be governed by market demand (move towards more intensive and integrated processing); heat stress will be faced by livestock, particularly in arid and semiarid areas; the salinity of the underground water and the soil will increase; thus the vegetation in the grassland area will change as a result of salinity and drought and the woody plants and plants with long roots will dominate grasslands; because of increased evotranspiration, there will be increased consumption of water by livestock; The Bay of Bengal will experience problems through the rise in sea level. Microlivestock and its Advantages The human population is increasing and the available space for the rearing of livestock will decrease, which will encourage the consideration of the alternative of adopting microlivestock to maintain productivity against a background of less land, water and nutrients and of climate variation. In this scenario, microlivestock will be an

CAB International 2010

Pramod Kumar Rout and M.C. Sharma

important contribution to meet human protein requirements. Moreover, the research and development programmes have overlooked microlivestock, with the exception of poultry. In the speciﬁc drive towards large animals, the gene pool of small species and breeds has been mostly bypassed. Microlivestock is generally used for subsistence agriculture and has never attracted adequate attention for commercial production. Microlivestock is the main source of food and income for poor farmers in different parts of the world. Microlivestock primarily ensures nutritional security for small farmers as they manage to rear the animals in their homes or backyards with available foodstuffs from their homes. The marginal poor farmers are able to manage small animals because of their cheaper price and the lower ﬁnancial risk of maintaining them. The landless farmer can afford to raise such microanimals in their backyard under various climatic conditions. Small size generally signiﬁes high reproductive capacity and fast turnover. Microlivestock provide ﬂexibility to farmers in terms of space and feed management. Moreover, they can be sold at any point of time to cater for the needs of the family. They are efﬁcient converters of food energy. They increase the chances of successful breeding and provide a steady source of income. Small animals tend to ﬁt well into existing farming systems and can thrive on available nutrient/by-products of the area. Small animals have high reproductive capacity with short gestation periods and produce a large number of offspring and have high juvenile growth. They attain early sexual maturity and have a short generation interval [11, 12]. Microlivestock has not been given sufﬁcient research attention and its potential to bring a livestock revolution has not been considered. There is considerable scope for developing microlivestock production to improve the income level of resource-poor farmers. Small animals are well adapted to hostile weather, ravaging pestilence and poor diets. Especially in remote areas and in areas of extreme climate, they are often vital for basic subsistence. In situations of less land and forage, higher stress and disease environment, microlivestock may be the only practical livestock to enhance productivity [11]. The individual productivity may be low, but it is an efﬁcient production system with respect to the environment and other inputs. Microlivestock production should be integrated with different rural development projects in the areas where no other mechanism is available for improving the living standard of farmers. The farmers are raising small animals with the locally available feed and fodder to produce quality protein to fulﬁl their need. It is not a question of small or large livestock; it is a question of how effectively we can utilize the resources to meet the requirement of local farmers. Microlivestock never competes with traditional large livestock; rather, it complements traditional large livestock because of its unique physical, physiological, behavioural and economic characteristics [11, 12].

g

159

Microgoat Breeds in India Goat breeds with varying specialties have evolved in different agro-climatic zones, primarily through natural selection. On the basis of agro-climatic and geographical conditions, the country has been divided into four major eco-regions: (i) the temperate Himalayan region, (ii) the Northwestern region, (iii) the Southern region and (iv) the Eastern region. Goats are widely distributed all over the country in different agro-climatic regions [13, 14]. The goats of the temperate Himalayan region (which has low rainfall) produce the ﬁnest quality of undercoat, known as ‘Cashmere’ or ‘Pashmina’. The goat breeds found in the North and Northwestern regions are reasonably large in size and are primarily for dairy purpose. In the southern and peninsular parts of India, goat breeds with dual utility (meat and milk) are found. Highly proliﬁc meat breeds are found in the Eastern region of the country. Marwari in the Western dry region of Rajasthan, Kutchi in the Gujarat Plains and Black Bengal in the Lower Gangetic Plains of West Bengal are the important breeds for the purpose of meat production. There are potential breeds of goat for milk production like Jamunapari in the Upper Gangetic Plain, Jakhrana in the central plateau and Osmanabadi in the western plateau. Microgoat breeds in India have been described in a US National Research Council (NRC) report [11]; however, some further breeds with potential application to various climate have been described. Barbari Goat Barbari is a small- to medium-sized breed in the semi-arid region in the Northwest part of India and is distributed over Etah, Agra, Mathura and Aligarh districts of Uttar Pradesh. The coat colour is white with light to dark brown spots all over the body. Ears are short and erect. The body weight at 1 year of age is around 15–18 kg in female and 16–20 kg in male. The twining percentage is about 56% and milk yield varies from 0.5 to 1 litre per day. Black Bengal Black Bengal is a small-sized breed, distributed in the Eastern region in West Bengal, Assam, Jharkhand and Orissa. Coat colours are mainly black, white and brown, but black is the most predominant colour. The face is small, with a beard. Ears are short, ﬂat and horizontally placed. The body weight at 1 year of age was 11–13 kg in female and 12–14 kg in male. Multiple birth percentage is about 78%. It produces less milk, but the milk is rich with high nutritive value. Chegu Chegu is found in the temperate Himalayan region. White in colour, it has twisted long horns, short ears, long

CAB International 2010

160

Animal Science Reviews 2010

lustrous hair with ﬁne undercoat, and its forehead is concave, while its horns are long and ﬂat. The tail is short and straight. The breed is distributed in Lahul, the Spiti valley of Himachal Pradesh and hilly regions of Uttar Kashi, Chamoli and Pithoragarh districts of Uttaranchal. The body weight at 1 year of age was 12–13 kg in female and 14–15 kg in male. Multiple birth percentage is very low (1%). It is best known for ﬁbre production and the yield is about 120 g per year.

Changthangi The natural habitat of the Changthangi breed is the Kargil, Leh and Changthang tehsils of Ladakh in Jammu and Kashmir state. The body is covered with long lustrous hair and the colour is mainly white, but grey and brown are also observed, and the ears are small and pointed. The body weight at 1 year of age was 15 kg in female and 16 kg in male. Multiple birth percentage is very low (1%). It produces best-quality ﬁbre of 402 g per year with a diameter of 12.86 mm.

Surti The Surti breed is distributed in the Surat and Navsari districts in Gujarat and Nasik district in Maharashtra. The coat colour is white, with short lustrous hair, ears are medium-sized and drooping, the forehead convex and medium-sized horns are directed upward. The body weight at 1 year of age was 12–15 kg in females. Multiple birth percentage is 25–37%.

Khasi (Assam local) The Khasi (Assam local) is also locally known as Khasi goat and typically a dwarf breed and is white with silver or grey in colour. The body weight is around 10 kg at 1 year of age. Multiple births are high and the breed is known for quality meat production. It is distributed over Naga, Lushai Hills, Meghalaya and Nagaland.

Gaddi The Gaddi breed is distributed over Kangra, Kullu valley, Chamba, Sirmur and Simla districts of Himachal Pradesh. The breed has a well-built body with long hair, and is white and black brown in colour, with a convex nose, alert eyes, long and drooping ears, and long spiral horns. Twining percentage is 10–15%. The breed is best known for ﬁbre and meat production. Besides these breeds, there are some medium-sized breeds such as Marwari and Kutchi in the arid region, Ganjam in the Eastern region and Kannai adu, Malabari,

g

Osmanabadi and Sangamneri in the southern peninsular region.

Microgoats are a Species of Choice in the Face of Climate Change The goat as a species has become important because of concern over sustainable development and for the uplift of the rural population in terms of economic and nutritional stability. The goat has a wide genetic base, as it is distributed over varied geographical and climatic conditions with differential production capability. Because of its wide genetic variability, goats can be manipulated to develop suitable adaptable animals in relation to different ecotypes. In certain disadvantaged places, goat is a priority animal as it fulﬁls the basic nutritional requirement of families through its milk and meat and works as a cash buffer in times of need. The livestock species or breeds will be remunerative if they have a diverse gene pool to ﬁt into a desired climatic region, are adaptable and ecologically stable, are economically viable and are energy-efﬁcient in a particular production environment. Microgoats in different production system can be examined based on the above criteria for their suitability as a species for providing nutritional security as well as income in the face of environment change.

Microgoats as a Resource for the Farmer Microgoats are distributed in all types of extreme climate and are well adapted to the speciﬁc environments that India is going to face in the near future as a result of climate change. Microgoats are well known for their high reproduction rates, rapid growth, early maturity, tasty meat, rich milk and for their adaptability over a wide climatic range. Microgoats are a poor person’s source of milk, meat and cash income. Goats have great potential for diverse ecological adaptability and are spreading over a wide range of agroclimatic situations in India. Goats are found at high altitudes in the Himalayas, arid regions of Rajasthan, the semi-arid region of Uttar Pradesh, Andhra Pradesh and Tamilnadu as well as in the high rainfall and humid areas of Assam, West Bengal and Orissa. Goats are concentrated mainly in the high-rainfall lowlands, including the semiarid and arid zones, and in the upland areas. In the Himalayan region, microgoats are especially important for ﬁbre and meat production and also for transportation. In other parts of the country it is equally important for meat and milk also. Goat farming has more signiﬁcance in regions with extreme climate and topographical conditions. The importance is evident in extreme ecology and arid zones and upland regions of India. Goats are of special economic importance in arid and semiarid areas and in dryland agriculture, where other livestock farming is not economic.

CAB International 2010

Pramod Kumar Rout and M.C. Sharma

Microgoats are the main source of meat, and goat meat is widely accepted by religious societies in India. Black Bengal goats are known for their quality meat and skin production. Some of the microgoats such as Barbari and other indeterminate breeds produce both milk and meat, and they produce milk in stressful environmental conditions when other livestock are dry. Microgoats also produce the ﬁnest and most valuable ﬁbres. Microgoats also produce durable and ﬁne-textured leather. Manure is also easy to collect and has a high commercial value. Goats are most efﬁcient and economical for smallholders in all types of production systems. These animals have faster rate of growth and early reproductive age under harsh climatic conditions.

Microgoats are Ecologically Stable and Adaptable to Various Environment Ecological stability of the farming system should be the foremost criterion for sustainable development. Goats can subsist on many types of foodstuff that are not utilized by any other species of livestock. Microgoats have high reproductive performance as they rarely face any reproductive problem and produce twins, triplets and quadruplets. In hot and dry areas, goats have the added advantage of better heat dissipation [15]. The small size is signiﬁcant in dry areas; the wide body area is given preference because of the higher evaporation rate in a hot and humid climate. Goat breeds in the eastern region are also able to manage in hot, humid and waterlogged conditions. They are resilient to heavy parasitic load and are able to produce with a heavy load of liver ﬂuke and nematodes. Microgoats can graze in rough terrain such as desert ﬁelds for longer periods without water. They can derive most or all of their diet from roughage unusable by humans. Goats also show a considerable degree of resistance to foot and mouth disease, mange and internal parasites. Goats generally have a long snout and upright tail. The mouth is unusual in having a mobile upper lip and grasping tongue, which permits the animal to nibble tiny leaves from thorny trees. However, the goat, among all the domestic animals, has been labelled as an enemy of ecology with regard to its grazing habit. But this concept is no longer valid as several scientiﬁc works have indicated that small ruminants do not play a major role in ecological destruction and their role in ecological degradation has been exaggerated [16]. Goats are not responsible for land degradation, but rather perform an important function in land management by improving the grazing land by dispensing seeds and controlling many weeds such as Nassella trichotoma, Rubus fruticosus and Rosa rubiginosa. Goats integrate well in mixed agriculture as they consume leafy wastes, clear weeds and contribute fertilizer. Mixed farming systems dominate the livestock production system in both arid/semiarid and humid/sub-humid climatic zones of India. Microgoat rearing is also proﬁtable in

g

161

mixed rainfed farming systems and the region is especially difﬁcult for livestock because of high temperatures and high humidity. In rainfed and irrigated mixed farming systems, goats are widespread and integrated with the crop/animal system in India [17]. Goats also have an important role in the livelihood of the grassland production system and are mainly used for meat and milk production [18]. There is no correlation between goat density and deforestation. Higher concentrations of goats are found in semiarid and rainfed regions of India and these regions are targets for agricultural exploitation through increasing productivity. It is necessary to get maximum beneﬁt by providing efﬁciency to the system by putting an animal component. Goat is one of the most suitable animal components for dryland agriculture systems. So our target must be to increase the efﬁciency of the system, which will contribute towards sustainability through beneﬁcial crop animal interaction. Goats can be reared along with other livestock without any serious competition because of their preferential grazing habit. Goats spend more time in browsing, whereas sheep and cattle graze grass.

Viable Economy The goat farming system is generally capable of fulﬁlling most of the basic needs and providing work for all the local inhabitants with the guarantee of a satisfactory living standard. Goat rearing continues to be a very remunerative source of income. Goats are generally used as an adjunct to a cropping system where they can make use of natural vegetation and crop residues supplemented by tree leaves. Mostly Indian farmers invest their maximum time and labour in food production. Usually goats are raised as an extra investment without a major labour input, based on roadside and fence vegetation, which provide a sizeable portion of the energy consumed by them. Goat farming is very remunerative among all the livestock farming systems in the entire farm situation. As it is a low-investment proposition for farmers it is a very assured way of gaining income for marginal and small farmers who survive with part-time employment as agricultural labour. The small size is especially signiﬁcant, as it relates directly to economic, managerial and biological advantages. This is especially true in marginal and rangeland areas where rainfall and available feeds are sparse. Goats provide the means of livelihood in such marginal areas and harsh environments. They provide food and economic security and proper nutrition to the survival of human population. The economics of a 1-buffalo unit versus a 5-goat unit in Rajasthan was studied and it was observed that the proﬁt per annum was higher from a 5-goat unit than from a 1-buffalo unit (Rs 1945 from the goat unit and Rs 1755 from the buffalo unit) [19]. Similarly, a study of a drought-prone area showed that the total gross income of

CAB International 2010

162

Animal Science Reviews 2010

small and marginal farmers was accounted for by small ruminants [19]. The average gross income per household per rupee of expenditure was highest in goat farming as compared with other livestock and crop farming systems [20] (One rupee i.e smallest unit of Indian currency which is equivalent to .025 USD). In arid regions, the goats were more economical than sheep and cattle. Goats in the desert kept under range conditions are 40–160% more economical than sheep and an indigenous goat may provide cash beneﬁts besides 2 quintal of manure and beneﬁts of clearing obnoxious weeds and thorny bushes [21].

Microgoats are Energy-Efﬁcient in the Production Environment Generally, yield is considered as a major indicator of economic gain. But whether a livestock production system in a particular local ecological condition is efﬁcient will be measured as output–input energy ratios, i.e. the energy efﬁciency of the system. So a deep understanding of the ecosystem and the farmer’s indigenous practice and knowledge are required to design highly energy-efﬁcient systems. Input such as support energy, land and labour are not that costly for the goat production system. As far as the grazing habit is concerned, goats can thrive on types of vegetation that are not suitable for other species. As ruminants they can use roughage and non-conventional feeds for energy utilization and growth. The lands mainly targeted for goat grazing are rangeland and wasteland. Another major input is labour, which is at a low level in goat farming [22]. Goat production does not require skilled labour, as it is mainly handled by women and children in village conditions. Similarly, the output of the system is milk, meat, ﬁbre, manure, skin and breeding stock. Demand for the meet is high, so the question of overproduction of animal does not arise in the present situation. Other byproducts are hides, offal and bones, which do not cause pollution from the processing point of view. However, the excreta of goat urine and dung have been proved to be rich sources of nitrogen of soil fertility and also release only low levels of greenhouse gas to the environment.

Research and Development Needed for Improvement and Conservation Goat productivity in India varies greatly in all respects in different agro-climatic zones. To enhance productivity, it is equally important to give greater attention to the biological attributes of indigenous breeds and the need to exploit them for local advantage. The productivity of different goat breeds in different production systems is illustrative of their importance. So it is necessary to evaluate the performance of the different breeds in different production systems in relation to feed resources available. In India, microcattle and microsheep are known for their

g

attributes. Microcattle such as Pugnaur and Vechur are now in a critical stage and are almost facing extinction. Microsheep, such as Garole and Kendrapara, are genotypes adaptable to hot and humid environments that produce twins for their multiplication. We must evaluate microbreeds for the best gene pool for future global application for their greater use and conservation, and their unusual or special characteristics should be noted and the genetic origin of breed should be established. It is necessary to evaluate microbreeds or goats of undetermined breeds that are adapted to speciﬁc climatic conditions. Breeds from the Eastern region need to be evaluated in hot, humid, waterlogged areas and high-rainfall areas with respect to disease resistance and feed requirement. Similarly, in hilly areas of Assam and the northeast region, Khasi (Assam local) goats should be evaluated in different production environments with respect to heavy rainfall and grazing ability. Microbreeds from arid and semiarid regions should be evaluated for heat stress, water requirement and endurance of grazing for a whole day in the desert. Himalayan breeds and goats of indeterminate breed should be evaluated in cold dry and cold humid environments with respect to feed availability. Similarly, breeds from the southern peninsular region should be evaluated in speciﬁc environments for their valued characteristics besides production performance. The characterization of production performance as well as ﬁeld condition should be carried out in the speciﬁc production environment. Characterization of milk protein polymorphism should be carried out in all speciﬁc production environments in all the breeds including non-descriptive goats. Parasitic characterization of ﬁeld condition as well as host response in kids up to 1 year of age and in different physiological conditions should be carried out. Adaptability to different stresses, water requirement of different breeds with respect to environmental conditions and endurance ability of goats in desert regions and cold regions should be analysed at the molecular level. Molecular characterization with respect to DNA markers, epigenetic analysis and the role of different RNAs in expression of these traits in different environmental conditions should be carried out. Food and fodder resource characterization is also necessary along with disease pattern analysis in emerging situations. The major goal in the current situation is to reduce poverty through sustainable development and to prioritize local problems with national and global applications. We must focus on research on how microlivestock can help to reduce poverty and it is necessary to analyse the problem using geographical information systems.

Conclusion Microgoats managed in the natural ecosystem are an efﬁcient and sustainable method to produce high-quality protein with minimal environmental impacts. Microgoats

CAB International 2010

Pramod Kumar Rout and M.C. Sharma

are adaptable to emerging environmental situations caused by climatic change and therefore they will be proﬁtable for rearing and multiplying for gaining more income as well as generating more employment. As they are ecologically stable to different types of environmental conditions, microgoat rearing will be a proﬁtable option for small farmers. Keeping in view the emerging challenges such as climate change, demand for biofuel, pressure on land and water resources, microgoat rearing is proﬁtable and will be least affected by these factors. More importantly, microgoat rearing is economically viable in all types of production systems and is therefore a viable option for the smallholder to rear it and multiply for commercial use. Microgoat rearing will enhance the farmer’s income and provide more employment to women as it is proﬁtable and less vulnerable to shocks with respect to feed requirement and their management. Microgoats will play a major role in poverty reduction by providing a better source of income and nutrition to the poor people in the world. Therefore, microgoats will play a signiﬁcant role in future animal production as they are proﬁtable, sustainable and well managed by farmers. Microgoats have the potential for delivering a sustainable increase in income and providing safe high-quality and affordable livestock products.

163

Vanderlindan PJ, Hanson CE, editors. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK; 2007. 8. Sathaye J, Shukla PR, Ravindranath NH. Climate change, sustainable development and India: global and national concerns. Current Science 2006;90(3):314–25. 9. Siddique K. Challenge and opportunity in agriculture. The Hindu, 2 May 2008. 2008. Available from: URL: http://www. thehindu.com/2008/05/02/stories/2008050251651000.htm 10. IPCC. Climate Change 1995: Impacts, Adaptation and Mitigation of Climate Change: Scientiﬁc-Technical Analysis. Report of the Working Group II of the Intergovernmental Panel on Climate Change. Cambridge University Press, London and New York; 1996. 11. National Research Council. Microlivestock: Little-Known Small Animals with Promising Economic Future. National Academy Press, Washington, DC; 1991. Available from: URL: http://www.nap.edu/openbook.php?isbn=030904295X 12. Devendra C. Small ruminants potential value and contribution to sustainable development. Outlook on Agriculture 1994;23(2):97–103. 13. Acharya RM. Sheep and goat breeds of India. FAO Animal Production and Health Paper No. 30. Food and Agriculture Organization, UN, Rome; 1982. Available from: URL: http://www.fao.org/docrep/004/x6532e/X6532E00.htm 14. Gall C. Goat Production. Academic Press, London, 1981. p. 619. 15. Cole HH, Garret WN. Animal Agriculture. W.H. Freeman and Company, San Francisco, CA, USA; 1980.

References 1. FAO. Geographical trends in livestock densities and nutrient balance in South, East and South-East Asia. By P. Gerber, P. Chilonda, G. Franceschini & H. Menzi. LEAD. Electronic News Letter V3N1, March 2005. Available from: URL: http://pigtrop.cirad.fr/subjects/environment_and_ natural_resources_protection/geographical_trends_in_ livestock_densities_and_ nutrient_balances 2. Delgado C, Rosegrant M, Steinfeld H, Ehui S, Courbois C. Livestock to 2020: The Next Food Revolution. Food, Agriculture and Environment Discussion Paper 28, International Food Policy Research Institute (IFPRI), Washington, DC; 1999. http://www.ifpri.cgia.org. Available from: URL: ftp://ftp.fao.org/docrep/nonfao/lead/x6155e/ x6155e00.pdf 3. FAOSTAT. http://www.faostat.fao.org/ 4. Speedy AW. Global production and consumption of animal source foods. Journal of Nutrition 2003;133:40485–535. 5. FAO. World Agriculture: Towards 2015/2030. An FAO Perspective. FAO, Rome; 2003. Available from: URL: http:// www.fao.org/docrep/005/Y4252E/y4252e00.htm 6. Country Report. Country Report on the State of Animal Genetic Resources. (Available in DAD-IS Library; available from: URL: http://www.fao.org/dad-IS). FAO, Rome; 2004. 7. IPCC. Climate change 2007: impacts, adaptation and vulnerability. In: Pary ML, Canziani OF, Palutikof JP,

g

16. Government of India. Report of the Task Force to Evaluate the Impact of Sheep and Goat Rearing in Ecologically Fragile Zones. Government of India, Ministry of Agriculture, Department of Agriculture and Cooperation, New Delhi; 1987. 17. Devendra C, Morton JF, Rischowsky B. Chapter 3: Livestock systems. In: Owen E, Kitalyi A, Jayasuria N, Smith T, editors. Livestock and Wealth Creation: Improving the Husbandry of Animals Kept by Resource Poor People in Developing Countries. Nottingham University Press, Nottingham, UK; 2005. 18. Robbins P. Goats and grasses in western Rajasthan: interpreting change. In A Collection of Papers from Gujarat and Rajasthan. ODI, London; 1992. p. 612. (Available from: URL: http://www.odi.org.uk/papers/369.pdf). 19. Pasha SA. Sustainability and viability of small marginal farmers with reference to animal husbandry and common property resources. Paper presented in National Seminar held on October 8 to October 10 at NIRD, Hyderabad, India; 1990. 20. Singh K, Ram K. Economic analysis of goat keeping in the goat breeding areas of Punjab. Indian Journal of Animal Sciences 1987;57:317–323. 21. Ghosh PK, Khan MS. The goat in desert environment. Research Bulletin No. 12, CAZRI, Jodhpur; 1980. 22. Sastry NSR. Livestock Sector of India: Regional Aspects. International Book House, Lucknow; 1995.

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Assessing biosafety of GM plants containing lectins Morten Poulsen* and Jan W. Pedersen Address: Department of Toxicology and Risk Assessment, The National Food Institute, Technical University of Denmark, Mørkhøj Bygade 19, DK-2860 Søborg, Denmark. *Correspondence: Morten Poulsen, The National Food Institute, Technical University of Denmark, Mørkhøj Bygade 19, 2860 Søborg, Denmark. Email: [email protected] 16 November 2009 22 April 2010

Received: Accepted:

Abstract The introduction of genetic engineering has already shown its beneﬁts in transferring genes into crop plants and conferring resistance towards pests. Most of these crop plants on the market have been transformed with the cry genes from Bacillus species, conferring resistance towards certain insects. However, since the cry genes are not active against all insects, e.g. sap-sucking insects, other genes coding for proteins such as lectins show promise of complementing the cry genes for insect resistance. As with other novel plants, lectin-expressing plants will need to be assessed for their potential risks to human and animal health and the environment. The expressed lectin protein should be assessed on its own for potential toxicity and allergenicity as for any other new protein. Although not many lectins have been thoroughly tested for their toxicity, our evaluation suggests that most of the lectins that are potentially useful for insect resistance will pose no health risk in genetically modiﬁed (GM) plants. Since some lectins are known for their toxicity to humans, the insertion of lectin genes in food crop plants will have to be assessed carefully. It is expected that in some cases there will be a need to perform animal tests of such GM plants in order to eliminate any uncertainties about potential safety issues for these plants. A 90-day study designed and optimized for this purpose is suggested as one way to cope with these uncertainties. Keywords: GM plant, Lectin, Cry gene

Introduction The need to improve our crops for higher and more constant yield will never end. One of the everlasting challenges in this continuous effort to improve our crops is to ﬁnd new traits or ways to ﬁght pests that are responsible for high losses in our crops. Genetic engineering has already shown its value in this area through the introduction into crop plants of genes from the natural entomotoxic bacterium Bacillus thuringiensis (Bt). These genes encode various proteins, the so-called Bt toxins, which have a high degree of speciﬁcity in the sense that they are only active against certain groups of insects. This high speciﬁcity is highly useful in that it avoids negative effects on beneﬁcial insects such as bees. On the other hand, their speciﬁcity means that they cannot be used to confer resistance towards all pests. However, another

g

group of naturally occurring proteins, the lectins, found in some plants, seems promising in the effort to expand the possibility of producing GM plants with resistance towards other pests such as aphids, and given the fact that they are proteins encoded by single genes, in principle they should be straightforward for use in genetic engineering. This review focuses on the promising results obtained using lectin genes for insect resistance, some toxicological studies with these GM plants, and the special issues that the use of lectin genes might raise. The focus will be on the health risk assessment that will have to take place before marketing, e.g. in the EU. The need for this is indicated by the fact that ﬁeld trials are already under way, e.g. in China with a GM rice containing the snowdrop GNA (Galanthus nivalis) lectin gene that has conferred resistance towards various hopper pests [1]. The review

CAB International 2010

166

Animal Science Reviews 2010

will not include a discussion of the advantages and disadvantages of GM plants in general.

Plant Lectins Lectins are a large heterogeneous group of proteins that bind carbohydrates reversibly and non-covalently without inducing any change in the carbohydrate. Lectins bind to a variety of cells having cell-surface glycoproteins or glycolipids and they typically agglutinate certain animal cells and/or precipitate glycoconjugates [2, 3]. Lectins can be classiﬁed in different ways. One way is to divide them into haemagglutinating and non-agglutinating lectins according to their ability to agglutinate red blood cells [4]. Haemagglutinating and especially non-agglutinating lectins may be present in almost every plant species [5]. In plants, lectins are found at highest concentration in seeds and other storage organs and at lower concentrations in the vegetative tissue. The levels in seeds normally range from 0.1 to 0.5% on a dry matter basis. The levels of lectins on a protein basis normally range from 0.1 to 5% in both seeds and vegetative tissue [4]. There may be different types of lectins, coded by different genes, expressed in different parts of the same plant. The biological role of lectins is not well understood. In plants, certain lectins may protect the plant against pests [6]. The protective effect is passive, as the plant tissue needs to be ingested before the lectins can be released and elicit their harmful effects on the pest [7]. The exact toxicological mechanism of lectins on pests is not yet elucidated but seems to involve binding to glycoproteins in the midgut and possibly passage into the haemolymph [8, 9]. Not only have they proved effective against speciﬁc insects, lectins can in addition have marked biological effects on intestinal structures and function as well as adverse systemic effects in mammals [4, 10, 11]. In a food context, the most well-known lectins come from the Phaseolus (bean) genus, with the kidney bean Phaseolus vulgaris lectin agglutinin (PHA lectin) having the highest known toxic potential in mammals [4].

Use of Insect Resistance Genes in GM Plants Production of food from plants often faces a considerable reduction in yield of the crop from weeds, insects, viruses, fungi, etc. Losses can be 50% or more [12] and various approaches are used to reduce this potential loss. Reducing losses from pests is today highly dependent on the use of chemicals (insecticides and fungicides) that are not always considered as the best solution for the environment. Using genetic engineering, some of these losses might be reduced by providing the plant with genes that give rise to additional inherent defence mechanisms. For this purpose, scientists and breeders are looking to

g

Nature to ﬁnd useful solutions. These include looking for suitable genes known to be involved in an organism’s defence mechanisms. Such genes are abundant in plants and could serve the purpose of reducing damage from insects in crop plants [13]. To reduce damage from insects, GM crop plants have been made to express an entomotoxic protein (Cry proteins) from the bacterium B. thuringiensis. Different variants of B. thuringiensis produce different Bt proteins, each with their speciﬁc insect target group, such as Coleoptera (Cry3), Lepidoptera (Cry1 and Cry2) and Diptera (Cry2 and Cry4). Although successfully used in GM plants, such as maize and cotton, for reducing damage from insects, they are not effective against all groups of insects, e.g. sap-sucking insects [14], and they might lose their effectiveness if resistance develops in the target insects. Plants using several different strategies for ﬁghting the pest are considered to be less vulnerable to resistance development in the pest and therefore there is a continuous need for breeding and searching for new resistance genes. Lectins are one group of natural proteins that have the potential to be used for insect resistance in plants (e.g. Powel et al. [15]). Although the mechanisms by which they act on insects are still not fully understood, their insecticidal effects have been clearly shown [13]. The combination in plants of insect resistance genes such as the cry and lectin genes seems to be an obvious goal for providing insect resistance in future crop plants. One widely studied lectin is the snowdrop lectin, Galanthus nivalis agglutinin (GNA). Early results show that GNA is active against sap-sucking insects such as planthoppers, leafhoppers and aphids [15–18] and might be very useful in developing GM crops with resistance towards insects that are not targeted by the Bt toxins [13]. Legaspi et al. [19] gives a comprehensive review of plant lectins expressed in transgenic crops with the focus primarily on the GNA lectin. The advantage of using GNA as an inserted trait is its speciﬁc binding to the gut epithelia of insects, eliciting toxicity, while at the same time it is considered non-toxic to mammals, because of its low binding capacity in the jejunum [20]. GNA has now been tested for its effects on various pests in plants such as wheat, rice, potato, maize and papaya [14, 18, 21–23]. The results from tests with these GM plants are promising for the development of new crop varieties with improved pest resistance, but there are to our knowledge no GM plants with inserted lectin genes on the market or approved by regulatory authorities.

Risk Assessment of GM Plants Basically, the international guidelines for safety assessment of GM plants proposed by the Organisation for Economic Co-operation and Development (OECD), the World Health Organization/Food and Agriculture Organization (WHO/FAO), the Codex Alimentarius Commission (Codex) and the European Food Safety Authority (EFSA)

CAB International 2010

Morten Poulsen and Jan W. Pedersen

[24–28] are almost identical in their approach and are based on the information on: the host; the donor and new product from inserted genes; comparison of the GM plant and a non-GM counterpart. The EU regulation on GM plants for food is based on regulation 1829/2003, which states in article 4.1 that: Food referred to in Article 3(1) must not (a) have adverse effects on human health, animal health or the environment; (b) mislead the consumer; (c) differ from the food which it is intended to replace to such an extent that its normal consumption would be nutritionally disadvantageous for the consumer. Since there are considerable differences among GM plants given their different hosts, different genes, different transformation events and different functions of the genes, safety assessment is based on a case-by-case evaluation focusing on the differences between the GM plant and the conventional counterpart. Often the GM plant will contain a new gene producing a new protein that has to be evaluated for its potential toxicity, anti-nutritional characteristics and allergenicity. The amount of data or studies needed for safety assessment of the newly expressed protein will depend on the extent of existing knowledge with respect to the source of the protein (or gene), function, activity and history of consumption. So far, none of the applications in the EU for approval of a GM plant have received a negative opinion from EFSA based on health risk. Some of these applications include GM plants with the well-known Bt toxins (Cry-1Ab, -1F, -3Bb1, -3A, -34Ab1, -35Ab1 and -2Ab2), which have been thoroughly tested in animals for their toxicity. Since there are, so far, no applications for the marketing of GM plants with inserted lectin genes, there are no examples of what information would be considered necessary before a risk assessment can be completed. However, given some similarities between lectins and Cry proteins (they are proteins and active against insects), it is expected that the risk assessment of lectins for insect resistance will be comparable to the assessment of the Cry proteins not only in relation to the environment, including determination of the target and non-target organisms, but also with respect to potential health risk. EFSA in their report from 2008 suggested, following a comparative approach using (for example) molecular and compositional analysis, to test the safety of the GM plants. If these analyses raise any doubts about the safety of the plant, it is recommended that an animal feeding study, e.g. a 90-day study in rodents, be performed to ensure safety [29]. Experiments with GM Plants Expressing Lectins The level at which the lectins will work effectively in reducing plant damage from the target insect will be crucial

g

167

for determining how low expression levels can be. Published data suggest that a level of about 1–2% of soluble protein would be expected to protect adequately [30]. Transgenic rice expressing the leaf lectin ASAL from Allium sativum (garlic) at about 0.7% of the total soluble protein in leaves signiﬁcantly reduced both the survival and fecundity of the brown planthopper and the green leafhopper [31]. Since much of the focus on GM plants with inserted lectin genes has been on the effects on pests living on leaves, there is some information on the expression level in leaves. However, little or no information exists on expression levels in the seeds that are used as food. The constitutive promoter CAMV35S used for expression of the lectin genes in several of the GM plants generated so far would be expected to allow expression in seeds also. Using tissue-speciﬁc promoters to restrict expression to the leaves or phloem could confer better resistance towards leaf-sucking insects [14]. At the same time this might reduce risk assessment requirements if these are not the plant parts used for food, e.g. the seeds or the roots, and the expression levels in these parts are relatively low. This will certainly be the case where the target tissue for expression is the leaves, and the seeds are used for food. However, it cannot be excluded that lectins might also be used for targeting insects feeding on seeds that are used for food and where the expression level would need to be relatively high to have an effect. A few GM plants expressing lectins have been tested for their safety in laboratory animal studies. These plants have been made for non-commercial purposes and there is so far no experience with commercial plants expressing lectins. Within the EU project SAFOTEST (‘New methods for the safety testing of transgenic food’), two rice varieties expressing GNA or P. vulgaris erythroagglutinin (PHA-E) lectin were tested [20, 32]. The purpose of the testing was not to perform an individual hazard assessment of each lectin-expressing GM rice variety but to test and improve the sensitivity and speciﬁcity of the OECD test guideline 408 (90-day oral toxicity study in rodents) when used for hazard assessment of GM plants [33]. The OECD test guideline 408 has routinely been used as the basis to ensure the safety of GM plants [34–36], but it has been questioned whether it has sufﬁcient sensitivity towards whole GM crops [37]. In the SAFOTEST approach, a 90-day study in rats was considered as the core safety study and the design of this was guided by data on the chemical and toxicological characteristics of the new gene product and information on changes in the lectinexpressing GM rice. This information was obtained by a comprehensive analytical characterization of both rice varieties [20, 32, 33]. In the 90-day rat study using 60% GNA rice in the feed, there was no difference in the number of adverse health effects between the group receiving the GM rice and the group receiving control rice. With the estimated average

CAB International 2010

168

Animal Science Reviews 2010

Table 1 Presentation of selected parameters from the 90days feeding study with PHA-E rice1

Parameter Plasma sodium Plasma protein Plasma albumin Plasma creatinine Plasma ALAT Plasma urea Mesenteric lymph nodes weight Small intestine weight Stomach weight Pancreas weight Small intestine length

GM rice with inherent PHA lectin (30 mg/kg body weight/day) cw. control

GM rice with inherent PHA lectin (30 mg/kg body weight/day) and spiked with PHA lectin (70 mg/kg body weight/day) cw. control

$ $ $ $ $ # $

# # # # " ## ""

""" """ " $

""" """ " """

1 The arrows "/# indicate statistically (P < 0.05) signiﬁcant differences from the control group. The arrows "" indicate statistically (P < 0.01) differences from the control group. The arrows """ indicate statistically (P < 0.001) differences from the control group. The arrows $ indicate no signiﬁcant differences from the control group.

of 1.25% GNA of the total soluble protein in the rice, the inclusion level of 60% GNA rice in the diet corresponded to a mean daily GNA-lectin intake of approximately 62 mg/kg body weight. The study was designed to detect both expected and unexpected effects, although it can be debated whether it was sufﬁciently sensitive to detect these effects and distinguish between them. In the 90-day study with PHA-E rice, a new animal test design was therefore used. In this setup, a third group given PHA-E-expressing rice was included that in addition received pure PHA-E lectin in an amount that is known from a preliminary in vivo study to elicit adverse effects in the rats. The mean daily intake of PHA-E lectin in the three groups was 0, 30 and 100 mg/kg body weight. A number of signiﬁcant effects were seen in the groups given GM rice compared with the group fed control rice. The differences that were either statistically signiﬁcant or showed a tendency towards an effect were more prominent in the group fed PHA-E rice+pure PHA-E lectin compared with the group fed PHA-E rice (Table 1). The results from this study using the new test design suggested that the majority of effects were as expected and were the result of the presence of the gene product and not caused by unexpected effects of the genetic modiﬁcation per se. This kind of design is not limited to GM plants containing lectins but can be used for risk assessment of other GM plants expressing gene products with a high bioactivity as well. In both studies the animal diets were carefully balanced with regard to micro- and

g

macronutrients to ensure that any observed differences were not caused by improper nutrition.

Consideration of How to Assess the Safety of GM Plants Containing Lectins GM plants with inserted genes coding for lectins will as a starting point be assessed as any other GM plants and the lectin protein should be assessed on its own for its potential toxicity and allergenicity as any other new protein. In the Codex and EFSA guidelines, as well as other international guidelines on genetically modiﬁed organisms (GMOs), it is stated that the toxic potential of a new protein, with no history of safe use, will focus on sequence similarity to known toxic proteins and information about stability to heat or degradation in gastric or intestinal model systems. In parallel, this will also be the information used for assessment of the potential for allergenicity where similarities to known allergens are analysed. However, as the lectins bind to carbohydrates and this may cause toxic or anti-nutritional effects in the gut, this will require special consideration. Information about the donor organism may in many cases give some indication of toxicity of the lectin. This will be the case where the donor organism is a known food plant or the donor is a non-food plant known to be toxic to the lectins. For the known food plant, the lectins can either be known as non-toxic to humans or as toxic to humans but with knowledge of how to avoid their toxic effects. The agglutinin lectins of garlic leaf (ASAL) and bulb (ASA) have a history of safe use and, together with information on the expression level in a GM plant compared with the level in the garlic, there will be a good basis for the risk assessment. Some lectins such as PHA from the kidney bean (P. vulgaris) are known for their toxicity to humans, while others such as GNA are considered non-toxic to humans. In the case of using the PHA lectin gene for resistance towards insects, the risk assessment should take into account the level of expression in the GM plant compared with the level found in beans currently on the market. Since the toxicity of the kidney bean lectins is normally removed by proper soaking and cooking of the beans, it cannot be excluded that the risk assessment would be based on the precondition that the GM plant should be handled in the same way to minimize the risk. However, this will probably not be an acceptable outcome, since the risk is that people may not handle the new food plant the right way to minimize the toxicity of the PHA lectin. As already indicated earlier, a more acceptable solution would be if the expression of the lectin gene is limited, e.g. by the use of a tissue-speciﬁc promoter, so the level of lectin in the edible part of the plant is below that which might cause concern for toxicity. In such a situation, it will be essential that the applicant for approval of the product can document the stability of the expression or more

CAB International 2010

Morten Poulsen and Jan W. Pedersen

precisely the lack of or low expression in the edible part of the plant. This should include assessment of its stability by further breeding of the GM plant where this will be a part of the scope for approval. It has been stated that, in principle, lectins from an edible species are more acceptable for use in transgenic crop plants than, for example, the GNA lectin from snowdrop [38]. However, for risk assessment, information on whether the lectin is from a food source or not should only have an inﬂuence in the sense that the history of use could contribute to the scientiﬁc risk assessment.

Conclusion Naturally occurring lectins expressed in GM plants seem promising in the effort to expand the group of target insects susceptible to insect-resistant plants. As they are proteins coded by single genes, they are a straightforward subject for genetic engineering. A few GM plants expressing lectins have been tested for their safety in laboratory animal studies. None of these studies showed any adverse effects on animals after feeding relatively large amounts of the GM plants. These GM plants were made for non-commercial purposes and there are no GM plants with inserted lectin genes on the market or approved by regulatory authorities at present. GM plants with inserted genes coding for lectins should be subjected to the same risk assessment procedures as other GM plants. The lectin protein should be assessed on its own for potential toxicity and allergenicity, as should any other new protein. It has been difﬁcult to show by the use of conventional animal models whether these are sensitive enough to detect unintended effects in the GM plants. Attempts have therefore been taken to increase the sensitivity and speciﬁcity by using spiking of the expressed proteins into the animals’ feed, which could be an option if the expressed protein has a visible biological activity. It can further be anticipated that the sensitivity and speciﬁcity of the animal studies could be further enhanced e.g. by the use of new ‘omics’ technologies. Although not many lectins have been thoroughly tested for toxicity, our evaluation is that most of the lectins that are potentially useful for insect resistance will pose no health risk in GM plants. This evaluation is based on the fact that they have either been tested and found safe as with the GNA lectin from snowdrop or because they have a history of safe use from other food sources. However, it is expected that in the future, GM plants will express a combination of different insect and other pest resistance genes of both bacterial and plant origin. Expression levels could, furthermore, exceed the amount that is known to be of safe use. This will be a challenge for the safety assessment procedure where the use of a more sensitive animal test approach may prove to be helpful.

g

169

References 1. Deka S, Barthakur S. Overview on current status of biotechnological interventions on yellow stem borer Scirpophaga incertulas (Lepidoptera: Crambidae) resistance in rice. Biotechnology Advances 2010;28:70–81. 2. Pusztai A, Ewen SWB, Grant G, Peumans WJ, van Damme EJM, Rubio L, et al. Relationship between survival and binding of plant lectins during small intestinal passage and their effectiveness as growth factors. Digestion 1990;46 Suppl. 2:308–16. 3. Grant G, Dorward PM, Pusztai A. Pancreatic enlargement is evident in rats fed diets containing raw soybeans (Glycine max) or cowpeas (Vigna unguiculata) for 800 days but not in those fed diets based on kidney beans (Phaseolus vulgaris) or lupinseed (Lupinus angustifolius). Journal of Nutrition 1993;123:2207–15. 4. van Damme EJM, Peumans WJ, Pusztai A, Bardocz S. Handbook of Plant Lectins: Properties and Biomedical Applications. John Willey and Sons, West Sussex, UK; 1998. 5. Collinge DB, Kragh KM, Mikkelsen JD, Nielsen KK, Rasmussen U, Vad K. Plant chitinases. Plant Journals 1993;3:31–40. 6. Gatehouse AMR, Powell KS, Peumans WJ, van Damme EJM, Gatehouse JA. Insecticidal properties of plant lectins: their potential in plant protection. In: Pusztai A, Bardocz S, editors. Lectin: Biomedical Perspectives. Taylor and Francis, London; 1995. p. 35–57. 7. Pusztai A. Can science give us the tools for recognizing possible health risks of GM food? Nutrition and Health 2002;16:73–84. 8. Fitches E, Woodhouse SD, Edwards JP, Gatehouse JA. In vitro and in vivo binding of snowdrop (Galanthus nivalis agglutinin; GNA) and jackbean (Canavalia ensiformis; Con A) lectins within tomato moth (Lacanobia oleracea) larvae; mechanisms of insecticidal action. Journal of Insect Physiology 2001;47:777–87. 9. Hogervorst PAM, Ferry N, Gatehouse AMR, Wa¨ckers FL, Romeis J. Direct effects of snowdrop lectin (GNA) on larvae of three aphid predators and fate of GNA after ingestion. Journal of Insect Physiology 2006;52:614–24. 10. Lafont J, Rouanet JM, Gabrion J, Assouad JL, Zambonino Infante JL, Besancon P. Duodenal toxicity of dietary Phaseolus vulgaris lectins in the rat: an integrative assay. Digestion 1988;41:83–93. 11. De Moya CC, Grant G, Fru¨hbeck G, Urdaneta E, Garci´a M, Marzo F, et al. Local (gut) and systemic metabolism of rats is altered by consumption of raw bean (Phaseolus vulgaris L. var. athropurpurea). British Journal of Nutrition 2003;89:311–8. 12. Oerke EC. Crop losses to pests. Journal of Agricultural Science 2006;144:31–43. 13. Chen MS. Inducible direct plant defense against insect herbivores: a review. Insect Science 2008;15:101–14. 14. Rao KV, Rathore KS, Hodges TK, Fu X, Stoger E, Sudhakar D, et al. Expression of snowdrop lectin (GNA) in transgenic rice plants confers resistance to rice brown planthopper. Plant Journal 1998;15:469–77. 15. Powell KS, Gatehouse AMR, Hilder VA, Gatehouse JA. Antimetabolic effects of plant-lectins and plant and fungal enzymes on the nymphal stages of 2 important rice pests,

CAB International 2010

170

Animal Science Reviews 2010 Nilaparvata lugens and Nephotettix cinciteps. Entomologia Experimentalis et Applicata 1993;66:119–26.

16. Hilder VA, Powell KS, Gatehouse AMR, Gatehouse JA, Gatehouse LN, Shi Y, et al. Expression of snowdrop lectin in transgenic tobacco plants results in added protection against aphids. Transgenic Research 1995;4:18–25. 17. Sauvion N, Rahbe Y, Peumans WJ, van Damme EJM, Gatehouse JA, Gatehouse AMR. Effects of GNA and other mannose binding lectins on development and fecundity of the peach-potato aphid Myzus persicae. Entomologia Experimentalis et Applicata 1996;79:285–93. 18. Down RE, Gatehouse AMR, Hamilton WDO, Gatehouse JA. Snowdrop lectin inhibits development and decreases fecundity of the glasshouse potato aphid (Aulacorthum solani) when administered in vitro and via transgenic plants both in laboratory and glasshouse trials. Journal of Insect Physiology 1996;42:1035–45. 19. Legaspi JC, Legaspi Jr BC, Setamou M. Insect-resistant transgenic crops expressing plant lectins. In: Koul O, Dhaliwal DS, editors. Transgenic Crop Protection: Concepts and Strategies. Science Publishers, Enﬁeld, NH, USA; 2004. p. 85–116. 20. Wang ZY, Zhang KW, Sun XF, Tang KX, Zhang JR. Enhancement of resistance to aphids by introducing the snowdrop lectin gene GNA into maize plants. Journal of Biosciences 2005;30:627–38. 21. Poulsen M, Kroghsbo S, Schrøder M, Wilcks A, Jacobsen H, Miller A, et al. A 90-day safety in Wistar rats fed genetically modiﬁed rice expressing snowdrop lectin Galanthus nivalis (GNA). Food and Chemical Toxicology 2007;45:350–63. 22. Stoger E, Williams S, Christou P, Down RE, Gatehouse JA. Expression of the insecticidal lectin from snowdrop (Galanthus nivalis Agglutinin; GNA) in transgenic wheat plants: effects on predation by the grain aphid Sitobion avenae. Molecular Breeding 1999;5:65–73. 23. McCafferty HRK, Moore PH, Zhu YJ. Papaya transformed with the Galanthus nivalis GNA gene produces a biologically active lectin with spider mite control activity. Plant Science 2008;175:85–393.

27. Codex Alimentarius. Codex principles and guidelines on foods derived from biotechnology. Codex Alimentarius Commission, Joint FAO/WHO Food Standards Programme, Food and Agriculture Organisation, Rome; 2003. 28. EFSA. Guidance document of the scientiﬁc panel of genetically modiﬁed organisms for the risk assessment of genetically modiﬁed plants and derived food and feed. European Food Safety Authority, May 2006. ISBN: 92-9199-019-1. 29. EFSA. Safety and nutritional assessment of genetically modiﬁed plants and derived food and feed: the role of animal feeding trials. Food and Chemical Toxicology 2008;46:S2–70. 30. Boulter D. Insect pest-control by copying nature using genetically-engineered crops. Phytochemistry 1993;34:1453–66. 31. Saha P, Majumder P, Dutta I, Ray T, Roy SC, Das S. Transgenic rice expressing Allium sativum leaf lectin with enhanced resistance against sap-sucking insect pests. Planta 2006;223:1329–43. 32. Poulsen M, Schrøder M, Wilcks A, Kroghsbo S, Lindecrona RH, Miller A, et al. Safety testing of GM-rice expressing PHA-E lectin using a new animal test design. Food and Chemical Toxicology 2007;45:364–77. 33. Knudsen I, Poulsen M. Comparative safety testing of genetically modiﬁed foods in a 90-day rat feeding study design allowing the distinction between primary and secondary effects of the new genetic event. Regulatory Toxicology and Pharmacology 2007;49:53–62. 34. Malley LA, Everds NE, Reynolds J, Mann PC, Lamb I, Rood T, et al. Subchronic feeding study of DAS-59122-7 maize grain in Sprague-Dawley rats. Food and Chemical Toxicology 2007;45:1277–92. 35. Hammond B, Lemen J, Dudek R, Ward D, Jiang C, Nemeth M, et al. Results of a 90-day safety assurance study with rats fed grain from corn rootworm-protected corn. Food and Chemical Toxicology 2006;44:147–60.

24. OECD. Safety evaluation of foods derived by modern biotechnology: concept and principles. 1993. http:// www.oecd.org/home

36. Hammond B, Dudek R, Lemen JK, Nemeth M. Results of a 90-day safety assurance study with rats fed grain from corn borer-protected corn. Food and Chemical Toxicology 2006;44:1092–9.

25. WHO/FAO. Safety aspects of genetically modiﬁed foods of plant origin. Report of a joint FAO/WHO expert consultation on foods derived from biotechnology. 2000. Available from: URL: http://www.who.int/foodsafety/publications/biotech/en/ ec_june2000_en.pdf

37. Ko¨nig A, Cockburn A, Crevel RWR, Debruyne E, Grafstroem R, Hammerling U, et al. Assessment of the safety of foods derived from genetically modiﬁed (GM) crops. Food and Chemical Toxicology 2004;42:1047–88.

26. WHO/FAO. FAO/WHO expert consultation on foods derived from biotechnology. Evaluation of allergenicity of genetically modiﬁed foods. WHO/FAO, Rome; 2001. Available from: URL: http://www.who.int/foodsafety/ publications/biotech/en/ec_jan2001.pdf

38. Sadeghi A, Smagghe G, Broeders S, Hernalsteens JP, De Greve H, Peumans WJ, et al. Ectopically expressed leaf and bulb lectins from garlic (Allium sativum L.) protect transgenic tobacco plants against cotton leafworm (Spodoptera littoralis). Transgenic Research 2008;17:9–18.

g

CAB International 2010

Animal Science Reviews 2010

Ovine and caprine brucellosis (Brucella melitensis) Assadullah Samadi1*, M.M.K. Ababneh1, N.D. Giadinis2 and S.Q. Laﬁ1 Address: 1 Faculty of Veterinary Medicine, Jordan University of Science and Technology, PO Box 3030, Irbid 22110, Jordan. 2 Clinic of Farm Animals, School of Veterinary Medicine, Aristotle University, 546 27 Thessaloniki, Greece. *Correspondence: Assadullah Samadi. Email: [email protected] Received: Accepted:

7 April 2010 24 May 2010

Abstract Brucellosis, caused by Brucella melitensis, remains one of the most common zoonotic diseases worldwide with more than 500 000 human cases reported annually. The bacterial pathogen is classiﬁed, by the US Centers for Disease Control (CDC), as a category (B) pathogen and belongs to the World Organisation for Animal Health (OIE) list B diseases. Brucella melitensis, the ﬁrst species in the genus Brucella to be described, causes abortions in female goats and sheep, unilateral orchitis in males and Malta fever in humans. The natural hosts may be goats and sheep, but the organism is the least species-speciﬁc of the brucellae. B. melitensis biovars 1 and 3 are the most frequently isolated in Mediterranean countries. In most circumstances, the primary excretion sources are foetal ﬂuids, vaginal discharges after abortion or full-term parturition, in milk and semen. The epidemiology of human brucellosis has drastically changed over the past decade because of various sanitary, socio-economic and political reasons, together with the evolution of international travel. Besides the public health concern of such an important zoonosis, Brucella infections in animals have an important economic impact especially in developing countries as they cause abortion in pregnant animals, reduce milk production and cause fertility problems. The most reliable and the only unequivocal method for diagnosing animal brucellosis is the isolation of Brucella spp., but polymerase chain reaction (PCR) has the potential to meet the need for better diagnostic tools. It is highly sensitive, very speciﬁc, inexpensive and easily adapted to high-volume demands. The process is rapid, simple and requires little manual labour. A sensitive and dependable diagnostic PCR assay for B. melitensis is crucial for controlling the spread of Brucella in animal and human population. Immunological testing for brucellosis among livestock is usually conducted as a component of the disease eradication and surveillance programme rather than as diagnostic support and each country has a different policy for testing livestock. The serological tests commonly used for the diagnosis of B. melitensis infection are the Rose Bengal Test (RBT), Serum Agglutination Test (SAT), Complement Fixation Test (CFT) and enzyme-linked immunosorbent assays (ELISA). The RBT and CFT are the most widely used tests for the serological diagnosis of sheep and goat brucellosis; they are also the ofﬁcial tests for international trade. B. melitensis infection is particularly problematic, because Brucella abortus vaccines do not protect effectively against B. melitensis infection. In regions with a high prevalence of the disease, the only way of controlling and reducing the prevalence of zoonosis is by the vaccination of all susceptible hosts and elimination of infected animals. B. melitensis strain Rev-1, although highly infectious to humans, is considered as the only vaccine available for the control of ovine and caprine brucellosis, especially when administrated at the standard dose by the conjunctival route. However, the Rev-1 vaccine shows a considerable degree of virulence and induces abortions when administered during pregnancy. Also, the antibody response to vaccination cannot be differentiated from the one observed after ﬁeld infection, which impedes control programmes. Attempts have been made to develop new live attenuated rough B. melitensis vaccines, which are devoid of the O-side chain. Those vaccines await further evaluation in ﬁeld experiments. The aetiology, isolation, epidemiology, global distributions and

g

CAB International 2010

172

Animal Science Reviews 2010

new methods of diagnosis and control programmes for ovine and caprine brucellosis have been reviewed in this article.

Introduction Brucellosis, especially that caused by Brucella melitensis, remains one of the most common zoonotic diseases worldwide with more than 500 000 human cases reported annually. The bacterial pathogen is classiﬁed by the Centers for Disease Control (CDC) as a category (B) pathogen and belongs to OIE list B diseases [1]. Brucellosis is present throughout the ﬁve continents and it is still an uncontrolled serious public health problem in many developing countries, although information on its distribution is rather sparse in many parts of the world, probably because of the fact that it remains insidious, while many countries with limited resources face other priority diseases that are more spectacular, such as foot-andmouth disease, sheep pox, Rift Valley fever, peste des petits ruminants, etc. [2]. B. melitensis accounts for most recorded cases globally with cattle emerging as an important reservoir. Isolated cases of non-terrestrial brucellosis and continuing transmission from wild animals have raised important epidemiological issues [3]. Animal brucellosis poses a barrier to the trade of animals and animal products and could seriously impair socio-economic development, especially for livestock owners, which represent a vulnerable sector in many rural populations. As an indication of the importance of this disease, it sufﬁces to mention that plans to eliminate ovine, caprine and bovine brucellosis from the European Union (EU) were expected to receive over half of the total European Commission funding for animal disease control measures in 1997 [2]. Brucellosis is an important infectious disease of sheep and goats and is considered to be the most serious zoonotic disease for humans [4]. In sheep and goats, brucellosis is mainly caused by B. melitensis, a Gram-negative coccobacillus or short rod. This organism is a facultative intracellular pathogen. B. melitensis contains three biovars (biovars 1, 2 and 3). All three biovars cause disease in small ruminants, but their geographic distribution varies [5]. Brucellosis is endemic in sheep and goats in most countries of the Mediterranean basin, the Middle East and Central Asia [6, 7]. The distribution of B. melitensis has long been associated with the Mediterranean littoral; however, it is now known to be much more widely distributed with only North America, North Europe, SouthEast Asia and Oceania being spared [8]. The majority of abortions in sheep and goats occurred in the last month of gestation. It can be of considerable economic importance to the sheep and goat industry [9]. Control measures are based on strict hygiene and the vaccination of susceptible animals such as sheep and goats. Vaccination is regarded as a measure for reducing the

g

prevalence of the disease eventually to a level where eradication by test and slaughter can be considered. Of the vaccines now used for immunizing small ruminants against B. melitensis, Rev-1 vaccine is generally preferred [10, 11]. The Rev-1 vaccine is indicated to protect small ruminants against brucellosis and to protect females from abortion in regions where the disease occurs. Conjunctival vaccination is safer than subcutaneous vaccination, but is not safe enough to be applied regardless of the pregnancy status of animals [12] and the duration of immunity conferred by this method of vaccination is a subject of controversy. The public health and economic impact of brucellosis remains a particular concern in developing countries throughout Africa, West Asia and some parts of Latin America, given that the danger of infected animals constitute the transmission of severe zoonosis to humans as well as to the economic losses associated with the disease in animals, and the serious constraints imposed to the improvement of animal husbandry and genetic resources in the affected areas [2]. The disease still occurs in several EU countries. While common efforts to eradicate B. melitensis infection are undertaken in ﬁve EU countries, namely France, Spain, Portugal, Italy and Greece; this is far from being achieved or even approached except in France [2, 13]. The geographical distribution of human brucellosis is constantly changing with new foci emerging or reemerging [1]. Human brucellosis caused by B. melitensis, the most pathogenic species in humans, constitutes a public health priority. Although a notiﬁable disease in many countries, the disease remains underestimated by the medical authorities, as ofﬁcial ﬁgures do not reﬂect the number of human infections that occur each year and which may be as high as 10 or 25 times the reported ones. Given the lack of awareness and of adequate laboratory support to detect the infection, it is often considered as ‘fever of unknown origin’ [1]. Although a number of successful vaccines are being used for immunization of animals, no satisfactory vaccine against human brucellosis is available [1]. Since the treatment of animal brucellosis is very expensive, one should encourage the mass vaccination of livestock [3].

Aetiology B. melitensis, the ﬁrst species in the genus Brucella to be described, causes abortions in female goats and sheep, unilateral orchitis in males and Malta fever in humans [14]. This micro-organism is a facultative intracellular Gramnegative cocco-bacilli that may appear in pairs, short chains, or groups, non-spore-forming and non-capsulated,

CAB International 2010

Assadullah Samadi, M.M.K. Ababneh, N.D. Giadinis and S.Q. Laﬁ

grow in nutrient-rich media within 48–72 h of incubation at 37 C in a 5% CO2 atmosphere, aerobic, catalase- and oxidase-positive [15]. They do not ferment carbohydrates and have variable urease activity [16, 17] and belong to the alpha-2 subdivision of the Proteobacteria, along with Ochrobactrum, Rhizobium, Rhodobacter, Agrobacterium, Bartonella and Rickettsia [18]. B. melitensis has three biovars (1–3), is highly pathogenic for humans, grows in the presence of fuchsine and thionine, does not produce H2S, has urease activity, does not need CO2 for its growth and is negative on Tiblisi phage lysis [17]. From an investigation conducted on the genomes of 51 strains of B. melitensis, all the strains appeared to be very similar, so that it seems justiﬁable to consider them as a single species, which, for priority reasons, could be called B. melitensis [13]. 10 genome sequences representing ﬁve species of Brucella (B. melitensis, Brucella suis, Brucella abortus, Brucella ovis and Brucella canis) are available and about 25 additional Brucella strains/species are being sequenced. The genomes of the members of Brucella are very similar in size and gene make-up [19]. Each species within the genus has an average genome size of approximately 3.29 Mb and consists of two circular chromosomes. Chromosome I is approximately 2.11 MB on an average and Chromosome II is approximately 1.18 Mb. The G+C content of all Brucella genomes is 57.2% for Chromosome I and 57.3% for Chromosome II [20–22]. Based on a comparison of 10 published Brucella genomes, what is striking are the shared anomalous regions found in both chromosomes, consistent with horizontal gene transfer in spite of a predominantly intracellular lifestyle [23]. The completion of the genome sequences has led to studies focused on evaluating the functional annotation of products from predicted coding sequences, giving researchers an insight into genes and gene functions. It has also created an opportunity to study the products produced by novel genes and to evaluate their implications for survival, replication and virulence. Proteomic studies to deﬁne the biochemical pathways associated with stress responses, host speciﬁcity, pathogenicity, virulence and vaccine development have also expanded. The comparison of genomes and proteomes among the species has shed signiﬁcant light on the Brucella genus [21, 24–26]. The Brucellae have no classic virulence genes encoding capsules, plasmids, pili or exotoxins and, compared with other bacterial pathogens, relatively little is known about the factors contributing to the persistence in the host and multiplication within phagocytic cells. Also, many aspects of interaction between Brucella and its host remain unclear [27].

Antigenic composition A substantial number of antigenic components of Brucella have been characterized. However, the antigen that

g

173

dominates the antibody response is the lipopolysaccharide (LPS). B. melitensis, which is one of the smooth phase strains (S), has the S-LPS, which is comprised of a lipid A (containing two types of aminoglycose), distinctive fatty acids (excluding b-hydroxymyristic acid), a core region containing glucose, mannose and quinovosamine and an O-chain comprising a homopolymer of approximately 100 residues of 4-formamido-4,6-dideoxymannose (linked predominantly a-1,2 in A-epitope-dominant strains with every ﬁfth residue linked a-1,3 in M-dominant strains) [28]. The difference in linkage inﬂuences the shape of the LPS epitopes. The A-dominant type is rod-shaped and is determined by ﬁve consecutive a-1, 2 linked residues, whereas the M-dominant type is kinked and determined by four residues, including one linked a-1,3 [29]. Antigen A predominates in B. abortus and B. suis, while antigen M is the major antigen in B. melitensis. Strains that react with antisera to both A and M epitopes produce LPS of both types in approximately equal proportions [30], consistent with the original hypothesis of Wilson and Miles [31]. The presence of 4-amino, 4, 6 dideoxymannose in the LPS is also responsible for the antigenic cross-reactivity with Escherichia hermanni, Escherichia coli O : 157, Salmonella O : 30, Stenotrophomonas maltophilia, Vibrio cholerae O : 1 and Yersinia enterocolitica O : 9 LPS [28]. Numerous outer and inner membrane, cytoplasmic and periplasmic proteins have also been characterized [3]. Some Brucella strains (B. abortus 2308, B. abortus S19, B. abortus 02 and B. melitensis 03) tested, showed a haemagglutination activity on red blood cells (RBC) from the various sources, with B. melitensis 03 showing the highest haemagglutination titres against all RBC tested and B. abortus 2308 showing the lowest titre [32].

Global Distributions Brucellosis is still present throughout the ﬁve continents (although its presence is anecdotal in Oceania), but information on its distribution is rather sparse in many parts of the world, probably because of the disease, which belongs to OIE list B diseases, remains insidious, while many countries with limited resources face other priority diseases that are more spectacular, such as foot-and-mouth disease, sheep pox, Rift Valley fever and peste des petits ruminants [2, 33]. Up to 2003, the disease caused by B. melitensis has been reported in Angola, Burkina Faso, Cape Verde, Djibouti, Eritrea, Ethiopia and Kenya in South-Saharan Africa. The last outbreaks of B. melitensis infection reported were in 1993 in Guamand in 2000 in French Polynesia. In South Asia, the disease is endemic in India, Pakistan and Bangladesh. Some countries in South-East Asia (Thailand, Malaysia and Myanmar) reported the disease sporadically. Mongolia is heavily contaminated and countries of the former Asian USSR are endemically infected. Latin

CAB International 2010

174

Animal Science Reviews 2010

America, Mexico, Peru and Argentina are the three countries, where the disease is regularly reported, although it is also present in Bolivia, Brazil, Columbia, Costa Rica and the Dominican Republic. In the countries of the EU, the following Member States and regions have been recognized as being free from the infection: Belgium, Denmark, Finland, Germany, Irish Republic, Luxembourg, The Netherlands, Sweden, the UK, 17 departments of France and two provinces of Spain. The disease still occurs in several EU countries. While common efforts to eradicate B. melitensis infection are undertaken in ﬁve EU countries, namely France, Spain, Portugal, Italy and Greece, this is far from being achieved or even approached, except in France. Cyprus (a new EU Member country) was given as an example of success for the eradication of B. melitensis until 1996, when the disease reappeared and has been reported annually since then. In addition to the EU countries, brucellosis is prevalent in Albania, Andorra, Armenia, Azerbaijan, Bosnia and Herzegovina and Serbia and Montenegro. The disease is endemic in almost all African and Asian countries [2, 13]. However, North America (except Mexico) is believed to be free of the disease, as are South-East Asia, Australia and New Zealand [8]. Of the three different biovars, biovar 3 predominates almost exclusively in the Mediterranean countries and the Middle East (B. melitensis biovar 3 is the most commonly isolated species from animals in Egypt, Jordan, Israel, Tunisia and Turkey. B. melitensis biovar 2 was reported in Turkey and Saudi Arabia, and B. melitensis biovar 1 in Libya, Oman and Israel [34], while biovar 1 seems to predominate in Latin America. Biovars 1 and 2 have also been reported in some southern European countries. However, the precise recognition of biovar 3, especially its differentiation from biovar 2 sometimes appears equivocal [33].

