Disease Markers Editor-in-Chief Sudhir Srivastava, Ph.D, MPH Office address: Chief Cancer Biomarkers Research Group Division of Cancer Prevention National Cancer Institute E-mail:
[email protected] Correspondence address: Disease Markers C/O IOS Press, Inc. 5795-G Burke Centre Parkway Burke, VA 22015, USA Tel.: +1 703 323 5554 Fax: +1 703 323 3668 E-mail:
[email protected] Disease Markers C/OIOS Press Nieuwe Hemweg 6B 1013 BG Amsterdam, The Netherlands Tel.: +31 20 688 33 55 Fax: +31 20 620 24 19 E-mail:
[email protected] Editors D.S. Alberts Tuscon, AZ, USA P.G. Anker Geneva, Switzerland S.G. Baker Bethesda, MD, USA B. Bapat Toronto, Canada
D.K. Bol Princeton, NJ, USA
B. Levin Houston, TX, USA
P. Boffetta Lyon, France
Bethesda, MD, USA
H.B. Burke Valahalla, NY, USA
D. Lo Hong Kong
W BUSSE Madison, Wl, USA
P. Lynch Houston, TX, USA
R.G.H. Cotton Fitzroy, VIC, Australia G.J. Downing Bethesda, MO, USA Z. Feng Seattle, WA, USA J. Greenman Hull, UK S.M. Hanash Ann Arbor, Ml, USA G. Haroske Dresden, Germany M. Harrington Pasadena, CA, USA N.K. Hayward Herston, QLD, Australia D.E. Henson Rockville, MD, USA
LA. Liotta
H. Magdelenat Paris, France H. Malm Chicago, IL, USA
L. Mao Houston, TX, USA F. Olopade Chicago, IL, USA R.B. Parekh Abingdon, UK G. Rennert Haifa, Israel R. Schleimer Baltimore, MD, USA J.W. Shay Dallas, TX, USA I. Shoulson Rochester, NY, USA M. Steel St. Andrews, UK
P.E. Barker Gaithersburg, MD, USA
W.N. Hittelman Houston, TX, USA
R.L. Strausberg Bethesda, MD, USA
M.J. Birrer Rockville, MD, USA
S. Kaneko Kanazawa, Japan
E. Tahara La Jolla, CA, USA
T.M. Block Doylestown, PA, USA
M. von Knebel Doeberitz Heidelberg, Gemany
H.Vainio Lyon, France
Aims and Scope The journal publishes original research findings (and reviews solicited by the Editor) on the subject of the identification of markers associated with the disease processes whether or not they are an integral part of the pathological lesion. The disease markers may be a genetic host factor predisposing to the disease or the occurrence of cell-surface markers, enzymes or other components, either in altered forms, abnormal concentrations or with abnormal tissue distribution. This journal is designed to provide a forum for publications dealing with original observations in this developing field on any aspect of the general topic including: • •
Identification of new genetic or non-genetic markers (e.g. cell-surface antigens, serum proteins, intra- and extra-cellular enzymes, cytogenic markers and DNA-sequences) Population studies of new and existing markers, designed to elucidate information on their normal distribution as well as that in disease states. Amplification of knowledge about existing markers. Family studies of markers in disease. New techniques for identification and/or isolation of important marker molecules. Use of monoclonal antibodies for the definition of molecular structures associated with disease markers. Identification of disease-associated abnormalities in DNA using recombinant DNA techniques, gene-cloning and DNA restriction enzyme fragment polymorphisms. Identification of markers identifying malignantly transformed neoplastic cells, including precancerous lesions.
Published quarterly
0278-0240/01/$8.00
Printed in The Netherlands
203
Foreword
Sudhir Srivastavaa and Sam Hanashb a
Cancer Biomarkers Research Group, Division of Cancer Prevention, National Cancer Institute, Bethesda, MD 20892, USA b University of Michigan Medical Center, Ann Arbor, MI 48109, USA The field of proteomics holds great promise for identifying nontargeted, global molecular profiles "signatures" of normal and diseased cells. While genomics offers a wide array of tools for identifying mutated or dysregulated genes, proteomics offers the ability to measure post-translational modifications, protein stability, phosphorylation state, protein-protein interactions, and protein DNA-binding affinities. These biological events play major roles in the pathogenesis of disease and cannot be studied by DNA and mRNA efforts alone. Recent studies have shown that there is no specific correlation between mRNA abundance and protein expression levels in a cell at a given time. The discrepancy could arise from the control of mRNA translation, and the stability of mRNAs and proteins. Because almost all therapeutic intervention strategies or early detection technologies target expressed proteins, proteomic-based studies can provide fundamental information to characterize disease progression at the molecular level. Therefore, proteomics has particular relevance to diagnostics and to the identification of disease markers, as proteins can be assayed in serum and other biological fluids. Since the completion of human genome sequence, the discipline of proteomics has taken a center stage in defining the future of molecular diagnostics. This is due, to some extent, to the fact that the sequencing of the human genome and other important genomes has opened the door for proteomics by providing a sequence-based framework for mining the proteome. The growth in proteomic tools and proteomics is in an extraordinary growth phase. As a consequence, proDisease Markers 17 (2001) 203-204 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
teomics is witnessing an extraordinary growth in the number of new investigators and biotechnology companies that are taking an active interest in the field. Proteomics is evolving at a fast pace, as is evident from this special issue. Strategies for protein fractionation prior to analysis are increasingly being relied upon. Selective enrichment for a subset of proteins of interest can be achieved using a variety of techniques from centrifugation procedures to affinity capture. In addition to reducing sample complexity and increasing sensitivity, strategies based on the separate analysis of subcellular compartments provide the means to determine protein location in a cell. There is clearly interest in developing robust "industrial strength" proteomic platforms to achieve high-throughput and high sensitivity. During the lean years of proteomics, the field was largely dominated by a single approach, namely twodimensional (2-D) gels, backed up by an assortment of related tools, ranging from software for image analysis to mass spectrometry. Several technologies for protein profiling are emerging that are not 2-D gel based, including direct profiling by mass spectrometry and the use of protein microarrays. High throughput technologies, such as Surface-Enhanced Laser Desorption/Ionization Time-of-Flight (SELDI-TOF) and Matrix-Assisted Laser Desorption/Ionization Time-ofFlight (MALDI-TOF) technologies, are providing us with a non-biased, global discovery approach using patient serum, plasma, urine or other sources of secretions, to identify - in a single run - the expression patterns in thousands of small molecular mass proteins (< 20 kDa) based on their molecular mass and charge, offering a tool to determine which proteins are secreted from tumor cells or are measurable in the bodily fluids of patients. However, it is likely that no one technology platform will meet the needs of the wide spectrum of applications encompassed by proteomics. Diagnosing disease through proteomics may well require all the tools at our disposal to identify and validate the most promising markers.
S. Srivastava and S. Hanash / Global Strategies for Disease Detection and Treatment: Proteomics
While the emergence of proteomics brings hope to disease detection and therapeutic intervention, it poses several analytical challenges. Analyzing voluminous data generated by high-throughput proteomics arrays is a fairly new endeavor for statisticians for which there is not extensive literature. Another complexity to these analyses is how to analyze proteomics data in reference to disease outcomes. The maturation of proteomics technology, with its global, non-directed ability to analyze serum proteins, will add to the markers or patterns of markers that are able to predict the presence of ovarian cancer. For example, if the protein products
of the amplified DNA or RNA are secreted from tumor cells and migrate to serum, plasma, urine, or other accessible fluids, proteomics will offer a rapid advance in the identification of novel biomarkers to facilitate the development of noninvasive, sensitive assays for earlier cancer detection. This special issue of Disease Markers covers all aspects of proteomics with special emphasis on clinical applications. Some of the proposed approaches may not be ready for a clinical use, however, they are justified and desirable in the spirit of the developing field of proteome research.
205
Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus Allan E. Karlsen, Thomas Sparre, Karin Nielsen, J0rn Nerup and Flemming Pociot* Steno Diabetes Center, Gentofte, Denmark Type 1 (insulin-dependent) diabetes mellitus (T1DM) is associated with a specific destruction of the insulin-producing beta-cells in the islets of Langerhans. Several factors, e.g. genetic, environmental and immunologial, may be involved in the etiology and pathogenesis of T1DM. Autoreactive Tand B-lymphocytes, together with macrophages infiltrate the islets during the pathogenesis, releasing a mixture of cytokines, demonstrated to be specifically toxic to the beta-cells within the islets. Our goal is to understand the molecular mechanisms responsible for the beta-cell specific toxicity enabling us to design novel intervention strategies in T1DM. The proteome approach allows us to get a detailed picture of the beta-cell proteins, which change expression level or are post-translationally modified in different in vitro and in vivo models of TIDM-associated beta-cell destruction. Combining the information obtained from this extended proteome approach, with that of genetic-, transcriptome- and candidategene approaches, we believe that it is possible to reach this goal.
1. Introduction Diabetes mellitus represents a heterogeneous group of disorders. Some distinct diabetic phenotypes can be characterized in terms of specific etiology and/or pathogenesis, but in many cases overlapping phenotypic characteristics make etiological and pathogenetical classification difficult. Type 1 (insulin-dependent) diabetes mellitus (TlDM) is characterized by absolute insulin deficiency, abrupt onset of symptoms, proneness to ketosis and depen* Address for correspondence: Flemming Pociot, Steno Diabetes Center, Niels Steensensvej 2, DK-2820 Gentofte, Denmark. Fax: +45 4443 7313; E-mail:
[email protected]. Disease Markers 17 (2001) 205-216 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
dency on exogenous insulin to sustain life. It is the most common form of diabetes among children and young adults in populations of Caucasoid origin, where the prevalence is approximately 0.4%. T1DM is traditionally regarded as a disease of the childhood and adolescent period. The age-specific incidence pattern is similar worldwide within this age segment, although the worldwide incidence differs markedly for the age group up to 15 years. The disease is rare before 6 months of age, but the incidence increases up to puberty, and thereafter declines. The overall age-adjusted incidence of T1DM varies from 0.1/100.000 per year in China and Venezuela to 36/100.000 per year in Sardinia and Finland [1]. An increasing incidence of T1DM for individuals < 15 years of age has been observed in many countries. Studies have suggested that more than 50% of all T1DM cases develop after the age of 21 years, and that the cumulative incidence (up to age 80 years) may reach approximately 1% [2,3]. T1DM is the result of pancreatic beta-cell destruction. The islets of Langerhans containing the insulinproducing beta-cells are scattered throughout the pancreas and comprise only 1-2% of the total tissue. They contain in addition to the beta-cells, the 3 other endocrine cell-types, the glucagon-producing alphacells, the somatostatin producing delta-cells and the PPcells producing pancreatic polypeptide, with the alpha(~ 20%) and beta-cells (~ 70%) constituting the major part of a normal islet (Fig.1). All the 4 endocrine cell-types are believed to differentiate from stem-cell progenitor cells through a common pathway, which late in this maturation process branches out into the specific cell-types. Since T1DM is caused by a specific destruction of the beta-cells in the islets, a major goal within T1DM research is to understand the mechanisms responsible for the beta-cell maturation and associated phenomena making the beta-cell specifically recognized and destroyed in T1DM. The etiology of T1DM is not yet fully understood. It is believed that genetic factors are a major compo-
206
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
Fig. 1. The pancreatic islets of Langerhans. Table 1 Candidate regions for Type 1 diabetes mellitus susceptibility genes identified through linkage analyses Locus IDDM1 (HLA) IDDM2 (INS) IDDM3 IDDM4 IDDM5 IDDM6 IDDM7 IDDM8 IDDM9 IDDM10 IDDM11 IDDM12 IDDM13 IDDM15 IDDM17 IDDM18 DXS1068 GCK
Chromosome 6p21 11pl5 15q26 11ql3 6q25 18q 2q31 6q25-q27 3q21-q25 10pll.2-qll.2 14q24.3-q31 2q33 2q34 6q21 10q25 5q33-q34 Xq 7p
nent in the etiology of T1DM. Genes within the MHC region on chromosome 6p21 comprise the major genetic locus determining predisposition to T1DM [46]. However, over the last 6-7 years extensive efforts have been invested in order to genetically map other chromosomal regions demonstrating evidence of linkage to T1DM [7-10]. This has led to the identification of at least 16 non-HLA loci demonstrating some evidence of linkage to T1DM (Table 1). Nevertheless, although large resources have been put into the genome scan approach, no evidence for any etiological mutation(s) has yet been obtained. We therefore believe that novel approaches or combinations of multidisciplinary approaches are needed to unravel the molecular basis of the disease. Combining genetic and proteomebased strategies may represent a valuable approach as
discussed below. In addition to the well-established candidate gene and linkage approaches for identifying disease predisposing genes and proteins, methods for expression profiling have been developed over the last years. Expression profiling is the evaluation and comparison of expression levels of transcripts and/or proteins in diseased versus non-diseased tissues or in experimental model systems hereof. Although the trigger(s) of T1DM is not yet identified, diet, environmental exposures and viral infections are putative etiologic agents involved in the pathogenesis. Most studies on diet components in the aetiology of T1 DM have focused on the role of early exposure to bovine serum albumin in cow's milk [11]. Other dietary components may include nitrite and nitrates from food and drinking water [12]. Although a single major component of the diet still has to be demonstrated of importance in T1DM, studies of the increasing incidence rate in low-risk populations migrating to high-risk areas strongly suggest an environmental component in the disease etiology. Viruses, as an environmental factor, have been implicated in T1DM pathogenesis based on both case reports of virus isolation from pancreata of acutely diabetic deceased patients and induction of diabetes in animal models by infection with virus as well as on epidemiological studies examining recent-onset T1DM patients for virus-specific antibodies [13]. An etiological model reconciling these different aspects is shown in Fig. 2. Research to elucidate the pathogenesis of T1DM has for many years mainly focused on the genetic and immnological markers associated with the disease and to a lesser degree on environmental factors. However, recently increasing focus has been directed towards the target cell, the beta-cells in the islets. Why are the beta-
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
207
Fig. 2. Etiological model of Type 1 diabetes mellitus. If the total beta-cell mass in a healthy individual is set to be 100%, some yet unrecognized environmental triggers may activate pathogenetic mechanisms leading to progressive loss of pancreatic islet beta-cells. This process is associated with a variety of immune phenomena, including infiltration of the islets with macrophages and auto-reactive T-lymphocytes (insulitis), and both T-cell (cell-mediated) and B-cell (humoral, antibody mediated) reactivity have been demonstrated prior to the onset of disease. However, the immunological mechanisms directly involved in beta-cell killing have not yet been clearly defined in man. Whereas only few diabetes-associated T-cell clones have been isolated, several autoantibodies to beta-cell proteins have been demonstrated up to several years before the onset of disease in blood from prediabetic and recent onset diabetic individuals. This autoimmune mediated destruction may progress slowly (over several years) or more rapid, dependent on the frequency and level of insults, influencing the balance between beta-cell destruction and regeneration in the individual islets. The end-point is the clinical onset of T1DM, when approximate 10% of the beta-cell mass remains. Until this time-point, the residual beta-cell mass is sufficient to maintain normal blood-glucose under normal conditions, however, the reduced beta-cell mass may become insufficient to challenges with higher demand for insulin secretion, as demonstrated by the loss of the first phase insulin response (FPIR).
cells specifically destroyed, when the other pancreatic endocrine cell-types of common origin in the islets, e.g. the alpha-cells, are not? The original demonstration that cytokines [1416], especially interleukin-l-ß (IL-ß) released from macrophages and T-lymphocytes during insulitis were specifically toxic to the beta-cells, enabled us in 1988 to construct a testable model of the pathogenesis of T1DM [17], the so-called "Copenhagen-model", outlined in details below and shown in its present form in Fig. 3. The model predicts that anything from the external or internal environment which can destroy a beta-cell (nutrients? virus? chemicals? cytokines?) will lead to the release of beta-cell proteins. These proteins will be taken up by residing antigen presenting cells (APC; e.g. macrophages (M0) and dendritic cells) in the islets, will be processed to antigenic peptides and as such be presented by MHC-Class II molecules on the cell surface. This activates the APCs to produce and secrete monokines (IL-1ß and tumor necrosis factor (TNF)a and co-stimulatory signal(s)), which, if T-helper lymphocytes with receptors specifically recognizing the antigenic peptide are present in the islet, induce the transcription of a series of lymphokine genes.
One of these, interferon (IFN)g, feedback stimulates the APC to increase expression of MHC-Class II molecules and IL-1ß and TNFa. In addition, other cells of the APC lineage present in the islet are induced to secrete monokines. IL-ß, potentiated by TNFa; and IFNg is cytotoxic to beta-cells, primarily through the induction of free radical (FR: NO, 02) formation in the islet. It has been confirmed by recent studies from several laboratories, including our own [18-24] that a role of FR is important, by demonstrating that nitric oxide (NO), and oxy gen-derived FR are induced in islet cells by cytokines, and that the inducible nitric oxide synthase (iNOS), responsible for the induced NO production is induced by IL-1ß in beta-cells (not present in other endocrine islet cells) from which we were able to clone it out [25]. In addition, also FR independent toxicity exists as shown in cytokine exposed human islets, where toxicity is still induced in the presence of NO-inhibitors as well as by combinations of cytokines, which do not induce FR production [26]. As part of the beta-cell destructive mechanism betacell proteins may be modified by the FR and in more antigenic forms presented to the immune system, thereby closing the loop in a self-perpetuating and selflimiting fashion. Thus, self-proteins presented by the MHC-Class-I molecules on the beta-cell or released during the beta-cell destruction by FR and taken up and
208
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
Fig. 3. Copenhagen model of beta-cell destruction.
presented by MHC-Class-H, need not be antigenic in their native conformation, but rather antigenecity may be a result of FR modification(s) [27]. This implies, that beta-cell proteins may become antigens when and because beta-cells are destroyed by cytokines. In addition the cytokine exposure has been shown to induce expression of FAS on the surface of the beta-cells. Interaction with the FAS-ligand on T-lymphocytes may therefore also represent a mechanism of induced betacell lysis/destruction [28,29]. The magnitude of beta-cell destruction will depend upon a) the velocity of the feed-back circuit between the APC and the T-helper lymphocyte, i.e. on the efficacy of antigen transport presentation/recognition, b) the magnitude and type of cytokine production, and c) on the capacity of beta-cell defense mechanisms during cytokine exposure [30]. The model has important implications, since it eliminates the need for the existence of specific environmental trigger(s), T1DM specific genes or gene mutations, and targeted exposure of beta-cells to the cytotoxic action of monokines. Pro-inflammatory cytokines induce both protective and deleterious mechanisms in all cells with cytokine receptors. The resulting race between protective and deleterious mechanisms determines the outcome, which is whether the cell will die or survive. Indeed, both protective and deleterious mechanisms have been demonstrated in response to cytokines in different cell types, however, the response varies between cell types, and within a cell-type between species and probably also between humans with different genetic background. Thus, we hypothesize that in genetically predisposed individuals, T1DM develops when the deleterious events in the beta-cells prevail.
This would suggest that the pathogenesis of T1DM, i.e., the earliest events and mechanisms affecting the target cell, has two distinct phases: a non-antigen driven, non-lymphocyte dependent initiation phase, and an antigen driven (by multiple antigens?), lymphocyte dependent amplification and perpetuation phase. The model further offers a possible explanation for the existing multitude of "autoantibodies" in the blood of newly diagnosed and prediabetic individuals (see below). On this background we postulate that the remarkable specificity of cellular destruction in the islets in T1DM is due to an inherent vulnerability of beta-cells to cytokine-induced FR damage, acquired during its differentiation into the highly developed and metabolically sophisticated beta-cell. In other words, when exposed to proinflammatory cytokines the beta-cells die because they are beta-cells. Hence, detailed studies of the intracellular molecular events resulting from the interaction between immune effector molecules/cells and the target cell are in demand, and proteome analysis is a valuable tool to address this aspect.
2. Antibody-markers of T1DM As mentioned above, T1DM is associated with the presence of one or more of a multitude of autoantobodies present in the blood of the majority of prediabetic and recent onset T1DM individuals. Whereas these antibody markers may show different degree of disease specificity [31,32], none of the known T1DMassociated autoantigens clearly qualifies for the position as the major or primary driving antigen responsi-
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
ble for the amplification and perpetuation of beta-cell destruction. Rather they may serve as disease markers of a progressive beta-cell destruction. Nevertheless, proteome analysis of cytokine exposed islets/betacells may help identify new auto-antigen candidates among native (e.g., de novo synthesized or changed cellular localization) or FR-modified beta-cell specific proteins, to be tested for potential reactivity with antibodies present in the blood of prediabetic and recent onset diabetic individuals.
3. Proteome analysis A cell is dependent upon a multitude of metabolic and regulatory pathways for survival. There is no strict linear relationship between the genes and the protein complement of a cell (i.e., mRNA levels may or may not correlate with the protein level) [33]. The ability of characterizing the protein expression profile in a cell or tissue at a given time point, reflecting the metabolic and functional status at that particular time point, is the hallmark of proteome analysis. Therefore, proteome technology is a suitable tool in the search for disease associated protein/protein pathways in T1DM and other diseases with a known target organ/cell. The possibility of comparing the status of the protein expression profile between healthy and diseased tissue/cells creates opportunities for new strategies to identify at the molecular level proteins and mechanisms involved in, or responsible for, disease progression. With focus on the beta-cell destruction associated with T1DM, this is discussed below.
4. Proteome analysis in T1DM - an in vitro approach As detailed above, it is believed that TlDM is associated with the destructive effects of cytokines released in the islets during the insulitis process occurring before clinical onset of the disease. The Copenhagen model (Fig. 3) details this process, and is a testable model, which can be approached using proteome analysis as follows: Pancreatic islets of Langerhans are isolated, kept in culture in vitro and exposed to different cytokines for fixed periods of time. Extensive research has documented that the metabolic function of such isolated islets (from either rat, mouse and humans) may be severely impaired by the co-culture with cytokines, which dependent on combination and concentration,
209
may also induce cell death [29]. Using this in vitro culture system with rat islets allowed us to make the first proteome analysis of cytokine-exposed islets [34-37]. Figure 4 illustrates the experimental design for these studies. First, isolated islets (from rat, mouse or human organ donors) are cultured in the absence or presence of cytokines for a set period of time (e.g. 24 h, or ideally, for several different time-points to allow analysis of the kinetics of changes in protein expression). Different strategies may be applied in order to subsequently identify and quantify the different proteins (e.g. radiolabelling, labelling with fluorescent dyes or silver staining), each method with its pros and cons. We have used 35-S labelled methionine to radiolabel the proteins for 4 h at the end of the cytokine exposure period. The next step in the procedure is the separation of the proteins from the control and cytokine exposed islets, respectively, by two-dimensional electrophoresis (2DGE) according to charge (pI) and size (Mw) by isoelectrical focusing (IEF) or non-equilibrium pHgradient electrophoresis (NEPHGE) and SDS-page gel electrophoresis, respectively. The individual proteins may now be visualized using a phosphoimager (or on a X-ray film) based on the incorporation of the 35-S labeled methionine, allowing quantification of the spots over a large dynamic range. An important point in the experimental design is to reduce the number of type 2 errors, e.g. minor irrelevant changes in protein expression, which may be caused by the experimental procedure as such. Therefore, it is essential to perform a number of independent experiments with islets material obtained from different batches of islets preparations (Panel 2, Fig. 4). This strategy is evident when thinking of analyses using human islets isolated from persons with different genetic background, where results obtained from one individual islet isolation, in addition to inherent influence of the genetics, may be influenced by variations in the tissue procurement and experimental setup, and therefore may not be reproducible in future experiments. Unfortunately, this necessary analysis of several gels from each group to be compared against each other represents a bottleneck in the experimental design. Whereas computer assisted alignment of the protein spots between the gels may be performed using different available software, even the most optimized 2DGE system will not consistently run gels 100% identical. Hence, careful manual evaluation and completion of the computer mediated alignment of the individual gels within the different experimental groups and between different groups is necessary, why a composite image is created and the different gels
210
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
Fig. 4. Experimental design for proteome analysis in Type 1 diabetes.
aligned to this. Having done that, it is possible to make statistical evaluation of the expression levels of the different proteins to identify those, which significantly change expression level in response to e.g. cytokine treatment (Panel 2, Fig. 4). The next step is to identify the proteins in the spots of changed expression level. This is effectively done by mass spectrometry (MS), as illustrated in Panel 3 (Fig. 4). The individual protein spots are excised from preparative gels, rehydrated and digested by a protease, e.g. trypsin, in situ in the gel plug. Following up-concentration of the material, it is subjected to MS-analysis and the resulting MS-profile analyzed in silico against all the theoretical MS-profiles calculated from available amino acid sequences present in public sequence databases. Based on this analysis and the available information of Mw and pI from the gel, the protein may be identified by its "fingerprint" if its sequence is present in the sequence databases. With completion of the human genome project, the success rate of identifying human proteins by mass spectrometry is now very high. Knowing the identity of all the proteins of changed expression levels, the next task is to clarify how these proteins may be involved in cytokine-mediated beta-
cell destruction and which changes may be primary and which may be secondary to this process. This part, in Fig. 4, Panel 4, referred to as bioinformatics, now a day involves computer-assisted searches of information e.g. with the aim to group the proteins according to function and pathways. Examples of this are given below. Identification of proteins and protein-pathways of primary interest for further functional characterization also involve genetic analyses, e.g. does the gene encoding the protein map to a chromosomal region demonstrating linkage to T1DM. The final step in the experimental design is to characterize the selected proteins/protein modifications for the putative involvement in cytokine mediated beta-cell destruction and TlDM by functional and further genetic analysis. This involves cloning and recombinant expression of the protein in beta-cell lines to elucidate if the protein influences the known effect of cytokines. If so, further studies in transgenic mice, including the animal model of diabetes, the nonobese diabetic (NOD) mouse, either by knock-out or over-expression of the gene of interest in the beta-cells under control of an insulin-promoter. Others and we have demonstrated the value of this method by overexpressing protective proteins in beta-cell lines result-
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
211
Fig. 5. Fluorograph of a 2-dimensional gel of rat islets of Langerhans. The proteins marked represent proteins changing expression level in response to IL-ß exposure. IEF gel (pH 3.5-7) on the right side; NEPHGE gel (pH 6.5-10.5) on the left side.
ing in protection of the cells against subsequent cytokine exposure [24,38,39]. The genetic analyses further may involve screening for polymorphisms of the gene, tests for disease association and linkage, and possible functional significance of identified gene variations. We have previously demonstrated the value of this approach [40-43].
5. In vitro analysis of il-ß exposed rat islets Using the experimental design detailed above, we set out to produce a database of rat islet proteins containing about 2,200 protein spots characterized by Mw and pI [34,35]. This demonstrated that IL-1ß exposure of the islets in culture resulted in reproducible and statistically significant modulation of protein expression levels or de novo synthesis of 105 of these proteins. Fiftytwo proteins were up-regulated, 47 down-regulated and 6 synthesized de novo [35] (Fig. 5). Subsequently, we have attempted to positively identify these 105 proteins in order to better understand beta-cell destruction and T1DM at the molecular level, since this would provide the first larger scale assessment of IL-1ß-mediated beta-cell damaging processes at the protein level.
A success rate of approximately 60% for positive identifications was obtained in this study [37]. Most identified proteins could be assigned classes according to their known or putative functions: i) energy transduction and redox potentials; ii) glycolytic pathway; iii) protein synthesis, chaperones, and protein folding; and iv) signal transduction, regulation, differentiation, and apoptosis [37]. These findings strongly support the hypothesis that that islet exposure to cytokines induces a complex pattern in beta-cells comprising protective (e.g. up regulation of stress proteins) as well as deleterious (e.g. iNOS induction and NO production) events. The overall picture is complex and reflects the range of cellular responses to the cytokine challenge. It is not known which protein changes should be considered 'primary' or 'secondary' in importance in time and in sequence. Furthermore, it has been shown that IL-1ß may induce both NO dependent as well as NO independent beta-cell impairment [34,44]. To further elucidate this, we have recently demonstrated that of the 105 protein spots, which changed expression levels in response to IL-1ß [35], only 23 were dependent on NO production [36]. In addition, the effect of the chemical NO donor S-nitroglutathione (GSNO) on protein expression was addressed, demonstrating changes in
212
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
In vitro analysis (rat and human islets)
Fig. 6. Extended proteome approach in the study of Type 1 diabetes.
expression levels of 19 islet proteins, which were not identified by MS [36]. This suggests that the majority of protein changes observed in rat islets exposed to IL1ß are independent of NO production. It is not clear whether the plethora of changes observed is secondary to the early effects of changes in just one or a few proteins, e.g. involved in mitochondrial energy generation. We are presently finalizing our analysis of isolated human islets exposed to Il-1ß alone or IL-1ß in combination with IFNg and TNFa. The data obtained in this study are even more complex, but the picture emerging show that the same pathways as in the IL-1ß exposed rat islets are affected.
6. In vivo analysis of IL-1ß exposed rat islets Whereas the experimental design and data presented above was designed to characterize the proteins involved in cytokine responses in vitro, we have applied proteome-based strategies also to address the influence of immune mediators on beta-cell destruction in vivo during disease progression using an animal model of
T1DM, the inbred diabetes prone BioBreeding (BB) rat [45,46]. The overall aim of this extended proteome approach (Fig. 6) is to include and combine information from different experimental procedures with the same focus to narrow down the number of proteins to be functionally characterized for their potential involvement in T1DM development. In this model, syngeneic islets, isolated from newborn rats are transplanted under the kidney-capsule of 30 days old BB rats (time of weaning), approximately 7 weeks before the onset of disease (Fig. 6). Beta-cells of the syngeneic islets transplanted under the kidney capsule are exposed to the same immunological events (e.g. infiltration with macrophages, T- and B-lymphocytes) as the host islets of the pancreas, and presumably they are destroyed in a process identical to that in the host islets. At different time points before disease onset, at the day of onset of disease and in the few animals escaping from diabetes, the transplanted islets are excised and used for proteome as well as histochemical analyses. This allows us to follow the changes in protein expression occurring in the islets until the onset of diabetes. Furthermore, analvses of the islets from the rats that es-
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
cape diabetes development, may also give valuable information. This complex set of data is presently being completed, and several of the proteins and pathways found in the in vitro experiments are also identified in the in vivo situation.
7. A model of beta-cell maturation An additional component in our extended proteome approach, is the use of a pluripotent cell line, which dependent of culture condition may mature from a glucagon-producing pre-beta-cell phenotype into an insulin producing beta-cell phenotype [47]. We have recently demonstrated that this maturation is accompanied by an acquired sensitivity to IL-1ß [22]. Therefore, this cell-system is useful for characterization of the mechanisms responsible for cytokine sensitivity. Proteome analysis is presently performed of the IL-1ß non-sensitive (pre-beta-cell phenotype) vs. the sensitive beta-cell phenotype in the presence or absence of cytokines (Fig. 6). As an example of a preliminary finding from this analysis, one spot was reduced fivefold in expression level following maturation into the cytokine sensitive beta-cell phenotype. After exposure to IL-lß, the spot was up-regulated in both phenotypes by a factor of approximately 3.5. This suggests that the protein could represent a scavenger protein involved in scavenging the IL-lß induced toxic FR. Insufficient up-regulation of a scavenger protein in the beta-cell phenotype may result in its destruction. However, at present we do not know the identity of this protein.
8. Perspective in proteome analysis As described above the use of proteome analysis in characterization of processes involved in or responsible for disease progression is a valuable and a rapidly developing area. One major advantage is that it allows analysis of proteins both in their native and modified form. This is in contrast to the rapidly developing transcriptome approach, focusing on the gene-expression profile (the steady state mRNA expression, Fig. 7), where modifications of proteins are not apparent from the mRNA. Thus, whereas the transcriptome technique clearly complements the proteome analysis [48,49], it does not provide all the information obtained by proteome analysis. It is well known that several cellular processes are dependent on posttranslational modifications of an already existing, readily available protein
213
pool in the cell. This includes phosphorylation, glycosylation, sulphatation or cleavage of precursor sequences as they can determine activity, stability, localization and turnover. The ability to detect the same protein in several different spots on a 2D-gel makes it possible to obtain information about potential posttranslational modifications. Whereas this "state of the art" technology for identifying the posttranslational modifications is rapidly developing, it is at present not easily included in the experimental strategies outlined above. A limitation of proteome analysis is the inability to separate all the proteins, even on a large gel. In line with the increasing sensitivity of the MS-identification, this becomes an increasing problem, since contamination with closely neighboring or partly overlapping proteins may result in identification of more than one protein in the same spot. Since MS is not quantitative, it is not possible to determine which protein is the one of changed expression without further analysis, e.g. Western blotting with specific antibodies. This problem may, however, be solved with the use of so called zoom-gels making it possible to look at only small parts of the pI or Mw area in separate gels, and finally combine the information form the different gels, obtaining an increased separation of the protein-spots. Indeed the area is in rapid progression and continued development and what is a bottleneck today may be easily solved tomorrow.
9. Conclusion The use of proteome analysis may elucidate relevant mechanisms for beta-cell death after cytokine exposure. Whether IL-lß induced changes in protein expression levels in rat islets in vitro will also reflect pathogenically important changes in beta-cells in rats spontaneously developing T1DM remains to be determined. Preliminary observations using 2D-gel studies of excised syngeneic islet transplants from different time points posttransplantation in BB-DP rats [46] suggest that the approach will be useful for studies of T1DM pathogenesis in vivo. Interestingly, the exquisite beta-cell sensitivity to cytokine toxicity may be an acquired trait developed during beta-cell maturation [22]. We believe that data obtained by the methodology presented here alone or in combination with e.g. genetic studies, transcriptome analyses and studies of candidate genes e.g. identified through the Copenhagen model (Fig. 8) will lead us to a detailed and complex picture of the molecular processes producing beta-cell
214
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
Genome
Transcriptome primary RNA transcript
DNA
mRNA
mRNA processing control
transcriptional control
Proteome
mR
mRNA transport control
mRNA degradation control
protein product
protein
i translational protein protein . , maturation modification(s) control control
protein degradation control
I
Fig. 7. Approaches in Type 1 diabetes research. Research addressing mechanisms involved in cytokine-mediated beta-cell destruction may be divided in three areas, focusing on the genome, the transcriptome or the proteome as detailed in the text.
Candidate approach
Genetic approach
Virus ?
Chemicals Nutrition 7
IDDM'X' human
Molecular mechanisms in ß-cell destruction in T1DM
Proteome approach
Idd'y'
mouse
Transcriptome
2-D gel
Fig. 8. Identification of molecular mechanisms in beta-cell destruction and Type 1 diabetes. A combined approach taken advantages from genomics, proteomics and biochemical studies is most likely to reveal the molecular mechanisms involved in the beta-cell destruction.
destruction in vitro. Although the picture is complicated and far from complete we are looking at the ailing beta-cell through a new window and the challenge now is to learn to fully understand what we see. Further proteome and transcriptome analyses may eventually complete the picture. The long-term perspective of this is the development of new and specific intervention
modalities in beta-cell destruction in T1DM. Making the beta-cell more resistant to immunological mediators may increase the survival time of transplanted islets or engineered beta-cells with reduced immunosuppressive modalities in treatment of T1DM patients, and potentially prevent the ongoing beta-cell destruction in predisposed individuals.
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus
References
[1] M. Karvonen, M. Viik-Kajander and E. Moltchanova et al., Incidence of childhood Type 1 diabetes worldwide, Diabetes Care 23 (2000), 1516-1526. [2] T. Lorenzen, F. Pociot, P. Hougaard and J. Nerup, Long-term risk of IDDM in first-degree relatives of patients with IDDM, Diabetologia 37 (1994), 321-327. [3] A.G. Moelbak, B. Christau, B. Mamer, K. Borch-Johnsen and J. Nerup, Incidence of insulin-dependent diabetes mellitus in age groups over 30 years in Denmark, Diabetic Med 11 (1994), 650-655. [4] J. Nerup, P. Platz and O.O. Andersen et al., HLA antigens and diabetes mellitus, Lancet 2 (1974), 864-866. [5] F. Pociot, Insulin-dependent diabetes mellitus - a polygenic disorder? Dan Med Bull (Thesis) 43 (1996), 216-248. [6] J.-X. She, Susceptibility to Type 1 diabetes: HLA-DQ and DR revisited, Immunol Today 17 (1996), 323-329. [7] L. Hashimoto, C. Habita and J. Beressi et al., Genetic mapping of a susceptibility locus for insulin-dependent mellitus on chromosome llq, Nature 371 (1994), 161-164. [8] J.L. Davies, Y. Kawaguchi and S.T. Bennett et al., A genomewide search for human susceptibility genes, Nature 371 (1994), 130-136. [9] P. Concannon, K. Gogolinewens and D. Hinds et al., A secondgeneration screen of the human genome for susceptibility to insulin-dependent diabetes-mellitus, Nature Genet 19 (1998), 292-296. [10] C. Mein, L. Esposito and M. Dunn et al., A search for type1 diabetes susceptibility genes in families from the United Kingdom, Nature Genet 19 (1998), 297-300. [11] H. Gerstein, Cow's milk exposure and Type 1 diabetes, Diabetes Care 17 (1994), 13-19. [12] G. Dahlquist, Environmental risk factors in human Type 1 diabetes - an epidemiological perspective, Diabetes Metab Rev 11 (1995), 37-46. [13] H.S. Jun and J.W. Yoon, The role of viruses in Type 1 diabetes: two distinct cellular and molecular pathogenic mechanisms of virus-induced diabetes in animals, Diabetologia 44 (2001), 271-285. [14] T. Mandrup-Poulsen, K. Bendtzen, J.H. Nielsen, G. Bendixen and J. Nerup, Cytokines cause functional and structural damage to isolated islets of Langerhans, Allergy 40 (1985), 424429. [15] T. Mandrup-Poulsen, K. Bendtzen and J. Nerup et al., Affinity purified human interleukin 1 is cytotoxic to isolated islets of Langerhans, Diabetologia 29 (1986), 63-67. [16] K. Bendtzen, T. Mandrup-Poulsen, J. Nerup, J.H. Nielsen and C. Dinarello, Cytotoxicity of human pI 7 interleukin-1 for pancreatic islets of Langerhans, Science 232 (1986), 15451547. [17] J. Nerup, T. Mandrup-Poulsen and J. M01vig et al., Mechanisms of pancreatic beta-cell destruction in Type 1 diabetes, Diabetes Care 11(1) (1988), 16-23. [18] W. Suarez-Pinzon, K. Strynadka, R. Schulz and A. Rabinowitch, Mechanisms of cytokine-induced destruction of rat insulinoma-cells: The role of nitric oxide, Endocrinology 134 (1994), 1006-1010. [19] D.L. Eizirik, D.G. Pipeleers and Z. Ling et al., Major species differences between humans and rodents in the susceptibility to pancreatic ß-cell injury, Proc NatlAcadSci USA 91 (1994), 9253-9256.
215
[20] J.A. Corbett and M.L. McDaniel, Intraislet release of interleukin 1 inhibits ß cell expression of inducible nitric oxide synthase, J Exp Med 181 (1995), 559-568. [21] M. Di Matteo, A. Loweth and S. Thomas et al., Superoxide, nitric oxide, peroxynitrite and cytokine combinaitons all causes functional impairment and morphological changes in rat islets of Langerhans and insulin secreting cell lines, but dictate cell death by different mechanisms, Apoptosis 2 (1997), 164-177. [22] K. Nielsen, A.E. Karlsen and M. Deckert et al., Beta-cell maturation leads to in vitro sensitivity to cytotoxins, Diabetes 48 (1999), 2324-2332. [23] A.E. Karlsen, D. Pavlovic and K. Nielsen et al., Interferongamma induces interleukin-1 converting enzyme expression in pancreatic islets by an interferon regulatory factor-1dependent mechanism, J Clin Endocrinol Metabol 85 (2000), 830-836. [24] S. Lortz, M. Tiedge and T. Nachtwey et al., Protection of insulin-producing RINm5F cells against cytokine-mediated toxicity through overexpression of antioxidant enzymes, Diabetes 49 (2000), 1123-1130. [25] A.E. Karlsen, H.U. Andersen and H. Vissing et al., Cloning and expression of cytokine inducible nitric oxide synthase cDNA from rat islets of Langerhans, Diabetes 44 (1995), 753-758. [26] C.A. Delaney, D. Pavlovic, A. Hoorens, D.G. Pipeleers and D.L. Eizirik, Cytokines induce deoxyribonucleic acid strand breaks and apoptosis in human pancreatic islet cells, Endocrinology 138 (1997), 2610-2614. [27] A.E. Karlsen and T. Dyrberg, Molecular mimicry between non-self, modified self and self in autoimmunity, Sem Immunol 10 (1998), 25-34. [28] J. Nerup, T. Mandrup-Poulsen and S. Helqvist et al., On the pathogenesis of IDDM, Diabetologia 37(2) (1994), S82-S89. [29] T. Mandrup-Poulsen, The role of interleukin-1 in the pathogenesis of insulin-dependent diabetes mellitus, Diabetologia 39 (1996), 1005-1029. [30] B. Freiesleben-De Blasio, P. Bak, F. Pociot, A.E. Karlsen and J. Nerup, Onset of Type 1 diabetes: a dynamical instability, Diabetes 48 (1999), 1677-1685. [31] W.A. Hagopian and A. Lernmark, Islet cell autoantibodies (1996), 441-448. [32] W.E. Winter, The use of islet autoantibody markers in the prediction of autoimmune Type 1 diabetes, Clin Immunol Newsletter 19 (1999), 25-39. [33] S.P. Gygi, Y. Rochon, B.R. Franza and R. Aebersold, Correlation between protein and mRNA abundance in yeast, Mol Cell Biol 19 (1999), 1720-1730. [34] H.U. Andersen, P.M. Larsen and S.J. Fey et al., Twodimentionel gel electrophoresis of rat islets proteins: Interleukin-1 beta induced changes in protein expression are reduced by L-argenine depletion and nicotinamide, Diabetes 44 (1995), 400-407. [35] H.U. Andersen, S.J. Fey and P. Mose Larsen et al., InterleukinIbeta induced changes in the protein expression of rat islets, Electrophoresis 18 (1997), 2091-2103. [36] N. John, H. Andersen and S. Fey et al., Cytokine or chemicallyderived nitric oxide alters the expression of proteins detected by two-dimensional gel electrophoresis in neonatal rat islets of Langerhans, Diabetes 49 (2000), 1819-1829. [37] P.M. Larsen, SJ. Fey and M.R. Larsen et al., Proteome analysis of interleukin-1 beta-induced changes in protein expression in rat islets of Langerhans, Diabetes 50 (2001), 1056-1063. [38] A. Rabinovitch, W. Suarez-Pinzon and K. Strynadka et al., Transfection of human pancreatic islets with an and- apop-
216
[39]
[40]
[41]
[42]
[43]
A.E. Karlsen et al. / Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus totic gene (bcl-2) protects beta-cells from cytokine- induced destruction. Diabetes 48 (1999), 1223-1229. S. Grey, M. Avelo, W. Hasenkamp, F. Bach and C. Ferran. A20 inhibits cytokine-induced apoptosis and nuclear factor kB-dependent gene activation in islets, J Exp Med 190 (1999). 1135-1145. F. Pociot, J. M01vig, L. Wogensen, H. Worsaae and J. Nerup, A TaqI polymorphism in the human interleukin-1ß (IL-13) gene correlates with IL-1ß secretion in vitro, EurJ Clin Invest 22(1992), 396-402. F. Pociot, A.G. Wilson, J. Nerup and G.W. Duff, No independent association between a TNFa promoter region polymorphism and insulin-dependent diabetes mellitus (IDDM), Eur J Immunol 23 (1993), 3043-3049. O.P. Kristiansen, M. Zamani and J. Johannesen et al.. Linkage and association between a CD4 gene polymorphism and IDDM in Danish IDDM patients. Diabetes 47 (1997), 281283. R. Bergholdt, Z.M. Larsen and N.A. Andersen et al., Characterization of new polymorphisms in the 5' UTR of the human interleukin-1 receptor Type 1 (IL1R1) gene: linkage to Type 1 diabetes and correlation to IL-1RI plasma level. Genes Immunol 1 (2000). 495-500.
[44]
[45]
[46]
[47]
[48]
[49]
D.L. Eizirik, M. Flodstrom. A.E. Karlsen and N. Welsh, The harmony of the spheres: inducible nitric oxide synthase and related genes in pancreatic beta cells. Diabetologia 39 (1996). 875-890. U. Christensen, T. Sparre and A. Cooke et al., Syngeneic islet transplantation in prediabetic BB-DP rats - a synchronized model for studying beta-cell destruction during the development of IDDM, Autoimmunity 28 (1998). 91-107. U.B. Christensen, P.M. Larsen and SJ. Fey et al., Islet protein expression changes during diabetes development in islet syngrafts in BB-DP rats and during rejection of BB-DP islet allografts, Autoimmunity 32 (2000). 1-15. O.D. Madsen, C. Karlsen and P. Serup et al., in: Insulin secretion and pancreatic B-cell research, PR. Flatt and S. Lenzen, eds, Smith-Gordon, UK, 1994, pp. 61-68. K. Rieneck, L. Bovin and K. Josefsen et al., Masssive parallel gene expression profiling of RINm5F pancreatic islet ß-cells stimulated with interleukin-1ß, APMIS 108 (2000). 855-872. A. Cardozo, M. Kruhoeffer, R. Leeman, T. Omtoft and D. Eizirik, Identification of novel cytokine-induced genes in pancreatic beta-cells by high density oligonucleotide arrays. Diahetes 50 (2001). 909-920.
217
Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma Philip Shalhoub, Sarah Kern, Sophie Girard and Laura Beretta* Department of Microbiology and Immunology and Cancer Center, University of Michigan Medical School, Ann Arbor, MI 48109, USA There is increasing evidence for an immune response to cancer in humans, demonstrated in part by the identification of autoantibodies to tumor antigens. The identification of panels of tumor antigens that elicit a humoral response may have utility in cancer screening, diagnosis or in establishing prognosis. Several approaches are currently available for the identification of tumor antigens. We have used a proteomic-based approach for the identification of tumor antigens that induce an antibody response which we have applied to hepatocellular carcinoma, a major type of cancer worldwide. Twodimensional gel electrophoresis allows simultaneous separation of several thousand individual proteins from tumor tissue or tumor cell lines. Proteins eliciting a humoral response in HCC were identified by 2-D Western blotting using sera from patients with hepatocellular carcinoma, followed by mass spectrometry analysis and database search. The common occurrence of autoantibodies to specific proteins may have utility for HCC screening and diagnosis.
1. Introduction Primary liver cancer is a prevalent tumor worldwide. The incidence of hepatocellular carcinoma (HCC), the major histolpgical form of primary liver cancer, has substantially increased in Japan [1], Western Europe [2, 3], and the United States [4] over the past two decades. The age-specific incidence of this cancer has also shown a progressive shift towards younger people (40 * Address for correspondence: Laura Beretta, Dept. of Microbiology and Immunology, University of Michigan, 1510 MSRB-I, 1150 W. Medical Ctr. Dr., Ann Arbor, MI 48109-0666, USA. Tel: +1 734 615 5964; Fax: +1 734 615 6150; E-mail:
[email protected]. Disease Markers 17 (2001) 217-223 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
to 60 years old). Human hepatocarcinogenesis is a multistage process with the involvement of a multifactorial aetiology and many gene-environment interactions. It is important to emphasize the heterogeneity of the histological background on which the tumor develops. The majority of HCCs are associated with cirrhosis (at least 90% in America and Europe). These tumors often have a poor prognosis, with five-year survival rates of less than 5%. Some tumors will occur in livers with minimal histological changes and "benign" adenomas can even develop in normal livers. This heterogeneity likely reflects different environmental, as well as possibly genetic factors. A large number of epidemiological and molecular studies have clearly indicated the major importance of environmental factors in the development of primary liver cancers in humans. The role of genetic factors has also been raised but is difficult to properly address, due to confounding variables such as intrafamilial transmission of HBV [5].