Transmission Brucellosis is usually introduced into a herd (by or with infected animals [35, 36]). Brucellosis of small ruminants affects mainly sexually mature individuals with abortion; most importantly in the later stages of pregnancy while sexually immature animals are resistant. The receptivity of ewes to B. melitensis varies according to the breed. Maltese sheep are resistant, while the Awassi breed of the Middle East is quite susceptible. The spread of an infection from country to country or within the same country generally follows the transfer of infected animals. After the Second World War, a vast movement of sheep and goats took place in Europe, contributing to the spread of the infection. Brucellosis is also transmitted from farm to farm through wild animals and dogs, responsible for carrying around aborted foetuses. Mixing herds at pasture and keeping the animals in shelter during the night, particularly if in such areas parturition takes place, represent

g

major factors for transmission of the infection. Cattle can also be infected from sheep. Dogs and rodents in contact with infected animals may acquire infection, but this mode of transmission is of little importance from an epidemiological point of view. In the transmission cycle, insects and ticks may also be involved [13]. The phenomenon of latency, so common in cattle, has also been conﬁrmed in B. melitensis-infected sheep [37], even if it seems that the latently infected ewes rarely transmit the infection to their lambs. However, in spite of the low frequency of transmission, the existence of such latent infections greatly increases the difﬁculty of eradicating brucellosis. Material from an abortion represents the main source of transmission, with the excretion of enormous numbers of bacteria; the placenta, foetuses and foetal ﬂuids are highly infective. After delivery or abortion, the excretion of brucellae in vaginal discharge continues for about 3 weeks but may last up to 2 months. Therefore, the soil where parturitions take place becomes massively contaminated. The number of brucellae excreted in milk is generally not relevant for sheep-to-sheep transmission, but is important for transmission of the infection to humans. The nature of the material that is contaminated by the brucellae is of some importance. Sand and straw used for bedding may absorb a considerable number of bacteria. Impervious material, such as concrete, keep bacteria on the surface and animals may therefore become infected through inhaling the contaminating micro-organisms. It is a general belief that the male does not play an important role in the epidemiology of brucellosis. It is possible, however, that it may transmit the infection through mechanical means. In males, the infection may affect the reproductive organs and, quite often, orchitis develops. The resistance of Brucella in the environment is not easily determined, because the conditions in which the bacteria may be found are very variable. The organism can survive in dust from 3 to 44 days, on sterile surfaces for 20 days and in tap water for 30 days. Resistance in wooden houses and on the ﬂoor of shelters is about 4 months. In pastures exposed to the sun, survival is up to 15 days, while in the shade it is 35 days. B. melitensis is killed by pasteurization, and it is sensitive to common disinfectants [13]. Sheep and goat excreta as well as their milk products are the main source of infection, but B. melitensis in cattle has emerged as an important problem in some southern European countries, Israel, Kuwait and Saudi Arabia [38]. Also, B. abortus vaccines do not protect effectively against B. melitensis infection. Thus, bovine B. melitensis infection is emerging as an increasingly serious public health problem in some countries with the spread of the disease through unpasteurized dairy products received from these infected cows have not been reported in hinds [36]. In nonpregnant ewes, B. melitensis is not excreted from the vagina. However, in pregnant animals, the excretion starts at the time of parturition or abortion and may last for months. Following infection with B. melitensis, the mammary glands are often colonized, and infection of the

CAB International 2010

Assadullah Samadi, M.M.K. Ababneh, N.D. Giadinis and S.Q. Laﬁ

udder interferes with the production of milk, thus reducing or arresting milk output. Brucellae are not always excreted during lactation, although it is believed that infection of the udder is the means by which infection persists to the following pregnancies. Excretion of the bacteria may last for as long as 180 days after parturition or abortion [13]. Infection in lactating non-pregnant goats is likely to lead to colonization of the udder with excretion of B. melitensis in the milk [39]. In goats, about two-third of acute infections acquired naturally during pregnancy lead to infection of the udder and excretion of bacteria in milk during subsequent lactation [40]. In some goats, excretion may cease during this lactation, but in many it persists and often continues during the next [41]. Greatly reduced milk yield follows abortion, and infection of the udder following a normal birth also leads to a considerable reduction in yield. In spite of this, clinical signs of mastitis are seldom detectable in naturally infected goats [14]. Sheep that abort often excrete bacteria in milk, but generally for not more than 2 months [14]. However, exceptionally, excretion may continue for 140 days and even 180 days [42]. In goats, excretion of the organisms from the vagina is prolonged and copious (2–3 months generally). In sheep, excretion is generally less prolonged, usually ceasing within 3 weeks after abortion or full-term parturition [43]. Dissemination of the Rev-1 vaccine strain from vaccinated to non-vaccinated and healthy animals also plays a role in transmission of the infection. Recently, Kojouri and Gholami [44] reported that bacteraemia can be prolonged: >60 days after vaccination with Rev-1 vaccine, and can cause dissemination from vaccinated animals to healthy ones. It is also reported that, using standard subcutaneous doses (110972109) of the vaccine, Rev-1 yields a 2- to 3-month widespread and persistent infection of the animal by the organism, actively colonizing the spleen and several lymph nodes, which can transmit the disease to non-vaccinated animals [45].

Clinical Signs The main route of entry of the bacteria is the nasopharynx. However, the cutaneous route must not be excluded. The bacterium spreads via the lymphatics and is arrested in the lymph nodes. In animals resistant to infection, the brucellae are killed by macrophages, the active cells of the immune system, with the intervention of the antibodies and lymphocytes. There is no evidence that the clinical features of B. melitensis infection in sheep and goats vary according to the biovar involved [46]. In susceptible animals, on the other hand, the bacterium survives phagocytosis and replicates inside cells. Following lysis of the phagocytic cell, the brucellae are liberated and infect other cells. Bacteraemia may eventually develop. In fully susceptible pregnant and non-pregnant animals, B. melitensis cells are present in blood for 30–45 days

g

175

after infection. In virgin females from endemically infected areas, bacteraemia is rare and seen only in a small number of animals. In pregnant animals, the bacterium enters the uterus where it reproduces in the placenta and foetal tissues, inducing an infective state not necessarily followed by abortion. The percentage of aborting animals varies according to circumstances. In non-pregnant animals, Brucella can cause a chronic infection, which is of epidemiological importance, because after an initial serological reaction in the animal, the infection becomes non-apparent, thus creating problems in diagnosis [13]. B. melitensis mainly causes abortions, stillbirths and the birth of weak offspring. Retained placentas can be observed. Sheep and goats usually abort only once, but reinvasion of the uterus and shedding of organisms can occur during subsequent pregnancies [47]. However, clinical signs of mastitis are uncommon. Acute orchitis and epididymitis can occur in males, and may result in infertility. Arthritis is seen occasionally in both sexes. Many non-pregnant sheep and goats remain asymptomatic [48].

Morbidity and Mortality B. melitensis is associated with a high morbidity rate in naı¨ve herds, and a much lower morbidity rate in chronically infected herds. In naive ruminant herds, B. melitensis spreads rapidly, and 30–80% of the herd may abort. In herds where this organism becomes endemic, only sporadic abortions can occur in animals at their ﬁrst pregnancies. Fertility can be permanently impaired after infection with some species of Brucella. Deaths are rare in adult animals of most species [35].

Post-Mortem Lesions Some aborted foetuses appear normal; others are autolysed or have variable amounts of subcutaneous oedema and blood-stained ﬂuid in the body cavities. In small ruminant foetuses, the spleen and/or liver may be enlarged, and the lungs may exhibit pneumonia and ﬁbrous pleuritis. Abortions caused by B. melitensis are typically accompanied by placentitis. The cotyledons may be red, yellow, normal or necrotic. In small ruminants, the intercotyledonary region is typically leathery, with a wet appearance and focal thickening. There may be exudate on the surface [35, 36]. In adults, granulomatous to purulent lesions may be found in the male and female reproductive tract, mammary gland, supramammary lymph nodes, other lymphoid tissues, bones, joints and other tissues and organs. Mild to severe endometritis may be seen after an abortion, and males can have unilateral or bilateral epididymitis and/or orchitis [35].

CAB International 2010

176

Animal Science Reviews 2010

Diagnosis Direct (Bacteriological) Diagnosis The most reliable and the only unequivocal method for diagnosing animal brucellosis is the isolation of Brucella spp. [43]. The bacteriological diagnosis of B. melitensis can be made by means of the microscopic examination of stained smears from vaginal swabs, placentas or aborted foetuses (Stamp’s method). However, the isolation of B. melitensis on appropriate culture media is recommended for accurate diagnosis. Vaginal excretion of B. melitensis is usually copious and persists for several weeks after abortion [14], as does udder infection [49]. Thus, taking vaginal swabs and milk samples is the best way to isolate B. melitensis from sheep and goats. The spleen and lymph nodes (iliac, supramammary and prefemoral) are the best sites for obtaining samples for isolation during post-mortem examination [49]. B. melitensis does not require serum or CO2 for growth and can be isolated on ordinary solid media under aerobic conditions at 37 C. Nevertheless, because of the overgrowing contaminants usually present in ﬁeld samples, selective media are needed for isolation purposes. The Farrell selective medium is also recommended for B. melitensis [43, 50]. However, nalidixic acid and bacitracin, at the concentration used in that medium, have inhibitory effects for some B. melitensis strains [51] and the isolation rate increases signiﬁcantly by the simultaneous use of both the Farrell and the modiﬁed Thayer–Martin media [49, 51]. Blood samples and tissue homogenates are added to agar-serum glucosate. If low numbers of bacteria are foreseen, or if antibiotics have been added, it is advisable to enrich the culture with agar-blood or agar Border–Gengou. When Rev-1 vaccine is used, it is necessary to distinguish the vaccinal strain from the virulent B. melitensis. Rev-1 has a low virulence for the guinea pig, which, if inoculated subcutaneously with a 103 dose of Brucella organisms, gives negative cultures from the spleen 3–5 months after inoculation, whereas virulent B. melitensis induces an infection lasting 6–12 months [13]. While culturing is a speciﬁc method, its sensitivity depends on the viability of Brucella in the sample, the kind of sample and the number of specimens tested from the same animal [51].

Polymerase chain reaction (PCR) As with any disease, the control of brucellosis would beneﬁt from improvements in diagnostic methods. PCR has the potential to meet the need for better diagnostic tools. It is highly sensitive, very speciﬁc, inexpensive and easily adapted to high-volume demands. The process is rapid, simple and requires little manual labour. As long as careful attention is given to avoid contamination, the method is very reliable and usually highly reproducible at any

g

properly equipped laboratory. Since PCR was ﬁrst introduced in 1987, researchers have made excellent progress in developing quality PCR-based tests for Brucella. A sensitive and dependable diagnostic PCR assay for B. melitensis is crucial for controlling the spread of Brucella in animal and human populations [52]. The time required for culturing can be long and samples that are only contaminated with a low number of Brucella may not be detected. PCR assay has been shown to be a valuable method for detecting DNA from different microorganisms and provides a promising option for diagnosis of brucellosis [53–57]. Despite the high degree of DNA homology within the genus Brucella, several molecular methods, including PCR, PCR restriction fragment length polymorphism (RFLP) and Southern blot, have been developed that allow, to a certain extent, the differentiation between Brucella species and some of their biovars. Pulseﬁeld gel electrophoresis has been developed that allows the differentiation of several Brucella species [58]. Several authors reported good sensitivity with PCR, based on different molecular markers (16S rRNA, bscp31, IS 6501/ 711) [59] for detecting Brucella DNA with pure cultures [60–63]. Speciﬁc molecular markers have speciﬁcally been developed for distinguishing the Rev-1 strain from B. melitensis wild strains [64]. Recently, a new method has been described for ﬁngerprinting Brucella isolates based on the multi-locus characterization of a variable number, 8-bp, and tandem repeat. The technique is highly discriminatory among Brucella species or strains [54]. Recently, Gupta et al. [52] reported a reliable, highly sensitive and speciﬁc single-step PCR test for the detection of B. melitensis in the tissue and blood obtained from infected goats. The report also describes a simpliﬁed method for extracting Brucella DNA from tissue and blood samples and described the evaluation of a single-step PCR for the detection of the omp31 gene sequence of B. melitensis in tissues and blood of naturally infected goats. In addition, a PCR assay identifying B. melitensis in the semen of infected sheep has been developed [65]. A new method of PCR has been developed: single-tube, multi-locus variable number tandem repeat analysis (MLVA) assay for simultaneous speciation and strain typing of Brucella, the aetiologic agent of brucellosis. This MLVA assay consists of eight loci, two of which are species-speciﬁc markers that allow deﬁnitive identiﬁcation of B. melitensis, and other Brucella species, while ruling out related pathogenic bacterial genera. This new PCR method can differentiate all strains and biotypes of B. melitensis [66]. Recently, a robust and rapid multiplex PCR assay has been introduced which allows the differentiation of all nine currently recognized Brucella species including the recently described Brucella species, Brucella microti, Brucell inopinata, Brucella ceti and Brucella pinnipedialis [67].

CAB International 2010

Assadullah Samadi, M.M.K. Ababneh, N.D. Giadinis and S.Q. Laﬁ

Indirect (serological and allergic) diagnosis Immunological testing for brucellosis among livestock is usually conducted as a component of the disease eradication and surveillance programme rather than as diagnostic support and each country has a different policy for testing livestock [48]. Generally, the indirect diagnosis of disease is recommended for large-scale surveillance and/or eradication purposes. The detection of antibodies (and at a lesser degree the measure of the cell-mediated immunity) against relevant Brucella epitopes is a more practical approach. However, precise antigens and adequate tests have to be used for a proper efﬁcacy and reliability. Particularly relevant is the problem of the speciﬁcity of serological tests, since antibodies against Brucella epitopes may be present in the animal population because of vaccination and/or of contacts with other Gram-negative bacteria (mainly Yersinia enterocolitica O : 9) sharing cross-reactive epitopes with Brucella. Although the Rev-1 vaccine is an essential tool to control small ruminants brucellosis, when applied under standard conditions (i.e. full dose via the subcutaneous route), it induces long-lasting serological responses that interfere with subsequent serological screening [14, 68]. No single serological test is appropriate in all epidemiological situations; all have limitations especially when it comes to screening individual animals [69]. The RBT and CFT are the most widely used tests for the serological diagnosis of sheep and goats brucellosis [68]; they are also the ofﬁcial tests for international trade [70].

177

nor induces an increase in the titre of antibodies, and thus does not interfere with allergic or serological tests. In sheep, 0.5 ml of INRA allergen is inoculated intradermally in the lower eyelid. Reading is done after 48 h from the front of the animal, in order to have the untreated eye as a control. The method is suggested for screening a large number of animals, thus eliminating the time-consuming collection of blood samples. An increment in capillary permeability of the product is obtained by adding hyaluronidase to the allergen. This causes a quicker and more intense reaction, allowing a reduction in the allergic response in those sheep with a doubtful positive reaction [13, 73]. Rose Bengal Test (RBT) This test is internationally acknowledged as the choice for the screening of brucellosis in small ruminants [50]. However, standardized conditions suitable for diagnosing cattle infection [50, 68] are not adequate in sheep and goats [74, 75] and account for the low sensitivity of RBT antigens in small ruminants along with the fact that a high proportion of animals in infected areas give negative results in RBT, but positive in CFT [74], which brings into question the efﬁcacy of the present RBT as an individual test. A simple modiﬁcation of increasing slightly the amount of sera for the test dose from 25–30_l to 75–90_l, at the same time maintaining the antigen volume (25–30 _l), signiﬁcantly increases the sensitivity without affecting speciﬁcity [76]. Complement ﬁxation test (CFT)

Serum agglutination test (SAT) This test is generally regarded as being unsatisfactory for the purposes of international trade [71] and may be inﬂuenced by Rev-1 and other antigens, and the response can be variable even in the same animal. Therefore, the SAT is recommended only as a screening test, and, in cases in which a low titre is found, additional methods are necessary [13]. Allergic skin test A brucellin allergic skin test can be used in unvaccinated small ruminants for B. melitensis [35, 71]. Many other substances for the allergic diagnosis of brucellosis have been produced through the years. Among these, the ‘Mirri’ allergen has been widely employed in Italy. INRA allergen is extensively used nowadays [72]. It is a product rich in proteins, lacking in LPS, prepared from a strain of B. melitensis (B 115 in rough phase). This allergen induces neither local nor generalized reactions in non-sensitized animals, but gives cutaneous reactions in animals sensitized either by infection or vaccination with members of the genus Brucella. The INRA allergen neither sensitizes

g

This is the most widely used conﬁrmatory test. Despite its complexity and the heterogeneity of techniques used in the different countries, it is agreed that this test is effective in small ruminants [14, 68]. However, CFT lacks sensitivity and does not fully discriminate between infection and Rev-1 vaccination [74]. It must be mentioned that when the Rev-1 vaccine is administered conjunctivally, the problem of interferences is signiﬁcantly reduced in all serological tests, including CFT [77]. Also, the sensitivity of CFT has been reported to be lower in sheep in ﬁeld conditions (88.6%) than those of RBT (92.1%) and 100% of the indirect ELISA (iELISA) [74, 77]. Enzyme-linked immunosorbent assays (ELISA) Good diagnostic results have been obtained in sheep and goats with iELISA or, at a lesser degree, competitive ELISA (cELISA) using various antigens, but generally those with a high content of S-LPS are the most reliable. These ELISA provide similar or better sensitivity than both RBT and CFT, but like classical tests, ELISA are unable to differentiate infected animals from animals recently vaccinated with the Rev-1 vaccine [74, 76–78] or animals infected

CAB International 2010

178

Animal Science Reviews 2010

with cross-reacting bacteria. However, the association of the conjunctival vaccination procedure and the presence of a moderate interval after vaccination minimize or abrogate the speciﬁcity problems. A similar indirect technique has also been proposed for diagnosing sheep brucellosis in individual or pooled milk samples, but the test lacks sensitivity when compared with serological tests. A highly immunogenic periplasmic protein from B. abortus or B. melitensis [79] has been applied to brucellosis diagnosis in different host species. Indirect and competitive ELISA with this antigen could be sensitive and speciﬁc tests for diagnosing B. melitensis infection in sheep and have been reported to be useful in differentiating Rev-1-vaccinated animals from infected animals [50]. In the last few years, several attempts have been made to standardize a test that eventually could clearly differentiate an antibody response of infected sheep from Rev1-vaccinated sheep. Apparently, the goal has been reached by testing sheep sera in an ELISA with partially puriﬁed Brucella-cytosoluble 20-kDa protein, which seems to have the potential for detecting B. melitensis-infected ewes and their differentiation from B. melitensis Rev-1-vaccinated ones [13, 80].

Fluorescence polarization assay (FPA) For detecting antibodies against Brucella spp. FPA has recently been developed based on Perrin’s theory and improvements have been made by Weber and Steiner [81]. The test measures the return photons in the planes parallel and perpendicular to the polarized light that excites the ﬂuorescence molecule (ﬂuorophore), and thus allows the assessment of the ﬂuorophore’s rotation rate, which is inversely proportional to its size. Thus, the rotation rate of the Brucella O-polysaccharide molecule, labelled with ﬂuorescein isothiocyanide (FITC), changes if anti-Brucella LPS antibody binds to it, because of the increase in size of the molecule [82]. FPA is a rapid, homogenous, species-independent assay, which was initially developed and validated for the detection of antibodies to B. abortus in cattle. Information on its performance in detecting antibodies to B. melitensis in sheep and goats is limited. Till date, in the studies conducted to validate FPA for the diagnosis of B. melitensis infection in small ruminants, the assay was carried out in single glass test tubes using a volume of 1 ml, and the results read by an instrument designed to read a single test tube. Recently, new instruments that can read microplates in ﬂuorescence polarization mode have been developed, and these have opened up new prospects for FPA [82].

Diagnostic antigens There is no agreement on what should be the nature and characteristics of a universal antigen for diagnosing

g

brucellosis. One of the most critical and controversial points concerning serological diagnosis of B. melitensis infection in small ruminants is related to which Brucella species and biovars are used in the production of antigens [83]. The antigenic suspensions (whole cells) used in both tests are made with an A-dominant B. abortus biovar 1 [43] and, theoretically, infections due to M-dominant strains (i.e. B. melitensis biovar 1) could be misdiagnosed [43, 68]. However, the existence of a common (C) epitope in the immunodominant S-LPS can account for the high sensitivity of the B. abortus biovar 1 antigens to detect B. melitensis biovar 1 infections and vice versa [68]. No signiﬁcant differences have been found in the sensitivity of the classical B. abortus 1 RBT antigen (AC) between ovine populations infected either with B. melitensis biovar 1 (MC) or 3 (AMC) [75]. Moreover, the iELISA sensitivity in sheep, goats and cattle is not affected by the epitopic composition (AC or MC) of the antigens used [84]. Several authors have attempted to identify the main speciﬁcities of the antibody response to outer membrane proteins (omp) extracts of B. melitensis by using either immunoblotting or cELISA with speciﬁc monoclonal antibodies [85]. However, while omp of 10, 16, 19, 25–27 and 31–34 kDa were found suitable as potential antigens by immunoblotting or ELISA, the antibody responses detected in infected sheep were scant and heterogeneous [85].

Identiﬁcation of vaccine strain B. melitensis strain Rev-1 has the normal properties of a biovar 1 strain of B. melitensis, but develops smaller colonies on agar media, does not grow in the presence of basic fuchsine, thionin (20 mg/ml) or benzyl penicillin (3 mg/ml) (ﬁnal concentrations), but grows in the presence of streptomycin at 2.5 or 5 mg/ml (5 IU/ml) [43, 86]. Vaccine strain Rev-1 may also be identiﬁed using speciﬁc PCR [87]. PCR-RFLP is used successfully to differentiate all vaccine strains from ﬁeld infection using outer membrane proteins 2 gene (omp2) of brucellae which has 2 alleles omp2a and omp2b. This method can differentiate ﬁeld infection with Rev-1 vaccines by producing a different band pattern using PstI endonuclease enzyme [88]. The PstI is a type-II restriction endonuclease (or restriction enzyme) from Providencia stuartii. Omp2a does not have the restriction site of PstI and therefore is not a good target for the differentiation of vaccine strains with ﬁeld strain in PCR RFLP, but omp2b has the mentioned site for the PstI enzyme and can be used successfully for the differentiation of all Brucella vaccine strains from the ﬁeld strains infection [88]. Vaccine strains produce three bands (282, 238 and 44 bp), while ﬁeld strains produce two bands (238 and 44 bp), on PCR-RFLP using PstI endonuclease enzyme [88].

CAB International 2010

Assadullah Samadi, M.M.K. Ababneh, N.D. Giadinis and S.Q. Laﬁ

Treatment The widespread distribution of the bacteria in the body and their ability to survive inside cells render chemotherapy ineffective [13, 89]. There is no practical treatment for infected animals with brucellosis, but long-term antibiotic treatment was sometimes successful in infected dogs. Antibiotic treatment has also been used successfully in some valuable rams, but it is usually not economically feasible. Fertility may remain low even if the organism is eliminated from treated rams [35, 36].

Prevention Brucellosis is usually introduced into a herd of infected animals, but can also be introduced with infected semen. Herd additions should come from brucellosis-free areas or accredited herds. Animals from other sources should be isolated and tested before adding them to the herd. Domesticated animals should always be kept away from contact with wild animal reservoirs [35, 36]. Till date, little has been accomplished with the control and eradication of brucellosis in small ruminants. The best scheme to follow is the identiﬁcation and culling of infected animals. Prophylactic campaigns aimed at eradicating the disease have been successful in the most advanced European countries, but have fallen short of this aim in developing countries. Three strategies are available: (1) Vaccination as a preliminary intervention. (2) Vaccination associated with the culling of infected animals. (3) The identiﬁcation and culling of infected animals with no vaccination. Where the level of infection is not known, a trial investigation is necessary before selecting the most appropriate prophylactic method. However, the following general sanitary measures are considered to be of some beneﬁcial effect in controlling the disease: (1) Introduction of new sheep or goats to the herd should be rigorously controlled, and mating should also be with animals from non-infected herds. (2) Strict isolation is necessary on the introduction of animals susceptible to brucellosis. (3) The mixing at markets or at pasture of healthy animals with infected animals or those of unknown status must be avoided. (4) Periodic clinical and serological examination of rams selected for mating is necessary. (5) Precautions should be taken during the transit of sheep to avoid infection [13, 89]. The main concern of the authorities must be to curb transmission through the elimination of all sources of infection, employing mandatory identiﬁcation of the infection sites and enforcement of all measures aimed at maintaining the health status of the area. Cases of abortion must be reported, conﬁrmed by serological and allergic tests. Infected animals must be culled. If possible, lambing should take place in isolation. The destruction of infected materials, i.e. by incineration, as well as disinfection of the areas where deliveries take place, is necessary

g

179

[36]. Attendants must disinfect hands after handling infected material. It is also important that milking be performed with the maximum hygiene. The use of vaccine is useful only as an addition to the rules mentioned, as the immunizing level given by the vaccine is never complete. The persistence of virulent brucellae in vaccinated sheep may cause a chronic infection that goes undetected, but contributes to the dissemination of infection. In some countries, the test and slaughter policy together with the vaccination of young females is adopted, in others, particularly with regard to sheep and goats; mass vaccination has been recently started [3, 34]. Controlling animal brucellosis in developing countries requires a considerable effort to build infrastructure that educates people about the risks of brucellosis, provides proper laboratory facilities and trained personnel to collect and test samples, keeps a record and arranges for active surveillance programmes [90]. The farmers, milk and milk product industry, breeding companies, consumers and the politicians must work together and ﬁnd a practical eradication effort that is suitable for each country [90]. The control of B. melitensis infection in small ruminants can be achieved if the population’s resistance to disease is increased by vaccination. It is accepted that vaccination is more acceptable and effective than other methods applied for this purpose. Also, the vaccines used practically eliminate the clinical signs of brucellosis, and so contamination of the environment and exposure of population at risk to the infectious agent are reduced [89].

Vaccines Rev-1 strain of B. melitensis-attenuated vaccine is mainly used against brucellosis in sheep and goats [40]. The Rev-1 vaccine was developed by [91] from a nondependent reverse mutant of the virulent streptomycindependent strain of B. melitensis 5056 as a live vaccine against B. melitensis in goats and sheep [92]. Rev-1 is a smooth bacterium with a complete LPS, which induces a similar antibody response to that caused by the ﬁeld strains and cannot be easily differentiated by conventional serology. The vaccine is a streptomycin-resistant B. melitensis, and can become an important health problem if accidents occur involving people working with the vaccine [45]. Vaccinated animals produce antibodies that can be demonstrated by serological tests. Complement-ﬁxing antibodies disappear 6–8 weeks after vaccination, in most cases. For this reason, Rev-1 is employed only in young animals before they reach their reproductive age. However, in some instances, the vaccine is also administered to adult ewes, in which, in order to reduce the likelihood of abortions and excretion in the milk, a reduced dose of the vaccine and subconjunctival administration have been used. Considering the fact that subcutaneous administration confers a longer persistence of antibodies and the obvious difﬁculty in performing two separate vaccinations,

CAB International 2010

180

Animal Science Reviews 2010

it has been suggested that the animals should be vaccinated only once, intraconjunctivally, with a 109 dose of vaccine [72]. Following Rev-1 vaccination, the bacteria are disseminated widely, followed by their localization in the prescapular lymph nodes on the side of inoculation, with a possible spread to the cranial lymph nodes. In most cases, the organism disappears after 3 months [93]. Rev-1 induces a very efﬁcient immunity, lasting more than 2 years. A deletion mutant of strain Rev-1 was recently obtained which, according to the results of tests conducted in mice, would allow serological differentiation between infected and vaccinated sheep [94]. The deleted gene codes for the periplasmic protein BP26, the immunodominant antigen in the serological response of B. melitensis in sheep. The authors suggest, if proven safe and effective in the target species (sheep), the use of the Rev-1 bp26 deletion mutant as a vaccine for the eradication of B. melitensis infection in sheep. Several other prospective candidate vaccines are also under investigation [95]. In areas with widespread infection or the likelihood of re-infection, it is recommended that adult animals receive reduced doses of the vaccine [96]. Ideal qualities of a vaccine candidate include: (1) long duration of immunogenesis; (2) minimum interference with diagnostic tests; (3) easy production and storage of the vaccine with long stability and (4) minimum adverse effects in vaccinated animals with no danger to humans in the event of exposure [97]. Vaccination alone will not eradicate Brucella as the immunity produced by Brucella vaccines are not absolute and can be circumvented by increasing the level of infection. It is obvious, therefore, that a policy of vaccination is more likely to succeed if combined with good measures of husbandry. The ability of Rev-1 vaccine to produce a high level of immunity against both experimentally induced and natural infections has been convincingly demonstrated for both sheep and goats [14]. It has been well established that a large proportion of vaccinated animals are protected against infection, and in those vaccinated animals where infection occurred, it is often transitory and the period of B. melitensis excretion from the udder or vagina is shorter. Because of this, the degree of microbial contamination of the surroundings is reduced and consequently, disease transmission within and between ﬂocks is signiﬁcantly reduced [98]. The routes of administration of vaccine have valuable effect on the duration of immunity and safety consideration in pregnant animals. Subcutaneous vaccination with Rev-1 is recommended for goats and sheep between 4 and 6 months of age. Subcutaneous vaccination with standard doses (1–2109) of Rev-1 produces a generalized and persistent infection (2–3 months) with an active colonization of spleen and several lymph nodes, inducing an intense and durable serological response, which can interfere in the diagnostic tests [45, 77, 99]. This should not be relevant in epidemiological situations, where eradication is not feasible and the only objective is to induce

g

the highest level of immunity in the highest number of susceptible animals. The problem of serological interference is only important when implementing a combined vaccination and eradication programme. However, it is possible to use diagnostic tests with a reasonable speciﬁcity to discriminate serological responses caused by infection from those caused by vaccination [77]. Moreover, conjunctival vaccination (a single dose of 1–2109 in a volume of 30–50 litres) confers an adequate protection in replacement (3–6-month-old) animals, minimizing the serological interferences [77]. It is important to remember that the younger the Rev-1-vaccinated animals are, the lower serological interference is produced [45, 77]. Conjunctival vaccination induces sufﬁcient protection in animals of 3–6 months of age and also reduces the possibility of serological interference. Both standard and reduced doses of Rev-1 induce abortions in sheep and goats vaccinated during pregnancy. Studies have demonstrated that even reduced doses as low as 106, used either subcutaneously or conjunctivally, have been demonstrated to induce abortions and milk excretion of the vaccine strain [12] in sheep and goats [100]. Nevertheless, conjunctival vaccination of animals with Rev-1 during breeding periods can effectively provide protection and reduce the risk of vaccine-induced abortions [12]. A ﬁeld study to evaluate the serological response and the safety of different doses and administration routes of the Rev-1 vaccine was carried out on two Churra breed ﬂocks. Reduced doses of live organisms were administered by the subcutaneous or the conjunctival route. In those animals which were seropositive before vaccination, the percentage of positive sera declined progressively in a similar way in all groups over the 36 weeks that the study lasted; the antibody titres also dropped continuously in the group vaccinated by the conjunctival route with the lower dose, while in the remaining three groups there was a transitory increase in the fourth week after vaccination. In those animals which were serologically negative prior to vaccination, the percentage of positive sera and the antibody titres generally reached their peak in the fourth week after vaccination, followed by a progressive decline in succeeding weeks. Similarly, titres were higher in animals vaccinated subcutaneously than in those vaccinated by the conjunctival route. The differences between the frequencies of positive sera and the levels of antibodies were important when routes were compared. No cases of abortion were reported in the 461 vaccinated ewes. With both these vaccination techniques, subsequent serological response was lower and of shorter duration than with the standard vaccination [101].

Eradication of B. melitensis infection in small ruminants Eradication of an infection means the extinction of infectious agent completely, and under this concept, very few

CAB International 2010

Assadullah Samadi, M.M.K. Ababneh, N.D. Giadinis and S.Q. Laﬁ

diseases until now have been eradicated. Especially for veterinary medicine, eradication implies the disappearance of an infectious agent from an area or a deﬁned population. It is achieved after the implementation of a highly organized campaign, so that no further outbreaks occur, unless the infectious agent had been reintroduced in the area. Eradication of brucellosis in small ruminants can be achieved mainly by the depopulation of infected ﬂocks. On a national scale, it can be achieved only by identifying infected ﬂocks and slaughtering all animals in these ﬂocks. This procedure is feasible where B. melitensis infection has been newly introduced into a previously non-infected area or where the prevalence of infection is very low [102]. Nationwide eradication programmes for B. melitensis include quarantines of infected herds, vaccination, test-andslaughter and/or depopulation techniques, cleaning and disinfection of infected farms, various forms of surveillance and trace backs [36]. Eradication with the implementation of test-andslaughter policy is possible under certain conditions, when sheep and goat ﬂocks are under strict control, the type of husbandry is not extensive or nomadic, there is not common grazing or transhumance of the ﬂocks, an efﬁcient identiﬁcation system of the animals is in place, the ﬁnancial and other resources are available for a long period of time and the veterinary service responsible for the programme is well organized. In case the B. melitensis infection is endemic or its prevalence is high in an area, the ﬁrst step for its eradication must be the control of the disease by immunization of the susceptible animals [89]. The most efﬁcient strategy for the control of the disease is the vaccination of young and adult animals with Rev-1 vaccine administered conjunctivally. This strategy is considered as the most appropriate in cases that the prevalence of infection is high among animals and the type of husbandry is extensive or nomadic. This strategy has been proved to be the most effective way for the control of Brucella infection in small ruminants since the immunity of the ﬂock develops rapidly. Any strategy for the control or eradication of brucellosis in order to be effective must have the support and cooperation of the farmers, must be well planned so that all the resources needed to be available on time, the veterinary service that is responsible must be well organized and has the ability and patience for implementing long-term control and eradication programmes [89]. Only a few countries in which B. melitensis infection is present have the resources required for eradication implementing this strategy. If brucellosis is well-established in a deﬁned area, eradication is nearly always achieved at an enormous cost in terms of resources and patience. Since many countries have succeeded in eradicating B. abortus from their cattle population by test-andslaughter strategy, it was thought that B. melitensis infection could also be eradicated from small ruminant ﬂocks with the same procedure. The experiences gained from many countries on this issue suggest that the feasibility of

g

181

brucellosis eradication in small ruminants implementing test-and-slaughter strategy appears to depend largely on the conditions under which they are kept. The chances of success are high if the ﬂocks are small, isolated and kept under close control and low in large ﬂocks, especially when they are in close contact with other ﬂocks, with common grazing and transhumance practised [102]. International agencies have, therefore, proposed that whole ﬂock vaccination should precede any testand-slaughter programme, until the disease prevalence is signiﬁcantly reduced. Only then should test-and-slaughter be implemented as part of a national eradication scheme [103]. It has to be mentioned that vaccination is not allowed in Member States or regions of the EU in which ofﬁcial brucellosis-free status (B. melitensis) has been achieved or is being sought [13]. It must be borne in mind that the diagnostic tests used are unable to reveal all infected animals and may give false negative because of the incubation period, latency or because of the criteria used to interpret the results [104]. The full cooperation of farmers is also essential, as slaughter of seropositive animals can be resisted by owners, because of the lack of clinical signs, inadequate compensation or lack of replacement animals. It is usually considered that a brucellosis eradication programme by test-and-slaughter policy is justiﬁed on economic grounds only when the prevalence of infected animals in an area is 2% or below and the ﬂocks are maintained under closely controlled conditions and protected efﬁciently against re-entry of infection [104].

References 1. Seleem MN, Stephen M, Sriranganathan N. Brucellosis: a re-emerging zoonosis. Veterinary Microbiology 2010;140: 392–8. 2. Benkirane A. Ovine and caprine brucellosis: world distribution and control/eradication strategies in West Asia/North Africa region. Small Ruminant Research 2006;62:19–25. 3. Mantur BG, Amarnath SK. Brucellosis in India – a review. Journal of Bioscience 2008;33:539–47. 4. Gul ST, Khan A. Epidemiology and epizootology of brucellosis: a review. Pakistan Veterinary Journal 2007;27(3):145–51. 5. Enright FM. The pathogenesis and pathogenicity of Brucella infection in domestic animals. In: Nielson KH, Duncan JR, editors. Animal Brucellosis. CRC Press, Boston, MA; 1990. p. 301–20. 6. Omer MK, Skjerve E, Holstad G, Woldehiwet Z, Macmillan AP. Prevalence of antibodies to Brucella spp. in cattle, sheep, goats, horses and camels in the State of Eritre; inﬂuence of husbandry systems. Epidemiology and Infection 2000;125:447–53. 7. Radostits OM, Blood D, Gay CC. Veterinary Medicine, 9th edn. W.B. Saunders Co., Philadelphia, PA; 2000. 8. FAO/OIE/WHO. 1995 Animal Health Yearbook. FAO Animal Production and Health Series. FAO, USA; 1997.

CAB International 2010

182

Animal Science Reviews 2010

9. Aldomy F, Hussain NO, Sawalha L, Khatatbeh K, Aldomy A. A national survey of perinatal mortality in sheep and goats in Jordan. Pakistan Veterinary Journal 2009;29(3):102–6.

Brucella melitensis 16M. Journal of Proteome Research 2006;5:1731–40.

10. FAO/WHO. Joint FAO/WHO Expert Committee on Brucellosis: Sixth Report, Technical Report Series 740. WHO, Geneva, Switzerland; 1986.

26. Mujer CV, Wagner MA, Eschenbrenner M, Horn TA, Kraycer JA, Redkar R, et al. Global analysis of Brucella melitensis proteomes. Annals of the New York Academy of Sciences 2002;969:97–101.

11. OIE. Manual of Diagnostic Tests and Vaccines for Terrestrial Animals, 5th edn. Ofﬁce International des Epizooties, Paris, France; 2004. p. 598–606.

27. Seleem MN, Boyle SM, Sriranganathan N. Brucella: a pathogen without classic virulence genes. Veterinary Microbiology 2008;129:1–14.

12. Blasco JM. A review of the use of B. melitensis Rev-1 vaccine in adult sheep and goats. Preventive Veterinary Medicine 1997;31:275–83.

28. Perry MB, Bundle DR. Lipopolysaccharide antigens and carbohydrates of Brucella. In: Adams LG, editor. Advances in Brucellosis Research. Texas A& M University, Austin, TX; 1990. p. 76–88.

13. Aitken ID. Diseases of Sheep, 4th edn. Blackwell, Ames, IA; 2007. p. 137–42. 14. Alton GG. Brucella melitensis. In: Nielsen K, Duncan JR, editors. Animal Brucellosis. CRC Press, Boca Raton, FL; 1990. p. 383–409.

29. Bundle DR, Cherwonogrodzky JW, Gidney MAJ, Meikle PJ, Perry MB, Peters T. Deﬁnition of Brucella A and M epitopes by monoclonal typing reagents and synthetic oligosaccharides. Infection and Immunity 1989;57:2829–36.

15. Fretin D, Fauconnier A, Kohler S, Halling S, Leonard S, Nijskens C, et al. The sheathed ﬂagellum of Brucella melitensis is involved in persistence in a murine model of infection. Cellular Microbiology 2005;7:687–98.

30. Minnick MF, Stiegler GL. Nucleotide sequence and comparison of the 5S ribosomal genes of Rochalimaea henselae, R. quintana and Brucella abortus. Nucleic Acids Research 1993;21:2518.

16. Young EJ. An overview of human brucellosis. Clinical Infectious Diseases 1995;21:283–90.

31. Wilson GS, Miles AA. The serological differentiation of smooth strains of the Brucella group. British Journal of Experimental Pathology 1932;13:1–13.

17. Pappas G, Bosilkovski M, Akritidis N, Mastora M, Krteva L, Tsianos E. Brucellosis. New England Journal of Medicine 2005;352:2325–36. 18. Yanagi M, Yamasato K. Phylogenetic analysis of the family Rhizobiaceae and related bacteria by sequencing of 16S rRNA gene using PCR and DNA sequencer. FEMS Microbiology Letters 1993;107:115–20. 19. Sriranganathan N, Seleem MN, Olsen SC, Samartino LE, Whatmore AM, Bricker B, et al. Brucella. In: Vishvanath N, Chittaranjan K, editors. Genome mapping and genomics in animal-associated microbes. Springer, New York; 2009. Chapter 1. 20. DelVecchio VG, Kapatral V, Elzer P, Patra G, Mujer CV. The genome of Brucella melitensis. Veterinary Microbiology 2002;90:587–92. 21. Halling SM, Peterson-Burch BD, Bricker BJ, Zuerner RL, Qing Z, Li LL, et al. Completion of the genome sequence of Brucella abortus and comparison to the highly similar genomes of Brucella melitensis and Brucella suis. Journal of Bacteriology 2005;187:2715–26. 22. Paulsen IT, Seshadri R, Nelson KE, Eisen JA, Heidelberg JF, Read TD, et al. The Brucella suis genome reveals fundamental similarities between animal and plant pathogens and symbionts. Proceedings of the National Academy of Sciences, USA 2002;99:13148–53. 23. Wattam AR, Williams KP, Snyder EE, Almeida NF Jr, Shukla M, Dickerman AW, et al. Analysis often Brucella genomes reveals evidence for horizontal gene transfer despite a preferred intracellular lifestyle. Journal of Bacteriology 2009;191:3569–79.

32. del C Rocha-Gracia R, Castaneda-Roldan EI, Giono-Cerezo S, Giron JA. Brucella sp. binds to sialic acid residues on human and animal red blood cells. FEMS Microbiology Letters 2002;213:219–24. 33. Garin-Bastuji B, Lasco JM, Grayon M, Verger JM. B. melitensis infection in sheep: present and future. Veterinary Research 1998;29:255–74. 34. Refai M. Incidence and control of brucellosis in the Near East region. Veterinary Microbiology 2002;90(1–4):81–110. 35. Center for Food Security and Public Health (CFSPH). Brucellosis. Iowa State University, College of Veterinary Medicine; 2007. p. 1–13. Available from: URL http://www. cfsph.iastate.edu/Brucellosis (last updated June 2007). 36. Center for Food Security and Public Health (CFSPH). Ovine and Caprine Brucellosis, Iowa State University, College of Veterinary Medicine; 2007. p. 1–5. Available from: URL: http// www.cfsph.iastate.edu/Brucellosis (last update 29 July 2009). 37. Grillo´ MJ, Barbera´n M, Blasco JM. Transmission of Brucella melitensis from sheep to lambs. Veterinary Records 1997;140:602–5. 38. Renukaradhay GJ, Isloor S, Rajasekhar M. Epidemiology, zoonotic aspects, vaccination and control/eradication of brucellosis in India. Journal of Veterinary Microbiology 2002;90:183–95. 39. Renoux G, Alton GG, Sacquet E. Etudes sur la brucellose ovine et caprine. X. Elimination de B. melitensis par le lait de chevres suedises artiﬁciellement infectess. Archives de l’Institut Pasteur de Tunis 1953;33:413.

24. Ding XZ, Paulsen IT, Bhattacharjee AK, Nikolich MP, Myers G, Hoover DL. A high efﬁciency cloning and expression system for proteomic analysis. Proteomics 2006;6:4038–46.

40. Alton GG. The epidemiology of B. melitensis in sheep and goats. In Verger JM, Plommet M, editors. B. melitensis: A CEC Seminar. Martinus Nijoff, Dordrecht–Boston– Lancaster; 1985. p. 187–96.

25. Eschenbrenner M, Horn TA, Wagner MA, Mujer CV, Miller-Scandle TL, Del Vecchio VG. Comparative proteome analysis of laboratory grown Brucella abortus 2308 and

41. Alton GG. The reactions of goats naturally infected with B. melitensis to vaccination with living attenuated vaccine. Research in Veterinary Science 1962;3:326.

g

CAB International 2010

Assadullah Samadi, M.M.K. Ababneh, N.D. Giadinis and S.Q. Laﬁ 42. Biggi P. Indagini sulla batteriemia brucellare in pecore sperimentalmente infettate. Annali dello Facolta di Medicina Veterinaria – Universita di Pisa 1956;9:264. 43. Alton GG, Jones LM, Angus RD, Verger JM. Techniques for the Brucellosis Laboratory. Institut National de la Recherche Agronomique, Nouzilly, France; 1988. 44. Kojouri GA, Gholami M. Post vaccination follow-up of Brucella melitensis in blood stream of sheep by PCR assay. Comparative Clinical Pathology 2009;18:439–42. 45. Blasco JM. Existing and future vaccines against brucellosis in small ruminants. Small Ruminant Research 2006;62:33–7. 46. Fensterbank R, Verger JM, Grayon M. Conjunctival vaccination of young goats with B. melitensis strain Rev-1. Annales de Recherches Ve´te´rinaires 1987;18(4):397–403. 47. Lilenbaum W, de Souza GN, Ristow P, Moreira MC, Fraguas S, Cardoso Vda S, et al. A serological study on Brucella abortus, caprine arthritis-encephalitis virus and Leptospira in dairy goats in Rio de Janeiro. Brazilian Veterinary Journal 2007;173:408–12. 48. Glynn MK, Lynn TV. Brucellosis. Journal of American Veterinary Medical Association 2008;233:900–8. 49. Marin CM, Alabart JL, Blasco JM. Effect of antibiotics contained in two Brucella selective media on growth of Brucella abortus, B. melitensis and B. ovis. Journal of Clinical Microbiology 1996;34:426–8. 50. Garin-Bastuji B, Blasco JM. Caprine and ovine brucellosis (excluding B. ovis). In: Manual of Diagnostic Tests and Vaccines for Terrestrial Animals, 5th edn. OIE, Paris, France; 2004. p. 598–606. 51. Marin CM, Jime´nez de Bagu¨e´s MP, Barbera´n M, Blasco JM. Comparison of two selective media for the isolation of Brucella melitensis from naturally infected sheep and goats. Veterinary Records 1996;138:409–11.

183

porpoises, and dolphins. Journal of Veterinary Diagnostic Investigation 1999;11:152–7. 59. Halling SM, Tatum FM, Bricker BJ. Sequence and characterization of an insertion sequence, IS711, from Brucella ovis. Gene 1993;133:123–7. 60. Baily GG, Krahn JB, Drasar BS, Stoker NG. Detection of Brucella melitensis and B. abortus by DNA ampliﬁcation. Journal of Tropical Medicine and Hygiene 1992;95:271–5. 61. Da Costa M, Guillou JP, Garin-Bastuji B, ThiJbaud M, Dubray G. Speciﬁcity of six gene sequences for the detection of the genus Brucella by DNA ampliﬁcation. Journal of Applied Bacteriology 1996;81:267–75. 62. Fekete A, Bantle JA, Halling SM, Sanborn MR. Preliminary development of a diagnostic test for Brucella using polymerase chain reaction. Journal of Applied Bacteriology 1990; 69:216–27. 63. Romero C, Gamazo C, Pardo M, Lopez-Gon˜i I. Speciﬁc detection of Brucella DNA by PCR. Journal of Clinical Microbiology 1995;33:615–7. 64. Cloeckaert A, Grayon M, Gre´pinet O. Identiﬁcation of Brucella melitensis vaccine strain Rev-1 by PCR-RFLP based on a mutation in the rpsL gene. Vaccine 2002;20:2546–50. 65. Amin AS, Hamdy MER, Ibrahim AK. Detection of B. melitensis in semen using the polymerase chain reaction assay. Veterinary Microbiology 2001;83:37–44. 66. Rees RK, Graves M, Caton N, Ely JM, Probert WS. Single tube identiﬁcation and strain typing of Brucella melitensis by multiplex PCR. Journal of Microbiological Methods 2009;78:66–70. 67. Anne MS, Draeger A, Go¨llner C, Scholz HC, No¨ckler K. Advancement of a multiplex PCR for the differentiation of all currently described Brucella species. Journal of Microbiological Methods 2010;80:112–4.

52. Gupta VK, Deepak K, Singh VK, Ranjeeta Kumari SV, Singh Vihan VS. Single-step PCR for detection of Brucella melitensis from tissue and blood of goats. Small Ruminant Research. 2006;66:169–74.

68. MacMillan A. Conventional serological tests. In: Nnielsen K, Dnuncan JR. editors, Animal Brucellosis. CRC Press, Boca Raton, FL; 1990. p. 153–98.

53. Bricker BJ. PCR as a diagnostic tool for brucellosis. Veterinary Microbiology 2002;90:435–46.

69. Nielsen K, Smith P, Yu W, Nicoletti P, Elzer P, Robles C, et al. Towards single screening tests for brucellosis. Reviews in Science and Technology 2005;24:1027.

54. Bricker BJ, Ewalt DR, Olsen SC, Jensen AE. Evaluation of the Brucella abortus species speciﬁc polymerase chain reaction assay, an improved version of the Brucella AMOS polymerase chain reaction assay for cattle. Journal of Veterinary Diagnostic Investigation 2003;15:374–8. 55. Bricker BJ, Halling SM. Differentiation of Brucella abortus bv 1, 2, and 4, Brucella melitensis, Brucella ovis, and Brucella suis bv 1 by PCR. Journal of Clinical Microbiology 1994;32:2660–6. 56. Bricker BJ, Halling SM. Enhancement of the Brucella AMOS PCR assay for differentiation of Brucella abortus vaccine strains S19 and RB51. Journal of Clinical Microbiology 1995;33:1640–42. 57. Ocampo-Sosa AA, Agu¨ero-Balbı´ n J, Garcı´ a-Lobo JM. Development of a new PCR assay to identify Brucella abortus biovars 5, 6 and 9 and the new subgroup 3b of biovar 3. Veterinary Microbiology 2005;110:41–51. 58. Jensen AE, Cheville NF, Thoen CO, Macmillan AP, Miller WG. Genomic ﬁngerprinting and development of a dendrogram for Brucella spp. isolated from seals,

g

70. Collective Manual. Bovine brucellosis. In Manual of Diagnostic Tests and Vaccines for Terrestrial Animals, 5th edn. OIE, Paris, France; 2004. p. 409–38. 71. OIE. Terrestrial Manual. Caprine and Ovine Brucellosis (excluding Brucella ovis), Volume II. Ofﬁce International des Epizooties, Paris, France; 2009. Chapter 2.7.2. p. 974–83. 72. Fensterbank R. Allergic diagnosis of brucellosis. In: Plommet M, Verger JM, editors. B. melitensis. Martinus Nijhoff Publisher, Dordrecht; 1985. p. 167–72. 73. Farina R. Allergiereazione brucellinica e jaluronidasi. Annali della Facolta´ di Medicina Veterinaria di Pisa 1953;VI:3–7. 74. Blasco JM, Garin-Bastuji B, Marin CM, Gerbier G, Fanlo J, Jime´nez de Bagu¨e´es MP, et al. Efﬁcacy of different rose bengal and complement ﬁxation antigens for the diagnosis of Brucella melitensis in sheep and goats. Veterinary Records 1994;134:415–20. 75. Blasco JM, Marı´ n CM, Jime´nez de Bagu¨e´s MP, Barbera´n M, Hernande´z A, Molina L, et al. Evaluation of allergic and

CAB International 2010

184

Animal Science Reviews 2010 serological tests for diagnosis of B. melitensis in sheep. Journal of Clinical Microbiology 1994;32:1835–40.

76. Ferreira AC, Cardoso R, Travassos Dias I, Mariano I, Belo A, Rolao Preto I, et al. Evaluation of a modiﬁed Rose Bengal test and an indirect enzyme-linked immunosorbent assay for the diagnosis of Brucella melitensis infection in sheep. Veterinary Research 2003;34:297–305. 77. Marin CM, Moreno E, Moriyo´n I, Dı´ az R, Blasco JM. Performance of competitive and indirect ELISAs, gel immunoprecipitation with native hapten polysaccharide and standard serological tests in diagnosis of sheep brucellosis. Clinical Diagnostic Laboratory Immunology 1999;6:269–72. 78. Ficapal A, Alonso B, Velasco J, Moriyo´n I, Blasco JM. Diagnosis of B. ovis infection of rams with an ELISA using protein G as conjugate. Veterinary Records 1995;137:145–7. 79. Cloeckaert A, Salih-Alj A, Debarrh H, Zygmunt MS, Dubray G. Polymorphism at the dnak locus of Brucella species and identiﬁcation of a Brucella melitensis speciesspeciﬁc marker. Journal of Medical Microbiology 1996;45:200–13. 80. Debbarh HSA, Zygmunt M, Dubray G, Cloeckaert A. Competitive enzyme-linked immunosorbent assay using monoclonal antibodies to the B. melitensis BP26 protein to evaluate antibody responses in infected and B. melitensis Rev-1 vaccinated sheep. Veterinary Microbiology 1996;53:325–37. 81. Dandliker WB, De Saussure VA. Fluorescence polaization in immunochemistry. Immunochemistry 1970;7:799. 82. Minas A, Stournara A, Minas M, Stack J, Petridou E, Christodoulopoulos G, et al. Validation of a ﬂuorescence polarization assay (FPA) performed in microplates and comparison with other tests used for diagnosing B. melitensis infection in sheep and goats. Journal of Immunological Methods 2007;320:94–103. 83. Garin-Bastuji B, Blasco JM, Marı´ n C, Albert D. The diagnosis of brucellosis in sheep and goats, old and new tools. Small Ruminant Research 2006;62:63–70. 84. Alonso-Urmeneta B, Marin CM, Aragon V, Blasco JM, Diaz R, Moriyon I. Evaluation of lipopolysaccharides and polysaccharides of different epitopic structures in the enzyme-linked immunosorbent assay for diagnosis of brucellosis in small ruminants and cattle. Clinical and Diagnostic Laboratory Immunology 1998;5: 749–54. 85. Zygmunt MS, Debbarh HSA, Cloeckaert A, Dubray G. Antibody response to Brucella melitensis outer membrane antigens in naturally infected and Rev-1 vaccinated sheep. Veterinary Microbiology 1994;39:33–46. 86. Diaz R, Garatea P, Jones LM, Moriyon I. Radial immunodiffusion test with a Brucella polysaccharide antigen for differentiating infected from vaccinated cattle. Journal of Clinical Microbiology 1979;10(1):37–41. 87. Garcia D, Marin CM, De Miguel PM, Munoz PM, Vizmanos JL, Lopez-Goni I. Multiplex PCR assay for the identiﬁcation and differentiation of all Brucella species and the vaccine strains Brucella abortus S19 and RB51 and Brucella melitensis Rev-1. Clinical Chemistry 2006;52: 779–81.

g

88. Bardenstein S, Mandelboim M, Ficht AT, Baum M, Banai M. Identiﬁcation of the Brucella melitensis vaccine strain Rev-1 in animals and humans in Israel by PCR analysis of the PstI site polymorphism of its omp2 gene. Journal of Clinical Microbiology 2002;40(4):1475–80. 89. Minas A. Control and eradication of brucellosis in small ruminants. Small Ruminant Research 2006;62:101–7. 90. Seleem MN, Stephen M, Sriranganathan N. Brucellosis: a re-emerging zoonosis. Veterinary Microbiology 2009;10:1016. 91. Elberg SS, Faunce K. Immunization against Brucella infection VI. Immunity conferred on goats by a nondependent mutant from a streptomycin-dependent mutant strain of Brucella melitensis. Journal of Bacteriology 1957;73:211–7. 92. Sanmartino L. Brucellosis Vaccines. In The 58th Brucellosis Conference, Merida, Yucatan, Mexico; 2005. p. 31–41. 93. Lantier F, Fensterbank R. Kinetics of Rev-1 infection in sheep. In: Verger JM, Plommet M, editors Brucella melitensis. Martinus Nijhoff, Dordrecht; 1985. p. 247–51. 94. Cloeckaert A, Jacques I, Grillo MJ. Development and evaluation as vaccines in mice of Brucella melitensis Rev-1 single and double deletion mutants of the bp26 and omp31 genes coding for antigens of diagnostic signiﬁcance in ovine brucellosis. Vaccine 2004;22:2827–35. 95. Blasco JM. Existing and future vaccines against brucellosis in small ruminants. In Proceedings of the 6th International Sheep Veterinary Congress, Hersonissos, Crete; 2005. p. 44–6. 96. Adone R, Francia M, Ciuchini F. Evaluation of Brucella melitensis B115 as rough-phenotype vaccine against B. melitensis and B. ovis infections. Vaccine 2008;26: 4913–7. 97. Nicoletti P. Vaccination. In: Nielsen K, Duncan JR, editors. Animal Brucellosis. CRC Press, Boca Raton, FL; 1990. p. 283–99. 98. Garrido F. Rev-1 and B-19 vaccine control in Spain. Observations on the handling and effectiveness of Rev-1 vaccine and the immune response. In: Plommet, M. editor, Prevention of Brucellosis in the Mediterranean Countries. Pudoc Scientiﬁc Publishers, Wageningen, The Netherlands; 1992. p. 223–31. 99. Fensterbank R, Pardon P, Marly J. Comparison between subcutaneous and conjunctival route of vaccination of Rev-1 strain against B. melitensis infection in ewes. Annales de Recherches Ve´te´rinaires 1982;13:295–301. 100. Alton GG. Vaccination of goats with reduced doses of Rev-1 B. melitensis vaccine. Research in Veterinary Science 1970;11:54–9. 101. Delgado S, Ferna´ndez M, Ca´rmenes P. Evaluation of an enzyme-linked immunosorbent assay for the detection of sheep infected and vaccinated with B. melitensis. Journal of Veterinary Diagnostic Investigation 1995;7:206–9. 102. WHO. Joint FAO/WHO Expert Committee on Brucellosis, Technical Report 740. WHO; 1986. 103. Banai M. Control of small ruminant brucellosis by use of Brucella melitensis Rev-1 vaccine: laboratory aspects and ﬁeld observations. Veterinary Microbiology 2002;90: 497–519. 104. Nicoletti P. The eradication of brucellosis in animals. Saudi Medical Journal 1993;14:288–92.

CAB International 2010

Animal Science Reviews 2010

Plants as reservoirs for human enteric pathogens Nicola J. Holden* Address: Plant Pathology, The Scottish Crop Research Institute, Invergowrie, Dundee DD2 5DA, UK. *Correspondence: Email: [email protected] 22 February 2010 30 June 2010

Received: Accepted:

Abstract Outbreaks of enteric pathogens associated with fresh produce, in the form of raw or minimally processed fruit and vegetables, have increased over recent times. The pathogens may be bacterial, viral or protoctist parasites and many are considered to be zoonotic. Contamination of fresh produce occurs by a number of routes, in particular from irrigation water or manure. While some pathogens are vectored into the food chain by plants, others are able to colonize plants and use them as alternative hosts. Enteric bacteria normally associated with animal hosts can colonize and proliferate on or within plants, in particular in the rhizosphere. Some have the ability to internalize into plant tissue, from where they can be detected in edible foliage. Speciﬁc interactions have been demonstrated that are dependent on both the microbe and plant host. Colonization of non-animal hosts by enteric pathogens has, until recently, been largely neglected. However, alternative environmental reservoirs such as plants need to be considered in the context of transmission of enteric pathogens through the food chain. Keywords: Enteric pathogens, Food-borne, Outbreak, Fresh produce, Fruit and vegetables, Plant hosts Review Methodology: The following databases were searched: ISI Web of Knowledge and PubMed (keyword search terms and pathogen species names). In addition, references from the articles obtained by this method were used to check for additional relevant material.

Introduction Fresh fruits and vegetables are vital sources of nutrition and fundamental components of our diet. These food items are collectively referred to as fresh produce. The microbiological safety of fresh produce is a major area of concern, since these foods are often consumed as readyto-eat or raw. The incidence of food-borne outbreaks caused by human enteric pathogens has markedly increased over recent times, which has necessitated the revision of hazard analysis and critical control point protocols, and resulted in changes to agricultural and food production practices. It has also triggered a series of studies within the scientiﬁc community to investigate the biological basis of the problem. However, fresh produce acting as a source of enteric pathogens is not a recent phenomenon. Although the increase in incidence can partly be accounted for by increased consumption of

g

fresh produce as part of a ‘healthy’ diet, and improved detection and surveillance methods, these aspects do not fully account for the increase [1]. What is becoming clear from the scientiﬁc community is that some pathogens are not merely transported by plants into the food chain in an inert manner, instead the microbes interact with plants and are able to utilize them as hosts, akin to a parasitic or commensal relationship. As such, plants can represent an alternative environmental reservoir for many of the most important human enteric pathogens. The evidence for active interactions to support this assertion is discussed.

Causative Organisms A wide range of microbial enteric pathogens from almost all the microbial kingdoms have been associated with fresh produce and have been the cause of many serious

CAB International 2010

186

Animal Science Reviews 2010

outbreaks (described elsewhere [2–7]). Bacterial enteric pathogens of note that have been associated with fresh produce are toxigenic Escherichia coli, non-typhoidal Salmonella enterica, Shigella species, Bacillus cereus, Campylobacter species, Clostridium species and Listeria monocytogenes. Protoctists associated with fresh produce include the apicomlexans Cryptosporidium and Cyclospora and enteric viral pathogens include hepatitis A virus and norovirus (NoV). (Fungi and invertebrate parasites are not considered as they do not cause enteric disease in humans.) The most prevalent food-borne enteric pathogens in the developed countries, for all food groups, are S. enterica, Campylobacter species and NoV [8–10].

Bacteria The bacterial pathogens associated with fresh produce represent a diversity of Gram-negative and Gram-positive bacteria. Several important human pathogens belong to the Enterobacteriaceae, including pathogenic E. coli, S. enterica and Shigella species. There are four distinct classes of enterovirulent E. coli, but the majority of reported outbreaks from fresh produce are almost all caused by bacteria within the Verotoxigenic E. coli (VTEC) group. This group is responsible for haemorrhagic colitis and sometimes fatal haemolytic uremic syndrome in humans, and those isolates associated with symptomatic disease are termed enterohaemorrhagic E. coli (EHEC). The serotype O157:H7 belongs to the VTEC group and caused more that half of the human VTEC infections in the European Union member states in 2008 [8]. EHEC have been associated with fresh produce for a reasonably long time, with one of the earliest reports of entertoxigenic E. coli from 1974 [11]. Notable EHEC outbreaks associated with fresh produce include a multi-state outbreak in the USA from spinach [12], a large-scale Japanese outbreak from radish sprout salad [13] and an outbreak in Northern Europe from lettuce [14]. Genomic sequence data are available for three EHEC isolates, two of which are associated with fresh produce. It is of note that both the radish sprout-outbreak isolate and the spinach-outbreak isolate may be more pathogenic than the third sequenced EHEC [15], a clinical isolate linked to ground beef [16]. We and others have suggested that pathogens associated with fresh produce encode genes that promote adaptation to plant hosts [15, 17, 18]. Epidemiological data show that fresh produce can be associated with approximately 21% of VTEC outbreaks that have occurred in the USA over a 20-year period [19]. Internationally, fresh produce accounted for 19.5% of all the food groups associated with E. coli outbreaks [20]. Shigella species are closely related to toxigenic E. coli and both groups encode Shiga toxin responsible for the severe disease outcomes. Although food-borne shigellosis is less frequently reported than other Enterobacteria [8], as a genus Shigella account for 4.9% of all outbreaks associated

g

with fresh produce, while the fresh produce accounted for 29% of all the food groups associated with Shigella outbreaks [20]. S. enterica is one of the most prevalent human enteric pathogens and is responsible for the majority of foodborne illnesses in the developed world. The species comprises over 200 different serovars and all are considered pathogenic to humans. In 2008, Salmonella caused 35.4% of all food-borne outbreaks in European Union countries; a far greater proportion than any other causative agent [8]. Salmonellosis is rarely fatal, but considered to be one of the most important enteric infections on the basis of incidence and the resulting morbidity burden. Recent data show that S. enterica serovars are responsible for 38.4% of food-borne outbreaks associated with fresh produce [20]. Notable outbreaks include a multi-state outbreak in the USA from tomatoes [21] and an international outbreak from basil [22]. Out of the many S. enterica serovars, some appear to be more commonly associated with fresh produce, suggesting a difference in their colonization potential. Occasionally, other serovars cause outbreaks from fresh produce, for example S. enterica serovar Paratyphi from spinach and baby leaf salad in northern European countries in 2007 [23] and from alfalfa spouts, in Canada in 2001 [24]. S. enterica serovar Paratyphi is a cause of gastroenteritis and the causative biovar, Java, that has been associated with fresh produce is distinct from ‘classical’ paratyphoidal serotypes in that it is known to have an animal reservoir, which probably accounts for the source of contamination of fresh produce. It is of note that there is an apparent distinction between the food types linked with S. enterica and those linked with EHEC. S. enterica is commonly associated with fruits (tomatoes and melons) and EHEC with leafy salad vegetables (spinach and lettuce), although there is some cross-over between these groups [25]. Some studies have shown differential colonization ability of S. enterica serovars to different crop plants, although the basis to these differences is not yet clear [26, 27]. Campylobacter species persist under microaerophilic conditions and are frequently isolated from poultry, pigs and cattle. Large-scale outbreaks tend to be rare and the majority of cases are sporadic. Occasional outbreaks have been reported from consumption of fresh produce, including cucumber in Australia [28] and peas in the USA [29]. Campylobacter have been isolated from bagged salad on a number of occasions [2, 30]. Fresh produce associated outbreaks from Gram-positive bacteria account for less than 10% of the total number from all food-borne pathogens [20]. Clostridium botulinum is the most prevalent pathogen in this group. Other Gram-positive bacteria include B. cereus, Clostridium perfringens, L. monocytogenes and Staphylococcus aureus. Most Clostridia are considered to be saprophytes and persist under strict anaerobic conditions. C. botulinum causes food poisoning through the production of a

CAB International 2010

Nicola J. Holden

187

Table 1 Summary of hosts and environments that support proliferation of selected enteric pathogens. The pathogens that are not able to use plants as true hosts are still effectively vectored through the food chain by plants Host/environment Enteric pathogen

Plant

Animal

Invertebrate

Soil

Water

EHEC Salmonella enterica Campylobacter species Clostridium botulinum Cryptosporidium species Enteric viruses

3 3 8 ?2 8 8

3 3 3 3 3 3

3 ?1 ?1 ?1 8 8

3 3 3 3 8 8

3 3 3 8 8 8

1

It is unknown whether these pathogens can use invertebrates as a host, but in some cases the pathogens have been isolated from ﬂies. Evidence has yet to be reported.

2

potent toxin and the pathogen is historically linked with preserved foods that have been canned. Clostridia endospores are resistant to mild heat treatment and can survive inadequate processing. The anaerobic environment in canned foods allows the out-growth of germinative cells and toxin production. Recent vegetableassociated outbreaks have included home-canned tomatoes in Canada [31], home-canned mushrooms in the UK [32] and commercially produced carrot juice [33]. There is a good possibility that C. botulinum associates with plants prior to harvest or production but interactions between C. botulinum and growing plants are yet to be reported. Clostridium difﬁcile is normally associated with nosocomial infections and is not considered as a food-borne pathogen. However, a recent study found that it was possible to isolate C. difﬁcile from a small number of bagged salads that had originated in countries within the European Union [34], which suggests a potential for fresh produce as a vehicle of infection. Outbreaks with commercial products, such as the carrot juice outbreak, have resulted in changes in production practices to eliminate endospores.

Protoctists Cryptosporidium and Cyclospora belong to the Apicomlexa and both are obligate intracellular pathogens. They have similar life cycles and form oocysts that are notoriously resistant to standard microbical treatments. Both are considered to be water-borne pathogens, although the cryptosporidia are also zoonotic. As such, contaminated water is the primary source of transmission into fresh produce. The oocysts retain their infectivity for several months and have been associated with outbreaks from fresh produce on several occasions [35]. Once the oocysts are ingested and sporozoites are released, the life cycle can be completed within the intestinal tract of the host. A recent outbreak of Cryptosporidium parvum in Helsinki was thought to have arisen from the consumption of a salad mixture [36]. Cyclospora cayetanensis is a recently recognized parasite that is transmitted through water and contaminated food.

g

Whereas Cryptosporidium species are associated with only sporadic food-borne outbreaks, C. cayetanensis has been the cause of several large-scale outbreaks [37]. Notable outbreaks associated with fresh produce have arisen from the consumption of soft fruit [38, 39].

Virus Enteric viruses are one of the most common causes of infectious disease in man and NoV is one of the most prevalent in the developed world. While the primary mode of transmission is through person-to-person spread, a proportion can also be attributed to the consumption of contaminated food [40]. Outbreaks have been associated with fresh produce on several occasions, in particular salads [9, 41–43]. Hepatitis A virus has also caused outbreaks from the consumption of green onions [44].

Plant–Microbe Interactions There is great diversity in the types of enteric pathogens that can contaminate fresh produce and cause food-borne outbreaks. However, to consider plants as true hosts, a distinction needs to be made between those microbes that demonstrate the capacity for proliferation on or within a plant host and those microbes that are vectored by the plant in a manner that does not support proliferation. The ability of enteric pathogens to proliferate on plants, as well as other hosts or environments that directly link to growing produce crops, has been summarized for some of the key pathogens in Table 1. Protoctists and viruses that have been associated with produce are obligate intracellular pathogens that require a mammalian host to complete their life cycle. Although they can be found on plants, they cannot use plant hosts for proliferation, and are only transiently associated with plants. Virus particles have been found within plant tissue, but this is likely to be driven by plant-based mechanisms. When lettuce plants were irrigated with water containing either enteric canine calicivirus (CaCV) or NoV, a small proportion of the plants contained internalized CaCV, but

CAB International 2010

188

Animal Science Reviews 2010

no NoV [45]. It is thought that the differences in the uptake of virus particles was largely down to differences in virus surface properties, or potentially selective virus internalization or virus clearance by the plant [45]. Hepatitis A virus has also been demonstrated to be internalized by green onions [46]. However, the uptake and internalization of inert ﬂuorescent microspheres via the root system were demonstrated, in this study and elsewhere [46, 47], which suggests that the internalization of virus particles is mediated by the plant. Protoctist contamination of produce is similar to virus contamination, in that the parasites do not use plants as true hosts. Spinach plants that had been irrigated with water containing C. parvum oocysts were found to contain oocysts within the stomata, below the level of the guard cells [48]. Interestingly, there was evidence of excystation as a single sporozoite was observed within the stomata immediately adjacent to an oocyst [48]. The signiﬁcance of this observation needs to be conﬁrmed as spontaneous excystation is known to occur in the absence of host triggers and has been induced in the presence of a polyphenol-rich blueberry extract [49]. The life cycle of colonization of any host requires a sequence of well-established stages: initial adherence to host cells; establishment of a colony (either on the external tissue or following internalization which may be intra- or extracellular); at the same time, the bacterial population needs to adapt to the host response; ﬁnally, the bacteria are disseminated to other hosts. Similarities in these successive stages occur in the colonization of both plant and animal hosts for several bacterial enteric pathogens, especially those within the Enterobacteriaceae. Others, such as Camplyobacter species, interact with plant tissue and can persist on plants for a period of time, but do not appear to be able to proliferate on or within plant hosts [50]. The molecular mechanisms that govern bacterial colonization of plants have been described in detail elsewhere [18] and subsequent studies have increased our knowledge of the area. It is becoming apparent that overand-above the general similarities in colonization differences at the molecular level in both the host and bacterial species deﬁne the nature of speciﬁc interactions. Plants present two distinct macroscopic niches for microbial colonization: the phylloplane is deﬁned by the leaf surface, while the rhizosphere is deﬁned by the environment that immediately surrounds the root system and is directly inﬂuenced by root exudation. Colonization of the phylloplane or phyllosphere (the total aboveground plant surfaces) is less likely than in the rhizosphere, but some recent studies have shown speciﬁc bacterial interactions with leaf tissue. Bacterial ﬂagella are known to mediate bacterial adherence to leaf tissue, for example, adherence of S. enterica to basil leaves [51] and EHEC to spinach leaves is reduced in the respective ﬂagella mutants [52]. In addition, this study conﬁrmed a potential role for the type-three secretion system, ﬁrst demonstrated in adherence of EHEC to rocket leaves

g

[53]. Microscopic analysis shows that the bacterial factors mediate adherence of leaf tissue in what appears to be a very speciﬁc manner. Indeed, differences have been shown for different bacterial isolates of the same pathotypes [51, 53, 54]. These data show that some enteric bacterial colonization factors mediate adherence to plant tissue in a speciﬁc manner. Beyond the initial stages of adherence, some enteric bacteria have been shown to internalize into plant tissue, where they are potentially protected from external sanitation treatments currently used during production. Flagella-mediated internalization has been previously demonstrated for S. enterica on Arabidopsis thaliana [55]. Functional ﬂagella were also required for bacterial migration from infected roots to aerial portions of the plant [55]. An S. enterica ﬂagella mutant was found to be deﬁcient in colonization to lettuce leaf tissue, and signiﬁcantly reduced in its ability to internalize [56]. Furthermore, the same study found that ﬂagella-mediated internalization was linked to chemotaxis in S. enterica. Work to suggest that bacterial factors per se are not required for internalization and uptake by the plant host has shown comparable levels of internalization for ﬂuorescent microspheres as for EHEC in growing lettuce plants [47]. Although the internalization work described above appears to contradict this study, it is most likely that there is some redundancy such that no single bacterial factor mediates adherence or internalization into plant tissue. Indeed, some plant endophytes that have a well-established relationship with plant hosts lack the genes to encode ﬂagella yet are still located on and within plant tissue [57]. The majority of studies show that internalized bacteria are extracellular, within the apoplastic ﬂuid, which is similar to internalization of the plant isolate of Klebsiella pneumoniae [57]. Enteric pathogenic bacteria appear to be able to exploit light-induced opening of stomata to gain access to internal tissue, in a manner similar to plant pathogens. The plant pathogen Pseudomonas syringae was shown to selectively migrate towards open stomata of A. thaliana but not to closed stomata [58]. Open stomata then closed within an hour of infection. The same phenomenon was observed with EHEC infection, although stomata remained closed for more than 8 h, in contrast to just 3 h for P. syringae infection [58]. Stomatal closure in response to plant and enteric pathogens was found to be triggered by pathogen-associated molecular patterns (PAMPs)dependent responses. Thus, it appears that the specialized plant pathogen is able to manipulate the host response in a manner not exhibited by the enteric pathogen [58]. S. enterica was also shown to enter open stomata, on detached lettuce leaf sections [56]. Given the pathogentriggered response for EHEC and P. syringae in A. thaliana, it is intriguing then to note that S. enterica did not induce stomatal closure in lettuce leaf sections. It is possible that the differences have arisen because of speciﬁc bacteria and plant species differences.

CAB International 2010

Nicola J. Holden

189

Figure 1 Routes of transmission of enteric pathogens into or onto fresh produce. Some of the pathways may occur pre-harvest, e.g. from contaminated irrigation water or animal manure, while others can occur post-harvest, e.g. during the production or food preparation. The pathway between animals and plants is bi-directional because of the possibility of contaminated feed or pasture

The phylloplane of different plant species has been shown to induce a stress response in the resident epiphytic bacteria [59]. Physical parameters that contribute to stressful conditions include ultraviolet radiation, temperature shifts and the presence of reactive oxygen species. Furthermore, bacterial transport systems required for the uptake of various nutrient substrates were consistently detected from different phylloplane environments, presumably in response to the limited carbon availability [59]. The plant response to colonization of enteric bacteria has been documented for EHEC and S. enterica infection of A. thaliana [60, 61]. In both cases, the host mounted a PAMP-triggered response, whereas an additional effector-triggered response was observed for infection with P. syringae [60]. One of the most potent PAMPs is a highly conserved domain of ﬂagella (Flg22), which signals the presence of motile bacteria [62] via recognition by the ﬂagellin-sensitive 2 (FLS2) receptor in complex with receptor kinase BRI1 (Brassinosteroid Insensitive 1)-associated receptor kinase 1 (BAK1) [63]. While specialised plant pathogens can subvert FLS2 recognition with effector proteins or modiﬁcation of ﬂagella [64, 65], it is as yet unknown whether motile enteric pathogens can also avoid Flg22 recognition. Work on the plant defence response to enteric pathogens is in its infancy and further investigations are required to reveal the nature of the plant–microbe interactions. Colonization characteristics for enteric pathogens in the root system have been poorly studied in comparison with colonization of leaf tissue. This is most likely down to practical reasons of obtaining ‘clean’, soil-free tissue

g

to perform detailed molecular and microscopic analysis. However, previous studies have demonstrated a capacity of enteric bacteria to colonize the rhizosphere and root system [18] and it is recognized as a reservoir for opportunistic human pathogens [66, 67]. Work on plantassociated Pseudomonas species describes roles for a number of bacterial factors in the colonization of the rhizosphere [68]. Lateral root junctions present potential entry points for enteric bacteria in the root system [66, 69, 70]. These areas have discontinuities in the exodermal Casparian band which may allow bacterial entry into the apoplast. Finally, nutrients within the rhizosphere have been shown to support microbial growth [66, 69, 70].

Sources of Contamination There are multiple routes of transmission which result in the contamination of fresh produce by enteric pathogens, represented diagrammatically in Figure 1, and the pathways are discussed below.

Animal reservoir The majority of infectious agents associated with fresh produce are considered zoonotic and various farm animals represent a primary pathogen reservoir. While cattle are the primary host for pathogenic E. coli [71], cattle, pigs and poultry are reservoirs for many of

CAB International 2010

190

Animal Science Reviews 2010

the salmonellae [72]. Farmyard manure is an important potential source of transmission of enteric pathogens into ﬁelds used to grow fresh produce. However, wild animals and birds also need to be considered as a risk factor in the transmission of zoonoses. For example, the USA spinach outbreak strain of E. coli O157:H7 was identiﬁed in feral swine found in close proximity to the implicated spinach ﬁeld [73]. Guidelines have been put in place in many countries to account for animals as a source of enteric pathogens, such as the inclusion of a substantial gap (24 months) between the application of farm yard manure onto ﬁelds and their subsequent use for growing readyto-eat fresh produce [74].

Water Water-borne transmission is a primary means of dissemination for many infectious agents. In particular, parasitic pathogens are often associated with contaminated water [35]. Bacterial pathogens are also known to be transported by water [75] and some outbreaks have occurred as a direct result of water-borne transport [76]. The presence of pathogenic microbes in irrigation water generally arises through contamination from manure [77–81]. Tracking of E. coli O157 in a region associated with outbreaks from spinach indicated that the bacteria could be transported 16–30 km distances by the water system [82]. River water not only transports bacteria but can also support proliferation, as E. coli O157 was found to grow in sterilized river water over a range of temperatures (15–35 C), and the growth rates were comparable to a river-water isolate of E. coli [83]. Campylobacter jejuni also demonstrated a potential for growth in water [84]. Water-borne transport of enteric pathogens deserves serious consideration, in particular in the context of ﬂooding and changes in watersheds as a consequence of changing climate.

Persistence in soil and the rhizosphere Once enteric pathogens have been (inadvertently) applied to growing fresh produce, they invariably end up on the leaves and/or in the soil. While the interaction of microbial pathogens with leaves has been the subject of several studies, the phylloplane is a relatively harsh environment and less likely to support microbial growth to the same extent as the soil and rhizosphere [85, 86]. Several studies have shown that the rhizosphere and root system support colonization by enteric pathogenic bacteria [55, 66, 67, 87–89]. In general, the levels of bacteria in soil decline over time and are affected by soil-type, temperature and water availability. E. coli O157:H7 was found to persist for up to 36 days in the rhizosphere of lettuce plants grown in sandy soil and, to a lesser extent, in clay soils. The bacteria were applied to the soil 2 weeks prior to planting and

g

detected by a combination of direct plating and the presence of bacterial genes [89]. Bacteria were detected in the lettuce phylloplane 7 days after rhizosphere detection, although always in lower levels than was observed in the rhizosphere. Interestingly, no internalized bacteria were detected, which suggested that bacteria had been transferred to external leaf tissue from contaminated soil, as the seedlings emerged [89]. Likewise, marked differences were found when comparing persistence of C. jejuni in the phylloplane and rhizosphere of different edible plant species [50]. The detection of bacteria from either plant environment was temperaturedependent, with the greatest persistence found at lower temperatures. At 10 C, bacteria were detected for 6 days from infected spinach leaves (9 days after enrichment), whereas bacteria were detectable for at least 28 days from spinach rhizosphere, at 10 C and 16 C [50]. Longterm persistence of S. enterica, of greater than 7 weeks, was demonstrated in soil that had been irrigated with bacteria-spiked water [90]. The study found that the phylloplane contained a consistently lower level of bacteria compared to the rhizosphere, of tomato plants grown in contaminated soil [90]. E. coli O157:H7 was shown to persist for more than 10 weeks in soil that was amended with manure [91]. The site was used to grow carrots and onions and in both cases, bacteria were detected from the internal tissue of the root vegetables [91]. A greater number of bacteria persisted, and for a longer period, on soil used to grow carrots compared to onions. The differences were attributed to the phenolic toxicity of onions compared to carrots. Long-term persistence of E. coli has been found under extreme environmental conditions, tracking E. coli bovine faecal contaminants in Alpine diary grassland soil [92]. As expected, the highest number of bacteria was detected from cowpats, but snow-cover and subsequent snow-melt did not adversely affect the number of bacteria isolated from top-soil, which stayed at a consistent level over the course of the 2-year study [92]. Preliminary sequence analysis showed that the E. coli in the top-soil was in the same phylogenetic group as cowpat isolates, suggesting that E. coli that adapted to persist in the top-soil had originated from bovine manure [92].

Insect and microbe vectors Insects are well-known vectors of some of the most devastating diseases of man, including cholera and malaria. They also have a role in the transmission of food-borne enteric pathogens. Proliferation of E. coli O157:H7 was shown on houseﬂies, which were subsequently able to disseminate bacteria for up to 3 days after feeding [93]. Although the large-scale outbreak of E. coli O157:H7 associated with spinach in the Salinas valley (USA) has been largely attributed to contamination from feral swine [73], a recent study has also implicated ﬂies in the

CAB International 2010

Nicola J. Holden

dissemination of bacteria onto the spinach crop. Insect trapping revealed E. coli O157:H7 in ﬁlth ﬂies and laboratory tests demonstrated transfer of marked bacteria from contaminated ﬂies onto growing spinach plants [94]. C. botulinum was shown to grow within the carcasses of dead honey bees and bee pupae, which potentially presents another route of transmission onto growing, ﬂowering plants [95]. Insects are also known as vectors for other enteric pathogens, for example, houseﬂies captured from different poultry farms were shown to contain either Campylobacter species or S. enterica [96, 97]. In addition to winged insects, enteric bacteria have also been associated with invertebrates normally found within the soil. Caenorhabditis elegans supported the growth of ingested enteric bacteria and transmitted the bacteria to their progeny [98]. Further studies of E. coli O157:H7 ingested by C. elegans suggest that the nematodes can disperse enteric bacteria in agricultural soil [99]. Ciliate protozoae may also act as vectors and transient hosts for enteric bacteria. Viable S. enterica were detected within the vesicles of Tetrahymena following ingestion, and remained viable in secreted vesicles [100]. In contrast, although L. monocytogenes was egested by Tetrahymena, viable bacterial cells were only rarely detected, suggesting a mechanism exists that maintains viable salmonellae [100]. Ciliate protozoa may present yet another mechanism to host and transport some enteric bacteria into crop plants from soil and water.

Post-harvest A further source of contamination arises during harvest and production, either from food-handlers or contaminated equipment [7]. For example, an outbreak of Cryptosporidium hominis in Denmark in 2005 was linked to fresh carrots and peppers and thought to have occurred through cross-contamination from an infected kitchen worker [101]. Contamination from infected food-handlers is also thought to be a route of infection of enteric viruses given the high proportion of person-to-person spread. However, in some cases, the outbreaks reports have concluded that the virus was associated with the produce prior to food preparation [41, 42, 44]. Stringent regulations are in place in many countries to reduce this risk both during harvest and production, and also during food preparation [74, 80].

Conclusions Outbreaks of enteric pathogens from the consumption of fresh produce have increased in recent times. For plants to act as true alternative hosts for enteric pathogens, the microbes must demonstrate the ability to colonize the host, either on external surfaces or internally, where they are able to proliferate. These conditions are not

g

191

met for enteric pathogens that are obligate mammalian intracellular pathogens, such as protoctists parasites and viruses. Nevertheless, the transmission of these pathogens by plants into the food chain is an important food safety issue that cannot be overlooked. Human and animal pathogenic enterobacteria have been demonstrated to interact with plant tissue and proliferate on or within different plant organs. Parallels on the mechanism of plant colonization can be drawn with bacteria normally associated with plants; whether endophytic, epiphytic, beneﬁcial or pathogenic. The potential to colonize non-animal hosts must therefore be considered in the context of agriculture and food safety. Furthermore, this is an area of science that has only recently been appreciated and breaks the traditionally held dogmas that zoonotic pathogens do not persist and proliferate with their animal and human hosts.

Outstanding questions Many outstanding questions on the role of plants as hosts for enteric pathogens remain, which have implications for issues of food safety and colonization potential. Firstly, when considering the mechanisms that underpin enteric colonization of plants, it is necessary to establish a role for microbial factors on living plants rather than cut leaf sections, under conditions representative of agronomy. While some studies have been carried out with growing plants, for example, those that show an internalization role for ﬂagella, others have used leaf tissue cut from samples of bagged salad. Given the interactions that occur between bacteria and plant host are bi-directional, a systems approach is more appropriate. In this manner, further questions can then be addressed on differences in (i) pathogen type; (ii) plant species and varieties; (iii) plant organs and tissue types, e.g. leaves and roots and (iv) environmental and physical parameters. However, experimentation with human pathogens imposes quite severe restrictions, since by their very nature the pathogens are subject to legislation under health and safety laws. Although containment in a laboratory and glasshouse setting is feasible with highly pathogenic organisms, previous studies in an agricultural setting have all been carried out with bacteria modiﬁed to reduce their pathogenicity and thus, their hazard category. However, subsequent ﬁeld work still carries the burden of either decontamination of any site where bacteria have been artiﬁcially added, or continued contamination of the site. Secondly, the ability of some bacteria to internalize into plant tissue requires further studies in an agricultural context. The current FAO/WHO microbiological risk assessment of the hazards from fresh produce recognizes internalization, but does not see it as a signiﬁcant problem in practice [102]. While the small number of ﬁeld studies that have been carried out show very little or no internalization, others have shown (mostly on seedlings under

CAB International 2010

192

Animal Science Reviews 2010

laboratory conditions) that a proportion of the bacterial population can internalize into plant tissue. The signals (bacterial or plant) and mechanisms that underpin internalization is of great interest, both to the scientiﬁc community and for food safety. A third issue concerns variation in microbial pathotype, and whether some bacteria that have greater potential to colonize plants are as pathogenic to humans/animals as those with a lower potential. Gene carriage and allele variation are known to affect disease outcome in some animal infections, e.g. carriage of stx variants and other genetic markers in EHEC [103, 104]. Comparative genomic analysis has already highlighted how the acquisition of horizontally acquired DNA islands contributes to host specialization of plant pathogenic bacteria that are very closely related to human pathogenic enterobacteria [105]. Exploitation of low-cost and high-throughput sequencing techniques is, and will be, instrumental for further genomic comparison. In turn, this information will show whether there is a genetic basis to a life style more adapted to one host species over another, which will be an important step in the prevention of enteric infections from fresh produce.

Acknowledgements N.H. is supported by Scottish Government Rural and Environmental Research and Analysis Directorate. Thanks go to Ian Toth (SCRI) and Adrian Newton (SCRI) for critical reading of the manuscript.

References 1. Tauxe RV, Doyle MP, Kuchenmu¨ller T, Schlundt J, Stein CE. Evolving public health approaches to the global challenge of foodborne infections. International Journal of Food Microbiology 2010;In press. 2. Little CL, Gillespie IA. Prepared salads and public health. Journal of Applied Microbiology 2008;105:1729–43. 3. Beuchat LR. Pathogenic microorganisms associated with fresh produce. Journal of Food Protection 1996;59:204–16. 4. Sivapalasingam S, Friedman CR, Cohen L, Tauxe RV. Fresh produce: a growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997. Journal of Food Protection 2004;67:2342–53. 5. Brandl MT. Fitness of human enteric pathogens on plants and implications for food safety. Annual Review of Phytopathology 2006;44:367–92. 6. Warriner K, Huber A, Namvar A, Fan W, Dunﬁeld K. Recent advances in the microbial safety of fresh fruits and vegetables. Advances in Food and Nutritional Research 2009;57:155–208. 7. Allerberger F, Sessitsch A. Incidence and microbiology of salad-borne disease. CAB Reviews: Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources 2009;4:1–13.

g

8. European Food Safety Authority. The community summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in the European Union in 2008. EFSA Journal 2010;1496:288. 9. Long SM, Adak GK, O’Brien SJ, Gillespie IA. General outbreaks of infectious intestinal disease linked with salad vegetables and fruit, England and Wales, 1992–2000. Communicable Disease and Public Health 2002;5:101–5. 10. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, Shapiro C, et al. Food-related illness and death in the United States. Emerging Infectious Diseases 1999;5:607–25. 11. Merson MH, Morris GK, Sack DA, Wells JG, Feeley JC, Sack RB, et al. Travelers’ diarrhea in Mexico. A prospective study of physicians and family members attending a congress. New England Journal of Medicine 1976;294:1299–305. 12. Centers for Disease Control and Prevention. Ongoing multistate outbreak of Escherichia coli serotype O157:H7 infections associated with consumption of fresh spinach-United States, September 2006. Morbidity and Mortality Weekly Report 2006;55:1045–6. 13. Watanabe Y, Ozasa K, Mermin JH, Grifﬁn PM, Masuda K, Imashuku S, et al. Factory outbreak of Escherichia coli O157:H7 infection in Japan. Emerging Infectious Diseases 1999;5:424–8. 14. Friesema I, Sigmundsdottir G, van der Zwaluw K, Heuvelink A, Schimmer B, de Jager C, et al. An international outbreak of Shiga toxin-producing Escherichia coli O157 infection due to lettuce, September–October 2007. Euro Surveillance 2008;13:19065. 15. Kulasekara BR, Jacobs M, Zhou Y, Wu Z, Sims E, Saenphimmachak C, et al. Analysis of the genome of the Escherichia coli O157:H7 2006 spinach-associated outbreak isolate indicates candidate genes that may enhance virulence. Infection and Immunity 2009;77:3713–21. 16. Riley LW, Remis RS, Helgerson SD, McGee HB, Wells JG, Davis BR, et al. Hemorrhagic colitis associated with a rare Escherichia coli serotype. New England Journal of Medicine 1983;308:681–5. 17. Fouts DE, Tyler HL, DeBoy RT, Daugherty S, Ren Q, Badger JH, et al. Complete genome sequence of the N2-ﬁxing broad host range endophyte Klebsiella pneumoniae 342 and virulence predictions veriﬁed in mice. PLoS Genetics 2008;4:e1000141. 18. Holden N, Pritchard L, Toth I. Colonization outwith the colon: plants as an alternative environmental reservoir for human pathogenic enterobacteria. FEMS Microbiology Reviews 2009;33:689–703. 19. Rangel JM, Sparling PH, Crowe C, Grifﬁn PM, Swerdlow DL. Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982–2002. Emerging Infectious Diseases 2005;11:603–9. 20. Greig JD, Ravel A. Analysis of foodborne outbreak data reported internationally for source attribution. International Journal of Food Microbiology 2009;130:77–87. 21. Bidol SA, Daly E, Rickert R, Hill TA, Al Khaldi S, Taylor TH, et al. Multistate outbreaks of Salmonella infections associated with raw tomatoes eaten in restaurants – United States, 2005–2006. Journal of the American Medical Association 2007;298:1753–5.

CAB International 2010

Nicola J. Holden 22. Pezzoli L, Elson R, Little C, Fisher I, Yip H, Peters T, et al. International outbreak of Salmonella Senftenberg in 2007. Euro Surveillance 2007;12:E070614.3. 23. Denny J, Threlfall J, Takkinen J, Lo¨fdahl S, Westrell T, Varela C, et al. Multinational Salmonella paratyphi B variant Java (Salmonella Java) outbreak, August–December 2007. Euro Surveillance 2007;12:3332. 24. Stratton J, Stefaniw L, Grimsrud K, Werker DH, Ellis A, Ashton E, et al. Outbreak of Salmonella paratyphi B var java due to contaminated alfalfa sprouts in Alberta, British Columbia and Saskatchewan. Canada Communicable Disease Report 2001;27:133–7.

193

39. Herwaldt BL, Ackers ML. An outbreak in 1996 of cyclosporiasis associated with imported raspberries. The Cyclospora Working Group. New England Journal of Medicine 1997;336:1548–56. 40. Lopman BA, Adak GK, Reacher MH, Brown DW. Two epidemiologic patterns of norovirus outbreaks: surveillance in England and Wales, 1992–2000. Emerging Infectious Diseases 2003;9:71–7. 41. Makary P, Maunula L, Niskanen T, Kuusi M, Virtanen M, Pajunen S, et al. Multiple norovirus outbreaks among workplace canteen users in Finland, July 2006. Epidemiology and Infection 2009;137:402–7.

25. Warriner K, Namvar A. The tricks learnt by human enteric pathogens from phytopathogens to persist within the plant environment. Current Opinion in Biotechnology 2010; In Press, Corrected Proof

42. Mesquita JR Nascimento MS. A foodborne outbreak of norovirus gastroenteritis associated with a Christmas dinner in Porto, Portugal, December 2008. Euro Surveillance 2009;14:19355.

26. Shi X, Namvar A, Kostrzynska M, Hora R, Warriner K. Persistence and growth of different Salmonella serovars on pre- and postharvest tomatoes. Journal of Food Protection 2007;70:2725–31.

43. Ethelberg S, Lisby M, Bo¨ttiger B, Schultz AC, Villif A, Jensen T, et al. Outbreaks of gastroenteritis linked to lettuce, Denmark, January 2010. Euro Surveillance 2010;15:19484.

27. Barak JD, Liang A, Narm K-E. Differential attachment to and subsequent contamination of agricultural crops by Salmonella enterica. Applied and Environmental Microbiology 2008;74:5568–70. 28. Kirk M, Waddell R, Dalton C, Creaser A, Rose N. A prolonged outbreak of Campylobacter infection at a training facility. Communicable Diseases Intelligence 1997;21:57–61.

44. Dentinger CM, Bower WA, Nainan OV, Cotter SM, Myers G, Dubusky LM, et al. An outbreak of hepatitis A associated with green onions. Journal of Infectious Diseases 2001;183:1273–6. 45. Urbanucci A, Myrmel M, Berg I, von Bonsdorff CH, Maunula L. Potential internalisation of caliciviruses in lettuce. International Journal of Food Microbiology 2009;135:175–8.

29. ProMED-mail. Campylobacteriosis, raw peas – USA: (Alaska), waterfowl source. ProMED-mail 2008;20081010:3216.

46. Chancellor DD, Tyagi S, Bazaco MC, Bacvinskas S, Chancellor MB, Dato VM, et al. Green onions: potential mechanism for hepatitis A contamination. Journal of Food Protection 2006;69:1468–72.

30. Evans MR, Ribeiro CD, Salmon RL. Hazards of healthy living: bottled water and salad vegetables as risk factors for Campylobacter infection. Emerging Infectious Diseases 2003;9:1219–25.

47. Solomon EB, Matthews KR. Use of ﬂuorescent microspheres as a tool to investigate bacterial interactions with growing plants. Journal of Food Protection 2005;68:870–3.

31. Loutfy MR, Austin JW, Blanchﬁeld B, Fong IW. An outbreak of foodborne botulism in Ontario. Canadian Journal of Infectious Diseases 2003;14:206–9.

48. Macarisin D, Bauchan G, Fayer R. Spinacia oleracea L. leaf stomata harboring Cryptosporidium parvum oocysts: a potential threat to food safety. Applied and Environmental Microbiology 2010;76:555–9.

32. Health Protection Agency. Botulism associated with home-preserved mushrooms. Communicable Disease Report Weekly 1998;8:159–62. 33. Sheth AN, Wiersma P, Atrubin D, Dubey V, Zink D, Skinner G, et al. International outbreak of severe botulism with prolonged toxemia caused by commercial carrot juice. Clinical Infectious Diseases 2008;47:1245–51. 34. Bakri MM, Brown DJ, Butcher JP, Sutherland AD. Clostridium difﬁcile in ready-to-eat salads, Scotland. Emerging Infectious Diseases 2009;15:817–8.

49. Anthony JP, Fyfe L, Stewart D, McDougall GJ, Smith HV. The effect of blueberry extracts on Giardia duodenalis viability and spontaneous excystation of Cryptosporidium parvum oocysts, in vitro. Methods 2007;42:339–48. 50. Brandl MT, Haxo AF, Bates AH, Mandrell RE. Comparison of survival of Campylobacter jejuni in the phyllosphere with that in the rhizosphere of spinach and radish plants. Applied and Environmental Microbiology 2004;70:1182–9.

35. Dawson D. Foodborne protozoan parasites. International Journal of Food Microbiology 2005;103:207–27.

51. Berger CN, Shaw RK, Brown DJ, Mather H, Clare S, Dougan G, et al. Interaction of Salmonella enterica with basil and other salad leaves. Journal of the International Society for Microbial Ecology 2008;3:261–5.

36. Ponka A, Kotilainen P, Rimhanen-Finne R, Hokkanen P, Hanninen ML, Kaarna A, et al. A foodborne outbreak due to Cryptosporidium parvum in Helsinki, November 2008. Euro Surveillance 2009;14:19269.

52. Xicohtencatl-Cortes J, Sanchez Chacon E, Saldana Z, Freer E, Giron JA. Interaction of Escherichia coli O157:H7 with leafy green produce. Journal of Food Protection 2009;72:1531–7.

37. Herwaldt BL. Cyclospora cayetanensis: a review, focusing on the outbreaks of cyclosporiasis in the 1990s. Clinical Infectious Diseases 2000;31:1040–57.

53. Shaw RK, Berger CN, Feys B, Knutton S, Pallen MJ, Frankel G. Enterohemorrhagic Escherichia coli exploits EspA ﬁlaments for attachment to salad leaves. Applied and Environmental Microbiology 2008;74:2908–14.

38. Fleming CA, Caron D, Gunn JE, Barry MA. A foodborne outbreak of Cyclospora cayetanensis at a wedding: clinical features and risk factors for illness. Archives of Internal Medicine 1998;158:1121–5.

g

54. Kroupitski Y, Pinto R, Brandl MT, Belausov E, Sela S. Interactions of Salmonella enterica with lettuce leaves. Journal of Applied Microbiology 2009;106:1876–85.

CAB International 2010

194

Animal Science Reviews 2010

55. Cooley MB, Miller WG, Mandrell RE. Colonization of Arabidopsis thaliana with Salmonella enterica and enterohemorrhagic Escherichia coli O157:H7 and competition by Enterobacter asburiae. Applied and Environmental Microbiology 2003;69:4915–26. 56. Kroupitski Y, Golberg D, Belausov E, Pinto R, Swartzberg D, Granot D, et al. Internalization of Salmonella enterica in leaves is induced by light and involves chemotaxis and penetration through open stomata. Applied and Environmental Microbiology 2009;75:6076–6086. 57. Dong Y, Iniguez AL, Triplett EW. Quantitative assessments of the host range and strain speciﬁcity of endophytic colonization by Klebsiella pneumoniae 342. Plant and Soil 2003;257:49–59. 58. Melotto M, Underwood W, Koczan J, Nomura K, He SY. Plant stomata function in innate immunity against bacterial invasion. Cell 2006;126:969–80. 59. Delmotte Nl, Knief C, Chaffron S, Innerebner G, Roschitzki B, Schlapbach R, et al. Community proteogenomics reveals insights into the physiology of phyllosphere bacteria. Proceedings of the National Academy of Sciences, USA 2009;106:16428–33. 60. Thilmony R, Underwood W, He SY. Genome-wide transcriptional analysis of the Arabidopsis thaliana interaction with the plant pathogen Pseudomonas syringae pv. tomato DC3000 and the human pathogen Escherichia coli O157:H7. Plant Journal 2006;46:34–53. 61. Schikora A, Carreri A, Charpentier E, Hirt H. The dark side of the salad: Salmonella typhimurium overcomes the innate immune response of Arabidopsis thaliana and shows an endopathogenic lifestyle. PLoS ONE 2008;3:e2279. 62. Felix G, Duran JD, Volko S, Boller T. Plants have a sensitive perception system for the most conserved domain of bacterial ﬂagellin. Plant Journal 1999;18:265–76. 63. Chinchilla D, Zipfel C, Robatzek S, Kemmerling B, Nurnberger T, Jones JD, et al. A ﬂagellin-induced complex of the receptor FLS2 and BAK1 initiates plant defence. Nature 2007;448:497–500. 64. Xiang TT, Zong N, Zou Y, Wu Y, Zhang J, Xing WM, et al. Pseudomonas syringae effector AvrPto blocks innate immunity by targeting receptor kinases. Current Biology 2008;18:74–80. 65. Taguchi F, Suzuki T, Takeuchi K, Inagaki Y, Toyoda K, Shiraishi T, et al. Glycosylation of ﬂagellin from Pseudomonas syringae pv. tabaci 6605 contributes to evasion of host tobacco plant surveillance system. Physiological and Molecular Plant Pathology 2009;74:11–17. 66. Dong Y, Iniguez AL, Ahmer BM, Triplett EW. Kinetics and strain speciﬁcity of rhizosphere and endophytic colonization by enteric bacteria on seedlings of Medicago sativa and Medicago truncatula. Applied and Environmental Microbiology 2003;69:1783–90. 67. Berg G, Eberl L, Hartmann A. The rhizosphere as a reservoir for opportunistic human pathogenic bacteria. Environmental Microbiology 2005;7:1673–85. 68. Lugtenberg BJ, Dekkers L, Bloemberg GV. Molecular determinants of rhizosphere colonization by Pseudomonas. Annual Review of Phytopathology 2001;39:461–90. 69. Bais HP, Weir TL, Perry LG, Gilroy S, Vivanco JM. The role of root exudates in rhizosphere interations with plants and other organisms. Annual Review of Plant Biology 2006;57:233–66.

g

70. Badri DV, Weir TL, van der Lelie D, Vivanco JM. Rhizosphere chemical dialogues: plant–microbe interactions. Current Opinion in Biotechnology 2009;20:642–50. 71. Spears KJ, Roe AJ, Gally DL. A comparison of enteropathogenic and enterohaemorrhagic Escherichia coli pathogenesis. FEMS Microbiology Letters 2006;255:187–202. 72. Stevens MP, Humphrey TJ, Maskell DJ. Molecular insights into farm animal and zoonotic Salmonella infections. Philosophical Transactions of the Royal Society of London. Series B, Biological sciences 2009;364:2709–23. 73. Jay MT, Cooley MB, Carychao D, Wiscomb GW, Sweitzer RA, Crawford-Miksza L, et al. Escherichia coli O157:H7 in feral swine near spinach ﬁelds and cattle, central California coast. Emerging Infectious Diseases 2007;13:1908–11. 74. Chilled Foods Association Ltd. Microbiological Guidance for Produce Suppliers to Chilled Food Manufacturers. Chilled Foods Association Ltd, Kettering, UK; 2007. 75. Ferguson C, Husman AMD, Altavilla N, Deere D, Ashbolt N. Fate and transport of surface water pathogens in watersheds. Critical Reviews in Environmental Science and Technology 2003;33:299–361. 76. Auld H, MacIver D, Klaassen J. Heavy rainfall and waterborne disease outbreaks: the Walkerton example. Journal of Toxicology and Environmental Health. Part A 2004;67:1879–87. 77. Solomon EB, Yaron S, Matthews KR. Transmission of Escherichia coli O157 : H7 from contaminated manure and irrigation water to lettuce plant tissue and its subsequent internalization. Applied and Environmental Microbiology 2002;68:397–400. 78. Ibekwe AM, Watt PM, Shouse PJ, Grieve CM. Fate of Escherichia coli O157:H7 in irrigation water on soils and plants as validated by culture method and real-time PCR. Canadian Journal of Microbiology 2004;50:1007–14. 79. Cheong S, Lee C, Song SW, Choi WC, Lee CH, Kim S-J. Enteric viruses in raw vegetables and groundwater used for irrigation in South Korea. Applied and Environmental Microbiology 2009;75:7745–51. 80. Monaghan JM, Thomas DJI, Goodburn K, Hutchison ML. A Review of the Published Literature Describing Foodborne Illness Outbreaks Associated with Ready to Eat Fresh Produce and an Overview of Current UK Fresh Produce Farming Practices. Food Standards Agency Project B17007. Editor. Food Standards Agency, Newport, UK; 2009. 81. Greene SK, Daly ER, Talbot EA, Demma LJ, Holzbauer S, Patel NJ, et al. Recurrent multistate outbreak of Salmonella Newport associated with tomatoes from contaminated ﬁelds, 2005. Epidemiology and Infection 2008;136:157–65. 82. Cooley M, Carychao D, Crawford-Miksza L, Jay MT, Myers C, Rose C, et al. Incidence and tracking of Escherichia coli O157:H7 in a major produce production region in California. PLoS ONE 2007;2:e1159. 83. Vital M, Hammes F, Egli T. Escherichia coli O157 can grow in natural freshwater at low carbon concentrations. Environmental Microbiology 2008;10:2387–96. 84. Tatchou-Nyamsi-Konig JA, Moreau A, Federighi M, Block JC. Behaviour of Campylobacter jejuni in experimentally contaminated bottled natural mineral water. Journal of Applied Microbiology 2007;103:280–8.

CAB International 2010

Nicola J. Holden 85. Hutchison ML, Avery SM, Monaghan JM. The air-borne distribution of zoonotic agents from livestock waste spreading and microbiological risk to fresh produce from contaminated irrigation sources. Journal of Applied Microbiology 2008;105:848–57. 86. Loper JE, Suslow TV, Schroth MN. Lognormal distribution of bacterial populations in the rhizosphere. Phytopathology 1984;74:1454–60. 87. Jablasone J, Warriner K, Grifﬁths M. Interactions of Escherichia coli O157:H7, Salmonella typhimurium and Listeria monocytogenes plants cultivated in a gnotobiotic system. International Journal of Food Microbiology 2005;99:7–18. 88. Kutter S, Hartmann A, Schmid M. Colonization of barley (Hordeum vulgare) with Salmonella enterica and Listeria spp. FEMS Microbiology Ecology 2006;56:262–71. 89. Ibekwe AM, Grieve CM, Papiernik SK, Yang CH. Persistence of Escherichia coli O157:H7 on the rhizosphere and phyllosphere of lettuce. Letters in Applied Microbiology 2009;49:784–90.

195

96. Hald B, Skovgard H, Pedersen K, Bunkenborg H. Inﬂuxed insects as vectors for Campylobacter jejuni and Campylobacter coli in Danish broiler houses. Poultry Science 2008;87:1428–34. 97. Holt PS, Geden CJ, Moore RW, Gast RK. Isolation of Salmonella enterica serovar Enteritidis from houseﬂies (Musca domestica) found in rooms containing Salmonella serovar Enteritidis-challenged hens. Applied and Environmental Microbiology 2007;73:6030–5. 98. Kenney SJ, Anderson GL, Williams PL, Millner PD, Beuchat LR. Persistence of Escherichia coli O157:H7, Salmonella Newport, and Salmonella Poona in the gut of a free-living nematode, Caenorhabditis elegans, and transmission to progeny and uninfected nematodes. International Journal of Food Microbiology 2005;101:227–36. 99. Anderson GL, Kenney SJ, Millner PD, Beuchat LR, Williams PL. Shedding of foodborne pathogens by Caenorhabditis elegans in compost-amended and unamended soil. Food Microbiology 2006;23:146–53.

90. Barak JD, Liang AS. Role of soil, crop debris, and a plant pathogen in Salmonella enterica contamination of tomato plants. PLoS ONE 2008;3:e1657.

100. Brandl MT, Rosenthal BM, Haxo AF, Berk SG. Enhanced survival of Salmonella enterica in vesicles released by a soilborne Tetrahymena species. Applied and Environmental Microbiology 2005;71:1562–9.

91. Islam M, Morgan J, Doyle MP, Jiang X. Fate of Escherichia coli O157:H7 in manure compost-amended soil and on carrots and onions grown in an environmentally controlled growth chamber. Journal of Food Protection 2004;67:574–8.

101. Ethelberg S, Lisby M, Vestergaard LS, Enemark HL, Olsen KE, Stensvold CR, et al. A foodborne outbreak of Cryptosporidium hominis infection. Epidemiology and Infection 2009;137:348–56.

92. Texier S, Prigent-Combaret C, Gourdon MH, Poirier MA, Faivre P, Dorioz JM, et al. Persistence of culturable Escherichia coli fecal contaminants in dairy Alpine grassland soils. Journal of Environmental Quality 2008;37:2299–310. 93. Kobayashi M, Sasaki T, Saito N, Tamura K, Suzuki K, Watanabe H, et al. Houseﬂies: not simple mechanical vectors of enterohemorrhagic Escherichia coli O157:H7. American Journal of Tropical Medicine and Hygiene 1999;61:625–9. 94. Talley JL, Wayadande AC, Wasala LP, Gerry AC, Fletcher J, DeSilva U, et al. Association of Escherichia coli O157:H7 with ﬁlth ﬂies (Muscidae and Calliphoridae) captured in leafy greens ﬁelds and experimental transmission of E. coli O157:H7 to spinach leaves by house ﬂies (Diptera: Muscidae). Journal of Food Protection 2009;72:1547–52. 95. Nakano H, Kizaki H, Sakaguchi G. Multiplication of Clostridium botulinum in dead honey-bees and bee pupae, a likely source of heavy contamination of honey. International Journal of Food Microbiology 1994;21:247–52.

g

102. FAO/WHO. Microbiology Hazards in Fresh Leafy Vegetables and Herbs, in Meeting Report. Microbiological Risk Assessment Series No. 14. Food and Agriculture Organisation of the United Nations, World Health Organisation, Rome; 2008. p. 1–151. 103. Wu G, Carter B, Mafura M, Liebana E, Woodward MJ, Anjum MF. Genetic diversity among Escherichia coli O157:H7 isolates and identiﬁcation of genes linked to human infections. Infection and Immunity 2008;76:845–56. 104. Whitworth J, Zhang Y, Bono J, Pleydell E, French N, Besser T. Diverse genetic markers concordantly identify bovine origin Escherichia coli O157 genotypes underrepresented in human disease. Applied and Environmental Microbiology 2010;76:361–5. 105. Toth IK, Pritchard L, Birch PR. Comparative genomics reveals what makes an enterobacterial plant pathogen. Annual Review of Phytopathology 2006;44:305–36.

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Use of high-density marker genotyping for genetic improvement of livestock by genomic selection Jack C.M. Dekkers Address: Department of Animal Science and Center for Integrated Animal Genomics, 239D Kildee Hall, Iowa State University, Ames, IA 50011, USA. Correspondence: Email: [email protected] 9 December 2009 30 April 2010

Received: Accepted:

Abstract Genomic or whole-genome selection represents a technological development in the use of a large number of genetic markers for genetic improvement, which is expected to lead to a paradigm shift in the design and implementation of livestock breeding programmes. Genomic selection (GS) is based on the simultaneous use of large numbers (>50 000) of genetic markers for selection of individuals without phenotype and has been made possible by recent developments in genomics and in statistical methods for analysis and prediction. The objectives of this review are to describe the principles of whole-genome selection analyses, opportunities that are provided by GS to enhance and redesign breeding programmes, summarize results of recent implementations of GS analyses in dairy cattle and describe issues associated with current implementations of GS analyses. Finally, future avenues of research that can be pursued to address these issues and develop statistical methods and breeding programme designs that capitalize on the opportunities provided by GS are described. Keywords: Animal breeding, Genetic selection, Molecular genetics, Marker-assisted selection

Introduction The most important traits in animal agriculture are the so-called ‘quantitative’ or ‘complex’ traits that are determined by multiple to many genes, along with environmental factors [1]. Traditional approaches to genetic improvement of animal breeding populations have focused on using observable phenotypes to derive estimated breeding values (EBV) for use in selection of animals to produce the next generation. Sophisticated statistical methods based on mixed linear methodology and best linear unbiased prediction (BLUP) [2, 3] have been implemented in most livestock species to utilize phenotypes on both the animal itself and its relatives to maximize the accuracy of the resulting EBV, in order to increase response to selection. Here, accuracy is deﬁned as the correlation between true and EBV [1]. Although these phenotype-based selection strategies have resulted in tremendous rates of genetic gain in many livestock species for multiple traits, the limitations of this

g

phenotype-based approach to selection are well recognized [4]. These limitations primarily stem from the need to obtain phenotypes on the individual itself and/or its relatives to enable accurate prediction of an individual’s EBV, which is problematic for traits that are sex limited (e.g. milk production, reproduction), that can only be obtained late in life (e.g. longevity and reproductive performance), that require sacriﬁcing the animal (e.g. meat quality) or that are expensive to record (e.g. disease resistance). Because DNA can be collected at any age and on all animals, several of these limitations can be removed by marker- or gene-assisted selection [4]. Gene-assisted selection involves genotyping selection candidates for genes that are known to affect traits of interest, whereas markers-assisted selection (MAS) involves genotyping selection candidates for genetic markers that may themselves not have a direct effect on the trait but that have been shown to have a statistical association with the trait because of their proximity in the genome to genes

CAB International 2010

198

Animal Science Reviews 2010

that affect the trait: the so-called quantitative trait loci (QTLs). The use of gene- or marker-assisted selection requires knowledge of genes or markers that are associated with traits of interest and quantitative estimates of these associations in the population of interest. With the advent of molecular genetics in the 1970s, many so-called QTL mapping studies were conducted in livestock, with the aim of ﬁnding such genes or markers and estimating their effects on traits (e.g. [5–7]). Although these studies have resulted in the identiﬁcation of many QTLs, their application in commercial breeding programmes has been limited for the following reasons [8–10]: (i) effects identiﬁed explained only a limited amount of genetic variation for the trait; (ii) a limited number of genetic markers were available and the cost of genotyping was relatively high; (iii) several results were false positives or were discovered in experimental populations and did not replicate in populations of interest; (iv) marker or QTL effects were estimated on a within-family basis, which made it more difﬁcult to incorporate them into breeding programmes. These limitations of previous applications of QTL mapping results for MAS have, however, been lifted by recent developments in technology. These include genome sequencing, the identiﬁcation of large numbers of genetic markers across the genome in the form of single nucleotide polymorphisms (SNPs), and cost-effective highthroughput genotyping of tens of thousands of such SNPs on individual animals (e.g. [11, 12]). Combined with further development of statistical methods for analysis of molecular data, this is leading to substantial changes in how genetic markers are used for prediction of breeding values in the form of what has been termed ‘genomic selection’ or whole-genome selection [13]. Genomic selection (GS) is an enhanced version of MAS that involves selection of animals for breeding on the basis of their genotype for tens of thousands of ‘random’ SNPs that cover the genome. In GS, the association of each SNP with phenotype is estimated using sophisticated statistical and quantitative genetics models without pre-screening markers based on signiﬁcance. This is in contrast to ‘traditional’ MAS, which involves a two-step approach, with screening of markers based on signiﬁcance of their association with phenotype as the ﬁrst step, followed by the use of only the signiﬁcant markers for selection [14]. Meuwissen et al. [13] showed by simulation that GS could result in accuracies of EBV of young individuals without phenotype of up to 80%. The ﬁrst widespread application of GS to animal breeding practice was in early 2009, through its implementation in national genetic evaluations for dairy cattle in the USA. This implementation was based on the analysis of statistical associations of phenotypes from over 1 million daughter records from over 4000 sires with the genotype of each of these sires for 40 000 SNPs across the genome [15]. Additional applications of GS in dairy cattle and other livestock

g

species are under way [16, 17] and similar methods are also being considered in plant breeding [18]. Against this background, the purpose of this review is to outline the basic principles of GS, the opportunities GS provides for genetic improvement of livestock and to discuss the challenges of current applications of GS. Although GS also has opportunities to improve mating strategies that capitalize on non-additive genetic effects (dominance), this review will primarily focus on the use of GS for genetic improvement through additive genetics. This speciﬁc focus is in part motivated by the fact that selection on additive genetic effects is the most effective strategy for genetic improvement in livestock and by the arguments of Hill et al. [19] that most genetic variation for quantitative traits is expected to be additive in nature.

Principle of Genomic Selection GS involves the following steps [13]: (i) genotype individuals that have been phenotyped for the trait or that have progeny with phenotypes for a large number of SNPs across the genome, (ii) use the resulting data to estimate the statistical association of each SNP with phenotype for the trait and (iii) use the resulting estimates to predict the breeding values of selection candidates in a new generation or in another population. This process is illustrated in Figure 1. In general, the prediction model for GS is developed by ﬁtting the following linear model to phenotypes of individuals that make up the training population [13]: X yi =m+ j Xij bj +ei , where yi is the phenotype (or progeny mean phenotype) of individual i in the training data, m represents ﬁxed effects, summation Sj is over all genotyped SNPs, Xij is the number (0, 1 or 2) of copies of allele ‘1’ (versus ‘0’) that individual i carries at SNP j, bj is the allele substitution effect [1] for SNP j, and ei is a random residual. One statistical challenge to ﬁt this model is that the number of markers (>30 000 is common) typically is much greater than the number of animals with phenotypic records that are available for estimating their effects (typically less than 2000). Although multiple methods are available to deal with this [13, 20, 21], the most common method is to ﬁt the effect of each SNP as a random effect and then implement it via Bayesian methods using certain prior distributions for the variance of SNP effects [13, 20]. An important distinction is between models that use prior distributions that assume genetic variance is equally distributed across all genotyped SNPs, the so-called genomic or G-BLUP method of Meuwissen et al. [13], and methods that assume that a large proportion of SNPs have zero or very small effects, such as the Bayes-B method of Meuwissen et al. [13] and other versions and implementations of these methods [20, 22]. Non- and semi-parametric

CAB International 2010

Jack C.M. Dekkers

Genotype for highhigh-density SNP panel

Progeny

Training data

Phenotype

Genotype for highhigh-density SNP panel

199

Develop genomic prediction model

Predict BV from marker genotypes

Progeny

Genotype

Predict BV

for highhigh-density SNP panel

from marker genotypes

Figure 1 Illustration of the concept of genomic selection

methods [23–26], dimension reduction methods [27, 28], and other Bayesian regression methods [29] have also been used to develop prediction equations based on the large number of SNPs that are genotyped with GS. Because the accuracy of the ‘genomic’ EBV (G-EBV) that are derived using GS models is difﬁcult to predict, a validation step is usually included. This involves separating the data set into training and validation data sets, developing the GS prediction model on the training data set, applying the prediction model to compute G-EBV for individuals in the validation data set and correlating the resulting G-EBV with the phenotypes of the latter individuals. Whereas initial GS analyses separated the data randomly into training and validation sets (e.g. [27]), it is now well understood that this can result in severe upward biases in correlations that are obtained because of genetic relationships between individuals in the training and validation data [21]. To properly assess the beneﬁt of G-EBV and the impact of pedigree relationships between individuals in the validation and training data sets, in addition to computing G-EBV for individuals in the validation data based on phenotypes in the training data, regular BLUP EBV should also be computed for validation individuals based on these same training data phenotypes. The G-EBV and phenotype-based EBV can then be compared based on their correlations with phenotype of individuals in the validation data. To assess the value of G-EBV in practice, construction of the validation data sets should be driven by how the G-EBV will be used in practice. Thus, if G-EBV will be used to select the next generation of individuals within a breed, the training and validation data sets should be separated by year of birth of individuals, with the older animals used for training and the younger animals for validation, as has been done in more recent GS analyses [15].

g

Genomic Versus Marker-assisted Selection In principle, GS represents an extension of traditional strategies for MAS [14] but with some important differences: (i) MAS is based on the premise of using a limited number of markers that have been shown to have large effects on the trait(s) of interest in prior QTL detection or marker association analyses. In contrast, GS uses all SNPs for prediction, without prior screening of associations for signiﬁcance, although some GS analysis methods do implement statistical methods that aim to reduce the number of markers used for prediction (see later). Combined with ﬁtting marker effects as random factors, which regresses estimates toward zero depending on the amount of information available in the data to estimate the effect and the strength of prior information, this reduces the upward biases that have been observed in studies that only focus on signiﬁcant associations, i.e. the so-called Beavis effect [30]. (ii) GS ﬁts all SNPs in the model simultaneously, rather than one at a time or by genomic region, as is done in typical QTL detection studies. Although the main focus of GS analyses is on prediction of G-EBV, it should be noted that this same strategy of ﬁtting all markers simultaneously as random effects can also be used for QTL detection and leads to improved power, as demonstrated by Xu [31]. (iii) Current strategies for GS focus on utilizing associations that exist between markers and QTLs across the population, based on population-wide linkage disequilibrium (LD) between markers and QTLs. In contrast, apart from candidate gene studies [32], most QTL detection studies have utilized co-segregation of markers linked to QTLs within breed crosses (F2, backcross) or within families [8]. Here, co-segregation refers to the joint transmission from parents to progeny of alleles at loci that are close

CAB International 2010

200

Animal Science Reviews 2010

together in the genome. In contrast, LD refers to the nonrandom association of alleles at two loci across individuals in the population [1]. For example, co-segregation refers to the fact that, for a parent with genotype Mm at a marker and Qq at a linked QTL, which are arranged in coupling phase on the parent’s two homologous chromosomes, i.e. as MQ/mq rather than Mq/mQ, most progeny that receive marker allele M from this parent will also receive allele Q at the QTL because the marker and QTL ‘co-segregate’. Thus, with linkage and co-segregation, the marker allele that is transmitted by the parent is predictive of the QTL allele, allowing the use of the marker in MAS. Note, however, that which marker allele is ‘favourable’ depends on whether the marker and QTL are in coupling or repulsion phase in the parent, requiring MAS based on co-segregation to be implemented on a within-family basis. In contrast, LD refers to whether the frequency of coupling haplotypes in the population (MQ and mq) differs from the frequency of repulsion haplotypes (Mq and mQ) [1]. If a marker is in high LD with the QTL, then marker genotype will be predictive of QTL genotype across families. Because of the ability to use information across families, the use of markers that are in LD with QTL across the population is much more beneﬁcial than the use of markers that only co-segregate with QTL, but this does require availability of markers that are much closer to the QTL in the genome for LD to exist [8]. The need for many markers that are in LD with QTL across the population for MAS to become useful for genetic improvement was ﬁrst pointed out in 1993 by Smith and Smith [33], but has just now become feasible with the availability of cost-effective genotyping of large numbers of SNPs across the genome and development of statistical methods for the use of such data for prediction.

Opportunities for GS for Genetic Improvement of Livestock Based on the well-known breeder’s equation: response per year=i r s g =L the rate of genetic improvement in a population is roughly determined by the following factors [1]: 1. Accuracy of selection, r, that is, the accuracy with which the breeding value of a selection candidate can be predicted for the trait(s) of interest. 2. Generation interval, L, which is related to the age of parents when they are used for breeding. 3. Intensity of selection, i, which is related to the proportion of all candidates available for selection that are selected to be used for breeding. 4. Genetic standard deviation, sg, which is primarily a property of the trait and of the population but is reduced by inbreeding. The structure of breeding programmes that currently exist for livestock is primarily based on the requirement

g

that selection candidates must either be phenotyped themselves or have close relatives with phenotypes in order to obtain accurate estimates of breeding values for use in selection. Combined with the limitations and opportunities that are set by the male and female reproductive rates, this has, for example, resulted in the development and extensive use of progeny testing programmes in dairy cattle through the use of artiﬁcial insemination and frozen semen [34]. The size of the populations that are employed for genetic improvement is, apart from market and cost considerations, driven by inbreeding considerations. Inbreeding reduces genetic variance and, thereby, future responses to selection, and reduces viability and performance through inbreeding depression [1]. The rate of inbreeding is determined primarily by the number of males and females that are used for breeding and the extent to which they are related to each other [35]. Based on the factors that drive and limit genetic gain, one of the main challenges in animal breeding programmes is to manage and optimize the following three antagonistic relationships between these factors: i Accuracy of selection versus generation interval: accuracy of selection is determined by the amount of information that is available to estimate an individual’s breeding value, which generally increases with age of the animal (i.e. with more phenotypes collected on the individual itself and/or its relatives, including progeny). ii Intensity of selection versus inbreeding: although increasing the selection intensity by reducing the proportion of selection candidates to be used for breeding increases response in the short term, the smaller number of individuals used for breeding also increases rates of inbreeding. iii Accuracy of selection versus inbreeding: accuracy of selection can be increased by using phenotypes from relatives for calculation of EBV. However, because family members share family information, their EBVs tend to be more correlated if more family information is used [36]. This leads to greater emphasis on selection of families rather than of individuals from within a family, which results in the selected individuals to be more related to each other and increased rates of inbreeding. In fact the only sources of information that allow the EBV of an individual to be differentiated from that of its family members are those that express the effect of the speciﬁc sample genes that the individual received from its parents, rather than those that express the average effect of the family. In classical selection, phenotypic records that provide information on these so-called Mendelian sampling terms are limited to own phenotype and phenotype of progeny and further descendants [37]. In contrast, phenotypes of ancestors and collateral relatives, such as full- and half-sibs, provide information to estimate the average EBV of a family, but

CAB International 2010

Jack C.M. Dekkers

cannot be used to differentiate among family members. There are two additional antagonistic relationships that are relevant and these are related to the cost of breeding programmes: iv Accuracy of selection versus the cost of phenotyping: accuracy of selection can be increased by collecting more phenotypic data. Collecting phenotypic data is, however, expensive and is also limited by the requirement that phenotypes must be on the selection candidates themselves and/or their close relatives in order for them to be useful for estimation of breeding values of selection candidates. The latter is problematic for many traits. v Inbreeding and size of the breeding programme: rates of inbreeding can be reduced by increasing the number of selection candidates and maintaining the same proportion selected, such that the number of individuals used for breeding is increased without affecting selection intensity. Both the number of selection candidates and the number of breeding animals, however, have large impacts on costs of breeding programmes. GS can in principle address each of the antagonisms described above. These opportunities are predicated on the following properties of G-EBV: (a) G-EBV can be obtained as soon as DNA can be collected from selection candidates. This allows greater accuracy to be obtained at a younger age, allowing generation intervals to be reduced. (b) G-EBV are based on estimates of effects of the speciﬁc sample of alleles (for markers) that the individual has received from its parents, that is, they provide information on Mendelian sampling terms. This in theory reduces the emphasis that is placed on family information compared with phenotype-based EBV [38]. This reduces the rate of inbreeding for a given number of animals selected, or allows intensity to be increased without increasing rates of inbreeding, because individuals from the same family are less likely to have similar G-EBV. (c) G-EBV can in principle be predicted using estimates of SNP effects that are obtained from phenotypes on individuals that are not closely related to the selection candidates. With GS, in contrast to traditional selection, the phenotype that is collected on an individual cannot only be used to estimate the breeding value of the animal itself or its relatives, but also that of unrelated or less-related individuals. Thus, with GS, the requirement to collect phenotypes on selection candidates or their close relatives is removed and phenotypes collected can be leveraged for breeding value estimation across different families and generations and even to different populations and breeds. This, in principle, also reduces the amount of phenotypic data that may need to be collected.

g

201

In addition, phenotypes can be collected on individuals and in environments that are the target of genetic improvement, e.g. on crossbred individuals in commercial environments (see later). The above-mentioned properties of GS in principle also apply MAS. Thus, GS has opportunities to enhance breeding programmes in similar manners as have been described and to some extent implemented for MAS [8, 39, 40]. These include opportunities to increase accuracies of selection, in particular for difﬁcult-toimprove traits, and capitalizing on selection space that is not fully utilized in current breeding programmes, e.g. by pre-selecting individuals for entry into further testing programmes based on genetic markers [14, 41]. In most of these situations, information of markers is used to enhance selection decisions without drastically changing the programme. However, because GS has the potential to explain a much greater proportion of genetic variance than has been the case with MAS, and implementation is across rather than within families, GS has the potential to lead to substantial changes in animal breeding programmes and their operation. Some examples of the use of GS to dramatically change the structure of breeding programmes are presented in the following.

Reducing Generation Intervals Using deterministic simulations, Schaeffer [42] showed that implementation of GS in dairy cattle could substantially increase rates of genetic gain by allowing selection of bulls for breeding at a young age based on their G-EBV, rather than having to wait for progeny test results to become available. By reducing generation intervals on the bull side from over 5 to 2 years, Schaeffer [42] predicted that genetic gain per year could be doubled. Because GS can provide accurate EBV at a young age, there would not be a need to collect phenotypic records on large numbers of daughters of large numbers of young bulls through progeny testing which in addition to increasing generation intervals, has a substantial cost. Use of GS on the female side also has the potential to reduce generation intervals compared to current breeding programmes. Konig et al. [43] simulated and compared several strategies for GS for the German Holstein population and also identiﬁed opportunities for substantial increases in genetic gain by reducing generation intervals and substantial reductions in costs by eliminating bull housing. Opportunities to reduce generation intervals through GS also exist and have been investigated in other species, for example in layer poultry programmes [44]. Prior to the advance of GS, Georges and Massey [45] proposed the integration of MAS with the use of advanced reproduction techniques to more effectively capitalize on the beneﬁts of MAS by reducing generation intervals in dairy cattle. In their so-called ‘velogenetics’ strategies,

CAB International 2010

202

Animal Science Reviews 2010

short generations with selection on markers alone are facilitated by the use of reproductive technologies such as the recovery of oocytes from the unborn foetus, in vitro maturation of oocytes and in vitro fertilization. Enhancements to further reduce generation intervals in such programmes were suggested by Haley and Visscher [46] and Visscher et al. [47]. Apart from the feasibility of the reproductive technology components of these strategies, their feasibility has to date been hampered by the limited amount of genetic variation explained by available markers. GS has the potential to remove this latter limitation.

Selection for Traits with Limitations on Phenotype Recording As mentioned previously, selection for many traits is complicated by limitations on data collection and the requirement of phenotype-based selection approaches to obtain phenotypes on the selection candidates themselves and/or their close relatives. Examples are traits that require individuals to be sacriﬁced (e.g. carcass and meat quality traits) or exposed to conditions that make them unsuitable for breeding (e.g. disease traits). In addition, many of these traits are expensive to record, which often limits the amount of phenotypic data that are available to estimate the breeding value of selection candidates. GS allows phenotypic data that are collected within a population to be leveraged across all selection candidates and potentially across generations. As a result, requirements for the amount of phenotypic data that must be collected on a routine basis can be reduced and there is greater ﬂexibility with regard to which individuals those records are collected on. Opportunities to enhance genetic improvement of traits for which phenotypes cannot be recorded on selection candidates were previously investigated for MAS [40, 48, 49] and also apply to GS. Recently, opportunities to replace current sib testing approaches in ﬁsh breeding programmes with GS were investigated by Sonesson et al. [50]. Repeated collection of phenotypes may, however, be required to enable estimates of marker effects to be updated (re-training) over time. In addition, for traits with low heritability and that are affected by many QTLs of small effect, large data sets will be needed to obtain accurate estimates for prediction (see later).

Use of Data Across Breeds Most selection programmes are carried out within purebred populations, utilizing phenotypic data obtained on individuals from that breed. GS can, in principle, be utilized across breeds by training on combined data from multiple breeds. This has been demonstrated in simulation studies by Toosi et al. [51] and de Roos et al. [52], and in

g

real data by Hayes et al. [53], who combined data from Holstein and Jersey cattle. Results from these studies indicate that GS predictions developed in one breed tend to be poor predictors of breeding values of individuals from another breed. However, when training is on a data set that combines data from both breeds, GS predictions generally are at least as accurate for each breed as GS predictions developed on a within-breed basis and can be more accurate for breeds with limited amounts of data [53], despite differences in the QTL that segregate within each breed. In addition, SNP effects that are identiﬁed in across-breed training analyses tend to be more tightly linked to the QTL [51, 53], which may lead to the estimates remaining effective over more generations. Combining data is also beneﬁcial when the aim is to predict breeding values for crossbred individuals; Harris et al. [54] reported highest accuracies of G-EBV for crossbred HolsteinJersey bulls when data from both breeds were combined for training versus using data from each breed individually. The use of GS for training across breeds does require denser panels than may be needed within a breed (e.g. over 300 000 versus 50 000 SNPs, depending on the distance between the breeds involved) because only markers that are very close to the QTL are expected to have LD associations that are consistent across breeds [55, 56]. In addition, this approach would fail if effects of QTL are not consistent across breeds because of non-additive genetic effects or genotype by environment interactions.