2. Risk factors for the development of HCC Some factors in the pathogenesis of HCC have been defined and almost all tumors occur in the context of chronic liver-cell injury, inflammation, and increased turnover of hepatocytes. Cirrhosis is a very important risk factor for HCC [6], with the risk increasing by a factor of approximately 200 after the onset of cirrhosis, and up to 30-50% of patients with cirrhosis developing an HCC upon 10 years follow up. Chronic infection by hepatitis B (HBV) or C (HCV) virus is also a major risk factor for HCC, and development of a chronic carrier state is a most frequent event following acute viral infection [7,8]. The global distribution of HCC correlates well with the geographic prevalence of chronic carriers of HBV, who number 400 million worldwide. With persistent HBV infection, the risk of HCC increases by a factor of 100 [9].
218
P. Shalhoub et al. / Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma
Among those who become infected with HBV at birth, men have an estimated lifetime risk of HCC of 50%, while women have a risk of 20% [10]. Persistent infection with HCV is also an important risk factor for HCC development. As is true for HBV, the relative risk of HCC among persons with chronic HCV infection is approximately 100 times the risk in uninfected persons. Persistent HCV infection is the cause of 70%of the cases of HCC in Japan [11], and the most likely reason for the rising incidence of HCC in the United States is the spread of HCV infection in the population. Four million people in the United States have chronic HCV infection and persistent HCV infection is the cause of approximately 30-50% of the HCC cases in the United States. The mechanisms involved in virally related liver carcinogenesis still remain largely undefined. There is evidence for direct effects of HBV in this process; in contrast, there is presently only very preliminary information regarding the carcinogenic role of HCV. Persistent viral infection causes inflammation, increased cell turnover and cirrhosis. Furthermore, the HBV genome may be integrated into the chromosomes of hepatocytes, contributing to genomic instability. As in persistent HBV infection, persistent HCV infection also initiates inflammation, cellular injury, regeneration, and cirrhosis, all of which may contribute to the oncogenic process. Prevention of chronic infection with these viruses by immunization is a high priority, and childhood vaccination against HBV carries the greatest potential for reducing the liver cancer burden. Safe and effective HBV vaccines are available, both plasmaderived and recombinant DNA-derived, although they are not yet used universally because of cost. An HCV vaccine is not yet available and may be difficult to develop because of the high mutation rate of the viral genome and variability of genotypes. Epidemiological studies have demonstrated a strong association between exposure to aflatoxin and an increased incidence of HCC [12]. Aflatoxins have been found as contaminants in food, particularly in corn, peanut oil, soya sauce and fermented soya beans. Excessive alcohol consumption and tobacco smoking may also be associated with increased risk of HCC.
3. Prevention of HCC A goal of cancer prevention is the detection of latent premalignant or malignant clones before they expand to a clinically detectable rumor. Resection of small HCC
remains an important approach in achieving long-term HCC survival and to improving 5-year survival rates. It is more effective than treatment of large HCC, surgical cure being rarely possible. Therefore, it is accepted that early detection, diagnosis and treatment of HCC remains an important target to be achieved before a breakthrough appears on the primary prevention of HCC. As high risk factors for developing HCC have been established, mass health screening for early detection of HCC is available. Initial screening should be done as early as possible on patients with chronic HBV and HCV infection, cirrhosis, and persons who had blood transfusions or a family history of HCC. The screening test currently used and widely practiced for HCC includes serum alpha-fetoprotein (AFP) levels and abdominal ultrasound examinations. Annual screening with ultrasound and AFP fails to identify potentially curable tumors because the diagnosis is often only made at a late stage of the disease [13]. Therefore a follow-up every 3 to 6 months is necessary. Although the screening can identify small tumors, survival may not be improved because the presence of cirrhosis may limit the number of patients who can undergo resections. In addition, recurrences or second primary tumors are common. Screening all liver cirrhosis patients is a questionable approach because it is very expensive and its benefit in terms of patient survival is poor. Conflicting data on the utility and efficacy of screening patients with cirrhosis for early detection of HCC and failures in the screening of chronic hepatitis B or C virus-infected patients with ultrasound and AFP have been reported [14]. Additional targeted screening programs with definite risk factors are dramatically needed.
4. Tumor markers Several tumor markers of HCC have been identified. However, no evidence has been obtained indicating that the detection of these markers precedes clinical diagnosis of HCC. Aberrant expression of the AFP gene in serum is characteristic of a majority of HCC cases and is widely used as tumor marker in the evaluation of prognosis and management of patients with HCC [15]. Long considered the fundamental marker for diagnosis of HCC, the usefulness of AFP as a marker has slowly become overshadowed by its inability to efficiently diagnose early-stage tumors. A high false-negative diagnosis rate, sometimes reaching 40% [16], is also a problem associated with this marker. A fucosylated
P. Shalhoub et al. / Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma
subtype of AFP, termed Lens culinaris agglutinin A reactive AFP, has been shown to be significantly better in identifying early HCC than general AFP levels [17-19]. The search for new tumor markers has long been sought. Many searches are not aimed at replacing this rather valuable marker, but instead hope to couple AFP serum levels with another marker to produce a more definitive diagnosis. An example of such a marker is des-gamma-carboxy prothrombin (DCP), an abnormal prothrombin which has long been considered a reasonable indicator of HCC, particularly when used as a complementary marker with AFP [20-28]. It has been suggested that elevation of DCP levels in HCC patients are due to a vitamin K deficiency in the cancerous tissue [29]. DCP serum level was found to be one of the most significant indicators of recurrence of HCC in patients [30]. DCP levels have also been found to be a good indicator of portal venous invasion (PVI) in HCC patients, which heralds the progression of the disease [31]. One issue concerning DCP is the difficulty in detecting a small level in the serum, which is often a diagnostic indicator in early stage HCC patients. New generations of sensitive immunoassays for detection of minute serum DCP levels have recently been reported to be effective in patients with small-sized HCC [26,27,32]. The diagnostic value of serum gamma-glutamyl transferase (GGT) has also been under evaluation [33], particularly that of an isoenzyme specific to HCC. Recently, hepatoma-specific bands of this enzyme (HSGGT) were identified and found to be significantly increased in patients with HCC as compared to acute hepatitis, chronic hepatitis, cirrhosis, and extrahepatic tumors, and thus were found to be a useful tumor marker for AFP-negative HCC patients. Its use in combination with AFP serum levels has been proposed for monitoring chronic liver disease patients and diagnosing HCC [16]. The methylation status of the GGT gene has also been given consideration as hypomethylation of the CCGG sites of the GGT gene has been implicated in the abnormal expression of GGT in HCC patients [16]. Platelet-derived endothelial cell growth factor (PDECGF) has been demonstrated to be highly expressed in advanced stage HCC patients and has been suggested to induce rich neovascularization of tissue in HCC [34]. Although there is little usefulness for PD-ECGF measurement as an early diagnostic marker for HCC, its value lies in the prospect of using PD-ECGF as a prognostic marker of tumor development in advanced stage patients. Plasma vascular endothelial growth factor (VEGF) levels were also shown to be significantly in-
219
in patients with HCC. More specifically, the largest increase was seen after the disease has metastasized, which suggests some utility as a cancer stage specific marker [35], It has been found to be a useful marker in detecting metastasis of HCC, and may have valuable prognostic value in patient care. A recent study has confirmed both VEGF and PD-ECGF as useful markers for identifying HCC, and has also reported a strong correlation between portal vein tumor thrombus (PVTT) development in HCC patients and serum levels of these markers [36]. A few additional tumor markers have been proposed, such as: plasma nitrate/nitrite [37,38], MXR7 [39] fibroblast growth factor [40], the CD24 gene [41], alpha-L-fucosidase [42], serum C-reactive protein [43], and activin-A [44]. Further research will determine whether or not these markers will prove valuable in diagnosing HCC.
5. Identification of new HCC tumor-associated antigens using proteomics The common occurrence of autoantibody formation to certain cancer-related proteins may also have value in cancer screening and diagnosis. For example, mutations in the p53 tumor suppressor gene are present in up to 37% of patients with HCC. Conformational change and cellular accumulation can initiate an immune response with generation of circulating autoantibodies to p53 protein. However, the presence of p53 autoantibodies in patients with chronic liver disease is not completely specific for HCC and no direct evidence has been obtained that indicates p53 autoantibody formation precedes the clinical diagnosis of HCC [45,46]. Antigens that have been shown to induce a humoral response in HCC include diverse other nuclear proteins [47-51], cyclin Bl [52] and a novel cytoplasmic protein with RNA-binding motifs [53]. Methods have been developed to identify tumor associated antigens such as molecular cloning in expression systems [54,55] or using a biochemical strategy, based on the extraction of antigenic peptides bound to major histocompatibility complex class I molecules from tumor cells [56,57]. These methods have allowed the recognition of several human tumor antigens [58-62]. A method called SEREX, 'serological identification of antigens by recombinant expression cloning', has been recently used for the identification of tumor antigens [63]. The SEREX analysis is based on screening of autoantibodies in sera from cancer patients against an
220
P. Shalhoub et al. / Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma
Fig. 1. 2-D PAGE protein profile of hepatoma cell lines and tumor tissues. Total proteins from the hepatoma cell lines Huh7 and HepG2 (A) and from tumor tissues obtained from 2 patients with HCC (B) were separated by 2-D electrophoresis and subsequently silver-stained.
Fig. 2. Detection of autoantibodies in sera from patients with HCC. The proteins from the hepatoma cell line Huh7 were separated by 2-D electrophoresis and subsequently transferred on PVDF membranes for Western-blotting experiments using sera from a healthy individual (left panel) or from a patient with HCC (right panel) as a first antibody.
expression library made with the RNA from the autologous tumor. Through application of this strategy, an unexpected frequency of tumor antigens that elicit specific immune responses in the autologous host has been observed [63-66]. The SEREX approach is limited by the necessity to construct expression libraries and the analysis is usually restricted to one or a few patients. In addition, the approach does not allow the identification of antibodies directed against post-translational modifications. A SEREX study of hepatocellular carcinoma
has uncovered reactivity to diverse proteins involved in the transcription/translational machinery as well as to chaperone proteins [67]. In order to identify proteins eliciting humoral responses in HCC patients, we used a proteome-based approach that has been recently implemented in other studies [68-70]. Several thousand cellular proteins from hepatoma cell lines or from liver tumor tissues, were separated by 2-D PAGE. Figure 1 shows the protein profile of 2 commonly used hepatoma cell lines,
P. Shalhoub et al. / Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma
Huh7 and HepG2 (Fig. 1(A)) and of tumor tissues isolated from 2 patients with HCC (Fig. 1(B)). Proteins from hepatoma cell lines or from liver tumor tissues were then transferred onto membranes. Sera from HCC cancer patients were screened individually for antibodies that react against separated proteins. The autoantigens were detected using a secondary antibody directed against human IgM or IgG, followed by autoradiography. As shown in Fig. 2, a greater number of reactive proteins were detected in general with sera from patients with HCC than with sera from healthy individuals. Proteins that specifically reacted with sera from HCC patients were located on silver-stained 2-D gels after super-imposition with the blots, extracted from the gel, digested and identified by mass spectrometric analysis and/or amino acid sequencing. We identified 8 proteins for which autoantibodies were detected in sera of more than 10% of 37 patients with HCC tested, but not in sera from healthy individuals (manuscript submitted). Autoantibodies against four of these proteins (beta-tubulin, hsp60, cytokeratin 18 and creatine kinase-B) were detected at a comparable frequency in sera from patients with chronic hepatitis. The other four proteins, which consisted of calreticulin isoforms, cytokeratin 8, nucleoside diphosphate kinase A and Fl-ATP synthase beta subunit all induced autoantibodies among patients with HCC independently of their HBV/HCV status. The protein Fl-ATP synthase beta subunit was previously reported to be antigenic in patients with HCC, by SEREX [67]. Calreticulin and a protein spot with an estimated MW of 32 kDa most frequently elicited autoantibodies among patients with HCC (27%). We previously identified this 32 kD protein as a new truncated form of calreticulin corresponding to the C-terminal end of the protein, and designated this novel form Crt32 [71]. The protein calreticulin has been identified as an autoantigen in various rheumatic diseases [72]. However, whereas the epitopes eliciting a humoral response in patients with autoimmune diseases have been reported to be located in the N-terminal part of the molecule, the epitopes eliciting a humoral response in patients with HCC in our study, are located in the C-terminal portion. In addition, autoantibodies against Crt32 were largely restricted to liver cancer patients among the different cancer sera we have analyzed. 6. Conclusion A proteome-based approach allows individual screening of a large number of patient sera as well as detection
221
of autoantibodies directed against post-translational modifications of specific targets. We observed a distinct repertoire of autoantibodies associated with HCC, reflecting the heterogeneity of the tumor. These autoantibodies may have utility in early diagnosis of HCC among high-risk subjects with chronic hepatitis and/or cirrhosis and may be used in combination assay with serum AFP or other marker levels. This global analysis also emphasizes the need for specific markers used in targeted screening programs with defined risk factors. References [1]
[2]
[3] [4]
[5]
[6]
[7]
[8]
[9]
[10]
[11] [12]
[13]
[14]
K. Okuda, I. Fujimoto, A. Hanai AND Y. Urano, Changing incidence of hepatocellular carcinoma in Japan, Cancer Res 47 (1987), 4867-4972. S.D. Taylor Robinson, G.R. Foster, S. Arora, S. Hargreaves and H.C. Thomas, Increase in primary liver cancer in the UK, 1979-94, Lancet 350 (1997), 1142-1143. S. Deuflic, T. Poynard, L. Buffat and A.J. Valleron, Trends in primary liver cancer, Lancet 351 (1998), 214-215. H.B. El-Serag and A.D. Mason, Rising incidence of hepatocellular carcinoma in the United States, N Engl J Med 340 (1999), 745-750. N.R. Drinkwater and G.H. Lee, Genetic susceptibility to liver cancer, Liver Regeneration and Carcinogenesis (1995), 301314. K. Dceda, S. Saitoh and Koida et al., A multivariate analysis of risk factors for hepatocellular carci, ogenesis: A prospective observation of 795 patients with viral and alcoholic cirrhosis, Hepatology 18 (1993), 47-53. C. Brechot, F. Jaffredo and D. Lagorce et al., Impact of HBV, HCV and GBV-C/HGV on hepatocellular carcinomas in Europe: results of a european concerted action, J Hepatol 29 (1998), 173-183. K. JJceda, S. Saitoh and Y. Suzuki et al., Disease progression and hepatocellular carcinogenesis in patients with chronic viral hepatitis: a prospective observation of 2215 patients, J Hepatol 28 (1998), 930-938. R.P. Beasley, L.Y. Hwang, C.C. DLin and C.S. Chien, Hepatocellular carcinoma and hepatitis B virus: a prospective study of 22 707 men in Taiwan, Lancet 2 (1981), 1129-1133. D. Moradpour and J.R. Wands, Hepatic oncogenesis, in: Hepatology: a textbook of liver disease, (Vol. 2), (3rd ed.), D. Zakim and T.D. Boyer, eds, W.B. Saunders, Philadelphia, 1996, pp. 1490-1512. T. Heintges and J.R. Wands, Hepatitis C virus: epidemiology and transmission, Hepatology 26 (1997), 521-526. C.J. Chen, L.Y. Wang and S.N. Lu et al., Elevated aflatoxin exposure and increased risk of hepatocellular carcinoma, Hepatology 24 (1996), 38-42. F. Tradati, M. Colombo and P.M. Mannucci et al., A prospective multicenter study of hepatocellular carcinoma in italian hemophiliacs with chronic hepatitis C. The study group of the association of italian hemophilia centers, Blood 91 (1998), 1173-1177. F. Izzo, F. Cremona, F. Ruffolo, R. Palaia, V. Parisi and S.A. Curley, Outcome of 67 patients with hepatocellular cancer detected during screening of 1125 patients with chronic hepatitis, Ann Surg 227 (1998), 513-518.
P. Shalhoub et al. / Proteomic-based approach for the identification of tumor markers associated with hepatoceliular carcinoma [15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
S. Ho and P.J. Johnson, Detection of alphafetoproteinexpressing cells in the blood of patients with hepatoma and hepatitis, Br J Cancer 77 (1998), 2059. D. Yao, D. Jiang and Z. Huang et al., Abnormal expression of hepatoma specific gamma-glutamyl transferase and alteration of gamma-glutamyl transferase gene methylation status in patients with hepatoceliular carcinoma, Cancer 88(4) (15 Feb. 2000), 761-769. K. Taketa, Y. Endo and C. Sekiya et al., A collaborative study for the evaluation of lectin-reactive a-fetoproteins in early detection of hepatocellular carcinoma, Cancer Res 53 (1993), 5419-5423. Y. Aoyagi, Carbohydrate-based measurements on alphafetoprotein in the early diagnosis of hepatocellular carcinoma, Glycoconj J 12(3) (June 1995), 194-199. F. Yamashita, M. Tanaka, S. Satomura and K. Tanikawa. Prognostic significance of Lens culinaris agglutinin A-reactive alpha-fetoprotein in small hepatocellular carcinomas, Gastmenterology 111(4) (Oct 19%), 996-1001. H.A. Liebman, B.C. Furie and M. Tong et al., Des-7-carboxy (abnormal) prothrombin as a serum marker of primary hepatocellular carcinoma, N Engl J Med 310 (1984), 1427-1431. S. Fujiyama, T. Morishita, K. Sagara, T. Sato, K. Motohara and I. Matsuda, Clinical evaluation of plasma abnormal prothrombin (PIVKA-II) in patients with hepatocellular carcinoma, Hepatogastroenterology 33(5) (Oct 1986), 201-205. C.H. Ho, S.D. Lee, H.T. Chang, J.C. Wu, Y.T. Tsai and K.J. Lo, Application of des-gamma-carboxy prothrombin as a complementary tumor marker with alpha-fetoprotein in the diagnosis of hepatocellular carcinoma, Scand J Gastroenterol 24(1) (Jan 1989), 47-52. A. Nakao, A. Virji, Y. Iwaki, B. Carr, S. Iwatsuki and E. Starzl, Abnormal prothrombin (DES-gamma-carboxy prothrombin) in hepatocellular carcinoma, Hepatogastroenterology 38(5) (Oct 1991), 450-453. I.C. Weitz and H.A. Liebman, Des-gamma-carboxy (abnormal) prothrombin and hepatocellular carcinoma: a critical review, Hepatology 18(4) (Oct 1993), 990-997. G.L. Grazi. A. Mazziotti and C. Legnani et al., The role of tumor markers in the diagnosis of hepatocellular carcinoma, with special reference to the des-gamma-carboxy prothrombin, Liver Transpl Surg 1(4) (July 1995), 249-255. H. Okuda, T. Nakanishi and K. Takatsu et al., Measurement of serum levels of des-gamma-carboxy prothrombin in patients with hepatocellular carcinoma by a revised enzyme immunoassay kit with increased sensitivity, Cancer 85(4) (15 Feb 1999), 812-818. K. Soga, T. Watanabe, K. Aikawa, M. Toshima, K. Shibasaki and Y. Aoyagi, Serum des-gamma-carboxyprothrombin level by a modified enzyme immunoassay method in hepatocellular carcinoma: clinical significance in small hepatocellular carcinoma, Hepatogastroenterology 45(23) (Sep-Oct 1998), 1737-1741. T. Sassa, T. Kumada, S. Nakano and T. Uematsu, Clinical utility of simultaneous measurement of serum high-sensitivity des-gamma-carboxy prothrombin and Lens culinaris agglutinin A-reactive alpha-fetoprotein in patients with small hepatocellular carcinoma, Eur J Gastroenterol Hepatol 11(12) (Dec 1999), 1387-1392. T. Miyakawa, Y. Kajiwara, A. Shirahata, K. Okamoto, H. Itoh and K. Ohsato, Vitamin K contents in liver tissue of hepatocellular carcinoma patients, Jpn J Cancer Res 91(1) (Jan 2000). 68-74.
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
M. Shimada, Y. Yamashita and T. Hamatsu et al., The role of des-gamma-carboxy prothrombin levels in hepatocellular carcinoma and liver tissues, Cancer Lett 159(1) (16 Oct 2000). 87-94. Y. Koike, Y. Shiratori and S. Sato et al., Des-g-Carboxy Prothrombin as a useful predisposing factor for the development of portal venous invasion in patients with hepatocellular carcinoma: a prospective analysis of 227 patients. Cancer 91(3) (1 Feb 2001), 561-569. F. Nomura, M. Ishijima, K. Kuwa, N. Tanaka, T. Nakai and K. Ohnishi, Serum des-gamma-carboxy prothrombin levels determined by a new generation of sensitive immunoassays in patients with small-sized hepatocellular carcinoma, Am J Gastroenterol 94(3) (March 1999), 650-654. K. Xu, X.Y. Meng, J.W. Wu, B. Shen, YC. Shi and Q. Wei, Diagnostic value of serum gamma-glutamyl transferase isoenzyme for hepatocellular carcinoma: a 10-year study. Am J Gastroenterol 87(8) (Aug 1992), 991-995. K. Jin-no, M. Tanimizu and I. Hyodo et al., Circulating platelet-derived endothelial cell growth factor increases in hepatocellular carcinoma patients. Cancer 82(7) (1 April 1998). 1260-1267. K. Jin-no, M. Tanimizu and I. Hyodo et al., Circulating vascular endothelial growth factor (VEGF) is a possible tumor marker for metastasis in human hepatocellular carcinoma. J Gastroenterol 33(3) (June 1998), 376-382. J. Zhou, Z.Y. Tang and J. Fan et al., Expression of plateletderived endothelial cell growth factor and vascular endothelial growth factor in hepatocellular carcinoma and portal vein tumor thrombus, J Cancer Res Clin Oncol 126(1) (Jan 2000). 57-61. A. Moriyama, A. Masumoto and H. Nanri et al., High plasma concentrations of nitrite/nitrate in patients with hepatocellular carcinoma, Am J Gastroenterol 92(9) (Sep 1997), 1520-1523. A. Moriyama, A. Tabaru, H. Unoki, S. Abe, A. Masumoto and M. Otsuki, Plasma nitrite/nitrate concentrations as a tumor marker for hepatocellular carcinoma, Clin Chim Acta 296(12) (June 2000), 181-191. H.C. Hsu, W. Cheng and P.L. Lai, Cloning and expression of a developmentally regulated transcript MXR7 in hepatocellular carcinoma: biological significance and temporospatial distribution, Cancer Res 57(22) (15 Nov 1997), 5179-5184. P.I. Hsu, N.H. Chow and K.H. Lai et al., Implications of serum basic fibroblast growth factor levels in chronic liver diseases and hepatocellular carcinoma, Anticancer Res 17(4A) (JulAug 1997), 2803-2809. L.R.Huang and H.C. Hsu, Cloning and expression of CD24 gene in human hepatocellular carcinoma: a potential early tumor marker gene correlates with p53 mutation and tumor differentiation. Cancer Res 55(20) (15 Oct 1995), 4717-4721. P. Tangkijvanich, P. Tosukhowong and P. Bunyongyod et al., Alpha-L-fucosidase as a serum marker of hepatocellular carcinoma in Thailand, Southeast Asian J Trop Med Public Health 30(1) (March 1999), 110-114. Z.Y. Lin, L.Y. Wang and M.L. Yu et al., Role of serum Creactive protein as a marker of hepatocellular carcinoma in patients with cirrhosis, J Gastroenterol Hepatol 15(4) (Apr 2000), 417-421. M. Pirisi, C. Fabris and S. Luisi et al., Evaluation of circulating activin-A as a serum marker of hepatocellular carcinoma, Cancer Detect Prev 24(2) (2000), 150-155. J. Raedle, G. Oremek and M. Truschnowitsch et al., Clinical evaluation of autoantibodies to p53 protein in patients with
P. Shalhoub et al. / Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
chronic liver disease and hepatocellular carcinoma, Eur. J. Cancer 34 (1998), 1198-1203. R. Saffroy, J.C. Lelong and D. Azoulay et al., Clinical significance of circulating anti-p53 antibodies in European patients with hepatocellular carcinoma, BrJ Cancer 79 (1999), 604-610. H. Imai, E.K. Chan, K. Kiyosawa, X.D. Fu and E.M. Tan, Novel nuclear autoantigen with splicing factor motifs identified with antibody from hepatocellular carcinoma, J Clin Invest 92 (1993), 2419-2426. Y. Muro, E.K. Chan, G. Landberg and E.M. Tan, A cellcycle nuclear autoantigen containing WD-40 motifs expressed mainly in S and G2 phase cells, Biochem Biophys Res Commun 207 (1995), 1029-1037. H. Imai, R.L. Ochs, K. Kiyosawa, S. Furuta, R.M. Nakamura and E.M. Tan, Nucleolar antigens and autoantibodies in hepatocellular carcinoma and other malignancies, Am J Pathol 140 (1992), 859-870. G. Covini, C.A. von Muhlen, S. Pacchetti, M. Colombo, E.K. Chan and E.M. Tan, Diversity of antinuclear antibody responses in hepatocellular carcinoma, J Hepatol 26 (1997), 1255-1265. H. Imai, K. Furuta, G. Landberg, K. Kiyosawa, L.F. Liu and E.M. Tan, Autoantibody to DNA topoisomerase II in primary liver cancer, Clin Cancer Res 1 (1995), 417-424. G. Covini, E.K. Chan, M. Nishioka, S.A. Morshed, S.I. Reed and E.M. Tan, Immune response to cyclin B1 in hepatocellular carcinoma, Hepatology 25 (1997), 75-80. J.Y. Zhang, E.K. Chan, X.X. Peng and E.M. Tan, A novel cytoplasmic protein with RNA-binding motifs is an autoantigen in human hepatocellular carcinoma, J Exp Med 189 (1999), 1101-1110. E. De plaen, C. Lurquin and A. van Pel et al., Immunogenic (tum-) variants of mouse tumor P815: cloning of the gene of turn- antigen P91A and identification of the tum- mutation, Proc Natl Acad Sci USA 85 (1988), 2274-2278. P. Van der Bruggen, C. Traversari and P. Chomez et al., A gene encoding an antigen recognized by cytolytic T lymphocytes on a human melanoma, Science 254 (1991), 1643-1647. O. Mandelboim, G. Berke, M. Fridkin, M. Feldman, M. Eisenstein and L. Eisenbach, CTL induction by a tumour-associated antigen octapeptide derived from a murine lung carcinoma, Nature 369 (1994), 67-71. A. Cox, J. Skipper and Y. Chen et al., Identification of a peptide recognized by five melanoma-specific human cytotoxic T cell lines, Science 264 (1994), 716-719. T. Boon and P. Van Der Bruggen, Human tumor antigens
[59] [60] [61]
[62]
[63]
[64] [65]
[66] [67]
[68]
[69]
[70]
[71]
[72]
223
recognized by T lymphocytes, J Exp Med 183 (1996), 725729. P.P. Robbins and Y. Kawakami, Human tumor antigens recognized by T cells, Curr Opin Immunol 8 (1996), 628-636. Y. Kawakami and S.A. Rosenberg, Human tumor antigens recognized by T-cells, Immunol Res 16 (1997), 313-339. S.A. Rosenberg, Cancer vaccines based on the identification of genes encoding cancer regression antigens, Immunology Today 18 (1997), 175-182. S.A. Rosenberg, A new era for cancer immunotherapy based on the genes that encode cancer antigens, Immunity 10 (1999), 281-287. U. Sahin, O. Tureci and H. Schmitt et al., Human neoplasms elicit multiple specific immune responses in the autologous host, Proc Natl Acad Sci USA 92 (1995), 11810-11813. L.J. Old and Y.T. Chen, New paths in human cancer serology, J Exp Med 187 (1998), 1163-1167. O. Tureci, U. Sahin, C. Zwick, F. Neumann and M. Pfreundschuh, Exploitation of the antibody repertoire of cancer patients for the identification of human tumor antigens, Hybridoma 18 (1999), 23-28. Y.T. Chen, Cancer vaccine: identification of human tumor antigens by SEREX, Cancer J Sci Am 6 (2000), S208-217. F. Stenner-Liewen, G. Luo and U. Sahin et al., Definition of tumor-associated antigens in hepatocellular carcinoma, Cancer Epidemiol Biomarkers Prev 9 (2000), 285-290. L. Prasannan, D.E. Misek, R. Hinderer, J. Michon, J.D. Geiger and S.M. Hanash, Identification of beta-tubulin isoforms as tumor antigens in neuroblastoma, Clin Cancer Res 6 (2000), 3949-3956. F. Brichory, D.E. Misek and A.M. Yim et al., An immune response manifested by common occurence of annexins I and II autoantibodies and high circulating levels of interleukin 6 in lung cancer, Proc Natl Acad Sci USA, in press. F. Brichory, D. Beer, F. Le Naour, T.J. Giordano and S.M. Hanash, Proteomics-based identification of PGP 9.5 as a tumor antigen that induces a humoral immune response in lung cancer, Cancer Research, in press. F. Le Naour, L. Hohenkirk and A. Grolleau et al., Profiling changes in gene expression during differentiation and maturation of monocytes-derived Dendritic cells using both oligonucleotide microarrays and proteomics, J Biol Chem 276 (2001), 17920-17931. P. Eggleton, F.J. Ward and S. Johnson et al., Fine specificity of autoantibodies to calreticulin: epitope mapping and characterization, Clin Exp Immunol 120 (2000), 384-391.
This page intentionally left blank
225
The application of protein microarrays to serum diagnostics: Prostate cancer as a test case Jeremy C. Millera, E. Brian Butlerb, Bin Sing Tehb and Brian B. Haaba,* a
The Van Andel Research Institute, 333 Bostwick NE, Grand Rapids, MI 49503, USA b Baylor College of Medicine, Houston, TX 77030, USA
1. Introduction Reliable and specific serum disease markers have great value as non-invasive, rapid and inexpensive assays. The discovery of new disease markers is particularly necessary for diseases that are difficult to detect or diagnose at an early and curable stage. For example, the early detection of pancreatic cancer and the differentiation of malignant from benign disease are extremely difficult using current imaging and cytological methods. Improved screening tools would permit the avoidance of unnecessary pancreaticoduodenectomies and allow the opportunity to perform the procedure at a curative stage [1]. The challenge to the discovery of new serum markers lies in the difficulties of highthroughput detection and quantitation of proteins. A new tool that is potentially well suited to meet this challenge is the protein microarray. The feasibility of accurate, sensitive and specific protein microarray detection of multiple proteins in a serum background was recently demonstrated [2], and efforts are now underway to apply this technology to marker discovery. The technology as described by Haab et al. [2] was built upon the existing DNA microarray platforms that are present in many labs (see http://cmgm.stanford.edu/pbrown/), making the method practical and easy to implement. * Corresponding author: Tel.: +1 616 234 5268; Fax: +1 616 234 5269; E-mail:
[email protected]. Disease Markers 17 (2001) 225-234 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
In addition, further availability of protein microarray technology is coming through the many commercial ventures that are actively working to get various types of protein chips to market. This article addresses the use of protein microarrays for serum marker detection and discovery, using prostate cancer as a model disease. We initially describe protein microarray technology and its suitability for serum analysis, then discuss the existing serum markers for prostate cancer and the potential advantages of using multiple markers, and finally describe serum protein studies using protein microarrays.
2. Protein microarray technology for highly-parallel serum protein detection The microarray format has many beneficial features for protein analysis, such as highly parallel detection, low sample consumption, and the potential for highly accurate and sensitive detection in multiple wavelength regions using scanning fluorescence microscopy, as recently demonstrated [2]. Certain aspects of the technology make it particularly well suited to the analysis and discovery of serum markers. For example, the ability to run many microarray experiments rapidly enables studies on the large populations of samples that are needed for good statistics on new markers. Additionally, the sophisticated software tools that are continually under development to analyze DNA microarray data also may be used to analyze protein microarray data. Many of these tools are specifically designed for the identification of genes or sets of genes that have diagnostic utility. It has been noted that new markers may be comprised of combinations of genes rather than individual genes [3]. Microarrays provide a highly effective tool to analyze the relationship between many genes to evaluate their combined value. Multiplexed protein detection using spotted antibodies and antigens has been demonstrated for a variety
226
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case
of applications with diverse technological implementations. Protein arrays on poly(vinylidene fluoride) (PVDF) and nitrocellulose membranes have been used to screen binding specificities of a protein expression library [4-6] and to detect DNA, RNA, and protein binding targets [7]. Phage displayed antibodies were arrayed onto filters for high-throughput screening of their specificities [8]. Derivatized glass slides have been used to attach microarrays of antibodies and antigens for high-throughput ELISA [9], to detect autoantibodies [10], and to detect protein-protein [2] and proteinsmall molecule interactions [11]. Since the technology is relatively new, most of the published reports focus on feasibility studies and technological characterization rather than biological studies. Efforts are underway to apply the technology to biological studies and to address the issues necessary to make the method more robust and practical. The primary experimental challenge in obtaining useful protein microarray data is the acquisition of high specificity and high affinity protein capture reagents. The specificity and affinity of the capture reagents define the sensitivity and accuracy of the assay. Having many high quality capture reagents adds to usefulness of protein microarray data, but several aspects of protein chemistry make the collection of a such a set difficult. Unlike nucleic acids, for which binding interactions are well characterized and predictable, protein binding interactions must be identified empirically. Since protein-protein interactions have a wide variety in binding strengths, stabilities and specificities, finding a suitable binding partner to a particular protein may be difficult in some cases. Additionally, proteins are expensive and time-consuming to produce and purify. Several approaches have been put forth to address the generation of protein capture reagents for arrays. High-throughput protein expression and purification methods have been developed, based on recombinant baculoviruses [12] or the GatewayTM recombinant cloning system [13]. The proteins are produced in 96well microtiter plates and efficiently purified through the amino- or carboxyl-terminal attachment of an epitope tag, such as poly-histidine or Glu-Glu. An efficient method to test for proper protein expression and folding is based on the arraying of individual bacterial colonies of a cDNA library onto membranes [4]. The arrayed colonies were induced for protein expression, the cells were lysed on the membrane, and the proteins were tested for proper expression, folding, and antibody specificity by antibody staining. Highthroughput, rapid and less expensive antibody produc-
tion for microarrays may be possible using phage display libraries [14]. Antibodies to specific antigens can be selected from a diverse library of antibodies displayed on the surface of phage clones, and after selection the selected clones can be amplified. The feasibility of multiplexed antigen detection using arrayed scFv phage display clones on membranes was recently shown [8]. There are other technological challenges in the development of practical protein microarrays. Because proteins have an almost unlimited variety in charges, polarities and structures, the efficient attachment of specific spotted proteins while repelling the adsorption of nonspecific background proteins can be difficult. Various protein attachment methods, surface blocking methods and new surfaces that are resistant to non-specific protein binding have been evaluated. A particularly effective method appears to be the attachment of biotinylated proteins through a streptavidin-biotin bridge on the end of poly(ethylene glycol) (PEG) polymer strands [15]. The PEG, which is attached to a poly-1-lysine coating on glass, efficiently repels non-specific background proteins, yet specific attachment of the capture proteins is achieved through the biotin-streptavidin junction. A simpler strategy for protein attachment is the adsorption of spotted proteins to poly-1-lysine coated glass, followed by the blocking of the surface with milk or BSA proteins [2]. No modification of the spotted protein is required with this approach, but non-specific binding to the poly-1-lysine may be higher than to PEG. The most widely used method to attach proteins to glass is the covalent reaction of protein amine groups to silane cross-linkers [9,11 ]. For reviews on various implementations of protein microarray technology and advantages and disadvantages for particular applications, see references [16-19].
3. Serum markers for the diagnosis of prostate cancer The best demonstration of the utility of serum markers for cancer diagnosis is the prostate cancer marker prostate specific antigen (PSA). PSA tests are used to screen men over age 50 in the general population and at younger ages in patients who have a familial history of prostate cancer or other risk factors. 67-80% of men with developing prostate cancers are identified [20], depending on patient age and mode of the PSA test, making the PSA test the most sensitive serum test available. PSA is a useful marker of recurrence in post-operative
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case
men who have received a radical prostatectemy, and the test is used to monitor the disease state in men who are undergoing chemical castration, or who are undergoing a "watchful waiting" treatment regime [20]. The main shortcoming of the PSA test is low specificity, leading to many otherwise unnecessary biopsies. Since PSA is an organ-specific rather than a cancerspecific marker, conditions such as simple hypertrophy, prostatitis, or other benign conditions produce positive PSA tests. In concentrations between 4 and 10 ng/ml, considered abnormal levels, PSA has only a 25% specificity for prostate cancer [21]. Above 10 ng/ml, PSA is more specific for prostate cancer, giving accurate diagnoses in about 67% of cases [21]. Another shortcoming of the PSA test is a lack of information about the stage and aggressiveness of the cancer, regardless of the PSA concentration. At present, doctors often have little indication whether a radical prostatectemy is necessary to prevent aggressive growth of the cancer and to prolong the patient's life, or whether surgery is unnecessary and would give no benefit to the patient. Due to the limitations of the PSA test, much research is being devoted to the discovery of an improved prostate cancer serum test. The proteins Keratinocyte growth factor (KGF) [22], Human glandular kallikrein 2 (hK2) [23,24], PSA complexed with alpha (2)-macroglobulin (PSA-A2M) [25] and Interleukin-8 (IL-8) [26] have been investigated as markers to improve the differentiation between benign prostatic hyperplasia (BPH) and prostate cancer. A recent study found an increase in specificity from 9% to 28% (at a 95% sensitivity) to differentiate benign prostatic hyperplasia from prostate cancer using hK2 combined with free and total PSA measurements [27]. Another study similarly found that measurements of hK2 along with free and total PSA improved the identification of prostate cancer in patients with low total PSA (2.54.5 ng/mL) [24]. Others serum proteins have been investigated for information on prognosis or stage of prostate cancer. The carboxy-terminal propeptide of type I procollagen (PICP), a biochemical marker of bone formation, was shown to be a significant marker for bone metastatis and poor prognosis [28]. In addition, serum levels of urokinase-type plasminogen activator [29], interleukin 6 [30] and neuron-specific enolase [31] were shown to have prognostic value for prostate cancer. The serum levels of testosterone [32] and hK2 [33] have been evaluated to assess the stage of prostate cancer. hK2, which seems to have a higher serum concentration in men with prostate cancer as compared to men with BPH (see
227
above), also appears to have a higher serum concentration in patients with non-prostate confined cancer as compared to prostate confined cancer [33]. These findings, taken together, show that many changes in addition to high PSA levels occur in the serum of prostate cancer patients. Although no single test is sensitive or specific for all situations, the combination of the markers could yield a test with greatly enhanced diagnostic utility. A summary of serum markers and their potential utility in diagnosing prostate cancer is provided in Table1. 4. The use of combined markers in diagnosis As noted above, the use of combinations of markers has great potential to improve diagnostic specificity and sensitivity over individual markers. Table 2 summarizes a number of examples of the uses of multiple markers to enhance the diagnosis or prognosis, discussed in more detail below. A study of prognostic markers for small cell lung cancer found that patients could be classified into four groups of 5-year survival rates based on the combined expression of cyclin E, Ki-67, and ras p21 [34]. Patients with no expression alterations in the three markers had a 96% survival rate, and patients with all three altered had a 41% survival rate. Although each marker individually provided some prognostic information, the three together significantly enhanced the accuracy of risk stratification. CA-125 has been used for many years as a serum marker for malignant pelvic masses. When taken alone, CA-125 at an abnormal level (> 35 U/ml) gives a fairly high sensitivity and specificity of 78.1% and 76.8%, respectively. However, a significant enhancement in diagnostic performance was shown using a panel of five markers (CA-125, OVX1, LASA, CA15-3, and CA724) [35]. When two of these markers were elevated, the sensitivity and specificity increased to 83.3% and 84.0% respectively. These markers were further enhanced using a regression analysis of the values of all five of the markers, giving a sensitivity of 90.6% and a specificity of 93.2%. The specificity of prostate cancer diagnosis seems to be improved by the use of multiple markers, using the measurements of both the free and bound forms of PSA and the measurement of hK2 (described above). When biopsies were not performed on patients with free/total PSA ratio of > 0.25, there was a 20% reduction in the number of unnecessary biopsies performed [36]. The sensitivity of the test for prostate cancer remained at levels similar to that of PSA alone.
228
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case Table 1 Examples of reported prostate cancer serum markers and their potential uses Comments Detection of prostate cancer or risk Prostate specific antigen (PSA) Insulin-like growth factor 1 (IGF-1) 25-hydroxyvitamin D (25-VD) Prostatic acid phosphatase (PAP) To improve specificity Human glandular kallikrein 2 (hK2) Keratinocyte growth factor (KGF) PSA complexed with alpha(2)-macroglobulin (PSA-A2M) Interleukin-8 (IL-8) Information on grade, stage and prognosis Carboxy-terminal propeptide of type I procollagen (PICP) Urokinase-type plasminogen activator (uPA) Interleukin 6 (IL-6) Neuron-specific enolase (NSE) Testosterone Human glandular kallikrein 2 (hK2) Serum luteinizing hormone (LH) Bone alkaline phosphatase (BAP) Matrix metalloproteinase-2 (MMP-2)
Reference
High sensitivity for prostate cancer, but not specific or prognostic Low serum levels correlated with prostate cancer Low serum 25-VD correlated with increased risk for prostate cancer The first marker for prostate cancer, supplanted by PSA Higher in prostate cancer than BPH, used with free and total PSA Lower in prostate cancer than BPH The percentate of PSA complexed with A2M was lower in prostate cancer than BPH Serum IL-8 was higher in prostate cancer than BPH: also may correlate with stage
[42] [43] [44] [45] [23.24,27] [22] [25] [26]
High PICP correlated with bone metastasis and poor outcome
[28]
High uPA correlated with disease progression and poor prognosis High IL-6 correlated with stage and poor prognosis Pretreatment-elevated NSE correlated with poor prognosis Low free serum testosterone correlated with extent and grade of disease Predicted non-organ confined versus organ confined cancer Low levels correlated with advanced disease Correlated with the extent of bone disease Higher in patients with metastasis
[29] [30] [31 ] [32] [33] [46] [47] [48]
Table 2 Examples of investigations of the use of combinations of markers to enhance the diagnostic/prognostic characteristics of single markers Comments Improve sensitivity/specificity of diagnosis Colorectal Carcinoma: Creatine kinase-BB, Homoarginine-sensitive alkaline phosphatase. Salivary-type amylase. Macro-creatine kinase type 2. Ferritin, Alpha 1-acid glycoprotein, Creactive protein. Alpha 1-antitrypsin. Ceruloplasmin. CEA, and Beta human choriogonadotropin Ovarian Cancer: CA-125. OVX1. LASA. CA15-3. and CA72-4
Pancreatic Cancer: CA 19-9. CEA
Improve prognosis: Non-Small Cell Lung Cancer: Cyclin E, Ki-67. and Ras p21
Reference
Improved sensitivity of diagnosis of colorectal carcinoma by 17% and 64% for early- and late-stage disease respectively.
[49]
Improved sensitivity and specificity over CA-125 for the identification of malignant pelvic masses from 78.1% and 76.8%. respectively, to 90.6% and 93.2%. respectively.
[35]
Combined serum levels of CA 19-9 and CEA provided 85.8% sensitivity and 91.2% specificity. This was compared with 30% and 83% sensitivity for the individual markers respectively.
[50]
Segregated patients based on five-year survival patients with 0 altered - 96% survival, all three altered 41% survival.
[34]
5. Protein microarrays for highly multiplexed protein detection A recent study showed the feasibility of sensitive and accurate protein microarray detection of multiple specific antibodies and antigens in a serum background [2]. A robotic device (identical to that used to spot cDNA arrays [37]) was used to print hundreds
of specific antibody or antigen solutions in an array on the surface of derivatized microscope slides. Two complex protein samples, one serving as a standard for comparative quantitation, and the other representing an experimental sample in which the protein quantities were to be measured, were labeled by covalent attachment of spectrally-resolvable fluorescent dyes. Specific antibody-antigen interactions localized specific com-
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case
§
p38 delta
Numb
AIM-1
ABR
ARNT1
TEF-1
FIN13
MST3
G3BP
VLA-3alpha
Kanadaptin
229
-1 -2
100
10
1
0.1
100
10
1
0.1
100
10
Antibody concentration (ng/mL) Fig.1. Relationship between the Cy5/Cy3 fluorescence ratios measured using antigen microarrays and the concentration ratio of the cognate antibodies. The solid line represents the median of the log10 transformed Cy5/Cy3 fluorescence ratios from 6-12 replicate antigen spots, with the error bars representing the standard deviation between the replicate spots. The dashed line represents the ideal linear relationship between R/G ratio and concentration ratio. (From BB Haab, MJ Dunham, PO Brown. Genome Biology 2:1-13, 2001.)
ponents of the complex mixtures to defined cognate spots in the array, where the relative intensity of the fluorescent signal representing the experimental sample and the reference standard provided a measure of each protein's abundance in the experimental sample. The specificity, sensitivity and accuracy of the assay were evaluated using 115 antibody/antigen pairs. Six different mixtures of the 115 antibodies and six different mixtures of 115 antigens were prepared so that the concentration of each species varied in a unique pat-
tern across the protein mixtures over a range of three orders of magnitude. Each of the six protein mixtures was labeled with the dye Cy5 (red fluorescence) and mixed with a Cy3-labeled (green fluorescence) "reference" mixture containing each of the same 115 proteins at a constant concentration. The variation across the six microarrays in the red-to-green (R/G) ratio measured for each antibody or antigen spot should reflect the variation in the concentration of the corresponding binding partner in the set of mixes. By comparing the
230
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case
Fig. 2. Protein microarray analysis of human prostate cancer serum. (a) Proteins from a prostate cancer serum sample were labeled red (Cy5) and incubated on the array with a green-labeled (Cy3) pooled reference sample. (b) The labeling of the two samples from panel (a) was reversed, and the samples were incubated on an identical microarray. Colored rectangles in both images indicate areas expanded below, showing examples of spots that alternately appear red or green in the two arrays. The yellow square contains one of the spotted PSA antibodies (labeled in figure).
observed variation in the concentration ratios with the known variation in the concentration ratios, the performance of each antibody/antigen pair could be assayed. Figure 1 presents this relationship for 12 different arrayed antigens detecting their respective cognate antibodies in complex solutions. The dashed line represents the ideal linear response in R/G ratio with respect to analyte concentration, and the solid line is the median log10(R/G) ratio of 6-9 replicate spots, with the error bars representing the standard deviation in the log10(R/G) ratio. For many of the antigens, the experimental data very closely followed the ideal response (represented by the dashed line). For antigens such as P38 delta. Numb, and AIM-1, the measurements were reproducible and accurate over the entire three orders of magnitude con-
centration range. These antigens have detection limits of less than 1 ng/mL for their respective antibodies. The ratios measured at replicate spots were highly consistent and exhibited low standard deviations, except in some cases at low concentrations where the dispersion appeared more random (e.g. G3BP and ARNT1). These data demonstrated accurate and specific quantitation of protein ligands in a complex, physiologically relevant background. The detection limit of the assay depends on the level of background protein binding and on the affinity and specificity of the antibody/antigen interaction. Some of the antibody/antigen pairs allowed detection of the cognate ligands at absolute concentrations below 1 ng/ml, sensitivities sufficient for measurement of many clinically important proteins in patient blood samples. Since many potentially inter-
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case
231
4
-1 8 -2
-3
-3
-2
-1
0 1 2 Log R/G, Experiment (a)
3
4
Fig. 3. Investigation of experimental reproducibility and the accuracy in reverse labeling of protein microarray experiments. The R/G ratio of each antibody spot in experiment (a) of Fig. 2 was plotted with respect to the R/G ratio of its counterpart spot in experiment (b) of Fig. 2. The diagonal line (not fitted to the data) indicates the ideal inverse relationship between the "color swapped" experiments.
esting proteins have serum concentrations below that level, it will be important to further improve the sensitivity of the protein microarray assay. Approaches to increase detection sensitivity include using surfaces that are more resistant to background protein binding, selecting antibodies that have optimized affinities, and amplifying the fluorescent signal of the bound antibody. Showing particular promise for improving the sensitivity of protein microarray detection is rolling circle amplification (RCA), which was demonstrated to lower detection limits in solid phase immunoassays by over 100-fold [38]. Application of RCA to protein microarrays, along with the use of optimized antibodies and surfaces, should allow the quantification of many low abundance and potentially important serum proteins.
6. Highly parallel serum protein detection using protein microarrays Preliminary studies are under way in our laboratory to apply protein microarray technology to prostate cancer serum marker discovery. Prostate cancer serum samples with characterized PSA concentrations provide a positive control, and other serum proteins that
have been studied in connection with prostate cancer give many good leads for testing the utility of highly multiplexed marker detection. Over 200 antibodies to putative serum markers, known serum proteins and known cancer genes were collected through collaborations and printed in arrays on poly-1-lysine derivatized microscope slides. As a reference protein solution, 34 prostate cancer serum samples and 20 serum samples from healthy patients were pooled together. The individual serum samples were then fluorescently labeled and co-incubated on the microarrays with the differentially labeled reference. Initial experiments have confirmed multiplexed specific detection of proteins in the serum samples. Figure 2 presents two microarrays in which the fluorescent labeling of a prostate cancer sample and the reference pool were swapped. In panel (a), the prostate cancer serum and the reference pool were labeled with Cy5 (red fluorescence) and Cy3 (green fluorescence), respectively, and in panel (b) the labeling is reversed. Many antibody spots have fluorescence clearly above background in red, yellow and green colors. The expanded regions of the images in the lower panels of Fig. 2 show that many of the antibody spots that are primarily red in one array are primarily green in the
J,C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case
other, consistent with reproducible and specific labeling. Significantly, the PSA antibody appears more red when the prostate cancer serum was labeled red, and more green when the same sample was labeled green. The analysis of additional prostate cancer serum samples is needed to confirm quantitative detection of this protein. To quantitatively compare the color swapped experiments, the R/G ratios of matched spots on the two arrays were plotted with respect to each other (Fig. 3). The solid diagonal line represents the ideal inverse relationship. The general trend of the R/G ratios follows the solid line, showing that for most of these spots, the labeling and detection was consistent and reproducible. The scatter around the solid line reflects the level of noise inherent in the measurement, which can be evaluated to determine thresholds for accepting or rejecting spots for further analysis. Some of the spots fell well outside the scatter on the inverse diagonal, such as those in the lower left of the figure, and would be rejected from further analysis. These data demonstrate high signal-to-noise and reproducible detection of multiple proteins in human cancer serum. After a large set of prostate cancer sera samples have been analyzed, the aim of the analysis will be to identify patterns of proteins that significantly correlate with particular clinical parameters. Methods developed for the analysis of RNA expression profiles from cDNA microarrays will be applied to the data. For example, individual proteins that distinguish two sample sets (such as BPH versus prostate cancer) could be identified using a permutation t-test f39], and patterns of proteins that make the same distinction could be identified using methods such as 'tree harvesting' [40] or a 'cluster identification tool' [41]. With the right antibodies on the arrays, previous demonstrations of the value of multiple markers and power of microarrays indicate that significant advances in serum marker discovery and validation should be achievable with this new tool.
Acknowledgements J.C.M. is funded through a Van Andel Institute fellowship. We thank the Van Andel Institute for funding and the staff of the Van Andel Institute for helpful interactions.
References [ 1 ] A. Riker, S.K. Libutti and D.L. Bartiett, Advances in the early detection, diagnosis, and staging of pancreatic cancer. Surgical Oncology 6 (1998), 157-169. [2] B.B. Haab, M.J. Dunham and P.O. Brown. Protein microarrays for highly parallel detection and quantitation of specific proteins and antibodies in complex solutions. Genome Biologv 2(2001), 1-13. f31 CM. Perou. T. Sorlie. M.B. Eisen. M. van de Rijn, S.S. Jeffrey. C.A. Rees. J.R. Pollack, D.T. Ross, H. Johnsen, L.A. Akslen, O. Fluge, A. Pergamenschikov, C. Williams, S.X. Zhu, RE. Lonning, A.L. Borresen-Dale, P.O. Brown and D. Botstein. Molecular portraits of human breast tumours. Nature 406 (2000). 747-752. [4] K. Buessow, D. Cahill, W. Nietfeld, D. Bancroft, E. Scherzinger, H. Lehrach and G. Walter, A method for global protein expression and antibody screening on high-density filters of an arrayed cDNA library. Nucleic Acids Research 26 (1998), 5007-5008. [5] A. Lueking, M. Horn, H. Eickhoff, K. Buessow. H. Lehrach and G. Walter, Protein microarrays for gene expression and antibody screening, Anal. Biochem. 270 (1999), 103-111. [6] K. Buessow. E. Nordhoff, C. Luebbert. H. Lehrach and G. Walter, A human cDNA library for high-throughput protein expression screening, Gtnomics 65 (2000), 1-8. [7] H. Ge, UPA, a universal protein array system for quantitative detection of protein-protein, protein-DNA, protein-RNA and protein-ligand interactions. Nucleic Acids Research 28 (2000). 3, [8] R.M.T. deWildt, C.R. Mundy. B.D. Gorick and T.M. Tomlinson. Antibody arrays for high-throughput screening of antibody-antigen interactions. Nature Biotechnology 18 (2000), 989-994. (9] L.G. Mendoza, P. McQuary. A. Mongan, R. Gangadharan. S. Brignac and M. Eggers, High-throughput microarray-based enzyme-linked immunosorbent assay (ELISA), Biotechniques 27(1999), 778-788. [10] T.O. Joos, M. Schrenk, P. Hopfl, K. Kroger, U.D.S. Chowdhury, D. Schorner. M. Durr, K.R.S. Herick. K. Sohn and H. Hammerle. A microarray enzyme-linked immunosorbent assay for autoimmune diagnostics, Electrophoresis 21 (2000), 2641-2650. [11] G. MacBeath and S.L. Schreiber. Printing proteins as microarrays for high-throughput function determination. Science 289 (2000), 1760-1763. [12] J.S. Albala, K. Franke. I.R. McConnell, K.L. Pak, PA. Folia, B. Rubinfeld, A.H. Davies, G.G. Lennon and R. Clark, From genes to proteins: high-throughput expression and purification of the human proteome, J Cell Biochem. 80 (2000). 187-191. [13] A.J. Walhout, G.F. Temple. M.A. Brasch. J.L. Hartley. M.A. Lorson. S. van den Heuvel and M. Vidal. GATEWAY recomhinational cloning: application to the cloning of large numbers of open reading frames or ORFeomes. Methods Enzymol. 328 (2000), 575-592. [141 G. Winter, A.D. Griffiths. R.E. Hawkins and H.R. Hoogenboom. Making antibodies by phage display technology, Annu. Rev. fmmunol. 12 (1994), 433-455. [151 L.A. Ruiz-Taylor, T.L. Martin. F.G. Zaugg, K. Wine, P. Indermuhle, S. Nock and P. Wagner. Monolayers of derivatized poly(L-lysine)-grafted poly(ethylene glycol) on metal oxides as a class of biomolecular interfaces. Proc. Natl. Acad. Sci. USA 98(2001). 852-857.
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case [16]
[17]
[18]
[19] [20] [21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
A.Q. Emili and G. Cagney, Large-scale functional analysis using peptide or protein arrays, Nat. Biotech. 18 (2000), 393397. G. Walter, K. Buessow, D. Cahill, A. Lueking and H. Lehrach, Protein arrays for gene expression and molecular interaction screening, Curr. Opin. Microbio. 3 (2000), 298-302. B.B. Haab, Advances in protein microarray technology for protein expression and interaction profiling, Curr. Opin. Drug Disc. Dev. 4 (2001), 116-123. D. Cahill, Protein and antibody arrays and their medical applications, J. Immunol. Methods 250 (2001), 81-91. Prostate-specific antigen (PSA) best practice policy, Oncology 14 (1998), 267-286. M. Lukes, M. Urban, M. Zalesky, R. Zachoval, J. Heracek and E. Zdarsky, Prostate-specific antigen: current status, Gynecol. Oncol. 47 (2001), 41-49. P.B. Mehta, C.N. Robson, D.E. Neal and H.Y. Leung, Serum keratinocyte growth factor measurement in patients with prostate cancer, The Journal of Urology 164 (2000), 21512155. M.K. Kwiatkowski, F. Recker, T. Piironen, K. Pettersson, T. Otto, M. Wernli and R. Tscholl, Inprostatism patients the ratio of human glandular kallikrein to free PSA improves the discrimination between prostate cancer and benign hyperplasia within the diagnostic "gray zone" of total PSA 4 to 10 ng/mL, Urology 52 (1998), 360-265. A. Magklara, A. Scorilas, W.J. Catalona and E.P. Diamandis, The combination of human glandular kallikrein and free prostate-specific antigen (PSA) enhances discrimination between prostate cancer and benign prostatic hyperplasia in patients with moderately increased total PSA, Clinical Chemistry 45 (1999), 1960-1966. W.-M. Zhang, P. Finne, J. Leinonen, J. Salo and U.-H. Stenman, Determination of prostate-specific antigen complexed to alpha2-macroglobulin in serum increases the specificity of free to total PSA for prostate cancer, Adult Urology 56 (2000), 267-272. R.W. Veltri, M.C. Miller, G. Zhao, A. Ng, G.M. Marley, G.L.J. Wright, R.L. Vessella and D. Ralph, Interleukin-8 serum levels in patients with benign prostatic hyperplasia and prostate cancer, Urology 53 (1999), 139-147. C. Becker, T. Piironen, K. Pettersson, T. Bjork, K.J. Wojno, I.E. Oesterling and H. Lilja, Discrimination of men with prostate cancer from those with benign disease by measurements of human glandular kallikrein 2 (hK2) in serum, The Journal of Urology 163 (2000), 311-316. S. Akimoto, H. Inomiya, Y. Furuya, K. Akakura and H. Ito, Prognostic value of the serum levels of bone formation and bone resorption markers in prostate cancer patients with bone metastasis, European Urology 34 (1998), 142-147. H. Miyake, I. Hara, K. Yamanaka, K. Gohji, S. Arakawa and S. Kamidono, Elevation of serum levels of urokinasetype plasminogen activator and its receptor is associated with disease progression and prognosis in patients with prostate cancer, The Prostate 39 (1999), 123-129. J. Nakashima, M. Tachibana, Y. Horiguchi, M. Oya, T. Ohigashi, H. Asakura and M. Murai, Serum interleukin 6 as a prognostic factor in patients with prostate cancer, Clinical Cancer Research 6 (2000), 2702-2706. W Lilleby, E. Paus, E. Skovlund and S.D. Fossa, Prognostic value of neuroendocrine serum markers and PSA in irradiated patients with pN0 localized prostate cancer, The Prostate 46 (2001), 126-133.
[32]
[33]
[34]
[35]
[36]
[37] [38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
233
M.A. Hoffman, W.C. DeWolf and A. Morgentaler, Is low serum free testosterone a marker for high grade prostate cancer? The Journal of Urology 163 (2000), 824-827. A. Haese, C. Becker, J. Noldus, M. Graefen, E. Huland, H. Huland and H. Lilja, Human glandular kallikrein 2: a potential serum marker for predicting the organ confined versus nonorgan confined growth of prostate cancer, The Journal of Urology 163 (2000), 1491-1497. H. Dosaka-Akita, F. Hommura, T. Mishina, S. Ogura, M. Shimizu, H. Katoh and Y. Kawakami, A Risk-Stratification Model of Non-Small Cell Lung Cancers Using Cyclin E, Ki67, and ras p21, Cancer Research 61 (2001), 2500-2504. R.P. Woolas, M.R. Conaway, F. Xu, I.J. Jacobs, Y. Yu, L. Daly, A.P. Davies, K. O'Briant, A. Berchuck, J.T. Soper, D.L. Clarke-Pearson, G. Rodriguez, D.H. Oram and R.C.J. Bast, Combinations of multiple serum markers are superior to individual assays for discriminating malignant from benign pelvic masses, Gynecologic Oncology 59 (1995), 111-116. W.J.P.A. Catalona, K.M. Slawin, M.K. Brawer, R.C. Flanigan, A. Patel, J.P. Richie, J.B. deKernion, P.C. Walsh, P.T. Scardino, PH. Lange, E.N. Subong, R.E. Parson, G.H. Gasior, K.G. Loveland and P.C. Southwick, Use of the percentage of free prostate-specific antigen to enhance differentiation of prostate cancer from benign prostatic disease: a prospective multicenter clinical trial, JAMA 279 (1998), 1542-1547. M.B. Eisen and P.O. Brown, DNA arrays for analysis of gene expression, Methods Enzymol 303 (1999), 179-205. B. Schweitzer, S. Wiltshire, J. Lambert, S. O'Malley, K. Kukanskis, Z. Zhu, S.F. Kingsmore, P.M. Lizardi and D.C. Ward, Imunoassays with rolling circle DNA amplification: a versatile platform for ultrasensitive antigen detection, Proc. Natl. Acad. Sci. USA 97 (2000), 10113-10119. I. Hedenfalk, D. Duggan, Y. Chen, M. Radmacher, M. Bittner, R. Simon, P. Meltzer, B. Gusterson, M. Esteller, O.P. Kallioniemi, B. Wilfond, A. Borg and J. Trent, Gene-expression profiles in hereditary breast cancer, N. Engl. J. Med. 344 (2001), 539-548. R. Hastie, R. Tibshirani, D. Botstein and P.O. Brown, Supervised harvesting of expression trees, Genome Biology 2 (2001), research0003.1-research0003.12. M. Takahashi, D. Rhodes, K.A. Purge, H. Kanayama, S. Kagawa, B.B. Haab and B.T. Teh, Gene expression profiling of clear cell renal cell carcinoma: gene identification and prognostic classification, Proc. Natl. Acad. Sci. USA 2001, in press. T.A. Stamey, N. Yang, A.R. Hay, I.E. McNeal, F.S. Freiha and E. Redwine, Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate, N. Engl. J. Med. 317 (1987), 909-916. R. Baffa, K. Reiss, E.A. El-Gabry, J. Sedor, M.L. Moy, D. Shupp-Byrne, S.E. Strup, W.W Hauck, R. Baserga and L.G. Gomella, Low serum insulin-like growth factor 1 (IGF-1): a significant association with prostate cancer, Tech. Urol. 6 (2000), 236-239. M.H. Ahonen, L. Tenkanen, L. Teppo, M. Hakama and P. Tuohimaa, Prostate cancer risk and prediagnostic serum 25hydroxyvitamin D levels, Cancer Causes Control 11 (2000), 847-852. R. Mazzucchelli, P. Colanzi, R. Pomante, G. Muzzonigro and R. Montironi, Prostate tissue and serum markers, Adv. Clin. Path. 4 (2000), 111-120. H. Hilz, M. Graefen, J. Noldus, P. Hammeler, C. Knabbe, E. Huland and H. Huland, Advanced prostate cancer is associated with a decrease in serum luteinizing hormone, Eur. Urol. 38 (2000), 243-249.
234
J.C. Miller et al. / The application of protein microarrays to serum diagnostics: Prostate cancer as a test case
[47] J.A. Lorente, H. Valenzuela, J. Morote and A. Gelabert, Serum bone alkaline phosphatase levels enhance the clinical utility of prostate specific antigen in the staging of newly diagnosed prostate cancer patients, Eur. J. Nucl. Med. 26 (1999), 625632. [48] K. Gohji, N. Fujimoto, I. Hara, A. Fujii, A. Gotoh, H. Okada, S. Arakawa, S. Kitazawa, H. Miyake, S. Kamidono and M. Nakajima, Serum matrix metalloproteinase-2 and its density in men with prostate cancer as a new predictor of disease
extension, Int. J. Cancer 79 (1998), 96-101. D.W.T.T. Mercer, Multiple markers of malignancy in sera of patients with colorectal carcinoma: preliminary clinical studies, Clin Chem 31 (1985), 1824-1828. [50] W.H.E.W. Schmiegel, C. Kreiker, H. Kalthoff, G.H. Butzow, K. Jessen, R. Klapdor, N. Soehendra, M. Wargenau and M. Classen, Multiparametric tumor marker (CA 19-9, CEA, AFP, POA) analyses of pancreatic juices and sera in pancreatic diseases, Hepatogastroenterology 32(1985). 141-145. [49]
235
Protein expression analysis: From 'tip of the iceberg' to a global method Peter James Wallenberg Laboratory II, Lund University, P.O. Box 7031, SE-220 07 Lund, Sweden Tel.: +41765585802; E-mail: peter.james @ elmat.lth.se In this review I will describe the advances that have recently been made in 'traditional' two-dimensional gel based protein expression analysis. A major jump has been made toward the automation of gel image analysis and comparison, one of the major bottlenecks in the analysis chain as well as the automation of spot excision and preparation for mass spectrometric analysis. Currently the gel-based 'proteome mapping' approach is highly effective and 300 gels and over 10,000 spots a week can be analysed. Very recently, viable alternatives to the use of two-dimensional gel electrophoresis have emerged and these approaches are discussed here. In combination with the recently developed stable isotopic tagging methods for peptide quantitation and new mass spectrometers, this emerging technology will be a rapid and highly effective alternative to gel-based methods with few of the latter's shortcomings.
1. Introduction 1.1. The starting impulse, whole genome availability Recently biological vernacular has been expanded with a series of 'omes': The genome; the DNA sequence of an organism, the transcriptome; the mRNA being expressed at a given time in a cell and the proteome; the protein equivalent. The latest in the family has been dubbed the metabolome and is a catchall term for all small molecules that are a product of enzymatic and chemical activity within the cell. In contrast to the genome, which is fairly inert, the latter three molecular groups are highly dynamic and vary greatly according to the endo- and exogenous conditions and throughout the life cycle of an organism. The human genome for example consists of the forty-six chromosomes, which encode somewhere between 30 and 100,000 genes. These can either directly transcribed Disease Markers 17 (2001) 235-246 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
1:1 or can be recombined into various different combinations by gene rearrangements as with T-cell receptors and immunoglobulins. The nucleotide sequence of the human genome (at the first draft) is now available in databases, yet only a small fraction of the genes found have a known role. The mouse genome is roughly the same size as that of the human, about 3.1 billion base pairs and as of March 2001, the coverage of the public area mouse sequencing effort was around 95% with x3 coverage. This will greatly facilitate the interpretation of the human sequence since only about 5% of the human genome contain genes and the gene sequences in mouse and human that encode the same proteins show a high degree (85%) of sequence identity. The DNA sequences in the vast regions between genes are much less similar (50% sequence identity or less). The availability of complete genome sequences and extensive Expressed Sequence Tag (EST [1]) libraries potentially allows the entire potential protein complement of organisms to be defined. Interest is now focussed on trying to interpret the massive influx of new sequence data and to understand how the vast array of chemical species in the cell interact with one another to create the molecular machinery of the cell. The focus of biological problem solving must now move from a reductionist to a global approach and methodologies must be developed to allow genome wide monitoring of gene expression at the mRNA, protein and metabolite (protein activity) levels. 1.2. The development of genome-wide expression studies The dynamic expression of genes as mRNA, the transcriptome can be followed both in a quantitative and qualitative manner. This has been made possible by the development of a variety of mRNA expression analysis methods that allow genome wide studies [2,3] to be carried out. A prerequisite to these large scale mRNA expression studies and to the sequencing of genomic DNA is a high degree of automation. The key to the development of these large scale mRNA expression stud-
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
ies were technological advances in analytical biochemistry such as the development of polymerase chain reaction amplification (PCR), shotgun sequencing, and fluorescent-tagged DNA capillary sequencing [1,4,5]. In order to obtain the reproducibility and data accuracy at the high-throughput levels necessary for the assembly of these complex data sets, the process had to be automated by robotics and new algorithms for data assembly and sequence evaluation had to be developed. The combination has made the genome sequencing not only possible, but also almost routine. A 128 capillary array DNA sequencer can produce 128,000 bases per run, giving an output of 3 million bases, -equivalent to a bacterial genome per day. Reproducibility is essential for a statistical analysis of these complex data sets and automation and high throughput data accumulation are essential for such large undertakings. DNA and mRNA are physico-chemically very homogenous and 'easy' to handle, can be amplified by polymerase chain reaction methods and are hence amenable to automation. mRNA analysis methods such as DNA hybridisation arrays and Serial Amplification of Gene Expression have made the quantitative and qualitative analysis of mRNA into an extremely high-throughput technique. 1.3. The old 'newcomer' proteomics The same cannot be said for methods for global analysis of the protein complement of the cell, which is in its infancy. Since proteins are vastly more physicochemically diverse than nucleic acids, a universal separation method is unlikely to be found and this is further compounded by the lack of an amplifying method analogous to PCR. The only partially satisfactory methods for analysing the state of expression of the majority of proteins in a cell are those based on twodimensional polyacrylamide gel electrophoresis (2DPAGE [6]). Proteins are separated in the first dimension according to their isoelectric point, i.e. by migrating to a point in the gel where the pH causes the net charge on the protein to become neutral. In the second dimension they are separated according to their mobility in a porous gel, which is proportional to the amount of detergent, Sodium Dodecyl Sulphate bound which is approximately mass dependent. However 2D-gel technology suffers many drawbacks. The separation on a single gel can show up to 10,000 species, however many of these are due to post-translational modifications and the number of gene products being visualised is probably only of the order of 1,000. The increase in reproducibility that has been brought about by the
introduction of commercial immobilised pH gradient first dimension gels (IPG [7]) allows very accurate and quantitative comparative 2D gel mapping. Detailed 'proteome maps' can be created with advanced computer imaging programs and then analysed by subtractive or cluster methods to find relationships between the protein spots. The weakness of 2D-PAGE lies in it's inability to deal with certain classes of proteins, mostly highly hydrophobic ones (membrane and cytoskeletal especially) and those with isoelectric points at either extreme of the pH scale (such as acidic hyperphosphorylated and alkaline DNA binding proteins). There are also problems with quantitation due to the low dynamic range of stains. These and other problems must be solved before proteomics can truly become a global approach. I will address the advances being made in the 2D field and then compare these with the new non-gel based methods under development. 7.4. What is proteomics ? Before going in to detail as to how protein expression analysis can be automated, one should define the term Proteomics. Originally it was defined as the protein complement of the genome, however since the whole genome is never expressed in a cell, a more restricted definition must be used. The proteome is the set of gene products and their covalent modifications that occur within a given type of cell at a specific stage and time in its development. Proteome analysis can be subdivided into expression proteomics which analyses protein expression and modification and cell-map proteomics which attempts to define all protein-protein interactions occurring in a cell under given conditions [8]. Expression proteomics relies heavily on quantitative 2D-PAGE to map protein expression in defined cells and is used to follow how protein expression changes in response to perturbation, be it genetic modification or environmental. Cell-map proteomics can be carried out either by high-throughput genetic screening using twohybrid systems [91 or by isolation and characterization of protein complexes. These new methods for proteome and gene expression analysis are quantitative and will allow new systematic approaches to investigation the function and regulation of unknown genes. N.L. Anderson has defined three major areas for the analysis of gene function and regulation: molecular anatomy (protein composition of cells and tissues); molecular pathology (analysis of disease in terms of changes in protein expression and modification); and molecular pharmacol-
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
ogy/toxicology (the effects of drugs and xenobiotics on protein expression and modification). A fourth area, molecular physiology, can be added, the change in protein expression in response to changes in the cells micro- or macro-environment.
2. Why bother with proteomics? In order to fully understand the workings of such a complex system as a cell, global analyses (both spatial and temporal) of transcription, translation, posttranslational modifications and metabolites must be carried out. There is an obvious need to complement the well-established genome-wide mRNA expression methods with global analyses of protein expression and post-translational modification [10,11]. There have been very few comprehensive analyses of the correlation between mRNA profiles and protein expression in any biological system [12-14]. The initial evaluations seem to indicate that there is only a significant correlation between mRNA and protein levels for half of the genes being expressed. The reason(s) for this discrepancy is entirely unknown at the moment. There are several key objections to the reduction of biological studies to following changes in mRNA: (i) the level of mRNA does not allow one to predict the level of protein expression, (ii) protein function is controlled by many post-translational modifications, and (iii) protein maturation and degradation are very dynamic processes which dramatically alter the final amount of active protein independent of mRNA level. A large-scale protein expression study would be an invaluable aid to understanding this phenomenon as well as for identifying markers missed by mRNA studies. These studies can also indicate defects in cell signalling mechanisms by showing the changes occurring, for example, in phosphorylation patterns. This would be an important tool in understanding the mechanism underlying the development and progression of a disease. Finally, in a similar vein to analysing to the relationship of mRNA to protein, the level of metabolites in a tissue is only partially related to the protein expression profile [15]. The analysis of metabolite profiles may provide a very useful tool for diagnostics and prognostics [16]. The success of the genomes projects when measured by the sheer amount of sequence data that has been generated is immense. However the number of genes for which a function can be assigned is rather meagre and hence the discipline functional genomics was created to describe the analysis of gene expression and func-
237
tion. The genome of the yeast Saccharomyces cerevisiae contains at least 6,200 genes. Despite intensive genetic work over the past years, 60% of yeast genes have no assigned function and half of those encode putative proteins without any homology with known proteins [17]. In order to describe the functions of the yeast genes, a systematic large-scale approach is being taken using a combination of mutant generation and analysis by transcriptomics, proteomics and metabolomics [18].
3. Advances in the automation of 'traditional' two-dimensional gel electrophoresis based analysis 3.1. Prefractionation and gel-running The extremely high degree of complexity of eukaryotic tissues often requires that a pre-fractionation step be carried out in order to reduce the complexity and allow the resolution and analysis of minor components. When dealing with a mixed cell population such as a tissue, pre-fractionation of cells using a fluorescence activated cell sorter [19] can allow small sub-populations to be specifically isolated, greatly increasing the sensitivity of the analysis. Similarly, preconcentration of the proteins to be analysed can be carried out using methods orthogonal to 2D gel separation such as native PAGE [20] or by an affinity preenrichment such as heparin chromatography for DNA binding proteins [21], immobilised metal ion affinity chromatography for phosphoproteins [22] or by antibody precipitation to select for a specific protein complex. Alternatively a series of increasingly powerful solubilising buffers may be used to obtain a series of protein fractions [23] or the various cell compartments and/or organelles may be isolated [24]. All of the above methods can be automated and before any large-scale study is started, a systematic study of sample preparation reproducibility should be carried out. An alternative to prefractionation, if sufficient material is available, is the use of a series of overlapping narrow pH range first dimension strips. This allows greatly increased loading amounts and greater separation efficiency. These are called zoom gels and can be used to determine the pI of a protein to within 0.001 pH units on a narrow pH range gel covering say 0.5 pH units over a 20 cm separation range. A non-orthogonal approach which can be combined with zoom gels is the prefractionation of large amounts of cell extract into defined pH ranges using isoelectric membranes mounted
238
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
to form a series of pH defined chambers [25]. This is commercially available under the name IsoPrime and was intended for the purification of individual proteins but has proven a useful first step for zoom gels which can save large amounts of material which would otherwise be lost off the ends of the gels as well as improving the loading and separation of the zoom gels. Up-to-now, no fully automated commercial method is available for running multiple 2D gels. A detailed study using a prototype semi-automated instrument clearly showed the advantages of mechanical reproducibility [26]. Until recently the only developments in large scale automated production of 2D gels has been carried out in commercial enterprises such as Large Scale Biology, Proteome Systems and Oxford Glycosciences and given the developments in non-gel based systems, such systems are unlikely to be developed. 3.2. Automated gel matching and analysis Recently a commercial version of the fluorescencebased Differential Gel Electrophoresis (DIGE) technique described by Unlu et al. [27] has become available. The major advance in the technology offered by DIGE is that the entire gel-image analysis procedure can be fully automated. There is no longer any need for time-consuming manual gel matching or editing. The basis of the technique is the covalent fluorescence labelling of the samples with cyanine dyes prior to electrophoresis and the generation of a universal master gel (see Fig. 1). For example, say an experiment involves the analysis of 10 normal breast tissue and 40 tumour samples. A master pool sample is created by mixing half of the protein extracts from all of the samples together and labelling the mixture with dye 1 (red). Sample 1 is labelled with dye 2 (blue) and sample 2 with dye 3 (green) and after labelling both are mixed with the master sample and loaded on the gel. By using a fluorescence detector, a gel image can be obtained from each sample according to the marker dye and thus intra-gel matching of the three samples is trivial since identical spots occur always at the same place since the dyes have almost identical masses and the same charge. Since each gel contains a master image which contains all possible protein spots found in all samples, intergel matching becomes trivial and can carried out automatically without user intervention. This technique coupled with the use of zoom gels will greatly extend and speed up the traditional approach to gel-based proteomics. A single person can easily run and analyse up to 250 samples a week using a four-dye system.
3.3. Spot-cutting and preparation for analysis Once the data system has matched, quantitated and analysed a gel series, a spot cut-list can be generated based on the criteria fed into the statistical analysis program by the user. Currently there are three commercially available spot-cutting systems available (Amersham-Pharmacia Biotech, BioRad, and Genome Solutions). I will describe the Amersham-Pharmacia approach since it is currently the most fully automated though the others differ only in the degree of automation, and not the approach. The gels are run on a plastic backing to allow ease of handling by robots. After imaging, the stained gels are placed in a 'gel hotel', a temporary storage area which prevents them drying out and cracking. Each gel is picked up in turn by a robot arm and then identified by a bar-coding system and the appropriate cut-list is downloaded from the gel analysis system. The plastic backing has two landmark spots that are at either end of the gel. The spots serve to allow an automatic alignment of the gel once it is placed on the X-Y cutting board with the image of the gel used for the analysis to generate the spot-cutting file. The spots are then cut out with a cutter head and placed in bar-coded 96 well plates. The robot arm transfers the plate to a liquid handling station where the spots are washed (destained if non-fluorescent) and the plate is placed in a drying station. The plate is then returned to the liquid handling station for the addition of enzyme. After digestion the spots are extracted and an aliquot is spotted onto a mass spectrometer target plate for subsequent protein fingerprinting. The rest of the extract is kept cooled in the 96 well plate ready for transfer to an autosampler for HPLC-MS/MS analysis should the protein fingerprinting not deliver a high confidence result. All of this runs in a fully automated manner in a closed environment, allowing around 1.500 spots to be prepared for MS analysis per day. The group of Hochstrasser [28] has described an alternative approach to protein identification on 2D gels. After image analysis, the entire 2D gel is rehydrated with trypsin and allowed to digest before being electroblotted through an immobilised trypsin membrane onto a PVDF membrane. The membrane can then be soaked with matrix and analyses directly by scanning in a MALDI mass spectrometer. 3.4. Hierarchical mass spectrometric analysis Mass spectrometer manufacturers have also been focussing on increasing throughput. For gel-based pro-
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
239
Fig.1. (A) Using conventional technology, matching all four gels is very time consuming (ca. 4 hours) and difficult as there is a lot of gel to gel variation. Each gel contains a different amount of proteins in different locations and one cannot tell if there are real differences or if it is due to gel to gel variation. (B) Since the samples are running on the same gel, intragel matching is 100% efficient and automatic. Since there is a pool image in every gel, only these have to be matched to linked the samples and these are identical so matching is very easy and can be automated (ca. 1 min.). (C) The excitation and emission spectra of the various cyanine dyes are shown.
240
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
teome analysis, a two-tiered approach is most commonly used. An aliquot of the digestion extract (between one third and one tenth of the total) is used for protein fingerprinting by Matrix-assisted laser desorption and ionisation mass spectrometry (MALDI MS). The MALDI target plates produced by the gel cutting system are stacked in a storage array and a robot arm automatically loaded into the mass spectrometer. Spectra are automatically obtained for each position and the proteolytic masses are stripped out and used to search the database for proteins that generate a similar theoretical mass profile. This occurs in real-time and one spot can be analysed/identified per minute, matching the output of the cutter system. Those spots not identified by protein fingerprinting at a high enough confidence level can be automatically scheduled for MS/MS analysis. The plate containing the remaining digest is loaded into an autosampler for injection onto a LCMS/MS system. Here the throughput is slower and only 96 samples can be analysed/identified a day. However two different non-commercially available systems have been described by the group of Barry Karger [29,30] which allow rapid electrospray sampling in the order of tens of seconds per sample. The rate-limiting factor here becomes the accumulation of enough MS/MS spectra and the database search. Realistically, when applied to protein digests, a throughput of one sample per minute is obtainable. 3.5. Data analysis Possibly the most critical aspect of the automation procedure is the development of a robust Laboratory Information Management System (LIMS) to deal with the logistics of handling large numbers of samples and collecting and collating the results from the mass spectrometry and gel analyses with the sample types. The LIMS system should also facilitate the scheduling and running of all the samples so that user intervention and error introduction is kept to an absolute minimum. Above this, another layer of software must be available which allows one to interrogate the data from a higher level, cross-matching sets of experiments e.g. match results of 100 sample analysis of liver cirrhosis with a set of results from the result of liver damage by a prescribed medicine. Several firms are developing high-level analysis software based on Oracle data structures. This will allow one to analyse different data types, for example to correlate the findings from a proteomic, transcriptomic and metabolomic analysis of a disease progression and to match these to genotypic and demographic data.
4. The development of non-gel based proteomics approaches Currently there is no universal method allowing the separation and visualisation of all the proteins and their modifications in a cell [31]. Attempts have been made to carry out multi-dimensional chromatographic separation of proteins, however these methods suffer the same drawbacks as the 2D electrophoresis in not being able to handle low abundance and poorly soluble proteins. However one area in which this approach is useful, is for the analysis of the low molecular weight components of human fluids such as urine, blood, cerebrospinal and synovial fluids. Vast numbers of bioactive peptides have been identified and the development of libraries of peptide profiles may become a very useful diagnostic tool [32,33]. However by digesting the proteins into smaller fragments, they are much easier to handle since they are much more homogenous in their physico-chemical properties [10,13]. This removes the problem of very large or small proteins or membrane proteins since once reduced in size to peptides, a good deal, if not all of the peptides can be separated by standard chromatographic means. Thus sample preparation and handling can be simplified and do not have to be optimised for each cell or tissue type as is the case for 2D PAGE. Also, since HPLC can be directly coupled to mass spectrometers and hence to peptide identification and quantification, the entire process can potentially be completely automated with no user intervention necessary. 4.1. Basic requirements What is the total amount of protein needed to observe all peptides in a cell and what degree of separation is needed? If one assumes the maximum sensitivity level for peptide detection and MS/MS is 1 fmol. There are thus 6 x 10 -23 moles of a single copy protein per cell; hence 1.6 x 107 cells (0.25 mg protein extract) are needed, assuming no losses, to obtain 1 fmol signal. Thus the first dimension separation will have to be carried out on a 500 m ID column and the second dimension can then be done with a 150 mm column. Given the human genome is assumed to have 30,000 genes, of which 10% are expressed in any one cell line at a given time and assuming there are on average 20 variants of each protein due to alternative splicing, post-translational modification etc. there will be approximately 200,000 tryptic peptides per cell given an average protein molecular weight of 50 kDa. In order
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
241
to avoid too much signal suppression, one should aim to have a separation method that produces individual spectra containing 10 peptides or less. Given 10 fractions from the first dimension and a second dimension flow rate of 200 nl/min, the peak width will be about 5 sec. Thus a single gradient will have to be around 2.7 hours if a maximum of 10 peptides are to be observed per scan on average, giving a total analysis time of 27 hours. If MS/MS analysis is required for all peaks, then the time increases to 100 hours.
analysis covers only 410 spots corresponding to 282 gene products. The distribution of identified proteins ranged from very low copy number proteins such as transcription factors, through a significantly high number of membrane proteins (131) as well as the extremes of mass (from < 10 to > 550 kDa) and pI (from < 3.9 to > 12.5).
4.2. Initial reports on 2D peptide chromatography based proteomics
A commonly used technique in protein analysis is the use of radiolabelled ammo acids (such as 35S methionine or less commonly cysteine) to increase the detection sensitivity of 2D gel electrophoresis and to allow pulse-chase experiments to determine rates of protein synthesis and degradation. Recently the use of stable isotopes for MS analysis of proteins has been an area of intense interest. One approach is to use isotopically depleted media (i.e. media low in heavy isotopes such as 15N, 2 and 3H and 13C). This allows the isotopic cluster of an ion to be collapsed into a single peak, thereby increasing sensitivity and accuracy [35]. Whole-cell isotopic labelling has been used as a method to quantitate relative changes in protein expression and modification [36]. Cells are grown in either a 'light' medium that consists of compounds with a normal isotope distribution or in 'heavy' medium that is highly enriched in 15N (95%). The protein extracts from the cells grown under different conditions are pooled and partially separated, usually by 2-PAGE. The spot of interest is excised, digested and the extracted peptides analysed by mass spectroscopy. This allows one to differentiate between the peptides originating from the two cell pools since the 15N incorporation shifts one of the pools upwards by one mass unit. The peptides thus appear as doublets and can be quantitated by the relative height of the peaks. This ratio was found to be linear over an abundance ratio of two orders of magnitude. This procedure also lends itself well to defining changes in the level of post-translational modifications. This method is essentially limited by material costs since it is impractical to label entire animals and is often not applicable to eukaryotic cell cultures due to the need for serum derived factors usually obtained from foetal calf serum. An alternative approach is to use post-separation isotopic labelling for relative quantification. A very promising alternative to 2D gel analysis as a comprehensive method for comparative proteomics was recently describing by the group of Ruedi Aebersold and
The basic requirements of such a system are an effective two-dimensional chromatographic separation of peptides with a rapid HPLC-MS/MS analysis and subsequent peptide identification using the fragmentation spectra. The group of John Yates using the yeast ribosome as a model system [34] has made an initial proof of principle that this approach is viable. The first dimension is a strong cation exchange material run in tandem with a second dimension reversed-phase C18 material. The unique feature of this construction is that the separation phases can be packed into the same column, with the first 5 cm being SCX and the last 10 cm up to the nanospray tip being C18 material. The digest material is loaded onto the column and the peptides that do not bind to the SCX column are caught on the C18 column are eluted with a reverse-phase gradient of 080% acetonitrile. Then the peptides are eluted from the SCX onto the RP column by successive cycles of step elution with increasing amounts of ammonium acetate following by a reversed phase gradient. The eluent is directly electrosprayed into an ion-trap mass spectrometer programmed to carry out as many MS/MS analyses as possible. A dynamic exclusion rule is built into the analysis that excludes a peptide mass from being analysed by MS/MS more that once in a defined time window to prevent highly abundant peptides swamping the analysis. The method allowed the identification of all of the predicted ribosomal and ribosomal associated proteins (> 100). Recently, the group has extended this method they term MudPIT (MultiDimensional Protein Identification Technology) to the analysis of the yeast proteome [10]. In a twenty-seven hour chromatography run, 5,540 peptides could be identified by their MS/MS spectra, corresponding to 1,484 proteins, representing a very significant part of the yeast proteome in logarithmically growing cells. The most complete annotated yeast 2D
4.3. Expression quantitation by stable isotope labelling
242
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
B
Fig. 2. (A) This shows schematically how the ICAT (isotope coded affinity tagging) technology works. (B) This shows the chemical structure of the isotopic reagents.
is called Isotope-coded affinity tagging (ICAT [37]). Essentially whole cell protein extracts are digested and labelled with either a 'light' or 'heavy' deuteratedbiotin label which has a thiol specific reactive group (Fig. 2). The mixture is digested and the biotinylated peptides are recovered using an avidin affinity column. This much simpler peptide mixture is then analysed by LCMS using an RP column. The isotopically labelled pairs of peptides elute almost simultaneously and the isotope peak ratios can give the relative amount of each. It is then necessary to perform MS/MS analysis of the peptides to determine their sequence and thus identify the parent protein. An analogous method for isotopic labelling for quantitation has been described by ourselve for proteins isolated by 2D SDS-PAGE but is can also be used directly for peptide mixtures [38]. The proteins are denatured and then succinylated prior to digestion (Fig. 3). The distinct cell digests are either labelled with D4 or H4 nicotinic acid prior to HPLC-MS analysis. Those peptides showing a change in expression or modification level are then chosen for MS/MS analysis, either in a second HPLC run or dynamically in the
first. The combination of any of these isotopic labelling approaches for protein expression quantitation with the MuDPIT technology described above should provide the basis for a broadly applicable non-gel proteomics method and represents the most viable alternative to quantitative 2D-PAGE available today. 4.4. Protein chips The concept of the proteome, if restricted to the set of proteins being expressed in a cell at any given time, yields a fairly static picture. In reality, proteins can only exert their functions in a cell as a result of highly dynamic interactions with other proteins. The cell can be regarded as a series of interacting molecular machines that are formed from large protein complexes [39]. The spatial and temporal modulation of these interactions is the key to defining cell functions in molecular terms. Mass spectrometry is now being explored as a tool to explore the dynamics of protein interactions. Following changes in the phosphorylation state of proteins.
243
Mass/Charge m/z Fig. 3. This show schematically the N-terminal isotopic labelling methodology.
especially those involved in signalling, can help define the set of interactions occurring. A more direct method of defining protein:protein contacts is by direct observation of the complexes and determination of the binding affinities. One method that has recently been developed is surface-enhanced laser desorption/ionisation (SELDI) affinity mass spectrometry [40]. The MALDI target is chemically modified to allow attachment of a 'bait' molecule (analogous to the bait protein used in the yeast two-hybrid system) which is then used to fish for prey proteins, i.e. proteins that bind to the immobilised molecule. The surface can then be washed and the target placed in the mass spectrometer to analyse what has bound to the immobilised molecule. A similar approach has been put forward to map the epitopes of antibodies by immobilising the antibodies on a target and presenting a digest of the target protein. The
non-binding peptides are washed away leaving the peptides that form the epitope. The completion of the human genome will allow the development of antibodies against all of the theoretical open reading frames and these should be commercially available in the not-todistant future. There are many groups working on the development of chip-based systems like those used for mRNA analysis. It should be possible to interface these chips directly with a mass spectrometer and determine which proteins have bound to the bait, beit an antibody, other protein or ligand. In order to identify the bound proteins, either the spot has to be digested which may be problematic if very small amounts of material are bound or the protein can be fragmented directly in the mass spectrometer (see later section on FT-ICR mass spectrometers).
244
P. James / Protein expression analysis: From "tip of the iceberg' to a global method
4.5. Mixed approaches
5. Summary
2D-PAGE separates proteins according to their isoelectric points in the first dimension and by mass in the second. Since stable isotopic labelling methods have been introduced for quantitation by mass spectrometry, it is no longer necessary to run a second dimension SDS-PAGE gel for quantitation by staining and scanning. It would seem logical therefore that one should replace the very low mass accuracy and resolution second dimension gel with a high resolution mass spectrometer. Initial steps in this direction have been carried out using direct scanning of first-dimension IPG strips with an Infra Red Laser [41]. If one wishes to identify the proteins, one can then digest the proteins in situ and repeat the MALDI analysis to obtain protein fingerprints [41]. One can even replace the first dimension gel by capillary isoelectric focussing which can be directly connected to a mass spectrometer with an electrospray ionisation interface [42]. Fourier-Transform Ion Cyclotron Resonance mass spectrometry (FT-ICR MS) allows not only high accuracy mass measurement of the eluting proteins as well as quantitation by isotope distribution but the proteins can be rapidly identified directly using MSn techniques. It has already been demonstrated that intact proteins can be identified from their fragmentation in an FT-ICR by using a combination of the exact intact mass with a series of sequence tags extrapolated from the MSn experiments [43]. Recently Li and Marshall have demonstrated on-line identification of proteins by LC/ESI FT-ICR MS. A normal scan is first used to extract the exact mass of the intact protein and on alternate scans infra-red multi-phonon dissociation (IRMPD) is used to fragment one selected m/z ion from the protein. The intact mass is used with a wide mass window to select a subset of the database entries. The list of mostly b- and y-fragment ions, as well as any small sequence tags obtained from IRMPD, is then matched against this set to identify the protein [44]. Jensen et al. [42] described the analysis of E. coli cell extracts by capillary IEF-FT-ICR using total injections of only 300 ng of protein, which is equivalent to 3 million bacteria (or 3,000 human cells). 400-1,000 putative proteins were found with a mass range between 2 and 100 kDa. The sensitivity is now coming into the range where it will be possible to analyse individual cells. Why look for the needle in the whole haystack if you know which bale it is in? For example, instead of analysing a whole tissue biopsy, individual cell types can now be isolated by laser capture micro-dissection to select only those cells showing a morphology typical for cancer [45].
The main drawback of the non-gel techniques as described is the lack of quantitation of protein expression levels. However MuDPIT is fully compatible with the isotopic labelling techniques described below and should form the basis for a comprehensive proteome analysis tool. The length of time required for the separation is also somewhat limiting if it is to be extended to human proteome analysis. There are however new mass spectrometers under development that may solve this problem by allowing extremely rapid scanning rates (1000s of spectra per second) which are compatible with high speed chromatographic methods. The new mass spectrometers should also show an increase in the dynamic range of detection (and absolute sensitivity) from the current 3-4 orders to seven orders of magnitude, in line with the range of protein expression found in cells. The combination of fast scanning and chromatography could reduce the time from 27 hours to less than 2 hours within a few years. An alternative to on-line analysis of peptides after multi-dimensional separations has been described by the group of Barry Karger [46] using a vacuum deposition interface for coupling capillary electrophoresis with MALDI-TOF MS. The eluent together with matrix is deposited on a moving tape in the evacuated source chamber of a TOF mass spectrometer. The advantage of the method is that the interesting peptides (determined by isotopic ratios) can be analysed by MS/MS after post-run analysis, greatly reducing the number of MS/MS spectra to be accumulated. Maybe now we are verging on the edge of being able to harness the flood of information coming from the genome projects, to put it in order using proteome and microarray/SAGE projects, in a way that we may finally see how all the fine threads are pulled together to make the biochemical web which defines life. As an amateur detective once succinctly put it [47]: "My dear fellow", said Sherlock Holmes, "life is infinitely stranger than anything which the mind of man could invent. If we could fly out of that window and hover over this great city, gently remove the roofs, and peep in at the queer things which are going on, the strange coincidences, the plannings, the cross-purposes, the wonderful chain of events, working through generations, and leading to the most outre results, it would make all fiction with its conventionalities and foreseen conclusions most stale and unprofitable."
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
References [1] M.D. Adams, J.M. Kelley and J.D. Gocayne et al., Complementary DNA sequencing: expressed sequence tags and human genome project, Science 252 (1991), 1651-1656. [2] V.E. Velculescu, L. Zhang, W. Zhou, J. Vogelstein, M.A. Basrai, D.E. Jr. Bassett, P. Hieter, B. Vogelstein and K.W. Kinzler, Characterization of the yeast transcriptome, Cell 88 (1997), 243-251. [3] D.A. Lashkari, J.L. DeRisi, J.H. McCusker, A.F. Namath, C. Gentile, S.Y. Hwang, P.O. Brown and R.W. Davis, Yeast microarrays for genome wide parallel genetic and gene expression analysis, Proc Natl Acad Sci USA 94 (1997), 1305713062. [4] R.K. Saiki, S. Scharf, F. Faloona, K.B. Mullis, G.T. Horn, H.A. Erlich and N. Arnheim, Enzymatic amplification of betaglobin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia, Science 230 (1985), 1350-1354. [5] J.R. Scherer, I. Kheterpal, A. Radhakrishnan, W.W. Ja and R.A. Mathies, Ultra-high throughput rotary capillary array electrophoresis scanner for fluorescent DNA sequencing and analysis, Electrophoresis 20 (1999), 1508-1517. [6] P.H. O'Farrell, High Resolution Two-Dimensional Electroporesis of Proteins, J. Biol. Chem. 250 (1975), 4007-4021. [7] B. Bjellqvist, K. Ek, P.G. Righetti, E. Gianazza, A. Gorg, R. Westermeier and W. Postel, Isoelectric focusing in immobilized pH gradients: principle, methodology and some applications, J. Biochem. Biophys. Meth. 6 (1982), 317-339. [8] W.P. Blackstock and M.P. Weir, Proteomics: quantitative and physical mapping of cellular proteins, Trends Biotechnol. 17 (1999), 121-127. [9] M. Fromont-Racine, J.C. Rain and P. Legrain, Toward a functional analysis of the yeast genome through exhaustive twohybrid screens, Nat Genet. 16 (1997), 277-282. [10] M.P. Washburn, D. Wolters and J.R. Yates 3rd., Large-scale analysis of the yeast proteome by multidimensional protein identification technology, Nat Biotechnol. 19(2001), 242-247. [11] H. Zhou, J.D. Watts and R. Aebersold, A systematic approach to the analysis of protein phosphorylation, Nat Biotechnol. 19 (2001), 375-378. [12] L. Anderson and J. Seilhamer, A comparison of selected mRNA and protein abundances in human liver, Electrophoresis 18 (1997), 533-537. [13] S.P. Gygi, Y. Rochon, B.R. Franza and R. Aebersold, Correlation between protein and mRNA abundance in yeast, Mol Cell Biol. 19(1999), 1720-1730. [14] P.S. Nelson, D. Han, Y. Rochon, G.L. Corthals, B. Lin, A. Monson, V. Nguyen, B.R. Franza, S.R. Plymate, R. Aebersold and L. Hood, Comprehensive analyses of prostate gene expression: convergence of expressed sequence tag databases, transcript profiling and proteomics, Electrophoresis 21 (2000), 1823-1831. [15] O. Fiehn, J. Kopka, P. Dormann, T. Altmann, R.N. Trethewey and L. Willmitzer, Metabolite profiling for plant functional genomics, Nat Biotechnol. 18(2000), 1157-1161. [16] P. Duez, A. Kumps and Y. Mardens, GC-MS profiling of urinary organic acids evaluated as a quantitative method, Clin Chem. 42 (1996), 1609-1615. [17] A. Goffeau, B.C. Barrell and H. Bussey et al., Life with 6000 genes, Science 274(546) (1996), 563-567. [18] S.G. Oliver, M.K. Winson, D.B. Kell and F. Baganz, Systematic functional analysis of the yeast genome, Trends Biotechnol. 16(1998), 373-378.