Selection for Commercial Crossbred Performance In several livestock species, in particular pigs and poultry, animals that are used for commercial production are crossbreds but selection is conducted within the parental pure lines or breeds. In addition, the populations that are used for genetic improvement are typically kept under greater biosecurity than is required or economical in the ﬁeld. As a result of this pyramidal genetic improvement structure, although the ultimate goal is to improve performance of crossbreds in the ﬁeld, most phenotypic data that are used for selection are collected on purebreds in high-health environments. This has been shown to lead to substantial loss of genetic gain that is created in purebred populations when it is disseminated to the ﬁeld because of genotypeenvironment interactions and non-additive effects [57]. A potential solution is to utilize combined purebred and crossbred selection [58], which involves collecting phenotypic data on crossbred offspring in the ﬁeld and utilizing these data to estimate breeding values for animals in the purebred nucleus, along with data collected within the purebred nucleus populations. Such programmes are, however, expensive and require pedigree recording at the commercial level, which is often problematic, for example because of the use of mixed semen in pigs.

CAB International 2010

Jack C.M. Dekkers

GS can overcome these limitations by estimating SNP effects on a training data set that consists of phenotypes and genotypes collected on crossbred individuals in the ﬁeld and using the resulting SNP effect estimates to estimate breeding values of individuals in the purebred nucleus populations based on their high-density SNP genotypes [59]. By training on data from crossbreds in the ﬁeld, estimates of marker effects will be for performance of crossbreds under ﬁeld conditions, rather than for purebred performance in nucleus herds. Because GS does not require pedigree information, training can, in principle, be on an unpedigreed sample of individuals. Dekkers [59] showed that, if accurate G-EBV for crossbred performance could be obtained, this has the potential to substantially increase accuracy of selection for commercial crossbred performance, without the increases in generation intervals or rates of inbreeding that are associated with phenotype-based combined crossbred purebred selection [60]. Simulation studies by Ibanez et al. [61] and Toosi et al. [51] have shown that it is possible to obtain accurate G-EBV from training on crossbred data, even without speciﬁcally accounting for breed composition or breed origin of marker alleles.

Genetic Improvement in Under-developed Countries Genetic improvement in developing and under-developed countries is often limited by lack of infrastructure, that is, a lack of programmes that record phenotype on pedigreed animals in the ﬁeld. To overcome this, the use of nucleus breeding programmes has been proposed, in which selection is within a small elite breeding population that produces breeding stock or germplasm for use in the ﬁeld [62]. Although this has led to some successes [63], these programmes are often faced with extensive genotypeenvironment interactions and incomplete acceptance of nucleus breeding stock in the ﬁeld [64]. Similar to the situation described in the previous section for commercial crossbred selection, GS offers opportunities to overcome these limitations by collecting phenotypes and DNA from a large sample of animals in the ﬁeld, without requiring them to be pedigreed or even to have known breed composition. The resulting estimates could be used to either select within a nucleus herd or to identify superior breeding stock in the ﬁeld on the basis of a DNA sample. Although this does not remove the requirement for collecting phenotypic data because large data sets are required for training, record collection can be without requiring pedigree on the phenotyped individuals that are genotyped. In addition, through GS, the collected phenotypes can be used to estimate breeding values of individuals across the population and across generations, and are not limited to estimation of breeding values for individuals that are direct relatives of the recorded individuals. Combined with the typically low heritability of traits in these conditions, which increases

g

203

the size of training data needed, the cost to implement GS in under-developed countries can, however, not be understated. Without availability of pedigree, opportunities to use low-density SNP panels (see later) for training and prediction (see later) are also reduced.

Reducing Rates of Inbreeding or Size of Breeding Programmes Inbreeding is an important consideration in the design of breeding programmes and often sets a lower limit on the number of individuals that are selected to produce the next generation. Combined with selection intensity and cost considerations, this drives the size of breeding populations that are required to achieve desired rates of genetic improvement. With GS, the rate of inbreeding for a given number of individuals used for breeding is expected to be lower because GS does not rely on family information and increases the accuracy of estimates of Mendelian sampling terms [38, 65]. This property of GS can be used to reduce the rate of inbreeding for the existing breeding programme design or to increase the efﬁciency and effectiveness of the breeding programme while maintaining the same rate of inbreeding, by reducing the number of individuals selected. The latter could be capitalized on by increasing selection intensities and thereby rates of genetic improvement, or by reducing the size of the breeding programme population, resulting in reduced animal-related programme costs. In addition, it is well known that factorial mating designs result in lower rates of inbreeding than the hierarchical mating programmes of mating one female to only one male [66]. Factorial mating designs through, e.g., the use of mixed semen in pigs or rotating roosters in chickens, are enabled by GS because marker genotypes allow parentage identiﬁcation, resulting in additional opportunities to capitalize on reduced rates of inbreeding in breeding programme design.

Results from implementation of GS analyses The opportunities for GS reviewed in the previous section were based on the premise that G-EBV can be derived on training data of manageable size (500 to 2000) and that the resulting prediction equations can be used across generations and populations. These premises were borne out in the initial simulation studies that were conducted on GS [13, 21, 67, 68]. GS analyses have now been conducted on a wide scale for dairy cattle in several countries [16, 17] and research to implement GS in other countries and species, including crops, is under way [7, 10, 18, 22]. Reported accuracies of G-EBV in dairy cattle based on large training data sets of progeny-tested bulls range from 0.36 to 0.85 [22]. Accuracies depended in part on the size of the training data, heritability and SNP

CAB International 2010

204

Animal Science Reviews 2010

density and typically required much larger training data sets than was indicated by the initial simulation results (several thousands of genotyped bulls with accurate progeny test information). In addition, in most of these studies, models that allowed for unequal variances for SNPs, such as Bayes-B [13] or comparable non-linear regression models implemented by Van Raden et al. [15], resulted in accuracies of G-EBV that were not much greater and sometimes lower than the G-BLUP model that assumes equal variance across SNPs [16, 22]. These ﬁndings are typically interpreted as evidence that the traits analysed are controlled by large numbers of QTLs of small effect [15, 16, 22], in contrast to most simulations on GS, which typically assumed a limited number of QTLs (100 or less) with larger effects. And indeed, in a recent theoretical work, Goddard [69] showed that much larger training data sets are needed when the number of QTLs is large and they have normally distributed effects, compared to when the number of QTLs is smaller and they have effects from an exponential distribution. Similar theoretical results were obtained by Meuwissen [70]. Goddard [69] also showed that if the number of QTLs is large, models that assume equal variance across SNPs are expected to do as well or better than models that assume unequal variance because most SNPs will be in LD with one or more QTLs of similar effect. However, the presence of a large number of QTLs does not explain the ﬁnding of Van Raden et al. [15] and of Weigel et al. [71] that randomly reducing the number of SNPs used in the analysis did not result in very large drops in accuracy of G-EBV in the US Holstein population; squared accuracies dropped by less than 4 percentage points for most traits when the number of SNPs was reduced from 38 400 to 9600 in Van Raden et al. [15] and by less than 9 percentage points (from 37.5 to 29.1%) when the number of SNPs was reduced from 32 518 to as few as 2000 equally spaced SNPs in Weigel et al. [71]. If the number of QTLs were indeed large, reducing the number of SNPs would result in many QTLs not to be in LD with an SNP, especially considering the level of LD that exists in typical dairy cattle populations [55].

Impact of Genetic Relationships Derived from Markers on G-EBV Although the theory of GS, as proposed by Meuwissen et al. [13], is predicated on capturing associations between markers and QTLs because of historic LD associations between markers and QTLs, it is now becoming clear that other factors can make major contributions to GS predictions from training data from populations with dense and complex pedigree and family structures, and may explain the above-mentioned ﬁndings of Van Raden et al. [15] and Weigel et al. [71]. These include genetic relationships [21] and recent LD and within-family effects [72]. In fact, there is some evidence that these factors

g

other than historic LD may form the most important contributions to the accuracy of G-EBV in the recent applications of GS to dairy cattle data; Van Raden [73] simulated SNP genotypes and trait phenotypes using the actual pedigree from the population used in the application of GS by Van Raden et al. [15] but without simulating historical LD; markers and QTLs were simulated to be in linkage equilibrium for founder individuals of the available pedigree. However, despite the absence of historic LD, accuracies of G-EBV from the simulated data were substantial and similar to those found in subsequent analysis of the real data. The impact of genetic relationships of individuals in the validation data with individuals in the training data on G-EBV in the validation data was demonstrated by Habier et al. [21]. In fact, it is well known that the so-called genomic relationship matrix that is based on genetic markers can replace the traditional pedigree-based relationship that is used in traditional phenotype-based BLUP prediction of breeding values and that the G-BLUP model of GS is equivalent to use of the genomic rather than the pedigree relationship matrix [21, 69, 74–77]. The implication of this equivalence is that part of the accuracy of G-EBV derives from the ability of the markers to capture pedigree relationships, similar to animal model BLUP. Estimation of genetic relationships from marker genotypes does not require markers to be associated with the trait through linkage or LD with QTLs. However, if the markers are linked to QTLs, then the so-called realized genomic relationship matrix will also account for deviations of the actual from the expected relationships that are captured by pedigree [78, 79], leading to more accurate EBV [15, 72, 80, 81]. This increase in accuracy from including the realized relationship matrix in BLUP was ﬁrst demonstrated by Nejati-Javaremi et al. [82]. They, however, computed the realized relationships under the assumption that all QTLs could be genotyped. Villanueva et al. [83] demonstrated that a realized relationship matrix that is derived using a limited number of markers that are linked to QTLs can also increase the accuracy of selection by capturing associations of markers with phenotype within a family, similar to linkage or cosegregation analysis methods. Increases in accuracy were, however, modest. Recently, Hayes et al. [84] showed that with large family sizes, use of the genomic relationship matrix can result in substantial increases in accuracies. They also showed that realized genomic relationships can capture LD that exists between markers and QTL across the population. Genomic relationships also do not suffer from missing or incorrect pedigree information, which can be sizeable in livestock breeding programmes and can result in substantial loss of accuracy of EBV. Genomic relationships do suffer from mislabelling of DNA samples and from genotyping errors, although the impact of the latter will be limited if large numbers of markers are used.

CAB International 2010

Jack C.M. Dekkers

Implications of the Impact of Relationships and Family Effects on G-EBV for GS The ﬁnding that different types of effects contribute to G-EBV, including historic LD, genetic relationships, and within-family effects, has important implications for the use and implementation of GS because effects that are the result of markers that are not closely linked to QTL will more rapidly decline over generations as their associations are lost by recombination [21]. Although the ability of G-EBV to capture genetic relationships and within-family effects contributes to the accuracy of G-EBV for close relatives of individuals used for training (e.g. their progeny), they limit the accuracy of G-EBV for subsequent generations, for families that are not well represented in the training data and for other less related populations. This was demonstrated in a real data example of mice by Legarra et al. [85], in which they showed that the accuracy of G-EBV was much higher when selection candidates were family members of individuals used for training than when they were not. Similarly, in an analysis of German Holstein data, Habier et al. [86] found that accuracies of G-EBV were much poorer for bulls that had low relationships with the training population than for bulls that were closely related to the training population. The implication is that G-EBV will be most accurate for selection candidates that are closely related to individuals that are in the training data and that accuracies may be much lower for individuals that are less related, in particular for individuals from other breeds. Thus, continuous phenotypic recording across families may be needed to maintain accuracies of G-EBV within a breed. The beneﬁt of continuous phenotype recording and retraining has been demonstrated in simulation studies [51, 67]. The contribution of family information to G-EBV also results in family members ranking more similarly on G-EBV. This leads to smaller opportunities to reduce rates of inbreeding than predicted by e.g. Daetwyler et al. [38].

Improved Estimation Procedures for Genomic Selection The issues outlined in the previous section suggest that, although current methods for statistical analysis for GS can result in substantial accuracies of G-EBV, there is substantial room to improve the accuracy of predictions. One approach would be to use models or data sets for GS analysis that emphasize effects of SNPs that are associated with the trait through short-range LD with the QTL, versus SNPs that are associated with the trait through genetic relationships and within-family effects, such that effects are more predictive across generations and families. Several approaches have been suggested to accomplish this, including using models that assume that only a small proportion of SNPs have effects, such as the Bayes-B

g

205

method of Meuwissen et al. [13], and using multi-breed populations for training. Although methods such as BayesB may result in poorer accuracy of G-EBV of progeny of individuals in the training data than G-BLUP, there is evidence that these methods result in better predictive ability across generations and of less related individuals. The beneﬁt of Bayes-B over G-BLUP for maintaining accuracy of G-EBV across generations was also demonstrated with simulation by Habier et al. [21] and Zhong et al. [99]. In their analysis of German Holstein data, Habier et al. [86] also showed that G-EBV of bulls that had no close relatives in the training data were better predicted by Bayes-B than G-BLUP. Recent multi-breed analyses of dairy cattle data by Hayes et al. [53] also demonstrated that Bayes-B resulted in more accurate G-EBV than G-BLUP when training was on combined data from Holstein and Jersey bulls. Properties of GS prediction across breeds were also investigated by Kizilkaya et al. [88]. Multi-breed data emphasize marker–trait associations that are consistent across breeds, i.e. associations that are based on historic LD rather than recent relationships or within-family effects. Using simulation, Fernando et al. [89] also showed that the Bayes-B method gave higher accuracies of G-EBV than G-BLUP if the number of SNPs was large compared to the number of QTLs. Other approaches that may help disentangle the components that contribute to G-EBV are inclusion of a polygenic effect [90] and using the so-called combined linkage and linkage disequilibrium (LD-LA) analysis approaches. The latter explicitly model the co-segregation of markers and QTLs within families based on the method of Fernando and Grossman [91], along with LD between markers and QTLs in the founders [92–94]. Calus et al. [95] found that inclusion of co-segregation compared to GS based on LD only increased accuracy of G-EBV for traits with high heritability but not for low-heritability traits. This may, however, be the result of their simulation representing balanced family structures and shallow pedigrees and further work is needed to evaluate the impact of LD-LA methods on the deep unbalanced pedigrees that characterize livestock populations. Although the initial work on GS by Meuwissen et al. [13] proposed effects to be estimated for haplotypes of SNPs, most subsequent work and most applications to real data have included separate effects for each SNP in the analysis. Although this simpliﬁes the analysis and can be implemented without knowledge of the order and position of SNPs in the genome, it may not be optimal because some QTLs may only be in high LD with a haplotype of SNPs [10]. Calus et al. [87], however, found little advantage in using haplotypes rather than individual SNPs for GS prediction. In contrast, Villumsen et al. [96] found that GS models based on haplotypes performed better than SNP-based models and that there is an optimal length of haplotypes that maximizes accuracy of G-EBV.

CAB International 2010

206

Animal Science Reviews 2010

Re-Design of Breeding Strategies Based on GS To date, most studies on the beneﬁt of GS for genetic improvement programmes have focused on the accuracy of G-EBV in generations immediately following training or, if the dynamics of accuracy over multiple generations were studied, it was in unselected populations [13, 21]. Only a few studies have evaluated opportunities to redesign breeding programmes to fully capitalize on the beneﬁts of GS and these were primarily based on deterministic analyses that may not accurately model responses to GS. Schaeffer [42] evaluated opportunities for redesign of dairy cattle breeding programmes with GS using simpliﬁed deterministic estimates of response to selection and these were extended to a deterministic economic evaluation by Konig et al. [43]. Selection index methods to evaluate the impact of GS on response and inbreeding were developed by Dekkers [65]. Deterministic predictions of inbreeding with GS were also developed by Daetwyler et al. [38]. Goddard [69] used analytical derivations to predict long-term responses to GS and developed selection strategies to optimize long-term response. Bernardo and Yu [97] and Heffner et al. [18] reviewed prospects and strategies for the use of GS in crop improvement. The only published studies that have evaluated the dynamics of GS over multiple generations by stochastic simulation under selection are by Muir [67] and Sonesson et al. [50]. Muir [67] simulated a breeding programme with discrete generations for a trait that allowed phenotypes to be obtained on both sexes prior to selection. Sonesson et al. [50] evaluated prospects for reducing sib testing in aquaculture breeding programmes for traits that cannot be evaluated on selection candidates. Both studies used G-BLUP to estimate marker effects and found that accuracy of G-EBV declined rapidly over generations. This decline in accuracy can be attributed to a combination of factors, including (i) the decline of relationships between selection candidates and individuals in the training data as the number of generations that separates them increases, (ii) the break-up of LD between markers and QTL over generations due to recombination, and (iii) changes in frequencies at the markers as a result of selection on G-EBV. Both studies found that the decline in accuracy could be prevented by continuing to collect phenotypes over generations and using these phenotypes to re-train the GS prediction model. In addition to increasing the size of the training data, this re-training also results in stronger genetic relationships between selection candidates and individuals in the training data, which increases accuracies of G-EBV. Sonesson et al. [50] found that GS reduced rates of inbreeding by up to 81% compared to the sib-test selection scheme, as a result of increased within-family selection. Neither study evaluated opportunities to reduce generation intervals with GS. Further work is needed to design and evaluate breeding programmes to fully capitalize on the beneﬁts of GS. Because GS removes many of the limitations that exist in

g

current phenotype-based breeding programmes with regard to when and on which individuals phenotypes must be evaluated, opportunities for a complete redesign of breeding programmes exist. However, given the impact that relationships of selection candidates with individuals in the training data have on accuracy of G-EBV, these limitations may not be completely removed. Thus, comprehensive strategies for selective recording of phenotypes with GS must be evaluated. Redesign of breeding programmes with GS should also capitalize on the greater emphasis that GS places on selection within families. This can be utilized in a number of ways: (i) reducing rates of inbreeding, (ii) increasing response to selection for a given rate of inbreeding by increasing selection intensities or (iii) reducing the size of breeding populations and, thereby, costs, while maintaining the same rates of response and inbreeding. Although the cost of high-density SNP genotyping has declined rapidly over the past years, costs are expected to remain high for implementation of GS on a large scale, in particular for species with many selection candidates, such as beef cattle, pigs, poultry, sheep and aquaculture. Two approaches have been proposed for developing smaller, less costly, low-density genotyping panels, namely: (i) by identifying a subset of the SNPs that have strong associations with the trait [71] or (ii) by using sparse evenly spaced SNPs across the genome and use these to impute high-density SNP genotypes that are not on the small panel, on a within-family basis [98]. In contrast to approach (i), the evenly spaced low-density approach results in panels that are not trait- or population-speciﬁc and that are robust to the underlying genetic architecture of the trait [98]. Additional work is, however, needed to optimize the application of these low-density approaches by determining which individuals should be genotyped using the low- versus high-density panels, both in the training and target populations.

Conclusions and Perspectives GS represents an advanced level of marker-assisted selection that is made possible by recent advances in marker genotyping technology and statistical methodology for the use of large numbers of markers across the genome for prediction of breeding values. The main difference with previous strategies of using molecular marker data in marker-assisted selection is that GS represents a technological development in animal breeding that has the potential to result in unprecedented changes in breeding programmes and opportunities to increase the efﬁciency and effectiveness of breeding programmes. The scope of the changes GS could bring about are paralleled perhaps only by the introduction of artiﬁcial insemination and semen freezing technology in dairy cattle breeding programmes but will extend across all livestock species. Although initial results of the application of GS to real

CAB International 2010

Jack C.M. Dekkers

data are promising, additional research is needed to better understand the factors that contribute to the accuracy of G-EBV, further development of statistical methods for prediction of G-EBV and design of breeding programmes that optimally capitalize on the opportunities provided by GS to increase response to selection and reduce rates of inbreeding. Although GS results in the ability to predict G-EBV of individuals without phenotype, large amounts of phenotypic data are needed to develop predictions with sufﬁcient accuracy. While inclusion of data on close relatives of selection candidates for prediction increases the accuracy of G-EBV, with sufﬁciently large training data sets and sufﬁcient markers, GS can be used to predict G-EBV across generations and even across breeds. This opens opportunities for development of G-EBV prediction equations using phenotypes collected on individuals kept in environments that are relevant to production, such as on crossbred individuals in ﬁeld environments. Acknowledgements

207

12. Van Tassell CP, Smith TPL, Matukumalli LK, Taylor JF, Schnabel RD, Lawley CT et al. SNP discovery and allele frequency estimation by deep sequencing of reduced representation libraries. Nature Methods 2008;5:247–52. 13. Meuwissen THE, Hayes BJ, Goddard ME. Prediction of total genetic value using genome-wide dense marker maps. Genetics 2001;157:1819–29. 14. Lande R, Thompson R. Efﬁciency of marker-assisted selection in the improvement of quantitative traits. Genetics 1990;124:743–56. 15. VanRaden PM, Van Tassell CP, Wiggans GR, Sonstegard TS, Schnabel RD, Taylor JF et al. Invited review: Reliability of genomic predictions for North American Holstein bulls. Journal of Dairy Science 2009;92:16–24. 16. Hayes BJ, Bowman PJ, Chamberlain AJ, Goddard ME. Invited review: Genomic selection in dairy cattle: progress and challenges. Journal of Dairy Science 2009;92:433–43. 17. Loberg A, Durr J. Interbull Survey on the Use of Genomic Information. Proc. Interbull International Workshop, 2009, www-interbull.slu.se, Uppsala, Sweden; 2009. p. 3–14. 18. Heffner EL, Sorrells ME, Jannink JL. Genomic selection for crop improvement. Crop Science 2009;49:1–12.

The author acknowledges his colleagues Rohan Fernando, Dorian Garrick and David Habier for stimulating discussions, which have contributed to many aspects of this review.

19. Hill WG, Goddard ME, Visscher PM. Data and theory point to mainly additive genetic variance for complex traits. PLoS Genetics 2008;4(2):e1000008. doi: 10.1371/journal.pgen. 1000008. 20. Gianola D, de los Campos G, Hill WG, Manfredi E, Fernando R. Additive genetic variability and the Bayesian alphabet. Genetics 2009;183:347–63.

References 1. Falconer DS, McKay TFC. Introduction to Quantitative Genetics. Longman, Harlow, UK; 1996. xii, 438p 2. Henderson CR. Applications of Linear Models in Animal Breeding. University of Guelph, Guelph, ON, Canada; 1984. 3. Lynch MWB. Genetics and Analysis of Quantitative Traits. Sinauer Associates Inc., Sunderland, MA; 1998. 4. Dekkers JCM, Hospital F. The use of molecular genetics in the improvement of agricultural populations. Nature Reviews Genetetics 2002;3:22–32. 5. Andersson L. Genetic dissection of phenotypic diversity in farm animals. Nature 2001;2:130–8. 6. Weller JL. Quantitative Trait Loci Analysis in Animals. CABI, Cambridge, MA; 2009.

21. Habier D, Fernando RL, Dekkers JC. The impact of genetic relationship information on genome-assisted breeding values. Genetics 2007;177:2389–97. 22. Calus MPL. Genomic breeding value prediction: methods and procedures. Animal 2010;4:157–64. 23. Gianola D, Fernando RL, Stella A. Genomic-assisted prediction of genetic value with semiparametric procedures. Genetics 2006;173:1761–76. 24. Bennewitz J, Solberg T, Meuwissen T. Genomic breeding value estimation using nonparametric additive regression models. Genetics Selection Evolution 2009;41:20, doi: 10.1186/1297-9686-41-20.

7. Goddard ME, Hayes BJ. Mapping genes for complex traits in domestic animals and their use in breeding programmes. Nature Review Genetics 2009;10:381–91.

25. Gonzalez-Recio O, Gianola D, Long N, Weigel KA, Rosa GJ, Avendano S et al. Nonparametric methods for incorporating genomic information into genetic evaluations: an application to mortality in broilers. Genetics 2008;178:2305–13.

8. Dekkers JCM. Commercial application of marker- and geneassisted selection in livestock: strategies and lessons. Journal of Animal Science 2004;82(E-Suppl):E313–28.

26. Gianola D, van Kaam JBCHM. Reproducing kernel Hilbert spaces regression methods for genomic assisted prediction of quantitative traits. Genetics 2008;178:2289–303.

9. Boichard D, Fritz S, Rossignol MN, Guillaume F, Colleau JJ, Druet T. Implementation of marker assisted selection: practical lessons from dairy cattle. 8th World Congress on Genetics Applied to Livestock Production; 13–18 August 2006; Belo Horizonte, MG, Brazil.

27. Raadsma HW, Zenger KR, Khatkar MS, Crump R, Moser G, Solkner J et al. Genome wide selection in dairy cattle based on high-density genome-wide SNP analysis: from discovery to application. In: Legarra A, editor. Proceedings XI QTLMAS 2007. Toulouse. INRA, Castanet Tolosan, France; 2007. p. 77. Available from: URL: http://germinal.toulouse.inra.fr/qtlmas/

10. Goddard ME, Hayes BJ. Genomic selection. Journal of Animal Breeding and Genetics 2007;124:323–30. 11. Wong GKS, Liu B, Wang J, Zhang Y, Yang X, Zhang Z et al. A genetic variation map for chicken with 2.8 million singlenucleotide polymorphisms. Nature 2004;432:717–22.

g

28. Solberg TR, Sonesson AK, Woolliams JA, Meuwissen TH. Reducing dimensionality for prediction of genome-wide breeding values. Genetics Selection Evolution 2009;41:29, doi: 10.1186/1297-9686-41-29.

CAB International 2010

208

Animal Science Reviews 2010

29. de los Campos G, Naya H, Gianola D, Crossa J, Legarra A, Manfredi E et al. Predicting quantitative traits with regression models for dense molecular markers and pedigree. Genetics 2009;182:375–85. 30. Beavis WD. QTL analyses: power, precision and accuracy. In: Paterson H, editor. Molecular dissection of complex traits, CRC Press, Boca Raton, FL; 1998. p. 145–59. 31. Xu SZ. Estimating polygenic effects using markers of the entire genome. Genetics 2003;163:789–801. 32. Rothschild MF, Soller M. Candidate gene analysis to detect traits of economic importance in domestic livestock. Probe 1997;8:13. 33. Smith C, Smith DB. The need for close linkages in markerassisted selection for economic merit in livestock. Animal Breeding Abstracts 1993;61:197–204. 34. Henderson CR. Selecting the young sire to sample in artiﬁcial insemination. Journal of Dairy Science 1964;47:439–41. 35. Meuwissen THE. Maximizing the response of selection with a predeﬁned rate of inbreeding. Journal of Animal Science 1997;75:934–40. 36. Wray NR, Wooliams JA, Thompson R. Methods for predicting rates of inbreeding in selected populations. Theoretical and Applied Genetics 1990;80:503–12. 37. Avendano S, Woolliams JA, Villanueva B. Prediction of accuracy of estimated Mendelian sampling terms. Journal of Animal Breeding and Genetics 2005;122:302–8. 38. Daetwyler HD, Villanueva B, Bijma P, Woolliams JA. Inbreeding in genome-wide selection. Journal of Animal Breeding and Genetics 2007;124:369–76. 39. Dekkers JCM, van der Werf JHJ. Strategies, limitations and opportunities for marker-assisted selection in livestock. In: Guimara˜es EP, Ruane J, Scherf BD, Sonnino A, Dargie JD, editors. Marker-assisted Selection: Current Status and Future Perspectives in Crops, Livestock, Forestry and Fish. Food and Agriculture Organization of the United Nations, Rome, Italy; 2007. 40. Meuwissen THE, Goddard ME. The use of marker haplotypes in animal breeding schemes. Genetics Selection Evolution 1996;28:161–76. 41. Kashi Y, Hallerman E, Soller M. Marker-assisted selection of candidate bulls for progeny testing programmes. Animal Production 1990;51:63–74. 42. Schaeffer LR. Strategy for applying genome-wide selection in dairy cattle. Journal of Animal Breeding and Genetics 2006;123:218–23. 43. Konig S, Swalve HH. Application of selection index calculations to determine selection strategies in genomic breeding programs. Journal of Dairy Science 2009;92:5292–303. 44. Dekkers JCM, Zhao HH, Habier D, Fernando RL. Opportunities for genomic selection with redesign of breeding programs. Montreal, Canada; Journal of Animal Science 2009;87(E-Suppl. 2), Abstract No. 275. 45. Georges M, Massey JM. Velogenetics, or the synergistic use of marker assisted selection and germ-line manipulation. Theriogenology 1991;25:151–9. 46. Haley CS, Visscher PM. Strategies to utilize markerquantitative trait loci associations. Journal of Dairy Science 1998;81:85–97.

g

47. Visscher P, Pong-Wong R, Whittemore C, Haley C. Impact of biotechnology on (cross)breeding programmes in pigs. Livestock Production Science 2000;65:57–70. 48. Sonesson AK. Within-family marker-assisted selection for aquaculture species. Genetics Selection Evolution 2007;39:301–17. 49. Hayes B, Goddard ME. Evaluation of marker assisted selection in pig enterprises. Livestock Production Science 2003;81:197–211. 50. Sonesson AK, Meuwissen TH. Testing strategies for genomic selection in aquaculture breeding programs. Genetics Selection Evolution 2009;41:37, doi: 10.1186/ 1297-9686-41-37. 51. Toosi A, Fernando RL, Dekkers JCM. Genomic selection of purebreds using data from admixed populations. Journal of Animal Science 2010;88:32–46. 52. de Roos AP, Hayes BJ, Goddard ME. Reliability of genomic predictions across multiple populations. Genetics 2009;183:1545–53. 53. Hayes BJ, Bowman PJ, Chamberlain AC, Verbyla K, Goddard ME. Accuracy of genomic breeding values in multi-breed dairy cattle populations. Genetics Selection Evolution 2009;41:51, doi: 10.1186/1297-9686-41-51. 54. Harris BL, Johnson DL, Spelman RJ. Genomic Selection in New Zealand and the Implications for National Genetic Evaluation. ICAR Technical Series: Identiﬁcation, Breeding, Production, Health and Recording of Farm Animals. ICAR, Rome, 2009. 55. de Roos APW, Hayes BJ, Spelman RJ, Goddard ME. Linkage disequilibrium and persistence of phase in Holstein-Friesian, Jersey and Angus cattle. Genetics 2008;179:1503–12. 56. Goddard ME. How Can We Best Use DNA Data in Selection of Cattle? Beef Improvement Federation, Sacramento, CA; 2009. 57. Wei M, Werf JHJVD, Brascamp EW. Relationship between purebred and crossbred parameters. II. Genetic correlation between purebred and crossbred performance under the model with two loci. Journal of Animal Breeding and Genetics 1991;108:262–9. 58. Wei M, van der Steen HAM. Comparison of reciprocal recurrent selection with pure-line selection systems in animal breeding (a review). Animal Breeding Abstract 1991; 59:281–98. 59. Dekkers JC. Marker-assisted selection for commercial crossbred performance. Journal Animal Science 2007;85:2104–14. 60. Bijma P, van Arendonk JAM. Maximizing genetic gain for the sire line of a crossbreeding scheme utilizing both purebred and crossbred information. Animal Science 1998;66:529–42. 61. Ibanez-Escriche N, Fernando RL, Toosi A, Dekkers JCM. Genomic selection of purebreds for crossbred performance. Genetics Selection Evolution 2009;41:12. 62. Kosgey IS, Kahi AK, Van Arendonk JA. Evaluation of closed adult nucleus multiple ovulation and embryo transfer and conventional progeny testing breeding schemes for milk production in tropical crossbred cattle. Journal of Dairy Science 2005;88:1582–94. 63. Bosso NA, Corr N, Njie M, Fall A, Van der Waaij EH, Van Arendonk J.A.M., et al. The N’Dama Cattle Genetic Improvement Programme: a review. Animal Genetic Resources Information 2007;40:65069.

CAB International 2010

Jack C.M. Dekkers

209

64. Cunningham EP. The application of biotechnologies to enhance animal production in different farming systems. Livestock Production Science 1999;58:1–24.

82. Nejati-Javaremi A, Smith C, Gibson JP. Effect of total allelic relationship on accuracy of evaluation and response to selection. Journal of Animal Science 1997;75:1738–45.

65. Dekkers JC. Prediction of response to marker-assisted and genomic selection using selection index theory. Journal of Animal Breeding Genetics 2007;124:331–41.

83. Villanueva B, Pong-Wong R, Ferna´ndez J, Toro MA. Beneﬁts from marker-assisted selection under an additive polygenic genetic model. Journal of Animal Science 2005;83:1747–52.

66. Woolliams JA. Modiﬁcations to MOET nucleus breeding schemes to improve rates of genetic progress and decrease rates of inbreeding in dairy cattle. Animal Production 1989;49:1–14.

84. Hayes BJ, Visscher PM, Goddard ME. Increased accuracy of artiﬁcial selection by using the realized relationship matrix. Genetical Research 2009;91:47–60.

67. Muir WM. Comparison of genomic and traditional BLUPestimated breeding value accuracy and selection response under alternative trait and genomic parameters. Journal of Animal Breeding and Genetics 2007;124:342–55. 68. Solberg TR, Sonesson AK, Woolliams JA, Meuwissen THE. Genomic selection using different marker types and densities. Journal of Animal Science 2008;86:2447–54. 69. Goddard M. Genomic selection: prediction of accuracy and maximisation of long term response. Genetica 2009;136: 245–57. 70. Meuwissen TH. Accuracy of breeding values of ‘unrelated’ individuals predicted by dense SNP genotyping. Genetics Selection Evolution 2009;41:35, doi: 10.1186/1297-9686-41-35. 71. Weigel KA, de los Campos G, Gonzalez-Recio O, Naya H, Wu XL, Long N et al. Predictive ability of direct genomic values for lifetime net merit of Holstein sires using selected subsets of single nucleotide polymorphism markers. J Dairy Sci 2009;92:5248–57. 72. Goddard ME. The use of high density genotyping in animal health. In: Pinard M-H, Gay C, Pastoret P-P, Dodet B, editors. Animal Genomics for Animal Health. Developments in Biologicals; 132. Karger, Basel, Switzerland; 2008. p. 383–90. 73. VanRaden PM. Efﬁcient methods to compute genomic predictions. Journal of Dairy Science 2008;91:4414–23. 74. Fernando RL. Genetic Evaluation and Selection Using Genotypic, Phenotypic and Pedigree Information. Animal Breeding and Genetics Unit, University of Australia, Armidale, Australia; 1998. p. 329–36. 75. Zhang Z, Todhunter RJ, Buckler ES, Van Vleck LD. Technical note: use of marker-based relationships with multiple-trait derivative-free restricted maximal likelihood. Journal of Animal Science 2007;85:881–5. 76. Legarra A, Misztal I. Technical note: computing strategies in genome-wide selection. Journal of Dairy Science 2008;91:360–6. 77. Stranden I, Garrick DJ. Technical note: derivation of equivalent computing algorithms for genomic predictions and reliabilities of animal merit. Journal of Dairy Science 2009;92:2971–5.

85. Legarra A, Robert-granie C, Manfredi E, Elsen J-M. Performance of genomic selection in mice. Genetics 2008;180:611–8. 86. Habier D, Tetens J, Seefried F-R, Lichtner P, Thaller G. The impact of genetic relationship information on genomic breeding values in German Holstein cattle. Genetics Selection Evolution 2010;42:5, doi: 10.1186/1297-9686-42-5. 87. Calus MPL, Meuwissen THE, Windig JJ, Knol EF, Schrooten C, Vereijkin ALJ et al. Effects of the number of markers per haplotype and clustering of haplotypes on the accuracy of QTL mapping and prediction of genomic breeding values. Genetics Selection Evolution 2009;41:11. 88. Kizilkaya K, Fernando RL, Garrick DJ. Genomic prediction of simulated multi-breed and purebred performance using observed 50k SNP genotypes. Journal of Animal Science 2010;88:544–51. 89. Fernando RL, Habier D, Stricker C, Dekkers JCM, Totir LR. Genomic selection. Acta Agricultura Scandinavica Section A, Animal Science 2007;57:192–5. 90. Calus MPL, Veerkamp RF. Accuracy of breeding values when using and ignoring the polygenic effect in genomic breeding value estimation with a marker density of one SNP per cM. Journal of Animal Breeding and Genetics 2007;124:362–8. 91. Fernando RL, Grossman M. Marker assisted selection using best linear unbiased prediction. Genetics Selection Evolution 1989;21:467–77. 92. Meuwissen THE, Karlsen A, Lien S, Olsaker I, Goddard ME. Fine mapping of a quantitative trait locus for twinning rate using combined linkage and linkage disequilibrium mapping. Genetics 2002;161:373–9. 93. Fernando RL, Totir LR. Incorporating Molecular Information in Breeding Programs: Methodology. Poultry Breeding and Biotechnology. CABI Publishing, Cambridge, UK; 2003. 94. Meuwissen THE, Goddard ME. Mapping multiple QTL using linkage disequilibrium and linkage analysis information and multitrait data. Genetics Selection Evolution 2004;36:261–79. 95. Calus MPL, Meuwissen THE, Roos APWD, Veerkamp RF. Accuracy of genomic selection using different methods to deﬁne haplotypes. Genetics 2008;178:553–61.

78. Hill WG. Variation in genetic identity within kinships. Heredity 1993;71:652–3.

96. Villumsen TM, Janss L, Lund MS. The importance of haplotype length and heritability using genomic selection in dairy cattle. Journal of Animal Breeding and Genetics 2009;126:3–13.

79. Guo SW. Variation in genetic identity among relatives. Human Heredity 1996;46:61–70.

97. Bernardo R, Yu JM. Prospects for genomewide selection for quantitative traits in maize. Crop Science 2007;47:1082–90.

80. Legarra A, Aguilar I, Misztal I. A relationship matrix including full pedigree and genomic information. Journal of Dairy Science 2009;92:4656–63.

98. Habier D, Fernando RL, Dekkers JCM. Genomic selection using low-density marker panels. Genetics 2009;182:343–53.

81. Hayes BJ, Goddard ME. Technical note: prediction of breeding values using marker-derived relationship matrices. Journal of Animal Science 2008;86:2089–92.

g

99. Zhong SQ, Dekkers JCM, Fernando RL, Jannink JL. Factors affecting accuracy from genomic selection in populations derived from multiple inbred lines: a barley case study. Genetics 2009;182:355–64.

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Natural antimicrobials for food processing Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso* Address: Department of Food Technology, University of Lleida – Av. Alcalde Rovira Roure, 191. 25198, Lleida, Spain. *Correspondence: Olga Martı´ n-Belloso. Email: [email protected] 29 April 2010 16 July 2010

Received: Accepted:

Abstract Natural antimicrobial agents are required to ensure that manufactured foods remain safe and unspoiled. This paper will focus on the deﬁnition, types, mechanisms of action and factors that affect the activity and application potential of food-natural antimicrobials for food processing. Naturally occurring antimicrobial compounds derived from animal, plant and microbial sources have potential for use alone, as adjuncts to traditional, regulatory-approved antimicrobial preservatives, or in combination with other food preservation processes. Among the many natural antimicrobials discussed in this paper, only a few have been tested or applied to foods. Some examples include the antibiotic natamycin, the bacteriocin nisin, lactoferrin, hydrogen peroxide, polysaccharides as chitosan and enzymes as egg-white lysozyme. This review will focus on the effects and efﬁcacy of the different compounds in real foods and food-system models, the mechanisms of action against micro-organisms, the interactions with food components and other preservative systems, as well as their inﬂuence on food quality, toxicology and safety. Keywords: Natural antimicrobials, Natural preservatives, Food additives, Plant-derived antimicrobials, Animal-derived antimicrobials, Antimicrobial agents Review Methodology: We searched the following databases: Scopus, Food Science and Technology Abstracts, Web of Science (keyword search terms combined: natural antimicrobials, natural food preservatives and mechanisms of action, food). In addition, we used the references from the articles obtained by this method to check for additional relevant material.

Introduction Following harvest, slaughter or manufacture, all foods lose quality at a rate that varies in a manner that is very dependent on food type, composition, formulation (for manufactured foods), packaging and storage conditions [1]. A number of non-conventional preservation techniques are under current development to satisfy consumer demands with regard to nutritional and sensory aspects of foods [2]. Micro-organisms present on or in foods may be inactivated by traditional processing methods such as heat or irradiation (e.g. gamma, electron beam, X-ray and ultraviolet) and novel methods such as high pressure or pulsed electric ﬁelds. Inhibition processes include chilling or freezing, drying/reduced water activity, modiﬁed atmosphere packaging, reduced pH/acidiﬁcation [3] or the addition of natural antimicrobials, which is the subject of this paper.

g

Antimicrobials used in food processing may be deﬁned as chemical compounds present in or added to foods, food packaging, food contact surfaces, or food processing environments that inhibit the growth or inactivate, pathogenic or spoilage micro-organisms. Antimicrobials are sometimes called ‘preservatives’; however, the latter term is more general and deﬁned as chemical agents that act as antimicrobials, antioxidants or antibrowning (e.g. prevention of enzymatic browning) agents [4]. Antimicrobial compounds can be added to either unprocessed or processed foods in order to extend their shelf life by reducing the microbial growth rate or viability [5]. Consumers perceive that the use of industrially synthesized food antimicrobials may be associated with potential toxicological problems, which has attracted interest in the food industry in the use of naturally occurring compounds. It is important to note, however, that commonly used antimicrobials, such as organic acids, that are

CAB International 2010

212

Animal Science Reviews 2010

routinely produced in large quantities through chemical synthesis, are also found naturally in many food products, and their toxicological safety as food additives is ensured by regulatory authorities. The extraction of these and other antimicrobials from natural sources, however, can be complex, inefﬁcient and expensive. Originally, spices and herbs were added to change or to improve ﬂavour. Some of these substances also exhibit certain antimicrobial activities, thus contributing to the self-defence of plants against infectious organisms [6]. This increasing demand for natural products has opened new dimensions for the use of natural preservatives derived from plants, animals or microbiota. Herbs, spices, fruits, milk, eggs and lactic acid bacteria (LAB) used in food fermentation are some of the most remarkable sources of antimicrobials extracts. However, the selection, manufacture and commercial application of a proper antimicrobial are challenging because of the complexity of food, the variety of factors inﬂuencing preservation and the complex chemical and sensory properties of natural antimicrobials [3, 7]. Numerous studies have evaluated the potential of naturally occurring food antimicrobials. Those with the greatest potential for commercial application are discussed here. While most compounds are not approved by regulatory agencies for their use as food additives, some of them have received approval, or are under current evaluation by governmental agencies.

Types of Natural Antimicrobials The classiﬁcation of antimicrobials is extremely complex. They can be divided into traditional and novel compounds (this classiﬁcation often conforms to the division between chemically synthesized compounds and those obtained from natural sources, regardless of whether they are novel or not). Nevertheless, this classiﬁcation does not imply that a ‘traditional’ preservative is less effective than one of natural origin from a microbiological point of view. Antimicrobials are called traditional when: (1) they have been used for many years, (2) they have been approved by many countries for inclusion as antimicrobials in foods, or (3) they have been produced by chemical synthesis. Ironically, many synthetic traditional antimicrobials are found in nature. This is the case of benzoic acid (in cranberries), sorbic acid (in rowanberries), citric acid (in lemons), malic acid (in apples) or tartaric acid (in grapes) [8].

Antimicrobials of Animal Sources Some antimicrobials from animal sources are shown in Table 1. Many of the antimicrobial agents inherent to animals are in the form of antimicrobial polypeptides. Over the past 20 years, more than 700 antimicrobial peptides (AMPs) have been discovered [41]. These peptides arise from virtually all species, from bacteria to human beings,

g

and possess a broad spectrum of antimicrobial activity, e.g. against bacteria, fungi, viruses and eukaryotic parasites [42–44]. These AMPs form an essential part of the ‘innate’ arm of host resistance, serving as a ﬁrst line of defence against infection. Importantly, AMPs are believed to have an entirely distinct mechanism of action from those of clinically used antimicrobial agents. Therefore, there has been great interest in the development of AMPs for the treatment of drug-resistant infections [44]. Recently, some AMPs of marine origin rather than of terrestrial origin have been identiﬁed. These AMPs are mainly from ﬁsh milt, crab, frog and include clupeines, tachyplesin, salmine, magainin and iridine and skin secretions such as pleurocidin from winter ﬂounder [9, 15, 16, 45, 46]. Crustacean tissue has also been reported to have AMPs. Some of the potential antimicrobials of animal origin, which could be used as food additives, are discussed below. Pleurocidin Pleurocidin and dermaseptins are a-helical AMPs isolated from winter ﬂounder and frog skin, respectively. When used at its minimal inhibitory concentration (MIC), the hybrid of pleurocidin and dermaseptin, P-der, inhibits Escherichia coli growth, but does not cause bacterial death within 30 min, and demonstrates a weak ability to permeabilize the bacterial membrane [9]. When used at 10 times the MIC, the peptide causes rapid depolarization of the cytoplasmic membrane and cell death, indicating that the cell membrane is a target for the peptide applied at high concentrations [9]. The fact that pleurocidin was found to be active against Gram-positive organisms, such as Staphylococcus aureus, Lactobacillus alimentarius and Listeria monocytogenes, Gram-negative organisms, such as E. coli, Salmonella typhimurium and Vibrio species, and against yeasts and moulds of food safety importance, indicates the versatility of this AMP as a potential food preservative [10]. Defensins Defensins are small bacterial cationic peptides that are quite abundant in the granules of polymorphonuclear neutrophils (PMNs). Studies have shown that human PMNs have four defensins (HNP-1 to HNP-4) and are responsible for the defence of the organism. The antimicrobial activities of several animal defensins, such as dermaseptin, antileukoprotease, protegrin and others, have been demonstrated, especially against bacteria, fungi and protists; there are also speciﬁc defensins from invertebrates, plants and bacteria [11, 12]. These have been reported to exhibit a broad spectrum of antimicrobial activity, including Gram-positive, Gram-negative bacteria, fungi and enveloped viruses [13, 14, 23, 47]. Protegrim, tachyplesin, protamine, salmine, iridine and others Peptides found in some animals may also affect microorganisms; for instance, protegrin, a porcine defensin, acts

CAB International 2010

Table 1 Some natural antimicrobial agents found in animal sources Origin

Antimicrobial agent (Reference)

Animal

Antimicrobial spectrum

Mechanism of action

Pleurocidin [9, 10]

Twenty-ﬁve amino acid peptides

Skin mucus membrane of the winter ﬂounder (Pleuronectes americanus)

Active against Gram-positive and Gram-negative bacteria

Defensins [11–14]

Cationic peptides

Mammalian neutrophils

Protegrim [14–16]

Cationic peptides

Porcine

Gram-positive and Gramnegative bacteria, fungi and enveloped viruses Escherichia coli, Listeria monocytogenes and Candida albicans

Strong membrane translocation and pore-formation ability, reacting with both neutral and acidic anionic phospholipid membranes Form voltage-gated channels

Tachyplesin [11, 13, 14] Protamine [16, 17]

Cationic peptides

Horseshoe crab (Tachypleus tridentatus) haemocyte Mice, humans and certain ﬁsh

Gram-negative and Gram-positive bacteria Bacteria, yeasts and moulds

Haemoprotein

Raw milk, colostrum, saliva and other biological secretions

Bacteria

Lactoferrin [23–29]

Globular multifunctional protein

Milk, animal tissue

Bacteria

Chitosan [30–40]

Linear polysaccharide of randomly distributed b-(1–4)-D-glucosamine and N-acetyl-D-glucosamine

Shellﬁsh waste

Fungi and yeasts

CAB International 2010

Enzymes: Lactoperoxidase [18–22]

Arginine-rich, nuclear proteins

Membrane permeabilizing/disrupting properties, Interactions with intracellular targets or disruption of key intracellular processes Permeabilizes the outer membrane and cytoplasmic membrane Disrupt the electron transport chain and oxidative phosphorylation of cellular membrane Oxidation of thiols groups of cytoplasmic enzymes and damage cell well or cytoplasmic membrane, transport systems Plays a central role in ferrokinetics: it binds free iron with great afﬁnity limiting the amount of ions available for micro-organism’s metabolism Permeabilization of the cell membrane to small cellular components, coupled to a signiﬁcant membrane depolarization interference with cell wall biosynthesis

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso

Source isolated

g

Major component or structure

213

214

Animal Science Reviews 2010

against E. coli, L. monocytogenes and Candida albicans. These compounds inhibit the growth of pathogenic microorganisms, without affecting the tissues of the host that produces them, and have a broad antimicrobial spectrum. It is well known that, under stress conditions, some bacteria produce bacteriocins that may cure infectious diseases [11]. Other peptides, such as tachyplesin protamine, iridine salmine and clupeine, have been effective against bacteria, yeasts and moulds [15–17, 48]. Magainin peptides isolated from frogs have been found to be effective against an array of food-borne pathogens [46, 49], implying a possible application as food antimicrobial [50–52]. Enzymes b-glucanases, which hydrolyse the b-glucan structure of fungal walls as well as oxidoreductases, such as lactoperoxidase, glucose oxidase and catalases, which catalyse reactions producing cytotoxic compounds such as H2O2, OSCN7 and OCI7 have a potential to inactivate microorganisms. Lactoperoxidase is a haemoprotein present in raw milk, colostrum, saliva and other biological secretions. Bovine milk naturally contains 10–60 mg of lactoperoxidase per litre [18, 19]. Lactoperoxidase catalyses the oxidation of thiocyanate (SCN7) and iodide ions to generate highly reactive oxidizing agents. The generated products have antimicrobial effects against bacteria, fungi and viruses [20]. In 2002, the Food Standards Australia New Zealand (FSANZ) permitted the use of the lactoperoxidase system (LPS) for the treatment of meat (including poultry), ﬁsh and milk products (infant formula, ice cream, cream and cheeses) at maximum levels of 20 mg/kg meat or 30 mg/l milk [21]. In 2006 [22], the US Food and Drug Administration (FDA) included lactoperoxidase in the list of generally recognized as safe (GRAS) substances, thus allowing its use as an ingredient in several foods, including dairy products (up to 1000 mg/kg litre), and fruit and vegetable juices (up to 167 mg/l). On the other hand, other enzymes such as lysozyme have a great potential to act as antimicrobial agents in foods. Lysozyme is a protein present in mammalian milk and avian eggs that catalyses the hydrolysis of b-1,4 linkages between N-acetylmuramic acid and N-acetylglucosamine in the peptidoglycan layer of the bacterial cell wall [4]. It is used to inactivate pathogenic and spoilage micro-organisms in fruit juices. Lysozyme (E1105) is the only antimicrobial enzyme that has achieved signiﬁcant commercial application [53, 54]. It is used in Japan to preserve seafood, vegetables, pasta and salads. It has been evaluated for its use as an antimicrobial in wines to inhibit LAB and as a component of antimicrobial packaging [19, 55]. Lysozyme is generally active against most Grampositive bacteria, particularly thermophilic spore formers [4]. Gram-positive bacteria are more susceptible to lysozyme than Gram-negative bacteria because of the different contents of peptidoglycan in their cell walls [54]. In 2000, lysozyme was considered as GRAS by the USFDA.

g

Lactoferrin It is the primary iron-binding protein in milk and colostrum, with two binding sites per molecule. Lactoferrin has antimicrobial activity against a wide range of Grampositive and Gram-negative bacteria [23] fungi and parasites [24]. Some Gram-negative bacteria may be resistant, because they adapt well to poor iron environments by producing siderophores, such as phenolates and hydroxamates [18]. Micro-organisms with a low iron requirement, such as LAB, would not be inhibited by lactoferrin. Since it is cationic, lactoferrin may increase the outer membrane permeability to hydrophobic compounds, including other antimicrobials. Lactoferrin has been applied in meat products [25–27] as it has recently received approval for application on beef in USA [28]. Lactoferrin B or hydrolysed lactoferrin (HLF) is a small peptide produced by acid-pepsin hydrolysis of bovine lactoferrin [29]. Polysaccharides Chitin is the second most abundant polysaccharide found on earth next to cellulose. It is the main component in the shells of crustaceans, such as shrimp, crab and lobster. Chitosan is a heteropolysaccharide obtained commercially by deacetylation of chitin. It is produced commercially from crab and shrimp shell wastes with different degrees of deacetylation and molecular weights, thus presenting a variety of properties and is active against yeasts and bacteria [30–35]. Chitosan is considered a biocompatible, non-antigenic, non-toxic and biofunctional food additive [36, 37]. In addition, shrimp-derived chitosan was admitted as GRAS in 2005 by the US FDA in 2007 [38, 39]. Chitosan has several biological properties that make it useful for the food industry, but the most attractive is its potential use as a food preservative of natural origin because of its antimicrobial activity against a wide range of food-borne micro-organisms [40]. Lipids Free fatty acids (FFA) can be released from lipids, typically by enzyme action, and show diverse and powerful biological activities. FFAs may come from the breakdown of a triglyceride into its components (fatty acids and glycerol). When they are not attached to other molecules, they are known as ’free fatty acids’ [56, 57]. FFAs are the most active antimicrobial agents present in human skin lipid samples [58]. There are about 10–15 mg of FFAs/cm2 on human skin, of which lauric acid (C12 : 0), myristic acid (C14 : 0), palmitic acid (C16 : 0), sapienic acid (C16 : 1n-10) and cis-8-octadecenoic acid (C18 : 1n-10) are the most abundant [58]. Some lipids may exhibit antimicrobial properties with interesting potential applications on food preservation. Currently, extensive literature exists concerning the antibacterial effects of various FFAs from a wide range of biological sources, including algae, animals and plants [58, 59]. The in vitro antibacterial activities of FFAs are typically a broad spectrum and have been shown to be comparable to those of AMPs [56]. Medium- and

CAB International 2010

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso

long-chain unsaturated FFAs tend to be more active against Gram-positive bacteria than against Gram-negative bacteria.

Mechanism of Antimicrobial Action There is now a widespread acceptance that AMPs, apart from their membrane-permeabilizing and disrupting properties, may also affect microbial viability by interactions with intracellular targets or disruption of key intracellular processes [60–62]. Research on this area has focused on the identiﬁcation of targets in the interior of the microbial cell and the mechanism by which AMPs can enter the microbial cell in a non-disruptive way [63]. However, it is unlikely that the speciﬁc abilities of some AMPs to enter microbial cells and impede cellular functions are also shared by the hundreds of AMPs that differ in length, amino acid composition, sequence, hydrophobicity, amphipathicity and membrane-bound conformation [57, 64]. From the limited data currently available, the speciﬁc translocating properties of some AMPs are likely to be speciﬁc and limited to particular peptide families [57]. Although their antibacterial mode of action is still poorly understood, the prime target of the peptides action is the cell membrane, where they induce disruption of the electron transport chain and oxidative phosphorylation. Besides interfering with cellular energy production, both the fatty acid and the peptide antimicrobial action may result from the inhibition of enzyme activity, impairment of nutrient uptake, generation of peroxidation and auto-oxidation degradation products or direct lysis of bacterial cells [57, 65, 66]. Monoacylglycerols decrease the heat resistance of certain bacteria and fungi; therefore, they may ﬁnd application in reducing the required heat treatment for certain foods [7]. Lysozyme has been used for several decades as an antimicrobial. Studies have shown that lysozyme effectively lyses the walls of Gram-positive bacteria by hydrolysis of b-1,4-glycosidic linkage in sugar polymers such as N-acetylmuramic acid and N-acetylglucosamine linkages found in bacterial peptidoglycan. In addition, it has exhibited a good effect on some fungi, and it is seen as an ‘endogenous antibiotic’ [67].

Factors Affecting Antimicrobial Activity Most studies related to the antimicrobial efﬁcacy of compounds obtained from animal sources have been conducted. In the case of chitosan, the minimum inhibitory concentrations for both bacteria and yeasts vary widely from 0.01 to 5.0% depending on polymer characteristics, pH, temperature and the presence of interfering substances, such as proteins and fats [30, 68, 69]. The addition of chelating agents such as EDTA enhanced the activity of hydrolysate lactoferrin (HLF) in trypticase soy

g

215

broth (TSB), indicating that the decreased activity of HLF may have been due, in part, to an excess of cations in the medium [70]. Consistently, Venkitanarayanan et al. [71] found that lactoferricin B was much less effective as an antimicrobial in ground beef. On the other hand, processing conditions can also be a limiting factor for the effectiveness of natural antimicrobials. For instance, inactivation of lactoperoxidase occurs at 80 C in 15 s, whereas residual lactoperoxidase activity is detected following treatment at 72 C [18, 21]. The Gram-negative cell susceptibility to lysozyme can be increased by combination with other antimicrobials. The activity of lysozyme was enhanced against the Gram-positive organisms in the presence of EDTA under conditions in which growth was not restricted by starvation [53]. There was no protective effect of nitrite or NaCl upon the inhibitory action of lysozyme, nisin and EDTA. The addition of lysozyme, nisin and EDTA to cured meat products containing NaCl and nitrite should result in a wider spectrum of antimicrobial activity [53].

Antimicrobials of Plant Source The main antimicrobials from plant sources are shown in Table 2. Plant extracts Antimicrobial activities of several plants used today as seasoning agents in foods and beverages have been recognized for centuries. Plants, herbs and spices, as well as their derived essential oils and compounds, contain a large number of substances that are known to inhibit bacteria, moulds and yeast [100, 101]. Many of these compounds are under investigation and are not yet exploited commercially. The antimicrobial compounds in plant materials are commonly found in the essential oil fraction of leaves (rosemary, sage, basil, oregano, thyme and marjoram), ﬂowers or buds (clove), bulbs (garlic and onion), seeds (caraway, fennel and parsley), rhizomes (asafoetida), fruits (pepper and cardamom) or other parts of plants [2, 99, 102, 103]. Different studies have demonstrated the effectiveness of EOs and their active compounds to control or inhibit the growth of pathogenic and spoilage microorganisms in both fresh-cut fruit and fruit juices [54, 104–108]. Generally, the oils consist of a mixture of esters, aldehydes, ketones, terpenes and phenolic compounds and harbour the characteristic ﬂavour and aroma of the particular spice or herb [109]. The traditionally best-known antimicrobial herbs and spices include clove, cinnamon, chilli, garlic, thyme, oregano and rosemary. These compounds are generally more inhibitory against Gram-positive than against Gram-negative bacteria [110–116]. While this is true for many EOs, there are some agents that are as much effective against both groups, such as oregano, clove, cinnamon and citral [72–74, 117–120]. Their antimicrobial activity depends

CAB International 2010

216

Origin Plant

Antimicrobial agent (reference)

Major component or structure

Essential oils (EOs): Citral [72–74]

Antimicrobial spectrum

Mechanism of action

Citral, geraniol, limonene

Lemongrass (Cymbopogon citretus)

Fungi, bacteria

Allyl isothiocyanate (AIT) [75–77]

AIT

Mustard (Brassica hirta) and onion (Allium cepa)

Carvacrol [78–81]

Carvacrol, thimol, a-pinene, p-cymene Thimol, carvacrol, L-linalool, geraniol Hexanal and 2-(E)-hexenal

Oregano (Origanum vulgari)

Gram-positive and Gram-negative bacteria, fungi Fungi, bacteria

Inhibition of various metabolic activities of bacteria by several enzymatic systems Cellular activity affected by interactions with membranes, enzymes and proteins Increased membrane permeability

Thyme (Thymus vulgaris)

Fungi, yeast, bacteria

Increased membrane permeability

Olive oil

Fungi, bacteria

Vanillin, p-hydroxybenzoic acid Jasmonic acid and Methyl jasmonate

Vanilla (Vanilla planifolio, V. pompona) Flowers of jasmine (Jasminum grandiﬂorum)

Yeast, bacteria

Modiﬁcations of the composition of fatty acid of cell membrane and the volatile compounds produced Ability to alter microbial cell permeability

Fungi

Not reported

Cinnamon aldehyde, eugenol

Cinnamon (Cinnamomum zeylanicum)

Clove [95–98]

Eugenol

Clove (Syzygium aromaticum)

Oregano [95–97, 99]

Carvacrol, thymol

Oregano (Origanum vulgare)

Yeast, bacteria Modiﬁcation of microbial permeability (Staphylococcus, Listeria) and interaction with membrane proteins, causing deformation Fungi, yeast, bacteria Cellular membranes sensitized, (Bacillus subtilis) and sites saturated, causing a serious damage and rapid collapse of cytoplasmatic membrane integrity Fungi, bacteria Alteration of microbial permeability and interaction with membrane proteins, causing deformation

g

Source isolated

CAB International 2010

Thymol [78–81] Aldehydes: [82–88]

Esters [89–93] Herbs and spices: Cinnamon [94, 95]

Animal Science Reviews 2010

Table 2 Some natural antimicrobial agents found in plant sources

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso

not only on the extraction method but also on the initial concentration of the plant extract [102, 121–123]. It has been reported that, when applied as vapours, most compounds seem to exhibit a higher antimicrobial potential than as liquids [102]. Allyl isothiocyanate (AIT) [75] and garlic oil are more effective against Gram-negative bacteria and fungi [76, 77]. The major antibacterial components of these oils are carvacrol and its isomer thymol. Studies have shown that the essential oils of thyme and clove are effective against viruses, bacteria and fungi [78]. On the other hand, olive leaf extracts have been historically used as a folk remedy for combating fevers and other diseases, such as malaria. The bitter compound oleuropein (heterosidic ester of elenolic acid and dihydroxyphenylethanol), the major constituent of the secoiridoid family in olive (Olea europaea L.) trees, has been shown to be a potent antioxidant endowed with antiinﬂammatory properties. Oleuropein has antimicrobial activity against viruses, retroviruses, bacteria, yeasts, fungi, moulds and other parasites [79]. Also many species of the genus Allium have traditionally been used for various purposes. Garlic (Allium sativum L.), onion (Allium cepa L.) and leek (Allium porrum L.) are commonly used worldwide for food preparations and as seasoning agents. Their extracts exhibit antifungal activities against Aspergillus niger, Fusarium oxysporum, C. albicans ATCC 10231 and Metschnikowia fructicola. Amounts of organosulphur compounds are concentrated in dehydrated Allium species and therefore contain higher amounts of antifungal compounds than the fresh forms [80]. Other studies clearly indicate that white A. cepa (diethyl ether, water and methyl alcohol) and red A. cepa (diethyl ether, methyl alcohol, water and acetone) extracts were inhibitory against different micro-organisms (Enterobacter aerogenes (ATCC 13048), E. coli (ATCC 25922), Salmonella enteritidis (ATCC 13076), S. typhimurium (ATCC 14028), S. aureus ssp. aureus (ATCC 29213 and C. albicans) and Bacillus subtilis (test micro-organism of sterilization)). White A. cepa extract of ethyl alcohol at 800 mg/ml dose was found to be able to inhibit C. albicans [81]. Aldehydes Recently, a series of aliphatic long-chain aldehydes have been demonstrated to exhibit a remarkable activity against several food-borne microfungal and bacterial strains [124]. These compounds, produced throughout the lipoxygenase pathway, have an important role in plant defence, with a protective action towards microbial proliferation in wounded areas [82]. Lanciotti et al. [83] pointed out that these compounds are good candidates to improve the safety and quality of minimally processed fruits because of their effectiveness at low concentrations, the natural occurrence in several fruits, and the nonregulated doses, in addition to its GRAS status as ﬂavouring. Nevertheless, high concentrations of these compounds may produce an undesirable hay-like ﬂavour given the oxidative rancidity of fatty acids through the

g

217

lipoxygenase pathway, thus limiting their use as antimicrobials in some foods [84]. Aldehydes such as hexanal, and 2-(E)-hexenal have been demonstrated to possess a noticeable activity against several microfungal and Grampositive and Gram-negative bacterial strains. Therefore, these compounds might be good candidates for being used as antimicrobial agents against bacteria responsible for human infections [82, 83, 85]. Vanillin (4-hydroxy-3-methoxybenzaldehyde) is a phenolic aldehyde, which is the major constituent of vanilla beans, the fruit of an orchid (Vanilla planifola, Vanilla pompona or Vanilla tahitensis). It is most active against moulds, yeasts and non-lactic Gram-positive bacteria [86]. In addition, it is used as a ﬂavouring agent in a wide variety of foods such as baked products, ice creams and drinks [87]. However, its use as an antimicrobial in fruit juices and fruit juice-containing drinks may be limited, because of the formation of guaiacol (an ‘off-ﬂavour’ metabolic compound) catabolized from vanillin by several microorganisms including Alicyclobacillus acidoterrestris, Bacillus megaterium, Streptomyces spp. and Rhodotorula rubra [88]. It has been reported that vanillin may represent a useful tool to increase thermal susceptibility of Listeria innocua (as surrogate for L. monocytogenes), thus increasing the safety of minimally processed orange juice [125]. Vanillin has been demonstrated to be effective against both pathogenic and spoilage micro-organisms in fresh-cut fruit and fruit juices; however, the strong ﬂavour that effective concentrations (>0.2%) may impart to food products is one of the major drawbacks of its use. Therefore, combinations of preservation methods are required for decreasing its impact on the ﬂavour [126]. Esters Jasmonic acid and its volatile esteriﬁed derivative, methyl jasmonate (MeJA), are natural compounds derived from jasmine [89], a GRAS status substance [31], which is found as a lipid of plant cell membranes, synthesized via the lipoxygenase pathway [90]. Jasmonic compounds have been shown to elicit defence responses from the plant as well as antibacterial and antifungal activities, regulate plant growth, promote the closing of stomas, act as second messenger and decrease the pathogen’s attack [91]. No references reporting the use of MeJA in fresh-cut fruits to control pathogenic micro-organisms are available, although different studies have been published about the effectiveness of MeJA to reduce the spoilage of whole products of plant origin [91]. In a recent research [92], it has been reported that a post-harvest MeJA treatment with vapour maintained higher levels of sugars and organic acids in fresh-cut kiwifruit and was effective for preventing mould growth. Likewise, the application of MeJA, either by vapour or by dipping, reduced microbial growth compared with control groups in pineapples. The application of MeJA may be practically implemented at an industrial level to enhance the quality and the safety of fruits [93]. Thus, future studies are needed to examine and to

CAB International 2010

218

Animal Science Reviews 2010

determine the mechanism by which MeJA reduces microbial growth as well as which are the effective concentrations for the treatment of different food products. Herbs and spices Herbs as well as spices have varying degrees of antimicrobial activity. Among spices, cloves, cinnamon, oregano, thyme, sage, rosemary, basil and vanillin have strongest antimicrobial activity. The genus Mentha of the family Lamiaceae, comprises more than 25 species, with three Mentha species, Mentha piperita L. (peppermint), Mentha arvensis L. (cornmint) and Mentha spicata L. (spearmint), which are herbs commonly cultivated worldwide. These species are used extensively in the liquor and confectionary industries, ﬂavouring, perfumes production and medicinal purposes. Since they are a rich source of polyphenolic compounds and other volatile compounds, they could possess strong antimicrobial and antioxidant properties [127]. However, studies about their sensory impact should be carried out. The addition of cinnamon to apple juice, both at natural and adjusted pH, substantially reduced (4.6-log CFU/ml inactivation) L. monocytogenes counts inoculated in the juice (roughly 104 CFU/ml) without letting the microorganism grow throughout storage [94]. Other herbs and spices with antimicrobial properties include anise, bay (laurel), black pepper, cardamom, cayenne (red pepper), celery seed, chilli powder, coriander, cumin, curry powder, dill, fenugreek, ginger, juniper oil, mace, marjoram, nutmeg, orris root, paprika, sesame, spearmint, tarragon and white pepper [78].