[19]
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
245
P.S. Madsen, M. Hokland, J. Ellegaard, P. Hokland, G.P. Ratz, A. Celis and J.E. Celis, Major proteins in normal human lymphocyte subpopulations separated by fluorescence-activated cell sorting and analyzed by two-dimensional gel electrophoresis, Leukemia. 2 (1988), 602-615. G.L. Corthals, M.P. Molloy, B.R. Herbert, K.L. Williams and A. A. Gooley, Prefractionation of protein samples prior to twodimensional electrophoresis, Electrophoresis 18 (1997), 317323. M. Fountoulakis, H. Langen, S. Evers, C. Gray and B. Takacs, Two-dimensional map of Haemophilus influenzae following protein enrichment by heparin chromatography, Electrophoresis 18 (1997), 1193-1202. J. Porath, J. Carlsson, I. Olsson and G. Belfrage, Metal chelate affinity chromatography, a new approach to protein fractionation, Nature 258 (1975), 598-599. M.P. Molloy, B.R. Herbert, B.J. Walsh, M.I. Tyler, M. Traini, J.C. Sanchez, D.F. Hochstrasser, K.L. Williams and A.A. Gooley, Extraction of membrane proteins by differential solubilization for separation using two-dimensional gel electrophoresis, Electrophoresis 19 (1998), 837-844. N.G. Anderson, Preparative zonal centrifugation, Methods Biochem Anal. 15 (1967), 271-310. B. Herbert and P.G. Righetti, A turning point in proteome analysis: sample prefractionation via multicompartment electrolyzers with isoelectric membranes, Electrophoresis 21 (2000), 3639-3648. M.G. Harrington, K.H. Lee, M. Yun, T. Zewert, J.E. Bailey and L. Hood, Mechanical precision in two-dimensional electrophoresis can improve protein spot positional reproducibility, Appl Theor Electrophor. 3 (1993), 347-353. M. Unlu, M.E. Morgan and J.S. Minden, Difference gel electrophoresis: a single gel method for detecting changes in protein extracts, Electrophoresis 18 (1997), 2071-2077. W.V. Bienvenut, J.C. Sanchez, A. Karmime, V. Rouge, K. Rose, PA. Binz and D.F. Hochstrasser, Toward a clinical molecular scanner for proteome research: parallel protein chemical processing before and during western blot, Anal Chem. 71 (1999), 4800-4807. H. Liu, C. Felten, Q. Xue, B. Zhang, P. Jedrzejewski, B.L. Karger and F. Foret, Development of multichannel devices with an array of electrospray tips for high-throughput mass spectrometry, Anal Chem. 72 (2000), 3303-3310. C. Felten, F. Foret, M. Minarik, W. Goetzinger and B.L. Karger, Automated high-throughput infusion ESI-MS with direct coupling to a microtiter plate, Anal Chem. 73 (2001), 1449-1454. S.P. Gygi, G.L. Corthals, Y. Zhang, Y. Rochon and R. Aebersold, Evaluation of two-dimensional gel electrophoresisbased proteome analysis technology, Proc Natl Acad Sci USA 97 (2000), 9390-9395. K. Wagner, K. Racaityte, K.K. Unger, T. Miliotis, L.E. Edholm, R. Bischoff and G. Marko-Varga, Protein mapping by two-dimensional high performance liquid chromatography, J ChromatogrA. 893 (2000), 293-305. P. Schulz-Knappe, H.D. Zucht, G. Heine, M. Jurgens, R. Hess and M. Schrader, Peptidomics: the comprehensive analysis of peptides in complex biological mixtures, Comb Chem High Throughput Screen 4 (2001), 207-217. A.J. Link, J. Eng, D.M. Schieltz, E. Carmack, G.J. Mize, D.R. Morris, B.M. Garvik and J.R. Yates 3rd., Direct analysis of protein complexes using mass spectrometry, Nat Biotechnol. 17 (1999), 676-682.
246
P. James / Protein expression analysis: From 'tip of the iceberg' to a global method
[35] T. Solouki, M.R. Emmett, S. Guan and A.G. Marshall, Detection, number, and sequence location of sulfur-containing amino acids and disulfide bridges in peptides by ultrahighresolution MALDIFTICR mass spectrometry. Anal Chem. 69 (1997), 1163-1168. [36] Y. Oda, K. Huang, E.R. Cross, D. Cowburn and B.T. Chait, Accurate quantitation of protein expression and site-specific phosphorylation, Proc Natl Acad Sci USA 96 (1999), 65916596. [37] S.P. Gygi, B. Rist, S.A. Gerber, F. Turecek, M.H. Gelb and R. Aebersold, Quantitative analysis of complex protein mixtures using isotope-coded affinity tags, Nat Biotechnol. 17 (1999), 994-999. [38] M. Munchbach, M. Quadroni, G. Miotto and P. James, Quantitation and facilitated de novo sequencing of proteins by isotopic N-terminal labeling of peptides with a fragmentationdirecting moiety, Anal Chem. 72 (2000), 4047-4057. [39] Cell 92 (1998), whole issue. [40] H. Kuwata, T.T. Yip, C.L. Yip, M. Tomita and T.W. Hutchens, Bactericidal domain of lactoferrin: detection, quantitation, and characterization of lactoferricin in serum by SELDI affinity mass spectrometry, Biochem Biophys Res Commun 245 (1997), 764-773. [41] R.R. Ogorzalek Loo, J.A. Loo and PC. Andrews, Obtaining molecular weights of proteins and their cleavage products by directly combining gel electrophoresis with mass spectrome-
[42]
[43]
[44]
[45]
[46]
[47]
try. Methods Mol Biol. 112 (1999), 473-485. P.K. Jensen, L. Pasa-Tolic, G.A. Anderson, J.A. Horner, M.S. Lipton, J.E. Bruce and R.D. Smith, Probing proteomes using capillary isoelectric focusing-electrospray iomzation Fourier transform ion cyclotron resonance mass spectrometry. Anal Chem. 71(1999), 2076-2084. E. Mortz, P.B. O'Connor, P. Roepstorff, N.L. Kelleher, T.D. Wood, F.W. McLafFerty and M. Mann, Sequence tag identification of intact proteins by matching tanden mass spectral data against sequence data bases, Proc Natl Acad Sci USA 93 (1996), 8264-8267. W. Li, C.L. Hendrickson, M.R. Emmett and A.G. Marshall, Identification of intact proteins in mixtures by alternated capillary liquid chromatography electrospray ionization and LC ESI infrared multiphoton dissociation Fourier transform ion cyclotron resonance mass spectrometry. Anal Chem. 71 (1999), 4397-4402. M.R. Emmert-Buck, R.F. Bonner, P.D. Smith. R.F. Chuaqui, Z. Zhuang, S.R. Goldstein, R.A. Weiss and L.A. Liotta, Laser capture microdissection. Science 274 (19%). 998-1001. J. Preisler, P. Hu, T. Rejtar and B.L. Karger, Capillary electrophoresis - matrix-assisted laser desorption/ionization timeof-flight mass spectrometry using a vacuum deposition interface. Anal Chem. 72 (2000), 4785-4795. A.C. Doyle, The complete Sherlock Holmes, Penguin Books, London, 1930, pp. 190.
247
Analysis of the proteomic profiling of brain tissue in Alzheimer's disease T. Tsuji and S. Shimohama* Department of Neurology, Faculty of Medicine, Kyoto University, 54 Shogoin-Kawaharacho, Sakyoku, Kyoto 606, Japan
In proteome analysis, it is necessary to separate proteins as a first step prior to characterization. Thus, the overall performance of the analysis depends strongly on the separation tool, which is usually two-dimensional electrophoresis (2DE). We have utilized 2DE to begin characterization of the complex pathologic processes in Alzheimer's disease (AD). In the present study, we show how a reliable 2-DE database of brain proteins in Alzheimer's disease was created, improving reproducibility by using an immobilized pH gradient (IPG) for the first dimension gel electrophoresis. The recent progress in this field, and future prospects in this area are also discussed. Preparation of brain proteins into a suitable solubilized state enabled us to separate over 1000 well-defined protein spots in each 2-DE. A comparison of the density of the spots identified on the reference map between the AD and control group, showed that 5 protein spots were significantly increased, 28 spots were significantly decreased and 7 spots were specifically detected in AD. Two spots among those significantly increased and one spot among those significantly decreased were identified as GFAP related. It is hoped that comparative studies to identify, quantitate, and characterize the proteins differentially expressed in normal brain versus diseased brain will give insight into the mechanisms of pathogenesis and allow the development of a strategy to control both the etiology and course of the diseases.
Keywords: Two-dimensional gel electrophoresis, Alzheimer's disease, database, Internet
* Address for correspondence: S. Shimohama, MD, PhD, Department of Neurology, Faculty of Medicine, Kyoto University, 54 Shogoin-Kawaharacho, Sakyoku, Kyoto 606, Japan. Tel.: +81 75 751 3767; Fax: +81 75 751 9541; E-mail:
[email protected]. kyoto-u.ac.jp. Disease Markers 17 (2001) 247-257 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
1. Introduction 1.1. Alzheimer's disease Neural cells are unique in as much as they are unable to multiply after birth and throughout an individual's lifetime. The reason for such longevity is thought to be the good post-translational control of proteins. The most important pathological feature of the brain in aging and Alzheimer's disease (AD) is the presence of aggregated ß-amyloid protein and neurofibrillary tangles (NFT), which are much more prominent in AD brains compared with age-matched non-AD brains. These pathological features have already been described in the original article written by Alzheimer himself. Since then, nearly one century, the precise characterization of these materials has caught the attention of many researchers, but the mechanism of accumulation of these proteins remains unknown. Clarification of the precise mechanism of protein metabolism is an important goal which is expected to be achieved during this century. Recent research on AD has yielded many fruitful and rapidly unfolding observations relating to its pathogenesis. Various proteins such as amyloid precursor protein (APP), ß-amyloid, tau, presenilin and apoE are likely to be involved in the development of this disease [1-7]. Recent advances in molecular biology techniques have enabled us to identify the possible candidate genes for familial AD [8-11]. These mutations have been thoroughly analyzed in relation to the pathological process. Studies at the protein level, however, have lagged partly due to the complexity of the techniques required for protein separation, analysis, and identification. We are developing amazing tools to enable us to investigate whole sets of translated proteins [12], called proteomes (protein complement expressed by the genome in a particular cell or tissue). This will make it possible to map the relationship of whole proteins and their changes in aging and neurodegenerative diseases, such as AD.
248
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease
1.2. Proteome analysis During the last few years, considerable effort has focused on genomic studies and large-scale DNAsequencing projects. The ultimate goal of these studies is to define the genome, the human genome in particular, with the hope of being able to identify and study key genes functional in both normal and aberrant pathways in diseases. Several genetic alterations must be considered in the development of neurodegenerative diseases such as AD, especially in cases with rapid progression. It has been estimated, however, that only 2% of human diseases result from single gene defects (i.e. expression of an altered gene or loss of function of a normal gene). Epigenetic and environmental factors are involved in the remaining 98% of human diseases [13]. While gene expression has traditionally been studied at both the messenger RNA (m RNA) level and the protein expression level, it should be emphasized that mRNA-based studies only identify message abundance and not the actual levels of proteins, which are the functional molecules within the cell wall. In fact, Anderson and Seihamer have recently shown that, in the human liver, there may not be a good correlation between mRNA abundance and the amount of protein present [14]. Analysis of protein expression may provide a better assessment of the metabolic state of cells. Protein analysis, however, is considerably more challenging than DNA/RNA-based analyses, since, in contrast to DNA, which is composed of only four different nucleotides, proteins are composed of at least 20 unmodified, and many more modified, amino acids; thus the physicochemical characteristics of proteins vary considerably. Adding to the complexity of analysis is the fact that many proteins undergo innumerable co- and posttranslational modifications, including phosphorylation, acetylation, sialation, glycosylation, myristoylation, conjugation with lipids, and proteolytic processing. In contrast to the genome, which represents the fixed digitized characteristics of an organism, the proteome is expressed as a constant state of protein metabolism and is dependent upon multiple factors, including the developmental state of the organism, tissue and organelle location, and metabolic stage. Comparative studies to identify, quantitate, and characterize the proteins expressed in normal versus diseased cells will give insight into the mechanisms of pathogenesis.
2. Methodology of proteome analysis 2.1. 2D-PAGE Two-dimensional gel electrophoresis (2-DE) has been developed as a method of protein separation combining isoelectric focusing gel electrophoresis (IEF) with sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis (PAGE). This combination can separate and characterize many thousands of proteins detected as spots on the gels or transferred membranes [15]. The measurement of changes in the expression of multiple proteins represents a powerful strategy for characterizing complex pathophysiologic processes and designing novel drug therapies. Numerous technological improvements and modifications have been made to the original 2D-PAGE protocol. Approaches have focused on the development of reproducible and standardized procedures for sample preparation and the running of gels, enhanced spot detection and accurate protein identification, development of computer software for the analysis of polypeptide spot patterns and database construction. 2.2. Reproducibility and standardization of spot patterns Recently, immobilized pH gradient (IPG) electrophoresis has been introduced to eliminate complications associated with the use of carrier ampholytes (Cas) in the first dimension IEF in 2D-PAGE. IPG gel strips (Pharmacia), formulated in a variety of narrow (pH < 1) and broad (pH 3-10) range gradients, offer many advantages not afforded by CA-IEF including (i) enhanced gel pattern reproducibility. thereby greatly facilitating inter- and intralaboratory gel pattern comparisons; (ii) increased spot resolution and detection sensitivity over wide pH ranges; (iii) enhanced loading capacity for subsequent micropreparative applications. Whereas CA-IEF is essentially limited to the separation of microgram quantities of proteins between pH 4.57.5, IPG-2D-PAGE permits the separation of milligram (1 -15 mg) quantities of proteins on a single IP strip [16. 17]. In addition. IPG strips are considerably easier to manipulate than tube gels, since the IPG gels are covalently bound to mylar backing strips and are not subject to gel stretching and breakage, minimizing geometric distortions in gel patterns accordingly. Intraor inter- laboratory studies comparing the reproducibility of 2D-PAGE patterns using IPG gels in the first dimension showed a less than 1-mm positional variability between spots observed in standardized protein samples [18-21].
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease
249
2.3. Enhanced spot detection/sensitivity
3. Brain proteome analysis in Alzheimer's disease
Since most physiological and pathological processes are associated with a quantitative modulation of gene expression, sensitive spot detection and accurate protein quantitation methodologies are of utmost importance. The use of broad range (pH 4-12) IPGs provides an excellent global overview of protein composition. Separations utilizing well-defined, narrow range (1-2 pH unit) IPGs, specifically designed for the proteins of interest, have permitted the detection and analysis of individual polypeptides and polypeptide isoforms (e.g. phosphorylation, glycoproteins, etc.) having pis differing by as little as 0.01 pH units.While the lower limit of spot detection using current silver staining methods is approximately 1 ng [22] or roughly 10 copies per cell [18], immunoblotting, using highaffinity monoclonal antibodies and visualization with enhanced chemiluminescence can detect as few as 110 copies per cell. Microheterogeneity of m-calpain in the cytosolic fraction and possible changes in AD were detected with high-resolution 2-DE [23]. In the detection of the m-calpain molecule, quantitative information was gained as well as identification of various isoforms that most likely result from posttranslational modification of the primary gene products represented by eight spots on the transferred membranes. Although no qualitative changes in the microheterogeneity were found in AD, (all the spots were detected in both AD and controls), calculation of the ratio of the four acidic spots to the total calpain spots on silver stained 2-DE gels showed significant increase in the ratio of acidic spots in AD.
Several laboratories are constructing 2-DE databases and providing them via the Internet to researchers world wide (http://www.expasy.ch; http://biobase.dk/cgi-bin/ celis; http://www-ips.ncifcrf.gov/lemkin/gellab.html; http://siva.cshl.org/index.html). Using such databases, 2-DE data in individual laboratories can be compared with normal standards in order to detect changes in spots. Although the specimens used to construct these databases have become more diversified, they are currently restricted mainly to cultured cells, blood components, cardiac muscle, and liver proteins. There is not a 2-DE database for human brain proteins presently available on the Internet, which may reflect the heterogeneity of brain tissues that could produce inconsistent 2-DE images as well as the relatively low reproducibility of 2-DE separation of brain proteins. The most important step to permit 2-DE analysis in AD is to establish a reference map for constructing 2-DE databases. In the present study, we established a reference map of human brain proteins and demonstrated specific changes in protein spots in AD by improved reproducibility of the 2-DE analysis. Compared with classical 2-DE using CA, 2-DE electrophoresis using IPG, which are integral parts of the polyacrylamide matrix, has produced significant improvements in 2-DE electrophoretic separation, permitting higher resolution and reproducibility [20,30-32]. Using IPG for the first dimension of electrophoresis, we could use SDS as a protein solubilization reagent for IEF [33]. We constructed a reference map with Melanie II software (Bio-Rad Inc. Richmond, CA) by collecting wellmatched spots within selected gels in order to reduce errors related to the process of making a reference gel. With these improvements, we could analyze protein changes in disease after constructing a 2-DE map of human brain proteins. In the present study, we sought to establish a 2-DE database by identifying several protein spots on the 2-DE map in control specimens and applied the database to the detection of specific changes in these spots in AD.
2.4. Computer analysis/database development/www access Although 2D-PAGE maps appear complex, the technique is highly reproducible, and simple analysis of the 2D-PAGE patterns can be performed by superimposing one photographic image over another on a light box. For more complex studies, sophisticated computer analysis software packages have been developed to aid in the scanning and digitalization of 2D-PAGE maps, quantitation of individual protein spots and the automatic matching of gel images [24-28]. The high reproducibility of IPG-2D-PAGE has permitted the generation of several high-resolution 2D-PAGE reference maps and comprehensive databases of qualitative and quantitative protein expression in a variety of cell types and tissues [29].
4. Materials and methods 4.1. Autopsy brain samples Brain tissues were obtained at autopsy from 15 patients diagnosed clinically and histopathologically with AD (63 to 94 yr, postmortem period 4 to 21 hr), and
250
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease
from 15 age-matched controls (60 to 87 yr, postmortem period 4 to 24 hr) with no clinical or morphologic evidence of brain pathology. In both groups the typical cause of death was cardiac failure or a terminal respiratory condition. Immediately after autopsy the brains were divided saggitally into halves with one half used for biochemical studies and the other half for histological examination. The temporal cortices were used in the present study. The influence of the postmortem period on the 2-DE pattern was examined in three groups with a different postmortem period; early (4-8 hr), middle (9-16 hr) and late (17-24 hr). 4.2. Reagents IPG gradient gel strips (pH 4.0-7.0) and RepelSilane were purchased from Pharmacia IPG (Bromma, Sweden). SDS and 2-D markers were obtained from Bio-Rad. Phenylmethylsulfonyl fluoride (PMSF) and iodoacetamide were obtained from Sigma Chemicals (St Louis, MO). Monoclonal anti- ß actin antibody (clone no. AC-15) and an anti-glial fibrillary acidic protein (GFAP) antibody (clone no. G-A-5) were purchased from Sigma Chemicals. All other chemicals were obtained from Nakarai (Kyoto, Japan). 4.3. Sample preparation Brain tissues were thoroughly sonicated with a hand sonicator in 1 v/w of lysis buffer which contained 10 mM Tris HC1 (pH 7.5), 2% SDS, and 2% mercaptoethanol. After centrifugation at 100 000 g for 1 hr, the supernatant was collected and diluted with sample buffer containing 9 M urea, 0.5% Triton-X 100, and 0.14% PMSF. 4.4. The 2-DE system The first dimension of gel electrophoresis was carried out using an immobilized pH gradient gel (immobilized dry strip gel, pH 4-7/18 cm, Pharmacia) with a horizontal electrophoresis apparatus (Multiphor II, Pharmacia) according to the method described by Gorg et al. [34]. The sample solution was applied on the anodic side of the gel and was run according to manufacturer's instructions. The second dimension of gel electrophoresis was carried out on a 15% running gel (20 cm x 20 cm x 0.1 cm) in the presence of SDS essentially as described by Laemmli [35]. When necessary, marker proteins (SDS-PAGE standards and 2D standards from Bio-Rad) were separated in the same way to estimate the isoelectric points and molecular weights.
4.5. Protein staining After gel electrophoresis in the second dimension, the protein spots were visualized by silver staining with a Wako silver stain kit II (Wako Pure Chem. Ins., Osaka, Japan) which can detect 15 ng protein on 2-D separated spots. 4.6. Immunoblotting Immunodetection is a powerful and sensitive technique which relies on the specificity of antibodies to identify single protein spots from 2-D PAGE. Immunoblotting using commercially available antibodies and the enhanced chemiluminescence (ECL) system (Amerciam, UK) was carried out to identify ß-actin and GFAP on the 2-DE reference map to establish landmarks to facilitate comparison with other 2-DE reference maps. 4.7. Data analysis Protein spots on silver-stained 2-DE gels were digitized using a flatbed scanner at 300 dpi (Agfa-Gevaert, Mortsel, Belgium). The image data were analyzed on a Macintosh computer (Power Macintosh 7600/132) with Melanie II software (Bio-Rad). We analyzed the spots without filtering images to avoid artificial effects on images in comparisons. Spots detected by the program were matched between each gel in each group, and a reference gel was produced by merging the spots from the gels studied. When all gels had been matched to the given reference gel, the latter provided a unique numbering scheme for spot features across all gels. Each paired spot feature in a gel image could then be compared with the corresponding feature in the reference gel. The spot features were characterized with respect to optical density, area and volume. The volume was calculated by integration of the optical density (OD) over the spot's area. Relative volume (% VOL) was the ratio of VOL to total VOL over the whole image. The % VOL of the spots was analyzed to detect specific spots representing significant differences between AD and control groups. Data were analyzed statistically with Statview IV on a Macintosh, enabling us to identify spots in AD significantly changed from the controls by one-way analysis of variance and Bonferroni/Dunne's t test, defining significance as p < 0.05.
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease
5. Results 5.1. Detection of protein spots on the gels A photograph of a 2-D gel, representing the unfiltered raw image demonstrates high-resolution separation of spots and low background staining (Fig. 1). About 700 spots were assigned to a synthetic reference gel, including well-matched spots shared with the control group. Protein spots heavier than 100 kDa were far fewer than proteins lighter than 100 kDa. No significant difference was noted between the characteristics of the spots in the samples from the different postmortem periods (data not shown). Two individual protein spots were respectively assigned to ß-actin and GFAP by immunoblot analysis. 5.2. Protein spots lost or significantly decreased in AD brain There were no protein spots present in the controls but absent in AD. Quantitative analysis using %VOL identified 28 spots decreased in AD (Fig. 2, Table). One of these spots was identified as GFAP. 5.3. Protein spots significantly increased in AD brain Five protein spots were identified as which are significantly increased in the AD brain. The molecular weights of four increased spots (213, 215, 221, 226) were approximately 52 kDa with isoelectric points ranging from pH 4.44 to 5.04. Two of these four spots were identified as GFAP (Figs. 2 and 3). One low molecular weight protein (about 3 to 10 kDa) was detected as an increased spot in AD (Fig. 2, Table 1).
251
mension of electrophoresis [31,37]. Other factors such as autopsy delay, conditions of sonication, and thawing in equilibration solution can influence the results as well. Therefore, an analysis system should allow the extraction of data from a reliable standard spot in often-variable 2-DE images; this was achieved using Melanie II as the data analysis program [25,38]. In the present study, reference gels were created by automatically merging a set of gel images that contained at least three pairwise-matched gels. Application of 2-DE analysis in AD research has focused mainly on the detection of genetic mutations or post-translational modification of proteins such as tau [2], APP [39], actin [40] and heat shock protein [41]. In a study investigating changes in the expression of brain proteins on 2-DE, Mattila et al. [30] applied IPG to the first dimension of electrophoresis and observed four protein spots in Alzheimer 2-DE which were different in the controls; one spot was undetectable, two spots were significantly weaker and one spot was stronger than those in the controls. Our present results included several more spots which were significantly changed in AD, most likely because we used a narrow pH range (pH 4 to 7) and a large (180 x 180 mm) second-dimension SDS gel, which results in better separation. However, we encountered several unresolved problems in the separation of proteins. We could not demonstrate proteins with a molecular weight greatly exceeding 100 kDa, which might be due to the limited ability of heavy proteins to enter the first dimension IPG gel. Studies with larger numbers of cases and investigating more acidic and basic proteins will also be necessary.
5.4. Protein spots detected only in AD brain
6. Separation of hydrophobic proteins as in the case of membrane proteins
We detected seven spots present in AD brains but absent in the controls. These spots were small in volume and faint in staining except for the spots designated A107 and A695 (Fig. 2, Table 1). Analysis with 2-DE has been applied mainly to the proteins from cultured cells, blood components, serum, or body fluids, which are relatively homogeneous [20, 35,36]. Several improvements have allowed the application of this method to human brain, a heterogeneous tissue containing many cellular components such as a variety of neurons and glia as well as microvessels. Improvement in the resolution and reproducibility of gel images was achieved by using IPG for the first di-
While IPGs have considerably improved IEF as described above, there are still some drawbacks. It was noted quite quickly that the resolution of some hydrophobic proteins (membrane proteins) was poor and others were lost when separated by IPGs. It is thought that this is because of hydrophobic interactions between the proteins and the basic acrylamide derivatives of the IPG matrix [42]. More recently the protein patterns of some membrane preparations were compared on CAIEF or IPG-IEF 2-D [43]. These experiments clearly showed that the abundance of some proteins in the second dimension was significantly diminished when IPGs were used. Conversely, a wider study by Wilkins et
252
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease
Fig. 1. Typical silver-stained image of 2-DE gels in control human brain. The pH range is 4 to 7 and the molecular weight markers represent, from top to bottom, 97.4, 66.2,45.0, 31.0. 21.5, and 14.5 (kDa). This image was obtained through digitization with ARCUS II (Agfa-Gevaert, Mortsel, Belgium), subsequently processed using the Melanie II program. Without processing of this image, such as subtraction of background staining, the spots are detected as well-separated areas throughout the gel. By immunoblofting, two groups of spots are identified on the image as ß-actin (single arrow), GFAP farrow-heads) and m-calpain (double arrow-heads).
al. [18] indicated that poor protein solubility could be a contributing factor accounting for the absence of hydrophobic proteins on 2-D gels. These authors examined proteins identified on 2-D gels and matched them to the proteins predicted by genome sequencing. Proteins whose overall ammo acid composition displayed a hydrophobic bias were almost completely absent from the 2-D gels. Interestingly, the mode of IEF (CA or IPG) did not appear to affect this result, suggesting that the initial sample solubilization was the primary point where hydrophobic proteins were lost. In recent years, a greater understanding of the biological and pharmacological importance of membranes has prompted significant efforts to improve the separation of these less soluble proteins using the highresolution method of 2-DE. The solubilization and separation of membrane proteins has proved more complex than that of soluble proteins, especially in IPGs for IEF because of their chemical character and membrane compartmentalization. However, with the benefits of IPGs outweighing most drawbacks, improved techniques and strategies for membrane protein separation have slowly evolved over the past two decades.
Reviews reflect upon the problem encountered with 2DE of membrane proteins and discuss the most recent advances that have been made to overcome the drawbacks identified above [22,44]. 6.1. Protein identification with mass spectrvmetry Since the introduction of MS in protein chemistry, peptide mass fingerprint (PMF) analysis has become the method of choice in high-throughput protein identification. In this approach, the protein of interest, which, in many cases, is purified by 2-D gel electrophoresis, is either enzymatically or chemically cleaved and an aliquot of the peptide mixture thus obtained is analyzed by mass spectrometric techniques. The resulting peptide mass fingerprint is subsequently compared with "virtual" fingerprints obtained by theoretical cleavage of the protein sequences stored in databases and the top-scoring proteins are retrieved as possible candidate proteins. In comparison to other ionization techniques such as Electrospray Ionization (ESI), matrix-assisted laser desorption ionization (MALDI) tolerates moderate buffer and salt concentration in the analyte mix-
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease
4.0
pH
253
7.0 !
Fig. 2. Profile of spots significantly changed in AD compared with the synthetic reference gel. Gray spots represent proteins unchanged in AD brain. Black spots identified by represent significantly increased (p < 0.05) proteins in AD brain, black spots identified by g represent significantly decreased (p < 0.05) proteins in AD brain. Black spots identified by y represent proteins detected only on gels in the AD group. Two groups of spots are identified on the image as ß-actin (single arrow), GFAP (arrow-heads) and m-calpain (double arrow-heads). Two increased spots and one decreased spot were identified as GFAP.
ture and produces almost exclusively singly charged ions [8]. For these reasons it has become the preferred ionization technique for PMF analysis. In the early 90s several research groups reported the use of MALDIPMF in sensitive protein identification [45-47]. It was shown that only a small amount of accurately measured peptide masses is required for unambiguous protein identification and that low pmole to high fmol amounts of gel-separated proteins could be identified using this technique. ESI mass spectrometry has been particularly successful in the structural and conformational characterization of peptides and proteins [48-50]. Electrospray sources are usually trap mass analyzers, and these instruments have been utilized in protein characterization for years. Recently, a new type of hybrid mass spectrometer has been developed, which combines a quadrupole (Q) analyzer and a time-of-flight (TOP) analyzer in one instrument [51]. In these instruments the quadrupole acts as either an ion guide (MS mode) or as an ion filter (MS/MS mode), whilst the TOF analyzer is responsible for measuring the actual mass of the sample ions. In MS mode the quadrupole simply acts as an
ion guide transmitting all ions to the TOF analyzer, and in MS/MS mode it works as an ion filter, allowing only those ions in a predefined mass range to pass through. After passing through the quadrupole analyzer, the filtered ions can be fragmented by collision-induced dissociation in hybrid instruments equipped with a gas collision cell. These daughter ions are then transmitted to the TOF analyzer, where their masses are measured. Using this design it is possible to simultaneously achieve a high resolution and a high sensitivity, due to the very small energy spread and the high transmission of the sample ions [52]. These characteristics give hybrid mass analyzers the capacity to produce high quality MS and MS/Ms spectra from a mol sample quantities [53,54]. Compared to MALDI-MS techniques, sample preparation for ESI-MS is more complex and therefore less well-adapted for high-throughput proteomics. Hybrid Q-TOF mass spectrometers, however, still have a brilliant future and may take a strategic position in the analysis of the site and nature of posttranslational modifications of different proteins. In addition, they remain crucial in the de novo sequencing
254
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease Table 1 List of 2-dimensional electrophoresis protein spots whose density is changed in the AD temporal cortex compared with control temporal cortex from patients with nonneurological disorders
ID
pI
MW
mean OD ratio (AD/control) PI/MW of spots decreased in AD brain 317 4.49 45 0.152 45 335 4.49 0.752 306 4.52 46 0.656 47 297 4.52 0.666 222 4.52 52 0.155 280 4.54 48 0.701 0.197 486 4.55 32 209 4.55 53 0.465 81 4.55 63 0.216 307 4.58 0.887 46 0.854 357 4.60 43 44 347 4.60 0.785 46 312 4.65 0.795 0.216 90 4.69 63 62 0.157 95 4.77 5.10 59 0.610 131 148 5.12 0.561 58 113 5.12 0.108 61 217 5.14 0.899 52 38 0.163 420 5.18 424 5.21 38 0.229 379 5.24 42 0.152 97 5.26 62 0.199 201 5.38 53 0.172 47 0.190 285 5.59 60 6.36 0.962 66 137 6.41 0.138 59 0.176 314 6.68 46 0.144 49 260 6.68
ID
pI
MW
mean OD ratio (AD/control) PI/MW of spots decreased in AD brain (cont,) 74 0.124 6.80 65 536 4.29 29 0.843 572 4.41 29 0.195 505 4.60 0.229 31 29 0.237 606 4.72 759 5.43 6 0.152 545 5.49 29 0.160 29 581 5.57 0.672 5.68 12 716 0.283 0.236 732 6.42 10 PI/MW of spots increased in AD brain 4.44 52 1.139 215 4.47 221 52 1.381 1.294 52 213 4.68 5.04 226 52 1.455 3 765 5.57 1.881 PI/MW of spots detected only in AD brain A877 4.27 31 A830 4.27 32 A695 4.29 38 4.44 A737 36 A638 4.69 42 A107 5.18 66 A10 6.38 98 A1023 6.44 28 6.44 98 A5
The list includes the spot identification number (ID), isoelectric point (pI) and molecular weight (MW). Quantification of spots was carried out by % VOL (see Materials and methods). The figure in the "OD ratio (AD/control)" column is the ratio of the mean value of optical density in AD vs control.
of proteins albeit at very low sensitivities which have not yet found their counterparts in databases.
7. Future direction 7.1. High throughput detection system for post-translational modification Mass spectrometry, in combination with advanced bioinformatics, has now made it possible to identify proteins present in very small amounts on a 2-D gel with accuracy comparable to traditional Edman sequencing [45-47,55]. Hence, with the exponentially growing protein sequence databases, it has become possible to identify the majority of proteins visible on a 2-D gel. The next challenge in proteome analysis is the characterization of post-translational modifications directly
from a spot on a gel. In higher eukaryotes a large proportion of proteins present in a given cell are modified and these modifications are often essential for the function of the protein. Modifications can for example, change the solubility or the stability of the protein, act as molecular switches to control the biological activity of the protein or be used to localize the proteins to different compartments in the cell through complex transport pathways. The most common modifications are phosphorylation and glycosylation [56]. Phosphorylation of different amino acids, mainly serine, threonine and tyrosine residues, is of key importance in the regulation of the activity of proteins involved in signal transduction, metabolism or apoptosis in the cell [57]. Therefore, characterization of protein phosphorylation is essential for understanding the true complexity of living cells. It is not surprising that perturbations in the equilibrium of kinase and
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease
phosphatase activity is fundamentally involved in many diseases such as cystic fibrosis [58], AD [59] and other neurological disorders. Both the quantity and quality of these disease-related phosphorylations should be examined.
255
for providing us autopsy brains. This work was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and grants from the Ministry of Health and Welfare of Japan and the Smoking Research Foundation.
7.2. Protein protein interaction References Attributing a general role to a gene product based on its homology to other proteins can be a relatively easy (automated) task [60-63]. However, determination of the exact physiological function of a protein often requires a tedious and long-lasting effort, underlining the need for a generic method. Additional knowledge of protein function can be acquired by identifying the physiological binding partners of known function. An important technology is the yeast two-hybrid system [64,65]. However, this approach suffers from the fact that yeast lacks some of the physiology of multicellular organisms, in particular tyrosine kinase. Two-dimensional far-Western immunoblotting is helpful in detecting candidates for binding substrates so as to identify protein-protein interactions.
8. Conclusion 2-D PAGE technology and related techniques have reached a new stage in which it is possible to begin to use these methods for clinical applications. While the pathophysiology of the changes found awaits further study, the reference map constructed in the present study is a useful initiation of the human brain proteome database. After establishing the 2-DE database of human brain proteins, we are expanding it by identifying more protein spots, including those specifically changed in AD. We are also interesting to apply the 2DE database to other neurological disorders. We have made these data available on the Internet hoping that all researchers interested in protein analysis of the human brain using 2-DE might take advantage of this 2-DE map as a reference image and collaborate in completing a comprehensive 2-DE database. This will be very helpful in the diagnosis using CSF examination and in the creation of new drug therapy.
Acknowledgments We thank Drs. George Perry and Peter J. Whitehouse at Case Western Reserve University (Cleveland, OH)
[1] C.L. Joachim, J.H. Morris and DJ. Selkoe, Diffuse senile plaques occur commonly in the cerebellum in Alzheimer's disease, Am J Pathol 135 (1989), 309-319. [2] C. Janke, M. Holzer, J. Klose and T. Arendt, Distribution of isoforms of the microtubule-associated protein tau in grey and white matter areas of human brain: a two-dimensional gelelectrophoretic analysis, FEBS Lett 379 (1996), 222-226. [3] P. Seubert, T. Oltersdorf and M.G. Lee et al., Secretion of beta-amyloid precursor protein cleaved at the amino terminus of the beta-amyloid peptide, Nature 361 (1993), 260-263. [4] M. Shoji, T.E. Golde and J. Ghiso et al., Production of the Alzheimer amyloid beta protein by normal proteolytic processing, Science 258 (1992), 126-129. [5] S.H. Yen, D.W. Dickson, A. Crowe, M. Butler and M.L. Shelanski, Alzheimer's neurofibrillary tangles contain unique epitopes and epitopes in common with the heat-stable microtubule associated proteins tau and MAP2, Am J Pathol 126 (1987), 81-91. [6] D.H. Cribbs, L.S. Chen, S.M. Bende and P.M. LaFerla, Widespread neuronal expression of the presenilin-1 earlyonset Alzheimer's disease gene in the murine brain, Am J Pathol 148 (1996), 1797-1806. [7] K. Duff, C. Eckman and C. Zehr et al., Increased amyloidbeta42(43) in brains of mice expressing mutant presenilin 1, Nature 383 (1996), 710-713. [8] J.B. Fenn, M. Mann, C.K. Meng, S.F. Wong and C.M. Whitehouse, Electrospray ionization for mass spectrometry of large biomolecules, Science 246 (1989), 64-71. [9] J.E. Taylor, J.R. Tinklenberg and L.F. Eng et al., Association study between Alzheimer's disease and restriction fragment length polymorphisms at the human amyloid beta protein gene locus, Mol Biol Med 5 (1988), 167-172. [10] H. Tanaka, S. Naruse and K. Seki et al., Absence of linkage disequilibrium at amyloid precursor protein gene locus in Japanese familial Alzheimer's disease with 717Val—>Ile mutation, Neurosci Lett 162 (1993), 63-66. [11] M. Mullan, Familial Alzheimer's disease: second gene locus located [editorial], BMJ 305 (1992), 1108-1109. [12] T. Tsuji, S. Shimohama, S. Kamiya, T. Sazuka and O. Ohara, Analysis of brain proteins in Alzheimer's disease using highresolution two-dimensional gel electrophoresis, J Neurol Sci 166 (1999), 100-106. [13] R. Strohman, Epigenesis: the missing beat in biotechnology? Biotechnology (N Y) 12 (1994), 156-164. [14] L. Anderson and J. Seilhamer, A comparison of selected mRNA and protein abundances in human liver, Electrophoresis 18 (1997), 533-537. [15] P. O'Farrell, High resolution two-dimensional elecrophoresis of proteins, J Biol Chem 250 (1975), 4007-4021. [16] B. Bjellqvist, J.C. Sanchez and C. Pasquali et al., Micropreparative two-dimensional electrophoresis allowing the separation of samples containing milligram amounts of proteins, Electrophoresis 14 (1993), 1375-1378.
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease [17]
[18]
[19]
[20]
[21]
[22] [23]
[24]
[25]
[26]
[27] [28]
[29]
[30]
[31]
[32]
[33]
[34]
P.J. Wirth, T.N. Hoang and T. Benjamin, Micropreparative immobilized pH gradient two-dimensional electrophoresis in combination with protein microsequencing for the analysis of human liver proteins, Electrophoresis 16 (1995), 1946-1960. M.R. Wilkins, E. Gasteiger, J.C. Sanchez, A. Bairoch and D.F. Hochstrasser, Two-dimensional gel electrophoresis for proteome projects: the effects of protein hydrophobicity and copy number, Electrophoresis 19 (1998), 1501-1505. A. Gorg, Two-dimensional electrophoresis with immobilized pH gradients, current state, Biochem Soc Trans 21 (1993), 130-132. J.M. Corbett, M.J. Dunn, A. Posch and A. Gorg, Positional reproducibility of protein spots in two-dimensional polyacrylamide gel electrophoresis using immobilised pH gradient isoelectric focusing in the first dimension: an interlaboratory comparison, Electrophoresis 15(1994), 1205-1211. A. Blomberg, L. Blomberg and J. Norbeck et al., Interlaboratory reproducibility of yeast protein patterns analyzed by immobilized pH gradient two-dimensional gel electrophoresis, Electrophoresis 16(1995), 1935-1945. T. Rabilloud, Solubilization of proteins for electrophoretic analyses, Electrophoresis 17 (1996), 813-829. T. Tsuji, S. Shimohama, J. Kimura and K. Shimizu, m-Calpain (calcium-activated neutral proteinase) in Alzheimer's disease brains, Neurosci Lett 248 (1998), 109-112. N.L. Anderson, J. Taylor, A.E. Scandora, B.P. Coulter and N.G. Anderson, The TYCHO system for computer analysis of two-dimensional gel electrophoresis patterns. Clin Chem 27 (1981), 1807-1820. R.D. Appel, D.F. Hochstrasser and M. Funk et al., The MELANIE project: from a biopsy to automatic protein map interpretation by computer, Electrophoresis 12 (1991). 722735. P.F. Lemkin, Y. Wu and K. Upton, An efficient disk based data structure for rapid searching of quantitative two-dimensional gel databases, Electrophoresis 14 (1993), 1341-1350. J.I. Garrels, The QUEST system for quantitative analysis of two-dimensional gels, J Biol Chem 264 (1989), 5269-5282. P.J. Monardo, T. Boutell, J.I. Garrels and G.I. Latter, A distributed system for two-dimensional gel analysis. Comput Appl Biosci 10 (1994), 137-143. J.E. Celis, H.H. Rasmussen and P. Gromov et al., The human keratinocyte two-dimensional gel protein database (update 1995): mapping components of signal transduction pathways, Electrophoresis 16 (1995), 2177-2240. B. Bjellqvist, K. Ek and P.G. Gighetti et al., Isoelectric focusing in immobilized pH gradients: principle, methodology and some applications. J Biochem Biophvs Methods 6 (1982), 317-339. A. Gorg, W. Postel and S. Gunther et al., Approach to stationary two-dimensional pattern: influence of focusing time and immobiline/carrier ampholytes concentrations. Electrophoresis 9 (1988). 37-46. D. Cade Treyer, A. Cade, A. Darjo and M. Jouvion Moreno, Isoelectric focusing and titration curves in biomedicine and in agrofood industries: a multimedia teaching program, Electrophoresis 17 (1996), 479-482. M. Michea-Hamzehpour, J.C. Sanchez and S.F. Epp et al., Two-dimensional polyacrylamide gel electrophoresis isolation and microsequencing of Pseudomonas aeruginosa proteins, En-yme Protein 47 (1993), 1-8. A. Gorg, W. Postel, A. Domscheit and S. Gunther, Twodimensional electrophoresis with immobilized pH gradients of leaf proteins from barley (Hordeum vulgare): method, re-
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
producibility and genetic aspects. Electrophoresis 9 (1988). 681-692. D.W. Cleveland, S.G. Fischer, M.W. Kirschner and U.K. Laemmli, Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J Biol Chem. 252 (1977). 1102-1106. D. Dreher. J.R. Vargas, D.F. Hochstrasser and A.F. Junod. Effects of oxidative stress and Ca2+ agonists on molecular chaperones in human umbilical vein endothelial cells. Electrophoresis 16 (1995), 1205-1214. P.K. Sinha, E. Kottgen, M. Stoffler Meilicke, E. Gianazza and P.G. Righetti, Two-dimensional maps in very acidic immobilized pH gradients, J Biochem Biophys Methods 20 (1990). 345-352. D.F. Hochstrasser. R.D. Appel and R. Vargas et al., A clinical molecular scanner: the Melanie project, MD Comput 8 (1991). 85-91. L. Autilio Gambetti, A. Morandi and M. Tabaton et al.. The amyloid percursor protein of Alzheimer disease is expressed as a 130 kDa polypeptide in various cultured cell types [published erratum appears in FEBS Lett 1989 Feb 27:244:509]. FEBS Lett 241 (1988), 94-98. W. Jabbour, A. Pouplard-Barthelaix, R. Houlgatte and J. Emile, Abnormal expression of actin in lymphocytes of Alzheimer's disease and Down's syndrome patients. J Neitroimmunol 38 (1992), 199-208. N. Perez, J. Sugar and S. Charya et al., Increased synthesis and accumulation of heat shock 70 proteins in Alzheimer's disease. Brain Res Mol Brain Res 11 (1991). 249-254. H. Kilias. C. Gelfi and P.G. Righetti. Isoenzyme analysis of lichen algae in immobilized pH gradients. Electrophoresis 9 (1988). 187-191. C. Adessi, C. Miege, C. Albrieux and T. Rabilloud. Twodimensional electrophoresis of membrane proteins: a current challenge for immobilized pH gradients. Electrophoresis 18 (1997), 127-135. P.G. Righetti and E. Gianazza, Isoelectric focusing in immobilized pH gradients: theory and newer methodology. Methods Biochem Anal 32 (1987), 215-278. W.J. Henzel, T.M. Billed and J.T. Stults et al., Identifying proteins from two-dimensional gels by molecular mass searching of peptide fragments in protein sequence databases,Proc Natl Acad Sci USA 90 (1993), 5011-5015. P. James. M. Quadroni, E. Carafoli and G. Gonnet, Protein identification by mass profile fingerprinting, Biochem Biophys Res Commun 195 (1993), 58-64. M. Mann, P. Hojrup and P. RoepstorfT, Use of mass spectrometric molecular weight information to identify proteins in sequence databases. Biol Mass Spectrom 22 (1993). 338-345. A. Shevchenko, M. Wilm, O. Vorm and M. Mann, Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem 68 (1996), 850-858. M. Wilm, A. Shevchenko and T. Houthaeve et al., Femtomole sequencing of proteins from polyacrylamide gels by nanoelectrospray mass spectrometry, Nature 379 (1996). 466-469. A. Shevchenko, M. Wilm and M. Mann, Peptide sequencing by mass spectrometry for homology searches and cloning of genes. J Protein Chem 16 (1997). 481-490. T. Rabilloud. C. Adessi. A. Giraudel and J. Lunardi, Improvement of the solubilization of proteins in two-dimensional electrophoresis with immobilized pH gradients. Electrophoresis 18(1997), 307-316. P. RoepstorfT. Mass spectrometry in protein studies from genome to function. Curr Opin Biotechnol 8 (1997). 6-13.