Mechanisms of Antimicrobial Action The effectiveness of phenolic compounds has been shown to be dependent on the applied concentration [94]. At low concentrations, phenols affect enzyme activities, particularly those associated with energy production, while at high concentrations, they cause protein denaturalization. Certain small terpenoids and phenolic compounds found in herbal plants (for example, essential oil components) have been reported to disintegrate the outer membrane of Gram-negative bacteria, releasing lipopolysaccharides and increasing the permeability of the cytoplasmic adenosine triphosphate (ATP). The exact mechanism of action is not clear. Phenolic compounds could also interfere with membrane function (electron transport, nutrient uptake, protein, nucleic acids synthesis and enzyme activity) and interact with membrane proteins, causing deformation in structure and functionality [128, 129]. The antimicrobial activity of essential oils on yeasts could be the result of disturbance in several enzymatic systems involved in energy production and components synthesis. Once the phenolic compound crosses the cellular membrane, interactions with membrane enzymes and proteins would cause an opposite ﬂow of

g

protons, affecting cellular activity [95–97]. On the other hand, aldehydes would react with nucleophilic groups playing a key role in living cells, namely sulphydryl groups present in proteins and lower-molecular-weight compounds such as glutathione. Although the precise targets in microbial cells remain unclear, the toxicity of these molecules seems to be dependent on their afﬁnity to the membrane phospholipidic bilayer [83].

Factors Affecting Antimicrobial Activity Many of the antimicrobial components of herbs and spices are hydrophobic. Therefore, the methods based on agar diffusion to asses their antimicrobial activity are not appropriate. Nevertheless, reports based on such methods persist in the scientiﬁc literature. Furthermore, factors such as temperature, pH and the presence of fats and proteins, surfactants, minerals and other food components can greatly inﬂuence the outcome of antimicrobial testing and render difﬁcult any predictions about applicability in foods. The time of harvesting, stage of development and method of extraction of herbs and spices are also very important. The lipophilicity of many phenolic compounds such as cinnamic acid and carvacrol plays an important role in determining their antimicrobial efﬁcacy. [98, 99, 130, 131]. Composition, structure as well as functional groups in the essential oils are also important in determining their antimicrobial activity. Usually compounds with phenolic groups are the most effective [5, 120]. Interactions between phenolic groups and proteins, lipids and aldehydes could explain, at least partially, the reduction of the antimicrobial effect of essential oils, which are mainly constituted by phenols [132, 133].

Antimicrobials of Microbial Source Table 3 introduces some antimicrobials obtained from microbial sources. Antimicrobial substances produced by food-related micro-organisms have been recognized for many years [151]. However, only recently, these substances have been allowed for use as food additives. A new generation of AMPs, bacteriocins, has been isolated from a full range of micro-organisms, especially from both Gram-negative and Gram-positive bacteria. These compounds are active against a large spectrum of microorganisms, including bacteria and ﬁlamentous fungi in addition to protozoan and metazoan [134]. Bacteriocins are proteinaceous toxins produced by bacteria to inhibit the growth of similar or closely related bacterial strain(s). They are categorized in several ways, including producing strain, common resistance mechanisms or mechanism of killing. There are several large categories of bacteriocins, which are only phenomenologically related. These include bacteriocins from Gram-positive and Gram-negative bacteria and bacteriocins from Archaea [135, 151]. Their

CAB International 2010

Streptomyces spp.

Bacteria, H2O2-generating oxidases

Macrolide antibiotic

Weak acid

Antibiotic: Natamycin [151]

Other fermentation products: Hydrogen peroxide [1, 152, 153]

Bacteria, yeast and fungi

Binding to sterols in the fungal cell membrane to produce a change in membrane permeability that allows loss of essential cellular constituents Oxidative decarboxylation of pyruvic acid involving DNA damage

Pore complex formation in target cell membranes Lantibiotic peptides

Cationic peptide

Pediocin [135, 140, 141] Nisin [140, 142–150]

Pediococcus acidilactici and P. pentosaceus Lactococcus lactis

In the presence of glycerol Lactobacillus reuteri b-hydroxypropionaldehyde Reuterin [136–139]

Gram-positive bacteria, bacterial spores Fungi

Permeabilization of the target-cell membrane

Inhibition of the activity of bacterial ribonucleotide reductase, an enzyme catalysing the ﬁrst step in DNA synthesis bacterial

Lactobacillus sakei Cationic, hydrophobic peptides Bacteriocins: Sakacins [20, 134, 135] Microbial

Source isolated Major component or structure Antimicrobial agent (Reference) Origin

Table 3

Some natural antimicrobial agents found in microbial sources

Antimicrobial spectrum

Active against species of lactobacilli and L. monocytogenes Gram-positive and Gram-negative bacteria, yeasts, moulds and protozoa Gram-positive

Mechanism of action

Pore formation in the cytoplasmic membrane

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso

g

219

effectivity as natural antimicrobials in food has been extensively demonstrated. Bacteriocins are ribosomally synthesized peptides or proteins with antimicrobial activity, produced by different groups of bacteria. They are characterized as cationic and hydrophobic substances, some of them presenting a disulphide bridge in their molecular structure. Most of them are small molecules, with 50 or less amino acids. Nisin, the most studied bacteriocin, has 34 amino acids. In recent years, bacteriocins have attracted interest for their potential as non-toxic antimicrobials, and a large number of chemically diverse compounds have been identiﬁed and characterized [140, 154]. Bacteriocins exhibit stability to heat, low pH, refrigeration and freezing temperatures and also to various weak organic solvents, salts and enzymes, but are sensitive to proteolytic enzymes [155]. On the other hand, they form a heterogeneous group with regard to the producing species, molecular size, antibacterial spectrum, stability, physical and chemical properties and action mode. The classiﬁcation of bacteriocins is currently being revised to reﬂect similarities and differences observed in the discovery of new molecules. As a function of the producing organism and classiﬁcation criteria, bacteriocins can be categorized into several classes [140, 154, 155]. Most of them fall into classes I and II, which are the most intensively researched to date. Class I peptides typically have from 19 to more than 50 amino acids. Nisin is a class I bacteriocin formed by cationic and hydrophobic peptides that account for the formation of pores in target membranes. Another group of class I bacteriocins contain globular peptides, whereas Pediocin PA-1 represents class II bacteriocins, which are small heat-stable peptides. Class II bacteriocins can be further subdivided. Bacteriocins in the subclass IIa (pediocin-like bacteriocins) contain a consensus motif (YGNGV), are active against Listeria spp. and comprise the largest and most studied group. Bacteriocins composed of two different peptides comprise Class IIb, needing both peptides to be fully active. Class IIc was originally proposed to contain the bacteriocins that are secreted by the general sec-system. The large and heat-labile bacteriocins make up the Class III bacteriocins for which there is much less information available. A fourth class consists of bacteriocins that form large complexes with other macromolecules [140, 156, 157]. Although many other bacteriocins have been reported to have potential for food preservation and safety applications, nisin and related compounds such as pediocin are the only bacteriocins widely used for food preservation purposes [158–161]. Nisin has been sufﬁciently well characterized to be used for this purpose. Sakacins These bacteriocins are produced by several strains of Lactobacillus sakei. They are small, cationic, hydrophobic peptides with an antibacterial mode of action against species of lactobacilli and certain food-borne pathogens such as L. monocytogenes. Environmental conditions,

CAB International 2010

220

Animal Science Reviews 2010

particularly the pH, affect sakacin stability and efﬁcacy. Applications are predominantly in the ﬁeld of sakacinproducing starter or co-cultures for meat fermentations since the addition of puriﬁed sakacin is not yet accepted in food additives regulations. Studies on physical and chemical properties, in addition to structure–function relationships of sakacins and similar compounds, as well as on their potential as food additives or as starter cultures, are necessary if their use for meat preservation is to be exploited [20]. Reuterin is a broad-spectrum antimicrobial produced by Lactobacillus reuteri (a normal inhabitant of the human intestine) during anaerobic fermentation of glycerol [162]. It is active towards enteropathogens, yeasts, fungi, protozoa and viruses [136]. Talarico and Dobrogosz, in 1989, isolated, puriﬁed and identiﬁed this antimicrobial agent [137]. Reuterin is a water-soluble substance, active at a wide range of pH values and resistant to proteolytic and lipolytic enzymes [138]. A dose of 8 AU/ml has been shown to present bacteriostatic activity against L. monocytogenes, whereas its activity is slightly bactericidal against S. aureus at 37 C. In contrast, it exhibits bactericidal activity against all Gram-negative pathogens [139]. Pediocin is a bacteriocin produced by LAB, generally Pediococcus acidilactici and Pediococcus pentosaceus, that has found practical applications as a food antimicrobial. Pediocins are small heat-stable peptides, synthesized in a form of a precursor which is processed after two glycine residues, are active against Listeria and have a sequence of YGNGV-C in the N-terminal [140, 141]. This AMP is recognized as GRAS. Investigations have revealed that Pediococcus spp. does not have application in cheese starter cultures. Therefore, the plasmid-encoding pediociin was expressed in Lactococcus lactis to aid in the preservation of Cheddar cheese and to assure the microbial quality of the fermentation process [140]. Most pediocins are thermostable proteins and function over a wide range of pH values. Pediocin A exhibited inhibition against species of Lactobacillus, Lactococcus, Leuconostoc, Pediococcus, Staphylococcus, Enterococcus, Listeria and Clostridium [163]. Nisin is commercially used as a natural agent for food biopreservation and considered safe [140]. It is produced by L. lactis and is used in approximately 40 countries worldwide for several applications in food and beverages. Unfortunately, nisin is not suitable for use in all food systems because of its limited solubility above pH 5. It also has a limited spectrum of activity and does not inhibit Gram-negative bacteria or fungi [140, 142–144]. The application of nisin as a food antimicrobial has been studied extensively in different food products (beef, sausages, liquid whole egg, ground beef, meat products and poultry), based upon target micro-organisms [145– 150]. Nisin was effective in fruit juices for the control of A. acidoterrestris at levels as low as 5 IU/ml that were effective under conditions that simulated normal commercial practice most realistically [164]. Furthermore, 6.25 mg/g and 25 mg/g of nisin could inhibit LAB growth in a

g

Bologna-type sausage for over 28 days and 35 days, respectively [165]. Natamycin (piramicin) is a macrolide antibiotic with a molecular weight of 665.7 Da, containing a large lactone ring, few double bonds and no nitrogen. The ring is substituted with one or more sugar residues. It is produced by the bacterium Streptomyces natalensis. Natamycin is effective primarily against yeasts and moulds and ineffective against bacteria, viruses, protozoa and actinomycetes. The effective concentration of natamycin is 5–10 mg/ml against most yeasts and moulds [151, 152]. Because of its antifungal properties, it is used on the surface of cuts and sliced cheese. It is believed that natamycin acts by binding to ergosterol in fungal cell membranes, which leads to increased cell permeability and usually results in cell death. Hydrogen peroxide is naturally produced in organisms as a by-product of oxygen metabolism. Nearly all living things possess enzymes known as peroxidases, which harmlessly and catalytically decompose low concentrations of hydrogen peroxide to water and oxygen. It is a GRAS substance for use in food products as a bleaching, oxidizing, reducing and antimicrobial agent [166], and has been used for many years for microbial control in milk, cheese and whey processing [167]. For these and other food applications, the FDA regulations have speciﬁed that residual H2O2 should be removed. In most processing situations, the effectiveness of H2O2 to kill or inhibit microbes very much depends on the pH, temperature and other environmental factors [153].

Mechanism of Antimicrobial Action The biological action of bacteriocins occurs through the speciﬁc receptors located on the target microbial cell surface. The pore formation in the cytoplasmic membrane of the target micro-organism induces microbial cell death through various mechanisms isolated or concomitant [168]. Microbial cell killing through bacteriocin action could occur as consequence of unbalanced cytoplasmic membrane function (affecting energy synthesis and permeability), inhibition of nucleic acid synthesis, interference on the protein synthesis and changing cell translation mechanism [155]. The 34-amino-acid peptide nisin is highly active against Gram-positive bacteria. This high activity is a result of a combination of pore-formation and a high-afﬁnity interaction with Lipid II, which is a membrane-anchored cell-wall precursor essential for bacterial cell-wall biosynthesis. Nisin inhibitory action on Gram-positive micro-organisms occurs through two stages. First, it involves an unspeciﬁc interaction between nisin and the target microbial plasmatic membrane, which is characterized as a reversible and pH-dependent phenomenon [168, 169]. During the second stage, a strong attraction of the nisin with the target cytoplasmic membrane occurs. Nisin initially interacts with Lipid II via its N-terminus. Then nisin assembles into a pore thereby

CAB International 2010

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso

inducing leakage of cytosolic contents, with channels of diameter 0.1–0.2 mm. The simultaneous membrane depolarization provokes a quick ﬂux of essential molecules (K+ ions, amino acids and ATP), leading to alterations that culminate with the cell lysis [34, 170, 171]. Evidence has been provided that reuterin induces oxidative stress in cells, most likely by modifying thiol groups in proteins and small molecules [139].

Factors Affecting Antimicrobial Activity The chemical composition and the physical conditions of food can have a signiﬁcant inﬂuence on the activity of bacteriocins. Nisin, for example, is 228 times more soluble at pH 2 than at pH 8 and its spectrum of activity can be expanded to include Gram-negative bacteria when it is used in combination with chelating agents (e.g. EDTA), heat or freezing [172, 173]. The presence of food components such as lipids and proteins may have an inﬂuence on the antimicrobial activity of bacteriocins. Also the emergence of bacteriocin-resistant bacteria and inactivation by other additives should be considered [134]. In general if a compound is to be useful as a natural food antimicrobial, it must act in a food system as a function of its composition. Because food matrices are usually very complex and can be substantially affected by microbial, food-related (intrinsic), environmental (extrinsic) and process conditions, this could result in a decrease or increase in the efﬁcacy of the bacteriocin. Hence, it is somewhat difﬁcult to generalize the methods and doses of application of these compounds.

Potential Application of Natural Antimicrobials for Food Processing In an attempt to satisfy the rapidly changing consumption patterns of the global market, the food processing industry is continuously developing various new and modiﬁed products. Such product development requires cuttingedge technology to ensure organoleptic distinction as well as other qualities such as wholesomeness, sanitary grade, shelf-life and safety. The development of convenience foods, functional diets and nutraceutical health supplements also require the use of additives and antimicrobials. For example combinations of natural antimicrobials and mild preservation techniques are attractive approaches to control pathogens and spoilage organisms in many foods. The addition of lysozyme to certain foods before heat processing may reduce the thermal requirements necessary to inactivate spores of some strains of thermophilic spore-forming spoilage bacteria that are particularly sensitive to lysozyme, such as Bacillus stearothermophilus and Clostridium thermosaccharolyticum [174]. The antimicrobial effectiveness of lysozyme and its commercial application as a food preservative need to be further investigated,

g

221

especially in refrigerated foods under conditions similar to actual processing and storage of products. Salt levels for optimum lysozyme activity need to be determined for individual food systems. Additional research is needed to verify that lysozyme does not interfere with the growth of beneﬁcial bacterial and yeast cultures in fermented foods. More research is also needed to develop optimal systems comprised of combinations of antimicrobials for maximal efﬁcacy. Many of the natural food antimicrobial systems also demonstrate a multifunctional physiological advantage and are emerging as value-added ingredients in various food products. Milk lactoferrin is now being explored as an antioxidant, an iron-delivery system and an immunomodulating agent in various sport drinks and infant formulations. Globulin-rich ingredients such as whey-protein concentrates and egg proteins seem to reduce gastric ulcers in controlled clinical trials. Among the plants widely consumed in the human diet, garlic, onion and leek have been recognized and studied extensively for their antimicrobial properties. Many food-borne bacteria are sensitive to onion and garlic extracts [122]. LAB are believed to induce a probiotic effect and thereby improve gut physiology that might even contribute to the prevention of intestinal carcinogenesis. Among the natural food antimicrobial systems used, a variety of compounds including oleoresins of spices, diacetyl and vanillin, antimicrobialcolorants such as yellow and orange from saffron and turmeric, green from mint, chlorophyll and parsley between others, could serve as dual-function food additive systems antimicrobial ﬂavorants including oleoresins of spices, diacetyl and vanillin. The general overview of potential interest for the application of natural food antimicrobial in food is large. The incorporation of these compounds into foods poses great technological and scientiﬁc challenges, given the shielding and complexing factors that may interact in a real food matrix.

Conclusion Naturally occurring antimicrobials are abundant in the environment and have great potential usefulness to the food industry. The need for increasing its use is enormous, particularly since consumers are increasingly demanding for minimally processed safe foods with an adequate shelf-life but also keeping nutritional quality and with added convenience. Some natural antimicrobials are currently used as excellent hurdles against food deterioration, but more systematic studies on multi-synergistic effects are scarce in real food systems, e.g. a combination of lactoferrin, organic acids and oregano extracts with modiﬁed atmosphere packaging and pulsed electric ﬁeld technology to prevent microbial growth or MeJA and ethylene, acting synergistically in the regulation of disease resistance and pathogen attack. Therefore, research, commercial development and economic production must occur. Moreover, more fundamental knowledge on the

CAB International 2010

222

Animal Science Reviews 2010

mechanisms of action of the natural antimicrobials is needed to establish an effective preservation strategy for each product type.

12. Zhao C, Nguyen T, Liu L, Sacco RE, Brogden KA, Lehrer RI. Gallinacin-3, an inducible epitheial beta-defensin I chicken. Infection and Immunity 2001;69:2684–91. 13. Ganz T. Defensin: antimicrobial peptides of intimate immunity. Nature Reviews Immunology 2003;3:710–20. 14. Kagan BL, Ganz T, Lehrer RI. Defensins: a family of antimicrobial and cytotoxic peptides. Toxicology 1994;87:131–49.

Acknowledgements The Spanish Ministry of Science and Technology supported this work through the project AGL 2006-04775. The Council University and Humanistic Development of University of Orient, Monagas, Venezuela, brings a grant for doctoral studies to Ana Yndira Ramos Villarroel, also the University of Lleida by Jade Plus grant. ICREA Academia Award to Professor Olga Martı´n-Belloso is also acknowledged.

References 1 Russell NJ, Gould GW. Food Preservatives. 2nd edn. Plenum Publishers, New York; 2006. 2. Tiwari B, Valdramidis V, Donnel CPO, Muthukumarappan K, Bourke P, Cullen PJ. Application of natural antimicrobials for food preservation. Journal of Agricultural and Food Chemistry Review 2009;57:5987–6000. 3. Davidson PM. Food antimicrobials: back to nature. In: Havkin-Frenkel D, Frenkel C, Dudai N, editors. Proceedings of the First International Symposium on Natural Preservatives in Food Systems. Acta Horticulturae 2006;709:29–33. 4. Beuchat LR, Golden DA. Antimicrobials occurring naturally in foods. Food Technology 1989;43(1):134–42. 5. Deans SG, Noble RC, Hiltunen R, Wuryani W, Penzes LG. Antimicrobial and antioxidant properties of Syzygium aromaticum (L.) Merr. & Perry: impact upon bacteria, fungi and fatty acid levels in ageing mice. Flavour and Fragrance Journal 1995;10:323–8.

15. Conte M, Aliberti F, Fucci L, Piscopo M. Antimicrobial activity of various cationic molecules on foodborne pathogens. World Journal of Microbiology and Biotechnology 2007;23:1679–83. 16. Uyttendaele M, Debevere J. Protamine evaluation of the antimicrobial activity of protamine. Food Microbiology 1994;11:417–27. 17. Johansen C, Gill T, Gram L. Antibacterial effect of protamine assayed by impedimetry. Journal of Applied Bacteriology 1995;78:297–303. 18. Ekstrand B. Lactoperoxidase and lactoferrin. In: Dillon VM, Board RG, editors. Natural Antimicrobials Systems and Food Preservation. CAB International, Wallingford, UK; 1994. p. 15–64. 19. Zeuthen P, Bagh-Sarensen L. Food Preservation Techniques. Woodhead Publishing Limited, Cambridge, England; 2000. 20. Naidu AS. Food Antimicrobials Systems. CRS Press LLC, Florida, USA; 2000. 21. FSANZ. Application A404. Lactoperoxidase system. Food Standards Annual Report Australian New Zealand; 2002. Available from: URL: http://www.foodstandards.gov. au/_srcﬁles/A404_DAR.pdf. 22. USFDA (U.S. Food and Drug Administration). Food Additive Status List. Available from: URL: http://www.cfsan.fda.gov/ dms/opa-appa.html#ftnH; 2006. 23. Murdock CA, Cleveland J, Matthews KR, Chikindas ML. The synergistic effect of nisin and lactoferrin on the inhibition of Lysteria monocytogenes and Escherichia coli O157:H7. Letters in Applied Microbiology 2007;44(3):255–61.

6. Sofos J, Beuchat LR, Davidson PM, Johnson EA. Naturally occurring antimicrobials in food. Task Force Report. Council of Agricultural Science and Technology, Ames, IA; 1998.

24. Naidu AS. Activated lactoferrin – a new approach to meat safety. Food Technology 2002;56(3):40–5.

7. Kim HY, Lee YJ, Hong KH, Kwon YK, Sim KC, Lee JY, et al. Isolation of antimicrobial substances from natural products and their preservative effects. Food Science and Biotechnology 2001;10(1):59–71.

25. Al-Nabulsi AA, Holley RA. Effects of Escherichia coli O157: H7 and meat starter cultures of bovine lactoferrin in broth and microencapsulated lactoferrin in dry sausage batters. International Journal of Food Microbiology 2007;113:84–91.

8. Doyle MP, Beuchat LR, Montville T. Microbiologı´ a de los Alimentos. Fundamentos y Fronteras. Editorial Acribia S.A., Zaragoza, Espan˜a; 2001.

26. Al-Nabulsi AA, Ran JH, Liu ZQ, Rodrigues-Vleira ET, Holley RA. Temperature-sensitive microcapsules containing lactoferrin and their action against Carnobacterium viridans on Bologna. Journal of Food Science 2006;71(6): M208–14.

9. Patrzykat A, Friedrich CL, Zhang L, Mendoza V, Hancock RE. Sublethal concentrations of pleurocidin-derived antimicrobial peptides inhibit macromolecular synthesis in Escherichia coli. Antimicrobial Agents and Chemotherapy 2002;46:605–14. 10. Burrowes OJ, Hadjicharalambous C, Diamond G, Lee TC. Evaluation of antimicrobial spectrum and cytotoxic activity of pleurocidin for food applications. Journal of Food Science 2004;69(3):66–71. 11. Coutinho H, Lobo K, Bezerra D, Lobo I. Peptides and proteins with antimicrobial activity. Indian Journal of Pharmacology 2008;40(1):3–9.

g

27. Del Olmo A, Morales P, Nunez M. Bactericidal activity of lactoferrin and its amidated and pepsin-digested derivatives against Pseudomonas ﬂuorescence in ground beef and meat fractions. Journal of Food Protection 2009;72(4):760–5. 28. USFDA (U.S. Food and Drug Administration). ‘Everything’ Added to Food in the United States (EAFUS): A Food Additive Database. 2008. Available from: URL: http:// www.cfsan.fda.gov/dms/eafus.html. 29. Bellamy W, Takase M, Wakabayashi H, Kawase K, Tomita M. Antibacterial spectrum of lactoferricin B, a potent

CAB International 2010

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso bactericidal peptide derived from the N-terminal region of bovine lactoferrin. Journal of Applied Bacteriology 1992;73:472–9. 30. Roller S, Covill N. The antifungal properties of chitosan in laboratory media and apple juice. International Journal of Food Microbiology 1999;47:67–77. 31. Ben-Shalom N, Ardi R, Pinto R, Aki C, Fallik E. Controlling gray mould caused by Botrytis cinerea in cucumber plants by means of chitosan. Crop Protection 2003;22:285–90. 32. Je JY, Kim SK. Chitosan derivatives killed bacteria by disrupting the outer and inner membrane. Journal of Agricultural and Food Chemistry 2006;54(18):6629–33. 33. Chung YC, Wang HL, Chen YM, Li SL. Effect of abiotic factors on the antibacterial activity of chitosan against waterborne pathogens. Bioresource Technology 2003;88:179–84. 34. Helander IM, Nurmiaho-Lassila EL, Ahvenainen R, Rhoades J, Roller S. Chitosan disrupts the barrier properties of the outer membrane of Gram-negative bacteria. International Journal of Food Microbiology 2001;71:235–44. 35. Liu XF, Guan YL, Yang DZ, Li Z, Yao KD. Antibacterial action of chitosan and carboxymethylated chitosan. Journal of Applied Polymer Science 2001;79:1324–35. 36. Novack K, Cupp MJ, Tracy TS. Chitosan. In: Cupp MJ, Tracy TS, editors. Dietary Supplements. Toxicology and Clinical Pharmacology. Humana Press Inc., Totowa, NY; 2003. p. 33–40. 37. No HK, Meyers SP, Prinyawtwatkul W, Xu Z. Applications of chitosan for improvement of quality and shelf life of foods: a reviews. Journal of Food Science 2007;72:87–100. 38. USFDA (U.S. Food and Drug Administration). Summary of all GRAS Notices. 2007. Available from: URL: http://www.cfsan.fda.gov/rdb/opa-gras.html. 39. Liu N, Chen XiG, Park H, Liu C, Liu C, Hong Meng X, et al. Effect of MW and concentration of chitosan on antibacterial activity of Escherichia coli. Carbohydrate Polymers 2006;64:60–5. 40. Fernandes JC, Tavaria FK, Soares JC, Ramos OS, Monteiro MJ, Pintado ME, et al. Antimicrobial effects of chitosans and chitooligosaccharides, upon Staphylococcus aureus and Escherichia coli, in food model systems. Food Microbiology 2008;25:922–8. 41. Delvesbroughton J. Nisin and its application as a food preservative. Journal of the Society of Dairy Technology 1990;43(3):73–6. 42. Park C, Lee J, Park I, Kim M, Kim S. A Novel antimicrobial peptide from loach, Misgurnus anguillicaudatus. FEBS Letters 1997;411:173–8. 43. Tossi A, Sandri L, Giangaspero A. Amphipathic,-helical antimicrobial peptides. Biopolymers 2000;55:4–30. 44. Hancock RE, Sahl HG. Antimicrobial and host-defense peptides as new antiinfective therapeutic strategies. Nature Biotechnology 2006;24:1551–7. 45. Cole AM, Weis P, Diamond G. Isolation and characterization of pleurocidin, an antimicrobial peptide in the skin secretions of winter ﬂounder. Journal of Biological Chemistry 1997;272:12008–13.

g

223

46. Zasloff M, Martin B, Chen H. Antimicrobial activity of synthetic magainin peptides and several analogues. Proceedings of the National Academy of Sciences, USA 1988;85:910–913. 47. Higazi AAR, Ganz T, Kariko K, Cines DB. Defensin modulates tissue-type plasminogen activator and plasminogen binding to ﬁbrin and endothelial cells. Journal of Biological Chemistry 1996;271:17650–5. 48. Islam NMD, Itakura T, Motohiro T. Antibacterial spectra and minimum inhibition concentration of clupeine and salmine. Bulletin of the Japan Society of Fisheries Science 1984;50:1705–78. 49. Abler LA, Klapes NA, Sheldon BW, Klaenhammer TR. Inactivation of food-borne pathogens with magainin peptides. Journal of Food Protection 1995;58:381–8. 50. Rydlo T, Rotem S, Mor A. Antibacterial properties of dermaseptin s4 derivatives under extreme incubation conditions. Antimicrobial Agents and Chemotherapy 2006;50(2):490–7. 51. Coote PJ, Holyoak CD, Bracey D, Ferdinando DP, Pearce JA. Inhibitory action of a truncated derivative of the amphibian skin peptide dermaseptin s3 on Saccharomyces cerevisiae. Antimicrobial Agents and Chemotherapy 1998;42:2160–70. 52. Yaron S, Rydlo T, Shachar D, Mor A. Activity of dermaseptin K4-S4 against foodborne pathogens. Peptides 2003;24(11):1815–21. 53. Gill AO, Holley RA. Interactive inhibition of meat spoilage and pathogenic bacteria by lysozyme, nisin and EDTA in the presence of nitrite and sodium chloride at 24 C. International Journal of Food Microbiology 2003;80(3):251–9. 54. Raybaudi-Massilia RM, Mosqueda-Melgar J, Soliva-Fortuny R, Martı´ n-Belloso O. Control of pathogenic and spoilage microorganisms in fresh-cut fruits and fruit juices by traditional and alternative natural. Comprehensive Reviews in Food Science and Food Safety 2009;8:157–80. 55. Padgett T, Han IY, Dawson PL. Incorporation of food-grade antimicrobial compounds into biodegradable packaging ﬁlms. Journal of Food Protection 1998;61:1330–5. 56. Georgel P, Crozat K, Lauth X, Makrantonaki E, Seltmann H, Sovath S, Hoebe K. A toll-like receptor 2-responsive lipid effector pathway protects mammals against skin infections with Gram-positive bacteria. Infection and Immunity 2005;73:4512–21. 57. Desbois AP, Smith VJ. Antibacterial free fatty acids: activities, mechanisms of action and biotechnological potential. Applied Microbiology and Biotechnology 2010;85:1629–42. 58. Wille JJ, Kydonieus A. Palmitoleic acid isomer (C16 : 1D6) inhuman skin sebum is effective against Gram-positive bacteria. Skin Pharmacology and Applied Skin Physiology 2003;16:176–87. 59. McGaw LJ, Ja¨ger AK, Van Staden J. Antibacterial effects of fatty acids and related compounds from plants. South African Journal of Botany 2002;68:417–23. 60. Zasloff M. Antimicrobial peptides of multicellular organisms. Nature 2002;415:389–95. 61. Brogden KA. Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria? Nature Reviews Microbiology 2005;3:238–50.

CAB International 2010

224

Animal Science Reviews 2010

62. Boman HG, Faye I, Gudmundsson GH, Lee JY, Lidholm DA. Cell-free immunity in Cecropia. A model system for antibacterial proteins. European Journal of Biochemistry 1991;201:23–31. 63. Tailor RH, Acland DP, Attenborough S, Cammue BP, Evans IJ, Osborn RW, et al. A novel family of small cysteine-rich antimicrobial peptides from seed of Impatiens balsamina is derived from a single precursor protein. Journal of Biological Chemistry 1997;272:24480–87. 64. Hancock REW. Peptide antibiotics. Lancet 1997;349:418–22. 65. Yoshida K, Mukai Y, Niidome T, Takashi C, Tokunaga Y, Hatakeyama T, et al. Interaction of pleurocidin and its analogues with phospholipid membrane and their antimicrobial activity. Journal of Peptide Research 2001;57:119–26. 66. Lampe MF, Isaacs CE. Lactolipids. In: Naidu AS, editor. Natural Food Antimicrobial Systems. CRC Press, Boca Raton, FL; 2000. p. 159–63. 67. Ramanauskiene K, Inke´niene´ AM, Savickas A, Masteikova R, Brusokas V. Analysis of the antimicrobial activity of propolis and lysozyme in semisolid emulsion systems. Drug Research 2009;66(6):681–8. 68. Chen JF, Zagzebski JA, Dong F, Madsen EL. Estimating the autocorrelation function for ultrasound scatterers in isotropic media. Medical Physics 1998;25(5):648–55. 69. Rhoades J, Roller S. Antimicrobial actions of degraded and native chitosan against spoilage organisms in laboratory media and foods. Applied and Environmental Microbiology 2000;66:80–6. 70. Branen J, Davidson PM. Activity of hydrolysed lactoferrin against foodborne pathogenic bacteria in growth media: the effect of EDTA. Letters in Applied Microbiology 2000;30: 233–7. 71. Venkitanarayanan KS, Zhao T, Doyle MP. Antibacterial effect of lactoferricin B on Escherichia coli O157:H7 in ground beef. Journal of Food Protection 1999;55:238–40. 72. Kim S, Fung DY. Antibacterial effect of water-soluble arrowroot (Puerariae radix) tea extract on foodborne pathogens in ground beef and mushroom soup. Journal of Food Protection 2004;67:1953–6. 73. Sivropoulou A, Papanikolaou E, Nikolaou C, Kokkini S, Lanaras T, Arsenakis M. Antimicrobial and cytotoxic activities of oregano essential oils. Journal of Agricultural and Food Chemistry 1996;44:1202–5. 74. Skandamis P, Tsigarida E, Nychas GJE. The effect of oregano essential oil on survival/death of Salmonella typhimurium in meat stored at 5 C under aerobic, VP/MAP conditions. Food Microbiology 2002;19:97–103. 75. Ward SM, Delaquis PJ, Holley RA, Mazza G. Inhibition of spoilage and pathogenic bacteria on agar and pre-cooked roasted beef by volatile horseradish distillates. Food Research International 1998;31:19–26. 76. Yin MC, Cheng WS. Antioxidant and antimicrobial effects of four garlic-derived organosulfur compounds in ground beef. Meat Science 2003;63:23–8. 77. Nielsen PV, Rios R. Inhibition of fungal growth on bread by volatile components from spices and herbs, and their possible application in active packaging, with special emphasis on mustard essential oil. International Journal of Food Microbiology 2000;60:219–29.

g

78. Davidson PM, Naidu AS. Phyto-Phenols. In: Naidu AS, editor. Natural Food Antimicrobial Systems. CRC Press, Boca Raton, FL; 2000. p. 265–94. 79. Benavente-Garcı´ a O, Castillo J, Lorente J, Ortun˜o A, Del Rı´ o JA. Antioxidant activity of phenolics extracted from Olea europaea L. leaves. Food Chemistry 2000;68:457–62. 80. Irkin R, Korukluoglu M. Control of some ﬁlamentous fungi and yeasts by dehydrated Allium extracts. Journal of Consumer Protection and Food Safety 2009;4:3–6. 81. Degirmencioglu N, Irkin R. Survival of some microorganisms in the presence of onion (Alliun cepa L.) extracts in vitro. Polish Journal of Food and Nutrition Sciences 2009;59:61–6. 82. Lanciotti R, Corbo MR, Gardini F, Sinigaglia M, Guerzoni ME. Effect of hexanal on the shelf life of fresh apple slices. Journal of Agricultural and Food Chemistry 1999;47: 4769–76. 83. Lanciotti R, Belletti N, Patrignani F, Gianotti A, Gardini F, Guerzoni ME. Application of hexanal, (E)-2-hexenal, and hexyl acetate to improve to safety of fresh-sliced apples. Journal of Agricultural and Food Chemistry 2003;51: 2958–63. 84. Lei Q, Boatright WL. Lipoxygenase independent hexanal formation in isolated soy proteins induced by reducing agents. Journal of Food Science 2008;73:C646–8. 85. Corbo MR, Lanciotti R, Gardini F, Sinigaglia M, Guerzoni ME. Effects of hexanal, trans-2-hexenal, and storage temperature on shelf life of fresh sliced apples. Journal of Agricultural and Food Chemistry 2000;48:2401–8. 86. Jay JM, Rivers GM. Antimicrobial activity of some food ﬂavouring compounds. Journal of Food Safety 1984;6: 129–39. 87. Walton NJ, Mayer MJ, Narbad A. Vanillin. Phytochemistry 2003;63:505–15. 88. Lee SY, Dougherty RH, Kang DH. Inhibitory effects of high pressure and heat on Alicyclobacillus acidoterrestris spores in apple juice. Applied and Environmental Microbiology 2002;68:4158–61. 89. Beno-Moualem D, Gusev L, Dvir O, Pessis E, Meir S, Lichter A. The effects of ethylene, methyl jasmonate and 1-MCP on abscission of cherry tomatoes from the bunch and expression of endo-1,4-b-glucanases. Plant Science 2004;167:499–507. 90. Ippolito A, Nigro F. Natural antimicrobials in postharvest storage of fresh fruits and vegetables. In: Roller S, editor. Natural Antimicrobials for the Minimal Processing of Foods. CRC Press, Boca Raton, FL; 2003. p. 201–34. 91. Ayala-Zavala JF, Villegas-Ochoa MA, Cuamea-Navarro F, Gonza´lez-Aguilar GA. Compuestos vola´tiles de origen natural. Nueva alternativa para la conservacio´n. In: Gonza´lez- Aguilar GA, Gardea AA, Cuamea-Navarro F, editors. Nuevas tecnologı´ as de conservacio´n de productos vegetales frescos cortados. CIAD AC, Sonora, Me´xico; 2005. p. 314–38. 92. Wang CY, Buta JG. Maintaining quality of fresh-cut kiwifruit with volatile compounds. Postharvest Biology and Technology 2003;28:181–6. 93. Martı´ nez-Ferrer M, Harper C. Reduction in microbial growth and improvement of storage quality in fresh-cut pineapple after methyl jasmonate treatment. Journal of Food Quality 2005;28:3–12.

CAB International 2010

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso 94. Yuste J, Fung DYC. Inactivation of Listeria monocytogenes Scott A 49594 in apple juice supplemented with cinnamon. Journal of Food Protection 2002;65:1663–66. 95. Conner DE, Beuchat LR. Effects of essential oils from plants on growth of food spoilage yeasts. Journal of Food Science 1984;49:429–34. 96. Helander IM, Alakomi HL, Latva-Kala K, Mattila-Sandholm T, Pol I, Smid EJ, et al. Characterisation of the action of selected essential oil components on gram-negative bacteria. Journal of Agricultural and Food Chemistry 1998;46:3590–5. 97. Connor DE, Beuchat LR. Sensitivity of heat stressed yeasts to essential oils of plants. Applied and Environmental Microbiology 1984;47:229–33. 98. Roller S, Board RG. Naturally occurring antimicrobial systems. In: Russell NJ, Gould GW, editors. Food Preservatives. Plenum Publishers, New York; 2006. p. 262–90. 99. Nychas GJE, Skandamis PN. Chapter 9: Antimicrobials from Herbs and Spices. In: Roller S, editor. Natural Anti-Microbials for the Minimal Processing of Foods. Woodhead Publishing, Cambridge, UK; 2003. p. 176–200. 100. Burt SA, Reinders RD. Antibacterial activity of selected plant essential oils against Escherichia coli O157:H7. Letters in Applied Microbiology 2003;36:162–7. 101. Chorianopoulos NG, Giaouris ED, Skandamis PN, Haroutounian SA, Nychas GJE. Disinfectant test against monoculture and mixed-culture bioﬁlms composed of technological, spoilage and pathogenic bacteria: bactericidal effect of essential oil and hydrosol of Satureja thymbra and comparison with standard acid base sanitizers. Journal of Applied Microbiology 2008;104:1586–1869. 102. Alzamora SM, Tapia MS, Lo´pez-Malo A. Minimally Processed Fruits and Vegetables. Fundamental Aspects and Applications. An Aspen Publication, Gaithersburg, MD; 2000. 103. Gutierrez J, Rodriguez G, Barry-Ryan C, Bourke P. Efﬁcacy of plant essential oils against foodborne pathogens and spoilage bacteria associated with ready-to-eat vegetables: Antimicrobial and sensory screening. Journal of Food Protection 2008;71(9):1846–54. 104. Raybaudi-Massilia RM, Mosqueda-Melgar J, Martı´ n-Belloso O. Antimicrobial activity of essential oils on Salmonella enteritidis, Escherichia coli and Listeria innocua in fruit juices. Journal of Food Protection 2006;69:1579–86. 105. Raybaudi-Massilia RM, Mosqueda-Melgar J, Sobrino-Lo´pez A, Soliva-Fortuny R, Martı´ n- Belloso O. Shelf-life extension of fresh-cut Fuji apples at different ripeness stages using natural substances. Postharvest Biology and Technology 2007;45:265–75. 106. Raybaudi-Massilia RM, Mosqueda-Melgar J, Martı´ n-Belloso O. Antimicrobial activity of malic acid against Listeria monocytogenes, Salmonella enteritidis and Escherichia coli O157:H7 in apple, pear and melon juices. Food Control 2008;20:105–12. 107. Raybaudi-Massilia RM, Mosqueda-Melgar J, Martı´ n-Belloso O. Edible alginate-based coating as carrier of antimicrobials to improve shelf life and safety of fresh-cut melon. International Journal of Food Microbiology 2008;121:313–27. 108. Raybaudi-Massilia RM, Rojas-Grau¨ MA, Mosqueda-Melgar J, Martı´ n-Belloso O. Comparative study on essential oils incorporated into an alginate-based edible coating to assure

g

225

the safety and quality of fresh-cut Fuji apples. Journal of Food Protection 2008;71:1150–61. 109. Meyer AS, Suhr KI, Nielsen P. Chapter 6: Natural food preservatives. In: Ohlsson T, Bengtsson N. editors. Minimal Processing Technologies in the Food Industry. Woodhead Publishing, Boston, New York, Washington DC; 2000. p. 124–61. 110. Smith-Palmer A, Stewart J, Fyfe L. Antimicrobial properties of plant essential oils and essences against ﬁve important foodborne pathogens. Letters in Applied Microbiology 1998;26:118–22. 111. Kim J, Marshall MR, Wei C. Antimicrobial activity of some essential oil components against ﬁve food borne pathogens. Journal of Agricultural and Food Chemistry 1995;43: 2839–45. 112. Packiyasothy EV, Kyle S. Antimicrobial properties of some herb essential oils. Food Australia 2002;54(9):384–7. 113. Alzoreky NS, Nakahara K. Antimicrobial activity of extracts from some edible plants commonly consumed in Asia. International Journal of Food Microbiology 2002;80:223–30. 114. Davidson PM. Chemical Preservatives and Naturally Antimicrobial Compounds. In: Doyle MP, Beuchat LR, Montville TJ, editors. Food Microbiology: Fundamentals and Frontiers. 2nd edn. ASM Press, Washington, DC; 2001. p. 593–628. 115. Marino M, Bersani C, Comi G. Impedance measurement to study antimicrobial activity of essential oils from Lamiaceae and Compositae. International Journal of Food Microbiology 2002;67:187–95. 116. Chorianopoulos N, Kalpoutzakis E, Aligiannis N, Mitaku S, Nychas GJ, Haroutounian SA. Essential oils of Satureja, Origanum, and Thymus species: Chemical composition and antibacterial activities against foodborne pathogens. Journal of Agricultural and Food Chemistry 2004;52:8261–7. 117. Katayama T, Nagai I. Chemical signiﬁcance of the volatile components of spices in the food preservative view point. VI. Structure and antibacterial activity of Terpenes. Bulletin of the Japan Society of Fisheries Science 1960;26:29–32. 118. Farag RS, Daw ZY, Hewed FM, El-Baroty GSA. Antimicrobial activity of some Egyptian spice oils. Journal of Food Protection 1989;52:665–67. 119. Nychas GJE. Natural Antimicrobials from Plants. In: Gould GW, editor. New Methods of Food Preservation. Blackie Academic and Professional, Glasgow, Scotland; 1995. p. 58–89. 120. Dorman HJD, Deans SG. Antimicrobial agents from plants: antibacterial activity of plant volatile oils. Journal of Applied Microbiology 2000;88:308–16. 121. Lambert RJW, Skandamis PN, Coote PJ, Nychas GJE. A study of the minimum inhibitory concentration and mode of action of oregano essential oil, thymol and carvacrol. Journal of Applied Microbiology 2001;91:453–62. 122. Lopez-Malo A, Alzamora SM, Palou E. Naturally Occurring Compounds: Plant Sources. In: Davidson PM, Sofos JN, Branen AL, editors. Antimicrobials in Food. 3rd edn. CRC Press, New York; 2005. p. 429–51. 123. Demetzos C, Perdetzoglou DK. Composition and antimicrobial studies of the oils of Origanum calcaratum Juss and O. scabrum Boiss from Greece. Journal of Essential Oil Research 2001;13:460–2.

CAB International 2010

226

Animal Science Reviews 2010

124. Kubo A, Lunde CS, Kubo I. Antimicrobial activity of the olive oil ﬂavor compounds. Journal of Agricultural and Food Chemistry 1995;43:1629–33. 125. Char C, Guerrero S, Alzamora SM. Survival of Listeria innocua in thermally processed orange juice as affected by vanillin addition. Food Control 2009;20:67–74. 126. Rojas-Grau¨ MA, Raybaudi-Massilia RM, Soliva-Fortuny RC, Avena-Bustillos RJ, McHugh TH, Martı´ n-Belloso O. Apple puree-alginate edible coating as carrier of antimicrobial agents to prolong shelf life of fresh-cut apples. Postharvest Biology and Technology 2007;45:254–64.

140. Cleveland J, Montville TJ, Nes IF, Michael L, Chikindas ML. Bacteriocins: safe, natural antimicrobials for food preservation. International Journal of Food Microbiology 2001;71:1–20. 141. Anastasiadou S, Papagianni M, Filiousis G, Ambrosiadis I, Koidis P. Pediocin SA-1, an antimicrobial peptide from Pediococcus acidilactici NRRL B5627: Production conditions, puriﬁcation and characterization. Bioresources Technology 2008;99:5384–90. 142. Hancock REW, Lehrer R. Cationic peptides: a new source of antibiotics. Trends in Biotechnology 1998;16:82–8.

127. Gulluce M, Sahin F, Sokmen M, Ozer H, Deferera D, Sokmen A, et al. Antimicrobial and antioxidant properties of the essential oils and methanol extract form Mentha longifolia L. ssp. Longifolia.Food Chemistry 2007;4:1449–56.

143. Branby-Smith FM. Bacteriocins: Applications in food preservations. Trends in Food Science and Technology 1992;3:133–7.

128. Bajpai VK, Rahman A, Dung NT, Huh MK, Kang SC. In vitro inhibition of food spoilage and foodborne pathogenic bacteria by essential oil and leaf extracts of Magnolia liliﬂora Desr. Journal of Food Science 2008;73(6):M314–20.

144. Geornaras I, Skandamis PN, Belk KE, Scanga JA, Kendall PA, Smith G, et al. Postprocess control of Listeria monocytogenes on commercial frankfurters formulated with and without antimicrobials and stored at 10 degrees C. Journal of Food Protection 2006;69(1):53–61.

129. Kabara JJ, Eklund T. Organic Acids and Esters. In: Russel NJ, Gould GW, editors. Food Preservatives. Blackie and Son Ltd, Glasgow, Scotland; 1991. p. 44–71.

145. Eckner KF. Bacteriocins and food applications. Dairy, Food and Environmental Sanitation 1992;12(4):204–9.

130. Chorianopoulos N, Evergetis E, Mallouchos A, Kalpoutzakis E, Nychas GJ, Haroutounian SA. Characterization of the essential oil volatiles of Satureja thymbra and Satureja parnassica: Inﬂuence of harvesting time and antimicrobial activity. Journal of Agricultural and Food Chemistry 2006;54:3139–45. 131. Delaquis PJ, Stanich K, Girard B, Mazza G. Antimicrobial activity of individual and mixed fractions of dill, cilantro, coriander and eucalyptus essential oils. International Journal of Food Microbiology 2002;74:101–9. 132. Singh N, Singh RK, Singh A, Bhuniab AK. Efﬁcacy of plant essential oils as antimicrobial agents against Listeria monocytogenes in hotdogs. LWT-Food Science and Technology 2003;36:787–94. 133. Gutierrez J, Barry-Ryan C, Bourke P. Antimicrobial activity of plant essential oils using food model media: Efﬁcacy, synergistic potential and interactions with food components. Food Microbiology 2009;26:142–52. 134. Daeschel MA. Antimicrobial substances from lactic acid bacteria for use as food preservatives. Food Technology 1989;43(1):164–7. 135. De Vugst L, Vandamme EJ. Bacteriocins of Lactic Acid Bacteria: Microbiology, Genetics and Applications. Blackie Academics and Professional, London; 1994. 136. Nom MJR, Rombouts FM. ‘Fermentative preservation of plant foods’ in 1. Applied Bacterial Symposium Supplement 1992;73:1365–478. 137. Talarico TL, Dobrogosz WJ. Chemical characterization of an antimicrobial substance produced by Lactobacillus reuteri. Antimicrobial Agents and Chemotherapy 1989;33:674–9. 138. El-Ziney MG, van den Tempel T, Debevere JM, Jakobsen M. Application of reuterin produced by Lactobacillus reuteri 12002 for meat decontamination and preservation. Journal of Food Protection 1999;62:257–61. 139. Arques JL, Fernandez J, Gaya P, Nunez M, Rodriguez E, Medina M. Antimicrobial activity of reuterin in combination with nisin against food-borne pathogens. International Journal of Food Microbiology 2004;95(2):225–9.

g

146. Hugas M, Garriga M, Aymerich MT, Monfort JM. Inhibition of Listeria in dry fermented sausages by the bacteriocinogenic Lactobacillus sake CTC494. Journal of Applied Bacteriology 1995;79(3):322–30. 147. Henning S, Metz R, Hammus WP. New aspects for the application of nisin to food products based on its mode of action. International Journal of Food Microbiology 1986;3:141–55. 148. Zhang SS, Mustapha A. Reduction of Listeria monocytogenes and Escherichia coli O157: H7 numbers on vacuum-packaged fresh. Journal of Agricultural and Food Chemistry 1999;57(14):1123–7. 149. Delves-Broughton J, Williams GC, Wilkinson S. The use of the bacteriocin, nisin as a preservative in Pasteurized liquid whole egg. Letters in Applied Microbiology 1992;15:133–6. 150. Nassar A, Farrag SA. Nisin as inactivator to L. monocytogenes in broth and in ground beef. Assiut Veterinary Medical Journal 1995;32:198–206. 151. Branen AL, Davidson PM, Salminen S. Food Additives. Marcel Dekker Inc, New York; 1990. 152. Stiles M. E. Potential for biological control of agents of food-borne disease. Food Research International 1994;27:245–50. 153. Beuchat CR. Surface decontamination of fruits and vegetables eaten raw: a review. WHO/FSF/FOST/98.2, World Health Organization. Food Safety Unit, Geneva; 1998. 154. Cotter PD, Hill C, Ross RP. Bacteriocins: Developing innate immunity for food. Nature Reviews Microbiology 2005;3:777–88. 155. Leite de Souza E, Antonio da Silva C, Paiva de Sousa C. Bacteriocins: molecules of fundamental impact on the microbial ecology and potential food biopreservatives. Brazilian Archives of Biology and Technology 2005;48(4):559–66. 156. Nes IF, Holo H. Class II antimicrobial peptides from lactic acid bacteria. Biopolymers 2000;55:50–61.

CAB International 2010

Ana Yndira Ramos-Villarroel, Robert Soliva-Fortuny and Olga Martı´ n-Belloso 157. Riley MA, Chavan MA. Bacteriocins, Ecology and Evolution. Springer, Berlin, Germany; 2006. 158. Leite de Souza E, Da Silva CA, Paiva de Sousa C. Bacteriocins: Molecules of fundamental impact on the microbial ecology and potential food biopreservatives. Brazilian Archives of Biology and Technology 2005;48(4):559–66. 159. Nettles CG, Barefoot SF. Biochemical and genetic characteristics of bacteriocins of food-associated lactic acid bacteria. Journal of Food Protection 1993;56:338–56.

227

166. Sapers GM, Simmons GF. Hydrogen peroxide disinfection of minimally processed fruits and vegetables. Food Technology 1998;52:48–52. 167. McNeillie A, Bieser J. Hydrogen peroxide uses for the year 2000. Food Processing 1993;October:59–65. 168. Montville TJ, Chen Y. Mechanistic action of pediocin and nisin: recent progress and unresolved questions. Applied Microbiology and Biotechnology 1998;50:511–9.

160. Delves Broughton J, Blackburn P, Evans RJ, Hugenholtz J. Applications of the bacteriocin, nisin. Antonie van Leeuwenhoek 1996;69:193–202.

169. Lee DU, Heinz V, Knorr D. Effects of combination treatments of nisin and high-intensity ultrasound with high pressure on the microbial inactivation in liquid whole egg. Innovative Food Science and Emerging Technologies 2003;4:387–93.

161. Montville TJ, Chen Y. Mechanistic action of pediocin and nisin: recent progress and unresolved questions. Applied Microbiology and Biotechnology 1998;50:511–9.

170. Crandall AD, Montville TJ. Nisin resistance in Listeria monocytogenes ATCC 700302 is a complex phenotype. Applied and Environmental Microbiology 1998;64:231–7.

162. Axelsson LT, Chung TC, Dobrogosz WJ, Lindgren SE. Production of a broad spectrum antimicrobial substance by Lactobacillus reuteri. Microbial Ecology in Health and Disease 1989;2:131–6.

171. Davies EA, Adams MR. Resistance of Listeria monocytogenes to the bacteriocin nisin. International Journal of Food Microbiology 1994;21(4):341–7.

163. Piva A, Headon DR. Pediocin A, a bacteriocin produced by Pediococcus pentosaceus FBB61. Microbiology 1994;140:697–702.

172. Stevens KA, Sheldon BW, Klapes NA, Klaenhammer TR. Nisin treatment for inactivation of Salmonella species and other Gram-negative bacteria. Applied and Environmental Microbiology 1991;57:3613–5.

164. Komitopoulou E, Boziaris IS, Davies EA, Delves-Broughton J, Adams MR. Alicyclobacillus acidoterrestris in fruit juices and its control by nisin. International Journal of Food Science and Technology 1999;34(1):81–5.

173. Davidson MP, Sofos JN, Branen AL. Antimicrobials in food. 3rd edn. CRC Press, Taylor and Francis Group, LLC, Boca Raton, FL; 2005.

165. Davies EA, Milne CF, Bevis HE, Potter RW, Harris JM, Williams G, et al. Effective use of nisin to control lactic acid bacterial spoilage in vacuum-packed Bologna-type sausage. Journal of Food Protection 1999;62(9):1004–10.

174. Johnson EA. Egg white lysozyme as a preservative for use in foods. In: Sim JS, Nakai S, editors. Egg Uses and Processing Technologies: New Developments. CAB International, Wallingford, UK; 1994. p. 177–91.

g

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Palatability: principles, methodology and practice for farm animals J. Michael Forbes* Address: Institute of Integrative and Comparative Biology, University of Leeds, LS21 1JY, UK. *Correspondence: Email: [email protected] 26 April 2010 9 July 2010

Received: Accepted:

Abstract Palatability is a complex concept, not amenable to concise deﬁnition because it depends not only on the organoleptic properties of the food but also on the experience and genetic background of the animal in question and its physiological state, as well as environmental conditions and social context. The learning of the association between the appearance, ﬂavour or texture of a food and the metabolic consequences of eating that food provides the animal with a conditioned response to that food, either preference or aversion, irrespective of the nature of the sensory cue. Some ﬂavours are innately preferred, others aversive, but conditioning can change these preferences. Critical to the understanding of palatability is an appreciation of the methods used to measure it. These range from the initial rate of eating of a food, through choices expressed between two (or more) foods offered simultaneously, to more sophisticated methods involving direct or operant conditioning. Many factors inﬂuence the outcome of palatability tests, including the length of the test, time of the day, sensory characteristics of foods, environmental factors, physiological state of the animal and opportunity to learn from conspeciﬁcs. Whether or not long-term intake can be predicted from short-term preference depends on how the latter is measured and whether the animals have innate or conditioned preferences for a food; thus, detailed knowledge of the animals’ history and the conditions under which a test has been carried out is necessary in order to interpret and make use of palatability measurements. Keywords: Palatability, Food Flavour, Food Composition, Cattle, Sheep, Poultry, Pigs Review Methodology: The following databases were searched: ‘palatability’ in Google; ‘palatability and (sheep or cattle or pig or swine or chicken or poultry)’ in Google Scholar, Web of Science, CABI’s Nutrition and Food Sciences database.

Introduction ‘. . . palatability is . . . meaningless as a food descriptor’. ‘. . . it could not be measured in any meaningful way’ [1]. ‘. . . there is no precise way of measuring palatability and there seems little point in using the term. However, it is very widely used and it may be better to qualify it rather than reject its use altogether’ [2]. Notwithstanding such caveats, the voluntary intake of food is of such paramount importance in animal agriculture that there is considerable interest in the ﬂavour and other sensory characteristics of foods, collectively often referred to ‘palatability’, which might inﬂuence intake. This is especially true in situations in which the amount of food eaten falls short of that required for the animal to fulﬁl its

g

potential rate of growth, milk production or reproduction, especially after weaning, in late pregnancy, in early lactation and with poor-quality or imbalanced foods. ‘Palatability’ features prominently in many advertisements and descriptions of commercial feeds and is commonly thought to be important in determining the level of voluntary food intake. There is, therefore, a need to clarify what is meant by ‘palatability’, how it is assessed, what factors affect it and whether it can be used to predict long-term intake.

Deﬁnitions ‘Palatability’ is a commonly used but much misunderstood term. The generally accepted meaning of ‘palatable’ is

CAB International 2010

230

Animal Science Reviews 2010

‘agreeable to the palate or taste’ and in the context of livestock ‘palatability’ can be deﬁned as ‘the appeal and acceptability of feedstuffs, including the taste, odour, texture and temperature of the feed’ [3]; to that list of food properties could be added ‘appearance’ and ‘location’. A more comprehensive deﬁnition is ‘the hedonic reward provided by foods or ﬂuids that are agreeable to the “palate” with regard to the homeostatic satisfaction of nutritional, water, or energy needs’ [4]. Importantly, this deﬁnition continues: ‘The palatability of a food or ﬂuid, unlike its ﬂavour or taste, varies with the state of an individual: it is lower after consumption and higher when deprived. Palatability of foods . . . can be learnt’. Again, in terms less accessible to the agricultural scientist: ‘a response measure which is based on the outcome of the central nervous system’s integration of taste and internal-state signals combined with cues arising from previous associations’ [5]. There is clearly no universally recognized deﬁnition of the term ‘palatability’, but as Marten [6] says, with particular reference to grazed herbage: ‘the concept of palatability is of more importance than any speciﬁc deﬁnition’. He also advises against the use of ‘acceptability’ as being synonymous with ‘palatability’. This discussion will not cover the complex subject of the ‘palatability’ of grazed herbage. Yeomans [7] provides an excellent critique of palatability for humans. He points out that a major difﬁculty in the use of this term arises because it is often used as the ability of sensory factors to increase intake, and so it would be a circular argument to say that the observation of an increase in intake is a consequence of palatability. Human attitudes to foods are often derived from questionnaires, but it has been argued that natural food choice is subconscious and inducing subjects to think about foods generates false conclusions. It may be, therefore, that the methodology for animal studies is more appropriate for human research than hitherto appreciated and Ko¨ster [8] recommends that ‘Observation of actual choice in typical situations should replace methods based on just asking about liking or wanting . . .’. Notwithstanding these deﬁnitions, too often the term is used as a proxy for ﬂavour, without appreciation of the complexities hinted at above. Henceforth, I will not put quotation marks around palatability but bear in mind that it is not simply a quality of the food because it depends so heavily on the experience and metabolic status of the animals in question. It also depends on the innate (inborn) preferences and aversions which, in many mammalian species are for sweet and against bitter taste, respectively; such preferences and aversions could be said to be unconditioned in the sense that no experience by the individual is necessary in order for them to inﬂuence palatability – they might indeed be thought of as preconditioned. Palatability may also be inﬂuenced by prior exposure to tastants in utero and/or in the mother’s milk, another example of preconditioning. As with other

g

associations between the sensory and metabolic properties of foods, genetic, pre- and post-parturition conditionings can be reconditioned by exposing animals to different associations, for example, by pairing an aversive bitter taste with pleasant metabolic consequences, as in the following example.

An Example The difﬁculties in the concept of palatability can be neatly illustrated by a much-quoted example: pigs have Bitrex, the bitterest substance known to the human palate, suddenly incorporated in their food [9]. Not surprisingly, they immediately reject the food. Pigs are innately averse to bitter food; toxins in food often have a bitter taste so evolution has provided a means of avoiding such compounds. A combination of hunger and inquisitiveness, however, leads these pigs to approach the food again after a while but after one nibble they again retreat. As hunger gets stronger they approach and nibble more often and, a few days after the introduction of the Bitrex, they are once again eating normal amounts of the food. Although the food becomes unpalatable, the pigs gradually learn that there has been no change to how they feel after eating the food and they quickly learn to eat it readily. Whether this means that it has now become palatable is open to debate (it has certainly become acceptable), the outcome of which depends on the deﬁnition of palatability. However, the deﬁnition of palatability is also informed by examples such as the one just quoted. Whatever the argument, it is clear that learning is a critical part of any discussion on palatability.

Aims The aim of this article is to encourage a discussion of the concept of palatability and the methods whereby it is measured. It should be of value to students, teachers and researchers in animal science and nutrition, especially those embarking on projects involving animals and their feeds; to nutritionists in animal feed companies and veterinarians; and to all those with an interest in how animals interact with their food. At one level, it will be seen as a catalogue of methods for measuring palatability and a list of factors affecting it; at another level, it should provide a reference for anyone contemplating studies of palatability; and, most importantly, it should stimulate deeper thinking about the relationship between the animal and its food. This article will not, however, go into the discussion of putative mechanisms such as the neurochemistry of the brain, nor will it discuss the chemical senses, olfaction and taste, which play major roles as cues whereby the animal can identify different foods – Roura and Tedo´ [10] describe the manner in which odours and ﬂavours are

CAB International 2010

J. Michael Forbes 80 % preference for test forage

sensed and discuss their importance in the identiﬁcation of foods. It will not cover in detail the acceptability of numerous ingredients [11] or make recommendations as to how to improve palatability under commercial conditions. Rather, it will deal with the principles and pitfalls involved in the assessment and understanding of palatability, with particular reference to pigs, poultry, cattle and sheep. The subjects of palatability and food choice have been reviewed for ruminants by Baumont [12] and for pigs and poultry by Meunier-Salaun and Picard [13].

231

70 60 50 40

AIR

30 20

PIR

10 0

0

5

10

15

20

25

30

PIR of standard forage (g/min)

Figure 1 Preference for a test forage determined against hay of different particle sizes [20]. See text for details

Assessment of Palatability Given that palatability is inﬂuenced by so many factors, and that there is no absolute deﬁnition, there equally can be no absolute method of assessing it. The following are methods that have been used.

Daily Intake of a Single Food The weight of food eaten on the ﬁrst day of offering a new food could be used as an index of palatability. However, by the end of the ﬁrst 24 h the animal has already learned something about the metabolic effects of eating that food, but complete learning takes several days, so the 24 h intake is intermediate between initial experience and longer-term learned response. A decision has to be made as to the purpose of the measurement; whether it is to assess initial response or long-term acceptance. As we will see, the two might be very different, depending on the circumstances.

Short-Term Intake of a Single Food The rate at which animals eat a novel food when it is ﬁrst offered has been used as a measure of palatability, because it is an indication of the degree of motivation, which is governed by the degree of hunger and by the anticipation of resulting pleasure or comfort of eating that food. With farm animals, the rate of eating during large meals following an enforced fast of several hours is initially rapid, but slows down as the meal progresses (pigs [14], sheep [15], cows [16] and chickens [17]). Both the initial rate of eating and the rate of deceleration are inﬂuenced by animal factors, such as degree of hunger and food factors, such as particle size and ﬂavour. Cumulative intake during a large meal follows the general form: y=a(17e7bt), where y is the cumulative weight eaten up to time t, a is the asymptote (cumulative weight eaten by the time feeding stops) and b is the relative time constant. In the study of Baumont et al. [15] in which sheep were offered lucerne, the values of a and b were 1975 g and 0.0044, respectively.

g

The short-term intake rate (STIR) has been proposed as a method for ranking intake potential of forages by cattle. The weight eaten in 4 min after a 4-h fast was used by Harrison et al. [18] and, while there was some evidence that STIR was related to daily intake in the longer term, there were marked effects on STIR of the order in which the forages were presented on test days. The initial rate of eating is inﬂuenced by many factors including, in particular, previous experience of eating the food in question, or foods with similar sensory properties. For example, sheep ate more quickly food of a ﬂavour that had previously been eaten after the introduction of starch into the rumen than food of another ﬂavour eaten after water infusion into the rumen [19], i.e. they ate more quickly a food that was anticipated to provide a greater and more rapid yield of absorbable energy. Until they have learnt the associations between the sensory and the metabolic properties of a food animals are not in a position to base their initial rate of eating on anything other than innate reactions or what they may have learnt from exposure to foods with similar sensory properties. A more sophisticated use of the initial rate of eating, combined with two-choice preference testing, has been developed by Colebrook et al. [20] with the aim of separating the role of physical characteristics from the other sensory factors. ‘Potential intake rates (PIR)’ of test forages and of a standard hay cut to different lengths are measured as the intake in the ﬁrst minute after offering each fresh forage without choice. A test forage is then offered as a choice with several chop lengths of the standard forage in turn and the preferences ascertained. Finally, the PIR of the standard food at the chop length which gives a 50% choice when offered with the test forage is calculated (Figure 1) to describe the ‘adjusted intake rate (AIR)’ and the difference between PIR and AIR is taken as the measure of the sensory factors affecting selection of the test forage, relative to the standard forage. This method is complex and really only suitable for research studies and not for routine testing of foods. The initial rate of eating by sheep is negatively related to degree of grinding, as expressed by particle size, as is

CAB International 2010

232

Animal Science Reviews 2010

preference when different particle sizes are offered in free choice [21].