T. Tsuji and S. Shimohama / Analysis of the proteomic profiling of brain tissue in Alzheimer's disease [53]
H.R. Morris, T. Paxton and A. Dell et al., High sensitivity collisionally-activated decomposition tandern mass spectrometry on a novel quadrupole/orthogonal-acceleration time-offlight mass spectrometer, Rapid Commun Mass Spectrom 10 (1996), 889-896. [54] H.R. Morris, T. Paxton, M. Panico, R. McDowell and A. Dell, A novel geometry mass spectrometer, the Q-TOF, for lowfemtomole/attomole-range biopolymer sequencing, J Protein Chem 16 (1997), 469-479. [55] M. Mann and M. Wilm, Error-tolerant identification of peptides in sequence databases by peptide sequence tags, Anal Chem 66 (1994), 4390-4399. [56] R.G. Krishna and F. Wold, Post-translational modification of proteins, Adv Enzymol Relat Areas Mol Biol 67 (1993), 265298. [57] R. Patarca, Protein phosphorylation and dephosphorylation in physiologic and oncologic processes, Crit Rev Oncog 7(1996), 343-432. [58] D.C. Gadsby and A.C. Nairn, Control of CFTR channel gating by phosphorylation and nucleotide hydrolysis, Physiol Rev 79 (1999), S77-S107. [59] D.P. Hanger, J.C. Betts, T.L. Loviny, W.P. Blackstock and B.H.
257
Anderton, New phosphorylation sites identified in hyperphosphorylated tau (paired helical filament-tau) from Alzheimer's disease brain using nanoelectrospray mass spectrometry, J Neurochem 71 (1998), 2465-2476. [60] R. Schaapveld, B. Wieringa and W. Hendriks, Receptor-like protein tyrosine phosphatases: alike and yet so different, Mol Biol Rep 24 (1997), 247-262. [61] J.A. Frearson and D.R. Alexander, Protein tyrosine phosphatases in T-cell development, apoptosis and signalling, Immunol Today 17 (1996), 385-391. [62] H. Sun and N.K. Tonks, The coordinated action of protein tyrosine phosphatases and kinases in cell signaling, Trends Biochem Sci 19 (1994), 480-485. [63] T.R. Jr. Burke and Z.Y. Zhang, Protein-tyrosine phosphatases: structure, mechanism, and inhibitor discovery, Biopohmers 47(1998),225-241. [64] J. Luban and S.P. Goff, The yeast two-hybrid system for studying protein-protein interactions, Curr Opin Biotechnol 6 (1995), 59-64. [65] S. Biffo and P.C. Marchisio, The identification of signaling molecules by the yeast two-hybrid system, Methods Mol Biol 96(1999),211-220.
This page intentionally left blank
259
Proteomics and the inner ear Isolde Thalmann* Washington University School of Medicine, Department of Otolaryngology, St. Louis, MO 63110, USA The inner ear, one of the most complex organs, contains within its bony shell three sensory systems, the evolutionary oldest gravity receptor system, the three semicircular canals for the detection of angular acceleration, and the auditory system - unrivaled in sensitivity and frequency discrimination. All three systems are susceptible to a host of afflictions affecting the quality of life for all of us. In the first part of this review we present an introduction to the milestones of inner ear research to pave the way for understanding the complexities of a proteomics approach to the ear. Minute sensory structures, surrounded by large fluid spaces and a hard bony shell, pose extreme challenges to the ear researcher. In spite of these obstacles, a powerful preparatory technique was developed, whereby precisely defined microscopic tissue elements can be isolated and analyzed, while maintaining the biochemical state representative of the in vivo conditions. The second part consists of a discussion of proteomics as a tool in the elucidation of basic and pathologic mechanisms, diagnosis of disease, as well as treatment. Examples are the organ of Corti proteins OCP1 and OCP2, oncomodulin, a highly specific calcium-binding protein, and several disease entities, Meniere's disease, benign paroxysmal positional vertigo, and perilymphatic fistula. Keywords: Cochlea, organ of Corti, gap junctions, proteomics, oncomodulin, Meniere's disease, benign paroxysmal positional vertigo, otoconia, perilymphatic fistula
1. Introduction to the ear The inner ear, one of the most complex organs, contains within its bony shell three sensory systems, the evolutionarily oldest gravity receptors, the three semi* Address for correspondence: Isolde Thalmann, Ph.D., Washington University School of Medicine, Department of Otolaryngology, 660 S. Euclid Avenue, St. Louis, MO 63110, USA. Tel.: +1 314 362 7505; Fax: +1 314 362 7568; E-mail: thalmanni@msnotes. wustl.edu. Disease Markers 17 (2001) 259-270 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
circular canals for the detection of angular acceleration, and within the same cavity connected by the endolymphatic fluid system, the phylogenetically far younger cochlea, the organ of hearing (Fig. 1(A)). The essential element of these microscopic sensory organs are the hair cells, specialized mechanoreceptors, which are in contact with endolymph, a fluid with highly unusual ionic composition (Figs 1(B)-(D)). It is obviously an enormous undertaking to cover in a short review how this highly specialized organ accomplishes its tasks, how it can be studied from a proteomics standpoint, what is known to go wrong, and where we stand on fixing problems. A didactic approach is used, frequently requiring oversimplification. Nevertheless, we try to present a realistic state-of-theart assessment of the system in terms of its individual components, and at an integrated level. The focus is on the auditory system with its highly specialized receptor, the organ of the Corti. However, due attention is given to two indispensable support systems:1. the endolymphatic sac, the main regulatory center that plays a key role in Meniere's disease, and 2. the gravity receptor with its mineral particles that provide its exceptional sensitivity. Some 30 million Americans suffer from hearing impairment; one of every 1000 infants is born deaf and many more individuals are afflicted by genetic disorders resulting in deafness later in life. An even greater number at some time experience problems of balance or dizziness. In fact, a certain degree of hearing impairment, and degenerative processes in the balance organs, are normal features of aging and affect our quality of life. Care for these disorders exceeds a billion per year. The role proteomics has played in basic and clinical inner ear research is discussed, as well as the challenges of the future. For each project the primary approach, discovery vs hypothesis-driven, global vs specific, is stated. Disease entities covered as examples are Meniere's disease, benign paroxysmal positional vertigo (BPPV), perilymphatic fistula, and non-syndromic genetic hearing loss. 2. Milestones in inner ear research The modern era of inner ear research was launched with the discovery of the organ of Corti around 1850 [1 ].
260
1. Thalmann / Proteomics and the inner ear
Fig. 1. (A) Semidiagrammatic drawing of the ear with the temporal bone cut away to reveal the inner ear consisting of semicircular canals, vestibule and cochlea. (B) Drawing of the cross-section of the gravity receptor of the saccule showing the hair cells, supporting cells and the sensory superstructure, composed of the gelatinous membrane with the embedded otoconia. (Adapted from [42].) (C) (RIGHT) Drawing of a midmodiolar section through a human cochlea showing the three turns. Note the huge fluid spaces and the modiolus in the center which contains the fibers of the auditory nerve. The enlarged cross section (LEFT) shows the three fluid spaces of the cochlea, endolymph of scala media, and the two perilymphatic compartments. The organ of Corti, the auditory receptor, rests on the hasilar membrane and is covered by the tectorial membrane. The stria vascularis, the main transport organ, is located on the lateral wall of the cochlear duct, and is backed by the spiral ligament. (Adapted from [43].) (D) A higher power cross-section through the cochlear duct showing details of the organ of Corti, and the stria vascularis (StV). The supporting cells of the organ of Corti and adjacent nonsensory cells are connected by gap junctions, constituting the epithelial gap junction system (yellow). The spiral ligament and spiral limbus comprise the connective tissue gap junction system (blue). The two gap junction systems interact and are believed to be responsible for the removal of K+ emanating from the hair cells. IHC - inner hair cells; OHC - outer hair cells: IP - inner pillar cells; OP - outer pillar cells; D - Deiters' cells; H - Hensen cells; C - Claudius' cells; IS - inner sulcus; ID - interdental cells; L - limbus: SP - spiral prominence: SL - spiral ligament; RM - Reissner's membrane: BM - basilar membrane: SV - scala vestibuli: ST scala tympani; SM - scala media. (Adapted from [44].)
I. Thalmann / Proteomics and the inner ear
The first electrophysiological studies in the 1920s led to the discovery of the cochlear microphonics, the auditory receptor potential, which is distinct from the action potential [2,3]. In contrast to the all or nothing response of the action potential, cochlear microphonics are essentially electrical analogs of the acoustical stimulus. Von Bekesy inferred that the transduced electrical signal contained significantly more energy than the acoustic stimulus. In search for the required amplifier he discovered the unprecedented positive potential (+80 mV) of the endolymphatic space, the endolymphatic potential [4]. This inspired the first targeted quantitative chemical study by Smith, Lowry and Wu who discovered the equally unprecedented ionic profile of endolymph, a reversed K/Na ratio with K being 150 mEq and Na 1 mEq [5]. The stria vascularis, as the name implies, a highly vascularized epithelial strip on the lateral boundary of the endolymphatic compartment (Fig. 1(D)), was tentatively implicated as the source of the electrochemical gradient and the generator of the endolymphatic potential. With the objective of localizing and characterizing the cochlear energy sources, a search for the metabolic characteristics of the structures in question was initiated. Matschinsky and Thalmann adapted the powerful analytical techniques developed by Lowry for studies of the brain [6], to the unusual anatomical features of the ear [7] incorporating: 1. An analytical technique of the necessary sensitivity and specificity, and 2. a preparatory technique satisfying two essential requirements: a. maintenance of the in vivo metabolic state, achieved by rapid freezing and freeze-drying. b. spatial resolution by dissection of microscopic tissue elements at a level dictated by the aims of the study. This preparatory scheme, originally used in energy studies, has withstood the test of time and has served unchanged as basis for high resolution proteomics to follow much later. Using the behavior of labile metabolites as indicators of the energy status, together with information from a simpler model system, the stria vascularis was confirmed as the cochlear power plant, the source of the K current that flows through the hair cells of the organ of Corti [8,9]. The stereocilia of the hair cells modulate the current flow by sound-synchronous deflections, thereby modulating the current flow and setting up the receptor potential. This essentially was the picture some 25 years ago. However, the situation was soon to be shaken up when Spoendlin found that only the inner hair cells are innervated by afferent fibers and that the outer hair cells are in essence connected only to efferent input [10]. This
261
defined the inner hair cells as the only true auditory receptor cells, cells incorporating all three requirements, the sensing, the transduction, and the transmission elements; it left the glamorous, highly vulnerable outer hair cells out in left field without an obvious role. However, shortly thereafter a series of discoveries uncovered an entirely unanticipated level of refinement of the auditory receptor, superimposed upon the traditional model formulated in the 70s. The following are the key developments: 1. In 1978 Kemp discovered the spontaneous and evoked acoustic emissions, which could not be explained by the classical model but revealed an active process in the cochlea [11]. 2. Brownell et al. in 1985 discovered the electrically induced motility of the outer hair cells [12]. 3. Finally, Zheng et al. (2000) identified and characterized the motor protein prestin in the lateral plasma membrane of the outer hair cells [13]. The picture as we see it today in a nutshell: The inner hair cells are the only cells able to "hear" in the true sense, albeit at a rather rudimentary level. The outer hair cells could, in a sense, be considered 'disabled' receptor cells because they, even though equipped with the sensing and transducer elements, are unable to trigger release of afferent transmitter. Instead, the receptor potential is converted by a process of reverse transduction into synchronous change in the length of the hair cells. In essence the minute acoustic (micromechanical) signals are converted back into mechanical energy, but on a far higher level. Ultimately the motions of the outer hair cells provide the inner hair cells with a vastly amplified mechanical input. This synergism of the outer hair cells with the inner hair cells is the basis of the superb performance of the mammalian auditory receptor which is characterized by high sensitivity and frequency selectivity, in combination with an immense dynamic range. The function of the efferent system is highly complex, but adjustment of the gain of the motor is one essential aspect. In the absence of outer hair cells, the inner hair cells are still able to perceive and transmit sound but at a far reduced level of sensitivity and selectivity. (For a comprehensive review see Dallos et al. [14].) It should be obvious that a great number of steps, predominately mediated by proteins, go on simultaneously within this tiniest of places. It must be evident even to the most enthusiastic proponent of the universal power of DNA technology, that the fine workings of this machinery can only be unraveled by means of a highly selective proteomics approach.
262
I. Thalmann / Proteomics and the inner ear
3. Proteomics of the ear 3.1. Global vs. specific approach Proteomics is at best a daunting project. A choice must be made whether to use a global approach, i.e. trying to take inventory of proteins in various structures - of necessity with rather diffuse goals - or rather focus on subsets of protein(s) with challenging biological or medical properties. As we shall see, the latter approach, derogatorily called the "cottage industry approach" - and claimed to be the principal strategy used by academia - proves to be the preferred one. In the case of the ear the choice of approach is dictated by two factors: 1. anatomical features of the inner ear ; and 2. high-throughput tools. The high-throughput technology is not quite advanced enough to apply it on a large scale to the ear. A hard bony shell, and by comparison huge fluid spaces surrounding the fragile organ of Corti (an epithelial structure devoid of connective tissue and of vascularization), pose extreme challenges to the ear researcher. Moreover, the organ of Corti is postmitotic, and therefore one of the most important tools of proteomics, the use of cell lines, is precluded. The problems are compounded by additional constraints when dealing with the human ear and with pathobiochemical events underlying the various disease entities. The examples to be described will indicate that an eclectic approach which combines both, global and specific approaches, is ultimately most profitable. 3.2. Importance of preparatory technique The availability of a powerful micropreparatory technique for inner ear tissues and fluids developed earlier for metabolic studies (see Introduction) proved to be a major asset for the proteomics application. The importance of a representative sample cannot be sufficiently emphasized. For instance, measurements based on the analysis of the whole cochlea, are entirely useless. The data contributed by the miniscule sensory structures of interest - each with its specialized role - are entirely obscured by the huge amounts of bone, connective tissue and the fluids. To briefly reiterate, the technique, originally developed for metabolic studies, faithfully maintains the existing metabolic state and allows correlation of the biochemistry of precisely defined microscopic tissues elements with the underlying electrophysiologically controlled functional parameters [15,16]. This is accomplished by rapid freezing, freeze-drying in toto (which preserves the three-dimensional structure of in-
ner ear tissues), and dissection of substructures of the organ of Corti in the freeze-dried state (if required to the level of single outer hair cells). A windfall of this technique is that, because the tissue is freeze-dried, no interferences are introduced with, buffers, salts, contaminants from extraction procedures, and there are no concentration problems. Lysis buffer can be added directly prior to 2D-PAGE, for instance. 3.3. Examples of the application of proteomics to basic and clinical inner ear research 3.3.1. Organ of Corti proteins (OCPI and OCP2) Shortly after 2D-PAGE became available in the late 70s, we applied it to the structure of greatest interest, the organ of Corti; we used the technique in a typical proteomics aim, the global discovery mode. We opted not to use the time-honored hypothesis-driven approach, but instead went on a "fishing expedition". In view of the limited sensitivity and a modest sample size (30 ug of total protein of organ of Corti from the ears of two guinea pig), we expected at best to detect the most prominent structural proteins, such as tubulin and actin, and some of the more abundant housekeeping proteins. Instead, as seen in Fig. 2 (left) the two most prominent spots corresponded to two unknown, low molecular weight, acidic proteins. Because they were not detectable in other tissues with 80's technology, we termed them OCP1 and OCP2 [17]. Jumping ahead 20 years, we now know that we had discovered in a most unlikely location, i.e. in the remote organ of Corti, two proteins of universal importance. OCP2 turned out to be a close homolog of SKP1, a protein involved in control of the cell cycle, originally characterized in yeast [18], and OCP1 the first representative of the family of F-box proteins [19]. We hypothesize, that together with cullin-1 OCP1 and OCP2 form the so called SCF complex which targets specific proteins for ubiquitination and proteolysis by the proteasome (Fig. 3). Typical targets of the SCF complex in other systems are cell cycle proteins, such as cyclins and their inhibitors. However, a role of OCP1 and OCP2 in cell cycle control in the organ of Corti can be ruled out because of: 1. the abundance of the proteins (5% of total protein); 2. the cells of the organ of Corti are postmitotic; 3. their location in the supporting cells rather than the hair cells, specifically within the cells of the epithelial gap junction system [21,22].
I. Thalmann / Proteomics and the inner ear
263
Fig. 2. Two-dimensional separation of polypeptides from organ of Corti of guinea pig obtained (A) with 1980's technology (reproduced from [17], with permission) and (B) today's technology. For the gel on the right IPG (pH 3-10L) was used in the first dimension, ISO-DALT (9-18% acrylamide gradient) in the second dimension. 30 ug and 50 ug of total protein were loaded onto gels A and B, respectively. OCP1 and OCP2 were excised from gel B for amino acid sequencing. Because of early conventions, the left gel represents a mirror image. Gel A was Coomassie stained only, gel B was Coomassie stained, followed by silver staining. (For details of method see our Washington University Inner Ear Protein Database http://oto.wustl.edu/thc/innerear2d.htm.)
The target protein of the SCF complex in the organ of Corti remains conjectural, but indirect evidence points to connexin 26 and/or 30 because: 1. of the exact colocalization of the OCPs with connexin 26 and 30; 2. connexins are the only substrates expressed at comparative levels; 3. most connexins are known to turn over rapidly, and are controlled by proteolytic degradation in some systems, e.g. the heart [23]. A rapid turnover has recently been documented also for the cochlear gap junctions [24]. Figure 3 shows a model of the hypothetical role of the OCP1/OCP2 complex in connexin ubiquitination. (For details see [22].) Upshot: The two proteins, OCP1 and OCP2, together with cullin-1, form a regulatory complex that is believed to control an essential service system of the cochlea, the gap junction system. This syncytium of support cells aids in the removal of toxic waste, specifically the large amounts of K emanating from the hair cells, thereby counteracting the biological hardships imposed by a hostile electrochemical environment. Moreover, this system undoubtedly also serves for metabolic equilibration and electrical synchronization in analogy to the syncytium of the myocardium [23]. The importance of this service system is evident from the fact that: a) mutations in connexin 26 and 30 represent the predominant pathogenetic substrate of nonsyndromic hearing loss (over 50%); b) the OCP1 gene is located on the short arm of chromosome 1
in a region that has been implicated in nonsyndromic hearing loss [25]. 3.3.2. Oncomodulin (OM) Detection of OM in the organ of Corti occurred by chance during the course of a targeted study aimed at confirmation of a potential calcium-binding function of OCP2 [26,27]. The experimental set up in this case involved in effect a global search directed at the subset of calcium-binding protein of the organ of Corti. Subsequently we determined that the expression of OM is limited to the outer hair cells and that the protein accounts for as much as 0.5% of the total [28,29]. Also the original discovery of OM was due to chance during the screening of human tumor cell lines [30]. This accounts for the 'onco' portion of the name, while the "modulin" portion derives from certain similarities with calmodulin. This combination of features spurred vigorous interest in cancer research circles. However, the search for a physiological expression site was largely unsuccessful since it yielded only one low expression site in preimplantation embryonic tissue. Consequently prior to discovery of cochlear expression, OM was considered an "onco-developmental" protein. This is a graphic example of the futility of a strictly global search since it cannot possibly detect expression in an isolated microscopic structure, sequestered within a protective bony shell.
264
I. Thalmann / Proteomics and the inner ear
External signal
2nd Messenger (s)
Kinase
©
26S Protessome
Degradation
Substrate (Connexin 26?/30/31/32)
OCPl Fig. 3. Model of the function of the SCF in ubiquitination. External signals activate the corresponding kinase, which modifies the substrate connexin at the phosphorylation site (S). The modified substrate links the target protein to the SCF-E2 ubiquidnation apparatus, leading to the formation of a multiple ubiquitin chain. The multiubiquitinated protein is subject to the proteolysis by the 26 S proteasome. This model could explain the high turnover of connexin. We speculate a similar mechanism for the putative OCPl/OCP2/cullin complex which could play an important role in the epithelial gap junction system. (See text for details.) (Adapted from [45].)
The precise function of OM remains uncertain. However, the high level of expression restricted to the outer hair cells implies involvement in the specialized function of this unique cell type. This could either concern involvement in the motor process per se, or a role in the modulation of the gain of the motor by the efferent system. A direct involvement in the operation of the motor per se is not a viable option, because of its extremely rapid time scale, which can only be accounted for by electrically induced conformational changes. The most obvious alternative, therefore, would be involvement in the modulation of the gain of motor. An attractive hypothesis would postulate a role of OM in the mediation of the "slow efferent" cholinergic effect which modifies the gain of the motor in a calcium-dependent manner [29,31]. 3.3.3. Meniere's disease The detection of the OCPs is an example of a discovery-driven, global approach, while that of OM was a chance observation during a targeted study based
on a global strategy of limited scope. The following example shows another version of a discovery-driven approach with a specific target in mind, the subset of glycosylated proteins of endolymph. An estimated 1/2 to 1 million people in the United States suffer from Meniere's disease. The symptoms include attacks of vertigo, fluctuant hearing loss, tinnitus, and fullness. Figure 4(A) shows a cross-section of a cochlea of a patient with this disease, showing severe hydrops, i.e. increase in endolymph volume. Compelling experimental evidence implicates the endolymphatic sac as the center for volume regulation of the endolymphatic compartment [32] (Fig. 4(B)). An electrophysiologically controlled fluid dynamic study, combined with histochemical evaluation of the endolymphatic sac, showed that the 'homogeneous substance' that normally occupies the sac lumen, rapidly disappears in response to expansion of endolymph volume by injection of artificial endolymph, while the opposite is true when the volume is contracted (Fig. 4(C)-(D)) [33]. The 'homogeneous substance' appears to be a composite of
I. Thalmann / Proteomics and the inner ear
265
Fig. 4. (A) Photomicropgraph of a cross-section of the right cochlea of a patient, showing severe cochlear hydrops. Note the enormous distention of the endolymphatic space which displaces perilymph of scala tympani. For comparison see the normal situation in Fig. 1. (From [46], with permission.) (B) Schematic drawing of the inner ear (see Fig. 1(A)) with parts of the membranous labyrinth exposed, including cochlear duct and endolymphatic sac. The endolymphatic sac is the structure controlling endolymph volume. (C, D) Light micrograph of a cross-section through the endolymphatic sac of the guinea pig. (C) control; (D) experimentally induced expansion of the endolymphatic space by injection of endolymph (injection time 7.5 min). Note almost total absence of the normal intraluminal 'homogeneous substance' on the injected side when compared with the control ear. (From [33], with permission.)
glycosylated macromolecule(s). Our objective was to identify the chemical nature of this substance(s). By treating the fluid content of the endolymphatic sac with various deglycosylating enzymes prior to 2D-PAGE, we arrived at the subset of proteins that is presumed to include the 'homogeneous substance' (Fig. 5) [34]. In this project the main difficulty was procurement of uncontaminated fluid samples. Not only is the endolymphatic sac exceedingly small, holding about 200 nl compared to 4 ul of endolymph and some 30 ul of perilymph (in guinea pig), but its wall consists of a highly vascularized soft tissue layer. Therefore liquid sampling routinely used for cochlear endolymph which is encased in the bony shell, cannot be used effectively in the sac. We developed a technique whereby the content of the endolymphatic sac can be collected in its entirety in the freeze-dried state, free of contaminants. The most fundamental difference between the endolymphatic sac and peripheral endolymph is the level of total protein which is 50 fold higher in the sac. However, there are also several marked differences in the
profile of the proteins (Fig. 5). We determined that midmolecular weight, acidic proteins, make up the majority of detectable proteins, and that most of these are degraded by deglycosylating enzymes. One or several of these proteins presumably correspond(s) to the histologically visible 'homogeneous substance'. At least two levels of reference need to be considered when evaluating the profiles of the fluid of the sac: 1. Comparison of the profile of sac endolymph with that of peripheral endolymph residing within the same compartment, and comparison with the profiles in contiguous compartments, serum, perilymph and CSF; 2. experimental manipulation of endolymph volume, which results in a corresponding increase and decrease in the density of the histologically visible 'homogeneous substance'. Using these approaches in combination, several proteins have been selected as candidates, and tentatively identified on the basis of their electrophoretic mobility; others are being processed for partial amino acid sequencing and/or mass spectrometry.
266
I. Thalmann / Proteomics and the inner ear
Fig. 5. A portion of a two-dimensional separation of guinea pig endolymph derived from: A. endolymphatic sac, and B. cochlea. Sample volumes were adjusted for equal total protein content to facilitate estimation of relative abundance of corresponding spots. Note the two relatively abundant spots, UP1 and UP2 (unidentified protein), and two that we termed USEP1 and 2 (unidentified sac protein) because we could not detect them in the cochlear endolymph. Apolipoproteins D and J (apo D and apo J) show exceptionally high levels in cochlear endolymph. They are known to be present also in high concentrations in CSF.
The currently most effective treatment for relief of the symptoms of the incapacitating vertigo of severe Meniere's disease is destruction of the vestibular sensory organs by judicious use of ototoxic drugs, attempting to spare the cochlea. By identifying the substance(s) that control(s) endolymph volume, we hope to lay the basis for a rational therapeutic scheme aimed at correcting the dysfunction of volume regulation in a targeted, nondestructive way. 3.3.4. Non-syndromic hearing impairment Most genetic hearing impairments are non-syndromic (hearing impairment without other clinical abnormalities). In most instances it was not the discovery of a protein that led directly to identification of the gene(s) responsible for the deafness (except for OCP1). Therefore we refer the reader to a database for more information [35]. 3.3.5. Otoconia and benign paroxysmal, positional vertigo (BPPV) The gravity receptor organ contains a gelatinous sensory superstructure into which biomineral particles, called otoconia, are embedded for enhancement of gravity perception (Fig. 1(B)). Morphogenesis of otoconia is a highly controversial issue, and it is not known whether otoconia once formed during the embryonic
stage persist throughout the entire life span, or whether they can be repaired or regenerated following damage. It is firmly established that human otoconia of the saccular portion of the vestibule tend to degenerate starting at age 50-60 and this undoubtedly contributes to the instability and dysequilibrium experienced by the majority of the elderly. Recent interest in the pathobiology of otoconia surged following the realization that otoconial pathology is causally involved in one of the most prevalent otoneiirologic entities BPPV which preferentially affects younger age groups [36]. Apart from the mechanisms involving morphogenesis, one of the clinically most important questions is whether and to what extent the organic matrix is involved in the maintenance and the demise of otoconia. Highly pertinent in this context is the purported ability of otoconia to regenerate following destruction by ototoxic drugs [37]. As a first step in our attempts to identify the factors responsible for maintenance, repair and regeneration of otoconia, and of devising remedies to prevent degeneration or promote repair, we set out to characterize the otoconial matrix proteins. Briefly, matrix proteins were extracted from decalcified otoconia, separated by 2D-PAGE and the main matrix protein otoconin 90 partially sequenced and cloned [38, 39]. The molecule in question turned out to be a close homolog of phospholipase A2, evidently drafted for its
I. Thalmann / Proteomics and the inner ear
role in the organization of a mineral structure for its rigidity imparted by seven pairs of disulfide bridges. The molecule was further adapted to its role in mineralization by numerous substitutions with acidic residues and incorporation of a massive endowment with complex polysaccharides. The subsequently obtained expression patterns of mRNA and protein in the embryonic ear were in conflict with traditional concepts and required revision of the theory of otoconial morphogenesis for the following reasons: 1. Morphogenesis of otoconia is completed prior to onset of function; 2. Contrary to traditional teachings, the protein is not synthesized and secreted in the sensory region, but in the adjacent nonsensory epithelia. This scheme is plausible from a teleological standpoint since it delegates the menial job of mass production of biomineral particles to less specialized areas of the gravity receptor organ, sparing the highly specialized sense organ. 3. expression of the matrix proteins is not limited to the gravity receptor organ, but is plentiful throughout the entire membranous labyrinth. The question therefore arises why the formation and/or persistence of otoconia is restricted to the gravity receptor organ. This question has not been answered conclusively, but obviously an interaction of otoconin 90 with one or more factors localized to the gravity receptor must be postulated. The most promising candidate is the so-called matrix vesicles secreted by the supporting cells of the sensory region, which are held in place by the gelatinous phase of the gravity superstructure (see [39] for details). In this hypothetical model, interaction of otoconin 90 with acidic phospholipids of the matrix vesicles would initiate nucleation and growth of the calcium carbonate crystals in the low calcium environment of endolymph. How can information about otoconin 90 and other minor matrix proteins help in the understanding of the biologic factors leading to degeneration and/or translocation of otoconia, and thereby clarify the pathogenesis of BPPV? 1. Ongoing experiments on the behavior of the otoconial matrix proteins and the regenerative events following streptomycin intoxication should provide insight into the processes governing the upregulation of embryonic otoconial genes, and thereby provide information how such processes could be induced biochemically without the need to injure the tissues with toxic agents. 2. Observation of the behavior of the biochemical properties of otoconin 90 of human otoconia prior to onset of degeneration should provide important insights into pathobiologic mechanisms occurring on an expanded time scale. The study of human material (obtained during the course of remedial surgery) is in-
267
dispensable: 1. Because it allows comparisons of the behavior of matrix proteins with corresponding pathologic events; and 2. because of the enormous life span, during which wear and tear of inorganic components becomes a significant factor. 3.4. Proteomics: A tool for diagnosis of diseases A limited global approach has been applied to the study of inner ear fluids that showed promise for the development of a diagnostic tool for specific inner ear disorders. 3.4.1. Perilymphatic fistula In the mid 90s an inventory (global approach) of the proteins of the inner ear fluids of the guinea pig, and comparison with CSF and plasma resulted in the important findings that ultrafiltration cannot be the sole mechanism of perilymph production, and that endolymph is derived from perilymph rather than directly from blood [40]. The 2-dimensional electrophoretic protein profile indicated several significant differences from human perilymph obtained during surgery or post mortem. Most conspicuous were extremely high levels of apolipoprotein J and D, and the presence of beta-2 transferrin, a protein absent in plasma, but also present in CSF [41] (Fig. 6). It was attempted to use these proteins as markers for the diagnosis of perilymph fistula, one of the most controversial and challenging problems facing otologists. Perilymph fistula is an abnormal communication between the middle and inner ear, and the clinical signs include vertigo, tinnitus, aural fullness, and sensorineural hearing loss. In the search for markers several unanticipated problems arose which proved difficult to overcome: 1. Perilymph samples obtained during surgery frequently contained traces of plasma, and/or middle ear transudate, and 2. none of the proteins (except for beta-2 transferrin) are entirely specific for perilymph. Note that a small contamination of plasma in terms of volume, is a pronounced contamination in terms of concentration (total protein of plasma:perilymph is 30:1). Beta-2 transferrin, while in theory a highly specific marker for perilymph (and CSF), was of little practical use in immunochemical studies, because commercially available antibodies recognized many other transferrin isoforms and masked the beta-2 isoform. However, with the advent of ProteinChip techniques such as SELDI (Surface Enhanced Laser Desorption/Ionization) beta-2 transferrin could reestablish itself as a viable candidate.
268
I. Thalmann / Proteomics and the inner ear
Fig. 6. Electrophoretic separation of proteins from perilymph of a 6.5 months old infant, collected 2.5 hours postmortem. 3 ul of perilymph were applied to the gel. The total protein content is approximately one-fourth of the adult level. Areas relevant to perilymph fistula diagnosis are enlarged on the left. (From [411, with permission.)
3.4.2. Is the pwteomics approach for diagnosis of other inner ear disorders meaningful? Abnormal protein profiles have been described for plasma/serum, cerebrospinal fluid, and urine (see other articles in this volume) for a variety of disorders. It is not envisioned that analysis of perilymph will be used for diagnostic purposes, but rather as an aid for the elucidation of the mechanisms underlying inner ear disease, whether localized or as part of systemic alterations. The major obstacle is procurement of inner ear fluid sample. Unlike plasma, urine, saliva, milk, sweat, and seminal fluid, an invasive and potentially dangerous procedure is involved. While procurement of CSF and amniotic fluid is invasive, the procedures are relatively straight forward and diagnostic benefits outweigh potential danger. A second obstacle is the minute volumes available, combined with their low protein levels. However, again, the new Protein Chip Technology theoretically is promising. However at this point, the
technique does not warrant the immensely difficult and not risk-free sampling of inner ear fluids.
Acknowledgments This work was supported by research grant 5 ROl DC01414 from the National Institute on Deafness and Other Communication Disorders, National Institutes of Health, and a grant from The National Organization for Hearing Research Foundation.
References [1]
[2]
A. Corti, Recherches sur l'organe de 1'ouie des mammiferes premiere partie Lima con, Zeiischr f wissenschaftl Zoology (1851), 3. E.G. Wever and C.W. Bray, Action currents in the auditory nerve in response to acoustical stimulation. Proc Nat Acad Sci 16(1930). 344-350.
I. Thalmann / Preteomics and the inner ear [3] [4] [5] [6] [7]
[8] [9] [10] [11]
[12]
[13]
[14] [15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
S.S. Stevens and H. Davis, Hearing, Wiley & Sons, New York, 1938. G. von Bekesy, Experiments in Hearing, McGraw-Hill, New York, 1960. C.A. Smith, O.H. Lowry and M.L. Wu, The electrolytes of the labyrinthine fluids, Laryngoscope 64 (1954), 141-153. O.H. Lowry, The quantitative histochemistry of the brain, Histological sampling, JHistochem Cytochem 1 (1953), 420-428. F.M. Matschinsky and R. Thalmann, Quantitative histochemistry of the organ of Corti, stria vascularis and macula sacculi of the guinea pig. I. Sampling procedure and analysis of pyridine nucleotides, Laryngoscope 66 (1967), 292-305. R. Thalmann, Metabolic features of auditory and vestibular system, Laryngoscope 81 (1971), 1245-1260. D.P. Corey and A.J. Hudspeth, Ionic basis of the receptor potential in a vertebrate hair cell, Nature 281 (1979), 675-677. H.H. Spoendlin, The innervation of the organ of Corti, JLaryngol 81 (1967), 717-738. D.T. Kemp, Stimulated acoustic emissions from within the human auditory system, J Acoust Soc Am 64 (1978), 13861391. W.E. Brownell, C.R. Bader, D. Bertrand and Y. de Ribaupierre, Evoked mechanical responses of isolated cochlear outer hair cells, Science 227 (1985), 194-196. J. Zheng, W. Shen, D.Z. He, K.B. Long, L.D. Madison and P. Dallos, Prestin is the motor protein of cochlear outer hair cells, Nature 405 (2000), 149-155. P. Dallos, A.N. Popper and R.R. Fay, eds, The cochlea, Springer, New York, 1996. R. Thalmann, Quantitative histo- and cytochemistry of the ear, in: The handbook of auditory and vestibular research methods, C.A. Smith and J.A. Vernon, eds, C.C. Thomas, Part III Biochemistry, Springfield, IL, 1976, pp. 359-419. R. Thalmann, I. Thalmann and T.H. Comegys, Quantitative cytochemistry of the organ of Corti. Dissection, weight determination and analysis of single outer hair cells, Laryngoscope 82 (1972), 2059-2078. I. Thalmann, H.L. Rosenthal, B.W. Moore AMDR. Thalmann, Organ of Corti-specific polypeptides: OCP-I and OCP-II, Arch Otorhinolaryngol 226 (1980), 123-128. C. Connelly and P. Hieter, Budding yeast SKP1 encodes an evolutionarily conserved kinetochore protein required for cell cycle progression, Cell 86 (1996), 275-285. C. Bai, P. Sen and K. Hofmann et al., SKP1 connects cell cycle regulators to the ubiquitin proteolysis machinery through a novel motif, the F-box, Cell 86 (1996), 263-274. H. Chen, I. Thalmann and J.C. Adams et al., cDNA cloning, tissue distribution, and chromosomal localization of Ocp2, a gene encoding a putative transcription-associated factor predominantly expressed in the auditory organs, Genomics 27 (1995), 389-398. E.R. Yoho, G.N. Thomopoulos, I. Thalmann, R. Thalmann and B.A. Schulte, Localization of organ of Corti protein II in the adult and developing gerbil cochlea, Hear Res 104 (1997), 47-56. T.M. Henzl, J. O'Neal, R. Killick, I. Thalmann and R. Thalmann, OCP1, an F-box protein, colocalizes with OCP2/SKP1 in the cochlear epithelial gap-junction region, Hear Res (2001), in press. J.E. Saffitz, J.G. Laing and K.A. Yamada, Connexin expression and turnover: Implications for cardiac excitability, Circ Res 86 (2000), 723-728.
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32] [33]
[34] [35] [36]
[37]
[38]
[39]
[40]
[41]
[42] [43]
269
J. Edwards, N. Marziano and D. Becker et al., Co-localisation of connexin b2 [26] and connexin b6 [30] in cochlear gap junctions, Abstr Assoc Res Otolaryngol (2001), 150. J.H. Xia, C.Y. Liu and B.S. Tang et al., Mutations in the gene encoding gap junction protein beta-3 associated with autosomal dominant hearing impairment, Nat Genet 20 (1998), 370-373. M. Senarita, I. Thalmann, O. Shibasaki and R. Thalmann, Calcium-binding proteins in organ of Corti and basilar papilla: CBP-15, an unidentified calcium-binding protein of the inner ear, Hear Res 90 (1995), 169-175. I. Thalmann, O. Shibasaki, T.H. Comegys, M.T. Henzl, M. Senarita and R. Thalmann, Detection of a beta-parvalbumin isoform in the mammalian inner ear, Biochem Biophys Res Commun 215 (1995), 142-147. N. Sakaguchi, M. Henzl, I. Thalmann, R. Thalmann and B.A. Schulte, Oncomodulin is expressed exclusively by outer hair cells in the organ of Corti, J Histochem Cytochem 46 (1998), 29-39. I. Thalmann, R. Thalmann and M.T. Henzl, Novel calcium sensor in the outer hair cells: Quantitation of oncomodulin. Primary Sensory Neuron, Highlights in Hair Cell Biology 2 (1998), 283-296. J.P. MacManus, Occurrence of a low molecular weight calcium-binding protein in neoplastic liver, Cancer Res 39 (1979), 3000-3005. T.S. Sridhar, M.C. Liberman, M.C. Brown and W.F. Sewell, A novel cholinergic 'slow effect' of olivocochlear stimulation on cochlear potentials in the guinea pig, J Neurosci 15 (1995), 3667-3678. A.N. Salt, Fluid homeostasis in the inner ear, in: Meniere's disease, J.P. Harris, ed.. Kugler, The Hague, 1999, pp. 93-101. H. Rask-Andersen, J.E. DeMott, D. Bagger-Sjoback and A.N. Salt, Morphological changes of the endolymphatic sac induced by microinjection of artificial endolymph into the cochlea, Hear Res 138 (1999), 81-90. R. Thalmann, Source and role of macromolecules of endolymph, Acta Otolaryngol 119 (1999), 293-296. G. VanCamp and R. Smith, http://www.uia.ac.be/dnalab/hhh/. R.W. Baloh, V. Honrubia and K. Jacobson, Benign positional vertigo: Clinical and oculographic features in 240 cases, Neurol 37 (1987), 371-378. Y. Harada, S. Kasuga and N. Mori, The process of otoconia formation in guinea pig utricular supporting cells, Acta Otolaryngol 118 (1998), 74-79. Y. Wang, P. Kowalski, I. Thalmann, D.M. Ornitz, D.L. Mager and R. Thalmann, Otoconin-90, the mammalian otoconial matrix protein contains two domains of homology to secretory phospholipase A2, Proc Natl Acad Sci (USA) 95 (1998), 15345-15350. R. Thalmann, E. Ignatova, B. Kachar, D.M. Ornitz and I. Thalmann, in: The vestibular system in health and disease, J.A. Goebel and S. Highstein, eds, 2001, in press. I. Thalmann, T.H. Comegys, S.Z. Liu, Z. Ito and R. Thalmann, Protein profiles of perilymph and endolymph of the guinea pig, Hear Res 63 (1992), 37-42. I. Thalmann and J.G. Neely, Progressive hearing loss and the perilymphatic fistula, Seminars in Hearing 16 (1995), 214219. L.P. Gartner and J.L. Hiatt, Color textbook of histology, Saunders, Philadelphia, 1997, pp. 436. A.J. Vander, J.H. Sherman and D.S. Luciano, Human Physiology, New York-St. Louis, McGraw Hill, 1990, pp. 245.
270 [44]
I. Thalmann / Proteomics and the inner ear
N.B. Slepecky, Structure of the mammalian cochlea, in: The cochlea, P. Dallos, A.N. Popper and R.R. Fay, eds, Springer, New York, 19%, pp. 46. [45] S. Hatakeyama, M. Kitagawa and K. Nakayama et al., Ubiquitin-dependent degradation of IkBa is mediated by a
ubiquitin ligase Skpl/Cul 1/F-box protein FWD1, Proc Natl Acad Sci 96 (1999), 3859-3863. [46] H.F. Schuhknecht, Pathology of the ear. Harvard University Press, Cambridge, MA, 1974, pp. 454.
271
Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage Isabelle Noel-Georisa, A. Bernardb, Paul Falmagnea and Ruddy Wattieza,* a
Laboratoire de Chimie Biologique, Universite de Mons-Hainaut, avenue du Champ de Mars 6, B-7000 Mons, Belgium b Unite de Toxicologie Industrielle et de Medecine du Travail, Universite Catholique de Louvain, Clos Chapelle-aux-Champs, B-1200 Bruxelles, Belgium Most lung disorders are known to be associated to considerable modifications of surfactant composition. Numerous of these abnormalities have been exploited in the past to diagnose lung diseases, allowing proper treatment and followup. Diagnosis was then based on phospholipid content, surface tension and cytological features of the epithelial lining fluid (ELF), sampled by bronchoalveolar lavage (BAL) during fiberoscopic bronchoscopy. Today, it appears that the protein content of ELF displays a remarkably high complexity, not only due to the wide variety of the proteins it contains but also because of the great diversity of their cellular origins. The significance of the use of proteome analysis of BAL fluid for the search for new lung disease marker proteins and for their simultaneous display and analysis in patients suffering from lung disorders has been examined.
1. History of surfactant analysis for lung diseases diagnosis The cellular interface between the lung and the environment is composed of a heterologous epithelium: pseudostratified in the proximal airways, cuboi'dal in the distal airways and very thin in the alveoli, the latter representing more than 95% of the lung surface. Besides its primary function of providing an extensive and thin surface for gas exchanges between the blood and the air, the lung epithelium also fulfils a multitude of functions: it provides a protective barrier against inCorresponding author. Tel.: +32 65 37 3311; Fax: +32 65 37 3320; E-mail:
[email protected]. Disease Markers 17 (2001) 271-284 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
fection and injury from the environment, it contributes to the maintenance of the lung fluid balance, thereby preventing lung oedema, it is capable of normal cell turnover and regeneration after injury and, finally, it produces secretions such as mucus, host-defence proteins and surfactant [1—4]. The clinical importance of pulmonary secretions was first demonstrated in the late fifties, on infants dying from respiratory distress syndrome due to surfactant deficiency [5]. Later, synthetic surfactant therapies were shown to restore surfactant function in humans and animals [6]. Nowadays, numerous other lung diseases are associated with biochemical abnormalities of surfactant (reviewed by Griese [7]): obstructive lung diseases (asthma, bronchiolitis, chronic obstructive pulmonary disease, lung transplantation), infectious and suppurative lung diseases (cystic fibrosis (CF), chronic bronchitis, pneumonia, AIDS-related lung disease), adult respiratory distress syndrome (ARDS), pulmonary oedema, interstitial lung diseases (sarcoidosis, idiopathic pulmonary fibrosis (IPF), hypersensitivity pneumonitis, asbestosis), pulmonary alveolar proteinosis (PAP), other diseases specific to infants (neonatal respiratory distress syndrome, meconium aspiration syndrome, congenital diaphragmatic hernia, nosocomial infection in ventilated preterm neonates, chronic lung disease of prematurity, sudden infant death syndrome). Cardiopulmonary bypass and tobacco smoking also result in biochemical modifications of surfactant. Early and specific detection of these surfactant abnormalities in patients is of outstanding interest to enable accurate diagnosis, follow-up, prognosis and treatment of all lung disorders. Since the discovery of the possibility of harvesting the surfactant by bronchoalveolar lavage (BAL), the epithelial lining fluid (ELF) has been the object of an impressive number of studies, allowing analyses of its components, their functions and their possible changes in patients suffering from lung diseases [8]. The first optimised ancestor of bronchoalveolar lavage, using a large volume of saline fluid, appeared
272
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage
at the beginning of the twentieth century and was designed to therapeutically remove lung secretions [9]. Growing interest in the study of the physiological effects of small volume bronchopulmonary lavage and in cellular immunology of the lower airways, along with technical improvements of lavage techniques, lead to the development of routine and standardized sampling, by bronchoalveolar lavage, of cells and secretions making up the thin layer of epithelial lining fluid that covers the airways and the alveoli [8]. Today, bronchoalveolar lavage is considered as a very safe procedure, with no reported lethal complication and minor side-effects compared to more invasive techniques such as transbronchial lung biopsy. Most recent publications now describe standardized washing of the right middle or lower lobe of the lung with 5 X 20 ml of sterile 0.9% w/v saline during fiberoptic bronchoscopy to obtain bronchoalveolar lavage fluid (BALF) after centrifugation [10]. Bronchoalveolar lavage is not, however, the only way of collecting ELF. Sputum induction has, for example, been proposed as a less invasive alternative to bronchoscopy for the collection of airway secretions from asthmatic subjects and patients with chronic obstructive pulmonary disease (COPD) [11]. It involves the inhalation of hypertonic saline aerosol, a stimulus known to cause bronchoconstriction in asthmatic subjects. This method, mainly restricted to the assessment of airway inflammation in asthmatic patients over 6 years old, is currently extended to the sampling of airways of subjects with cystic fibrosis [12], tuberculosis [12] and various interstitial lung diseases (ILD) [13]. Condensation of exhaled breath is, on the other hand, a newly described non-invasive way to collect material originating from the lung, including the lower respiratory tract. It is therefore fully applicable to the follow-up of airway inflammation in very young children and to the analysis of healthy subjects. Its current applications remain however limited to the analysis of hydrogen peroxide, lipid peroxidation and nitric oxide derivatives, and some inflammation parameters present in the exhaled breath condensate [14-16].
2. BAL-based diagnosis of lung diseases 2.1. Phospholipid analyses Biochemical analyses have revealed that the main components of surfactant are phospholipids (~ 90%), especially dipalmitoylphosphatidylcholine (~ 65%)
and phosphatidylglycerol, which are thought to be responsible for the decrease of alveolar surface tension. The role of the other lipid components, of which cholesterol is the most abundant (~ 10%) and other neutral lipids are in trace amounts only, remains to be understood [7]. Phospholipid composition and surface tension of surfactant are modified in a large number of lung diseases. An impressive number of studies have indeed demonstrated a significant decrease in the percentage of phosphatidylcholine and phosphatidylglycerol and an increase of phosphatidylinositol in total phospholipids in pulmonary surfactant of patients suffering from ARDS. Many other surfactant parameters in patients with various other lung diseases have been reviewed recently by Griese [7] and will not be discussed here. 2.2. Cell analyses Much emphasis has also been given to the analysis of the cellular content of BAL, with the aim of using lung cytology as a diagnostic tool: in the normal lung, alveolar macrophages account for 80-95% of the cells in BAL samples [17]. Other cell types include lymphocytes (< 10%), neutrophils (< 5%), eosinophils (< 5%) and sometimes plasma cells [17,18]. Squamous epithelial cells, bronchial epithelial cells, type II pneumocytes, basophils and mast cells are also found in BAL [19,20]. Morphological analyses of BAL cells reveal significant differences in a large number of lung disorders [18, 21-23]. For example, granulomatous and allergic lung diseases are characterized by an increase in the lymphocyte count and high neutrophil counts are characteristic of fibrosing processes or of occupational diseases caused by inhalation of inorganic dust. The percentage of mast cells is significantly increased in BAL of patients with asthma, sarcoidosis or fibrosis [18]. However, most cytological BAL properties are generally not specific of only one lung pathology and can therefore not be considered as disease markers by themselves. For example, a combination of multiple cytological parameters (total cell count, differential cell count, number of infected cells) was shown to provide more significant information than a single parameter for ventilatorassociated pneumonia (VAP) [19]. A computer program has also been used to distinguish between three common ILD by using a combination of several parameters originating from BAL cell analysis [24]. Some peculiar aspects of BAL cytology permits nonetheless an unambiguous diagnosis, for instance in the case of
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage
some infectious pneumopathies, revealed by the presence in BAL of infectious agents like pneumocystis carinii, mycobacteria, aspergillus, anguillulae [19]; non infectious pneumopathies, as BAL allows the visualisation of tumor cells, for example [19]; diffuse alveolar damage: a cellular content of BAL specific of abundant alveolar cell injury, combined with the appropriate clinical setting, allows a correct diagnosis [25]. Phenotypical analysis of the cellular content of BAL, using flow cytometry with fluorescence activated cell analysis and sorting capabilities (FACS) and with monoclonal antibodies directed against specific surface antigens, may give deeper insights into BAL cell subpopulations. For example, lymphocyte subpopulations are modified in various lung diseases, including sarcoidosis [26-28], IPF [29], idiopathic eosinophilic pneumonia [30], hypersensitivity pneumonitis [26,27], bronchiolitis obliterans organizing pneumonia (BOOP) [26]. Differential immunocytochemical staining of BAL alveolar macrophages was also described in IPF [31,32]. Finally, functional studies like spontaneous or induced proliferation assays, chemotaxic and cytotoxicity testings can be performed on BAL cells to give a more accurate information about the immunologic status of the diseased lung [17]. 2.3. Nucleic acid analyses BAL nucleic acids-based diagnosis encountered recently huge progresses thanks to the development of the polymerase chain reaction and its derivatives. Its most important applications today include the detection in BALF of pathogens via nested PCR [33,34] and dosage of cytokins via RT-PCR [35,36]. 2.4. Protein analyses Last, but not least, soluble proteins account for 2030% of surfactant weight and are, to our point of view, the most promising elements of ELF leading to accurate diagnosis of lung disorders since BAL fluid samples contain a large number of different proteins: to date, more than 100 different proteins have been identified using 2-DE gel electrophoresis of BALF samples [37,38]. Moreover, thanks to an outstanding number of sources of BALF proteins, analysis of even a single protein in BALF will enable the integrated assessment of multiple lung parameters at a time. As a mere example, BALF Interleukin-10 has four possible origins: production by pulmonary T cells [39], alveolar
273
macrophages [40], bronchial epithelial cells [41] and diffusion from serum across the air-blood barrier. The wide variety of the proteins found in BALF, together with the multitude of measurements it makes possible in healthy and diseased subjects, are depicted in the next chapter.