1.0

Newly weaned animals are not used to eating large amounts of solid food, especially when weaned at a very young age as is common with piglets. The latency to take a signiﬁcant meal has been used as an indication of the palatability of food, where ‘signiﬁcant’ is taken to be a meal of at least 0.5 min [22]. However, useful such a measure might be, it demands either observation by eye or video camera, or else complex automatic detection methods, and is not suitable for any other than early weaned animals. Older animals are always likely to begin eating immediately on the offer of fresh food, even if it is novel, unless this food has an innately strongly aversive ﬂavour or has been rendered highly aversive by prior experience of a toxic food with the same or similar ﬂavour.

0.6

Oats Barley

0.4

Maize Wheat

0.2 0 Oats

Given the numerous factors affecting palatability of a food, however, it is measured; it is not surprising that attention has been focused on the assessment of relative palatability, usually monitored by comparing the amounts eaten of two foods offered simultaneously. This is probably the most widely used method for assessing palatability even though the results are sometimes difﬁcult to interpret. If both the foods offered in free choice have a lower content of the nutrient of interest than optimal, then it is difﬁcult for animals to choose between them, because any mixture of the two foods is deﬁcient. In addition, offering two foods that have an excess and a deﬁciency, respectively, of a nutrient such that an optimal diet would be 50% of one food and 50% of the other would lead to difﬁcult interpretation, as 50 : 50 could be random eating rather than directed choice. Diet selection experiments should be designed so that a choice of at least two-third of one food is to be expected and the proportions chosen can then be compared statistically with 50 : 50. There may be difﬁculties with this approach, however. For example, it may be that it is required to use two ﬂavours in a conditioning experiment and to be sure that they are approximately equally preferred before conditioning is applied (e.g. [23]); in this case, preference close to 50 : 50 is required. Any pair of foods can be offered to any animal(s) to determine which is preferred, as measured by the weight eaten over a given period of time. To rank 10 foods, it would be necessary to make 45 comparisons. It should, however, be possible to deduce the relative palatabilities of foods that have not been compared directly from knowledge of how each relates to another, common food and this would reduce to 10 the number of pairs of foods

g

Maize

Wheat

0.8 0.6

Oats Barley

0.4

Maize Wheat

0.2 0 Oats

Choice between Two Foods

Barley

1.0

Preference

Time to First Meal

Preference

0.8

Barley

Maize

Wheat

Figure 2 Preference between pairs of cereal-based foods for (a) suckler and (b) weaner piglets [24]. Oats, ^; barley, *; maize, m; wheat, &. On the X-axis each food is considered in turn to be the reference food and the points vertically above are the preference for each of the foods offered in choice with this reference food

to be offered. If A is preferred to B, and B to C, then would A be preferred to C? While in theory the properties of the reference food should not be important, in practice it is usually bland and nutritionally balanced. Many examples of results of two-choice tests with a reference food are given for pigs by Torrallardona and Sola-Oriol [11], covering cereals, protein sources, amino acids, fats, ﬁbre sources, palatability agents, antibiotics, mixtures of ingredients, particle size, texture and nutritive value. It is important to know whether test foods would be ranked in similar order when compared with different reference foods. Bruneau and Chavez [24] offered foods containing 70% of wheat, maize, barley or oats in pairs, both with suckled and weaned piglets, allowing each to be considered in turn as the reference food. Figure 2 shows that the preference ranking is the same irrespective of the food considered as the reference food. What is more, the ranking is very similar for sucklers as it is for weaners. There are other cases in which the agreement is not so strong, especially when the preferences are not very marked (e.g. [11]). The general conclusion, however, is that the rank order of a series of test foods is very similar for all reference foods. A very comprehensive study of food preference has been carried out by Sola-Oriol et al. [25, 26] with many different foods. A broad conclusion is that short-term preference for a food is a good predictor of longer-term preference. However, their deﬁnition of short-term was the ﬁrst 4 days after offering the pair of foods in which

CAB International 2010

J. Michael Forbes

case the piglets had chance to learn a great deal about the metabolic effects of eating the foods. Thus, by the end of the 4 days conditioning would be well under way, if not complete, and the mean of the 4-day preference results would represent a relative palatability during conditioning, rather than the unconditioned or the conditioned palatability (see Conclusions section). Where 4-day intakes are presented, then the ﬁrst day, when intakes are very low, contributes very little to the 4-day mean – most of this is from days 3 and 4, when conditioning is well under way. The fact that preference for test foods relative to a reference food during a 4-day test were signiﬁcantly positively correlated with the rate of glucose release in an in vitro system (r=0.48) and with digestible energy content (r =0.54), and negatively with the crude ﬁbre content (r = 70.61) [25], is a strong suggestion that the pigs in these trials had learned, at least partially, the association between the organoleptic and the metabolic properties of the foods in question. The results of two-choice tests are inﬂuenced by the amounts of the foods on offer, because if the preferred food becomes exhausted before the end of the test, the animal is likely to eat some of the less-preferred food, especially if it is still very hungry, i.e. if the test is carried out after several hours of fasting. The duration of the test can also affect its outcome as in short tests the animal may only have time to eat the preferred food, even if it would take some of the other food if the test had been longer. Test results can also be inﬂuenced by the availability of feeder space. Pigs tend to eat in synchrony, and if the preferred feed is occupied, then some pigs go to the other ‘non-preferred’ feeder.

Two-‘Bottle’ Tests Preference tests in which individual animals are offered a choice between two drinker-bottles, one containing water, the other a test solution, have been widely used with laboratory animals and humans. The important difference between two-food tests and two-bottle tests is that in the former, the tastant in question is mixed with a large number of other substances, many with their own taste; while in the latter, the tastant alone is present (in the presence of tasteless water). The threshold for taste is therefore likely to be much lower for a substance in solution than in solid food. For pigs, responses to solutions of bitter compounds have been studied using two-bottle tests over 24 h [27], whereas tests with sweet compounds have been carried out, but using buckets rather than bottles, by Kennedy and Baldwin [28] with both short (1 h) and long (12 h) durations. The concentration of each substance tested (sucrose, glucose, saccharin and cyclamate) was progressively increased. Preference for the ﬁrst three substances increased with concentration and was similar for both lengths of test. Cyclamate did not generate any preference.

g

233

Glaser et al. [29] also used such a test for growing pigs which had been accustomed to the test procedure and were then offered a large number of carbohydrates, polyols and sweeteners. Each choice was offered for 1 min, during which behaviour was monitored and intake measured. All the carbohydrates were preferred over water, sucrose being most favoured and the rank order for the carbohydrates was similar to that for humans. The polyols were also attractive, as in humans, while the results for the sweeteners were variable, several being less preferred than is the case in man. The very short duration of the tests meant that there was only a small intake of the substances and thus little opportunity for post-ingestive effects. These results are likely to be a consequence of innate preferences for sweet taste. This type of test has been used to individual amino acids by Tedo´ [30], who found that pigs can readily detect many of them in solution with a threshold of around 5 mM (500–1000 mg/l), i.e. similar to the concentration in normal diets for pigs and poultry. As discussed above, this does not mean that they could detect such concentrations in solid foods and it is expected that stronger taste or visual cues need to be used to allow animals to learn whether a particular food contains insufﬁcient or an excess of an amino acid to meet its requirement for that amino acid.

Three or More Foods A few observations have been made where three or more foods have been offered. Sometimes the choice is so complex that animals have failed to achieve a diet that is close to being balanced. In other cases, a good balance appears to have been achieved, but the difﬁcultly of interpreting the results of complex tests is such that they are not to be recommended.

Electrophysiological Studies It has been suggested that the effects of solutions sprayed onto the tongue of anaesthetized animals on the frequency of impulses in afferent nerves might be useful in studies of palatability. The research on farm animals included pigs [31] and calves [32]. Although an advantage of this approach is that it provides information about the sensitivity of lingual receptors without interference from nasal odour receptors, it does not take any account of the interactions with learning and ‘need’ so important in the consideration of palatability.

Presentation of Food Preference Results Given that the two-choice test is the most widely used and accepted way of assessing relative palatability, it is

CAB International 2010

234

Animal Science Reviews 2010

Cumulative sum (HP + LP, kg)

120 100 80 60 40 20 0 –20 – 40 – 60 0

20

40

60

80

100

120

140

Cumulative difference (HP – LP, kg)

Figure 3 Diet selection pathways for six lactating cows fed a choice of HP and LP concentrate supplements with a maximum intake restricted to 6.1 kg/day [35]. Grass silage was offered ad libitum. Immediately before this choicefeeding period, two cows had been given only HP supplement (solid lines), two LP (dashed lines) and two the choice of both (dotted lines)

important that standard methods for expressing results are adopted. Proportional intakes The clearest way to present the result of a preference test is the intake of one of the foods as the proportion of total intake. Rather than using percentage, proportions should be expressed as parts of one, or possibly as g/kg, noting that SI units are in multiples of 1000. Thus, if 200 g of food A and 50 g of food B are eaten during a particular test then the proportion of food A is 200/(200+50)=0.8 or 800 g/kg. It is misleading to express the intake of one food as a proportion of the intake of the second food [11], both for statistical reasons as well as to avoid confusion with other methods of expressing proportional food intake. For the purposes of statistical analysis, proportions are unlikely to be normally distributed and should be converted by arcsine transformation or arcsine of the square root of the proportions [33]. If the proportion is 0 or 1, it is permissible to replace with 0.0005 or 0.9995, respectively, so as to allow the transformation. When presenting the results of analysis of transformed data they should be back-transformed. Diet selection pathway This method of presenting the results of food choice experiments was ﬁrst used by Kyriazakis et al. [34] and involves plotting the cumulative difference between the daily intakes of the two foods against the sum of the intakes of the two foods (Figure 3). If equal amounts of the two foods are eaten then the trajectory will be horizontal; a rising line indicates more of the ﬁrst food is eaten, a falling line that more of the second food was eaten. The method has the advantage of showing daily

g

choices by individual animals in a clear and simple manner. It could also be used for periods other than one day, e.g. minutes or hours in short-term tests. Data presented in this form could be analysed by regression analysis, preferably polynomial as the gradient is likely to ﬂuctuate due to changes in proportional intake as animals learn about the foods and/or as their requirements develop as they grow. Figure 3 shows the pathways for six lactating cows for a period of 21 days when given grass silage ad libitum and 6.1 kg/day of concentrate supplement within which they could choose between one containing 280 (high protein, HP) and the other with 180 CP/kg fresh matter (low protein, LP) [35]. The ﬁgure shows that initially there was considerable variation in choice between animals, from only HP to only LP being eaten, with some apparently making no choice between the two, i.e. eating 50% of each. Within a few days, however, all but one were eating similar proportions, approximately equal amounts of each, which allowed them to produce milk at a similar rate to cows offered only the HP concentrate food. The cow which persisted in eating only HP supplement was the one with the highest yield of milk protein. Another way to present results from two-food choices where there are differences in two of the resources provided by the foods (e.g. energy and protein) is to plot the intake of one resource against the intake of the other. This approach has been used for the intakes of metabolizable energy (ME) and crude protein (CP) by broiler chickens [36] and by growing sheep [37] offered HP and LP foods. While these examples show only a single point for each treatment, it would be a simple matter to show the results for several different time periods on the same graph, i.e. to plot trajectories rather than just ‘rails’. Where there are three or more foods on offer, it becomes more difﬁcult to present the results in graphical form. Statistical analysis also becomes more problematical, because the proportions are not independent – the greater the proportion of one food eaten, the less of one or both of the others. There is a case for using multivariate analysis in this type of situation, which can cope with variables that are inter-related, in which case one should be warned that interpretation is notoriously difﬁcult! However, the results of preference tests are presented, there must be sufﬁcient description of all the relevant conditions to allow the reader to interpret the results according to her/his needs (see Conclusions section).

Conditioning Animals can easily be conditioned to associate some sensory property of a food with how it feels internally after eating this food. For example, feeding or injecting lithium chloride causes nausea in a wide range of animals and a novel food offered for a short time after such

CAB International 2010

J. Michael Forbes

treatment causes aversion to that food after a few sessions. Similarly, aversion to some toxic substances commonly found in the diet can be conditioned by association with novel ﬂavours; oxalic acid is present in the leaves of many root crops and conditioned taste aversions to it can persist for at least 60 days in sheep [38]. In chickens, injection with cholecystokinin (CCK, a hormone released in response to feeding) followed by eating a coloured food causes conditioned aversion to that colour [39]. The external stimulus, in this case the colour of the food, is called the conditioned stimulus (CS), while the internal stimulus, in this case CCK, is the unconditioned stimulus (US). However, food colour and ﬂavour are food related and the strength of conditioned preferences and aversions would be tested more rigorously if the CS was not one normally associated with food or feeding. In addition, direct conditioning does not give the animal opportunity to control the delivery of the CS, whereas with operant conditioning (OC) the animal has the freedom to respond whenever it feels the need for reinforcement; its motivation to eat a particular food can thereby be monitored. Classical OC The willingness of animals to work for a reward is an index of how attractive (i.e. palatable, when applied to food) they ﬁnd the reward – how motivated they are. OC is a protocol in which animals are allowed to operate on the environment in order to achieve a reinforcement; the extent to which they have to operate in order to obtain a reinforcement can be varied in order to assess how much responding they are prepared to do in order to obtain a reinforcement. In everyday language, they have to work (e.g. press a lever, CS), in order to obtain a reward (e.g. a portion of food, US). An example of a very simple system is where sheep were trained to push open a weighted door to obtain food and did so more frequently when food-deprived [40]. The methodology is more complex than two-choice testing, but the result is a more secure measure of motivation to eat, at least for the animals in question. The number of responses required to activate reinforcement can be either a ratio (ﬁxed, FR; variable, VR) or time-limited. A progressive VR has been used to assess how hard sheep were prepared to work for different foods; the number of responses required to obtain a reinforcement of food was progressively increased until the animals ceased to respond and this provided a quantitative expression of motivation [41]. With pigs, a positive relationship was found between the concentration of sucrose in water and the maximum number of responses in a VR schedule [28]. Preferences under choice conditions can also be studied using OC techniques. For example, Dumont and Petit [42] offered sheep and cattle poor-quality forage ad libitum but a good-quality one for which they had to work, in this case by walking across the test area to obtain

g

235

a small portion of food. The more they were prepared to respond for the high-quality forage, the more palatable it was deemed to be. Although these procedures are relatively simple to set up and appear to be a promising way of comparing the palatability of different forages, the results will still depend very much on the previous experience of the foods in question, both during the animal’s lifetime and during the testing period. Two-food choice tests show that animals of many species in several physiological states make nutritionally wise choices between foods with protein : energy (P : E) ratios above and below the optimal, as long as they have been given adequate opportunity to learn the metabolic consequences of eating the two foods. OC methodology shows that sheep are willing to make at least 30 responses (FR30), i.e. to work quite hard, to obtain food reinforcement in order to obtain a ‘balanced’ diet [43]. Second-order OC Typically, OC studies on motivation to eat involve repeatedly allowing animals to respond for reinforcements of small amounts of food; this food is likely to alter the motivation to respond and thus to interfere with the interpretation of the results. Prolonged testing will increase distension of the gastro-intestinal (GI) tract and the supply of nutrients to the body. A ‘second-order’ schedule of food reinforcement was developed by Day et al. [44] involving conditioning pigs to associate a nonfood stimulus (the CS, either a tone or the noise of the food delivery equipment) with an eventual delivery of food reinforcement, i.e. they responded in order to obtain a reinforcement that was not itself food, but had become associated with the delivery of food. By this means, responses were obtained that did not diminish with time and were thought to be a better representation of motivational state than conventional OC. This methodology was then used to determine whether GI distension, as induced by bulky foods, reduced motivation to eat [45]. Portions of foods with different levels of addition of dried citrus-pulp, but equal contents of energy and nutrients were used in the second-order OC schedule, with feeder noise as the CS. The results showed that pigs responded less for the CS the more bulky the food, suggesting that feeding motivation is inversely proportional to the bulkiness of the meal. Another second-order procedure is the novel cognitive palatability assessment protocol (CPAP) claimed to be ‘. . . a robust and reliable means of assessing palatability’ [46]. Dogs were offered three objects, either boxes of different shapes or lights of different colours, one associated with no reward, the other two with rewards of 1 g of the two foods (dry or moist) under test. The animals learned to associate object with reward and consistently chose the object associated with the moist food. CPAP is more time-consuming than two-choice testing, but gives more robust results using fewer animals [47].

CAB International 2010

236

Animal Science Reviews 2010

Sham Feeding An assessment of palatability as affected by ﬂavour could be gained by preventing swallowed food from entering the stomach, avoiding the possibility of post-ingestive effects on learned responses to foods. Oesophageal ﬁstulation has been used in ruminants, involving plugging the ﬁstula to allow ingested food to pass into the rumen or removing the plug to allow food to drop from the ﬁstula (sham-feeding) when desired by the experimenter. This technique has been used by Grovum and Chapman [48] and Buchanan-Smith [49], who studied the role of ﬂavour on the intake of grass silage by sheep; it was found that acetic acid added to silage offered to sham-fed sheep depressed the amount eaten during 30 min tests administered after 5 h of fasting. Acetic and lactic acids added together, however, increased intake, as did acetic plus butyric acids. An extract of a poor-quality silage increased intake linearly with increasing concentration. Together, these results show that the intake of poor-quality silage is unlikely to be the result of low palatability.

Addition of Food Bypassing the Mouth The opposite of sham-feeding is the introduction of food (or other material) into the digestive tract without passing through the mouth, thereby modifying the metabolic status of the animal without inﬂuencing the special senses. In order to investigate the extent to which palatability inﬂuences the intake of forages, Greenhalgh and Reid [50] offered either straw or dried grass to sheep while introducing the same amount of the other food directly into the rumen through a ﬁstula, thereby keeping diet composition and digestibility constant. The difference in voluntary intake between the two feeds could then be attributed to palatability. The ﬁnding that voluntary intake of dried grass was much greater than of straw was taken to mean that palatability plays a major part in the control of food intake in ruminants. Clearly the sheep could differentiate the two foods by taste, appearance and texture but that is not the point. If, however, the sheep in this experiment had prior experience of foods of the type studied then previously learned associations might have clouded the results – experience of a food given as the sole diet would have allowed an association to be built up between the organoleptic properties of the food and its digestive and metabolic consequences, thereby confounding the aim of the experiment, which was to examine the palatability of foods in the absence of digestive differences. In addition, the procedure assumes that digestion would be exactly similar irrespective of whether food was eaten or introduced into the rumen and it was later shown that this assumption was highly questionable [51], an indication that this technique may not be suitable for measuring relative palatability of feeds. Recently, Favreau et al. [52] have

g

re-addressed this question using similar methodology and have concluded that ‘palatability’ has little effect on voluntary intake of forages by sheep even though it inﬂuences feeding behaviour. Addition of food into the rumen does not necessarily mean that the animal does not sense its ﬂavour because of regurgitation of digesta from the rumen as part of the rumination process. There exists the possibility that associations might be learned between the metabolic consequences of putting food directly into the rumen and the taste of this food via regurgitation.

Factors Inﬂuencing the Results of Palatability Tests As already made clear, there are numerous factors that can affect the outcome of palatability measurements. Some of these relate to the food, others to the animal and yet others to the animal’s experience and to the environment. Failure to take account of such factors when conducting palatability tests could result in misleading, and occasionally erroneous, conclusions being drawn.

Diurnal Variation There are many examples, particularly with grazing ruminants, of food choice varying between different times of the day. In some cases, these are likely to be correlated with changes in an animal’s metabolism and therefore optimal diet. In others, it may be environmental factors that are responsible while, especially in the grazing situation, diurnal changes in the food itself could inﬂuence choice. Even where no such diurnal variation in preference has been demonstrated, it is advisable to design experiments and tests in such a way that time of day is controlled for.

Necessity for Cues In tests where two or more foods are offered simultaneously, it is necessary for there to be some sensory differentiation between them in order that animals might learn to associate the characteristics of what they are eating with the consequences of eating those foods. Those ﬂavours naturally in foods are used by normal animals to help the learning process. In such natural conditions, especially on free range, the cues are likely to be complex and the number of foods may be large and their nutritional quality very variable. While ﬂavour is the most common cue used by mammals, with birds it is the appearance of the food that more usually serves this purpose. Even in the absence of detectable differences in ﬂavour or appearance, animals

CAB International 2010

can memorize the positions of foods in their environment to be used as cues. Differentiation by ﬂavour Many different ﬂavours have been used for the purpose of making different foods clearly identiﬁable by animals. Even if a particular ﬂavour is initially liked or disliked, animals learn its associated consequences within a few days, and then eat for nutrient balance rather than just ﬂavour; witness the example of Bitrex in the ﬁrst section of this article. Often the food has a sufﬁciently characteristic ﬂavour (or appearance) that artiﬁcial cues by way of added ﬂavour (or colour) are not necessary. In many of the examples quoted above, however, the two foods offered in choice have had different added ﬂavours (e.g. [35]). In case of doubt, it is appropriate to ﬂavour or colour the foods differently to ensure the test animals can clearly differentiate between them. There are ﬂavourconsequence interactions whereby a bitter taste can still be rejected even when animals have learned that it predicts favourable metabolic consequences (e.g. [53]) and in well-designed tests the effects of ﬂavours will be balanced in the design.

Ascorbic acid intake (mg/bird/d)

J. Michael Forbes

237

Hot

10 9

*

Hot

*

8

*

* *

*

7

*

*

6

4

*

Cool

5 0

1

2

3

4

* 5 6 Days

7

*

Cool

* 8

9

10

11

Figure 4 Choice between coloured foods by chicks for ascorbic acid before and after changes in environmental temperature [55]. Vertical arrow indicates when temperature was changed; *, signiﬁcant difference (P < 0.05) between the two groups

of an LP food, compared with birds kept in a thermoneutral environment (20 C) [56]. In addition, see the example of choice for ascorbic acid, immediately above.

Social Factors Differentiation by colour Whereas mammals predominantly use the chemical properties of foods to differentiate foods by their odour and/ or taste, birds make more use of the appearance of foods, particularly the colour. When laying hens were offered food deﬁcient in methionine, together with drinking water with or without methionine added, they did not redress the deﬁciency by drinking amounts of water sufﬁcient to alleviate the deﬁciency in the food [54]. When, however, the water was coloured differently according to whether it contained methionine or not, and the birds had learned the difference between the two, they balanced their diet for methionine, even when the position of the bottles was reversed. Similarly, exposing chicks to heat stress, which generates an increased need for ascorbic acid (vitamin C), did not inﬂuence their choice between foods sufﬁcient and deﬁcient in ascorbic acid [55]. However, when the two foods were coloured differently, the birds adjusted the ratio between their intakes of the two foods, in order to obtain approximately twice as much ascorbic acid when heat stressed as when unstressed (Figure 4).

Environmental Temperature Animals kept at environmental temperatures below their lower critical temperature have increased requirements for energy and might be expected to choose a higher proportion of a food with a lower P : E, when given in choice with a high P : E food, in order to maintain a balanced diet. Broilers kept in the cold (10 C) not only increased their total food intake but also the proportion

g

A recurring question related to palatability is how to encourage weaned animals, especially piglets, to accept food immediately after weaning. Many attempts have been made to familiarize piglets before weaning with the food, or the ﬂavour of the food, to be offered after weaning. These have included ﬂavouring the diet of the pregnant mother, the diet of the lactating sow or the creep food offered to the piglets before weaning and the subject has been reviewed by Bolhuis et al. [57]. Maternal effects Including oregano in the diet of the pregnant ewe has been shown to increase their lambs’ preference for oregano-ﬂavoured food 7 months after birth [58]. The only relevant research with pigs suggests beneﬁts, but has not yet been reported in detail [57]. In performing tests involving ﬂavours, therefore, it is not necessarily safe to expect an animal never to have been exposed to a particular ﬂavour just because it has not itself been given food with that ﬂavour. Flavouring the food for the lactating sow can inﬂuence the piglets in two ways: ﬁrstly via the milk and secondly by the piglets tasting their dam’s food. Even if they cannot get direct access to the mother’s food they might well pick up the scent, and even the taste, from the dam’s mouth and breath. A much quoted example of effects of ﬂavouring the sow’s food on piglet intake is that of Campbell [59], in which a ﬂavour was added to the sow’s lactation food or the piglets’ weaner food or both. When the ﬂavour was carried over from the sow’s to piglets’ diet the intake of the piglets was increased signiﬁcantly in the ﬁrst 2 weeks after weaning, compared with where the ﬂavour was only

CAB International 2010

238

Animal Science Reviews 2010

in the sow’s or only in the piglets’ food, or neither. By the third week, however, the difference was no longer present. It has to be said that not all the tests of this hypothesis have given positive results [57]. Group effects There can be both positive and negative effects on food choice of social interactions between conspeciﬁcs. It has been reported that simply keeping naı¨ve birds in groups, as opposed to singly, can speed up the rate at which they learn associations between food and metabolism [60]. In particular, mixing animals which have already learnt about a food with naı¨ve animals can speed up the rate at which the latter accept the food. Growing pigs given a choice of foods high and low in lysine more quickly adopted an appropriate choice between the two foods when they were penned with an individual already familiar with the foods than when penned with another naı¨ve pig [61]. However, it sometimes occurs in a free-choice situation that dominant animals ‘hog’ a favoured food, forcing shy ones to eat less-favoured food(s) particularly with groups of animals that like to eat together, i.e. with older pigs it is less of an issue than with younger pigs. In a study of the extent to which cows would trade-off proximity to a dominant individual with access to food of a high quality, cows were taught to expect a high-quality food in a coloured bin and low-quality food in a different-coloured bin [62]. When a dominant cow was eating from the high-quality bin the test cows chose to eat from the lowquality bin.

Age of Animal Newly weaned piglets have lower food intakes and more variability in their food choices than older animals but, in general, the ranking of preference of different foods in two-choice tests was very similar in both age groups [26], so that it may be more convenient to study palatability in older animals.

changes became more rapid, so selection for the highmethionine food was less evident, being 65, 62 and 58%, respectively, i.e. marginally deﬁcient [63]. It has sometimes been claimed that switching the foods in two-choice tests regularly between left and right is a good thing, because it will prevent animals eating one particular food because of its less-favoured position in the test environment. This might be true when the foods are clearly differentiated by ﬂavour or colour, but if they are indistinguishable to the senses then it can confuse the animals and confound the results of the tests. The fact that sometimes animals can overcome switching of food positions (e.g. pigs selecting for lysine [64]) should not persuade us that switching is a good practice in diet selection trials and experiments.

Physiological Condition of the Test Animals While short-term, unconditioned preference for a food is primarily determined by its sensory properties, over a longer period of testing the animal’s metabolic response to the food becomes more important and preference is determined by learned preferences and aversions. Changes in the physiology of the animal will then inﬂuence their dietary choices and preferences. There are numerous examples of how selection for protein changes in a consistent manner as animals grow [2], while pregnancy, lactation and egg-laying have more variable effects in view of the complex changes taking place during these processes. Manipulating demand by under- or overfeeding has repercussions on subsequent food choice. For example, growing pigs given LP food for several weeks choose a higher proportion of HP, when given a choice between HP and LP, than those given HP previously [65], while chickens force-fed with HP, which were leaner and had more body protein at the time of testing for choice, chose a lower proportion of HP than controls [66].

Patterns of Acceptance with Time Changing Feeder Positions Where two foods on offer differ only in the concentration of a single amino acid or micronutrient, there is likely to be no detectable sensory difference between them. Even so, animals can learn to differentiate between them if the two are consistently to be found in the same positions in the test environment; switching their positions is likely to confuse animals and result in less-effective choices. Thus, when methionine-deﬁcient broiler chickens were offered a choice of high- and low-methionine foods in the same position every day for 36 days, there was a 68% choice of the high-methionine food, which was sufﬁcient to give a balanced diet. Other treatments involved reversing the positions of the foods every 9, 6 or 3 days and as the

g

It will have become very clear by now that preferences and palatability are likely to change with the length of time elapsed since ﬁrst exposure to novel foods. The rate at which this change occurs, the time to stability and the retention of associations once conditioned, are all important in designing tests and experiments and in gaining a fuller understanding of the subject. Once again, the twochoice test is the method for which most examples are available.

Learning Time From the diet selection pathways of cows offered a choice between HP and LP concentrate foods (Figure 3) it can be

CAB International 2010

HM as a proportion of total intake

J. Michael Forbes

239

however, whereby a bitter taste can still be rejected even when animals have learned that it predicts favourable metabolic consequences (e.g. [53]).

1.0 0.9 0.8 0.7

Memory Retention Time

0.6 0.5 0.4 0

1

2

3

4

5

6

Week

Figure 5 Proportional intake of high-methionine (HM) foods (0.23%, ^; 0.26%, &) by young pigs over 6 weeks when given in choice with a low-methionine food (0.19%) [69]

seen that initially there is considerable variation in the choice made between individuals; in this case, there is a partial explanation associated with the foods to which each cow was accustomed before the period of choice feeding. Within a week, however, most were taking similar proportions, and to a signiﬁcant extent their choice was related to their demand for dietary protein [35]. Dairy cows offered two total mixed rations, one HP the other LP, also took 3–4 days to respond completely by changing their selection between the two when urea was added to or removed from both [67]. It might be expected that smaller animals, with more rapid rates of digesta transit and absorption of nutrients, would complete their learned response to the metabolic effects of novel foods in shorter times than cows. The examination of the results of offering choices to piglets [68] or chicks (Figure 4), with body weight many hundreds of times less than that of cows (Figure 3), shows that in many cases it takes only a few days for stability of choice to be established, however, irrespective of body size. Such examples of rapid acclimatization to choice feeding within a few days are in contrast to other cases in which it appears to have taken much longer for animals to stabilize. Piglets were offered two foods with different contents of methionine, one higher (2.3 or 2.6 g/kg) the other lower (1.9 g/kg) than the estimated optimal [69]. Initially, the animals ate equal amounts of each food (suggesting that neither had strongly objectionable ﬂavours) but gradually, over the 6-week period of observation, they increased the proportion of the high-methionine food until it was more than 0.8 of the total food intake (Figure 5). It is likely that the similarity between the methionine content of the two foods made learning more difﬁcult, because neither was seriously imbalanced. Those ﬂavours naturally in foods are used by normal animals to help the learning process. In such natural conditions, especially on free range, the cues are likely to be complex and the number and quality of foods may be large. There are ﬂavourconsequence interactions,

g

How long it takes for an animal to forget a learned association between sensory and metabolic properties of a food (extinction) depends on many factors, including the age of the animal, the length of the exposure and the degree of discomfort caused [70]. Lambs were conditioned for 32 days to receive starch infusions into the rumen and offered a novel-ﬂavoured food for a short time afterwards [19]. After the end of conditioning, ﬂavour preference tests were performed weekly and it took up to 8 weeks before the animals no longer showed any preference. On the other hand, young chicks reverted to no preference for food colour within a few days of cessation of conditioning with CCK [39].

Prediction of Intake from Preference The question as to whether long-term intake can be predicted from short-term measures of palatability can only be answered from a knowledge of the details of the palatability studies, particularly their time scale. It will have become clear by now that what is meant by ‘short-term’ is ill-deﬁned. If we were considering food intake then we might consider up to several days to be short-term as it is only when longer periods are examined that reasonable stability of intake is attained. With palatability, however, it might take only a few hours for animals to learn the consequences of eating foods with different sensory characteristics, and so ‘short-term’ could be as little as 6–12 h. ‘Long-term’ might be deﬁned as more than a few days, leaving ‘medium-term’ to cover the period from a few hours to a few days, roughly corresponding to the conditioning period referred to above. A compilation of the results of several published comparisons between palatability and intake has been assembled by ([26], No. 7892), an extract of which is shown in Figure 6. In many cases, preference in short-term choicefeeding tests showed the test food to be chosen to a much lesser extent than the reference food, even though they were eventually eaten in very similar amounts when offered singly in the long term. This is likely to be the result of neophobia, the fear of novelty, very commonly shown when a young animal is presented with an unfamiliar food. This demonstrates the danger of using the same reference food repeatedly in choice with different novel foods because the reference food becomes increasingly familiar with each successive test. In many other cases, foods that were initially aversive continued to be eaten at low levels even when offered alone; these were cases in which the test food was

CAB International 2010

240

Animal Science Reviews 2010 Reference

Intake (relative to intake of basal food)

1.0

Threonine (0.57) 0.9

Threonine (0.5%)

0.8

Lysine (0.7%) 0.7

Lysine (0.7%) 0.6

Methionine (0.19%) 0.5

Tryptophan (0.11%)

0.4 0.0

0.1

0.2

0.3

0.4

0.5

Preference (proportion of total intake)

Figure 6 Intake of amino acid-deﬁcient foods given singly for 6 weeks, plotted against the proportion of each food chosen when offered in choice with a reference food over the same periods (data from [11]). Numbers in brackets after each label are the concentration of the amino acid in the food (g/kg)

imbalanced in its amino acid composition, limiting growth and voluntary food intake (studies of Kirchgessner, Ettle and Roth, see [11] for references). Figure 6 shows preferences and intake for six amino acid-deﬁcient foods by growing pigs. The intakes of the foods were calculated weekly so that the reported preferences would have been predominantly conditioned.

Conclusions Conclusions can be drawn at three levels: principles, methodological and practical.

Principles We can see that if we are to use the word ‘palatability’ and to conduct palatability tests with animals then it is necessary to be quite speciﬁc about the conditions existing at the time of the measurements. What animals learn about foods as they gain experience of them is vital in determining how they accept these foods during testing. A completely novel food will be judged on its physical (appearance and texture) and chemical (odour and taste) properties – those properties that can be assessed before the food is swallowed. The palatability of such a food, however it is measured, can be considered to be unconditioned, i.e. there has been no opportunity to learn an association between ﬂavour/appearance and beneﬁcial/ harmful metabolic effects. (Note that it might be considered that innate preferences for ﬂavour or appearance are ‘conditioned’ during evolution.)

g

During the ﬁrst few days of exposure to a new food, when the association is being learnt, palatability could be said to be conditioning, i.e. in the process of being conditioned. The rate and duration of this phase will depend on the many factors discussed above. Once conditioning is complete, palatability is conditioned, i.e. there is stability in the relationship between the animal and its food. Subsequent changes in the physiology of the animal, in the nutritive value of the food, or in some aspects of the environment can put the situation into reconditioning, i.e. acclimatization to the new state. If animals make choices between foods and behave in ways that are different from our predictions, we must be very wary of thinking that they are ‘wrong’. The animal is always ‘right’ and it is up to us to ponder why our predictions are wrong.

Methodological The methodology to be adopted in any particular situation will depend on the aims of the study. In practice, the twochoice test with a reference food is easy to apply and to interpret. For more basic studies of motivation to eat, a second-order OC method is more desirable. Whatever test is used, however, there is a long list of things to be considered in design and included in the description of experiments. Animal factors Age, breed, physiological state (growing, pregnant and lactating), health, genetic potential, deviation from potential (e.g. under- or over-feeding), experience, both of test animals and their conspeciﬁcs, of food(s) to be offered. Food factors Appearance, ﬂavour (including human description), physical preparation of food, e.g. particle size, composition and nutritive balance. Environmental factors Environmental temperature, design of accommodation and feeders. Social factors Group size, competition for food, experience of other group members and maternal transmission. Testing regime Length of test and timing of monitoring during test.

Practical Is it even worthwhile to try to measure palatability? If so, which method is right for a particular situation?

CAB International 2010

J. Michael Forbes

The two-choice protocol, one of the foods being a reference food, is probably the most useful and practical method for selecting materials to be used in weaner foods for piglets. The aim is to choose food materials that will be quickly accepted and for this purpose very short-term choice tests are required, initiated at weaning and continued for no more than 24 h. There is, however, no substitute for ‘. . . conventional feeding trials . . .’ which ‘. . . should always be the ultimate test to validate the efﬁcacy of a given ingredient to improve feed intake’ [11]. In conclusion, therefore, there is a case for the use of the word ‘palatability’, however difﬁcult it might be to deﬁne the term, as long as the concepts are understood and the methodologies are appropriate and described in sufﬁcient detail. Acknowledgements The stimulus to prepare this review came from an invitation to speak on ‘Palatability’ at a conference organized by Phytobiotics Futterzusatzstoffe GmbH. I am pleased to acknowledge the very valuable comments and criticism of Fiona Reynolds, Ilias Kyriazakis, the anonymous referee and, especially, of Eugeni Roura. References 1. Rook AJ, Penning PD, Rutter SM. Limitations of palatability as a concept in food intake and diet selection studies. In: Forbes JM, Lawrence TLJ, Rodway RG, Varley MA, editors. Animal Choices. British Society of Animal Science, Edinburgh; 1997. p. 75–6. 2. Forbes JM. Voluntary Food Intake and Diet Selection in Farm Animals. 2nd ed. CAB International, Wallingford; 2007. 3. Ontario Ministry of Agriculture Food and Rural Affairs. 2008. No. 2009. Available from: URL: http://www.omafra.gov.on.ca/ english/livestock/dairy/facts/08-039.htm (accessed 16th December 2009). 4. Wikipedia. 2009. No. 2009. Available from: URL: http:// en.wikipedia.org/wiki/Palatability (accessed 16th December 2009). 5. Grill HJ, Berridge KC. Taste reactivity as a measure of the neural control of palatability. Progress in Psychobiology and Physiological Psychology 1985;11:1–61. 6. Marten GC. The animal–plant complex in forage palatability phenomena. Journal of Animal Science 1978;46:1470–7. 7. Yeomans MR. Taste, palatability and the control of appetite. Proceedings of the Nutrition Society 1998;57:609–15.

241

Food Intake in Pigs. Wageningen Academic Publishers, Wageningen, The Netherlands; 2009. p. 105–140. 11. Torrallardona D, Sola-Oriol D. Evaluation of free-choice feedstuffs preferences by pigs. In: Torrallardona D, Roura E, editors. Voluntary Feed Intake in Pigs. Wageningen Academic Publishers, Wageningen, The Netherlands; 2009. p. 215–42. 12. Baumont R. Palatability and feeding behaviour in ruminants: a review. Annales de Zootechnie 1996;45:385–400. 13. Meunier-Salaun MC, Picard M. Factors involved in feed choices in pigs and poultry. Productions Animales 1996;9:339–48. 14. Auffray P, Marcilloux JC. Studies of feeding behaviour in the adult pig. Reproduction, Nutrition, Development 1983;23:517–24. 15. Baumont R, Brun JP, Dulphy JP. Inﬂuence of the nature of hay on its ingestibility and the kinetics of intake during large meals in sheep and cows. In: XVI International Grassland Congress, Nice, France; 1989. p. 788–9. 16. Faverdin P. Regulation of food intake in lactating cows in early lactation: Studies of the role of insulin [PhD thesis]. Institut National Agronomique, Paris, France; 1985. 17. Masic B, Wood-Gush DGM, Duncan IJH, McCorquodale C, Savory CJ. A comparison of feeding behaviours of young broiler and layer males. British Poultry Science 1974;15:499–505. 18. Harrison S, Romney DL, Phipps RH, Owen E. Short term intake rate (stir) as a method of ranking the intake potential of forage mixtures by dairy cows. Proceedings of the British Society of Animal Science 1998;193(abstract). 19. Villalba JJ, Provenza FD. Preference for wheat straw by lambs conditioned with intraruminal infusions of starch. British Journal of Nutrition 1997;77:287–97. 20. Colebrook WF, Black JL, Kenney PA. Effect of sensory factors on diet selection by sheep. Proceedings of the Nutrition Society of Australia 1985;10:99–102. 21. Kenney P, Black JL. Factors affecting diet selection by sheep. I. Potential intake rate and acceptability of food. Australian Journal of Agricultural Research 1984;35:5511–63. 22. Reynolds FH, Forbes JM, Miller HM. Does the newly weaned piglet select a zinc oxide supplemented feed, when given the choice? Animal 2010;4:1359–67. 23. Favreau A, Baumont R, Duncan AJ, Ginane C. Sheep use preingestive cues as indicators of postingestive consequences to improve food learning. Journal of Animal Science 2010;88:1535–44. 24. Bruneau CD, Chavez ER. Dietary preference for cereals of nursing and weaned piglets. Livestock Production Science 1995;41:225–31.

8. Ko¨ster EP. Diversity in the determinants of food choice: a psychological perspective. Food Quality and Preference 2009;20:70–82.

25. Sola-Oriol D, Pujol S, van Kempen T, Roura E, Torrallardona D. Glucose release correlates with piglet’s preference for cereal based feeds. Journal of Animal Science 2008;86 Suppl.:97.

9. Blair R, Fitzsimons J. A note on the voluntary feed intake and growth of pigs given diets containing an extremely bitter compound. Animal Production 1970;12:529–30.

26. Sola-Oriol D, Roura E, Torrallardona D. Use of double-choice feeding to quantify feed ingredient preferences in pigs. Livestock Science 2009;123:129–37.

10. Roura E, Tedo´ G. Feed appetence in pigs: an oronasal perspective. In: Torrallardona D, Roura E, editors. Voluntary

g

27. Nelson SL, Sanregret JD. Response of pigs to bitter-tasting compounds. Chemical Senses 1997;22:129–32.

CAB International 2010

242

Animal Science Reviews 2010

28. Kennedy JM, Baldwin BA. Taste preferences in pigs for nutritive and non-nutritive sweet solutions. Animal Behaviour 1972;20:706–18. 29. Glaser D, Wanner M, Tinti JM, Nofre C. Gustatory responses of pigs to various natural and artiﬁcial compounds known to be sweet to man. Food Chemistry 2000;68:375–85. 30. Tedo´ G. The umami taste in pigs: L-amino acid preferences and in vitro recognition by the receptor dimer pt1r1/pt1r3 expressed in taste and non-taste tissues [PhD thesis]. University Auto`noma de Barcelona, Barcelona, Spain; 2009. 31. Hellekant G, Danilova V. Taste in domestic pig, Sus scrofa. Journal of Animal Physiology and Animal Nutrition 1999;82:8–24. 32. Hellekant G, Roberts T, Elmer D, Cragin T, Danilova V. Responses of single chorda tympani taste ﬁbers of the calf (Bos taurus). Chemical Senses Advance Access 2010: 35(5):383–94. 33. Oregon. 2010. The arcsine transformation. Available from: URL: http://darkwing.uoregon.edu/robinh/arcsin.txt (accessed 5th January 2010). 34. Kyriazakis I, Emmans GC, Whittemore CT. Diet selection in pigs: choices made by growing pigs given foods of different protein concentrations. Animal Production 1990;51:189–99. 35. Lawson RE, Redfern EJ, Forbes JM. Choices by lactating cows between concentrates high and low in digestible undegraded protein. Animal Science 2000;70:515–25. 36. Raubenheimer D, Simpson SJ. Integrative models of nutrient balancing: application to insects and vertebrates. Nutrition Research Reviews 1997;10:151–79. 37. Forbes JM. A personal view of how ruminant animals control their intake and choice of food: Minimal total discomfort. Nutrition Research Reviews 2007;20:132–46. 38. Kyriazakis I, Anderson DH, Duncan AJ. Conditioned ﬂavour aversions in sheep: The relationship between the dose rate of a secondary plant compound and the acquisition and persistence of aversions. British Journal of Nutrition 1998;79:55–62. 39. Covasa M, Forbes JM. Exogenous cholecystokinin octapeptide in broiler chickens: Satiety, conditioned colour aversion and vagal mediation. Physiology and Behavior 1994;56:39–49. 40. Jackson RE, Waran NK, Cockram MS. Effect of different lengths of food deprivation on the feeding motivation of sheep. In: Forbes JM, Lawrence TLJ, Rodway RG, Varley MA, editors. Animal Choices. British Society of Animal Science, Edinburgh; 1997. p. 88–9. 41. Hutson GD, Van Mourik SC. Food preferences of sheep. Australian Journal of Experimental Agriculture and Animal Husbandry 1981;21:575–82. 42. Dumont B, Petit M. An indoor method for studying the preferences of sheep and cattle at pasture. Applied Animal Behaviour Science 1995;46:67–80. 43. Hou XZ, Emmans GC, Anderson D, Illius AW, Oldham JD. The effect of different pairs of feeds offered as a choice on food selection by sheep. Proceedings of the Nutrition Society 1991;50:94A. 44. Day JEL, Kyriazakis I, Lawrence AB. The use of a 2nd-order schedule to measure feeding motivation in the pig. Applied Animal Behaviour Science 1996;50:15–31.

g

45. Day JEL, Kyriazakis I, Lawrence AB. The use of a second-order schedule to assess the effect of food bulk on the feeding motivation of growing pigs. Animal Science 1996;63:447–55. 46. Araujo JA, Milgram NW. A novel cognitive palatability assessment protocol for dogs. Journal of Animal Science 2004;82:2200–8. 47. Araujo J, Studzinski C, Larson B, Milgram N. Comparison of the cognitive palatability assessment protocol and the two-pan test for use in assessing palatability of two similar foods in dogs. American Journal of Veterinary Research 2004;65:1490–6. 48. Grovum WL, Chapman HW. Factors affecting the voluntary intake of food by sheep. 4. The effect of additives representing the primary tastes on sham intakes by oesophageal-ﬁstulated sheep. British Journal of Nutrition 1988;59:63–72. 49. Buchanan-Smith JG. An investigation into palatability as a factor responsible for reduced intake of silage by sheep. Animal Production 1990;50:253–60. 50. Greenhalgh JFD, Reid GW. Relative palatability to sheep of straw, hay and dried grass. British Journal of Nutrition 1971;26:107–16. 51. Van Niekerk AI, Greenhalgh JFD, Reid GW. Importance of palatability in determining the feed intake of sheep offered chopped and pelleted hay. British Journal of Nutrition 1973;30:95–105. 52. Favreau A, Ginane C, Baumont R. Feeding behaviour of sheep fed lucerne v. grass hays with controlled post-ingestive consequences. Animal 2010;4:1368–77. 53. Favreau A, Baumont R, Ferreira G, Dumont B, Ginane C. Do sheep use umami and bitter tastes as cues of post-ingestive consequences when selecting their diet? Applied Animal Behaviour Science 2010;125:115–23. 54. Cadirci S, Smith WK, Mc Devitt RM. Determination of the appetite of laying hens for methionine in drinking water by using colour cue. Archivs Fu¨r Geﬂugelkunde 2009;73:21–8. 55. Kutlu HR, Forbes JM. Self-selection of ascorbic acid in coloured foods by heat-stressed broiler chicks. Physiology and Behavior 1993;53:103–10. 56. Mastika IM, Cumming RB. Effect of previous experience, and environmental variations on the performance and pattern of feed intake of choice fed and complete fed broilers. Recent Advances in Animal Nutrition in Australia 1987:260–82. 57. Bolhuis JE, Ostindjer M, Van den Brand H, Gerrits WJJ, Kemp B. Voluntary feed intake in piglets: potential impact of early experience with ﬂavours derived from the maternal diet. In: Torrallardona D, Roura E, editors. Voluntary Feed Intake in Pigs. Wageningen Academic Publishers, Wageningen, The Netherlands; 2009. p. 37–60. 58. Simitzis PE, Deligeorgis SG, Bizelis JA, Fegeros K. Feeding preferences in lambs inﬂuenced by prenatal ﬂavour exposure. Physiology and Behavior 2008;93:529–36. 59. Campbell RG. A note on the use of feed ﬂavour to stimulate the feed intake of weaner pigs. Animal Production 1976;23:417–9. 60. Covasa M, Forbes JM. The effects of social interaction on selection of feeds by broiler chickens. British Poultry Science 1994;35:817. 61. Morgan CA, Kyriazakis I, Lawrence AB, Chirnside J, Fullam H. Diet selection by groups of pigs: Effect of a trained individual

CAB International 2010

J. Michael Forbes on the rate of learning about novel foods differing in protein content. Animal Science 2003;76:101–9.

243

broiler chickens. Proceedings of the Nutrition Society 1992;51:56A.

62. Rioja-Lang FC, Roberts DJ, Healy SD, Lawrence AB, Haskell MJ. Dairy cows trade-off feed quality with proximity to a dominant individual in y-maze choice tests. Applied Animal Behaviour Science 2009;117:159–64.

67. Tolkamp BJ, Kyriazakis I, Oldham JD, Lewis M, Dewhurst RJ, Newbold JR, et al. Diet choice by dairy cows. 2. Selection for metabolizable protein or for ruminally degradable protein? Journal of Dairy Science 1998;81:2670–80.

63. Steinruck U, Kirchgessner M, Roth FX. Selective methionine intake of broilers by changing the position of the diets. Archivs fur Geﬂugelkunde 1990;54:245–50.

68. Forbes JM. Integration of absorptive and post-absorptive factors in feed intake regulation and prediction with particular respect to the pig. In: Torrallardona D, Roura E, editors. Voluntary feed Intake in Pigs. Wageningen Academic Publishers, Wageningen, The Netherlands; 2009. p. 61–86.

64. Ettle T, Roth FX. Dietary selection for lysine by piglets at differing feeding regimen. Livestock Science 2009;122:259–63. 65. Kyriazakis I, Emmans GC. Diet selection in pigs: Dietary choices made by growing pigs following a period of underfeeding with protein. Animal Production 1991;52:337–46. 66. Shariatmadari F, Forbes JM. The effect of force-feeding various levels of protein on diet selection and growth of

g

69. Roth FX, Meindl C, Ettle T. Evidence of a dietary selection for methionine by the piglet. Journal of Animal Science 2006;84:379–86. 70. Riley AL, Freeman KB. Conditioned ﬂavor aversions: assessment of drug-induced suppression of food intake. In: Crawley J, et al. editors. Current Protocols in Neuroscience. John Wiley and Sons, New York; 2004. p. 1–12.

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Embryo cryopreservation in domestic mammalian livestock species C.R. Youngs1*, S.P. Leibo2, and R.A. Godke3 Address: 1 Animal Science Department, Iowa State University, 2356B Kildee Hall, Ames, IA 50011, USA. 2 Audubon Nature Institute, University of New Orleans, 14001 River Road, New Orleans, LA 70131, USA. 3 School of Animal Sciences, Louisiana State University, J.B. Francioni Hall, Baton Rouge, LA 70803, USA. *Correspondence: C.R. Youngs. Email: [email protected] 2 July 2010 4 October 2010

Received: Accepted:

Abstract The cryopreservation of preimplantation embryos of domestic mammalian livestock species has become a routine practice in animal production. Embryo cryopreservation is an integral part of genetic improvement programmes utilizing embryo transfer and also is vital to germplasm preservation programmes. The purpose of this review is to provide the reader with the background and history of embryo cryopreservation as well as insights into current trends in the cryopreservation of preimplantation embryos of cattle, sheep, goats, pigs and horses. Keywords: Embryo, Cryopreservation, Vitriﬁcation, Cattle, Sheep, Goats, Pigs, Horses Review Methodology: We searched CAB Abstracts, used references from the articles obtained by that search to check for additional relevant material, and spoke with colleagues to obtain suggestions for other sources of information pertaining to this topic.

Introduction Embryo cryopreservation is the preservation of preimplantation embryos by storage in liquid nitrogen at 7196 C. The ﬁrst successful cryopreservation of mammalian embryos was reported in 1972 by researchers working with mice [1]. This breakthrough was followed by successful cryopreservation of embryos from the cow [2], sheep [3], goat [4], horse [5] and pig [6]. Much of the impetus for research on cryopreservation of embryos from domestic mammalian livestock species stemmed from the desire to move ‘exotic’ breeds from one part of the world to another. Beneﬁts to the movement of germplasm resources as embryos rather than live animals include lower cost of transportation, minimal risk of disease transmission [7] and healthier newborn animals because of the passage (via colostrum) of maternal antibodies of locally adapted recipient females to their embryo transfer offspring. There are several reasons why embryo cryopreservation is used in embryo transfer programmes. From a practical perspective, any embryos that are not immediately transferred to recipient females after collection from the donor female may be cryopreserved and stored for

g

future use. Storage of embryos also is an option when donor females produce a greater-than-expected number of embryos following superovulation, when recipient females are unavailable or inadequate (e.g. they do not exhibit oestrus synchronously with the donor), or when it may simply be more convenient to perform embryo transfer at a later time. In species where breeding and birthing times are seasonal because of either biological or economic constraints (e.g. horses, sheep or goats), cryopreservation allows for storage of embryos until the desired season for transfer. Embryo cryopreservation is used by many livestock producers as an avenue for marketing of their purebred breeding stock. Owners of genetically valuable females are often reluctant to sell their genetically elite individuals, yet embryos from such valuable females can be harvested and sold while retaining ownership of the donor. Cryopreserved embryos from a single donor may be stockpiled prior to marketing through either live or internet auctions. In some instances, embryos produced by mating speciﬁc sires and dams may be sold prior to the time the embryos are harvested and cryopreserved. More than half of all bovine embryos transferred globally in the calendar year 2008 were cryopreserved [8].

CAB International 2010

246

Animal Science Reviews 2010

Embryo cryopreservation has become an integral part of germplasm preservation programmes. Early efforts to preserve genetic resources often focused on cryopreservation of semen because semen is easily and repeatedly obtained from males. The main disadvantage to this approach is that it requires females of the same breed as the male to generate purebred offspring. Attempts to cryopreserve oocytes of domestic mammalian livestock species have met with limited success [9], thus making embryo cryopreservation critical to germplasm preservation programmes. Initial efforts for preservation of animals in the form of embryos were geared toward minor breeds that were at risk for extinction (e.g. [10]). More recently, however, preservation efforts have been expanded to include embryos from a wide variety of both common and uncommon breeds [11]. Embryo cryobanks serve as a form of insurance against loss of biodiversity. As efforts in gene mapping progress, the preservation of rare and endangered breeds and species is becoming increasingly important to prevent the unintentional loss of valuable genes encoding disease resistance and other highly desirable traits. Embryos are also more advantageous than semen because they may be transferred to recipient females of different breeds (or, in some cases, different species [12–14]), while still resulting in the production of purebred offspring. Cryopreservation of domestic mammalian embryos has been reviewed previously [15–23], but recent advancements in embryo cryopreservation warrant an update. This review will focus on cryopreservation of preimplantation embryos of domestic mammalian livestock species, but will exclude discussion of cryopreservation of gametes (sperm and oocytes).

Two approaches have been used to minimize intracellular ice crystal formation during cryopreservation of mammalian embryos. Although the two methods differ substantially in detail, both approaches rely on the same principle: to dehydrate the embryo before it is plunged into liquid nitrogen at 7196 C. One method is referred to as ‘equilibrium cooling’ or ‘controlled slow freezing’, and the second method is referred to as ‘non-equilibrium cooling’ or ‘vitriﬁcation.’ In the ﬁrst method, the embryo is allowed to equilibrate with about 10–15% cryoprotective additive (CPA) and then is cooled to subzero temperatures at a rate low enough to assure that the embryo remains in osmotic equilibrium with the solution as it is cooled to low subzero temperatures. In the second method, the embryo is exposed sequentially ﬁrst to a relatively low concentration of CPA (usually 10–20%) and then very brieﬂy to a much more concentrated CPA solution (usually 40–50%). This concentrated CPA solution causes the embryo to undergo very rapid osmotic dehydration, equivalent to that achieved with slow cooling when the 10–15% CPA solution is slowly cooled to a low subzero temperature. But before the embryo equilibrates with the very concentrated CPA solution, it is plunged into liquid nitrogen, ‘capturing’ the embryo in the dehydrated state. Some of the intracellular solution will vitrify (change from liquid to solid without forming ice crystals) at this step in the process. Therefore, a key to successful cryopreservation by non-equilibrium cooling/vitriﬁcation is to warm the embryo at a very high rate to prevent the vitriﬁed, intracellular solution from undergoing crystallization during warming. The rationale of the vitriﬁcation method was explained in detail by the originator of the procedure many years ago [26].

Major Obstacle to Embryo Cryopreservation

Slow Cooling Equilibrium Method of Embryo Cryopreservation

The volume of preimplantation embryos consists of 75–80% osmotically free water and 20–25% non-osmotic solids and bound water [24]. As embryos are cooled below the freezing point of the solution in which they are cryopreserved, ice forms in the extracellular solution, resulting in an increase in the concentration of the solution. If cooled slowly enough, embryos respond osmotically by undergoing dehydration so as to remain in osmotic equilibrium with the solution as it is cooled to lower and lower subzero temperatures. If cooled at a higher rate, however, the embryos do not lose water quickly enough, and the intracellular solution becomes increasingly supercooled. Ultimately, this supercooled intracellular solution will freeze and form ice within the cytoplasm and cell organelles. Depending on the rate at which the frozen embryos are warmed at thawing, the intracellular ice may disrupt the delicate intracellular structure of the embryos and the embryos may die. This sequence of events has recently been reviewed brieﬂy [9] and in detail [25].

g

Freezing of Embryos The initial step in the conventional approach to preimplantation embryo cryopreservation is to place embryos into a hypertonic solution of a permeating CPA. This action creates an osmotic gradient that causes movement of osmotically free water inside the blastomeres to the extracellular spaces, causing partial dehydration of the embryo [17]. Embryos remain in the CPA for a period of 5–10 min (depending on the CPA) to allow equilibration to occur (i.e. to achieve the same concentration of CPA inside and outside the cells of the embryo). Once equilibration has occurred, each embryo is individually loaded into a 0.25 ml plastic straw [27] for freezing, and the straw is then placed into a controlledrate programmable freezer at a temperature of approximately 76 C. After allowing the embryo to cool to 76 C, ice crystal formation is induced in the solution

CAB International 2010

C.R. Youngs, S.P. Leibo and R.A. Godke

surrounding the embryo (but not inside the embryo itself) in a process known as ‘seeding.’ The formation of ice crystals ‘removes’ free water from the solution surrounding the embryo and causes an increase in the effective concentration of the CPA to which the embryo is exposed. This crystallization creates another osmotic gradient, forcing additional water to exit the embryo. As the water leaves the embryo it crystallizes, and the embryo eventually reaches osmotic equilibrium. The embryo is held at seeding temperature for approximately 10 min. After seeding, the embryo is cooled at a rate of approximately 1 C per minute to a temperature of approximately 734 C. During this slow cooling process, most osmotically active water leaves the embryo. The embryo remains at 734 C for 10 min before being plunged into liquid nitrogen for long-term storage at 7196 C. Excellent quality embryos cryopreserved and stored in an appropriate manner will maintain their viability indeﬁnitely. Minor variations in the cooling rate (from 0.3 C to 1.0 C/min) and pre-plunge temperatures (from 732 to 736 C) are used in commercial embryo transfer companies. Technicians may use either liquid nitrogen vapour or methanol bath controlled-rate freezers for conventional (equilibrium) embryo cryopreservation. A number of permeating CPAs have been utilized for embryo cryopreservation including glycerol [28], dimethylsulfoxide [29], propylene glycol [30] and ethylene glycol [31]. Each CPA has a different molecular weight, density and permeability coefﬁcient. As a result, optimal use of each CPA theoretically requires a slightly different protocol. In practice, however, most technicians have adopted a single protocol for embryo cryopreservation. Typically, a 1.4–1.5 M concentration of a permeating CPA is prepared in a base medium such as phosphate-buffered saline. Some prefer also to add a low (0.25 M) concentration of sucrose [32].

Thawing of Cryopreserved Embryos Cryopreserved embryos should be thawed in an appropriate manner to achieve optimal post-thaw viability. In contrast to the slow (>1 h) cooling used for conventional embryo cryopreservation, embryo thawing is performed at a moderately rapid rate. Typically, straws containing an embryo are removed from the liquid nitrogen tank, held in room temperature air for 3–5 s (to reduce incidence of cracks in the zona pellucida [33]), and placed into a water bath at a temperature of approximately 38 C for 25–30 s. Embryos are removed from straws and are placed in a 1 M (34.2%, w/v) solution of sucrose, a non-permeating compound that creates an osmotic gradient and causes the CPA to exit the embryo. After 10 min in the sucrose solution, embryos are transferred to a holding medium to

g

247

allow for embryo rehydration and short-term holding until transfer into a recipient female.

Modiﬁcations to the Conventional Embryo Cryopreservation Method One alternative to the embryo cryopreservation method described earlier is the one-step method [34]. In the onestep method of embryo cryopreservation, the embryo is prepared for cryopreservation in the manner previously described. However, when the embryo is loaded into the straw for freezing, the sucrose solution used to remove the CPA from the frozen-thawed embryo is loaded into the same straw, with an air bubble separating the sucrose solution from the equilibrated embryo in the CPA solution. The seeding and cooling procedures are as previously described. After thawing the embryo (also as previously described), the straw is shaken to dislodge the air bubbles and to cause the embryo to enter the sucrose solution. After approximately 10 min, the embryo is transferred into a recipient female. A second alternative to the standard embryo cryopreservation technique is the direct transfer method [35, 36]. As the name implies, embryos are thawed and then directly transferred into the uterus of a recipient female without ﬁrst removing the CPA from the embryo. Because uterine ﬂuids (histotroph) do not contain CPA and are present in a very large volume compared with the volume of the embryo, the CPA will exit the embryo after it is transferred into the uterus as it attempts to reach osmotic equilibrium. Water from the uterine histotroph will enter the embryo to restore it to its precryopreservation state. Embryos cryopreserved for direct transfer ideally are equilibrated in ethylene glycol as the CPA [37] during the freezing process. Ethylene glycol is a small molecule that rapidly crosses cell membranes. Direct transfer embryos are thawed in the same manner as previously described (3–5 s in air; 25–30 s in a 38 C water bath). However, they are transferred directly into the uterus of a synchronous recipient female as quickly as possible after thawing (without attempting to remove the CPA). The major advantage of the direct transfer method when compared with the conventional method for embryo cryopreservation is that the embryo transfer technician does not have to unload the embryo from the straw in which it was cryopreserved, remove the cryoprotectant from the thawed embryo by placement into sucrose solution, and then load the embryo into a new straw for transfer to a recipient female. As a result, this method is faster, easier and less cumbersome. Because all embryos cryopreserved using the direct transfer method are transferred without post-thaw evaluation, it may also lead to higher post-transfer pregnancy rates in instances where viable embryos are discarded because of inaccurate post-thaw subjective visual assessment of embryo

CAB International 2010

248

Animal Science Reviews 2010

viability [38]. A practical beneﬁt of the direct transfer method is that the embryo transfer technician does not need a stereomicroscope at the time of embryo thawing, which could be highly important when performing embryo transfer under extensive (low input) animal management conditions. Direct transfer is not a fad. The commercial bovine embryo transfer industry in the USA has more or less universally adopted the use of direct transfer methodology. More than 97% of beef and dairy cattle embryos collected during the calendar year 2008 in the USA were cryopreserved for direct transfer [39]. Embryos cryopreserved for direct transfer may be thawed and handled in accordance with either direct transfer or conventional approaches, giving the embryo transfer technician the option of directly transferring the embryo into a recipient or placing it into sucrose for post-thaw inspection using a stereomicroscope. There is a voluntary embryo transfer industry standard that any embryo cryopreserved for direct transfer should be frozen in an amber-coloured (yellow) straw. Many straw manufacturers imprint the initials ‘DT’ on the straw as an extra indication that the embryo within the straw was cryopreserved for direct transfer. One potential disadvantage of direct transfer may exist in locales at which ambient temperatures exceed 35 C at the time of embryo transfer. Field observations under such conditions suggest an inverse relationship between the pregnancy rate and the interval from thawing to transfer (i.e. pregnancy rates will decrease as the length of time from embryo thawing to embryo transfer increases). To overcome this potential problem, suitable recipient females should be identiﬁed and prepared to receive an embryo prior to the time that a direct transfer embryo is thawed. This action will prevent thawed direct transfer embryos from sitting in ethylene glycol at warm temperatures for several minutes prior to transfer. Controlled scientiﬁc studies are needed, however, to verify or refute those ﬁeld observations.

Factors Inﬂuencing the Pregnancy Rate of Frozen-Thawed Embryos Pregnancy and birthing rates following the transfer of embryos cryopreserved using conventional methods are inﬂuenced by numerous factors such as species, the stage of embryonic development and the breed. For example, pig embryos are very difﬁcult to successfully cryopreserve [15] without removal or relocation of intracellular lipids [40], whereas frozen-thawed cattle embryos can produce pregnancy rates comparable to those obtained after transfer of freshly harvested embryos. Bos taurus embryos survive freezing and thawing very well compared with Bos indicus embryos [41]. Small equine blastocysts survive freezing and thawing better than large equine blastocysts >300 mm in diameter [42]. Embryos of some breeds

g

(e.g. Jersey dairy cattle) do not survive freezing and thawing as well as embryos obtained from other breeds [43]. Other difﬁcult challenges for conventional cryopreservation include in-vitro-produced (IVP) embryos [20, 44] and embryos that have been bisected [45] or biopsied [46].

Ultra-Rapid Non-Equilibrium Method of Embryo Cryopreservation Vitriﬁcation is an ultra-rapid non-equilibrium method of cryopreservation. The underlying premise of vitriﬁcation is to use an extremely rapid cooling rate so that intracellular water changes from liquid to solid without forming ice crystals (i.e. it forms an amorphous ‘glass’). Other distinct advantages of vitriﬁcation are that is it is less timeconsuming (embryos spend less time in CPA solutions prior to cryopreservation) and easier to perform (although detail-oriented skilled technicians are required) than conventional embryo cryopreservation, it eliminates the need for an expensive embryo freezing machine (although vitriﬁcation machines are commercially available), and it can be practical for use under ﬁeld conditions. The initial success with vitriﬁcation of mammalian embryos was reported in 1985 by researchers working with mice [47]. Since that time, live offspring have been reported from vitriﬁed embryos of cattle [48], sheep [49], goats [50] and pigs [51], but not for horses (even though equine embryo vitriﬁcation kits are sold commercially). Although speciﬁc vitriﬁcation procedures vary among species and among laboratories, the general approach for vitriﬁcation involves placing embryos into a moderate concentration (2–3 M) of one or more CPAs for the initial dehydration of embryos. After a short time (3–5 min) in the initial vitriﬁcation solution, embryos are placed into a second vitriﬁcation solution containing a high concentration (6–8 M) of CPA for not more than 30–45 s, after which time they are placed into liquid nitrogen vapour or into liquid nitrogen itself. The very short duration of exposure of embryos to the high concentration vitriﬁcation solution is necessary to avoid CPA toxicity/ osmotic damage, and some protocols expose embryos to the second vitriﬁcation solution at lower temperatures (e.g. 4 C) to reduce the likelihood of toxicity/osmotic damage. Speciﬁc technical aspects of the vitriﬁcation procedure are well described in a recent review [52]. Many vitriﬁcation protocols use a small volume of the vitriﬁcation solution to facilitate an extremely high cooling rate. Use of small volumes necessitates specialized approaches for handling of embryos. A number of different devices have been devised for vitriﬁcation including open pulled-straw [53], cryo-loop [54] and cryo-top [55]. However, the widespread adoption of such methods by the commercial livestock embryo transfer industry seems uncertain at this time, as indicated by the fact that only 1 embryo of 219 828 in vivo-derived (IVD) bovine embryos

CAB International 2010

C.R. Youngs, S.P. Leibo and R.A. Godke

cryopreserved during the calendar year 2008 by American Embryo Transfer Association members was vitriﬁed [39]. Great promise for use of vitriﬁcation, however, does exist, especially for the application to embryos that do not cryopreserve well (e.g. IVP, porcine) when using the conventional methods described earlier. An indication of this promise was exempliﬁed in a large-scale study performed under commercial conditions [56] in which bovine IVD embryos were either conventionally cryopreserved or vitriﬁed in 0.25 ml straws for in-straw dilution. Results showed no difference in the pregnancy rate following embryo transfer. Improvement of in vitro culture conditions during in vitro embryo production will likely increase the cryotolerance of embryos [57] and enhance the use of vitriﬁcation protocols by the commercial livestock embryo transfer industry. Bovine Embryo Cryopreservation More than 55% of the 539 683 IVD embryos that were transferred worldwide during the calendar year 2008 were cryopreserved compared with less than 11% of the 254 714 IVP embryos [8]. This large difference in the proportion of IVD versus IVP cryopreserved embryos reﬂects that the commercial embryo transfer industry is: (a) relatively satisﬁed with pregnancy rates that may be obtained from the transfer of IVD cryopreserved embryos [58] and (b) reluctant to accept reduced pregnancy rates often observed with cryopreserved versus fresh IVP embryos. Given the large annual global increase in production of IVP bovine embryos, there has been increasing interest in developing cryopreservation protocols (e.g. closed pulled straw [59]) that yield comparable post-thaw survival to that obtained with conventionally cryopreserved IVD embryos. There is also need for cryopreservation protocols for biopsied embryos such as those undergoing preimplantation genetic diagnosis [60], as well as for cryopreservation media that do not contain products of animal origin as a means to reduce biosecurity concerns [61]. Under optimal transfer conditions (i.e. skilled technician, appropriately synchronized and well-managed recipient females, and excellent quality embryos), pregnancy rates of 60–70% [58, 62, 63] and 45–55% [56, 64] are expected after transfer of IVD embryos cryopreserved conventionally or via vitriﬁcation, respectively. Pregnancy rates of 40–50% [62, 65] and 45–60% [66–69] are expected after the transfer of IVP embryos cryopreserved conventionally or via vitriﬁcation, respectively.

249

embryo was implicated as the reason for the chilling sensitivity [72]. Speciﬁc biochemical compositional data [73] showed that oleic acid, which has a melting point of 13.4 C, is the predominant fatty acid present in porcine embryos. Studies on the osmotic characteristics of pig ova and embryos [74] revealed differences that may be related to embryonic stage-speciﬁc sensitivity to cryopreservation [75]. The removal of lipids from the embryo prior to conventional cryopreservation led to the birth of piglets from frozen-thawed embryos [76] and provided strong evidence that lipids hinder conventional cryopreservation of porcine embryos. Recently, a non-invasive method for removal of lipid material (delipidation) from porcine embryos prior to cryopreservation was developed [77]. Vitriﬁcation of porcine embryos has been successful (in vitro assessment) using superﬁne open pulled straws (OPS) [78] and OPS protocols using chemically deﬁned media [79]. Data on transfer of IVD embryos cryopreserved conventionally are sparse, but pregnancy rates ranged from 14 to 100% [6, 76, 80, 81]. Pregnancy rates of 60–80% are expected after the transfer of IVD embryos cryopreserved via vitriﬁcation [82–86]. Only four reports of live piglets produced following the transfer of cryopreserved IVP embryos have been published, and all utilized vitriﬁcation. Farrowing rates ranged from 33 to 60% [77, 87–89]. Caprine Embryo Cryopreservation A number of studies have been conducted on cryopreservation of caprine embryos since the ﬁrst success reported in 1976 [4]. The ease with which embryos may be obtained seems to differ among dairy, meat and ﬁbre goats. Conventional cryopreservation in glycerol [4, 90–92], dimethylsulfoxide [93, 94], or ethylene glycol [95–97] has resulted in the birth of live kids. Although cleavage and morula stage embryos have been successfully cryopreserved [94, 98], blastocyst stage embryos typically survive freezing better [99]. Conventional cryopreservation also may be used for preservation of demi-embryos produced via embryo bisection [100, 101] and embryos produced in vitro [102]. Vitriﬁcation has been used to cryopreserve IVD [50, 97, 103, 104], IVP [105], cleavage stage [106] and biopsied [107] caprine embryos. Pregnancy rates of 60–70% [91, 92, 97, 98] and 50–60% [97, 103, 104, 107] are expected after transfer of IVD embryos cryopreserved conventionally or via vitriﬁcation, respectively. Pregnancy rates of 30–40% [102] and 45–55% [105, 108] are expected after transfer of IVP embryos cryopreserved conventionally or via vitriﬁcation, respectively.

Porcine Embryo Cryopreservation Ovine Embryo Cryopreservation The pig was the last of the major farm animal species to have their embryos successfully cryopreserved. Pig embryos are extremely sensitive to temperatures below 15 C [70, 71], and the high lipid content of the porcine

g

More than three decades have passed since the ﬁrst lamb was born from a sheep embryo cryopreserved in dimethylsulfoxide [3]. Since that time, other CPAs such as

CAB International 2010

250

Animal Science Reviews 2010

glycerol [109, 110] and propylene glycol [30] have been used successfully for conventional cryopreservation of preimplantation sheep embryos. However, it seems that ethylene glycol may be the preferred cryoprotective agent [111, 112], and this is likely due to its greater permeability to embryos [113, 114]. As with other farm animal species, conventional cryopreservation may be used to preserve demi-embryos [115], although cleavage-stage embryos are difﬁcult to cryopreserve [116]. Early efforts on vitriﬁcation of sheep embryos were reported in 1990 [49, 117, 118], and more recent investigations on sheep embryo cryopreservation have focused on reﬁnement of vitriﬁcation methodologies [119–121]. Some investigators have reported equivalent pregnancy rates from embryos cryopreserved conventionally or via vitriﬁcation [38, 122, 123], including when vitriﬁed-warmed embryos were directly transferred to recipients [38, 124]. Stagespeciﬁc cryosensitivities have been observed [125], with morulae not surviving vitriﬁcation as well as the more developmentally advanced stages [126]. Some studies on vitriﬁcation show lower pregnancy rates with IVP than IVD embryos [127–129], whereas others do not [130, 131]. Vitriﬁed, biopsied embryos result in lower lambing rates than those for similar non-biopsied, vitriﬁed embryos in some [132] but not other studies [133]. Pregnancy rates of 65–75% [111, 112, 115, 123] and 50–60% [110, 123, 125, 128] are expected after transfer of IVD embryos cryopreserved conventionally or via vitriﬁcation, respectively. Pregnancy rates of 25–35% [129] and 40–50% [121, 130, 131, 134] are expected after transfer of IVP embryos cryopreserved conventionally or via vitriﬁcation, respectively.

Equine Embryo Cryopreservation Comparatively less progress has been made in cryopreservation of equine embryos [135] during the past three decades than for other farm animal species. Although breed association restrictions on registration of embryo transfer offspring may have impeded research on embryo cryopreservation, the relative inability to superovulate mares, the relative inability to produce embryos in vitro and the biology of the equine embryo itself have been major deterrents to progress. The large size of equine blastocysts (which have a large ﬂuid-ﬁlled blastocoele cavity and a small surface area to volume ratio) and the presence of an embryonic capsule [136] hinder water and CPA transport during cryopreservation. Some investigators have used trypsin to digest the capsule with varying success [137, 138]. Glycerol [139] (which is known to damage the embryo [140]), ethylene glycol [141] and methanol [142] have been used as CPA for conventional (equilibrium) freezing of IVD and IVP [143] embryos, including in modiﬁed protocols [144]. Although pregnancy rates are often lower than those obtained with fresh embryos [42, 142], frozen embryos are accepted in the

g

commercial embryo transfer industry [145] if cryopreserved at a size of less than 300 mm in diameter [138]. Vitriﬁcation has been pursued for cryopreservation of large (>300 mm in diameter) embryos since the ﬁrst pregnancy was reported in 1994 [146]. The method is well described [147] and has been reviewed recently [148]. Pregnancies (as early as day 16) have been established using OPS [149], cryoloop [150], direct transfer [151] and pre-cooled [152] protocols, but pregnancy rates can be lower than those observed with fresh embryos [152]. Pregnancy rates of 55–65% [42, 138, 144, 145] and 50–60% [151–153] are expected after transfer of IVD embryos < 300 mm in diameter cryopreserved conventionally or via vitriﬁcation, respectively. Although data are sparse for IVP embryos, pregnancy rates of 25–73% [143, 146] and 62% [149] and have been reported for embryos cryopreserved conventionally or via vitriﬁcation, respectively.

Summary Tremendous advances in cryopreservation of preimplantation embryos from domestic mammalian livestock species are clearly evident when comparing current results and scope of use in the embryo transfer industry with those from the ﬁrst few years when the technology became available [154]. Embryo cryopreservation is a routine and major part of the commercial bovine embryo transfer industry. Although research studies have shown that commercially viable post-thaw pregnancy rates may be obtained after transfer of vitriﬁed bovine embryos, the precision with which procedural steps must be followed, and the variable (and often very poor) pregnancy rates that result when technicians do not possess adequate experience, have limited adoption of vitriﬁcation technology. The development of cryopreservation methods for porcine embryos will be of great beneﬁt not only to the biomedical research community but also to the swine industry for movement of germplasm resources and for salvaging genetics at the time of a disease outbreak during depopulation/repopulation. The small ruminant (sheep and goat) purebred breeding industries have embraced embryo cryopreservation technologies and have achieved successes equal to or exceeding those observed with cattle. The tremendous growth in the equine embryo transfer industry in recent years has led to the development of commercial kits for vitriﬁcation, although conventional approaches to cryopreservation are still limited to use with embryos less than 300 mm in diameter.

Acknowledgements Partial funding from the USDA multi-state research project W-2171 ‘Germ Cell and Embryo Development

CAB International 2010

C.R. Youngs, S.P. Leibo and R.A. Godke

and Manipulation for the Improvement of Livestock’ is gratefully acknowledged.

251

16. Fahning ML, Garcia MA. Status of cryopreservation of embryos from domestic animals. Cryobiology 1992;29:1–18. 17. Leibo SP. Techniques for preservation of mammalian germ plasm. Animal Biotechnology 1992;3:139–53. 18. Palasz AT, Mapletoft RJ. Cryopreservation of mammalian embryos and oocytes: recent advances. Biotechnology Advances 1996;14:127–49.

References 1. Whittingham DG, Leibo SP, Mazur P. Survival of mouse embryos frozen to 7196 C and 7269 C. Science 1972;178:411–4. 2. Wilmut I, Rowson LEA. Experiments on the low-temperature preservation of cow embryos. Veterinary Record 1973;92:686–90. 3. Willadsen SM, Polge C, Rowson LEA, Moor RM. Preservation of sheep embryos in liquid nitrogen. Cryobiology 1974;11:560 (abstract). 4. Bilton RJ, Moore NW. In vitro culture, storage, and transfer of goat embryos. Australian Journal of Biological Sciences 1976;29:125–9. 5. Yamamoto Y, Oguri N, Tsutsumi Y, Hachinohe Y. Experiments in the freezing and storage of equine embryos. Journal of Reproduction and Fertility 1982;32(Suppl.): 399–405. 6. Hayashi S, Kobayashi K, Mizuno J, Saitoh K, Hirano S. Birth of piglets from frozen embryos. Veterinary Record 1989;125:43–4. 7. Stringfellow DA. Recommendations for the sanitary handling of in vivo derived embryos. In: Stringfellow DA, Givens MD, editors. Manual of the International Embryo Transfer Society. 4th ed. International Embryo Transfer Society, Champaign, IL; 2010. p. 65–8. 8. Thibier M. Data retrieval committee statistics of embryo transfer- year 2008. The worldwide statistics of embryo transfers in farm animals. Embryo Transfer Newsletter 2009;27(4):13–19. 9. Leibo SP. Cryopreservation of oocytes and embryos: optimization by theoretical versus empirical analysis. Theriogenology 2008;69:37–47. 10. Kobrinovitch FN, Kot VS, Gorbunov L, Youngs CR, Bezugly ND. Creation of an embryo cryobank for cattle of the Grey Ukrainian breed. Cryobiology 1999;39:349–50 (abstract). 11. Notter DR. The importance of genetic diversity in livestock populations of the future. Journal of Animal Science 1999;77:61–9. 12. Bunch TD, Foote WC, Whitaker B. Interspecies ovum transfer to propagate wild sheep. Journal of Wildlife Management 1977;41:726–30. 13. Summers PM, Shephard AM, Hodges JK, Kydd J, Boyle MS, Allen WR. Successful transfer of the embryos of Przewalski’s horses (Equus przewalskii) and Grant’s zebra (E. burchelli) to domestic mares (E. caballus). Journal of Reproduction and Fertility 1987;80:13–20. 14. Stover J, Evans J, Dolensek EP. Inter-species embryo transfer from the gaur to domestic Holstein. In Proceedings of the Annual Meeting of the American Association of Zoo Veterinarians, 3–9 October 1981, Seattle, Washington, USA; 1981. p. 122–4. 15. Niemann H. Cryopreservation of ova and embryos from livestock: current status and research needs. Theriogenology 1991;35:109–24.

g

19. Vajta G. Vitriﬁcation of the oocytes and embryos of domestic animals. Animal Reproduction Science 2000;60–61:357–64. 20. Massip A. Cryopreservation of embryos of farm animals. Reproduction in Domestic Animals 2001;36:49–55. 21. Dobrinsky JR. Advancements in cryopreservation of domestic animal embryos. Theriogenology 2002;57:285–302. 22. Guinot F. Cryopreservation of embryos of domestic species. Revue INRA Productions Animales 2005;18:27–35. 23. Vajta G, Kuwayama M. Improving cryopreservation systems. Theriogenology 2006;65:236–44. 24. Leibo SP. Cryobiology: preservation of mammalian embryos. In: Evans JW, Hollaender A, editors. Genetic Engineering of Animals: An Agricultural Perspective. Plenum Press, New York; 1986. p. 251–72. 25. Mazur P. A biologist’s view of the relevance of thermodynamics and physical chemistry to cryobiology. Cryobiology 2010;60:4–10. 26. Rall WF. Factors affecting the survival of mouse embryos cryopreserved by vitriﬁcation. Cryobiology 1987;24:387–402. 27. Massip A, van der Zwalmen P, Ectors F, De Coster R, D’Ieteren C, Hanzen C. Deep freezing of cattle embryos in glass ampoules or French straws. Theriogenology 1979;12:79–84. 28. Bilton RJ, Moore NW. Successful transport of frozen cattle embryos from New Zealand to Australia. Journal of Reproduction and Fertility 1977;50:363–4. 29. Willadsen S, Polge C, Rowson LEA. The viability of deep-frozen cow embryos. Journal of Reproduction and Fertility 1978;52:391–3. 30. Schiewe MC, Rall WF, Stuart LD, Wildt DE. Analysis of cryoprotectant, cooling rate and in situ dilution using conventional freezing or vitriﬁcation for cryopreserving sheep embryos. Theriogenology 1991;36:279–93. 31. Tervit HR. Deep freezing sheep embryos. New Zealand, Ministry of Agriculture and Fisheries. Agricultural Research Division. 1982/83 Annual Report. 1984, p. 51. 32. Massip A, van der Zwalmen P. Direct transfer of frozen cow embryos in glycerol-sucrose. Veterinary Record 1984;115:327–8. 33. Rall WF, Meyer TK. Zona fracture damage and its avoidance during the cryopreservation of mammalian embryos. Theriogenology 1989;31:683–92. 34. Leibo SP. A one-step method for direct nonsurgical transfer of frozen-thawed bovine embryos. Theriogenology 1984;21:767–90. 35. Voelkel SA, Hu YX. Direct transfer of frozen-thawed bovine embryos. Theriogenology 1992;37:23–37. 36. Voelkel SA. Cryopreservation process for direct transfer of embryos. US Patent 5,160,312 (ﬁled 6 February 1991; issued 3 November 1992).

CAB International 2010

252

Animal Science Reviews 2010

37. Voelkel SA, Hu YX. Use of ethylene glycol as a cryoprotectant for bovine embryos allowing direct transfer of frozen-thawed embryos to recipient females. Theriogenology 1992;37:687–97. 38. Baril G, Traldi AL, Cognie´ Y, Leboeuf B, Beckers JF, Mermillod P. Successful direct transfer of vitriﬁed sheep embryos. Theriogenology 2001;56:299–305. 39. Stroud B. Annual report of the AETA statistics committee for the calendar year 2008. A closer look. American Embryo Transfer Association Newsletter 2009;24(4):16–8, 20–9, 31–42. 40. Nagashima H, Kashiwazaki N, Ashman RJ, Grupen CG, Seamark RF, Nottle MB. Removal of cytoplasmic lipid enhances the tolerance of porcine embryos to chilling. Biology of Reproduction 1994;51:618–22. 41. Visintin JA, Martins JFP, Bevilacqua EM, Mello MRB, Nica´cio AC, Assumpca˜o MEOA. Cryopreservation of Bos taurus vs Bos indicus embryos: are they really different? Theriogenology 2002;57:345–59. 42. Slade NP, Takeda T, Squires EL, Elsden RP, Seidel Jr GE. A new procedure for the cryopreservation of equine embryos. Theriogenology 1985;24:45–58. 43. Steel R, Hasler JF. Pregnancy rates resulting from transfer of fresh and frozen Holstein and Jersey embryos. Reproduction, Fertility and Development 2004;16:182–3 (abstract). 44. Leibo SM, Loskutoff NM. Cryobiology of in-vitro derived bovine embryos. Theriogenology 1993;39:81–94. 45. Rorie RW, Pendleton RJ, Pool SH, Youngs CR, Godke RA. The viability of bovine ‘half’ embryos produced before and after liquid nitrogen (LN2) freezing. In: Feichtinger W, Kemeter P, editors. Future Aspects of Human in Vitro Fertilization. Springer-Verlag, Heidelberg, Germany; 1987, p. 26–35. 46. Lopatarova M, Cech S, Krontorad P, Holy L, Lalova H, Dolezel R. Conception rate after sex determination and cryopreservation of D7 bovine embryos. Veterinarni Medicina 2010;55:10–8. 47. Rall WF, Fahy GM. Ice-free cryopreservation of mouse embryos at 7196 C by vitriﬁcation. Nature (London) 1985;313:573–5. 48. Massip A, van der Zwalmen P, Scheffen B, Ectors F. Pregnancies following transfer of cattle embryos preserved by vitriﬁcation. Cryo-Letters 1986;7:270–3. 49. Schiewe MC, Rall WF, Stuart LD, Wildt DE. In situ straw dilution of ovine embryos cryopreserved by conventional freezing or vitriﬁcation. Theriogenology 1990;33:321 (abstract). 50. Yuswiati E, Holtz W. Work in progress: successful transfer of vitriﬁed goat embryos. Theriogenology 1990;34:629–32. 51. Kobayashi S, Takei M, Kano M, Tomita M, Leibo SP. Piglets produced by transfer of vitriﬁed porcine embryos after stepwise dilution of cryoprotectants. Cryobiology 1998;36:20–31. 52. Vajta G, Nagy ZP. Are programmable freezers still needed in the embryo laboratory? Review on vitriﬁcation. Reproductive BioMedicine Online 2006;12:779–96. 53. Vajta G, Holm P, Kuwayama M, Booth PJ, Jacobsen H, Greve T, et al. Open pulled straw (OPS) vitriﬁcation: a new way to reduce cryoinjuries of bovine ova and embryos. Molecular Reproduction and Development 1998;51:53–8.

g

54. Lane M, Bavister BD, Lyons EA, Forest KT. Containerless vitriﬁcation of mammalian oocytes and embryos. Nature Biotechnology 1999;17:1234–6. 55. Kuwayama M. Highly efﬁcient vitriﬁcation for cryopreservation of human oocytes and embryos: the Cryotop method. Theriogenology 2007;67:73–80. 56. van Wagtendonk-de Leeuw AM, den Daas JHG, Rall WF. Field trial to compare pregnancy rates of bovine embryo cryopreservation methods: vitriﬁcation and one-step dilution versus slow freezing and three-step dilution. Theriogenology 1997;48:1071–84. 57. Pereira RM, Carvalhais I, Pimenta J, Baptista MC, Vasques MI, Horta AEM, et al. Biopsied and vitriﬁed bovine embryo viability is improved by trans10, cis12 conjugated linoleic acid supplementation during in vitro embryo culture. Animal Reproduction Science 2008;106:322–32. 58. Spell AR, Beal WE, Corah LR, Lamb GC. Evaluating recipient and embryo factors that affect pregnancy rates of embryo transfer in beef cattle. Theriogenology 2001;56:287–97. 59. Yu XL, Deng W, Liu FJ, Li YH, Li XX, Zhang YL, et al. Closed pulled straw vitriﬁcation of in vitro-produced and in vivo-produced bovine embryos. Theriogenology 2010;73:474–9. 60. Chrenek P, Boulanger L, Heyman Y, Uhrin P, Laurincik J, Bulla J, et al. Sexing and multiple genotype analysis from a single cell of [sic] bovine embryo. Theriogenology 2001;55: 1071–81. 61. Hasler JF. Synthetic media for culture, freezing and vitriﬁcation of bovine embryos. Reproduction, Fertility and Development 2010;22:119–25. 62. Hasler JF. The current status of oocyte recovery, in vitro embryo production, and embryo transfer in domestic animals, with an emphasis on the bovine. Journal of Animal Science 1998;76(Suppl. 3):52–74. 63. Hasler JF. Factors affecting frozen and fresh embryo transfer pregnancy rates in cattle. Theriogenology 2001;56:1401–15. 64. Lopatarova M, Cech S, Holy L, Dolezel R. The effect of vitriﬁcation in open pulled straws on pregnancy rates after transfer of in vivo produced bovine embryos. Veterinarni Medicina 2006;51:454–60. 65. Machatkova M, Hulinska P, Reckova Z, Hanzalova K, Spanihelova J, Pospisil R. In vitro production of embryos from high performance cows and the development of frozen-thawed embryos after transfer: a ﬁeld study. Veterinarni Medicina 2008;53:358–64. 66. Pugh PA, Tervit HR, Niemann H. Effects of vitriﬁcation medium composition on the survival of bovine in vitro produced embryos, following in straw-dilution, in vitro and in vivo following transfer. Animal Reproduction Science 2000;58:9–22. 67. Martı´ nez AG, Valca´rcel A, de las Heras MA, de Matos DG, Furnus C, Brogliatti G. Vitriﬁcation of in vitro produced bovine embryos: in vitro and in vivo evaluations. Animal Reproduction Science 2002;73:11–21. 68. Chebel RC, Deme´trio DGB, Metzger J. Factors affecting success of embryo collection and transfer in large dairy herds. Theriogenology 2008;69:98–106. 69. Frers L, Hepburn J, Mandriaza-Munoz J, Forsyth J, Strongman K. Vitriﬁcation of bovine in vitro-produced and

CAB International 2010

C.R. Youngs, S.P. Leibo and R.A. Godke in vivo embryos. Reproduction, Fertility and Development 2009;21:131–2 (abstract). 70. Wilmut I. The low temperature preservation of mammalian embryos. Journal of Reproduction and Fertility 1972;31: 513–4 (abstract). 71. Polge C, Wilmut I, Rowson LEA. The low temperature preservation of cow, sheep and pig embryos. Cryobiology 1974;82:560 (abstract). 72. Niemann H. Sensitivity of pig morulae to DMSO/PVP or glycerol treatment and cooling to 10 C. Theriogenology 1985;23:213 (abstract). 73. Youngs CR, Knight TJ, Batt SM, Beitz DC. Phospholipid, cholesterol, triacylglycerol, and fatty acid composition of porcine blastocysts. Theriogenology 1994;41:343 (abstract). 74. Bezugly N, Youngs C, Medvedovsky A, Valigura E. Osmotic composition of pig ova and embryos. In Proceedings of the 11th Scientiﬁc Meeting of the European Embryo Transfer Association, Hannover, 8–9 September 1995. p. 128. 75. Berthelot F, Martinat-Botte´ F, Vajta G, Terqui M. Cryopreservation of porcine embryos: state of the art. Livestock Production Science 2003;83:73–83. 76. Nagashima H, Kashiwazaki N, Ashman RJ, Grupen CG, Nottle MB. Cryopreservation of porcine embryos. Nature (London) 1995;374:416. 77. Li R, Murphy CN, Spate L, Wax D, Isom C, Rieke A, et al. Production of piglets after cryopreservation of embryos using a centrifugation-based method for delipation without micromanipulation. Biology of Reproduction 2009;80:563–71. 78. Cuello C, Sanchez-Osorio J, Almin˜ana C, Gil MA, Parrilla I, Roca J, et al. Superﬁne open pulled straws vitriﬁcation of porcine blastocysts does not require pretreatment with cytochalasin B and/or centrifugation. Reproduction, Fertility and Development 2010;22:808–17.

253

86. Fujino Y, Kojima T, Nakamura Y, Kobayashi H, Kikuchi K, Funahashi H. Metal mesh vitriﬁcation (MMV) method for cryopreservation of porcine embryos. Theriogenology 2008;70:809–17. 87. Li R, Lai L, Wax D, Hao Y, Murphy CN, Rieke A, et al. Cloned transgenic swine via in vitro production and cryopreservation. Biology of Reproduction 2006;75:226–30. 88. Nagashima H, Hiruma K, Saito H, Tomii R, Ueno S, Nakayama N, et al. Production of live piglets following cryopreservation of embryos derived from in vitro-matured oocytes. Biology of Reproduction 2007;76:900–5. 89. Somfai T, Ozawa M, Noguchi J, Kaneko H, Nakai M, Maedomari N, et al. Live piglets derived from in vitro-produced zygotes vitriﬁed at the pronuclear stage. Biology of Reproduction 2009;80:42–9. 90. Pendleton RJ, Youngs CR, Rorie RW, Pool SH, Memon MA, Godke RA. Offspring born from frozen goat embryos. Louisiana Agriculture 1986;30(1):16–7, 23. 91. Chemineau P, Procureur R, Cognie´ Y, Lefe`vre PC, Locatelli A, Chupin D. Production, freezing and transfer of embryos from a bluetongue-infected goat herd without bluetongue transmission. Theriogenology 1986;26:279–90. 92. Baril B, Casamitjana P, Perrin J, Vallet JC. Embryo production freezing and transfer in Angora, Alpine and Saanen goats. Zuchthygiene 1989;24:101–15. 93. Fie´ni F, Beckers JP, Buggin M, Bruyas JF, Perrin J, Daubie´ M, et al. Evaluation of cryopreservation techniques for goat embryos. Reproduction, Nutrition, Development 1995;35:367–73. 94. Rao VH, Sarmah BC, Agrawal KP, Ansari MR, Bhattacharyya NK. Survival of goat embryos frozen and thawed rapidly. Animal Reproduction Science 1988;16:261–4.

79. Sanchez-Osorio J, Cuello C, Gil MA, Parrilla I, Maside C, Almin˜ana C, et al. Vitriﬁcation and warming of in vivo-derived porcine embryos in a chemically deﬁned medium. Theriogenology 2010;73:300–8.

95. Holm P, Petersen BA, Hepburn J, Krogh K, Dagnaes-Hansen F, Callensen H. Transfer of Angora goat embryos imported into Denmark from New Zealand under quarantine conditions. Theriogenology 1990;33:251 (abstract).

80. Mo¨dl J, Reichenbach H-D, Wolf E, Brem G. Development of frozen-thawed porcine blastocysts in vitro and in vivo. Veterinary Record 1996;139:208–10.

96. Le Gal F, Baril G, Vallet JC, Leboeuf B. In vivo and in vitro survival of goat embryos after freezing with ethylene glycol or glycerol. Theriogenology 1993;40:771–7.

81. Fujino Y, Kikuchi K, Nakamura Y, Kobayashi H, Yonemura I, Suzuki M, et al. Batchwise assessment of porcine embryos for cryotolerance. Theriogenology 2007;67:413–22.

97. Guignot F, Bouttier A, Baril G, Salvetti P, Pignon P, Beckers JF, et al. Improved vitriﬁcation method allowing direct transfer of goat embryos. Theriogenology 2006;66:1004–11.

82. Berthelot F, Martinat-Botte´ F, Locatelli A, Perreau C, Terqui M. Piglets born after vitriﬁcation of embryos using the open pulled straw method. Cryobiology 2000;41:116–24.

98. Nowshari MA, Holtz W. In vitro culture of goat morulae to blastocysts before freezing. Theriogenology 1995;44:983–8.

83. Cameron RDA, Beebe LFS, Blackshaw AW, Keates HL. Farrowing rates and litter size following transfer of vitriﬁed porcine embryos into a commercial swine herd. Theriogenology 2004;61:1533–43. 84. Cuello C, Berthelot F, Martinat-Botte´ F, Guillouet P, Furstoss V, Boisseau C, et al. Transfer of vitriﬁed blastocysts from one or two superovulated Large White hyperproliﬁc donors to Meishan recipients: reproductive parameters at day 30 of pregnancy. Theriogenology 2004;61:843–50. 85. Berthelot F, Venturi E, Cognie´ J, Furstoss V, Martinat-Botte´ F. Development of OPS vitriﬁed pig blastocysts: effects of size of the collected blastocysts, cryoprotectant concentration used for vitriﬁcation and number of blastocysts transferred. Theriogenology 2007;68:178–85.

g

99. Li R, Cameron AWN, Batt PA, Trounson AO. Maximum survival of frozen goat embryos is attained at the expanded, hatching and hatched blastocyst stages of development. Reproduction, Fertility and Development 1990;2:345–50. 100. Tsunoda Y, Tokunaga T, Okubo Y, Sugie T. Beneﬁcial effect of agar for the frozen storage of bisected embryos. Theriogenology 1987;28:317–22. 101. Nowshari MA, Holtz W. Transfer of split goat embryos without zonae pellucidae either fresh or after freezing. Journal of Animal Science 1993;71:3403–8. 102. Han Y, Meintjes M, Graff K, Denniston R, Zhang L, Ziomek C, et al. Caprine offspring born from fresh and frozen-thawed in vitro-produced embryos. Veterinary Record 2001;149: 714–6.

CAB International 2010

254

Animal Science Reviews 2010

103. El-Gayar M, Holtz W. Technical note: vitriﬁcation of goat embryos by the open pulled-straw method. Journal of Animal Science 2001;79:2436–8. 104. Hong Q-H, Tian S-J, Zhu S-E, Feng J-Z, Yan C-L, Zhao X-M, et al. Vitriﬁcation of Boer goat morulae and early blastocysts by straw and open-pulled straw method. Reproduction of Domestic Animals 2007;42:34–8. 105. Rodrı´ guez-Dorta N, Cognie´ Y, Gonza´lez F, Poulin N, Guignot F, Touze´ J-L, et al. Effect of coculture with oviduct epithelial cells on viability after transfer of vitriﬁed in vitro produced goat embryos. Theriogenology 2007;68:908–13. 106. Begin I, Bhatia B, Baldassarre H, Dinnyes A, Keefer CL. Cryopreservation of goat oocytes and in vivo derived 2- to 4-cell embryos using the cryoloop (CLV) and solid-surface vitriﬁcation (SSV) methods. Theriogenology 2003;59: 1839–50. 107. Chen A, Zhang R, Yu S. Comparative results of survival of vitriﬁed biopsied goat embryos and mouse morulae. Turkish Journal of Veterinary and Animal Sciences 2008;32:93–7. 108. Traldi AS, Leboeuf B, Cognie´ Y, Poulin N, Mermillod P. Comparative results of in vitro and in vivo survival of vitriﬁed in vitro produced goat and sheep embryos. Theriogenology 1999;51:175 (abstract). 109. Sakul H, Bradford GE, BonDurant RH, Anderson GB, Donahue SE. Cryopreservation of embryos as a means of germ plasm conservation in sheep. Theriogenology 1993;39:401–9. 110. Martı´ nez AG, Matkovic M. Cryopreservation of ovine embryos: slow freezing and vitriﬁcation. Theriogenology 1998;49:1039–49. 111. Heyman Y, Vincent C, Garnier V, Cognie Y. Transfer of frozen-thawed embryos in sheep. Veterinary Record 1987;120:83–5. 112. McGinnis LK, Duplantis Jr SC, Youngs CR. Cryopreservation of sheep embryos using ethylene glycol. Animal Reproduction Science 1993;30:273–80. 113. Sze´ll A, Shelton JN, Sze´ll K. Osmotic characteristics of sheep and cattle embryos. Cryobiology 1989;26:297–301. 114. Songsasen N, Buckrell BC, Plante C, Leibo SP. In vitro and in vivo survival of cryopreserved sheep embryos. Cryobiology 1995;32:78–91.

embryos at different stages of development. Animal Reproduction Science 1997;48:247–56. 121. Dattena M, Mara L, Bin TAA, Cappai P. Lambing rate using vitriﬁed blastocysts is improved by culture with BSA and hyaluronan. Molecular Reproduction and Development 2007;74:42–7. 122. Isachenko V, Alabart JL, Dattena M, Nawroth F, Cappai P, Isachenko E, et al. New technology for vitriﬁcation and ﬁeld (microscope-free) warming and transfer of small ruminant embryos. Theriogenology 2003;59:1209–18. 123. Bettencourt EM, Bettencourt CM, Silva JCE, Ferreira P, Matos CP, Rama˜o RJ, et al. Fertility rates following the transfer of ovine embryos cryopreserved using three protocols. Small Ruminant Research 2009;82:112–6. 124. Green RE, Santos BFS, Sicherle CC, Landim-Alvarenga FC, Bicudo SD. Viability of OPS vitriﬁed sheep embryos after direct transfer. Reproduction of Domestic Animals 2009;44:406–10. 125. Ali J, Shelton JN. Successful vitriﬁcation of day-6 sheep embryos. Journal of Reproduction and Fertility 1993;99: 65–70. 126. Shirazi A, Soleimani M, Karimi M, Nazari H, Ahmadi E, Heidari B. Vitriﬁcation of in vitro produced ovine embryos at various developmental stages using two methods. Cryobiology 2010;60:204–10. 127. Dattena M, Ptak G, Loi P, Cappai P. Survival and viability of vitriﬁed in vitro and in vivo produced ovine blastocysts. Theriogenology 2000;53:1511–9. 128. Papadopoulos S, Rizos D, Duffy P, Wade M, Quinn K, Boland MP, et al. Embryo survival and recipient pregnancy rates after transfer of fresh or vitriﬁed, in vivo or in vitro produced ovine blastocysts. Animal Reproduction Science 2002;74:35–44. 129. Martı´ nez AG, Valca´rcel A, Furnus CC, de Matos DG, Iorio G, de las Heras MA. Cryopreservation of in vitro-produced ovine embryos. Small Ruminant Research 2006;63:288–96. 130. Zhu SE, Zeng SM, Yu WL, Li SJ, Zhang ZC, Chen YF. Vitriﬁcation of in vivo and in vitro produced ovine blastocysts. Animal Biotechnology 2001;12:193–203.

115. Shelton JN. Factors affecting viability of fresh and frozen-thawed sheep demi-embryos. Theriogenology 1992;37:713–21.

131. Dattena M, Accardo C, Pilichi S, Isachenko V, Mara L, Chessa B, et al. Comparison of different vitriﬁcation protocols on viability after transfer of ovine blastocysts in vitro produced and in vivo derived. Theriogenology 2004;62:481–93.

116. Garcia-Garcia RM, Gonzalez-Bulnes A, Dominguez V, Veiga-Lopez A, Cocero MJ. Survival of frozen-thawed sheep embryos cryopreserved at cleavage stages. Cryobiology 2006;52:108–13.

132. Leoni GC, Ledda S, Bogliolo L, Naitana S. Novel approach to cell sampling from preimplantation ovine embryos and its potential use in embryonic genome analysis. Journal of Reproduction and Fertility 2000;119:309–14.

117. McGinnis LK, Youngs CR. Vitriﬁcation of ovine embryos. Theriogenology 1990;33:287 (abstr).

133. Guignot F, Baril G, Dupont F, Cognie Y, Folch J, Alabart JL, et al. Determination of sex and scrapie resistance genotype in preimplantation ovine embryos. Molecular Reproduction and Development 2009;76:183–90.

118. Sze´ll A, Zhang J, Hudson R. Rapid cryopreservation of sheep embryos by direct transfer into liquid nitrogen vapour at 7180 C. Reproduction, Fertility and Development 1990;2:613–8. 119. Sze´ll AZ, Windsor DP. Survival of vitriﬁed sheep embryos in vitro and in vivo. Theriogenology 1994;42:881–9.

134. Walmsley SE, Pollard JW, Randall AE, Morris LHA. Survival of in vitro produced ovine embryos of different developmental stages after vitriﬁcation in EFS-40. Theriogenology 1999;51:177 (abstract).

120. Naitana S, Ledda S, Loi P, Leoni G, Bogliolo L, Dattena M, et al. Polyvinyl alcohol as a deﬁned substitute for serum in vitriﬁcation and warming solutions to cryopreserve ovine

135. Rubio Martinez LM. La criopreservacio´n embrionaria en la especie equine [Embryo cryopreservation in horses]. Medicina Veterinaria 2001;18:527–546.

g

CAB International 2010

C.R. Youngs, S.P. Leibo and R.A. Godke 136. Bosquet D, Guillomot M, Betteridge KJ. Equine zona pellucida and capsule: some physiochemical and antigenic properties. Gamete Research 1987;16:121–32. 137. Legrand E, Bencharif D, Barrier-Battut I, Delajarraud H, Cornie`re P, Fie´ni F, et al. Comparison of pregnancy rates for days 7–8 equine embryos frozen in glycerol with or without previous enzymatic treatment of their capsule. Theriogenology 2002;58:721–3. 138. Maclellan LJ, Carnevale EM, Coutinho da Silva MA, McCue PM, Seidel Jr GE. Cryopreservation of small and large equine embryos pre-treated with cytochalasin-B and/or trypsin. Theriogenology 2002;58:717–20. 139. Czlonkowska M, Boyle MS, Allen WR. Deep freezing of horse embryos. Journal of Reproduction and Fertility 1985;75:485–90. 140. Bruyas J-F, Be´zard J, Lagneaux D, Palmer E. Quantitative analysis of morphological modiﬁcations of day 6.5 horse embryos after cryopreservation: differential effects on inner cell mass and trophoblast cells. Journal of Reproduction and Fertility 1993;99:15–23. 141. Hochi S, Maruyama K, Oguri N. Direct transfer of equine blastocysts frozen-thawed in the presence of ethylene glycol and sucrose. Theriogenology 1996;46:1217–24. 142. Bass LD, Denniston DJ, Maclellan LJ, McCue PM, Seidel Jr GE, Squires EL. Methanol as a cryoprotectant for equine embryos. Theriogenology 2004;62:1153–9. 143. Galli C, Colleoni S, Duchi R, Lagutina I, Lazzari G. Developmental competence of equine oocytes and embryos obtained by in vitro procedures ranging from in vitro maturation and ICSI to embryo culture, cryopreservation and somatic cell nuclear transfer. Animal Reproduction Science 2007;98:39–55. 144. Barﬁeld JP, Sanchez R, Squires EL, Seidel GE. Vitriﬁcation and conventional cryopreservation of equine embryos. Reproduction, Fertility and Development 2009;21:130 (abstr). 145. Lascombes FA, Pashen RL. Results from embryo freezing and post ovulation breeding in a commercial embryo transfer programme. In Proceedings of the Fifth International

g

255

Symposium of Equine Embryo Transfer Havenmeyer Foundation Monograph Series 3, 6–9 July, 2000 in Saari, Finland 2001. p. 95–6. 146. Hochi S, Fujimoto T, Braun J, Oguri N. Pregnancies following transfer of equine embryos cryopreserved by vitriﬁcation. Theriogenology 1994;42:483–8. 147. Carnevale EM. Vitriﬁcation of equine embryos. Veterinary Clinics of North America, Equine Practice 2006;22:831–41. 148. Grizelj J, Duchamp G, Guignot F, Vidament M, Plotto A, Mermillod P. Ultra rapid open pulled straw (OPS) vitriﬁcation is a perspective for freezing horse embryos. Veterinarski Arhiv 2009;79:105–17. 149. Campos-Chillo`n LF, Suh TK, Barcelo-Fimbres M, Seidel Jr GE, Carnevale EM. Vitriﬁcation of early-stage bovine and equine embryos. Theriogenology 2009;71: 349–54. 150. Oberstein N, O’Donovan MK, Bruemmer JE, Seidel Jr GE, Carnevale EM, Squires EL. Cryopreservation of equine embryos by open pulled straw, cryoloop, or conventional slow cooling methods. Therigenology 2001;55:607–13. 151. Eldridge-Panuska WD, Caracciolo di Brienza V, Seidel Jr GE, Squires EL, Carnevale EM. Establishment of pregnancies after serial dilution or direct transfer of vitriﬁed embryos. Theriogenology 2005;63:1308–19. 152. Hudson J, McCue PM, Carnevale EM, Welch S, Squires EL. The effects of cooling and vitriﬁcation of embryos from mares treated with equine follicle-stimulating hormone on pregnancy rates after nonsurgical transfer. Journal of Equine Veterinary Science 2006;26:51–4. 153. Araujo GHM, Rocha Filho AN, Burns SD, Burns CM, Moya-Araujo CF, Meira C. Pregnancy rates after vitriﬁcation, warming and transfer of equine embryos. Animal Reproduction Science 2010;121(Suppl.):S299–300. 154. Whittingham DG. Low temperature preservation of embryos. In: Betteridge KJ, editor. Embryo Transfer in Farm Animals: A Review of Techniques and Applications. Agriculture Canada, Ottawa, Canada; 1977, p. 50–3.

CAB International 2010

This page intentionally left blank

Animal Science Reviews 2010

Risk assessment of toxic contaminants in animal feed Alberto Mantovani* and Chiara Frazzoli Address: Food and Veterinary Toxicology Unit and WHO/FAO Collaborating Centre in Veterinary Public Health, Department of Veterinary Public Health and Food Safety, Istituto Superiore di Sanita`, Viale Regina Elena, 299 00161, Rome, Italy. *Correspondence: Alberto Mantovani. Fax: +39 (0)6 4990 2815/2658. Email: [email protected] 9 March 2010 8 June 2010

Received: Accepted:

Abstract Feed contaminants are a broad issue ranging from undesirable or unauthorized feed components or additives through to environmental pollutants and to contaminants related to speciﬁc steps of food production, such as storage or cross-contamination. In many cases, e.g. most mycotoxins and plant-derived compounds, feed contaminants may pose a risk mostly for farm animal welfare; however, feeds can also be a major vehicle for the presence of bioaccumulating pollutants in human diet, especially for vulnerable farm animal productions, such as aquaculture or grazing ruminants. Critical risk assessment issues include the characterization of toxicological hazards, the possible pathways of feed contamination and the carry-over of parent compound or metabolites to foods of animal origin as well as the pinpointing of situations that may require risk management measures. Some examples are considered in detail, taking into account the assessments performed by the European Food Safety Authority: cross-contamination by coccidiostats, the endocrinedisrupting mycotoxin zearalenone, the trace element Cr(III) and the persistent organic pollutant hexaclorobenzene. Emerging issues, such as the widespread, bioaccumulating brominated ﬂame retardants and the potentially undesirable high levels of phytooestrogens, are also discussed. Diagnostic health risk assessment, considered as a tool for decision-making in ﬁeld situations, is also reviewed with regard to unavoidable, long-standing (e.g. methylmercury) and short-term contamination instances. Management of feed contaminants relies primarily on the implementation of good practice in feed production; speciﬁc research needs may target feed sources less prone to contamination as well as safe and effective detoxifying agents. Overall, the risk assessment of feed contaminants has a critical role in veterinary public health both as a basis to deﬁne the reference thresholds for prevention, as well as guidance for risk diagnosis and ﬁeld intervention in contamination events. Keywords: Endocrine disrupters, Food safety, Veterinary public health, Toxicology, Diagnostic risk assessment

Introduction The ‘farm-to-fork’ approach promoted by the European Union [1] requires the assessment and control of major components of the food production chain, with emphasis on primary production. Feeds must satisfy the nutritional requirements of the relevant animal species; they are expected to support safe and cost-effective production of foods of animal origin as well as to ensure the welfare of farm animals [2]. Like all

g

organisms, food-producing animals depend on the quality of their living environment. Feed contamination by environmental pollutants has led to concerns in industrialized countries, e.g. the several instances of feed and/ or pasture pollution by polychlorinated biphenyls (PCB [3]). Such episodes have prompted greater attention to the environment–feed–food chains, from both the regulatory and the research standpoints. Accordingly, feed additives and contaminants feature prominently among the opinions delivered by the European Food Safety

CAB International 2010

258

Animal Science Reviews 2010

Authority (EFSA; http://www.efsa.europa.eu/), since its establishment in 2003. Main feed contaminants may include persistent compounds of environmental origin (chlorinated compounds and heavy metals), crosscontamination by feed additives intended for other species (major example are coccidiostats) and mycotoxins, a re-emerging topic as a result of globalization and climate change [4]. In addition, undesirable substances present in natural feed ingredients or pastures (e.g. glucosinolates and pyrrolizidine alkaloids) have to be considered, also since they may be quite important in certain areas. In some cases, substances deliberately added to feeds may be factors of environmental pollution, leading to ecotoxicity and, on occasion, to harmful effects on grazing farm animals as in the case of copper; this essential trace element has high requirement levels in pigs, but is also highly toxic to small ruminants grazing on pastures contaminated by pig excreta [5]. Accordingly, the EFSA panel dealing with feed additives (FEEDAP; http://www.efsa.europa.eu/EFSA/ScientiﬁcPanels/efsa_ locale1178620753812_FEEDAP.htm) has elaborated a tiered, exposure-driven approach to assess the impact on aquatic and terrestrial environments [6]. Within EFSA, the assessment of feed contaminants is carried out by the panel on contaminants in the food chain (CONTAM, http://www.efsa.europa.eu/EFSA/ScientiﬁcPanels/efsa_ locale1178620753812_CONTAM.htm). The role of feed ingredients, additives and contaminants in shaping the safety of human food has stimulated an update of the concept of zoonoses, to include long-term risk factors related to toxicological hazards; in fact, because of the broader worldwide epidemiological transition from infectious/parasitic diseases to non-infectious diseases, attention is growingly addressed also to zoonoses caused by toxic exposures of food-producing animals and related environment–feed–food chain contamination pathways [7]. Thus, the risk assessment of toxic contaminants in feeds is ultimately a key issue for veterinary public health.

Toxicological Risk Assessment of Feed Contaminants: Principles and Criteria Assessing toxicological risks to consumers derived from substances present in feeds involves the same essential steps as other areas of risk: hazard identiﬁcation and characterization, including the determination of dose– response curves, exposure assessment and risk characterization, including the identiﬁcation of potential vulnerable groups. In addition, substances present in feeds have two target populations, the animals that are primarily exposed and the consumers undergoing secondary exposure from edible animal products; thus, from the consumer’s standpoint, animal metabolism (absoption, biotransformation, deposition in tissues and products, etc.) is a major determinant of exposure [2, 3].

g

Feed additives undergo a stepwise assessment based on studies carried out according to regulatory requirements, whereas a case-by-case approach is adopted for most feed contaminants [2]. Critical issues include the characterization of toxicological hazards, the possible pathways of feed contamination as well as carry-over of parent compound or metabolites to foods of animal origin. Thus, besides the possible recommendations of maximum tolerable levels in feeds, a major issue is to pinpoint situations that may require risk management measures. Based on the aforementioned criteria, from 2004 to April 2009, the EFSA has carried out 30 assessments on natural compounds and xenobiotics of potential concern for animal feeds [8]. In the majority of cases, the levels currently experienced in feeds do not give rise to concerns for animal health and/or consumer safety. Potential adverse effects on animal health at current exposure levels could not be excluded in some instances, especially for natural plant products, e.g. gossypol in sheep; recommendations were issued to reduce the presence of several persistent organic pollutants (POP) in edible tissues and products, e.g. camphechlor. As a general point, the need for further research on carry-over from feed to foods of animal origin was identiﬁed in order to reﬁne exposure assessment [9–11]. Many potential contaminants are either natural substances or outdated chemicals that persist in the environment: thus, risk assessment cannot make use of up-to-date dossiers, or make further data requests to applicants. Available sources are mainly limited to the scientiﬁc literature (which may be scanty for ‘unfashionable’ topics) and results from monitoring plans (only for contaminants for which regulatory limits exist). This explains the many knowledge gaps that were outlined within EFSA assessments [8]. Risk assessment of toxic feed contaminants should consider both food security and food safety. In the case of most mycotoxins and undesirable plant-derived compounds, the risks at realistic levels of feed exposure are mainly for animal health and welfare [12, 13]; in turn, this may lead to impaired farm animal production and reduced availability of foods of animal origin. On the other hand, feeds can also be a major vehicle for the presence of PCB and other POP in human diet [3]. Such compounds have endocrine-disrupting properties, may pass through placenta as well into breast milk, and give rise to life-long body burden that, in its turn, may affect the next generation [14]; thus, some feed contaminants are potentially relevant to transgenerational exposure in humans [15]. The carry-over to human diet of PCB, dioxins and methylmercury through aquaculture feeds entrains a further speciﬁc point for concern; high levels of contaminants may somewhat reduce the recognized beneﬁcial action of nutrients (e.g. omega-3) for which ﬁsh is a substantial source [16, 17]. Indeed, the interactions between contaminants and nutrients are insufﬁciently understood, yet highly relevant to the safety assessment of foods and feeds [18].

CAB International 2010

Alberto Mantovani and Chiara Frazzoli

In many instances, the attention towards toxic feed contaminants has sharpened because of feed–food chain pollution episodes: PCB provide two telling examples concerning, respectively, the use of contaminated feed material (PCB-dioxin Belgian incident, [19]) and the contamination of pastures with industrial wastes, as occurred in the area of Brescia, Italy [20]. Contamination episodes have also highlighted the dual role of toxicological risk assessment in veterinary public health: a more consolidated use for the deﬁnition of reference thresholds for prevention and a newer diagnostic application as guidance for intervention activities in actual contamination events (‘diagnostic risk assessment’, [21, 22]). Issues related to risk assessment of different undesirable compounds in animal feeds are obviously different among diverse groups of compounds: for example, speciﬁc feed contaminants can be mainly associated with certain feed ingredients and the relevant areas of livestock production, such as methylmercury and ﬁsh meals [23] or corn (maize) meals and fusariotoxins [12]. Feed additives may also be dealt with as undesirable compounds, in some instances: contamination of feeds for non-target animals, as in the case of coccidiostats [8], or unauthorized use of compounds that present safety concerns [11]. Instances are also known of additives that are especially prone to contamination by toxic chemicals, e.g. some binding and anticaking agents (sepiolite, etc.) with regard to dioxins and dioxin-like compounds; however, the contaminants in sepiolite were much less bioavailable to chickens than when present in feeds [24]. This overview is supported by four examples of different substances assessed by EFSA, in order to discuss and outline potential risks to animal production and human consumers. The examples presented in the ensuing paragraphs include: (i) the cocciodiostat narasin, since cross-contamination of feeds by pharmacologically active antiprotozoal compounds is an internationally recognized issue; (ii) trivalent chromium, Cr(III), that has received less attention as compared to other non-essential elements in feeds from the safety standpoint; (iii) the oestrogenic fusariotoxin zearalenone; (iv) hexachlorobenzene (HCB), a POP that might deserve more attention because of toxicological properties and widespread exposure.

Cross-contamination by Feed Additives: The Coccidiostat Narasin It is noteworthy that coccidiostats, such as narasin, have been assessed both as additives by FEEDAP according to their intended use (up to 70 mg/kg complete feed in chickens for fattening, [25]) and by CONTAM with regard to the possible cross-contamination of feeds for nontarget species [26]. In 2006–2008, the 11 coccidiostats

g

259

authorized in the EU were also assessed by the CONTAM, with regard to the potential cross-contamination of non-target feeding stuffs [8]. The evaluation was justiﬁed by the concerns for toxicity in non-target species of compounds such as narasin, as well as by the likelihood of exposure under practical conditions: during the production of mixed feeds, a certain percentage of a coccidiostatcontaining feed batch remains in the production circuit and these residual amounts can contaminate subsequent feed batches intended for non-target animal species.

Safety Assessment of Narasin as Coccidiostat The assessment by the FEEDAP represented the basis to evaluate potential risks associated with cross-contamination. Narasin is a polyether carboxylic ionophore; similar compounds such as narasin changes ion gradients and electrical potentials in cell membranes, thus impairing cellular function and metabolism of coccidia; the toxicological effects stem directly from the pharmacodynamic properties. Narasin may be quite toxic for several species; even in chickens for fattening the margin of safety between recommended concentration in feed and that inducing initial adverse effects is only about 1.4. Toxicological tests in laboratory animals showed focal degeneration of skeletal muscles, including the diaphragm, and peripheral neuropathy in dogs: accordingly, the no-observedadverse-effect-level (NOAEL) of 0.5 mg/kg b.w./day seen in the 1-year dog study was used to set the acceptable daily intake (ADI) of 5 mg/kg body weight (equal to 300 mg/ day for a person of 60 kg body weight). On the other hand, narasin is rapidly metabolized and excreted, giving no major ground for concerns about consumer exposure. A maximum residue limit (MRL) of 0.05 mg narasin/kg for all tissues, with withdrawal time of 1 day, is sufﬁcient to keep the consumer’s exposure below 20% the ADI, even using highly conservative, default ﬁgures of human consumption, e.g. 300 g/day of chicken meat [25].

Effects on Non-target Domestic Species Toxicological data indicate the high sensitivity of some non-target domestic species. Adverse effects occurred in turkeys and rabbits at feed concentrations lower than the maximum level for chickens for fattening. Dogs, horses and cattle may also be particularly sensitive. Signs and lesions in target and non-target animals are consistent with the mode of action of polyether ionophores, including dyspnoea, lung oedema, liver cell necrosis and muscle ﬁbre damage; moreover, narasin elicits wellrecognized adverse interactions with tiamulin. Nevertheless, adverse effects are not expected to occur in nontarget animals, if cross-contamination is kept 10% the maximum amount permitted in the feed of target animals

CAB International 2010

260

Animal Science Reviews 2010

(i.e. 7 mg/kg complete feed). Assuming a feed intake of approximately 50 g/kg body weight per day, which is applicable to most monogastric food-producing animals, a 10% cross-contamination level would result in a daily exposure of 0.35 mg/kg b.w. per day; this is still below the 1-year NOAEL in the most sensitive laboratory species, i.e. 0.5 mg/kg b.w. in the dog [26].

Assessment of cross-contamination The contamination of non-target feed chains depends on physico-chemical characteristics of the substance, including adhesive strength, particle size and density, electrostatic properties and overall stability, as well as on the layout, design and operation of the facilities and equipment [27, 28]. Available data indicate that narasin is a rather stable compound, but with limited dusting potential. As for the likelihood of contamination, survey data from European states indicate that the frequency of contaminated non-target feeds is in the 1% range with concentrations generally in the 0.5 mg/kg range [26]. Narasin is extensively metabolized and does not bioaccumulate; thus, in general, cross-contamination does not raise serious concerns for consumer exposure. The liver is the target tissue for total residues, which comprise many narasin metabolites, whereas unchanged narasin shows a limited persistence in the chicken’s skin/fat. Narasin may concentrate in the egg yolk [29]; indeed, it is not intended for use in laying hens. According to kinetic studies, hen’s eggs and pig liver show the highest levels of narasin residues in edible tissues/products of non-target animals. Actual data from European surveys show that eggs are by far the animal products most likely to show narasin residues; however, both the incidence and levels of residues are generally low. In a survey of 320 egg samples, purchased from eight different European countries, narasin was found in four samples (1.25%), always at levels below 10 mg/kg [30]. Even using highly conservative, default ﬁgures of human consumption (e.g. 50 g/day of eggs), if cross-contamination is kept up to a level of 10%, the resulting human exposure would be below 50% of the ADI. The likely low frequency of signiﬁcant crosscontamination by narasin adds further reassurance to the estimation that no risk for consumers is expected with a cross-contamination level up to 10%. Good practice in the feed production chain, including responsible identiﬁcation and management of possible toxicological hazards, should be implemented to check coccidiostat cross-contamination.

Other Instances of Cross-contamination Besides coccidiostats and the well-known example of copper [5], other feed additives may give signiﬁcant problems of cross-contamination. Organoarsenicals are

g

widely used in non-European countries as antiprotozoal feed additives for pigs and poultry. Widespread use leads to signiﬁcant excretion of organic As that is further degraded into inorganic As compounds, such as arsenate [31]. Litters used as fertilizers can therefore be a source of inorganic As for agricultural soils, including pastures: the element from manure can leach into drinking water sources [32] as well as being speciﬁcally taken up by some plants [33]. Chronic exposure to inorganic arsenic is associated with a number of adverse effects in humans, including cancer, and efforts should be made to reduce the dietary intake [34]; thus, sources of additional As in the agricultural environment are not recommendable.

A Trace Element and its Organic Form: Chromium (III) and Cr(III)-Methionine Chromium(III) is the Cr-species ubiquitously present in environment and feeds, different from Cr(VI), an environmental contaminant coming mainly from industrial emissions and an established carcinogen. The background Cr(III) in feeding stuffs comes mainly from mineral sources and only to a limited extent from plant sources. Chromium(III) is currently not allowed as feed additive in Europe, but long-standing claims exist about its nutritional essentiality and potential to enhance farm animal production. Accordingly, in 2009 the FEEDAP evaluated the possible role of Cr(III) in feeds, considering both possible beneﬁts and risks [34]. As discussed in more detail below, the overall conclusion was that Cr(III) cannot be considered as an essential nutrient in either farm animals or humans; moreover, an increase of Cr(III) content in feeds may be related to safety concerns.

Bioavailability in Farm Animals Overall, dietary Cr is poorly absorbed by animals ( < 1% of the ingested dose): organic Cr(III) compounds (Cr-nicotinate, Cr-picolinate and Cr-yeasts) may be better absorbed, in the range of 1.5–10% of the ingested dose. There are no general consensus biomarkers to accurately assess whether farm animals are exposed to biologically effective concentrations of Cr(III). In order to evaluate bioavailability studies, the FEEDAP panel considered appropriate a panel of biomarkers related to glucose metabolism, as the main biological target of Cr(III), namely: increased glucose clearance rate in pigs and beef cattle, reduced plasma glucose in horses and dairy cows and reduced insulin in horses [11]. A recent study on dairy cows suggests that the metabolic effect of Cr(III) might depend on the grain source (maize or barley) in the diet, thus hinting at a possible modulation of bioavailability by feed ingredients [35].

CAB International 2010

Alberto Mantovani and Chiara Frazzoli

Endocrine-Metabolic Action and Toxicological Effects Cr(III) has a recognized endocrine-metabolic action, potentiating insulin-dependent glucose entry into the cells as well as, albeit with a lower level of evidence, modulating immune response, leptin balance and lipid metabolism [36]. However, no symptoms of Cr(III) deﬁciency have been demonstrated either in animals or humans; therefore, there is no evidence of essentiality and no requirements can be established. Accordingly, Cr(III) cannot be considered as a nutritional additive, although a series of studies intended to demonstrate favourable effects in farm animals: such favourable effects are especially evident in animals under stress conditions and depend also on natural Cr(III) background in feed, the sources and levels of supplementary Cr and the presence of other dietary factors, including other trace elements such as Fe, Mn, V and Zn [11]. Analogous limitations and uncertainties do not allow assessment of maximum tolerable levels of supplemental Cr(III) in feed based on tolerance in farm animals. Also, the toxicology of Cr(III) is not yet fully clariﬁed. Oxidative DNA damage and formation of DNA adducts are the main modes of action of Cr(VI); remarkably, following Cr(VI) entry into the cell, Cr(III) is the likely ultimate intracellular form eliciting the genotoxic mechanisms. On the other hand, Cr(III) itself has very low intracellular accessibility and is clearly much less toxic; however, recent literature suggests a possible genotoxic potential in vivo [37, 38]. Also, the intracellular accessibility of Cr(III) might be signiﬁcantly enhanced for organic forms [39]. Further to potential genotoxicity, undesirable endocrine-metabolic effects may also be considered [36]: recent studies on rats indicate that Cr(III) may modulate the programming of body fat distribution [40] as well as modulate the tissue concentrations of the essential trace elements Cu and Zn in the dam and foetus [41], although the actual relevance to safety assessment of these ﬁndings is yet to be determined.

Consumer Safety Assessment Taking into account the safety concerns and scientiﬁc uncertainties, as well as the lack of deﬁnition of an upper tolerable limit for humans, the FEEDAP Panel considered it prudent to avoid any additional exposure of the consumers resulting from the use of supplementary Cr in animal nutrition. Consumer carry-over from Cr(III) in feeds is also difﬁcult to assess, since the available data do not show a consistent pattern of tissue deposition, also because of signiﬁcant analytical uncertainties [11]. The consumer background dietary intake of Cr(III) is likely to be around 0.1 mg/day and not expected to exceed 0.3 mg/ day: the contribution of foodstuffs of animal origin from unsupplemented Cr(III) animals to the background dietary

g

261

intake for adults is estimated in the range of 16–26% [42]: the main contributors include offal, followed by muscle (including ﬁsh) and eggs, while milk is a minor source [11, 43]. However, no reliable data are available to assess the additional consumer’s exposure resulting from the use of supplementary Cr in feeds. Finally, the FEEDAP Panel noted that concerns about Cr(III) toxicity, and especially genotoxicity, should alert also about user’s safety, so that occupational exposure in the feed industry should be kept to a minimum. Against this rather unfavourable background, the FEEDAP concurrently carried out the speciﬁc assessment of the organic trace element compound, Cr(III)-methionine, that was proposed as a nutritional feed additive, at the intended Cr(III) supplementation level of 0.4–1.6 mg/kg complete feeding stuffs. The compound proved to be orally bioavailable in farm animals, on the basis of studies using the above-mentioned biomarkers related to glucose metabolism. However, concerns were raised by several data, namely: a negative effect on milk production in dairy cows at near-use levels; lack of genotoxicity studies assessing the most relevant endpoints, i.e. those typical of Cr(VI) genotoxicity; inadequate data on deposition in edible tissues and products, with consequent inability to assess consumer’s exposure upon the use of Cr(III) methionine in feeds. According to the current evidence, the presence of background inorganic Cr(III) in feeds is not a cause for concern. Any additional exposure, however, is not justiﬁed by any nutritional need, and it may entrain potential problems for farm animal and consumer safety, whether such an increase being the result of high-Cr(III) ingredients or of a deliberate, unauthorized use. Thus, any Cr(III) addition into feeds should be considered as an undesirable compound, and properly managed by good practice and surveillance.

A Mycotoxin: Zearalenone Zearalenone is a mycotoxin produced by several ﬁeld fungi, including Fusarium graminaerum and Fusarium culmorum [12]. The toxin is common in maize and maize products, but can be found in soyabeans and various cereals and grains and their by-products as well. Moreover, zearalenone seems to occur on grass, hay and straw resulting in additional exposure of animals from roughage and bedding. A Dutch study on dairy cow feed ingredients [44] showed that zearalanenone was quite prevalent, with silage and compound feed as the target feeds, whereas incidence in forage was low. Thus, there is signiﬁcant likelihood of exposure to zearalenone for most farm animal species. Noticeably, Fusarium spp. produce other toxins, particularly the immunotoxic deoxynivalenol [45] and the neurotoxic fumonisins [46]. Therefore, the cooccurrence of zearalenone with other mycotoxins is frequent and exposed animals may show mixed clinical signs.

CAB International 2010

262

Animal Science Reviews 2010

Endocrine Disruption and Other Toxicological Hazards

nevertheless, zearalenone is a real risk for farm animal welfare, as outbreaks are recorded, especially in pigs [12, 48].