3. BAL proteins-based diagnosis of lung diseases 3.1. Protein markers originating from serum The major proteins in BALF are albumin (about 50% of its total protein content), transferrin (about 5.6%), alpha 1-antitrypsin (about 3.5%) and the immunoglobulins A and G (together about 30%) [42]. All of these major proteins are identical to serum proteins and may originate from it by simple diffusion across the intact blood-air barrier. Indeed, the concentration of most serum-like proteins in BAL parallels their serum distribution in healthy subjects [42]. However, the molecular weight of serum-like proteins found in BALF range up to 160 kDa, indicating a possible size exclusion of higher molecular weight proteins [42]. The higher levels of IgG and IgA in BAL than in serum are probably a result of the secretion of these two immunoglobulins by lung tissue [42]. Together with transferrin, IgA and IgG may serve as protecting agents against bacterial and viral infections [43]. The occasional detection of alpha2-macroglobulin in BAL can be explained by its secretion by alveolar macrophages [44]. The presence of alpha 1-antitrypsin probably accounts for the prevention of protease-dependent lung damage due to protease release from inflammatory cells [45]. Other minor BAL proteins that are identical to serum antigens include IgM, IgD, IgE, betal-lipoprotein, plasminogen, complement components C3 and C4, ceruloplasmin, haptoglobin, hemopexin, prealbumin, Gcglobulin, alpha2-HS-glycoprotein, alpha1-acid glycoprotein and beta2-glycoprotein 1 [42]. It is not excluded that these serum-like proteins may also originate from blood contamination during the lavage procedure, thereby questioning the accuracy of concentration estimation of serum-like proteins in BAL. Most interstitial lung diseases are characterized by a massive and significant rise of serum-like proteins in BALF, indicating a probable increase in the permeability of the air-blood barrier. Examples of such marker proteins are numerous and include IgG, IgA, albumin and alphal protease inhibitor in sarcoidosis [46,47], alpha 1 antitrypsin, alpha2 macroglobulin and albumin in
274
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage
pneumonia [48], IgM, alpha2 macroglubulin and albumin in ARDS [49]. The most significant differences in serum-like proteins in BALF of patients with various lung disorders are summarized in Table 1. In conclusion, measuring the BALF levels of serumspecific proteins could well serve as an indicator of the integrity of the air-blood barrier, provided the synthesis of these proteins is unchanged in serum and that the airblood barrier is the unique source of these serum-like antigens in BALF, which is not always the case. 3.2. Lung-specific protein markers Proteins specifically produced by lung epithelial cells include four surfactant-associated proteins (SP-A, B, C and D), the Clara cell protein (CC-16) and mucinassociated antigens (reviewed in [7,50]). The most abundant protein (3-4% of total surfactant weight), SPA, is secreted by type II, Clara and bronchial cells. SPB and SP-D are produced by type II and Clara cells, and SP-C is exclusively synthesised by type II cells in the mature lung. Both SP-A and SP-B are involved in tubular myelin formation. SP-A and SP-D are involved in host defence mechanisms whereas the main function of SP-C is biophysical. CC16, which is exclusively produced by Clara cells, might be crucial for immunosuppression and inflammation down regulation [51]. Mucin KL-6 is expressed in alveolar type II and bronchiolar cells as well as on other somatic cells outside the lung. Tightly associated with cellular membranes, KL-6 is thought to be involved in lung fibrogenic processes, glycocalyx formation and host defence. Mucins 17-Q2 and 17-B1 are secreted by mucous cells in the tracheobronchial lumen, where they participate in the formation of the mucous gel and in host defence mechanisms. Growing interest is given today to these lung-specific proteins with the expectation that their differential expression and/or post-translational modification might be an additional and useful parameter in diagnosing lung diseases. ELISA's (enzyme-linked immunosorbent assay) specific for each of these proteins have therefore been developed, enabling the determination of their absolute amount in BALF of patients with various lung disorders and healthy controls [52-57]. Differences in the amounts of lung-specific proteins in BALF may result from a wide variety of phenomena. Reduction of BALF SP's may be caused by reduction in the amount of secreting cells, or by decreased synthesis and/or release by these cells. Likewise, increase of BALF SP's mav result from increase of the svnthesis
and/or release by secreting cells or by impaired clearance by alveolar macrophages, mucociliary transport, degradation, and absorption into the bloodstream. For instance, increased synthesis and/or release are the most plausible mechanisms explaining BALF SP-A increase in patients with sarcoidosis and HP [58]. BALF SP-A increase in pulmonary alveolar proteinosis is also due to impaired removal/degradation of surfactant [52]. Both SP-A and SP-D are significantly decreased in BALF of patients with idiopathic pulmonary fibrosis (IPF), due to decreased synthesis and/or release by lung epithelial cells [59]. Tobacco smoke was shown to significantly reduce the population of Clara cells, thereby reducing the amount of BALF CC-16 [55]. Increase in BALF KL-6 levels in various ILD's is explained by enhanced release of the mucin, due to inflammation-linked cytotoxicity [60]. All known significant differences in lung-specific proteins in BALF of patients with various lung disorders have been reviewed and are summarized in Table 1. Some of the lung-specific proteins described above are also found in serum, as a result of their spontaneous leakage across the air-blood barrier, and thereby provide another mean of assessing the integrity of the barrier, less invasively than by measuring the levels of these proteins in BALF (reviewed in [50,61]). SPA, SP-B, SP-D, CC-16 and mucin-associated antigens are immunodetected in serum by ELISA's and their respective amounts are modified according to specific lung disorders. Serum SP-A significantly increases in patients with IPF, asbestosis and PAP. Cardiac lung oedema and ARDS are accompanied by a significant increase in SP-A and SP-B in serum. SP-D is significantly higher in serum of patients with fibrosis, PAP, tuberculosis and sarcoidosis. Serum CC-16 is higher in patients with chronic bronchitis, sarcoidosis and pulmonary fibrosis. Mucin KL-6 is elevated in serum of patients with active sarcoidosis and interstitial lung disorders (HP, IPF, COPD). Mucins 17-B1 and 17-Q2 increase in serum from patients with cystic fibrosis and ARDS, respectively. In conclusion, the analysis of lung-specific proteins in BALF enables the evaluation of multiple lung parameters, including the number and integrity of secreting cells, their ability to synthesise and release proteins, the efficiency of the clearance systems. When performed in serum, the analysis of lung-specific proteins also enables the assessment of the air-blood barrier integrity in a less invasive fashion than by their measurement in BALF.
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage
275
Table 1 Protein modifications in BAL fluid of healthy subjects and patients with lung disorders. The first group of proteins are lung-specific proteins, the second correspond to serum-like antigens, the third group comprises proteins secreted by inflammatory proteins (ECP: Eosinophilic Cationic Protein; MCP: Monocyte Chemoattractant Protein) and the fourth group of proteins perform various functions and have been identified by proteome analysis (IgBF: Immunoglobulin Binding Factor; Uq-like: Ubiquitin-like protein; FABP: Fatty Acid Binding Protein). Up- and downarrows indicate proteins that are over- or under- represented in comparison to healthy non-smokers
276
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage Table 1 (cont.) Tobacco smoke
Asthma
CF
Pneumonia ARDS
Caketiculin Calcyclin
3.3. Protein markers produced by lung inflammatory cells Increasing interest in the understanding of the mechanisms involved in the establishment of inflammation during lung diseases in the last decades lead to such a flood of informations about the cytokinic factors secreted by inflammatory cells in the lung tissue it would be presumptuous to try to review them exhaustively. Some recent examples for the most studied diseases are cited hereafter and are outlined in Table 1. During ARDS, pro-inflammatory TNFalpha increases in BAL whereas anti-inflammatory cytokine IL10 decreases [40]. Consistent with this, increased IL10 in the BALF of ARDS patients is correlated with patient survival [62], On the contrary, persistent elevation of inflammatory cytokines in BALF predicts a poor outcome in ARDS [63]. During sarcoidosis, TNFalpha levels were shown to increase in BAL, with concomitant increase of its inhibitor, TNFalpha receptor [64]. Granulocyte-Colony stimulating factor (G-CSF) is detected in BALF of IFF patients while it remains under the detection limits in BALF of healthy volunteers [65]. In BALF of patients with pneumonia, IL8, the major polymorphonuclear neutrophils chemoattractant cytokine, is significantly increased, as well as myeloperoxidase, which is released by neutrophils degranulation [66]. In BALF of patients with tuberculosis, TNFalpha, IL-lbeta and IL-6 are significantly increased [67]. Chemotaxins MCP1, MCP4 and Eotaxin are increased in BAL fluid of asthma patients [6870]. Neutrophils are also responsible for the synthesis of elastase and two other matrix metalloproteinases: collagenase and gelatinase. The enzymatic activity of these three proteinases is increased in ELF of emphysematous patients, where it is responsible for the progressive destruction of extracellular matrix [71]. Tryptase and histamine in bronchoalveolar lavage fluid (BALF) are used as indicators of pulmonary mast cell activation [72]. In conclusion, inflammatory cells present in the lung are diverse and secrete a wide variety of proteins in ELF. Assaying these proteins in BALF by ELISA's and/or enzymatic activity tests will give insights on the current stage of any lung disease, in correlation with the immunological status of the patient. Furthermore,
Sarcoidosis
IFF
Tuberculosis PAP
[91] [91]
measuring cytokines in BALF has also proven to give insights into the comprehension of the mechanisms involved in lung pathogeneses. 3.4. Other protein markers Oxidative burst plays a crucial role in the genesis of inflammatory lung diseases: reactive oxygen metabolites are released by alveolar macrophages, neutrophils and eosinophils during inflammation. BALF Manganese SuperOxide Dismutase is highly expressed by pneumocytes II and alveolar macrophages in the granulomas of pulmonary sarcoidosis and extrinsic allergic alveolitis in response to pro-inflammatory cytokines [73]. Metal-catalysed oxidation of BALF proteins, revealed by their carbonyl content, has been observed in patients with ILD, ARDS and chronic bronchitis [74-77]. BALF alkaline phosphatase is produced by pneumocytes II and is a marker of tissue damage in IFF and BOOP [78]. sFas and sFasL levels, revealing apoptosis, have been measured in BALF of patients with IPF, interstitial pneumonia associated with collagen vascular diseases (CVD-IP), BOOP, and ARDS using ELISA [79, 80]. sFasL is increased in all four disorders compared to control subjects, especially elevated in ARDS patients who died [79]. Moreover, an elevation of the BALF sFas level in BOOP patients, abrogating sFasL cytotoxicity, is associated with better prognosis of BOOP, compared to IPF or CVD-IP [80]. sFas antigen was also analysed by immunocytochemistry in BAL cells: expression in alveolar macrophages was higher in patients with sarcoidosis, lung cancer and fibrosis than in healthy controls [81]. Despite the wide variety of available tumour markers in bronchoalveolar lavage (neuropeptides [82], carbohydrate antigen, carcinoembryonic antigen, neuron specific enolase, squamous cell carcinoma antigen, tissue polypeptide antigen, tissue polypeptide specific antigen, cytokeratin fragment 19, ferritin, antigens Cag25 and CanAg50 [83-85]), none could be defined as diagnostic by itself, most studies advising the use of a combination of several of them to unambiguously diagnose lung carcinomas. In summary, the protein content of BALF displays a huge complexity, due to the enormous diversity of
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage
proteins it contains and the variety of origins of each protein considered. Whereas measuring changes in the levels of one particular protein species gives insights into some particular mechanism affected in a defined lung disease, it is, in general, insufficient to establish a unambiguous diagnosis of one particular disease, allowing proper treatment and follow-up: the combinational analysis of several protein markers is most often recommended. In this context, the use of proteomics, as a means to simultaneously display and analyse large amounts of protein markers and/or as a search tool for new lung disease markers in BALF, is discussed in the last part of this review.
4. Simultaneous display and search for new lung disease markers by BAL proteomic analysis From the abundant studies describing known BAL proteins as potential lung disease markers, it appears obvious that there could also exist unidentified proteins, present amongst the wide variety of proteins of BAL, which concentration varies specifically with one or another lung disorder, and that could therefore become helpful markers for accurately diagnosing lung diseases. The search for such protein markers in the BAL proteome, i.e. all the proteins present in BAL, was therefore initiated and encountered important developments owing to the improvement of all the techniques involved in proteomic analysis. On the other hand, proteome-wide expression analyses also enables simultaneous examination of proteins levels which concentration can be modified as a consequence of changing physiological conditions or diseases. Numerous examples of the use of proteomics for the study of human diseases exist, including the development new drugs (expression proteomics [106]), the diagnosis of neurological disorders (Alzheimer's disease [107]), the identification of tumor associated protein markers (bladder, kidney, breast, lung, ovarian, prostate cancer, leukaemia, neuroblastoma (reviewed in [108]), the search for proteins associated with dilated cardiomyopathy (reviewed in [109]) and, in the context of infectious diseases, the hunt for new diagnostic markers, candidate antigens for vaccines [110] and determinants of virulence [111]. The current tool for displaying a proteome is twodimensional gel electrophoresis (2-DE), that has proven to be particularly suited to provide specific diagnostic information from various body fluids such as seminal fluid, aqueous humor, blood, cerebrospinal fluid, urine,
277
and BALF [112]. In 2-DE, proteins are separated according to their isoelectric point in the first dimension and to their molecular weight in the second. Once stained, the resulting two-dimensional protein maps can be compared, the objective being, in differentialdisplay proteomics [.113],to search for and identify proteins that are up- or downregulated in a disease-specific manner for further use as diagnostic markers. The first 2-DE map of lung lavage effluents was published more than 20 years ago [114]. It provided a comparison of BALF 2-DE of patients with PAP with normal subjects. The premises of differentialdisplay proteomics of BALF were set, along with the search for lung-specific disease markers. BALF protein identifications were originally performed with groupspecific staining, comparisons with purified references standards and pattern matching with 2-DE maps of serum samples. It resulted in the identification of 23 serum-like proteins [42]. At the beginning of the nineties, Lenz and coworkers used the improved method for Isoelectric Focusing developed by Gorg and coworkers (Immobilized PH Gradient, IPG [115]) for the first-dimensional separation of BAL fluid proteins and obtained satisfactory and well reproducible separation [90,116]. They identified SP-A among the 400 spots of a BALF 2-DE gel (pH 4-7) by comparison with 2-DE pattern of purified SP-A [90] and showed that BALF SP-A, along with IgG, IgA and 15 other spots of unknown identity, were modified in IPF, sarcoidosis and/or asbestosis [90]. These observed modified SP-A levels in 2-DE gels of BALF of patients with IPF [90] are in agreement with our 2-DE results [91] and also with previous reports showing that ELIS A-detected SPA was significantly decreased in IPF [117]. Working with IPG strips ranging from pH 3 to 10, Lindahl and coworkers could detect approximately 1000 spots [118] and rose to 29 the number of proteins identified in BALF, which corresponds to 35-40% of all spots [89]. The newly identified proteins included isoforms of lipocortin-1, CC-16, lysozyme and lactoferrin, which may all turn out to be useful markers of lung inflammation [89], especially anti-inflammatory lipocortin1 and CC-16, because their isoform distribution was shown to be altered in BALF from smokers [119]. Lindahl et al. also pinpointed other significant differences in BAL fluid 2-DE pattern between smokers and nonsmokers [120] including IgA, ceruloplasmin and proapolipoprotein A-l [89]. In 1999, the same authors described the identification, by N-terminal sequencing, of several new and clinically interesting proteins [38]. Lipocalin-1 may function as a scavenger against toxic
278
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar Iavage
and proinflammatory lipids [121] and has cysteine proteinase inhibitor function [122]. Bronchial secretions from patients with CF were shown to contain higher levels of lipocalin-1, due to an up-regulated expression of the LCN1 gene [121] and BALF from smokers contain more lipocalin-1 than BALF from nonsmokers [38]. Cystatin S is thought to protect against cystein proteinases inhibitors originating from invading microorganisms or lysosomes. The two identified forms of cystatin S, probably differing from each other by their phosphorylation status, were decreased in BALF from smokers versus nonsmokers [38]. Immunoglobulin binding factor (IgBF), which is abundant in human seminal plasma but has also been found in BALF, may act as a modulator of local immune reponses, and is higher in smokers than in nonsmokers [38,123]. Wishing to characterize the nature of protein BALF components, we undertook an exhaustive identification of all proteins present in human BAL fluid [37]. Our strategy to construct a protein map that contains the widest range of proteins was based on the analysis of individual and pooled BAL fluid samples from patients suffering from various lung disorders. The use of a wide nonlinear gradient IPG to separate proteins with pi between 3 and 10 in the first dimension and of a flat, ultra-thin, horizontal gradient polyacrylamide gel to separate proteins with molecular masses from 5 to 400 kDa in the second dimension strongly increased the amount of information obtainable from BALF 2DE gels. Our present BALF 2-DE map comprises over 1200 silver stained spots ([37,91]; our unpublished results). Over 900 protein spots were identified by matching with the human plasma reference 2-DE map or with other miscellaneous cell line maps available from SWISS-2D PAGE (macrophage-like, epidermal keratinocyte, liver), by N-terminal or internal amino acid microsequencing and mass spectrometry (peptide mass fingerprinting) ([37,91]; our unpublished results). A dynamic 2-DE database for human BALF was made available on the Worldwide Web in 2000 (http://w3.umh.ac.berbiochim/proteomic.htm [91]). The construction of such BAL 2-DE database would obviously not be of great use if it were not to be associated with comparative proteomic analysis of patients with lung disorders. In 2000, a comparison of BALF 2-DE gels of patients with ILD (hypersensitivity pneumonitis, idiopathic pulmonary fibrosis and sarcoidosis) was published [91] and we could identify, with the assistance of our BAL database, a minimum of 50 protein spots displaying modified expression levels for each lung disease compared to healthy subjects. As
an example, a comparison of 2-DE of BALF from IFF, sarcoidosis patients and healthy controls is shown in Fig. 1. The study was done on 15 healthy subjects, 10 patients with IFF and 15 with sarcoidosis, all being nonsmokers. Dramatic increases of some plasma proteins (transfenin, transthyretin, alpha-1 antitrypsin and immunoglobulins) are observed in patients with sarcoidosis. This result is a confirmation of previous observations reporting increases in serum proteins in BALF of sarcoidosis patients [46,47]. The most likely mechanism for this elevation is an increased protein leakage from the bloodstream to the lung tissues caused by inflammatory damage to the alveolar-capillary barrier. Concentration of SP-A, the most abundant pulmonary surfactant protein, is down regulated in BALF of patients with IPF, which has been observed previously in 2-DE gel experiments and ELISA's [90,117], and is probably the consequence of an alteration of the synthesis and/or release of this lung-specific protein, caused by alveolar type II cells damage. High levels of three forms of calgranulin A (spots 201, 202 and 206) are also observed in BALF of IPF patients compared to healthy ones as well as an increase in number and intensity of small and acidic proteins. These include calcium binding proteins (calcyclin, spot 212 and calvasculin, spot 216), lipid binding proteins (epidermal fatty acid-binding protein, spot 174 and adipocyte fatty acid-binding protein, spot 183), enzymes (cathepsin D, spots 108,174 and 207; saposin D chain, spot 210) and miscellaneous proteins (intestinal trefoil factor, spot 179 and ubiquitin-like protein, spot 212). Most of these proteins have been involved in a variety of processes related to cell proliferation, but their exact role in IPF pathogenesis remains to be discovered. The presence of cathepsin D in BALF correlates with reports of its expression by alveolar macrophages, bronchial epithelial cells and type I pneumocytes [124]. Cathepsin D increase in BALF of IPF patients is consistent with its potential role in remodelling processes occurring during fibrogenesis [91,124]. Our 2-DE studies not only allowed the confirmation of the status of disease markers for lung disorders diagnosis (SP-A, cathepsin D, for example), but also permitted the identification of new proteins in BALF. Human AOEB166 was for instance identified in BALF 2-DE gels (see Fig. 1) and microsequenced to enable reverse cloning of the complete gene. The protein, a peroxiredoxin widely expressed in human tissue, is thought to have a protective role in oxidation and inflammatory processes since LPS (lipopolysaccharide)-induced inflammation in rat lung is accompanied by an increase
I. Noe'1-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage
279
Fig. 1. Analytical silver-stained 2-DE gels of human BALF from a healthy subject (A) a patient with IPF (B) and a patient with sarcoidosis (C). 25 mg proteins dissolved in 9 M urea, 0.5% v/v Triton X-100, 2% v/v Pharmalyte 3-10, 65 mM DTE and 8 mM PMSF were loaded on pH 3-10 non-linear IPG strips for isoelectric focusing. Second dimensional separation of the proteins was done on ExcelGel XL 12-14% and detected by silver staining. SP-A is absent in BALF of patients with IPF. Circles indicated small proteins up regulated in IPF (108: Cathepsin D, heavy chains; 172: FABP-E; 174: Cathepsin D, light chain; 179: Intestinal Trefoil Factor; 183: FABP-E; 194: Cathepsin D, light chain; 201, 202 and 206: Calgranulin A; 207: Saposin, D chain; 210: Ubiquitin-like protein; 212: Calcyclin; 216: Calvasculin) and the matching spots in healthy control.
280
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoaJveolar lavage
of lung AOEB166 mRNA levels [125]. Other unknown proteins detected in BALF, their tissue-specificity, their involvement in various lung disorders and their validations as lung disease markers are currently under investigation. Finally, proteomic analysis of BAL fluid has also been recently described as enabling early, direct and sensitive monitoring of the actual impact of therapeutic interventions on specific BALF proteins, long before lung function parameters and structural changes are expected to be detected. BALF SP-A proteolysis levels were assessed, by immunoblotting of the SP-A degradation products displayed on 2-DE gels, to evaluate restoration, by alpha 1-protease inhibitor treatment, of the protease-antiprotease imbalance in cystic fibrosis [126]. In conclusion, recent advances in proteomic analysis have enabled the identification of about 70% of BAL fluid protein spots displayed on 2-DE gels. These investigations confirm previous observations obtained by ELISA measurements and are consistent between different laboratories, which validate the use of 2-DE gels to compare BAL fluid protein levels between patients with lung disorders and healthy controls. The systematic sequencing of BAL fluid proteins is leading to the identification of numerous proteins of unknown function. Their possible use as new disease markers are currently being studied. Moreover, the everyday enrichment of the BALF 2-DE map by new protein sequences further facilitates comparisons betweens 2-DE gels, thereby accelerating the process of differentialdisplay proteomics involved in the identification of new disease markers. Finally, the 2-DE display of BAL fluid proteins permits the integrated analysis of numerous clinical parameters, which is often required to unambiguously and accurately establish a diagnosis of lung diseases and/or to check the influence of therapies on these clinical parameters.
5. Future directions of BALF proteomics-based diagnosis of lung diseases
tion limits for identification (e.g., by fluorescence staining), and use of smaller pI ranges which increase the first dimension resolution of proteins. High-throughput mass spectrometry-based methods, such as SELDITOF (Surface-Enhanced Laser Desorption/Ionisation Time Of Flight [127,128]), able to determine, in total protein extracts, unique protein profiles describing the progression from healthy to diseased states and back, will have to be implemented for BALF-based early diagnosis and therapy follow-up of lung diseases and for high-throughput identification of BALF components. Multiplying the molecular biological approaches to unravel the functional roles and the tissue specificity of the putative disease markers identified by proteomics will further increase confidence in the relevance of our observations and knowledge of the mechanisms involved in lung pathogenesis. On the other hand, the use of in vitro generated recombinant antibodies directed against all lung disease markers identified in BALF by proteomic analyses will give a very sensitive readout of the relative abundance of each marker in BALF or in other body fluids such as serum, induced sputum or breath condensates, monitored simultaneously by western blots, flow cytometry or antibody arrays.
Acknowledgements Our work was supported by the Commission of the European Communities (QLK4-CT-1999-01308). R. Wattiez is a Research Associate at the National Funds for Scientific Research.
References [l]
[2]
[3]
Future research leading to efficient proteomics-based diagnosis of lung diseases has to be oriented in multiple directions. Completion of human BALF 2-DE protein map will require a significant increase in the number and intensity of displayed protein spots, for example through BAL fluid fractionation prior to 2-DE analysis, solubilization of hydrophobic proteins, increased staining sensitivity in conjunction with raised detec-
[41
[51
R. Robbins and S. Rennard. Biology of airway epithelial cells, in: The Lung: Scientific Foundations, R. Crystal. J. West, E. Weibel and P. Barnes, eds. Lippincott-Raven. Philadelphia, 19%. PP. 445-457. Shak S. Mucins and lung secretions, in: The Lung: Scientific Foundations, R. Crystal. J. West. E. Weibel and P. Barnes, eds, Lippincott-Raven. Philadelphia. 1996, pp. 479-486. R. Lubman, K. Kim and E. Crandall. Alveolar epithelial barrier properties, in: The Lung: Scientific Foundations. R. Crystal. J. West. E. Weibel and P. Barnes, eds. LippincottRaven. Philadelphia. 19%, pp. 585-602 Hawgood, Surfactant: composition, structure, and metabolism, in: The Lung: Scientific Foundations. R. Crystal, J. West, E. Weibel and P. Barnes, eds. Lippincott-Raven. Philadelphia, 19%. PP. 557-571. F. Adams. T. Fujiwara, G. Emmanouilides and A. Scudder, Surface properties and lipids from lungs of infants with hyaline membrane disease. J Pediatr 66 (1965). 357-364.
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage [6]
[7] [8]
[9]
[10] [11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
B. Lachmann, P. Berggren, T. Curstedt, G. Grossmann and B. Robertson, Surfactant replacement in experimental respiratory distress syndrome induced by lung lavage, in: Current Concepts in Surfactant Research, P. von Wichert, ed., Karger, Basel Munchen Paris, 1984, PP. 251-256. M. Griese, Pulmonary surfactant in health and human lung diseases: state of the art,EurRespirJl3(1999), 1455-1476. H.Y. Reynolds, Use of bronchoalveolar lavage in humans past necessity and future imperative, Lung 178 (2000), 271293. R.M. Rogers, M.S. Braunstein and J.F. Shurman, Role of bronchopulmonary lavage in the treatment of respiratory failure: a review, Chest 62 (1972), S95-106. F. Capron, Lavage Bronchoalveolaire, Arch Anat Cytol Path 45 (1997), 255-260. F.E. Hargreave and R. Leigh, induced sputum, eosinophilic bronchitis, and chronic obstructive pulmonary disease, Am J Respir Crit Care Med 160 (1999), S53-57. N.R. Henig, M.R. Tonelli, M.V. Pier, J.L. Burns and M.L. Aitken, Sputum induction as a research tool for sampling the airways of subjects with cystic fibrosis, Thorax 56 (2001), 306-311. D. Olivieri, R. D'Ippolito and A. Chetta, induced sputum: diagnostic value in interstitial lung disease, Curr Opin Pulm Med 6 (2000), 411^14. L. Scheideler, H.G. Manke, U. Schwulera, O. inacker and H. Hammerle, Detection of nonvolatile macromolecules in breath. A possible diagnostic tool? Am Rev Respir Dis 148 (1993), 778-784. G. Becher, K. Winsel, E. Beck, G. Neubauer and E. Stresemann, Breath condensate as a method of noninvasive assessment of inflammation mediators from the lower airways, Pneumologie 51(2) (1997), 456-459. P. Reinhold, A. Langenberg, G. Becher and M. Rothe, Breath condensate - a medium obtained by a noninvasive method for the detection of inflammation mediators of the lung, Berl Munch Tierarztl Wochenschr 112 (1999), 254-259. A.M. Tiroke, B. Bewig and A. Haverich, Bronchoalveolar lavage in lung transplantation. State of the art, Clin Transplant 13 (1999), 131-157. G. Mikuz and A. Gschwendtner, Value of bronchoalvoeolar lavage in the diagnosis of lung disease, Verh Dtsch Ges Pathol 84 (2000), 129-135. J.A. Jacobs and E. De Brauwer, BAL fluid cytology in the assessment of infectious lung disease, Hosp Med 60 (1999), 550-555. Z. Chlap, U. Jedynak and K. Sladek, Mast cell: its significance in bronchoalveolar lavage fluid cytologic diagnosis of bronchial asthma and interstitial lung disease, Pneumonol Alergol Pol 66 (1998), 321-329. R.P. Daniele, J.A. Elias, P.E. Epstein and M.D. Rossman, Bronchoalveolar lavage: role in the pathogenesis, diagnosis, and management of interstitial lung disease, Ann intern Med 102(1985), 93-108. H. Popper and M. Pongratz, Value and indications for bronchoalveolar lavage combined with transbronchial lung biopsy, Wien Klin Wochenschr 99 (1987), 848-855. D.G. James, G. Rizzato and O.P. Sharma, Bronchopulmonary lavage (BAL). A window of the lungs, Sarcoidosis 9 (1992), 3-14. M. Drent, M.A. van Nierop, F.A. Gerritsen, E.F. Wouters and P.G. Mulder, A computer program using BALF-analysis results as a diagnostic tool in interstitial lung diseases, Am J Respir Crit Care Med 153 (1996), 736-741.
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
281
C.O. Beskow, C.B. Drachenberg and P.M. Bourquin et al., Diffuse alveolar damage. Morphologic features in bronchoalveolar lavage fluid, Acta Cytol 44 (2000), 640-646. N. Satake, S. Nagai and A. Kawatani et al., Density of phenotypic markers on BAL T-lymphocytes in hypersensitivity pneumonitis, pulmonary sarcoidosis and bronchiolitis obliterans with organizing pneumonia, Eur Respir J 6 (1993), 477-482. M. Raulf, V. Liebers, C. Steppert and X. Baur, increased gamma/delta-positive T-cells in blood and bronchoalveolar lavage of patients with sarcoidosis and hypersensitivity pneumonitis, Eur Respir J 7 (1994), 140-147. J. Wahlstrom, M. Berlin, C.M. Skold, H. Wigzell, A. Eklund and J. Grunewald, Phenotypic analysis of lymphocytes and monocytes/macrophages in peripheral blood and bronchoalveolar lavage fluid from patients with pulmonary sarcoidosis, Thorax 54 (1999), 339-346. E. Fireman, N. Vardinon and M. Burke et al., Predictive value of response to treatment of T-lymphocyte subpopulations in idiopathic pulmonary fibrosis, Eur Respir J 11 (1998), 706711. C. Albera, P. Ohio, P. Solidoro, I. Mabritto, L. Marchetti and E. Pozzi, Activated and memory alveolar T-lymphocytes in idiopathic eosinophilic pneumonia, Eur Respir J 8 (1995), 1281-1285. J. Kiemle-Kallee, H. Kreipe and H.J. Radzun et al., Alveolar macrophages in idiopathic pulmonary fibrosis display a more monocyte-like immunophenotype and an increased release of free oxygen radicals, Eur Respir J 4 (1991), 400-406. I. Striz, Y.M. Wang, I. Svarcova, L. Trnka, C. Sorg and U. Costabel, The phenotype of alveolar macrophages and its correlation with immune cells in bronchoalveolar lavage, Eur Respir J 6 (1993), 1287-1294. F. Moonens, C. Liesnard, F. Brancart, J.P. Van Vooren and E. Serruys, Rapid simple and nested polymerase chain reaction for the diagnosis of Pneumocystis carinii pneumonia, Scand J infect Dis 27 (1995), 358-362. G. Cathomas, M. Tamm, C.E. McGandy, A.P. Perruchoud, M.J. Mihatsch and P. Dalquen, Detection of herpesvirus-like DNA in the bronchoalveolar lavage fluid of patients with pulmonary Kaposi's sarcoma, Eur Respir J 9 (1996), 17431746. M. Fattal-German, F.L. Ladurie, J. Cerrina, F. Lecerf and S. Benin-Aknin, Modulation of 1CAM-1 expression in human alveolar macrophages in vitro, Eur Respir J 9 (1996), 463471. Y. Yu and G. Yang, Growth factor gene expression in bronchoalveolar lavage cells from patients with lung fibrosis, Chin MedSciJ8(l993), 172-176. R. Wattiez, C. Hermans, A. Bernard, O. Lesur and P. Falmagne, Human bronchoalveolar lavage fluid: twodimensional gel electrophoresis, amino acid microsequencing and identification of major proteins, Electrophoresis 20 (1999), 1634-1645. M. Lindahl, B. Stahlbom and C. Tagesson, Newly identified proteins in human nasal and bronchoalveolar lavage fluids: potential biomedical and clinical applications, Electrophoresis 20 (1999), 3670-3676. P. Ralph, I. Nakoinz and A. Sampson-Johannes et al., IL10, T lymphocyte inhibitor of human blood cell production of IL-1 and tumor necrosis factor, J Immunol 148 (1992), 808-814.
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage [40]
[41]
[42]
[43] [44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
L. Armstrong and A.B. Millar, Relative production of tumour necrosis factor alpha and interleukin 10 in adult respiratory distress syndrome, Thorax 52 (1997), 442-446. T.L. Bonfield, M.W. Konstan, P. Burfeind, J.R. Panuska, J.B. Hilliard and M. Berger, Normal bronchial epithelial cells constitutively produce the anti-inflammatory cytokine interleukin-10, which is downregulated in cystic fibrosis. Am J Respir Cell Mol Biol 13 (1995), 257-261. D.Y. Bell, J.A. Haseman, A. Spock, G. McLennan and G.E. Hook, Plasma proteins of the bronchoalveolar surface of the lungs of smokers and nonsmokers. Am Rev Respir Dis 124 (1981), 72-79. B. Muller and P. von Wichert, Bronchoalveolar lavage proteins, Klin Wochenschr 63 (1985), 781-787. R. White, A. Janoff and H.P. Godfrey, Secretion of Alpha2-macroglobulin by human alveolar macrophages. Lung 158 (1980), 9-14. J.E. Gadek, G.A. Fells, R.L. Zimmerman, S.I. Rennard and R.G. Crystal, Antielastases of the human alveolar structures. Implications for the protease-antiprotease theory of emphysema, J Clin invest 68 (1981), 889-898. M. Spatafora, A. Mirabella and G.A. Rossi et al., Lung inflammation in sarcoidosis: analysis of immunoglobulin levels in bronchoalveolar lavage fluid in active and inactive disease, Respiration 48 (1985), 127-135. O. Vandenplas, S. Depelchin, L. Delaunois, J.P. Delwiche and Y. Sibille, Bronchoalveolar lavage immunoglobulin A and G and antiproteases correlate with changes in diffusion indices during the natural course of pulmonary sarcoidosis, Eur Respir J 7 (1994), 1856-1864. F. Ratjen, W. Havers and J. Braun, mtrapulmonary protein leakage in immunocompromised children and adults with pneumonia, Thorax 54 (1999), 432-436. J.F. Holler, J.E. Weiland, E.R. Pacht, J.E. Gadek and W.B. Davis, Protein permeability in the adult respiratory distress syndrome. Loss of size selectivity of the alveolar epithelium. J Clin invest 78 (1986), 1513-1522. C, Hermans and A. Bernard, Lung epithelium-specific proteins: characteristics and potential applications as markers, Am J Respir Crit Care Med 159 (1999), 646-678. P.G. Jorens, Y. Sibille and N.J. Goulding et al., Potential role of Clara cell protein, an endogenous phospholipase A2 inhibitor, in acute lung injury, Eur Respir J 8 (1995). 16471653. Y. Honda, H. Takahashi, N. Shijubo, Y. Kuroki and T. Akino, Surfactant protein-A concentration in bronchoalveolar lavage fluids of patients with pulmonary alveolar proteinosis, Chest 103 (1993), 496-499. Y. Honda, Y. Kuroki and E. Matsuura et al., Pulmonary surfactant protein D in sera and bronchoalveolar lavage fluids. Am J Respir Crit Care Med 152 (1995), 1860-1866. H.J. Kramer, R. Schmidt, A. Gunther, G. Becker, Y. Suzuki and W. Seeger, ELISA technique for quantification of surfactant protein B (SP-B) in bronchoalveolar lavage fluid. Am J Respir Crit Care Med 152 (1995), 1540-1544. N. Shijubo, Y. Itoh and T. Yamaguchi et al., Serum and BAL Clara cell 10 kDa protein (CC10) levels and CClO-positive bronchiolar cells are decreased in smokers, Eur Respir J 10 (1997), 1108-1114. T. Takahashi, M. Munakata, I. Suzuki and Y. Kawakami, Serum and bronchoalveolar fluid KL-6 levels in patients with pulmonary alveolar proteinosis, Am J Respir Crit Care Med 158(1998). 1294-1298.
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71 ]
N. Shijubo, Y. Itoh, T. Yamaguchi and S. Abe, Development of an enzyme-linked immunosorbent assay for Clara cell 10kDa protein: in pursuit of clinical significance of sera in patients with asthma and sarcoidosis. Ann N YAcad Sci 923 (2000), 268-279. H. Hamm, J. Luhrs, Y. Guzman, J. Rotaeche, U. Costabel, H. Fabel and W. Bartsch, Elevated surfactant protein A in bronchoalveolar lavage fluids from sarcoidosis and hypersensitivity pneumonitis patients. Chest 106 (1994), 1766-1770. Y. Kuroki, H. Takahashi, H. Chiba and T. Akino, Surfactant proteins A and D: disease markers, Biochim Biophys Acta 1408(1998), 334-345. N. Kohno, Y. Awaya and T. Oyama et al., KL-6, a mucin-like glycoprotein, in bronchoalveolar lavage fluid from patients with interstitial lung disease. Am Rev Respir Dis 148 (1993). 637-642. C. Hermans and A. Bernard, Pneumoproteinaemia: a new perspective in the assessment of lung disorders, Eur Respir J 11(1998), 801-803. S.C. Donnelly, R.M. Strieter and P.T. Reid et al., The association between mortality rates and decreased concentrations of interleukin-10 and interleukin-1 receptor antagonist in the lung fluids of patients with the adult respiratory distress syndrome, Ann intern Med 125 (19%), 191-1%. G.U. Meduri, S. Headley and G. Kohler et al., Persistent elevation of inflammatory cytokines predicts a poor outcome in ARDS. Plasma IL-1 beta and IL-6 levels are consistent and efficient predictors of outcome over time. Chest 107 (1995). 1062-1073. L. Armstrong, N.M. Foley and A.B. Millar, inter-relationship between tumour necrosis factor-alpha (TNF-alpha) and TNF soluble receptors in pulmonary sarcoidosis. Thorax 54 (1999), 524-530. J. Ashitani, H. Mukae and H. Taniguchi et al., Granulocytecolony stimulating factor levels in bronchoalveolar lavage fluid from patients with idiopathic pulmonary fibrosis. Thorax 54 (1999), 1015-1020. H. Abul. A. Abul. I. Khan, T.C. Mathew, A. Ayed and E. AlAthary, Levels of IL-8 and myeloperoxidase in the lungs of pneumonia patients, Mol Cell Biochtm 217 (2001), 107-112. T.C. Tsao, J. Hong, C. Huang, P. Yang. S.K. Liao and K.S. Chang, increased TNF-alpha, IL-1 beta and IL-6 levels in the bronchoalveolar lavage fluid with the upregulation of their mRNA in macrophages lavaged from patients with active pulmonary tuberculosis, Tuber Lung Dis 79 (1999), 279-285. K.J. Rozyk, T. Plusa, P. Kuna and E. Pirozynska, Monocyte chemotactic and activating factor/monocyte chemoattractant protein in bronchoalveolar lavage fluid from patients with atopic asthma and chronic bronchitis. Relationship to lung function tests, bronchial hyper-responsiveness and cytology of bronchoalveolar lavage fluid. Immunol Lett 58 (1997). 47-52. B. Lamkhioued. E.A. Garcia-Zepeda and S. Abi-Younes et al., Monocyte chemoattractant protein (MCP)-4 expression in the airways of patients with asthma, induction in epithelial cells and mononuclear cells by proinflammatory cytokines. Am J Respir Crit Care Med 162 (2000), 723-732. B. Lamkhioued, P.M. Renzi and S. Abi-Younes et al., increased expression of eotaxin in bronchoalveolar lavage and airways of asthmatics contributes to the chemotaxis of eosinophils to the site of inflammation. J Immunol 159 (1997), 4593-4601. G.A. Finlay, K.J. Russell and K.J. McMahon et al.,. Elevated levels of matrix metalloproteinases in bronchoalveolar lavage
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
[82]
[83]
[84]
[85]
[86]
fluid of emphysematous patients, Thorax 52 (1997), 502506. S.E. Wenzel, AA. Fowler 3rd and L.B. Schwartz, Activation of pulmonary mast cells by bronchoalveolar allergen challenge, in vivo release of histamine and tryptase in atopic subjects with and without asthma, Am Rev Respir Dis 137 (1988), 1002-1008. E. Lakari, P. Paakko and V.L. Kinnula, Manganese superoxide dismutase, but not CuZn superoxide dismutase, is highly expressed in the granulomas of pulmonary sarcoidosis and extrinsic allergic alveolitis, Am J Respir Crit Care Med 158 (1998), 589-596. K. Maier, L. Leuschel and U. Costabel, increased levels of oxidized methionine residues in bronchoalveolar lavage fluid proteins from patients with idiopathic pulmonary fibrosis, Am Rev Respir Dis 143 (1991), 271-274. K.L. Maier, L. Leuschel and U. Costabel, increased oxidized methionine residues in BAL fluid proteins in acute or chronic bronchitis, Eur Respir J 5 (1992), 651-658. A.G. Lenz, U. Costabel and K.L. Maier, Oxidized BAL fluid proteins in patients with interstitial lung diseases, Eur Respir J 9 (1996), 307-312. A.G. Lenz, P.G. Jorens and B. Meyer et al., Oxidatively modified proteins in bronchoalveolar lavage fluid of patients with ARDS and patients at-risk for ARDS, Eur Respir J 13 (1999), 169-174. J. Schildge, B. Klar and N. Weinstock, Concentration of alkaline phosphatase (AP) and AP/albumin ratio in bronchoalveolar lavage (BAL) for the diagnosis of interstitial lung diseases, Pneumologie 54 (2000), 385-391. G. Matute-Bello, W.C. Liles and K.P. Steinberg et al., Soluble Fas ligand induces epithelial cell apoptosis in humans with acute lung injury (ARDS), J Immunol 163 (1999), 22172225. K. Kuwano, M. Kawasaki and T. Maeyama et al., Soluble form of fas and fas ligand in BAL fluid from patients with pulmonary fibrosis and bronchiolitis obliterans organizing pneumonia, Chest 118 (2000), 451-458. J. Domagala-Kulawik, P. Droszcz, I. Kraszewska and R. Chazan, Expression of Fas antigen in the cells from bronchoalveolar lavage fluid (BALF), Folia Histochem Cytobiol 38(2000), 185-188. A.E. Calogero, R. Polosa, E. Neville and R. D'Agata, Measurements of hormonal peptides in the bronchoalveolar fluid as tumor markers of lung cancer, J Endocrinol invest 18 (1995), 354-358. V. Macchia, A. Mariano and M. Cavalcanti et al., Tumor markers and lung cancer: correlation between serum and bronchial secretion levels of CEA, TPA, CanAg CA-50, NSE and ferritin, Int J Biol Markers 2 (1987), 151-156. J. Niklinski, E. Chyczewska, M. Furman, E. Kowal, J. Laudanski and L. Chyczewski, Usefulness of a multiple biomarker assay in bronchoalveolar lavage (BAL) and serum for the diagnosis of small cell lung cancer, Neoplasma 40 (1993), 305-308. T.C. Pina, I.T. Zapata, F.C. Hernandez, J.B. Lopez, P.P. Paricio and P.M. Hernandez, Tumour markers in serum, bronchoalveolar lavage and biopsy cytosol in lung carcinoma: what environment lends the optimum diagnostic yield? Clin Chim Acta 305 (2001), 27-34. X.Y. Li, I. Rahman, K. Donaldson and W. MacNee, Mechanisms of cigarette smoke induced increased airspace permeability, Thorax 51 (1996), 465-471.
283
[87] R.E. Nocker, J.S. van der Zee, F.R. Weller, F.J. van Overveld, H.M. Jansen and T.A. Out, Segmental allergen challenge induces plasma protein leakage into the airways of asthmatic subjects at 4 hours but not at 5 minutes after challenge, J Lab Clin Med 134 (1999), 74-82. [88] A.S. Titanian, R.P. Selitskaia and A.C. Kim, Determination of the concentration of immunoglobulins and other proteins in bronchoalveolar fluid of patients with various types of pulmonary pathology, TerArkh 59 (1987), 48-51. [89] M. Lindahl, B. Stahlbom, J. Svartz and C. Tagesson, Protein patterns of human nasal and bronchoalveolar lavage fluids analyzed with two-dimensional gel electrophoresis, Electrophoresis 19 (1998), 3222-3229. [90] A.G. Lenz, B. Meyer, U. Costabel and K. Maier, Bronchoalveolar lavage fluid proteins in human lung disease: analysis by two-dimensional electrophoresis, Electrophoresis 14 (1993), 242-244. [91] R. Wattiez, C. Hermans, C. Cruyt, A. Bernard and P. Falmagne, Human bronchoalveolar lavage fluid protein twodimensional database: study of interstitial lung diseases, Electrophoresis 21 (2000), 2703-2712. [92] T. Plusa and Z. Wasek, Immunobiochemical evaluation of bronchoalveolar lavage (BAL) in atopic bronchial asthma, chronic bronchitis and bronchiectasis, Allergol 1mmunopathol (Madr) 15 (1987), 209-213. [93] K. Shigehara, N. Shijubo and M. Ohmichi et al., increased levels of interleukin-18 in patients with pulmonary sarcoidosis, Am J Respir Crit Care Med 162 (2000), 1979-1982. [94] M. Laan, I. Qvarfordt, G.C. Riise, B.A. Andersson, S. Larsson and A. Linden, increased levels of interleukin-16 in the airways of tobacco smokers: relationship with peripheral blood T lymphocytes, Thorax 54 (1999), 911-916. [95] D.S. Kim, Y.G. Jeon and T.S. Shim et al., The value of interleukin-12 as an activity marker of pulmonary sarcoidosis, Sarcoidosis Vase Diffuse Lung Dis 17 (2000), 271-276. [96] L. Borish, A. Aarons, J. Rumbyrt, P. Cvietusa, J. Negri and S. Wenzel, interleukin-10 regulation in normal subjects and patients with asthma, J Allergy Clin Immunol 97 (1996), 1288-1296. [97] T.L. Bonfield, J.R. Panuska and M.W. Konstan et al., inflammatory cytokines in cystic fibrosis lungs, Am J Respir Crit Care Med 152 (1995), 2111-2118. [98] M.G. Kiehl, H. Ostermann, M. Thomas, T. Birkfellner and J. Kienast, inflammatory mediators in BAL fluid as markers of evolving pneumonia in leukocytopenic patients, Chest 112 (1997), 1214-1220. [99] A. Itoh, E. Yamaguchi, K. Furuya and Y. Kawakami, Secretion of GM-CSF by inflammatory cells in the lung of patients with sarcoidosis, Respirology 3 (1998), 247-251. [100] M.S. Carraway, A.J. Ghio, J.D. Carter and C.A. Piantadosi, Detection of granulocyte-macrophage colony-stimulating factor in patients with pulmonary alveolar proteinosis, Am J Respir Crit Care Med 161 (2000), 1294-1299. [101] K. Jahnz-Rozyk, A. Chcialowski, E. Pirozynska and A. Rogalewska, Expression of adhesion molecules LFA-1 (CDlla and ICAM-1 (CD54) on lymphocytes and chemokines IL-8 and MCP-1 concentrations in bronchoalveolar lavage of patients with asthma or chronic obstructive pulmonary disease, Pol Merkuriusz Lek 9 (2000), 649-652. [102] B.D. Car, F. Meloni, M. Luisetti, G. Semenzato, G. GialdroniGrassi and A. Walz, Elevated IL-8 and MCP-1 in the bronchoalveolar lavage fluid of patients with idiopathic pulmonary fibrosis and pulmonary sarcoidosis, Am J Respir Crit Care Med 149 (1994), 655-659.
I. Noel-Georis et al. / Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage Y. Sugiyama, T. Kasahara, N. Mukaida, K. Matsushima and S. Kitamura, Chemokines in the bronchoalveolar lavage fluid of patients with sarcoidosis, Intern Med 36 (1997), 856-860. M. Ennis, G. Turner and B.C. Schock et al., inflammatory mediators in bronchoalveolar lavage samples from children with and without asthma, din Exp Allergy 29 (1999), 362366. R. Kalenderian, L. Raju, W. Roth, L.B. Schwartz, B. Gruber and A. Janoff, Elevated histamine and tryptase levels in smokers' bronchoalveolar lavage fluid. Do lung mast cells contribute to smokers' emphysema? Chest 94 (1988), 119123. S.M. Stahl, New drug discovery in the postgenomic era: from genomics to proteomics, J Clin Psychiatry 61 (2000), 894-895. G. Lubec. M. Nonaka, K. Krapfenbauer, M. Gratzer, N. Cairns and M. Fountoulakis, Expression of the dihydropyrimidinase related protein 2 (DRP-2) in Down syndrome and Alzheimer's disease brain is downregulated at the mRNA and dysregulated at the protein level, J Neural Transm Suppl 57 (1999), 161-177. A.A. Alaiya, B. Franzen, G. Auer and S. Linden Cancer proteomics: from identification of novel markers to creation of artifical learning models for tumor classification, Electrophoresis 21 (2000), 1210-1217. P.R. Jungblut, U. Zimny-Arndt and E. Zeindl-Eberhart et al., Proteomics in human disease: cancer, heart and infectious diseases, Electrophoresis 20 (1999), 2100-2110. C.P. McAtee, K.E. Fry and D.E. Berg, Identification of potential diagnostic and vaccine candidates of Helicobacter pylori by "proteome": technologies. Helicobacter 3 (1998), 163-169. P.R. Jungblut, D. Bumann and G. Haas et al., Comparative proteome analysis of Helicobacter pylori. Mol Microbiol 36 (2000), 710-725. S. Kennedy. Proteomic profiling from human samples: the body fluid alternative, Toxicol Lett 120 (2001), 379-384. A. Pandey and M. Mann, Proteomics to study genes and genomes, Nature 405 (2000), 837-846. D.Y. Bell and G.E. Hook, Pulmonary alveolar proteinosis: analysis of airway and alveolar proteins. Am Rev Respir Dis 119(1979), 979-990. R. Westermeier, W. Postel, J. Weser and A. Gorg, Highresolution two-dimensional electrophoresis with isoelectric focusing in immobilized pH gradients, J Biochem Biophys Methods 8 (1983). 321-330. A.G. Lenz. B. Meyer, H. Weber and K. Maier. Twodimensional electrophoresis of dog bronchoalveolar lavage
[117]
[118]
[119]
[120]
[121]
[122]
[123]
[124]
[125]
[126]
[127]
[128]
fluid proteins, Electrophoresis 11 (1990), 510-513. F.X. McCormack, T.E. Jr. King, D.R. Voelker, P.C. Robinson and R.J. Mason, Idiopathic pulmonary fibrosis. Abnormalities in the bronchoalveolar lavage content of surfactant protein A, Am Rev Respir Dis 144 (1991), 160-166. M. Lindahl, B. Stahlbom and C. Tagesson, Two-dimensional gel electrophoresis of nasal and bronchoalveolar lavage fluids after occupational exposure, Electrophoresis 16 (1995). 1199-1204. M. Lindahl, J. Svartz and C. Tagesson, Demonstration of different forms of the anti-inflammatory proteins lipocortin-1 and Clara cell protein-16 in human nasal and bronchoalveolar lavage fluids, Electrophoresis 20 (1999). 881-890. M. Lindahl, T. Ekstrom, J. Sorensen and C. Tagesson, Two dimensional protein patterns of bronchoalveolar lavage fluid from non-smokers, smokers, and subjects exposed to asbestos, Thorax 51 (19%), 1028-1035. B. Redl, P. Wojnar, H. Ellemunter and H. Feichtinger. Identification of a lipocalin in mucosal glands of the human tracheobronchial tree and its enhanced secretion in cystic fibrosis. Lab Invest 78 (1998), 1121 -1129. W. van't Hof, M.F. Blankenvoorde. E.C. Veerman and A.V. Amerongen, The salivary lipocalin von Ebner's gland protein is a cysteine proteinase inhibitor, J Biol Chem 272 (1997). 1837-1841. F. Ogushi, S. Sone and K. Tani et al., Identification and localization of immunoglobulin binding factor in bronchoalveolar lavage fluid from healthy smokers. Am J Respir Crit Care Med 152 (1995), 2133-2137. M. Kasper, P. Lackie, M. Haase, D. Schuh and M. Muller. Immunolocalization of cathepsin D in pneumocytes of normal human lung and in pulmonary fibrosis, Vtrchows Arch 428(19%). 207-215. B. Knoops, A. Clippe and C. Bogard et al., Cloning and characterization of AOEBI66, a novel mammalian antioxidant enzyme of the peroxiredoxin family. J Biol Chem 274 (1999). 30451-30458. M. Griese, C. von Bredow and P. Birrer, Reduced proteolysis of surfactant protein A and changes of the bronchoalveolar lavage fluid proteome by inhaled alpha 1-protease inhibitor in cystic fibrosis, Electrophoresis 22 (2001), 165-171. M. Merchant and S.R. Weinberger, Recent advancements in surface-enhanced laser desorption/ionization-time of flightmass spectrometry, Electrophoresis 21 (2000), 1164-1177. R.B. Rubin and M. Merchant, A rapid protein profiling system that speeds study of cancer and other diseases. Am Clin Lab 19 (2000). 28-29.
285
Pseudomonas aeruginosa and a proteomic
approach to bacterial pathogenesis Nicholas E. Sherman, Bjarki Stefansson, Jay W. Fox and Joanna B. Goldberg* Department of Microbiology, University of Virginia Health System, Charlottesville, VA 22908, USA Pseudomonas aeruginosa is a Gram-negative bacterium that is ubiquitous in the environment and can cause a variety of diseases in compromised patients. The genome of P.aeruginosa strain PAO1 has been reported to contain 5570 potential proteins. The value of this genomic database is that new proteins can be recognized to use as diagnostic markers, novel drug targets, and to better understand the physiology of this organism. However, similar to what has been observed in other sequenced bacterial genomes, approximately one third of the potential proteins have no known function. This is somewhat surprising given the long-standing interest in P. aeruginosa as an opportunistic pathogen. Obviously new tools, in addition to sequence similarity analysis, are needed to determine the role of these proteins. Proteomics using two-dimensional gel electrophoresis followed by mass spectrometry to detect and identify P. aeruginosa proteins represents a novel approach to address this gap.
1. Introduction Pseduomonas aeruginosa is an opportunistic pathogen that can cause acute infections in compromised patients including those undergoing chemotherapy, with burns, or with eye injury [1]. P. aeruginosa also causes chronic lung infections in patients with cystic fibrosis (CF); this chronic colonization is the major cause of death in these patients [2,3]. Unfortunately, this bacterium is naturally resistant to high levels of many commonly used antibiotics and is nutritionally diverse allowing it to compete and survive in a large variety of niches. The basis for the antibiotic resistance of P. aeruginosa is its lower outer membrane permeability and mechanisms such as inducible /3-lactamases and * Corresponding author. Tel.: +1 434 243 2774; Fax: +1 434 982 1071; E-mail:
[email protected]. Disease Markers 17 (2001) 285-293 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
antibiotic efflux pumps [4]. In addition to its inappropriate interactions with humans, P. aeruginosa is also a normal inhabitant of the environment, growing in aquatic ecosystems and on plants. P. aeruginosa is a well studied bacteria due to its ability to cause disease. This bacterium makes a large number of recognized virulence factors including lipopolysaccharide (LPS), polysaccharides, toxins, proteases, and other enzymes [5]. Some of these are bound to the cell surface, some are released, others are secreted, and still others are injected directly from this bacterium using a type III secretion apparatus. This bacterium can also change its phenotype: strains isolated initially from lung infections in CF patients have phenotypes similar to those from acute infections and from the environment. However, strains isolated from chronic lung infections make large amounts of an exopolysaccharide called alginate that is responsible for the characteristic mucoid phenotype. In addition, these chronic CF isolates express lower levels of extracellular enzymes and toxins, an altered lipid A, and a defective LPS [2]. In infections where it grows on surfaces, P. aeruginosa can form multicellular structures referred to as biofilms [6], the formation of which is dependent on quorum sensing [7]. Growth in this mode is responsible for increased resistance to host defenses and antibiotics. Many virulence factors have been noted by their effect in in vitro assays or in vivo infection models that mimic the human infectious process. Traditionally these factors have been identified by the construction of mutations in genes required for product production followed by their characterization in various models of P. aeruginosa infection. These include infection of neutropenic mice that reproduces infection after immunocompromise, the scratched mouse or rabbit eye model that mimics P. aeruginosa bacterial keratitis, the burned mouse model to simulate infection after thermal injury, and the mouse lung infection model which is patterned after initial infections in cystic fibrosis patients. Aspects of these models mirror steps likely to be important in infection, including avoidance of the immune sys-
286
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis
tem, adherence, invasion, and dissemination [5]. Mutagenesis techniques such as in vivo expression technology (IVET) have been developed for P. aeruginosa to identify genes that are induced specifically during infection [8]. Recently infection systems for P. aeruginosa have been developed that do not use animals. These include infections in plants [9], the nematode Caenorhabditis elegans [10,11], and insects [12,13]. Some P. aeruginosa virulence factors have been shown to be required in both mammalian and non-mammalian infection models while others were important in only one svstem.
2. Sequence analysis Because of the importance of P. aeruginosa as an opportunistic pathogen in both acute and chronic infections and for its environmental versatility, the sequence of the genome of the laboratory strain PAO1 was determined by the Pseudomonas Genome Project (www.pseudomonas.com) [14]. The release of this sequence represents a major breakthrough in the analysis of this important human pathogen. One of the largest bacterial genomes sequenced to date (6.3 million base pairs [Mbp]), the PAO1 genome sequence includes 5570 open reading frames (ORFs) encoding potential proteins. The homology of these ORFs to other proteins provides vital information on their proposed activities and should provide a framework to test these functions and determine their role in virulence. Proportionally, P. aeruginosa has the highest number of regulatory proteins of any bacteria sequenced so far [14,15]. This observation implicates these regulators in the ability of this organism to compete and survive in a variety of environments. It was surprising that nearly 50% of the "homology hits" of the P. aeruginosa ORFs were to those found in Escherichia coli; no other bacterium showed even 10% similarity. In addition, significant similarity in the order of genes between P. aeruginosa and E. coli was noted with increased genetic complexity in P. aeruginosa compared to E. coli. The greater complexity of the P. aeruginosa genome likely results in this bacterium's extreme metabolic diversity. While the sequence analysis of PAO1 provides scientists with an enormous amount of data on this organism, it is also appreciated that not all P. aeruginosa strains are PAO1. Overall approximately 3% of the DNA is unique between P. aeruginosa clinical isolates and PAO1. a number significantly lower than between
E. coli strains [16]. Pathogenic isolates of P. aeruginosa can contain an approximately 50 kb region of DNA not present in PAO1. This DNA may be considered a "pathogenicity island", which is an unstable region containing genes encoding products important for virulence or survival in specific environmental niches [17]. Strain PAO1 contains other DNA in this region [16]. With the release of the genome sequence, a genomewide mutagenesis approach has been developed for P. aeruginosa. This genomic analysis and mapping by in vitro transposition (GAMBIT) technology relies on allelic exchange mutagenesis to detect essential P. aeruginosa genes [18]. This functional genomic characterization should lead to the identification of essential gene products that may prove valuable drug targets. As mentioned, P. aeruginosa has been genetically well studied due to its importance as a human pathogen. Even so, similar to what has been found in other bacterial genome sequences, about 1/3 of the ORFs have no homology to any reported sequence. These potential gene products likely denote unique features of P. aeruginosa. Targeting these proteins represents a rational approach for making new drugs to specifically combat these bacteria. Similarly, ORF encoding conserved hypothetical proteins that are similar to potential gene products in other organisms may represent important non-P. aeruginosa-specific targets. However determining whether these ORF actually encode proteins and the functions of these gene products remains a challenge.
3. Proteomic analysis Genomic information coupled with protein analysis, referred to as proteomics, represents a new approach to address the gap in our understanding of previously undescribed gene products. Using the established techniques of two-dimensional (2-D) gel electrophoresis and mass spectrometry (MS) changes in protein expression or post-translational modification can be monitored between different samples, samples representing different physiological conditions, and other comparisons. 2-D gel electrophoresis separates protein mixtures by two techniques, isoelectric focusing in the first dimension and SDS-PAGE in the second dimension. The resulting gel provides a high-resolution separation of a complex mixture of proteins. The degree of staining of individual bands represents a quantitative measurement of the relative amounts of the protein, effectively providing a third dimension of information. The amino acid sequence of the selected proteins can be de-
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis
termined by MS and this information can be compared to any genetic databases including the Pseudomonas Genome Project database. Based on previous experience, use of information derived from the analysis of ORFs alone to calculate molecular weight (MW) and pI is moderately useful at best in assigning identity to proteins separated by 2-D gel electrophoresis; many proteins are products of processing/degradation or posttranslational modifications, which radically alter MW and/or pI. We have already taken advantage of the availability of the completed genome of P. aeruginosa PAO1 and the techniques of 2-D gel electrophoresis followed by MS. We have optimized procedures for the extraction of proteins from P. aeruginosa and found that DNase and RNase treatment followed by phenol extraction improved the resolution, separation, and reproducibility of the 2-D gel electrophoresis. This was true for both P. aeruginosa CF isolates as well as PAO1 and has allowed us to apply approximately 3 mg of protein onto each gel. After staining the gel with either Coomassie or silver, first pass mass mapping procedures were used. 2-D gel electrophoresis was coupled with Matrix Assisted Laser Desorption lonization Time-of-Flight (MALDI-TOF) mass spectrometry (Mass Mapping) to rapidly identify proteins. In addition, those proteins not identified by this method were automatically processed further through microcapillary column liquid chromatography tandem mass spectrometry (mLC/MS/MS) for identification or de novo sequencing. A local copy of MS-FIT was used to identify proteins against the P. aeruginosa genome or ORFs. Any proteins that fail to mass map were further analyzed by mLC/MS/MS and searched with the SEQUEST program (ThermoFinnigan) or BLAST (National Center for Biotechnology Information). Finally, any novel proteins or post-translationally modified peptides were de novo sequenced. We have detected variations in protein expression in P. aeruginosa strains with phenotypes corresponding to those initially and chronically infecting the lungs of a CF patient and identified 14 proteins during this analysis [19]. We noted proteins that were expressed at similar levels between the strains as well as those that were overexpressed in either strain. Of particular interest, we recognized that some protein spots represented specific degradation products. Whether this degradation is a differentially regulated event is currently under investigation. The eventual goal of this type of analysis is to determine how specific proteins may contribute to the observed phenotypic differences
287
between these strains. Changes in protein expression detected through this analysis may reflect changes that are induced by the CF lung environment, those that play a role in survival in the lung environment, or those that contribute to the activities of the organism that are responsible for its pathogenesis. Characterizing these differences will promote further studies on the function and regulation of these particular proteins, their role in virulence, and their potential as novel drug targets and as vaccine candidates. There are a few other groups that have looked at P. aeruginosa proteins by 2-D gel electrophoresis and other proteomic techniques. Quadroni et al. [20] have detected 13 proteins from P. aeruginosa that were induced during sulfate starvation. Most of these proteins were identified by N-terminal Edman and MS/MS sequencing. Three classes of proteins were identified including those for solute binding, sulfonate and sulfate ester metabolism, and general stress response. These were identified by homology and localized to sulfate controlled operons on the P. aeruginosa genome. The difference in expression levels of these proteins was confirmed by RT-PCR, which indicated that repression in the presence of sulfate was at the level of transcription. The value of this approach is that changes in protein expression under certain growth conditions can be monitored. This may help to suggest the function of unknown genes as well as steps in the response to varying environmental stimuli. Malhotra et al. [21 ] have detected protein differences between PAO1 and a mucoid mutant of PAO1, PDO300. This mucoid strain has a mucA22 mutation in the antisigma factor that controls the activity of sigma-22; this is the mutation that is responsible for the emergence of the mucoid phenotype and seen in most isolates from CF patients [22]. N-terminal sequencing of proteins after 2-D gel electrophoresis revealed increased expression of 6 proteins including AlgA, AlgD, DsbA, and OprF as well as decreased expression of 3 other proteins in PDO300 compared to PAO1 [21]. The genes encoding two of these proteins, algA and algD, are encoded by the alginate biosynthetic operon, and were known to have increased transcriptional activity in mucoid strains [23]. These authors confirmed that the transcription of dsbA, encoding a disulfide bond isomerase, was increased in PDO300 compared to PAO1, verifying the protein expression studies and suggesting the regulation by mucA22 [21]. Interestingly, oprF activity was not increased in PDO300. Whether this was due to degradation of OprF in PAO1 rather than overexpression in PDO300 as we observed for OprF in
288
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis
our analysis of initial and chronic CF isolates [19] was not determined. An "in house" resource to manage the enormous amount of data that can be generated from these sorts of analyses, called "The Microbial Proteomic Database", has been described [24]. This group has also isolated membrane fractions of PAO1 and identified about 189 spots on a 2-D gel [25]. These workers determined that the genes for 104 of these were present in the PAO1 genome. They designated these ORFs as "previously characterized in P. aeruginosa" (38%), those with "similarity to proteins in other organisms" (46%) and those with "unknown function" (16%). These authors also noted at least 15 proteins that may be glycosylated [25]. However how any of these proteins correspond to the genes of the Pseudomonas Genome Project was not reported. In an attempt to integrate the information available from the Pseudomonas Genome Project and our ability to detect proteins, we have randomly chosen 125 spots from a Coomassie-stained 2-D gel of soluble P. aeruginosa PAO1 proteins and performed mass mapping (Fig. 1 and Table 1). The phenol extraction, 2D gel electrophoresis, and processing of spots prior to mass spectrometry were accomplished as described previously [19]. Unlike the previous work, the digests were subjected to mass mapping (mass analysis alone without ammo acid sequence). After the tryptic digest, 20% of each sample was micropurified using a ZipTip U-C18 (Millipore, Bedford, MA) according to the manufacturer's instructions for direct elution with matrix. Peptide mass maps of each digest were generated by MALDI-TOF mass spectrometry using a PE Biosystems Voyager DE-PRO (PE Biosystems, Foster City, CA). The system was operated with the following parameters: reflective, positive ion, 20 kV extraction, 76% grid, 0.002% guide wire, 180 nsec delay, and CHC (alpha-cyano-4-hydroxycinnamic acid) matrix (8 mg/mL in 70% acetonitrile/0.1% trifluoroacetic acid). Spectra were acquired by averaging ~ 100 laser shots. Mass calibration was done using autodigestion tryptic peptides. Protein identifications were done automatically using PS 1/ Protein Prospector software (PE Biosystems, Foster City, CA) against the Pseudomonas ORF database from the Pseudomonas Genome Project (12/15/00; http://www.pseudomonas.com). For confident identification, parameters such as percent ion intensity matched, percent of ions matched, molecular mass, and pI were used. Those considered not matched, "none", did not sufficiently meet the above criteria upon manual inspection.
Out of the 125 spots chosen for analysis, 99 gave a dominant protein identification corresponding to an ORF in the PAO1 genome. In each case we identified the major protein component, however we acknowledge that many spots probably contain proteins at lower abundance and these were not identified in this analysis. We noted that for 85% of the proteins, the functions were defined in the PAO1 database annotation. These included membrane proteins, housekeeping proteins, regulators, and structural proteins. We recognized 6% of the spots as conserved hypothetical proteins indicating their identification in other organisms. Perhaps the most interesting group of proteins that we recognized was the 9% hypothetical proteins. As with the conserved hypothetical proteins, these proteins had only been identified as ORFs in the genome sequence. It was not previously known whether these genes were even translated into proteins. While the functions of these proteins are unknown, they are unique to P. aeruginosa suggesting a specific function for this bacterium. Uncovering the conditions under which these proteins are expressed should help to elucidate the functions of these gene products.
4. Future directions The post-genomic analysis of P. aeruginosa remains a formidable task. The resolution and detection of proteins by 2-D gel electrophoresis followed by MS is improving. In general, a silver-stained 2-D gel spot containing ~ 500 fmol of protein can be mass mapped (~ 0.5-1 mg total protein in the gel). In addition to 2-D gel electrophoresis, other methods of separating proteins have been developed. Isotope-coded affinity tags (ICAT) is a method to specifically label and compare proteins between samples [26]. A "heavy" (deuterium) ICAT reagent, with a biotin tag is used to label all cysteines in a sample. In another sample, a "light" (hydrogen) ICAT reagent is used. The samples are combined, fractionated, and proteolyzed. The ICAT-labeled peptides are isolated, quantified, and identified by MS/MS allowing comparisons of proteins levels between samples. The ICAT technique has been used to measure differences in protein expression in Saccharomyces cerevisiae grown under varying conditions [26]. Another method, multidimensional protein identification technology (MudPIT) is a technique to separate proteins by coupling 2-D LC with MS/MS; this has been used to detect and identify over 1400 proteins from S. cerevisiae [27]. The further standardization and optimiza-
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis
289
Fig. 1. 2-D gel of the major proteins of P. aeruginosa PAO1. Two milligrams of protein was focused in a nonlinear immobilized pH gradient strip of 3-10. After isoelectric focusing, the sample was run on a 10% SDS-PAGE gel and subsequently stained with Coomassie blue. The spots that were subjected to mass mapping analysis by MALDI-TOF mass spectrometry are numbered on the gel. Molecular masses were determined on a parallel gel; the sizes in kDa are indicated on the left of the gel. pi units were determined using a template provided by the manufacturer and are indicated on the bottom of the gel.
tion of both 1C AT and MudPIT should be applicable to proteomic investigations of P. aeruginosa. Proteomic studies ongoing for P. aeruginosa include the fractionation of PAO1 to identify those proteins secreted from the bacteria, those in the outer and inner membrane, periplasmic space and cytoplasm under standard laboratory conditions. This should allow the compartmentalization of components within PAO1. In addition, identifying surface proteins may provide new candidates for vaccine development. Subsequent studies include detection of PAO1 proteins during different growth phases, temperatures, and media, all conditions shown to alter protein expression. Conditions that mimic the biofilm mode of growth will be com-
pared to growth in liquid media. Hypothetical proteins expressed under specific conditions may have critical roles in these various circumstances. Defining the protein expression patterns may help suggest the function of these proteins. Protein expression in various PAO1 regulatory mutants will be compared to PAO1 to determine how proteins are controlled. Again any data obtained about regulation of hypothetical proteins will provide critical information concerning protein function. Proteins expressed from various P. aeruginosa clinical isolates will be compared to PAO1. Any differences between P. aeruginosa CF isolates and strains from other types of infection may suggest the importance of
290
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis Table 1 Proteomic identification of Pseudomonas aeruginosa PAOl proteins Spot#
I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
Source information
PA# _
MW -
pI -
4366 1777 3309 3723 5173 4495 1589 2765 1464
21.4 37.6 16.5 40.3 33.1 24.9 30.3 32.4 17.5
5.27 4.98 5.50 5.87 5.25 5.79 5.79 4.70 4.33
-
-
-
3397 4385
29.5 57.1
5.65 5.04
-
-
-
1562 4265 4277 4370 5171 5556 3014 4761 1588 1585 4595 4236 4502
99.1 43.4 43.4 47.2 46.4 55.4 77.0 68.4 41.5 105.9 61.3 55.6 58.6 62.7 45.0 38.1 43.3 51.2 47.8 50.1 36.4 22.0 33.1 20.4
5.43 5.23 5.23 4.68 5.52 5.34 5.67 4.79 5.83 6.10 5.47 6.21 6.16 7.75 6.07 6.13 5.84 8.57 8.92 7.74 8.45 5.81 8.30 9.69
519 2444 5172
447 3922
972 3001
283 4671 1342 4251
-
-
-
4352 4670
31.0 34.0
5.92 6.10
-
-
-
4352 4670 1584 3785
31.0 34.0 26.2 17.0
5.92 6.10 6.58 6.29
-
_
-
1010
31.5
6.00
-
_ -
-
1777 1777
37.6 37.6
4.98 4.98
-
-
-
4265 4277 4385
43.4 43.4 57.1
5.23 5.23 5.04
-
_
-
_
Name None Superoxide dismutase Outer membrane porin OprF Hypothetical (conserved) FMN oxidoreductase Carbamate kinase Hypothetical Succinyl-CoA synthetase alpha chain Hypothetical Purine-binding chemotaxis protein None Ferredoxin-NADP+ reductase GroEL protein None Aconitate hydratase 1 Elongation factor Tu Elongation factor Tu Hypothetical Arginine deimmase ATP synthase alpha chain Fatty-acid oxidation complex alpha subunit DnaK protein Succinyl-CoA synthetase beta chain 2-oxoglutarate dehydrogenase El subunit ATP-binding component of ABC transporter Catalase Binding protein component of ABC transporter Nitrite reductase Serine hydroxymethyltransferase Catabolic ornithine carbamoyltransferase Glutaryl-CoA dehydrogenase Hypothetical (conserved) TolB Glyceraldehyde-3-phosphate dehydrogenase Sulfate-binding protein Ribosomal protein L25 Binding protein component of ABC transporter 505 ribosomal protein L5 None Hypothetical (conserved) Ribose-phosphate pyrophosphokinase None Hypothetical (conserved) Ribose-phosphate pyrophosphokinase Succinate dehydrogenase B subunit Hypothetical (conserved) None Dihydrodipicolinate synthase None None Outer membrane porin OprF Outer membrane porin OprF None Elongation factor Tu Elongation factor Tu GroEL protein None None None
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis Table 1, continued Spot # 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
Source information PA# MW pI 5171 1588 1583 291 973 1337 1587 3570
46.4 41.5 63.5 49.7 17.9 38.6 50.2 53.7
5.52 5.83 6.04 8.67 5.95 6.67 6.48 5.98
2967
25.6
6.16
4938 5312 2553
46.8 53.1 41.4
5.70 5.41 5.74
4744 3257 5100 5110 5427
90.9 78.2 61.2 37.2 35.9
5.83 6.22 5.95 5.71 5.57
300
40.6
6.97
3349 3171 3569 745 3686 943 430 5200 1049 3004 5142 5557 1579
34.4 25.9 30.5 29.9 23.1 27.2 32.2 27.9 24.9 26.3 23.7 19.3 22.1
5.85 5.91 5.87 6.02 5.98 7.77 6.12 6.24 6.35 6.14 5.86 5.78 7.71
537 4314 4730 3313 4495 2204 3656 3692 1556 4495 5105 1556 3836 5505 1890 395 3244 -
22.2 32.4 30.8 36.5 24.9 29.3 27.3 28.5
8.62 6.04 6.11 9.62 5.79 9.06 8.54 9.45
22.8 24.9 27.8 22.8 34.2 28.1 23.4 38.0 29.6
7.73 5.79 6.80 7.73 7.75 7.79 6.43 6.39 5.58
Name None Arginine deiminase Succinyl-CoA synthetase beta chain None Succinate dehydrogenase A subunit Outer membrane porin OprE Outer membrane porin OprL Glutaminase-asparaginase None Lipoamide dehydrogenase-glc Methylmalonate-semialdehyde dehydrogenase None 4-oxoacyl-acyl-carrier protein reductase None Adenylsuccinate synthetase Aldehyde dehydrogenase Acyl-CoA thiolase None Translation initiation factor IF-2 Periplasmic tail-specific protease Urocanase Fructose-1,6-biphosphate Alcohol dehydrogenase None Polyamine transport protein None Chemotaxis protein 3-demethylubiquinone-9 3methyltransferase 3-hydroxyisobutyrate dehydrogenase Enoyl-CoA hydratase/isomerase Adenylate kinase Hypothetical 5,10-methylenetetrahydrofolate reductase Two-component response regulator OmpR Pyridoxine 5'-phosphate oxidase Nucleoside phosphorylase Glutamine amidotransferase ATP synthase delta chain Hypothetical None Hypothetical (conserved) Formyltetrahydrofolate deformylase Pantoate-beta-alanine ligase Hypothetical Hypothetical Binding protein component of ABC transporter 30S ribosomal protein S2 Outer membrane protein None None Cytochrome c oxidase subunit Hypothetical Histidine utilization represser HutC Cytochrome c oxidase subunit Hypothetical TonB-dependent receptor Glutathione S-transferase Twitching motility protein PUT Cell division inhibitor MinD None
291
292
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis Table 1, continued Spot # 117 118 119 120 121 122 123 124 125
Source PA# 4920 1342 314 5555 5167 3831 5261 -
information MW pI 29.7 5.42 8.30 33.1 9.11 28.0 7.69 31.6 37.1 8.36 8.68 52.3 6.72 27.6 -
Name NH3-dependent NAD synthetase None Binding protein component of ABC transporter Binding protein component of ABC transporter ATP synthase gamma chain C4-dicarboxylate-binding protein Leucine aminopeptidase Alginate biosynthesis regulatory protein None
1) Mass mapping identified 80% of the first 125 spots cored. "None" indicates that identification was not possible based on mass data alone. This could result from multiple proteins being present in the spot or too few peptides for an accurate match. In some cases, background keratin appeared to obscure relevant peptides to too great an extent. 2) "Hypothetical" indicates proteins that have not been identified previously or that cannot be identified by homology. "Hypothetical (conserved)" indicates proteins that have homology to only hypothetical ORFs in other organisms. 3) Bold MW or pI indicates significant deviation from theoretical values (those listed in the table) and those observed in the gel (those on figure).
particular proteins to a specific infectious process and indicate potential drug targets. In addition, P. aeruginosa proteins expressed during in vitro and in vivo infections can be followed. To further identify proteins important during infection, 2-D gel electrophoresis followed by Western immunoblots with patient sera will be used to detect which proteins are "seen" by the immune system. Antigens from the bacteria Helicobacter pylori have been identified in this manner; it is suggested that these may be useful for serological detection and monitoring infection [28]. The release of the P. aeruginosa genome coupled with microarray gene chip technology will help to define the transcriptional organization of this important pathogen. While this method would seem to make the proteomic characterization of P. aeruginosa superfluous, post-transcriptional modification and protein processing can not be determined by this transcriptome analysis. In fact, the correlation of gene expression and protein expression is often quite poor. Since protein localization and establishing environmental conditions regulating protein production are critical to understanding elaborate biological pathways and designing new drugs, the importance of the proteomic analysis should not be underestimated. Similarly, defining the patterns of protein expression should help to determine the function of unknown gene products. Likewise, investigating protein-protein interactions using techniques such as yeast two-hybrid analysis will provide a mechanism to integrate proteomics and protein complex formation [29]. Such genome-wide approaches will be essential to define the functions of unrecognized protein products.
5. Conclusions P. aeruginosa proteomic analysis allows the comparison of proteins expressed between bacterial strains or under varying growth or stress conditions. This examination should uncover proteins whose expression is critical in certain specific circumstances. This in turn will allow for the identification of patterns of protein expression that should help to define the functions of many hypothetical and conserved hypothetical proteins and provide a focus on those relevant in infection. This combination of genomic, proteomic. and informatic technologies applied to P. aeruginosa should rapidly advance the field, making rational and unique targets for drug design within closer reach. Acknowledgements This work was supported by grants from the NIH (R01-AI37632) and the Cystic Fibrosis Foundation (GOLDBE00G0 and GOLDBE00P0) to JBG. The W.M. Keck Biomedical Mass Spectrometry Laboratory and the Biomolecular Research Facility are funded by a grant from the University of Virginia Pratt Committee. We are grateful to Ms. Sheri Hanna and Dr. Michael Kinter for their technical support and helpful discussions concerning this work. References [1]
M. Pollack. Pseudomonas aeruginosa. in: Principles and Practice of Infectious Diseases. G.L. Mandell. J.E. Bennett
N.E. Sherman et al. / Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis
[2] [3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
and R. Dolin, eds, Churchill Livingstone, Philadelphia, 2000, pp. 2310-2335. G.B. Pier, Pseudomonas aeruginosa: a key problem in cystic fibrosis, Am Soc Microbiol News 64 (1998), 339-347. J.R.W. Govan and V. Deretic, Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia, Microbiologic Rev. 60 (1996), 539-574. R.E. Hancock, Resistance mechanisms in Pseudomonas aeruginosa and other nonfermentative gram-negative bacteria, Clin Infect Dis. 27(1) (1998), S93-S99. J.B. Lyczak, C.L. Cannon and G.B. Pier, Establishment of Pseudomonas aeruginosa infection: lessons from a versatile opportunist, Microbes Infect 2 (2000), 1051-1060. J.W. Costerton, PS. Stewart and E.P. Greenberg, Bacterial biofilms: a common cause of persistent infections, Science 284(1999), 1318-1322. D.G. Davies, M.R. Parsek, J.P. Pearson, B.H. Iglewski, J.W. Costerton and E.P. Greenberg, The involvement of cell-to-cell signals in the development of a bacterial biofilm, Science 280 (1998), 295-298. J. Wang, A. Mushegian, S. Lory and S. Jin, Large-scale isolation of candidate virulence genes of Pseudomonas aeruginosa by in vivo selection, Proc Natl Acad Sci USA 93 (1996), 10434-10439. L.G. Rahme, EJ. Stevens, S.F. Wolfort, J. Shao, R.G. Tompkins and F.M. Ausubel, Common virulence factors for bacterial pathogenicity in plants and animals, Science 268 (1995), 1899-1902. M.W. Tan, S. Mahajan-Miklos and F.M. Ausubel, Killing of Caenorhabditis elegans by Pseudomonas aeruginosa used to model mammalian bacterial pathogenesis, Proc Natl Acad Sci USA 96 (1999), 715-720. C. Darby, C.L. Cosma, J.H. Thomas and C. Manoil, Lethal paralysis of Caenorhabditis elegans by Pseudomonas aeruginosa, Proc Natl Acad Sci USA 96 (1999), 15202-15207. G. Jander, L.G. Rahme and F.M. Ausubel, Positive correlation between virulence of Pseudomonas aeruginosa mutants in mice and insects, J Bacteriol. 182 (2000), 3843-3845. D.A. D'Argenio, L.A. Gallagher, C.A. Berg and C. Manoil, Drosophila as a model host for Pseudomonas aeruginosa infection, J Bacteriol. 183(2001), 1466-1471. C.K. Stover, X.Q. Pham, A.L. Erwin, S.D. Mizoguchi, P. Warrener, M.J. Hickey, F.S.L. Brinkman, W.O. Hufnagle, D.J. Kowalik, M. Lagrou, R.L. Garber, L. Goltry, E. Tolentino, S. Westbrock-Wadman, Y. Yuan, L.L. Brody, S.N. Coulter, K.R. Folger, A. Kas, K. Larbig, R. Lim, K. Smith, D. Spencer, G.K.-S. Wong, Z. Wu, I.T. Paulsen, J. Reizer, M.H. Saier, R.E.W. Hancock, S. Lory and M.V. Olson, Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen, Nature 406 (2000), 959-964. A. Rodrigue, Y. Quentin, A. Lazdunski, V. Mejean and M. Foglino, Two-component systems in Pseudomonas aeruginosa: why so many? Trends Microbiol. 8 (2000), 498-503. X. Liang, X.-Q.T. Pham, M.V. Olson and S. Lory, Identification of a genomic island presnt in the majority of pathogenic isolates of Pseudomonas aeruginosa, J Bacteriol. 183 (2001), 843-853. B.B. Finlay and S. Falkow, Common themes in microbial
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
293
pathogenicity revisited, Microbiol Mol Biol Rev. 61 (1997), 136-169. S.M. Wong and J.J. Mekalanos, Genetic footprinting with mariner-based transposition in Pseudomonas aeruginosa, Proc Natl Acad Sci USA 97 (200), 10191-10196. S.L. Hanna, N.E. Sherman, M.T. Kinter and J.B. Goldberg, Comparison of proteins expressed by Pseudomonas aeruginosa strains representing initial and chronic isolates from a cystic fibrosis patient: an analysis by 2-D gel electrophoresis and capillary column liquid chromatography-tandem mass spectrometry, Microbiol. 146 (2000), 2495-2508. M. Quadroni, P. James, P. Dainese-Hatt and M.A. Kertesz, Proteome mapping, mass spectrometric sequencing and reverse transcription-PCR for characterization of the sulfate starvation-induced response in Pseudomonas aeruginosa PAO1, Euro J Biochem. 266 (1999), 986-996. S. Malhotra, L.A. Silo-Suh, K. Mathee and D.E. Ohman, Proteome analysis of the effect of mucoid conversion on global protein expression in Pseudomonas aeruginosa PAO1 shows induction of disulfide bond isomerase, DsbA. J Bacteriol. 182 (2000), 6999-7006. J.C. Boucher, H. Yu, M.H. Mudd and V. Deretic, Mucoid Pseudomonas aeruginosa in cystic fibrosis: characterization of muc mutations in clinical isolates and analysis of clearance in a mouse model of respiratory infection, Infect Immun. 65 (1997), 3838-3846. C.E. Chitnis and D.E. Ohman, Genetic analysis of the alginate biosynthetic gene cluster of Pseudomonas aeruginosa shows evidence of an operonic structure, Mol Microbiol. 8 (1993), 583-593. S.J. Cordwell, A.S. Nouwens, N.M. Verrills, J.C. McPherson, P.G. Hains, D.D. Van Dyk and B.D. Walsh, The microbial proteome database-an automated laboratory catalogue for monitoring protein expression in bacteria, Electrophoresis 20 (1999), 3580-3588. A.S. Nouwens, S.J. Cordwell, M.R. Larsen, M.P. Molloy, M. Gillings, M.D.P. Willcox and B.J. Walsh, Complementing genomics and proteomics: the membrane subproteome of Pseudomonas aeruginosa PAO1, Electrophoresis 21 (2000), 37973809. S.P. Gygi, B. Rist, S.A. Gerber, F. Turecek, M.H. Gelb and R. Aebersold, Quantitative analysis of complex protein mixtures using isotope-coded affinity tags, Nat Biotechnol. 17 (1999), 994-999. M.P. Washbum, D. Wolters and J.R. Yates III, Large-scale analysis of the yeast proteome by multidimensional protein identification technology, Nat Biotechnol. 19(2001), 242-247. C.P. McAtee, M.Y. Lim, K. Fung, M. Velligan, K. Fry, T. Chow and D.E. Berg, Identification of potential diagnostic and vaccine candidates of Helicobacter pylori by two-dimensional gel electrophoresis, sequence analysis, and serum profiling, Clin Diagn Lab Immunol. 5 (1998), 537-542. J.C. Rain, L. Selig, H. De Reuse, V. Battaglia, C. Reverdy, S. Simon, G. Lenzen, F. Petel, J. Wojcik, V. Schachter, Y. Chemama, A. Labigne and P. Legrain, The protein-protein interaction map of Helicobacter pylori, Nature 409 (2001), 211-215.
This page intentionally left blank
295
A proteomic approach to the identification of lung cancer markers Samir Hanasha,*, Franck Brichorya and David Beerb a
Department of Pediatrics, University of Michigan Medical Center, Ann Arbor, MI 48109, USA b Department of Surgery, University of Michigan Medical Center, Ann Arbor, MI 48109, USA We have developed a comprehensive approach to the identification of protein markers in lung cancer that includes profiling of tumor tissue and cell lines as well as the analysis of serum for autoantibodies to lung tumor antigens. A large number of proteins that are differentially expressed in the major subtypes of lung cancer have been identified by mass spectrometry. A database of protein expression in lung cancer and other types of cancer has been constructed that integrates two-dimensional gel profiles, mass spectrometry data, quantitative protein data and gene expression data at the RNA level, that serves as a resource for biomarker identification. Analysis of the serological response in lung cancer has led to the identification of novel markers detectable in serum of lung cancer patients at the time of diagnosis. The proteomic approach is likely to yield novel classification schemes and novel markers for early diagnosis of lung cancer. Keywords: Lung, cancer, proteomics, markers
1. Introduction We have implemented a comprehensive strategy for the molecular analysis of lung cancer that includes proteomic analysis of lung tumor tissue and cell lines, as well as serum analysis for the identification of lung tumor proteins that induce a serological response in the form of autoantibodies. Although we have relied to date primarily on two-dimensional (2-D) polyacry-
* Address for correspondence: Samir Hanash, University of Michigan Medical Center, 1150 W. Medical Center Drive, A520 Medical Science Research Building I, Ann Arbor, MI 48109-0656, USA. Tel.: +1 734 763 9311; Fax: +1 734 647 8148; E-mail: shanash® umich.edu. Disease Markers 17 (2001) 295-300 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
lamide gels for protein separations, the 2-D gel approach is being increasingly complemented with additional analyses using liquid based protein separations and protein microarrays. Proteomic analysis of tissues and cell populations uniquely contributes an understanding of protein post-translational modifications and of the distribution of protein gene products in subcellular compartments. An important objective of our lung cancer effort is the identification of novel markers for early detection. A large number of studies involving lung cancer have been independently performed in the laboratory. At the protein level, these studies have resulted in over 1000 samples related to lung cancer, that have been processed using 2-D gels and for which information has been recorded in the Lung Protein Database. This number represents a fraction of over 30,000 2-D Gels produced by our group for different studies, including studies of other cancer types. While lung adenocarcinomas represent a major portion of the lung cancer database, other lung tumor types including squamous cell carcinomas and small cell lung cancers are represented, as are control lung tissues. Other 2-D patterns were produced from studies of cell lines that have been manipulated by transfection or by treatment with specific agents, as well as patterns produced after different cell fractionation schemes. Mass Spectrometry and/or N-terminal sequencing of protein spots from 2-D gels of lung tumor samples or cell lines has led to the identification of a large number of proteins expressed in lung cancer. Also, most identifications made for proteins from a sample type can often be confidently transferred to matching protein spots on master images from lung studies. Figure 1 exhibits some of the progress we have made in identifying proteins in 2-D gels of lung samples. An important first step in mining lung cancer proteomic data for various applications is to determine the ability of proteomic profiling to distinguish between known types of lung cancer. Specific protein differences between different types of cancer have been identified by other groups. In a recent study of breast, ovary
296
S. Hanash et al. / A proteomic approach to the identification of lung cancer markers
and lung tumors, 20 differentially expressed proteins were identified [ 1 ] and in a prior study, 16 polypeptides were found to be associated with different histopathological features of lung cancer [2,3]. In a study of 25 adenocarcinomas of the lung, 12 small cell lung cancers, and 16 squamous cell tumors, by our group (manuscript submitted) an initial analysis of protein 2-D patterns uncovered a group of 52 protein spots that differed in average integrated intensity between the three groups in a statistically significant manner. We have identified 39 of this set of 52 spots by either Nterminal sequencing and/or mass spectrometry of spot digests.
2. Serological approaches for the identification of lung cancer markers There is increasing evidence for an immune response to cancer in humans, demonstrated in part by the identification of autoantibodies against a number of intracellular and surface antigens detectable in sera from patients with different cancer types [4-7]. The majority of tumor derived antigens that have been identified as eliciting a humoral response in lung cancer, as in other tumor types, are not the products of mutated genes. They include differentiation antigens and other proteins that are overexpressed in tumors [8]. The oncogenic proteins L-Myc and C-Myc have been found to elicit autoantibodies in a small percentage of patients [5.9J. There is some evidence that occurrence of autoantibodies in lung cancer is of prognostic relevance [10-15]. Remarkably, tumor regression has been demonstrated in some patients with small-cell lung carcinoma and autoantibodies to onconeural antigens [15,16]. It is not clear why only a subset of patients with a tumor type develop a humoral response to a particular antigen. Immunogenicity may depend on the level of expression, post-translational modification or other types of processing of a protein, the extent of which may be variable among tumors of a similar type. Other factors that influence the immune response may include variability among individuals and tumors in major histocompatibility complex molecules. Cytokines are also known to affect the immune response and may vary in concentration between tumors or in circulation [17-19]. The identification of panels of tumor antigens that elicit an antibody response may have utility in cancer screening, diagnosis or in establishing prognosis. Such antigens may also have utility in immunotherapy against the disease. There are several approaches for
the detection of tumor antigens that induce an immune response. A number of antigens have been detected by screening expression libraries with patient sera [4-6. 20-22]. The merits of our proteomic approach is that it allows proteins, in their modification states as they occur in cells, to be analyzed for their antigenicity. Given that proteins are subject to post-translational modifications, antibodies to epitopes that result from such posttranslational modifications can be detected. Additionally, the 2-D approach allows for serial serum samples to be analyzed much more readily than the screening of expression libraries. We have implemented a proteomic approach for the identification of tumor antigens that elicit a humoral response [23-25]. To this end. we have utilized 2-D PAGE to simultaneously separate several thousand individual cellular proteins from tumor tissue or tumor cell lines. Separated proteins are transferred onto membranes. Sera from cancer patients are screened individually, for antibodies that react against separated proteins, by Western blot analysis. Proteins that specifically react with sera from cancer patients are identified by mass spectrometric analysis and/or amino acid sequencing and further evaluated with respect to their specificity.
3. A combined serological and proteomic-based approach to the identification of novel lung cancer markers We have identified using our proteomic approach a battery of proteins that induce autoantibodies that are specific for different types of cancer including some that show specificity for lung cancer. The availability of a database of protein expression in lung cancer has facilitated the identification of proteins that induce autoantibodies. in addition to providing valuable information regarding the expression pattern of such protein antigens in different tumor types and cell lines. One such antigen we have identified in lung cancer is protein PGP 9.5 (Fig. 2) [24]. PGP 9.5 was identified as a protein in lung cancer that induces autoantibodies as part of a study in which sera from 64 newly diagnosed patients with lung cancer, from 99 patients with other types of cancer and from 71 non-cancer controls were analyzed for antibody-based reactivity against lung adenocarcinoma proteins resolved by 2-D PAGE. Gels containing separated proteins were blotted and subsequently hybridized with individual sera from patients or controls. Unlike controls, autoantibodies against a protein iden-
S. Hanash et al. / A proteomic approach to the identification of lung cancer markers
297
Fig. 1. A small cell lung cancer master image with identified proteins.
tified by mass spectrometry as protein gene product 9.5 (PGP 9.5) were detected in sera from 9 of 64 patients with lung cancer. Circulating PGP 9.5 antigen was detected in sera from two additional patients with lung cancer, without detectable PGP 9.5 autoantibodies. PGP 9.5 was first identified as a specific marker for neurons and neuroendocrine cells [26]. PGP 9.5 belongs to a family of ubiquitin C-terminal hydrolase (UCH) isoenzymes that play a regulatory role in the ubiquitin system [27].