In mammals, zearalenone is a powerful endocrine disrupter; it interacts with oestrogen receptors, inducing an apparent hyperoestrogenism with related effects, such as reduced fertility [47]. Besides the primary, endocrine effect, zearalenone may induce liver and immune toxicity, as well as DNA damage, in both laboratory and farm animals; such effects could also be related to the oestrogenic action [48]. Among farm animals, the young female pigs are the most sensitive species; bile concentration of the parent compound and its metabolites, alpha- and beta-zearalenol, are the best biomarkers of exposure in pigs; the zearalenol metabolites are more potent oestrogenic agents than the parent compound [49]. Rather limited data indicate that next to pigs, sheep are more sensitive than cattle, while poultry (chicken and turkey) are quite resistant to the hormonal effects of zearalenone. Insufﬁcient data exist on some so-called ‘minor’ species (nonetheless important in some countries) such as the rabbit. Zearalenone and deoxynivalenol have elicited additive effects on immune function in piglets [50]; since the exposure levels were only slightly higher than the maximum levels of deoxynivalenol and zearalenone recommended for pig feed in Europe (0.9 and 0.1 mg/kg feed, respectively [51]), further studies on the possible combined effects at or below the recommended level might be warranted. Conjugated mycotoxins are another emerging problem, since they can remain undetected by usual analytical protocols. They can occur either as soluble forms (‘masked’ mycotoxins) or bound to macromolecules, and emerge after metabolization by living organisms [52]. Zearalenone is a telling example; it has been long-recognized that the glycoside-conjugated form present in cereals can be cleaved by hydrolysis in the pig intestine, releasing the oestrogenic parent compound [53]; however, the impact of such a ﬁnding on feed safety still awaits a thorough assessment. A more recent, striking example is a study on maize naturally contaminated with the neurotoxic fumonisins: only 37–68% of the total fumonisin concentrations were found to be extractable [54]. Thus, conjugated forms may be regarded as a potential reservoir of toxicologically active mycotoxins, requiring the development of novel strategies for exposure assessment.

Approaches to Reduce Zearalenone Exposure or Bioavailability Implementing good practices along the whole feed production chain (from culture and harvesting of feed vegetable ingredients through manufacturing, storage, transport and usage in the animal farm) is the primary route for the prevention of mycotoxin contamination. Control of storage conditions of feed ingredients at risk (e.g. maize) may substantially reduce the growth chances of zearalanone-producing Fusarium spp. Monitoring of feeding stuffs will support prevention measures, allowing exposure assessment along the time course, so as to check, e.g., any effect on contamination by climate changes [4]. Also, further research is needed to establish safe levels of exposure for zearalenone in feed materials for such farm animal species as rabbits and small ruminants. Furthermore, the increasing interest in mycotoxins through climate change and, especially, global marketing of feed and food ingredients has prompted attention towards methods to combat the unavoidable presence of mycotoxins in feeds. One of the strategies for reducing the exposure to mycotoxins is to decrease their bioavailability by including various mycotoxin-adsorbing agents in the compound feed; another strategy is the degradation of mycotoxins into non-toxic metabolites by using biotransforming agents such as bacteria/fungi or enzymes. Several compounds drastically reduce the bioaccessibility of zearalenone from contaminated feeds in vitro, such as activated carbon and aluminosilicates [55, 56]. However, in a ﬁeld study on pig, the most sensitive species, the modiﬁed aluminosilicate montmorillonite (0.4% in feed) failed to reduce the bioavailability of zearalanenone present at a concentration of 1.2 mg/kg feed [57]. Interesting results concerning the biodegradation of zearalenone have been achieved using a soil bacterial mixture [58], soil fungi Rhizopus spp. [59] and the yeast strain Trichosporon mycotoxinivorans [60]. Thus, with regard to zearalenone and mycotoxin, in general, detoxifying agents are doubtlessly interesting, but much more information is needed.

Carry-over

Other Endocrine-Disrupting Natural Substances in Feeds

Given the rapid biotransformation and excretion of zearalenone in animals, secondary human exposure resulting from carry-over from feeds to foods of animal origin appears to be low [12, 49]. The available evidence indicates that animal products generally contribute only marginally to the human dietary exposure to zearalenone;

Natural plant compounds provide a range of other undesirable endocrine-disrupting substances in feeds that require a thorough assessment of feed composition in order to avoid any damage to animal production. Noticeably, the low feed-to-food carry-over is a feature shared by most such compounds: a well-known example

g

CAB International 2010

Alberto Mantovani and Chiara Frazzoli

is goitrogenic thiocianates deriving from glucosinolates in Brassica sp., which may impair animal welfare and production but is considered to be of limited signiﬁcance for the safety of foods of animal origin [13]. A rather different instance is represented by the socalled ‘phytooestrogens’, where the soya isoﬂavones genistein and daidzein feature prominently, since soybean is a major feed ingredient: phytoestrogens interact with oestrogen receptors and may impinge on several other biological mechanisms eliciting either beneﬁcial or adverse effects, in relation to factors such as dose and lifestage of the exposed organism [61]. Physiological status also inﬂuences the bioavailability of soybean phytooestrogens in cattle [62]. Isoﬂavones are not widely regarded as undesirable compounds in feeds, and they might also exert a positive effect on animal production through antioxidant activity [63]. However, recent studies raised possible concerns about the use of red clover, a widespread, phytooestrogen-rich forage, in cattle: a signiﬁcant presence of phytooestrogens was found in manure [64], whereas milk from cattle-fed red clover has detectable levels (in the 0.5 mg/l range) of the phytooestrogen metabolite equol [65]. Bovine milk with high equol level showed estrogenic activity in vitro and reduced the expression of oestrogen receptor-beta in the uterus of immature mice, which may be interpreted as an early molecular marker of endocrine disruption [66]. Thus, further research is warranted to assess whether the content of phytooestrogens in animal nutrition should be reduced in order to support the safety of certain animal products.

263

are associated with a multiplicity of effects. The liver is one major target organ; indeed, porphyria and the induction of cytochrome P-450 enzymes appear as the primary effects of HCB. Other main effects are immune dysregulation, involving macrophage and pro-inﬂammatory activation [67], as well as ovarian toxicity and apparent antioestrogenicity with reduced circulating oestradiol and tissue oestrogen receptors [68]. While HCB is negative in most genotoxicity tests, metabolism of HCB can result in the formation of reactive intermediates such as epoxides, which can covalently bind with proteins and DNA. HCB induces liver tumours in female rats, while male rats appear much less vulnerable: the mode of action may be related to long-term alterations in intercellular gap junctional communication. Noticeably, the gender-related susceptibility to hepatic tumours is not dependent on ovarian hormonal output, as it is observed also in ovariectomized rats [69]. Also in farm animals, HCB has a low acute toxicity, whereas medium- to long-term feeding exposure elicited liver effects in pigs and lambs and offspring mortality in mink: NOAEL were in the 1.0–0.1 mg/kg feed range, equivalent to a 0.05 mg/kg b.w. range. Mammals appear to be more susceptible than chickens and rainbow trout that show adverse effects only at feed concentrations higher than the 5–10 mg/kg range [9]. Feed Contamination and Carry-over

Hexachlorobenzene (HCB) was introduced as an agricultural pesticide in 1945, and was banned in 1981 for agricultural use in the European Community [9]. The current exposure to HCB is thus mainly because of its persistence and bioaccumulation. Indeed, HCB is quite volatile, highly lipophylic and among the most persistent environmental pollutants. As a result, it can be transported over long distances and can bioaccumulate in the lipid component of tissues; accordingly, HCB is included in the international protocols on POP (United Nation Environment Programme: http://www.chem.unep.ch/pops/). Nevertheless, also current release sources do exist as a by-product of incomplete incineration processes, leakage from old dump sites and inappropriate manufacturing, including waste disposal of a number of chlorinated compounds, such as solvents and pesticide formulations. Overall, HCB is ubiquitously present in environmental and biological samples world-wide.

The main exposure factors for farm animals are related to the use of animal-derived oils in feeds (as in aquaculture) and also to vegetable fats as well to grazing in or collecting feeds from highly polluted areas. Much more than adverse effects in farm animals, however, the main issue is the feed-to-food carry-over, that might lead to a signiﬁcant exposure of humans consuming foods derived from animals on contaminated feeds. HCB accumulation ratios (i.e. steady state of body accumulation/feed concentration) in milk and eggs are estimated in the 2.0–10.5% and 1.3–5.5% ranges of the ingested dose, respectively; in ﬁsh about 80–90% of HCB assumed through feed can be retained; for pigs and chickens, the accumulation ratio was estimated in the 8–11 and 11–30% range, respectively [9]. As with farm animals, human dietary exposure to HCB is linked to certain food commodities. Available data indicate that fatty ﬁsh and ﬁsh-derived products, particularly ﬁsh oils, generally contain the highest levels of HCB. Ruminants being reared in polluted areas might also be a major, local source of human exposure, as it has been observed in areas highly polluted by other POP [3].

Toxicological Hazards

Consumer Safety Assessment

HCB is readily absorbed in humans and animals; whereas acute toxicity is low, medium- and long-term exposure

In general, foods of animal origin are by far the major source of HCB dietary exposure in the general

A POP: Hexachlorobenzene

g

CAB International 2010

264

Animal Science Reviews 2010

population, although high levels may be occasionally found in lipid-containing plant products, such as pumpkin seeds and vegetable oils, from contaminated areas. Therefore, controlling the presence of HCB in animal feed is a critical step to protect the human population. A special instance of exposure to HCB, as with other POP, is breastfeeding: a recent Spanish study showed that serum levels in 4-yearold children who were breastfed were two-fold than in formula-fed children [70]. While the beneﬁcial effects of breastfeeding are not in question, maternal body burden related to previous dietary exposure is one major factor in POP contamination of human breast milk [71]. Notwithstanding its diffusion and persistence, monitoring of human HCB exposure has received a somewhat limited attention. Some biomonitoring studies indicate that the HCB body burden in the general population is comparable to that of other major POP [70, 72, 73]. Nevertheless, the available data also show a considerable decline of HCB presence in human foods over the last few decades; a constant decrease of the concentration of HCB and other POP was detected in the adipose tissue of Swedish cattle and especially pigs in the 1991–2004 period [74]. The suggested health-based guidance value for HCB, 170 ng/kg b.w. per day, is based on non-neoplastic liver effects (including enzyme induction and mitochondrial alterations) and dates back to 1997 [75]; thus, it might not account completely for effects of potential concern (endocrine disruption and immunotoxicity) that have been more thoroughly investigated afterwards. However, such reference value shows safety margins in the 50–100fold range compared with more recent dietary exposure estimates, except for breastfed infants [9]. Therefore, targeted monitoring programmes of the environment and feed ingredients are likely to reduce residual concerns related to HCB. On the other hand, one may not completely exclude additive effects on sensitive toxicological targets by the concurrent exposure to traces of different contaminants that can accumulate in certain food matrices: thus, risk reduction strategies towards the overall burden of POP may be developed for vulnerable sectors of the feed production chain, e.g. processing or blending with vegetable oil or ﬁsh oil used in aquaculture feeds [76, 77].

Emerging POP Other groups of organic pollutants with potential for persistence and bioaccumulation have emerged after the ban of PCB, HCB and chlorinated insecticides. In particular, brominated ﬂame retardants (BFR) deserve attention because of widespread environmental exposure, lipophyllicity, as well as carry-over in animal products [78, 79]. BFR include polybrominated diphenyl ethers (PBDE), tetrabromobisphenol A, hexabromocyclododecane and other chemicals [10].

g

BFR do not elicit marked systemic toxicity; however, rodents exposed to PBDE or hexabromocyclododecane during prenatal and postnatal development show signs of impaired thyroid, immune and neurological functions as well as reduced liver vitamin A, whereas tetrabromobisphenol A appears less potent [80–82]. Thus, available data suggest that lifestage and nutritional status may modulate the vulnerability to BFR effects. The contamination of edible tissues may show speciﬁc patterns as compared to other POP. A study on lactating cows exposed to a diet naturally contaminated with higher substituted brominated diphenyl ethers indicated that meat may be more signiﬁcant than dairy products to consumer exposure [79]; a signiﬁcant persistence in muscle was also observed in trout exposed through the feed to hexabromocyclododecane [83]. A feeding trial on salmon showed that PBDE mixtures have very high (close to 100%) accumulation efﬁciency, with skinned ﬁllet accumulating approximately 50% of the PBDE intake [78]. Similar to other lipophyllic contaminants, the PBDE feed-to-food carry-over may contribute to transgenerational exposure as well, in particular because of the widespread presence in breast milk [84]. The occurrence of several BFR in the environment, feeds, foods and human biological samples, has led to bans on the production and use of some PBDE technical products; accordingly, the EFSA has identiﬁed the most signiﬁcant BFR to be considered in feed and food monitoring programmes [10]. However, so far, only limited data are available to assess the possible risks for farm animals and consumers arising from feed contamination and maximum tolerable levels in feeds are yet to be set.

From Prevention to Field Intervention: Applying Diagnostic Assessment to Feed Contaminants Further to its key role in guiding prevention policies, standard setting and regulation of hazardous chemicals or practices, risk assessment can have a major role to guide intervention; in particular, a diagnostic assessment points at determining whether a problem exists and, if so, its magnitude and causes [85]. Thus, problems faced by the diagnostic assessment may range from environmental impacts on health population to the deﬁnition of diffuse or point pollution source(s). For instance, many environmental threats to health are systemic in nature, involving different compartments and food chains, as well as social and economic factors. Thus, scientiﬁc guidance is required to deal with challenges posed by integrated exposure impact assessment associated with multiple contamination pathways, possible cumulative effects and multifactorial health outcomes: a telling example is the complex pollution of food chains, including pastures and animal feeds, associated with the disposal of electronic waste in developing countries [21, 22]. Also, whereas direct causes may be clearly identiﬁable, impacts of contamination may be far reaching,

CAB International 2010

Alberto Mantovani and Chiara Frazzoli

because of the movement of environmental vectors and trade of goods; one recent example is the international melamine incident, where the global distribution of products and delay in reporting combined to give the unexpected proportions of the episode [86]. When real-life contamination of animal feeds is involved, the agricultural features of the area are of paramount importance, such as main farm animal species and products, as well as size, distribution and type (e.g. pasture rearing) of farming: for instance, the PCB pollution incident of Brescia (Italy) led to a signiﬁcant exposure of cattle and humans, because the pollution source was close to small farms of dairy cattle using local feed and milk was also mostly consumed locally [87]. Closeness in space and time is a proxy measure for environmental exposures and evidence of space-time clustering may suggest point-source exposures: for instance, in the case of the widespread POP, detection of scattered high-level samples in food surveillance programmes may suggest point contamination sources [88]. Both live animals grazing or consuming locally produced feed and their products (especially milk, since it is produced continuously by live animals) [89] are the more sensitive indicators for monitoring as they are high in the bioaccumulation chain; information can be also gained from different exposure levels of young and old animals, indicating whether contamination is recent or/and persistent [90]. Exposure patterns or contamination source(s) cannot be characterized solely by monitoring of animals and products; however, in such cases as dioxinlike chemicals, the distribution of congeners may also provide indication of pollution sources [88]. Feeds themselves may be an important source of information. For instance, locally produced feeds may reﬂect contamination by air deposition (e.g. dioxins from incinerators [89]) or pasture deposition following accumulation in sediments of waterways used for irrigation (e.g. PCB incident of Brescia, Italy [87]). Investigation on feeds may support the diagnostic assessment of yet insufﬁciently known issues, such as: the environmental transfer of toxic elements, e.g. cadmium and ﬂuorine, from fertilizers [91], or the transfer of endocrine disrupters (e.g. phthalates) from consumer products to urban waste spread over pastures [92, 93]. Therefore, whenever feasible, local feeds and farm animals should be valued as an essential component of an integrated diagnostic assessment of environmental risks; whereas measures on different environmental matrices and media may be too fragmented and human epidemiological data are often very limited, information on feeds and animals might provide a relevant ‘interface’ scenario between environment and human exposure. Diagnostic risk assessment of situations involving contamination of feeds or pastures should give attention to substances and exposure levels associated with markers of impaired function of target systems (e.g. nervous, immune and reproductive), especially in the developing

g

265

life stages; these effects, while of public health signiﬁcance in the long range, are more directly related to prolonged and/or repeated low-level exposures (such as those from the feed-food chain) rather than the more clinically ‘heavy’ effects such as cancer. One example, though not directly related to animal farming, is the association between altered thyroid markers and the consumption of POPpolluted ﬁsh in the USA Great Lake region [94]. As we already mentioned for HCB, feed contaminants may ultimately lead to trans-generational exposure in humans. A signiﬁcant example is the carry over from aquaculture feeds, which might interfere with the beneﬁcial effects of ﬁsh on human development, under speciﬁc consumption levels of women up to one year before pregnancy and during pregnancy and breastfeeding [16, 17]. Apart from speciﬁc pollution incidents, methylmercury is the main contaminant of concern in large predatory ﬁsh; methylmercury can affect both neurodevelopment and birth weight of offspring. Noticeably, the contamination levels (methyl-mercury, PCB and dioxins) of farmed ﬁsh do not differ signiﬁcantly from caught ﬁsh, this being primarily attributable to the use of ﬁsh meal in aquaculture feeds [16]. A secondary, but severe, consequence of the presence of methylmercury (and other contaminants, as well) is the potential impairing of the ﬁsh nutritional beneﬁts. In particular, the intake of omega-3 fatty acids from ﬁsh is recognized as beneﬁcial to brain development, i.e. the same target of methylmercury [16]. An analysis of ﬁsh marketed in Spain showed salmon, mackerel and red mullet as the species with the highest content of omega-3 fatty acids; the mean daily intake of mercury through ﬁsh consumption was 9.9 mg. It should be noted that this value was calculated for a ‘standard’ 70-kg adult (not the most vulnerable consumer) and that intakes of both omega-3 and mercury could be signiﬁcantly higher in high-consumption groups. Within these limits, the consumption of most ﬁsh species should not give rise to health risks; on the other hand, the authors recommend the issuing of recommendations to consumers about the risks and beneﬁts of main types of ﬁsh and the frequency of consumption [95]. In complete feeding stuffs for ﬁsh 8% of samples exceeded the maximum tolerated level in EU, 1 mg/kg total feed. The resulting contamination levels in farmed salmonids indicate that the weekly consumption of two ﬁsh meals, as recommended by nutritionists, does not pose any appreciable health risk to consumers; however, limited data exist for other farmed ﬁsh species, that can be, nonetheless, important for consumer’s intake [13]. Thus, in this case, diagnostic risk assessment serves as a basis for management and measures (including consumer’s advice) in order to reduce risks and preserve nutritional beneﬁts of such a major food of animal origin. Moreover, risks and beneﬁts are speciﬁcally relevant to prevent potential risks for the next generation(s), lending further support to the relevance of safe feeds to the new concept of ‘sustainable food safety’ [15].

CAB International 2010

266

Animal Science Reviews 2010

Besides providing the basis for management of longstanding problems, diagnostic health risk assessment can also set the basis for decision-making and follow-up of more acute feed contamination events. In the case of the Belgian PCB-dioxin incident of 1999, the assessment pivoted on the quantiﬁcation of the total PCB amount added through recycled fat (150 kg total PCB, containing 3 g of tetrachloro-dibenzodioxin (TCDD)-equivalent dioxin-like compounds), the extent of contamination (more than 2500 farms), the identiﬁcation of the most contaminated animals (reproduction poultry, both hens and chickens) and food products (chicken meat) and the most exposed population (Belgian consumers); as well as the proportion of the poultry chain affected [96]. Although chicken meat had PCB and dioxin levels that were more than 100 times above maximal recommended values, Bernard et al. [96] considered that a doubling of the PCB and dioxin burden of the young adult population (including women in childbearing age) would have required the consumption of 10 and 20 highly contaminated meals, respectively, which was quite unlikely, with the possible exception of farmers consuming their own products. On the other hand, van Larebeke et al. [28] considered that the exposure resulting from the incident could have summed up with a somewhat high background intake; thus, a conservative estimate of the health impact could not exclude an increase of cancer cases and an impact on neurobehavioural development of newborns, which is not only a sensitive endpoint but also difﬁcult to quantify. The impressive advances of the risk assessment science should also support the development of operational tools for screening toxicological markers, early identiﬁcation of pollutants and/or warning signals in animal populations, their products and/or living environments (e.g. BEST technological platform [21]): the transfer of such tools from the laboratory to the ﬁeld would strengthen the diagnostic application of risk assessment. In conclusion, implementing diagnostic health risk assessment has speciﬁc goals [22]: (i) to examine an ongoing alarm and understand its potential impact; (2) to generate and/or adjust a feasible and ﬂexible intervention system on ﬁeld situations, deﬁning more vulnerable targets (e.g. sectors of animal farming and food production chains) and modalities; (3) to evaluate and, if appropriate, generate a risk management plan for contaminants, such as BFR, for which no widespread monitoring system has existed until now; (4) to examine the efﬁcacy of risk management measures; and, last but not least, (5) to communicate appropriate messages to policy makers, enterprises and other stakeholders.

Conclusions Feed contaminants are a broad and diverse ﬁeld ranging from undesirable or unauthorized feed components or

g

Figure 1 In its version of sustainable food safety (SFS), the ‘one medicine’ covers the environment-feed-foodbreastmilk carry-over, where risk assessment (RA) is pivotal g 2009, Noodles.ONLUS

additives, through to environmental pollutants able to bioaccumulate and to pose hazards to vulnerable animal productions (e.g. aquaculture), as well as to contaminants related to speciﬁc feed production steps, such as storage (e.g., mycotoxins) or cross-contamination (e.g. coccidiostats). Feed contaminants are a recognized issue in food safety, and have to be included among veterinary public health topics in light of the consideration of toxicant exposures as novel zoonoses [7]. Securing safe feeds should be ultimately viewed in the context of the ‘one medicine’ approach of close cooperation between human and veterinary medicine and life sciences disciplines (Figure 1). Since feed contaminants are a persistent, unavoidable problem, risk assessment should support rational strategies for the control and management of feed production chains. The primary strategy is the implementation of good practice in feed production, to ensure quality of ingredients, appropriateness of storage and transport, avoidance of cross-contamination and correct use of authorized additives. Quality assurance of certain components may be important as they may be signiﬁcant sources of undesirable substances, e.g. corn for fusariotoxins [12, 45, 46], ﬁsh oils for POP [76] or mineral additives for Cr(III) [11] and dioxins [24]. In some

CAB International 2010

Alberto Mantovani and Chiara Frazzoli

267

4. Miraglia M, Marvin HJ, Kleter GA, Battilani P, Brera C, Coni E, et al. Climate change and food safety: an emerging issue with special focus on Europe. Food Chemistry and Toxicology 2009;47:1009–21.

instances, it is critical to develop feed sources less prone to contamination, such as vegetable oils for aquaculture feeds [77]: since the ultimate target is the reduction of exposure of vulnerable consumers group, these studies represent a telling example of the link between animal feed research and public health goals. Another aspect is the development of approaches to reduce the presence of contaminants or their bioavailability, as in the case of mycotoxin detoxifying agents. However, there is the need to transfer promising results (see e.g. [55–57]) to real-life ﬁeld situations; also, agents added to feed on purpose have to undergo a thorough safety assessment [2] and, in the case of mycotoxin detoxifying agents, possibly also a risk-to-beneﬁt assessment. Finally, the issue of feed contaminants emphasizes the need for integrating the different components of the environment and health context. Long-standing pasture contamination episodes have represented examples of non-epidemic emergencies, where livestock rearing was the victim of improper industrial practice (e.g. toxic waste disposal); this has led to both the disappearance of viable farming enterprises and potential risks to consumers (see, e.g. [20, 87, 88]). Overall, this short excursus indicates that risk assessment of feed contaminants is an important component of veterinary public health, interfacing with food safety, food security, animal and human nutrition as well as environmental sciences and technological innovation; the many remaining data need to provide stimulating issues for further research.

11. European Food Safety Authority. Safety and efﬁcacy of chromium methionine (Availa1 Cr) as feed additive for all species. EFSA Journal 2009;1043:1–53.

Acknowledgement

12. European Food Safety Authority. Opinion of the Scientiﬁc Panel on Contaminants in the Food Chain related to Zearalenone as undesirable substance in animal feed. EFSA Journal 2004;89:1–35.

This paper has been elaborated within the frame of: the Integrated Project (European 6th Framework Programme) AQUAMAX (http://www.aquamaxip.org), the National Health System project “PESCI” (http:// w3.uniroma1.it/PESCI/), the non-proﬁt organization Noodles, Nutrition and Food Safety and Wholesomeness (http://www.noodlesonlus.org) and the Italian Society for Environment and Health – ISEH (http://www.iseh.it).

References 1. Commission of the European Communities. White paper on food safety. Brussels: EC. 12 January 2000, COM (1999) 719 ﬁnal. 2000. Available from: URL: http://ec.europa.eu/dgs/ health_consumer/library/pub/pub06_en.pdf 2. Mantovani A, Maranghi F, Puriﬁcato I, Macrı` A. Assessment of feed additives and contaminants: an essential component of food safety. Annali dell’Istituto Superiore di Sanita` 2006;42:427–32. 3. La Rocca C, Mantovani A. From environment to food: the case of PCB. Annali dell’Istituto Superiore di Sanita` 2006;42:410–5.

g

5. Kerr LA, McGavin HD. Chronic copper poisoning in sheep grazing pastures fertilized with swine manure. Journal of the American Veterinary Medical Association 1991;198:99–101. 6. European Food Safety Authority. Opinion of the Scientiﬁc Panel on Additives and Products or Substances used in Animal Feed on the development of an approach for the environmental risk assessment of additives, products and substances used in animal feed. EFSA Journal 2007; 529:1–73. 7. Frazzoli C, Mantovani A. Toxicants exposures as novel zoonoses: reﬂections on sustainable development, food safety and veterinary public health. Zoonoses Public Health 2010 [Epub ahead of print]. 8. European Food Safety Authority. EFSA completes 30 risk assessments on undesirable substances in animal feed. 2009a. Available at http://www.efsa.europa.eu/EFSA/ efsa_locale-1178620753812_1211902444045.htm 9. European Food Safety Authority. Opinion of the Scientiﬁc Panel on Contaminants in the Food Chain on a request from the Commission related to hexachlorobenzene as undesirable substance in animal feed. EFSA Journal 2006;402:1–49. 10. European Food Safety Authority. Advice of the scientiﬁc panel on contaminants in the food chain on a request from the commission related to relevant chemical compounds in the group of brominated ﬂame retardants for monitoring in feed and food. EFSA Journal 2006;328:1–4

13. European Food Safety Authority. Opinion of The Scientiﬁc Panel on Contaminants in the Food Chain on Glucosinolates as undesirable substances in animal feed. EFSA Journal 2008;590:1–76. 14. Murphy LE, Gollenberg AL, Louis GM, Kostyniak PJ, Sundaram R. Maternal serum preconception polychlorinated biphenyl concentrations and infant birth weight. Environmental Health Perspectives 2010;118:297–302. 15. Frazzoli C, Petrini C, Mantovani A. Sustainable development and next generation’s health: a long-term perspective about the consequences of today’s activities for food safety. Annali dell’Istituto Superiore di Sanita` 2009;45:65–75. 16. European Food Safety Authority. Opinion of the CONTAM Panel related to the safety assessment of wild and farmed ﬁsh. EFSA Journal 2005;236:1–118. 17. European Food Safety Authority. Scientiﬁc opinion of the Scientiﬁc Committee. Guidance on human health risk beneﬁt assessment of foods. EFSA Journal 2010;1673:1–40. 18. Baldi F, Mantovani A. A new database for food safety: EDID (Endocrine disrupting chemicals – Diet Interaction Database). Annali dell’Istituto Superiore di Sanita` 2008;44:57–63. 19. van Larebeke N, Hens L, Schepens P, Covaci A, Baeyens J, Everaert K, et al. The Belgian PCB and dioxin incident of

CAB International 2010

268

Animal Science Reviews 2010 January-June 1999: exposure data and potential impact on health. Environmental Health Perspectives 2001;109:265–73.

20. Turrio-Baldassarri L, Alivernini S, Carasi S, Casella M, Fuselli S, Iacovella N, et al. PCB, PCDD and PCDF contamination of food of animal origin as the effect of soil pollution and the cause of human exposure in Brescia. Chemosphere 2009;76:278–85. 21. Frazzoli C, Mantovani A, Campanella L, Dragone R. BEST – Integrated Toxicity (Bio)sensors System for Hazard Analysis and Management in the Food Chain and the Environment. International Publication No. WO 2010/001432 A1. World Intellectual Property Organization; 2010. Available from: URL: http://www.wipo.int/pctdb/en/wo.jsp?WO=2010001432 22. Frazzoli C, Orisakwe OE, Dragone R, Mantovani A. Diagnostic health risk assessment of electronic waste on the general population in developing countries’ scenarios. Environmental Impact Assessment Review 2010;30:388–99. 23. European Food Safety Authority. Mercury as undesirable substance in animal feed. EFSA Journal 2008;654:1–74. 24. Parera J, Abalos M, Pe´rez-Vendrell AM, Brufau J, de Juan F, Escribano F, et al. Occurrence and bioaccumulation study of PCDD and PCDF from mineral feed additives. Chemosphere 2008;73(Suppl.):S252–60. 25. European Food Safety Authority. Opinion of the Scientiﬁc Panel on additives and products or substances used in animal feed (FEEDAP) on the re-evaluation of efﬁcacy and safety of the coccidiostat Monteban1 G100 in accordance with article 9 G of Council Directive 70/524/EEC. EFSA Journal 2004;90:1–44. 26. European Food Safety Authority. Cross-contamination of nontarget feedingstuffs by narasin authorised for use as a feed additive [1] – Scientiﬁc Opinion of the Panel on Contaminants in the Food Chain. EFSA Journal 2007;552:1–35. 27. Mc Evoy JDG, Smyth WG, Kennedy DG. Contamination of animal feedingstuffs with nicarbazin: investigations in a feed mill. Food Additives and Contamination 2003;20:136–140.

of barley grain with corn: Effects on performance and lactation in periparturient dairy cows. Journal of Dairy Science 2009;92:5411–8. 36. Mantovani A, Frazzoli C, La Rocca C. Risk assessment of endocrine-active compounds in feeds. Veterinary Journal 2009;182:392–401. 37. Kirpnick-Sobol Z, Reliene R, Schiestl RH. Carcinogenic Cr(VI) and the nutritional supplement Cr(III) induce DNA deletions in yeast and mice. Cancer Research 2006;66:3480–4. 38. Medeiros MG, Rodrigues AS, Batore´u MC, Laires A, Rueff J, Zhitkovich A. Elevated levels of DNA-protein crosslinks and micronuclei in peripheral lymphocytes of tannery workers exposed to trivalent chromium. Mutagenesis 2003;18:19–24. 39. Nguyen A, Mulyani I, Levina A, Lay PA. Reactivity of chromium(III) nutritional supplements in biological media: an X-ray absorption spectroscopic study. Inorganic Chemistry 2008;47:4299–309. 40. Padmavathi IJ, Rao KR, Venu L, Ganeshan M, Kumar KA, Rao ChN, et al. Chronic maternal dietary chromium restriction modulates visceral adiposity: probable underlying mechanisms. Diabetes 2010;59:98–104. 41. Deshmukh NS, Bagchi M, Lau FC, Bagchi D. Safety of an oxygen-coordinated niacin-bound chromium(III) complex (NBC): II. Developmental toxicity study in rats. Journal of Inorganic Biochemistry 2009;103(12):1755–60. 42. Leblanc JC, Gue´rin T, Noe¨l L, Calamassi-Tran G, Volatier JL, Verger P. Dietary exposure estimates of 18 elements from the 1st French Total Diet Study. Food Additives and Contamination 2005;22:624–41. 43. European Commission. Opinion of the Scientiﬁc Committee on food on the tolerable upper intake level of trivalent chromium. 2003. Available from: URL: http://ec.europa.eu/food/fs/sc/scf/ out197_en.pdf

28. Strauch W. Is contamination-free feed production realistic? Feed Technology 2003;7:23–25.

44. Driehuis F, Spanjer MC, Scholten JM, Te Giffel MC. Occurrence of mycotoxins in feedstuffs of dairy cows and estimation of total dietary intakes. Journal of Dairy Science 2008;91:4261–4271.

29. Rokka M, Eerola S, Perttila U, Rossow L, Venalainen E, Valkonen E, et al. The residue levels of narasin in eggs of laying hens fed with unmedicated and medicated feed. Molecular Nutrition and Food Research 2005;49:38–42.

45. European Food Safety Authority. Opinion of the Scientiﬁc Panel on contaminants in the food chain [CONTAM] related to deoxynivalenol (DON) as undesirable substance in animal feed. EFSA Journal 2004;73:1–35.

30. Mortier L, Huet AC, Charlier C, Daeseleire E, Delahaut P, Van Peteghem C. Incidence of residues of nine anticoccidials in eggs. Food Additives and Contamination 2005;22:1120–5.

46. European Food Safety Authority. Opinion of the Scientiﬁc Panel on contaminants in the food chain [CONTAM] related to fumonisins as undesirable substances in animal feed. EFSA Journal 2005;235:1–32.

31. Garbarino JR, Bednar AJ, Rutherford DW, Beyer RS, Wershaw RL. Environmental fate of roxarsone in poultry litter. I. Degradation of roxarsone during composting. Environmental Science and Technology 2003;37:1509–14. 32. Shah SB, Hutchison KJ, Hesterberg DL, Grabow GL, Huffman RL, et al. Leaching of nutrients and trace elements from stockpiled turkey litter into soil. Journal of Environmental Quality 2009;38:1053–65. 33. Yao LX, Li GL, Dang Z, He ZH, Zhou CM, Yang BM. Arsenic uptake by two vegetables grown in two soils amended with As-bearing animal manures. Journal of Hazardous Materials 2009;164:904–10. 34. European Food Safety Authority. Scientiﬁc opinion on arsenic in food. EFSA Journal 2009;1351:1–199. 35. Sadri H, Ghorbani GR, Rahmani HR, Samie AH, Khorvash M, Bruckmaier RM. Chromium supplementation and substitution

g

47. Minervini F, Dell’aquila ME. Zearalenone and reproductive function in farm animals. International Journal of Molecular Sciences 2008;9:2570–84. 48. Zinedine A, Soriano JM, Molto´ JC, Man˜es J. Review on the toxicity, occurrence, metabolism, detoxiﬁcation, regulations and intake of zearalenone: an oestrogenic mycotoxin. Food Chemistry and Toxicology 2007;45:1–18. 49. Goyarts T, Da¨nicke S, Valenta H, Ueberscha¨r KH. Carry-over of Fusarium toxins (deoxynivalenol and zearalenone) from naturally contaminated wheat to pigs. Food Additives and Contaminants 2007;24:369–80. 50. Chen F, Ma Y, Xue C, Ma J, Xie Q, Wang G, et al. The combination of deoxynivalenol and zearalenone at permitted feed concentrations causes serious physiological effects in young pigs. Journal of Veterinary Science 2008;9:39–44.

CAB International 2010

Alberto Mantovani and Chiara Frazzoli 51. Commission of the European Communities. Commission recommendation 2006/576/CE of 17 August 2006 on the presence of deoxynivalenol, zearalenone, ochratoxin A, T-2 and HT-2 and fumonisins in products intended for animal feeding. 2006. 52. Berthiller F, Schuhmacher R, Adam G, Krska R. Formation, determination and signiﬁcance of masked and other conjugated mycotoxins. Analytical and Bioanalytical Chemistry 2009;395:1243–52. 53. Gareis M, Bauer J, Thiem J, Plank G, Grabley S, Gedek B. Cleavage of zearalenone-glycoside, a “masked” mycotoxin, during digestion in swine. Zentralbl Veterinarmed B 1990;37(3):236–40. 54. Dall’Asta C, Mangia M, Berthiller F, Molinelli A, Sulyok M, Schuhmacher R, et al. Difﬁculties in fumonisin determination: the issue of hidden fumonisins. Analytical and Bioanalytical Chemistry 2009;395:1335–45. 55. Avantaggiato G, Solfrizzo M, Visconti A. Recent advances on the use of adsorbent materials for detoxiﬁcation of Fusarium mycotoxins. Food Additives and Contaminants 2005;22:379–88. 56. Avantaggiato G, Havenaar R, Visconti A. Assessment of the multi-mycotoxin-binding efﬁcacy of a carbon/aluminosilicatebased product in an in vitro gastrointestinal model. Journal of Agricultural and Food Chemistry 2007;55:4810–9. 57. Do¨ll S, Gericke S, Da¨nicke S, Raila J, Ueberschar KH, Valenta H, et al. The efﬁcacy of a modiﬁed aluminosilicate as a detoxifying agent in Fusarium toxin contaminated maize containing diets for piglets. Journal of Animal Physiology and Animal Nutrition 2005;89:342–358. 58. Styriak I, Conkova E. Microbial binding and biodegradation of mycotoxins. Veterinary and Human Toxicology 2002;44: 358–61. 59. Varga J, Toth B. Novel strategies to control mycotoxins in feeds: A review. Acta Veterinaria Hungarica 2005;53:189–203. 60. Schatzmayr G, Zehner F, Ta¨ubel M, Schatzmayr D, Klimitsch A, et al. Microbiologicals for deactivating mycotoxins. Molecular Nutrition and Food Research 2006;50:543–551. 61. U.K. Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment. Phytoestrogens and Health. 2003. Available from: URL: http://cot.food.gov.uk/pdfs/ phytoreport0503. 62. Woclawek-Potocka I, Piskula MK, Bah M, Siemieniuch MJ, Korzekwa A, et al. Concentrations of isoﬂavones and their metabolites in the blood of pregnant and non-pregnant heifers fed soy bean. Journal of Reproduction and Development 2008;54:358–63. 63. Jiang ZY, Jiang SQ, Lin YC, Xi PB, Yu DQ, Wu TX. Effects of soybean isoﬂavone on growth performance, meat quality, and antioxidation in male broilers. Poultry Science 2007;86: 1356–62. 64. Tucker HA, Knowlton KF, Meyer MT, Khunjar WO, Love NG. Effect of diet on fecal and urinary estrogenic activity. Journal of Dairy Science 2010;93:2088–94.

269

equol using immature mouse uterotrophic responses and an estrogen receptor transactivation assay. Cancer Epidemiology 2009;33:61–8. 67. Ezendam J, Vos JG, Pieters R. Research articles mechanisms of hexachlorobenzene-induced adverse immune effects in brown norway rats. Journal of Immunotoxicology 2005;1: 167–75. 68. Alvarez L, Randi A, Alvarez P, Piroli G, Chamson-Reig A, Lux-Lantos V, et al. Reproductive effects of hexachlorobenzene in female rats. Journal of Applied Toxicology 2000;20:81–7. 69. Plante I, Charbonneau M, Cyr DG. Decreased gap junctional intercellular communication in hexachlorobenzene-induced gender-speciﬁc hepatic tumor formation in the rat. Carcinogenesis 2002;23:1243–9. 70. Grimalt JO, Carrizo D, Garı´ M, Font-Ribera L, Ribas-Fito N, et al. An evaluation of the sexual differences in the accumulation of organochlorine compounds in children at birth and at the age of 4 years. Environmental Research 2010;110:244–50. 71. Campoy C, Olea-Serrano F, Jime´nez M, Baye´s R, Can˜abate F, Rosales MJ, et al. Diet and organochlorine contaminants in women of reproductive age under 40 years old. Early Human Development 2001;65(Suppl.):S173–82. 72. Attard Barbini A, Vanni F, Pelosi P, Generali T, Amendola G, Stefanelli P, et al. Low levels of organochlorine pesticides in subjects with metabolic disturbances: a survey taken in Roma in 2001–2002. Bulletin of Environmental Contamination and Toxicology 2004;73:319–26. 73. Dhooge W, Den Hond E, Koppen G, Bruckers L, Nelen V, van de Mieroop E, et al. Internal exposure to pollutants and sex hormone levels in Flemish male adolescents in a crosssectional study: associations and dose-response relationships. Journal of Exposure Analysis and Environmental Epidemiology 2009 [Epub ahead of print]. 74. Glynn A, Aune M, Nilsson I, Darnerud PO, Ankarberg EH, Bignert A, et al. Declining levels of PCB, HCB and p,p0 -DDE in adipose tissue from food producing bovines and swine in Sweden 1991–2004. Chemosphere 2009;74:1457–62. 75. World Health Organization-International Programme on Chemical Safety (WHO-IPCS). Hexachlorobenzene, Environmental Health Criteria 195. World Health Organization, Geneva, Switzerland; 1997. 76. Oterhals G, Kvamme B, Berntssen MH. Modeling of a shortpath distillation process to remove persistent organic pollutants in ﬁsh oil based on process parameters and quantitative structure properties relationships. Chemosphere 2010 [Epub ahead of print 3 May 2010]. 77. Sprague M, Bendiksen EA, Dick JR, Strachan F, Pratoomyot J, Berntssen MH, et al. Effects of decontaminated ﬁsh oil or a ﬁsh and vegetable oil blend on persistent organic pollutant and fatty acid compositions in diet and ﬂesh of Atlantic salmon (Salmo salar). British Journal of Nutrition 2010;1:1–10.

65. Mustonen EA, Tuori M, Saastamoinen I, Taponen J, Wa¨ha¨la¨ K, Saloniemi H, et al. Equol in milk of dairy cows is derived from forage legumes such as red clover. British Journal of Nutrients 2009;102:1552–6.

78. Isosaari P, Lundebye AK, Ritchie G, Lie O, Kiviranta H, Vartiainen T. Dietary accumulation efﬁciencies and biotransformation of polybrominated diphenyl ethers in farmed Atlantic salmon (Salmo salar). Food Additives and Contaminants 2005;22:829–837.

66. Nielsen TS, Nørgaard JV, Purup S, Frette´ XC, BonefeldJørgensen EC. Estrogenic activity of bovine milk high or low in

79. Kierkegaard A, Asplund L, de Wit CA, McLachlan MS, Thomas GO, Sweetman AJ, et al. Fate of higher brominated

g

CAB International 2010

270

Animal Science Reviews 2010 PBDEs in lactating cows. Environmental Science and Technology 2007;41:417–23.

80. Ellis-Hutchings RG, Cherr GN, Hanna LA, Keen CL. Polybrominated diphenyl ether (PBDE)-induced alterations in vitamin A and thyroid hormone concentrations in the rat during lactation and early postnatal development. Toxicology and Applied Pharmacology 2006;215:135–45. 81. Saegusa Y, Fujimoto H, Woo GH, Inoue K, Takahashi M, Mitsumori K, et al. Developmental toxicity of brominated ﬂame retardants, tetrabromobisphenol A and 1,2,5,6,9,10hexabromocyclododecane, in rat offspring after maternal exposure from mid-gestation through lactation. Reproductive Toxicology 2009;28:456–67. 82. van der Ven LT, van de Kuil T, Leonards PE, Slob W, Lilienthal H, Litens S, et al. Endocrine effects of hexabromocyclododecane (HBCD) in a one-generation reproduction study in Wistar rats. Toxicology Letters 2009;185:51–62. 83. Hauka˚s M, Mariussen E, Ruus A, Tollefsen KE. Accumulation and disposition of hexabromocyclododecane (HBCD) in juvenile rainbow trout (Oncorhynchus mykiss). Aquatic Toxicology 2009;95:144–51. 84. Daniels JL, Pan IJ, Jones R, Anderson S, Patterson DG Jr, Needham LL, et al. Individual characteristics associated with PBDE levels in U.S. human milk samples. Environmental Health Perspectives 2010;118:155–60. 85. Briggs DJ. A framework for integrated environmental health impact assessment of systemic risks. Environmental Health 2008;7:61–78. 86. Gossner CM, Schlundt J, Ben Embarek P, Hird S, Lo-Fo-Wong D, Beltran JJ, et al. The melamine incident: implications for international food and feed safety. Environmental Health Perspectives 2009;117:1803–8. 87. Turrio-Baldassarri L, Abate V, Alivernini S, Battistelli CL, Carasi S, Casella M, et al. A study on PCB, PCDD/PCDF industrial contamination in a mixed urban-agricultural area signiﬁcantly affecting the food chain and the human exposure. Part I: soil and feed. Chemosphere 2007;67:1822–30.

g

88. Esposito M, Cavallo S, Serpe FP, D’Ambrosio R, Gallo P, Colarusso G, et al. Levels and congener proﬁles of polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans and dioxin-like polychlorinated biphenyls in cow’s milk collected in Campania, Italy. Chemosphere 2009;77:1212–6. 89. Ma HW, Lai YL, Chan CC. Transfer of dioxin risk between nine major municipal waste incinerators in Taiwan. Environment International 2002;28:103–10. 90. Beldome´nico HR, Baroni E, Campagnoli DU, Sigrist ME, Rubio M, Boggio JC. Cadmium accumulation and distribution in slaughtered horse kidneys from the Argentine central region. Archives of Environmental Contamination and Toxicology 2001;41:100–3. 91. Loganathan P, Hedley MJ, Grace ND. Pasture soils contaminated with fertilizer-derived cadmium and ﬂuorine: livestock effects. Reviews in Environmental Contamination and Toxicology 2008;192:29–66. 92. Rhind SM, Kyle CE, Telfer G, Duff EI, Smith A. Alkyl phenols and diethylhexyl phthalate in tissues of sheep grazing pastures fertilized with sewage sludge or inorganic fertilizer. Environmental Health Perspectives 2005;113:447–53. 93. Rhind SM, Kyle CE, Mackie C, Telfer G. Effects of exposure of ewes to sewage sludge-treated pasture on phthalate and alkyl phenol concentrations in their milk. Science of the Total Environment 2007;383:70–80. 94. Turyk ME, Anderson HA, Freels S, Chatterton R Jr, Needham LL, Patterson DG Jr, et al. Great Lakes Consortium. Associations of organochlorines with endogenous hormones in male Great Lakes ﬁsh consumers and nonconsumers. Environmental Research 2006;102:299–307. 95. Domingo JL, Bocio A, Falco G, Llobet JM. Beneﬁts and risks of ﬁsh consumption Part I. A quantitative analysis of the intake of omega-3 fatty acids and chemical contaminants. Toxicology 2007;230:219–26. 96. Bernard A, Broeckaert F, De Poorter G, De Cock A, Hermans C, Saegerman C, et al. The Belgian PCB/dioxin incident: analysis of the food chain contamination and health risk evaluation. Environmental Research 2002;88:1–18.

CAB International 2010

Index ABCG2 64 accelerated lambing 93, 100–107, 119, 124–132 artiﬁcial rearing 99–100 breed 98–100, 101–102, 123–124, 124–132 Finnish Landrace (Finnsheep) 98–100, 125–126, 127, 130 Ile de France 100–105 Merino 100 Romanov cross-breeds 124–125 St Croix White (STX) 123–124 disease 96–97 mastitis 97 effective teaser goats (ETGs) 106–107 effective teaser rams (ETRs) 103–104, 106–107 FLASH system 106, 107, 124, 131 housing 93–107 beneﬁts 95–97 cost 97–98 exercising 97 lambing 96 nutrition 98 shepherding 97 involuting uterus 116–117 lactation 116–118 lamb age 117–118 luteinizing hormone (LH) 116–119 management 95–98 melatonin 129–130 ‘nanny’ goats 98 nutrition 98, 118 pests 96 photoperiod 115–116, 120 production 102–105 proliﬁcacy 98–100, 102–107 ram 118–119, 122 Rowett project 120–121 seasonal breeding 115–116 South Africa 100–109 STAR system 106, 121–124, 131 stock selection 97 systems FLASH 106, 107, 124, 131 Rowett project 120–121 STAR 106, 121–124, 131 weaning 117–118, 122 animal studies 167–169 animal-derived antimicrobials 211–212, 221–222 activity 215, 221 applications 221 chitosan 214 defensins 212

free fatty acids (FFA) 2114–215 lactoferrin 214 lactoperoxidase 214 lysozyme 214 mechanism 215, 220–221 pleurocidin 212 protegrim 212–214 antigenic drift 76 antimicrobials 211–212, 221–222 activity 215, 218, 221 animal-derived 212 chitosan 214 defensins 212 free fatty acids (FFA) 2114–215 lactoferrin 214 lactoperoxidase 214 lysozyme 214 pleurocidin 212 protegrim 212–214 applications 221 mechanism 215, 218, 220–221 microbial-derived 218–219 bacteriocins 218–220 hydrogen peroxide 220 natamycin 220 nisin 220 pediocin 220 reuterin 220 sakacins 219–220 plant-derived 212 aldehydes 217 Allium species 217 cinnamon 218 essential oils 215–217 methyl jasmonate (MeJA) 217–218 oleuropein 217 plant extracts 215–217 vanillin 217 avian inﬂuenza 73–74, 139–141, 144–145 chicken production 141–142, 143 culling 140, 143 ducks 75, 142–143 FAO classiﬁcation system 139, 141 government policy 144 vaccines 74, 88–89, 140, 143–144 beneﬁts 85–87 campaigns 77–81, 89 China 81 cost 76, 84–88, 89 duration of immunity 75–76 efﬁcacy 74–75, 77 Egypt 82

272

Index

emergency 77 Hong Kong 82 lag period 75–76 Indonesia 82–83, 84–85 Pakistan 83 preventive 76–77 production systems 78–80 routine 77 strategies 76–77 Vietnam 83–88 B. Fecundity gene 64–65 Bacillus cereus 186 Bayesian estimation 52–55, 198–199, 205 beef See Cattle bioinformatics 10 bovine genomics 23–29 Bovine Leukocyte Adhesion Deﬁciency (BLAD) 66 Bovine Spongiform Encephalopathy (BSE) 66 breeding 50, 59–60 cattle 49–56, 202, 203–204 congenital disorders 65–67 crossbred performance 202–203 daughter yield deviation (DYD) 51–54 estimated breeding values (EBV) 50–53, 197, 199, 200–201, 204 genome-wide association studies (GWAS) 51–52 genomic estimated breeding values (GEBV) 51–56, 199, 201, 202, 203–207 inbreeding 203 marker-assisted selection (MAS) 27, 49–56, 67–68, 197–207 microgoats 162 production traits 62–64 quantitative trait mutations (QTMs) 59–68 reproduction traits 64–65 Brucella melitensis See brucellosis brucellosis 171–172 antigenic composition 173, 178 diagnostic tools 171, 176 enzyme-linked immunosorbent assays (ELISA) 177–178 ﬂuorescence polarization assay (FPA) 178 polymerase chain reaction (PCR) 176 serological 177–178 strain identiﬁcation 178 distribution 173–174 eradication 180–181 genome 173 morbidity 175 prevention 179 transmission 174–175 symptoms 175 vaccination 171–172, 179–180 buffalo 1–6 Campylobacter species 186, 188, 190, 191 candidate gene approach 60–61

Canine Leukocyte Adhesion Deﬁciency (CLAD) 66 cattle 23–29, 39–40 ABCG2 64 Bovine Leukocyte Adhesion Deﬁciency (BLAD) 66 Bovine Spongiform Encephalopathy (BSE) 66 breeding 49–56, 202, 203–204 brucellosis 174–175 CMD 66–67 Crooked Tail Syndrome (CTS) 67 DGAT1 64 embryo cryopreservation 249 feed 234, 238–239 Foetal Ichytiosis 67 MSTN 63–64 chicken backyard multi-purpose ﬂocks 78, 80, 142 breeder ﬂocks 78, 79 broiler ﬂocks 78, 79–80, 141–142 feed 237 layer ﬂocks 78, 79, 141–142 China 81 Chromium (III) 260–261 climate change India 158 microgoats 160 Clostridium species 186–187, 191 CLPG 63 CMD 66–67 coat colour 17–20 congenital disorders 65 CMD 66–67 Crooked Tail Syndrome (CTS) 67 CVM 67 Factor XI (FXI) 67 Foetal Ichytiosis 67 Leukocyte Adhesion Deﬁciency (LAD) 66 Transmissible Spongiform Encephalopathy (TSE) 65–66 consumer behaviour 41–42 cow See cattle Crooked Tail Syndrome (CTS) 67 crossbred performance 202–203 crude protein (CP) 1–4 cryopreservation 245–246, 250 cattle 249 cryoprotective additive (CPA) 246–248 direct transfer 247–248 equilibrium cooling method 246–248 freezing 246–247, 248 goat 249 horse 250 in vitro-produced (IVP) 248–250 in vivo-derived (IVD) 248–250 non-equilibrium cooling method 246, 248–249 obstacles 246 pig 249 pregnancy rate 248

Index

sheep 249–250 thawing 247 vitriﬁcation 246, 248–249 Cryptosporidium parvum 187, 188, 191 CVM 67 Cyclospora cayetanensis 187 daughter yield deviation (DYD) 51–54 diacylglycerol O-acyltransferase (DGAT1) gene 26, diagnostic tools 171, 176 enzyme-linked immunosorbent assays (ELISA) 177–178 ﬂuorescence polarization assay (FPA) 178 polymerase chain reaction (PCR) 176 serological 177–178 strain identiﬁcation 178 disease avian inﬂuenza 73–89, 139–145 brucellosis 171–181 human enteric pathogens 185–192 parasites 161 sheep 96–97 dogs Canine Leukocyte Adhesion Deﬁciency (CLAD) 66 food 235 dry matter (DM) 1, 6 ducks 142–143 backyard multi-purpose ﬂocks 78, 80 breeder ﬂocks 78 broiler ﬂocks 78, 80 HPAI 75, 142–143 layer ﬂocks 78, 79 eat quality 35–37, 38–43 ecological stability 161, 162 Egypt 82 embryo cryopreservation 245–246, 250 cattle 249 cryoprotective additive (CPA) 246–248 direct transfer 247–248 equilibrium cooling method 246–248 freezing 246–247, 248 goat 249 horse 250 in vitro-produced (IVP) 248–250 in vivo-derived (IVD) 248–250 non-equilibrium cooling method 246, 248–249 obstacles 246 pig 249 pregnancy rate 248 sheep 249–250 thawing 247 vitriﬁcation 246, 248–249 endocrine disruption 262–263 enteric pathogens See human enteric pathogens enzyme-linked immunosorbent assays (ELISA) 177–178 equine embryo cryopreservation 250 Escherichia coli 186, 189–191, 212, 214

273

estimated breeding values (EBV) 50–53, 197, 199, 200–201, 204 See also genomic estimated breeding values (GEBV) European food safety authority (EFSA) 257–259 Factor XI (FXI) 67 FAO classiﬁcation system 139, 141 fat gustatory system 149–153 long-chain fatty acids (LCFA) 148–153 mobilization 3 polyunsaturated fatty acids (PUFA) 153 taste perception 147–153 taste receptor cells (TRC) 149, 151, 152–153 feed accelerated lambing 98 assessment of 231 addition of food 236 conditioning 234–236 daily intake 231 diet selection pathway 234–234 electrophysiological studies 233 proportional intakes 234 sham feeding 236 short-term intake 231–232 time to ﬁrst meal 232 two bottle test 233 two choice test 232–234 dry matter (DM) 1, 6 meat quality 38–39, 40–41 palatability 229–231, 240–241 patterns of acceptance 238 learning time 238–239 memory retention time 239 prediction of intake 239–240 toxic contamination 257–258, 266–267 brominated ﬂame retardants (BFR) 264 Chromium (III) 260–261 cocciodiostats 259–260 contamination 260, 263 diagnostic assessment 264–266 endocrine disruption 262–263 European food safety authority (EFSA) 257–259, 260, 261, 264 hexachlorobenzene (HCB) 263–264, 265 mycotoxin 261–262 narasin 259–260 persistent organic pollutants (POP) 258–259, 263–264, 265 polychlorinated biphenyls (PCB) 257–259, 265 risk assessment 258–259, 261, 263–266 toxicological effects 259–260, 261, 262, 263 zearalenone 261–262 Finnish Landrace (Finnsheep) 98–100 FLASH system 106, 107, 124, 131 ﬂuorescence polarization assay (FPA) 178

274

Index

Foetal Ichytiosis 67 food additives See antimicrobials animal See feed safety 257–259 gene expression 9–14, 27–29 genetic markers 49–50 quantitative trait loci (QTL) 9–14, 24, 49––53, 55– 56, 59–68, 198, 199–200, 202, 204, 205–206 heritability 11–12, 49 simple sequence repeats (SSRs) 49–50 single nucleotide polymorphisms (SNPs) 11, 50–56, 198–199, 202, 204, 205, 206 intramuscular fat (IMF) 24–29 genetic modiﬁcation (GM) 165–169 animal studies 167–169 insect resistance 165–166 risk assessment 166–167 genetical genomics 9–14 genome-wide association studies (GWAS) 51–52 genomic estimated breeding values (GEBV) 51–56, 199, 201, 202, 203–207 genomic selection 67–68, 197–207 goat effective teaser goats (ETGs) 106–107 embryo cryopreservation 249 microgoats 157–163 breeding 162 breeds 159–160 ecological stability 161, 162 economics of 161–162 productivity 160–161 ‘nanny’ 98 See also brucellosis growth hormone (GH) 2, 4 gustatory system 149–153 CD36 149–152, 153 H5N1 73–74, 139–141, 144–145 chicken production 141–142, 143 culling 140, 143 ducks 75, 142–143 FAO classiﬁcation system 139, 141 government policy 144 vaccines 74, 88–89, 140, 143–144 beneﬁts 85–87 campaigns 77–81, 89 China 81 cost 76, 84–88, 89 duration of immunity 75–76 efﬁcacy 74–75, 77 Egypt 82 emergency 77 Hong Kong 82 lag period 75–76 Indonesia 82–83, 84–85 Pakistan 83 preventive 76–77

production systems 78–80 routine 77 strategies 76–77 Vietnam 83–88 hexachlorobenzene (HCB) 263–264, 265 highly pathogenic avian inﬂuenza (HPAI) See H5N1 historical genetics 17–20 Hong Kong 82 human enteric pathogens 185–186, 191–192 Bacillus cereus 186 Campylobacter species 186, 188, 190, 191 Clostridium species 186–187, 191 contamination source 189–191 Cryptosporidium parvum 187, 188, 191 Cyclospora cayetanensis 187 Escherichia coli 186, 189–191, 212, 214 Listeria monocytogenes 186, 214, 218, 220 NoV 187–188 plant-microbe interactions 187–189 colonization 188–189 Salmonella enterica 186, 188, 191 Shigella species 186 transmission 189 animal 189–190 post harvest 191 soil 190 vector 190–191 water 190 See also antimicrobials IGF2 62–63 inbreeding 203 India 157–163 Indonesia 82–83, 84–85 inﬂuenza See highly pathogenic avian inﬂuenza (HPAI) insect resistance 165–166, 167, 169 Ile de France 100–105 intramuscular fat (IMF) 23–29, 36–37, 38–42 lactation period 1–2 lambing See accelerated lambing lectins 165–169 leptin 24–25 linkage disequilibrium (LD) 26–27, 50–52, 60, 61, 68, 199–200, 202, 204 lipid gustatory system 149–153 long-chain fatty acids (LCFA) 148–153 mobilization 3 polyunsaturated fatty acids (PUFA) 153 taste perception 147–153 taste receptor cells (TRC) 149, 151, 152–153 Listeria monocytogenes 186, 214, 218, 220 livestock buffalo 1–6 cattle 23–29, 39–40 ABCG2 64 Bovine Leukocyte Adhesion Deﬁciency (BLAD) 66

Index

Bovine Spongiform Encephalopathy (BSE) 66 breeding 49–56, 202, 203–204 brucellosis 174–175 CMD 66–67 Crooked Tail Syndrome (CTS) 67 DGAT1 64 embryo cryopreservation 249 feed 234, 238–239 Foetal Ichytiosis 67 MSTN 63–64 coat colour 17–20 genetic improvement 9–14, 49–56, 59–68 goat effective teaser goats (ETGs) 106–107 embryo cryopreservation 249 microgoats 157–163 ‘nanny’ 98 housing 93–109 microlivestock 158–159 pig 39 embryo cryopreservation 249 feed 230, 232–233, 235, 237–238, 239 IGF2 62–63 pale, soft and exudative (PSE) 39, 41, 62 PRKAG3 62 RYR1 62 poultry 139 FAO classiﬁcation system 139, 141 production systems 77–80, 141–143 sheep 19 B. Fecundity gene 64–65 CLPG 63 embryo cryopreservation 249–250 feed 236 lambing See accelerated lambing MSTN 63–64 scrapie 65–66 long-chain fatty acids (LCFA) 148–153 luteinising hormone (LH) 116–119 mammals 17–20 marker-assisted selection (MAS) 27, 49–56, 67–68, 197–207 marketing 41–42 meat beef 23–24, 27–29, 35, 38–43 goat 160–161 intramuscular fat (IMF) 23–29, 36–37, 38–42 marketing 41–42 organic 35–36 pork 35, 38–43 pale, soft and exudative (PSE) 39, 41 quality 35–37, 38–43 melatonin 129–130 Mendelian inheritance 17–19 Merino 100 microarrays 10–14, 26–28, 50

microbial-derived antimicrobials 218–219 activity 221 applications 221 bacteriocins 218–220 hydrogen peroxide 220 mechanism 220–221 natamycin 220 nisin 220 pediocin 220 reuterin 220 sakacins 219–220 microgoats 157–163 breeding 162 breeds 159–160 ecological stability 161, 162 economics of 161–162 productivity 160–161 milk 1–2 ABCG2 64 DGAT1 64 urea (MU) 2–4 mouse genetics 17–18 MSTN 63–64 MSTN 63–64 mycotoxin 261–262 narasin 259–260 nerves 151 nitrogen 2–6 nucleus of the solitary tract (NST) 149, 151–152 obesity 147 organic production 35–43 output-orientation 42 Pakistan 83 palatability 229–231, 240–241 affecting factors 236 animal age 238 colour 237 cue 236–237 diurnal variation 236 environmental temperature 237 feeder position 238 ﬂavour 237 physiological condition 238 social factors 237–238 assessment of 231 addition of food 236 conditioning 234–236 daily intake 231 diet selection pathway 234–234 electrophysiological studies 233 proportional intakes 234 sham feeding 236 short-term intake 231–232 time to ﬁrst meal 232 two bottle test 233 two choice test 232–234

275

276

Index

conditioned stimulus (CS) 235 cognitive palatability assessment protocol (CPAP) 235 patterns of acceptance 238 learning time 238–239 memory retention time 239 prediction of intake 239–240 pale, soft and exudative (PSE) 39, 41 RYR1 62 parasites 161 persistent organic pollutants (POP) 258–259, 263–264, 265 photoperiod 115–116 pig 39 embryo cryopreservation 249 feed 230, 232–233, 235, 237–238, 239 IGF2 62–63 pale, soft and exudative (PSE) 39, 41 RYR1 62 PRKAG3 62 plant-derived antimicrobials 212 activity 218, 221 aldehydes 217 Allium species 217 applications 221 cinnamon 218 essential oils 215–217 mechanism 215, 218, 220–221 methyl jasmonate (MeJA) 217–218 oleuropein 217 plant extracts 215–217 vanillin 217 polychlorinated biphenyls (PCB) 257–259, 265 polymerase chain reaction (PCR) 176 polyunsaturated fatty acids (PUFA) 153 pork 35, 38–43 pale, soft and exudative (PSE) 39, 41, 62 poultry 139 FAO classiﬁcation system 139, 141 production systems 77–78, 141–143 backyard multi-purpose ﬂocks 78, 80, 142 breeder ﬂocks 78, 79 broiler ﬂocks 78, 79–80 layer ﬂocks 78, 79 See also highly pathogenic avian inﬂuenza (HPAI) 73–74 PRKAG3 62 production systems 35–43 breed 39–40 lambing See accelerated lambing India 157–158 poultry 77–78, 141–143 backyard multi-purpose ﬂocks 78, 80, 142 breeder ﬂocks 78, 79 broiler ﬂocks 78, 79–80 layer ﬂocks 78, 79

production traits ABCG2 64 CLPG 63 DGAT1 64 IGF2 62–63 MSTN 63–64 PRKAG3 62 RYR1 62 proliﬁcacy 98–100, 102–107 protein 1–6 protein/energy (P/E) ratio 2–3 purine derivatives (PD) 5–6 quantitative trait loci (QTL) 9–14, 24, 49––53, 55–56, 59–68, 198, 199–200, 202, 204, 205–206 heritability 11–12, 49 quantitative trait mutations (QTMs) 59–68 reproduction B. Fecundity gene 64–65 diet 3–4 lambing See accelerated lambing seasonal 115–116 See also embryo cryopreservation 249 risk assessment 166–167 RNA expression 9–14 Rowett project 120–121 RYR1 62 Salmonella enterica 186, 188, 191 scrapie 65–66 sheep 19 B. Fecundity gene 64–65 CLPG 63 embryo cryopreservation 249–250 feed 236 lambing See accelerated lambing MSTN 63–64 scrapie 65–66 See also brucellosis Shigella species 186 signal transduction 151–153 simple sequence repeats (SSRs) 49–50 single nucleotide polymorphisms (SNPs) 11, 50–56, 198–199, 202, 204, 205, 206 intramuscular fat (IMF) 24–29 South Africa 100–109 STAR system 106, 121–124, 131 sterol regulation 26 sustainability 161–162 system approach 42–43 taste buds 148–149, 153 gustatory system 149–153 long-chain fatty acids (LCFA) 148–153 nucleus of the solitary tract (NST) 149, 151–152 perception 147–153 receptor cells (TRC) 149, 151, 152–153 thyroglobulin (TG) gene 25–26 toxic contamination 257–267

Index

Transmissible Spongiform Encephalopathy (TSE) 65–66 urea metabolism 2–3 vaccination beneﬁts 85–87 brucellosis 171–172, 179–180 campaigns 77–81, 89 China 81 Egypt 82 Hong Kong 82 Indonesia 82–83, 84–85 Pakistan 83 Vietnam 83–88 cost 76, 84–88, 89 duration of immunity 75–76

efﬁcacy 74–75, 77 emergency 77 H5N1 73–89 lag period 75–76 preventive 76–77 production systems 78–80 routine 77 strategies 76–77 Vietnam 83–88 whole-genome association (WGA) analysis (WGAA) 61–62, 197–207 studies 26–27 zearalenone 261–262 zoonotic disease See brucellosis

277