It has been implicated in the mechanism to remove ubiquitin from ubiquitinated proteins and thus preventing their degradation by proteasomes [28]. Ubiquitination of cellular proteins and their targeting for subsequent degradation via ubiquitin-mediatedproteolysis is an important mechanism that regulates the activity of a variety of genes, notably cell cycle genes [27,29]. In lung tumors, increased de-ubiquitination of cyclins by PGP 9.5 may contribute to uncontrolled proliferation [28]. In our study we demonstrated by 2-D PAGE
298
S. Hanash et al. / A proteomic approach to the identification of lung cancer markers
Fig. 2. A 2-D image of a lung adenocarcinoma. Boxed areas containing annexins I and II and PGP 9.5 proteins are presented in Fig. 3.
and Western blot analyses that 80% of the lung tumors we have studied contained detectable levels of PGP 9.5. In previous analyses by immunohistochemistry, PGP 9.5 was detected in 40-82.5% of NSCLC and 50-90% of SCLC [30-33]. Hibi et al reported ectopic expression of PGP 9.5 in lung cancers by SAGE analysis and by immunochemistry [31,34]. In primary NSCLCs, 54% of the cases had positive PGP 9.5 staining, and protein expression was associated with pathological stage (44% of stage I and 75% of stages II and IIIA). PGP 9.5 was observed in both SCLC and NSCLC cell lines, independent of neuronal differentiation. Using A549 lung adenocarcinoma cell line, we have demonstrated that PGP 9.5 was present at the cell surface, as well as secreted. Thus, the findings of PGP 9.5 antigen and/or antibodies in serum of patients with lung cancer suggest that PGP 9.5 may have utility in lung cancer screening and diagnosis, as part of a panel of such proteins or their corresponding antibodies, which we have identified. In another study, sera from 54 newly diagnosed patients with lung cancer and 60 patients with other cancers and from 61 non-cancer controls were analyzed for autoantibodies to lung tumor proteins. Sera from 60% of patients with lung adenocarcinoma, and 33% of patients with squamous cell lung carcinoma but none of the non-cancer controls exhibited IgG based reactivity against proteins identified as glycosylated annexins I
and/or II. Imiminohistochemical analysis showed that annexin I was diffusely expressed in neoplastic cells in lung tumor tissues, whereas annexin II was predominant at the cell surface. Annexin I is a 37 kDa protein which has been implicated in glucocorticoid induced inhibition of cell growth [35,36]. Annexin II is a 36 kDa protein that occurs in a monomeric form or as a tetramer, associated with the annexin II light chain (p 11), which is a member of the S100 family [37,38]. Annexin II has been implicated in cell-cell adhesion and in plasminogen activation and may function as a cell surface receptor [39]. Annexin II tetramers have been shown to interact with procathepsin B on the surface of tumor cells and may be involved in extracellular proteolysis, facilitating tumor invasion and metastasis [40]. Interestingly, annexin I is a target of autoantibodies in autoimmune diseases such as systemic lupus erythematosus [41,42] and rheumatoid arthritis [43]. Annexin II, specifically, has not been previously implicated as a target of autoantibodies in any disorders. Annexins are known to undergo post-translational modification including glycosylation [44]. Annexin I and annexin II are both phosphorylated by various kinases [45]. In our study, immunoreactivity against annexin I was found to be dependent on N-glycosylation. A potential N-linked glycosylation site is present at positions 42 and 61 from the N-terminus of annexins I and
S. Hanash et al. / A proteomic approach to the identification of lung cancer markers
299
for cancer diagnosis. The approach is currently being expanded in several ways. First, proteomic analysis is increasingly focused on the analysis of individual sub-cellular compartments such as surface membranes, nuclear proteins etc. Second the 2-D gel based approach is increasingly complemented with other separation modes such as the use of multi-dimensional liquid chromatography that is particularly suited for automation and for the analysis of small molecular weight proteins and peptides. Other emerging technologies include the use of protein microarrays in which antigen or antibody, representing probes, is deposited onto glass surfaces and interrogated with targets represented by reagents, sera, biological fluids or cell or tissue lysates. One would envisage the development of specialized microarrays to interrogate targets relevant to lung cancer. Proteomics is likely to contribute substantially to our understanding of the pathophysiology of lung cancer and to the development of novel diagnostics and therapeutics.
References Fig. 3. Boxed areas containing annexins I and II and PGP 9.5 proteins shown in Fig. 2. Panel A: Location of PGP 9.5 forms P1, P2 and P3 in a silver stained gel. Panel B: Reactivity of a lung adenocarcinoma patient serum with forms P1-P3. Panel C: Location of annexins I and II in a silver stained gel. Panel D: Reactivity of anti-annexin I monoclonal antibody with form identified as annexin I by mass spectrometry. Panel E: Reactivity of anti-annexin II monoclonal antibody with form identified as annexin II by mass spectrometry.
II, respectively [46,47]. Glycosylation may contribute to protein stability and may enhance signal transduction [44]. Interestingly, IL-6 levels were significantly higher in sera of antibody-positive lung cancer patients compared with antibody-negative patients and controls. This led us to conclude that an immune response manifested by annexins I and II autoantibodies occurs commonly in lung cancer and is associated with high circulating levels of an inflammatory cytokine.
4. Conclusion The initial proteomic approach we have implemented was based on the analysis of whole tissue or whole cell lysates. We have demonstrated that this approach has utility for the identification of differentially expressed proteins and for the development of serum based assays
[1] A.C. Bergman et al., Identification of gel-separated tumor marker proteins by mass spectrometry, Electrophoresis 21 (2000), 679-686. [2] T. Hirano et al., Detection of polypeptides associated with the histopathological differentiation of primary lung carcinoma, Br. J. Cancer 72 (1995), 840-848. [3] H.R. Schmid et al., Lung tumor cells: a multivariate approach to cell classification using two-dimensional protein pattern, Electrophoresis 16 (1995), 1961-1968. [4] A. Yamamoto et al., Detection of auto-antibodies against Lmyc oncogene products in sera from lung cancer patients, Int J Cancer 22 (1996), 283-289. [5] E. Stockert et al., A survey of the humoral immune response of cancer patients to a panel of human tumor antigens, J Exp Med 187 (1998), 1349-1354. [6] A.O. Gure et al., Human lung cancer antigens recognized by autologous antibodies: Definition of a novel cDNA derived from the tumor suppressor gene locus on chromosome 3p21.3, Cancer Res 58 (1998), 1034-1341. [7] T. Soussi, p53 Antibodies in the sera of patients with various types of cancer: a review, Cancer Res 60 (2000), 1777-1788. [8] A. Yamamoto et al., Infrequent presence of anti-c-Myc antibodies and absence of c-Myce oncoprotein in sera from lung cancer patients, Oncology 56(2) (1999), 129-133. [9] P. Maddison et al., Favourable prognosis in Lambert-Eaton myasthenic syndrome and small-cell lung carcinoma, Lancet 353(9147) (1999), 117-118. [10] F. Blaes et al., Antineural and antinuclear autoantibodies are of prognostic relevance in non-small cell lung cancer, Ann. Thorac. Surg 69(1) (2000), 254-258. [11] Y. Hirasawa et al., Natural autoantibody to MUC1 is a prognostic indicator for non-small cell lung cancer, Am J Respir Crit Care Med 161 (2000), 589-594.
S. Hanash et al. / A proteomic approach to the identification of lung cancer markers R.B. Darnell and L.M. DeAngelis, Regression of small-cell lung carcinoma in patients with paraneoplastic neuronal antibodies, Lancet 341(8836) (1993), 21-22. R.B. Darnell, Onconeural antigens and the paraneoplastic neurologic disorders: at the intersection of cancer, immunity, and the brain, Proc. Natl. Acad. Sci. USA 93(10) (1996), 45294536. H. Yanagawa et al., Serum levels of interleukin 6 in patients with lung cancer, Br J Cancer 71(5) (1995), 1095-1098. F. Martin et al., Cytokine levels (IL-6 and IFN-gamma), acute phase response and nutritional status as prognostic factors in lung cancer, Cytokine 11(1) (1999), 80-86. J.R. Strahler, R. Kuick and S.M. Hanash, Two-dimensional polyacrylamide gel electrophoresis of proteins, in: Protein Structure: A Practical Approach. T. Creighton, ed.. IRL Press Ltd., Oxford, 1989, pp. 65-92. F. Gharahdaghi et al.. Mass spectrometric identification of proteins from silver-stained polyacrylamide gel: a method for the removal of silver ions to enhance sensitivity. Electmphoresix 20(3) (1999), 601-605. K.C. Gatter and C.L. Oakley lecture. Diagnostic immunocytochemistry: achievements and challenges. J Pathol 159(3) (1989), 183-190. C. Asselin-Paturel et al., Quantitative analysis of Thl, Th2 and TGF-betal cytokine expression in tumor, TIL and PEL of non-small cell lung cancer patients, Int J Cancer77(1) (1998), 7-12. M.M. Gourevitch et al., Polymorphic epithelial mucin (MUC1 )-containing circulating immune complexes in carcinoma patients. Br. J. Cancer 72 (1995). 934-938. T. Soussi, The humoral response to the tumor-suppressor gene product p53 in human cancer: Implications for diagnosis and therapy. Immunol. Today 17 (1996), 354-356. L.J. Old and Y.T. Chen, New paths in human cancer serology, JExpMed 187 (1998), 1163-1167. L. Prasannan et al., Identification of beta-tubulin isoforms as tumor antigens in neuroblastoma, Clin. Cancer Res. 6(10) (2000), 3949-3956. F. Brichory et al., Proteomics-based identification of PGP 9.5 as a tumor antigen that induces a humoral immune response in lung cancer. Cancer Res., in press. P.M. Brichory et al., An immune response manifested by common occurrence of annexins I and II autoantibodies and high circulating levels of interleukin 6 in lung cancer, submitted. I.N. Day and R.J. Thompson, Molecular cloning of cDNA coding for human PGP 9.5 protein. A novel cytoplasmic marker for neurones and neuroendocrine cells. Febs Lett. 210(2) (1987), 157-160. A. Hershko and A. Ciechanover, The ubiquitin system. Ann. Rev. Biochem. 67 (1998), 425-479. V. Spataro, C. Norbury and A.L. Harris, The ubiquitinproteasome pathway in cancer, Br. J. Cancer 77(3) (1998). 448-455. M. Hochstrasser, Ubiquitin-dependent protein degradation. Anna. Rev. Genet. 30 (1996). 405-439.
f30]
[31] [32] [33] [34] [35]
[ 36]
[37]
[38]
[39]
[40]
[41 ]
[42]
[43]
[44]
[45]
[46] [47]
B.J. Addis et al., Immunohistochemical markers of small ceil carcinoma and related neuroendocrine tumours of the lung. 7. Pathol. 153(2) (1987), 137-150. K. Hibi et al., Serial analysis of gene expression in non-small cell lung cancer. Cancer Res. 58(24) (1998), 5690-5694. G. Abbona et al., Chromogranin A gene expression in nonsmall cell lung carcinomas, / Pathol. 186(2) (1998). 151-156. A.P. Dhillon et al., Neural markers in carcinoma of the lung. Br. J. Cancer 51(5) (1985). 645-652. K. Hibi et al., PGP9.5 as a candidate tumor marker for nonsmall-cell lung cancer. Am. J. Pathol. 155( 3) (1999). 711 -715. J. Mai et al.. Human procathepsin B interacts with annexin II tetramer on the surface of tumor cells. J Biol Chem 275( 17) (2000), 12806-12812. W. Pruzanski et al., Circulating group II phospholipase A2 activity and antilipocortin antibodies in systemic lupus erythematosus. Correlative study with disease activity. / Rheiunatol. 21(2X1994), 252-257. M.R. Podgorski et al., Autoantibodies to lipocortin-1 are associated with impaired glucocorticoid responsiveness in rheumatoid arthritis. J. Rheumatol. 19(11)(1992), 1668-1671. F. Goulet. K.G. Moore and A.C. Sartorelli. Glycosylation of annexin I and annexin II. Biochem. Bioph\s. Res. Commun. 188(2X1992). 554-558. T. Dubois et al., Annexins and protein kinases C. Biochim. Biophys. Acta 1313(3) (1996). 290-294. B.P. Wallner et al.. Cloning and expression of human lipocortin. a phospholipase A2 inhibitor with potential antiinflammatory activity. Nature 320(6057) (1986). 77-81. T. Kristensen et al.. Primary structure of bovine calpactin I heavy chain (p36), a major cellular substrate for retroviral protein-tyrosine kinases: homology with the human phospholipase A2 inhibitor lipocortin, Biochem 25( 16X 1986). 44974503. A. Gure et al., Serological identification of embryonic neural proteins as highly immunogenic tumor antigens in small cell lung cancer. Proc. Natl. Acad. Sci. USA 97(8) (2000). 41984203. A. Dowlati, N. Levitan and S.C. Remick, Evaluation of interleukin-6 in bronchoalveolar lavage fluid and serum of patients with lung cancer. J Lab Clin Med 134(4) (19991. 405409. E. Solito et al.. IL-6 stimulates annexin 1 expression and translocation and suggests a new biological role as class II acute phase protein, Cytokine 10(7X1998). 514-521. K. Li et al.. Tumour-specific MHC-class-II-restricted responses after in vitm sensitization to synthetic peptides corresponding to gplOO and Annexin II eluted from melanoma cells, Cancer Immunol Immunother 47(1)( 1998), 32-38. U. Boehm et al., Cellular responses to interferon-gamma. Ann. Rev. Immunol. 15 (1997). 749-795. M.D. Johnson et al., Lipocortin-1 immunoreactivity in central and peripheral nervous system glial tumors. Hum. Pathol. 20(81(1989). 772-776.
301
Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF: Potential for new biomarkers to aid in the diagnosis of breast cancer Cloud P. Paweletza,b,c, Bruce Trockd, Marie Pennanene, Theodore Tsangarise, Collette Magnantf, Lance A. Liottab and Emanuel F. Petricoin IIIa,* a
Tissue Proteomics Unit, Division of Therapeutic Proteins, CBER, Food and Drug Administration, Bethesda, MD 20892, USA b Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA c Department of Chemistry, Georgetown University, Washington, DC 20057, USA d Department of Urology, Johns Hopkins University, Baltimore, MD 21287, USA e Department of Surgery, Georgetown University, Washington, DC 20007, USA f Department of Surgery, Sibley Hospital, Washington, DC 20010, USA
Nipple aspirate fluid (NAF) has been used for many years as a potential non-invasive method to identify markers for breast cancer risk or early detection. Because individual markers have not been optimal, we are exploring the use of surface enhanced laser desorption and ionization time of flight (SELDITOF) mass spectrometry to identify patterns of proteins that might define a proteomic signature for breast cancer. SELDITOF was used to analyze a study set of NAF samples that included 12 women with breast cancer and 15 healthy controls (the latter included three women with an abnormal mammogram but subsequent normal biopsy). In this preliminary report, we present data showing that SELDI analysis of NAF is rapid, reproducible, and capable of identifying protein signatures that appear to differentiate NAF samples from breast cancer patients and healthy controls, including those with an abnormal mammogram who were later proven to be biopsy normal. Disease Markers 17 (2001) 301-307 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
1. Introduction Mammographic screening for breast cancer is currently the best available approach for early detection in the general population. However, additional approaches are needed. Current mammography is associated with a sensitivity of 75%-90% [1], but the positive predictive value at 25% is very low [2,3]. Although resolution continues to improve, mammography is still dependent on the existence of mass lesions. Because many breast tumors will already have metastasized by the time a mass is detectable [4], a significant portion of mammographically detected tumors in women undergoing regular screening will already be disseminated and incurable. Furthermore, mammography alone may not be sufficient for early detection in premenopausal women, particularly for young women with inherited susceptibility or other high risk profiles because of following reasons: (a) effectiveness of mammography has not been established in women younger than 40, and is controversial in women aged 40-49, (b) younger women have more dense breast tissue, which reduces mammographic sensitivity, (c) tumor growth rates may be higher in younger women, and (d) women carriers of some germ-line mutations such as ataxia telangiectasia and possibly BRCA1/2 may have increased sensitivity to radiation and conceivably could be harmed by frequent mammograms [5-7]. More widespread testing of young women for germline predisposing mutations (such as BRCA1/2) will soon result in thousands of young women at high risk for breast cancer, for whom conventional screening approaches may be inadequate. Thus, there is an
"Corresponding author. Tel: +1 301 827 1753; Fax: +1 301 480 3256; E-mail:
[email protected].
302
C.P. Paweletz et al. / Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF
Fig. 1. SELDI-TOF spectra of nipple aspirate fluids are reproducible, a) Three representative SELDI-TOF chromatograms between 2500 Da and 10000 Da from one case are shown. Good reproducibility between individually runs is achieved. Enlargement of mass ranges of proteins between 8000 Da and 10000 Da reveal additional individually resolved proteins that are reproducible. Chromatograms can be changed to density maps that represent the spectra as a gel view with each peak being a differently 'stained' band whose density is proportional to its height, b) Reproducibility of two different cases.
urgent need for additional methods of early detection that can provide an adjunct to mammography. An alternative to imaging technologies for breast cancer detection is to examine easily accessible biological fluids for evidence of molecular signatures associated with neoplastic changes. Ideally such an approach should reflect the relevant biology of breast epithelium, be relatively non-invasive, easy for the patient, and conducive to serial monitoring. Nipple aspirate fluid (NAF) may be promising for direct sampling of breast epithelial biology. This fluid is secreted continuously by the non-lactating breast and can be aspirated through duct openings in the nipple using a simple non-invasive pump. Despite many years of study, however, no single marker in NAF has been shown to have high sensitivity and specificity for breast cancer [8-14]. Most molecular based approaches that have been investigated for early detection are targeted at specific defects, such as oncogenes, tumor suppressor genes, growth factors, tumor antigens, or other gene products. The inherent problem is that none of these factors alone are present in a large majority of breast cancers, and some are not specific to cancer or to breast tissue, so the sensitivity and specificity of such approaches is low. Rather than targeting a specific abnormality that may
only be present in a small subgroup of patients, we propose to identify general patterns of protein expression that distinguish the invasive phenotype from preinvasive or normal tissue. Because such patterns could represent pathways commonly altered during neoplastic progression, it is highly likely that multiple elements of the pathway would be present in a majority of breast tumors, even if some individual protein alterations vary from case to case. Surface-enhanced laser desorbtion time of flight spectrometry (SELDI-TOF) is a highly sensitive, specific, and high throughput technology platform for the study of whole protein lysates [15]. Protein lysates are directly applied onto a chromatographic surface (i.e. hydrophobic. hydrophilic. cationic, or anionic). washed, and subsequently analyzed by time of flight mass spectrometry. Thus, the obtained chromatogram only presents a protein phenotype of shared common chemical or physical characteristics. SELDI-TOF differs from conventional MALDI-TOF spectrometry in that it does not require initial separation of complex biological mixtures by high performance liquid chromatography (HPLC) or gas chromatography (GC) [ 16. 17]. The speed and ease of SELDI-TOF makes this technology very adaptable for use in the clinical laboratory.
C.P. Paweletz et al. / Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF
303
Fig. 2. Nipple aspirate fluids are marked by a high degree of heterogeneity. Representative gel view images of nipple aspirate fluids of three patients that were biopsy proven to be normal, but showed abnormal mammograms, six uninvolved breasts, and six involved breast are shown.
In this report, we used SELDI-TOF to generate protein profiles from NAF's of 27 patients (12 women with breast cancer and 15 healthy women; the latter included three women with repeated abnormal mammograms for whom subsequent biopsies were normal) to test the hypothesis that there exist low molecular weight proteomic signatures that can potentially distinguish breast cancer patients from healthy individuals.
2. Patients and methods Study participants came from three sources: the Lombardi Cancer Center's Comprehensive Breast Center, the Georgetown University Breast Surgery clinic, and the office of a breast surgeon in private practice (CM). NAF was obtained using well-established methods [18] that we have successfully employed in previous studies. A modified breast pump is used, comprised of a finely polished glass cup with a Luer-lok end attached directly to a 20 mL slip-tip syringe. The cup is placed over the cleansed nipple of the breast, and the woman compresses her breast with both hands while the plunger of the syringe is withdrawn to 10 mL and held for 8-10 seconds. With this technique we have ob-
tained NAF from 50-70% of premenopausal women. Droplets of fluid that appear at any duct openings on the nipple are collected into capillary tubes. Three attempts of 8-10 seconds each were made to obtain fluid; if no fluid appears after three attempts the woman was considered a non-secretor. All fluid from an individual breast was pooled and expelled into 100 mL of PBS. Samples from each breast were stored separately at —70°C prior to analysis. SELDI-TOF spectra were obtained as previously described [15]. Briefly, 1 mL of acetonitrile was added onto a C-16 aliphatic chip (Ciphergen, Palo Alto, CA) followed by 1 mL of NAF. The specimen was allowed to incubate under a humid atmosphere for at least 5 minutes after which the chip was thoroughly washed with water. The chip was air dried and 2 * 0.5 mL of 3,5-dihydroxy cinnamic acid (diluted in 50% Acetonitrile, 0.5% TFA) was co-crystallized. The instrument was operated in linear mode. Three profiles totaling 35 shots were obtained from 3 of the 100 addressable regions for a given spectrum. All multiparametric analyses were performed using PEAKS 2.0.1 (Ciphergen, Palo Alto, CA).
304
C.P. Paweletz et al. / Proteomic patterns of nipple aspirate fluids obtained by SELD1-TOF
Fig. 3. Generation of Normal and Tumor specific tumor plates, a) SELDI-TOF spectra from 12 tumor and 15 normal NAF samples were run (in triplicates, with three representative GelViewTM spectra shown) and compared to each other. Proteins that were within 0.75% of each other were considered to be the same molecular identity and thereby placed into the protein template. Protein identities that were not present in all analyzed cases were deemed not suitable for diagnosis purposes. Four proteins were present in NAFs of normal and cancerous breasts at 3415.6 Da and 4149.7 Da, as well as 4233.09 Da and 9470.0 Da, respectively.
3. Results Previous experiments showed that body fluids such as saliva (unpublished data) and serum (manuscript submitted) are compatible with SELDI-TOF and yield very reproducible protein patterns under 20 kDa without the need to purify the samples by HPLC or size exclusion chromatography. These results are in line with previously published reports that show that MALDITOF spectra of serum are richer in ion species in molecular mass ranges that are smaller than 30 kDa [15,19]. While protein patterns of NAF's at higher molecular weights (up to 150 kDa) can easily be obtained the diagnostic utility diminishes rapidly due to well known ion suppression effects (data not shown). Coefficients of variances for inter and intra runs were previously shown to be less than 10% [15]. Concurrently, using an aliphatic C16 chip yields very reproducible spectra between 2 kDa and 15 kDa for NAFs with an average of over 50 individually resolved proteins in the mass range. Each sample was run in triplicates to ensure that proteins seen during individual runs were indeed
reproducible (Fig. 1). Day to day or instrument variations can be excluded as samples run on different days yielded identical spectra (data not shown). However, in contrast to serum and saliva chromatograms that do not show large variations (data not shown), NAFs, while being reproducible, are marked by unusual large variations in their spectra between different samples within a group (Fig. 2). Confronted with this observation, we decided against measuring individual protein abundance levels, but rather decided to generate a protein template that presents only the proteins that are consistently associated with cancerous or normal breasts. This was achieved by comparing all proteins and peptides that are greater than two standard deviations above background of all cancerous breasts and all normal NAF specimens, respectively. Proteins that were within 0.75% mass accuracy of each other were considered the same and thus compared with the next respective sample. A protein profile for normal and tumor NAF samples is shown in Fig. 3. The dramatic reduction in proteins compared to those detected without the above criteria i.e. from more than 50 pro-
C.P. Paweletz et al. / Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF
305
Fig. 4. Protein Profiles from patient matched NAFs from involved an uninvolved breasts. SELDI spectra were taken between 2500 Da and 15000 Da of patient matched NAFs with one of the breast being biopsy proven cancerous, and the other being normal. As expected, most of the proteins did not change, however protein changes track with generated template (arrow).
teins to 10 or fewer) is indicative of the large heterogeneous nature of these samples with respect to protein expression. However, two proteins were uniquely present in tumor-associated samples (at 4233.09 m/z and 9470.0 m/z) and two were uniquely associated with normal samples (at 3415.6 m/z and 4149.7 m/z). This template was tested in a variety of cases for whom NAFs were obtained from both the cancerous and contralateral breast (Fig. 4). As expected most proteins do not change between the involved and the contralateral uninvolved breast and appear very different from the generated templates; however, the individual identified proteins always tracked with the disease state. A powerful application of this technology would be as an adjunct to improve specificity of the diagnostic paradigm in women with abnormal mammograms. A protein template of NAFs from three women with abnormal mammograms was generated, and protein templates from the cancerous and normal breasts were compared (Fig. 5). All three of these women were biopsied and found to have normal breasts. The chromatogram
obtained from women with abnormal mammograms is distinct from that of both the normal and tumorassociated specimens. However, the protein identity at 4152.3 Da that is present in normal SELDI-TOF template, but not in the cancerous template, can be found. As the mass accuracy of the machine is between 0.75% and 1% the protein identities at 4152.3 Da and 4149.7 Da are likely to be the same. Furthermore, each protein identity that is present in the tumor template is not present in the abnormal mammogram spectra (Fig. 5). These results indicate that spectra obtained by SELDI may aid in the decision making process for patients that show an abnormal mammogram.
4. Discussion The need for rapid, inexpensive, noninvasive approaches to improve the specificity of breast cancer screening paradigms has led to the search for tumor markers in readily accessible body fluids. However, the search for such individual markers have been dis-
306
C.P. Paweletz et al. / Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF
Fig. 5. SELDI analysis of NAFs from biopsy proven normal patients with abnormal mammograms. SELDI templates of three patients mat were biopsy proven to be normal, but showed abnormal mammograms were compared to the normal and cancerous template. While providing an unique SELDI protein pattern in itself tumor unique proteins (4233.09 Da and 9470.0 Da) were absent, and proteins that were specific to normal samples (4197.9 Da) were present in NAFs of breasts that showed abnormal mammograms.
appointing [20,21]. We [15] and others [16,22,23] are utilizing rapid proteomic approaches that can easily be implicated to the clinic and aid in the diagnosis decision making. Toward this end, our laboratory currently employs two approaches utilizing affinity SELDI-TOF to identify breast cancer protein signatures in NAF and serum. Such a signature will be a pattern of protein expression that consistently is associated with cancer and discriminates it from samples from women with normal breasts. The simplest approach to identifying such a signature is based on identifying low molecular weight proteins and peptides (< 10.000 Da) that are present in either 100% of tumor-associated or normalassociated samples. The discovery of multiple proteins and patterns of protein expression that are consistently altered may increase the specificity and sensitivity of an eventual diagnosis. Here, we report the use of SELDI-TOF to generate protein profiles from NAF's obtained from 27 women (12 with breast cancer and 15 without breast cancer). This study group also included three women with ab-
normal mammograms but found to be normal at biopsy. Two peptides at 4233.09 Da and 9470.0 Da were consistently associated with tumor samples, while two peptides at 3415.6 Da and 4149.7 Da consistently appeared in normal samples, and thus may be of diagnostic importance. Moreover, overall protein patterns obtained from NAFs from women with abnormal mammograms did not clearly resemble protein patterns from normal or tumor patients. However, all three cases contained a 4152.3 Da that showed resemblence to the 4149.9 Da. Furthermore, no proteins that are unique to tumor samples were found in the abnormal cases. While matrix assisted laser desorption spectrometry (MALDI-TOF) has been used to analyze known tumor markers these studies are not really amenable to clinical settings [23,24]. The speed and ease of SELDI-TOF to generate protein patterns without the need to purify the specimen prior to application makes this technology very adaptable for use in the clinical laboratory. We believe that advances in this technology will allow
C.P. Paweletz et al. / Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF
clinical investigators to monitor proteomic patterns for diagnosis or prognosis.
[13]
[14]
References [15] [1] W.L. Donegan, Evaluation of a palpable breast mass,N' Engl JMed 327 (1992), 937-942. [2] J.R. Harris, M.E. Lippman, U. Veronesi and W. Willett, Breast cancer, N EnglJ Med327 (1992), 319-328. [3] J.G. Elmore, M.B. Barton, V.M. Moceri, S. Polk, P.J. Arena and S.W. Fletcher, Ten-year risk of false positive screening mammograms and clinical breast examinations, N Engl J Med 338 (1998), 1089-1096. [4] B. Fisher, The evolution of paradigms for the management of breast cancer: a personal perspective, Cancer Res 52 (1992), 2371-2383. [5] M. Swift, Ionizing radiation, breast cancer, and ataxia_telangiectasia, J Natl Cancer Inst 86 (1994), 15711572. [6] S.K. Sharan, M. Morimatsu and U. Albrecht et al., Embryonic lethality and radiation hypersensitivity mediated by Rad51 in mice lacking Brca2, Nature 386 (1997), 804-810. [7] R. Scully, J. Chen and R.L. Ochs et al., Dynamic changes of BRCA1 subnuclear location and phosphorylation state are initiated by DNA damage, Cell 90 (1997), 425-435. [8] M. Wrensch and N.L. Petrakis et al., Breast cancer risk associated with abnormal cytology in nipple aspirates of breast fluid and prior history of breast biopsy, Am J Epidemiol 137 (1993), 829-833. [9] E.R. Sauter, E. Ross and M. Daley et al., Nipple aspirate fluid: A promising non-invasive method to identify cellular markers of breast cancer risk, Br J Cancer 76 (1997), 494-501. [10] Y. Liu, J.L. Wang, H. Chang, S.H. Barsky and M. Nguyen, Breast-cancer diagnosis with nipple fluid bFGF, Lancet 356 (2000), 567. [11] Y. Zhao, J.V. Sigitas, N. Klar, N.L. Sadowsky, C.M. Kaelin and B. Smith, Nipple fluid carcinoembryonic antigen and prostate antigen in cancer bearing and tumor-free breasts, J Clin Oncol 19 (2001), 1462-1467. [12] E.R. Sauter, M. Daly and K. Linahan et al., Prostate Specific antigen levels In nipple aspirate fluid correlate with breast cancer risk, Cancer Epideminol Biomarkers Prev 5 (1996), 967-970.
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
307
E.B. King, K.L. Chew, N.L. Petrakis and V.L. Ernster, Nipple apirate cytology for the study of breast cancer precursor, J Natl Cancer Instil (1983), 1115-1121. L. Foretova, J. Garber and N. Sadowsky et al., Carcinoembryonic antigen in breast nipple spirate fluid, Cancer Epidemiol Biomarkers Prev 7 (1998), 195-198. C.P. Paweletz, J.W. Gillespie and D.K. Ornstein, Rapid protein display profiling of cancer progression directly from human tissue using a protein biochip, Drug Dev Res 49 (2000), 34-42. T. Nakanishi, A. Shimizu, N. Okamoto, A. Ingendoh and M. Kanai, Analysis of serum protein precipitated with antiserum by matrix-assisted laser desorption ionization/time of flight and electrospray ionization mass spectrometry as a clinical laboratory test, J Am Soc Mass Spectrom 6 (1995), 854-859. P. Chaurand, M. Stoeckli and R.M. Caprioli, Direct Profiling of proteins in biological tissue sections by MALDI mass spectrometry, Anal Chem 71 (1999), 5263-5270. N.L. Petrakis, ASO distinguished achievement award lecture. Studies on the epidemiology and natural history of benign breast disease and breast cancer using nipple aspirate fluid, Cancer Epidemiol Biomarkers Prev 2 (1993), 3—10. L. Ferrari, R. Seraglia, C.R. Rossi, A. Bertazzo, M. Lise and A. Graziella et al., Protein profiles in sera of patients with malignant cutaneous melanoma, Rapid Com in Mass Spec 14 (2000), 1149-1154. R.B. Dickson, J.A. Low, M.D. Johnson, M.J. Hawkins and BJ. Track, Blood-borne indicators of breast cancer and their use in experimental, medical oncologic, and prevention studies, The Breast 5 (1996), 379-384. M.A. Harding and D. Theodorescu, Prognostic markers in localized prostate cancer: from microscopes to molecules, Cancer and Met Reviews 17 (1999), 429-37. J. Khan, J.S. Wei, M. Ringer, L.H. Saal, M. Ladanyi and F. Westermann et al., Classification and diagnostic prediction of cancers using gene expression profiling and artifical neural networks, Nature (2001), 673-679. J.M. Lacey, H.R. Bergen, M.J. Magera, S. Naylor and J.F. O'Brien, Rapid determination of transferrin isoforms by immunoaffinity liquid chromatography and electrospray mass spectrometry, Clin Chem 47 (2001), 513-518. A. Lapolla, D. Fedele, R. Aronica, M. Garbeglio, M. Alpaos, R. Seraglia and P. Traldi, Evaluation of IgG Glycation levels by matrix-assisted laser desorption/ionization mass spectrometry, Rapid Commun Mass Spectrom 11 (1997), 1342-1346.
This page intentionally left blank
309
Author Index Volume 17 (2001) The issue number is given in front of the pagination
Alvarez, S., M.S. Mesa, F. Bandres and E. Arroyo, C282Y and H63D mutation frequencies in a population from central Spain (2) 111-114 Arroyo, E., see Alvarez, S. (2) 111-114 Bagni, R., see Mikovits, J. (3) 173-178 Bandres, F., see Alvarez, S. (2) 111-114 Beer, D., see Hanash, S. (4) 295-300 Ben-Dor, A., see Chen, Y. (2) 59-65 Beretta, L., see Shalhoub, P. (4) 217-223 Bernard, A., see Noel-Georis, I. (4) 271-284 Bittner, M., see Chen, Y. (2) 59-65 Block, T., see Steel, L.F. (3) 179-189 Bonnet-Duquenoy, M., see Touitou, R. (3) 163-165 Brentani, R.R., see Caballero, O.L. (2) 67-75 Brichory, F, see Hanash, S. (4) 295-300 Bubendorf, L., see Srivastava, M. (2) 115-120 Butel, J.S., Increasing evidence for involvement of SV40 in human cancer (3) 167-172 Butler, E.B., see Miller, J.C. (4) 225-234 Caballero, O.L., S.J. de Souza, R.R. Brentani and A.J.G. Simpson, Alternative spliced transcripts as cancer markers (2) 67-75 Chang, YE. and L.A. Laimins, Interferon-inducible genes are major targets of human papillomavirus type 31: Insights from microarray analysis (3) 139-142 Chanock, S., Candidate genes and single nucleotide polymorphisms (SNPs) in the study of human disease (2) 89-98 Chen, Y, Z. Yakhini, A. Ben-Dor, E. Dougherty, J.M. Trent and M. Bittner, Analysis of expression patterns: The scope of the problem, the problem of scope (2) 59-65 Dorjsuren, D., see Mikovits, J. (3) 173-178 Dougherty, E., see Chen, Y. (2) 59-65 Dwek, R., see Steel, L.F. (3) 179-189 Eidelman, O., see Srivastava, M. (2) 115-120 Disease Markers 17 (2001) 309-311 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
Evans, A.A., see Steel, L.F. (3) 179-189 Falmagne, P., see Noel-Georis, I. (4) 271-284 Fehrle, W., see Srivastava, M. (2) 115-120 Fox, J.W., see Sherman, N.E. (4) 285-293 Garcea,R.L., SV40: A human pathogen? (3) 149-151 Girard, S., see Shalhoub, P. (4) 217-223 Glasman, M., see Srivastava, M. (2) 115-120 Goldberg, J.B., see Sherman, N.E. (4) 285-293 Haab, B.B., see Miller, J.C. (4) 225-234 Hanash, S., F. Brichory and D. Beer, A proteomic approach to the identification of lung cancer markers (4)295-300 Hanash, S., see Srivastava, S. (4) 203-204 Hebestreit, H., see Steel, L.F. (3) 179-189 Herrmann, P.C., L.A. Liotta and E.F. Petricoin III, Cancer proteomics: The state of the art (2) 49-57 James, P., Protein expression analysis: From 'tip of the iceberg' to a global method (4) 235-246 Joab, I., see Touitou, R. (3) 163-165 Kallioniemi, O.P., see Srivastava, M. (2) 115-120 Karlsen, A.E., T. Sparre, K. Nielsen, J. Nerup and F. Pociot, Proteome analysis - A novel approach to understand the pathogenesis of Type 1 diabetes mellitus(4)205-216 Kern, S., see Shalhoub, P. (4) 217-223 Khalili, K., Human neurotropic JC virus and its association with brain tumors (3) 143-147 Knebel Doeberitz, von M., New molecular tools for efficient screening of cervical cancer (3) 123-128 Kutkat, L. and S. Srivastava, The Early Detection Research Network: A platform for communication and collaboration (1) 3-4 Laimins, L.A., see Chang, YE. (3) 139-142 Lambert, P.P., see Verma, M. (3) 191-201 Leighton, X., see Srivastava, M. (2) 115-120
310
Author Index Volume 17 (2001)
Lemoine, F.J., D.R. Wycuff and S.J. Marriott, Transcriptional activity of HTLV-I Tax influences the expression of marker genes associated with cellular transformation (3) 129-137 Liotta, L.A., see Herrmann, P.C. (2) 49-57 Liotta, L.A., see Paweletz, C.P. (4) 301-307 London, W.T., see Steel, L.F. (3) 179-189 Madden, C.R. and B.L. Slagle, Stimulation of cellular proliferation by hepatitis B virus X protein (3) 153-157 Magnant, C., see Paweletz, C.P. (4) 301-307 Marriott, S.J., see Lemoine, F.J. (3) 129-137 Mattu, T.S., see Steel, L.F. (3) 179-189 Mehta, A., see Steel, L.F. (3) 179-189 Mesa, M.S., see Alvarez, S. (2) 111-114 Mikovits, J., F. Ruscetti, W. Zhu, R. Bagni, D. Dorjsuren and R. Shoemaker, Potential cellular signatures of viral infections in human hematopoietic cells (3) 173-178 Miller, G., see Srivastava, M. (2) 1 15-120 Miller, J.C., E.B. Butler, B.S. Teh and B.B. Haab, The application of protein microarrays to serum diagnostics: Prostate cancer as a test case (4) 225-234 Nerup, J., see Karlsen, A.E. (4) 205-216 Nielsen, K., see Karlsen, A.E. (4) 205-216 Noel-Georis, I., A. Bernard, P. Falmagne and R. Wattiez, Proteomics as the tool to search for lung disease markers in bronchoalveolar lavage (4) 271284 Nolan. L., see Srivastava, M. (2) 115-120 Paweletz, C.P., B. Trock, M. Pennanen, T. Tsangaris, C. Magnant, L.A. Liotta and E.F. Petricoin III, Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF: Potential for new biomarkers to aid in the diagnosis of breast cancer (4) 301-307 Pennanen, M., see Paweletz, C.P. (4) 301-307 Perou, C.M., see Ross, D.T. (2) 99-109 Petricoin III, E.F, see Paweletz, C.P. (4) 301-307 Petricoin III, E.F, see Herrmann, P.C. (2) 49-57 Pociot, F, see Karlsen, A.E. (4) 205-216 Pollard, H.B., see Srivastava, M. (2) 115-120 Raffeld, M., see Srivastava, M. (2) 1 15-120 Riggins, G.J., Using Serial Analysis of Gene Expression to identify tumor markers and antigens (2) Rollison, D.E.M., see Shah, K.V. (3) 159-161
Ross, D.T. and C.M. Perou, A comparison of gene expression signatures from breast tumors and breast tissue derived cell lines (2) 99-109 Ruscetti. F, see Mikovits, J. (3) 173-178 Shah, K.V. and D.E.M. Rollison, Investigation of the SV40 - human cancer association: Look for the full signature of the virus (3) 159-161 Shalhoub, P., S. Kern, S. Girard and L. Beretta. Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma (4) 217-223 Sherman, N.E., B. Stefansson, J.W. Fox and J.B. Goldberg, Pseudomonas aeruginosa and a proteomic approach to bacterial pathogenesis (4) 285-293 Shimohama, S., see Tsuji, T. (4) 247-257 Shoemaker, R., see Mikovits, J. (3) 173-178 Simpson, A.J.G., see Caballero, O.L. (2) 67-75 Slagle, B.L., see Madden, C.R. (3) 153-157 Souza, de S.J., see Caballero, O.L. (2) 67-75 Sparre, T., see Karlsen, A.E. (4) 205-216 Srivastava, M., L. Bubendorf, L. Nolan, M. Glasman. X. Leighton, G. Miller, W. Fehrle, M. Raffeld, O. Eidelman, O.P. Kallioniemi, S. Srivastava and H.B. Pollard, ANX7 as a bio-marker in prostate and breast cancer progression (2) 115-120 Srivastava, S. and S. Hanash, Global Strategies for Disease Detection and Treatment: Proteomics (4) 203-204 Srivastava, S.. see Kutkat, L. (1) 3-4 Srivastava. S.. see Srivastava, M. (2) 115-120 Srivastava, S.. see Verma, M. (3) 121-122 Srivastava, S.K., see Verma, M. (3) 191-201 Steel. L.F, T.S. Mattu, A. Mehta, H. Hebestreit. R. Dwek, A.A. Evans, W.T. London and T. Block, A proteomic approach for the discovery of early detection markers of hepatocellular carcinoma (3) 179-189 Stefansson. B.. see Sherman. N.E. (4) 285-293 Teh. B.S.. see Miller, J.C. (4) 225-234 Thalmann, I., Proteomics and the inner ear (4) 259-270 Touitou, R., M. Bonnet-Duquenoy and I. Joab, Association of Epstein-Barr virus with human mammary carcinoma. Pros and cons (3) 163-165 Trent, J.M., see Chen, Y. (2) 59-65 Trock, B., see Paweletz, C.P. (4) 301-307 Tsangaris. T.. see Paweletz, C.P. (4) 301-307 Tsuji. T. and S. Shimohama, Analysis of the proteomic profiling of brain tissue in Alzheimer's disease (4) 247-257
Author Index Volume 17 (2001)
311
Verma, M. and S. Srivastava, Molecular Signatures of Infectious Agents in Human Cancer (3) 121-122 Verma, M., P.F. Lambert and S.K. Srivastava, Meeting highlights: National Cancer Institute workshop on Molecular Signatures of Infectious Agents (3) 191-201
Weinstein, J.N., Searching for pharmacogenomic markers: The synergy between omic and hypothesisdriven research (2) 77-88 Wycuff, D.R., see Lemoine, F.J. (3) 129-137
Wattiez, R., see Noel-Georis, I. (4) 271-284
Zhu, W., see Mikovits, J. (3) 173-178
Yakhini, Z., see Chen, Y. (2) 59-65
This page intentionally left blank
313
Keyword Index Volume 17 (2001) Alzheimer's disease
247
benign paroxysmal positional vertigo
259
markers MedMiner Meniere's disease microarray molecular marker mutation
295 77 259 77 77 89
omics oncomodulin organ of Corti otoconia
77 259 259 259
cancer cancer therapy cell line clustered image map cochlea
49, 77, 295 77 77 77 259
database disease susceptibility
247 89
gap junctions genetic genome genomics
259 89 89 77
human cancers
159
signal transduction SV40
49 159
Internet
247
two-dimensional gel electrophoresis
247
laser capture microdissection lung
49 295
variation
Disease Markers 17 (2001) 313 ISSN 0278-0240 / $8.00 © 2001, IOS Press. All rights reserved
perilymphatic pharmacogenomics pharmacology proteomics
fistula
259 77 77 49, 77, 259, 295
89
/OS
Press IOS Press is a rapidly expanding Scientific, Technical, Medical and Professional publishing house focusing on a broad range of subject areas, such as; medical science, healthcare, telecommunication, artificial intelligence, information and computer science, parallel computing, physics and chemistry, environmental science, and other subjects. IOS Press Nieuwe Hemweg 6B 1013 BG Amsterdam The Netherlands Tel.:+ 31 206883355 Fax:+ 31 206203419 IOS Press/Lavis Marketing 73 Lime Walk Headington Oxford OX3 7AD United Kingdom Fax.:+44 1865 75 0079 IOS Press, Inc. 5795-G Burke Centre Parkway Burke, VA 22015, USA Tel.: + 1 7033235554 Fax.: + 1 703 323 3668 E-mail:
[email protected] /OS Press/LSLde Gerichtsweg 28 D-04103 Leipzig, Germany Tel.: +49341 9954250 Fax.:+ 49 341 9954255 Ohmsha, Ltd. 3-1 Kanda Nishiki-cho Chiyoda-ku Tokyo 101, Japan Fax.: + 813 3233 2426
Journals Online IOS Press is pleased to be able to announce a new service, the IOS Press journals are now available electronically via the IOS Press website. As a subscriber to an IOS Press journal you can have free electronic access to the articles published in this journal. Electronic access is free with a print subscription. If you wish to access the system through IP access, please let us know and send us your institutes' IP ranges. We will set up access for these IP ranges and send you a confirmation e-mail. It is also possible to arrange access for your institution/university/library through Http Referrer, for this option, please send us your http details and we will set up access and send you a confirmation. The service on the IOS Press site is far more than only electronic access to journal articles. It is possible to search and browse journals and tables of contents, you can specify your favorite publications in order to facilitate quicker access to them, select publications for which you wish to receive notifications of new issues, instruct the system to automatically resubmit specified searches on newly published content and deliver notifications when relevant articles are found. Many of the above mentioned services are free to use, even if you are not a subscriber to the journal. Furthermore, individual articles can be purchased without being a subscriber to a journal. For an overview of all options and possibilities of this new system, please visit our website and use this service at: http://iospress.metapress.com or e-mail
[email protected]
Disease Markers Subscription information Disease Markers (ISSN 0278-0240) is published in one volume of six issues a year. The subscription price for 2002 (Volume 18) is EUR 480 + EUR 36 p.h. = EUR 516 (US$ 473). The Euro price is definitive. The US dollar price is subject to exchange-rate fluctuations and is given only as a guide. 6% VAT is applicable for certain customers in the EC Countries. Subscriptions are accepted on a prepaid basis only, unless different terms have been previously agreed upon. Personal subscription rates and conditions, if applicable, are available upon request from the Publisher. Subscription orders can be entered only by calendar year (Jan.–Dec.) and should be sent to the Subscription Department of IOS Press, or to your usual subscription agent. Postage and handling charges include printed airmail delivery to countries outside Europe. Claims for missing issues must be made within six months of our publication (mailing) date, otherwise such claims cannot be honoured free of charge.
Publisher IOS Press Nieuwe Hemweg 6B 1013 BG Amsterdam The Netherlands
E-mail: Subscription Department:
[email protected] Advertising Department:
[email protected] Desk Editorial Department:
[email protected]
Tel.: Fax:
Internet: www.iospress.nl
+31 20 688 33 55 +31 20 620 34 19
© 2001 IOS Press. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publisher, IOS Press, Nieuwe Hemweg 6B, 1013 BG Amsterdam, The Netherlands. No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, instructions or ideas contained in the material herein. Although all advertising material is expected to conform to ethical standards, inclusion in this publication does not constitute a guarantee or endorsement of the quality or value of such product or of the claims made of it by its manufacturer. Special regulations for readers in the USA. This journal has been registered with the Copyright Clearance Center, Inc. Consent is given for copying of articles for personal or internal use, or for the personal use of specific clients. This consent is given on the condition that the copier pays through the Center the per-copy fee stated in the code on the first page of each article for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. The appropriate fee should be forwarded with a copy of the first page of the article to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. If no code appears in an article, the author has not given broad consent to copy and permission to copy must be obtained directly from the author. This consent does not extend to other kinds of copying, such as for general distribution, resale, advertising and promotion purposes, or for creating new collective works. Special written permission must be obtained from the publisher for such copying.
