Contributors
Irina D. Burd (383) Department of Obstetrics, Gynecology, and Reproductive Sciences, Robert Wood Johnson ...
106 downloads
1512 Views
80MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Contributors
Irina D. Burd (383) Department of Obstetrics, Gynecology, and Reproductive Sciences, Robert Wood Johnson Medical School, University of Medicine and Dentistry of New Jersey, New Brunswick, New Jersey 08901
Numbers in parentheses indicate the pages on which the authors' contributions begin.
Sanjay K. Agarwal (591) Division of Reproductive Medicine, Department of Obstetrics and Gynecology, Cedars-Sinai Burns and Allen Research Institute, Cedars-Sinai Medical Center, Los Angeles, California 90048; and Department of Obstetrics and Gynecology, University of California, Los Angeles, School of Medicine, Los Angeles, California 90095 Deborah J. Anderson (353) Fearing Laboratory, Department of Obstetrics, Gynecology, and Reproductive Biology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115 Nancy E. Avis (339) Institute for Women's Research, New England Research Institutes, Watertown, Massachusetts 02472 Gloria A. Bachmann (383) Department of Obstetrics, Gynecology, and Reproductive Sciences, Robert Wood Johnson Medical School, University of Medicine and Dentistry of New Jersey, New Brunswick, New Jersey 08901 John A. Baron (583) Section of Biostatistics and Epidemiology, Dartmouth Medical School, Hanover, New Hampshire 03755 Steven Birken (61) Department of Medicine, Columbia University College of Physicians and Surgeons, New York, New York 10032 Julia E. Bradsher (203) Abt Associates, Inc., Cambridge, Massachusetts 02138 M. Brincat (261) Department of Obstetrics and Gynecology, St. Luke's Hospital Medical School, Gwardamangia MSD 07, Malta
Henry G. Burger (147) Prince Henry's Institute of Medical Research, Clayton, Victoria 3168, Australia John E. Buster (625) Department of Obstetrics and Gynecology, Baylor College of Medicine, Houston, Texas 77030 Peter R. Casson* (625) Department of Obstetrics and Gynecology, Baylor College of Medicine, Houston, Texas 77030
Sybil L. Crawford (159, 175) New England Research Institutes, Watertown, Massachusetts 02472 Susan R. Davis (445) The Jean Hailes Foundation, Clayton, Victoria 3168, Australia; and Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Victoria 3004, Australia Carol A. Derby (229) New England Research Institutes, Watertown, Massachusetts 02472 Christine Draper (287) Department of Medicine, University of Western Australia, and Department of Endocrinology and Diabetes, Sir Charles Gairdner Hospital, Nedlands, Western Australia 6009, Australia Gary A. Ebert (383) Department of Obstetrics, Gynecology, and Reproductive Sciences, Robert Wood Johnson Medical School, University of Medicine and Dentistry of New Jersey, New Brunswick, New Jersey 08901
* Current address: Department of Obstetrics and Gynecology, Ottawa Hospital, Ottawa, Canada
xiii
xiv
Gregory F. Erickson (13) Department of Obstetrics and Gynecology, University of California, San Diego, La Jolla, California 92093 Denis Evans (175) Rush Institute on Aging, Chicago, Illinois 60612 Patricia D. Finn (33) Ligand Pharmaceuticals, Inc., San Diego, California 92121 Jeanne Franck (309) Department of Dermatology, Cornell Medical College and New York Presbyterian Hospital, New York, New York, 10021 Robert R. Freedman (215) Departments of Psychiatry and Behavioral Neurosciences and Obstetrics and Gynecology, Wayne State University, Detroit, Michigan 48201 R. Galea (261) Department of Obstetrics and Gynecology, St. Luke's Hospital Medical School, Gwardamangia MSD 07, Malta Ellen B. Gold (175, 189) Department of Epidemiology and Preventive Medicine, School of Medicine, University of California, Davis, Davis, California 95616 Joseph W. Goldzieher (397) Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center, Amarillo, Texas 79106 Gail A. Greendale (175, 639) Departments of Medicine and Obstetrics and Gynecology, University of California, Los Angeles, School of Medicine, Los Angeles, California 90024 Francine Grodstein (543) Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115 Robert E Heaney (481) Creighton University, Omaha, Nebraska 68131 Victor W. Henderson (315) Department of Neurology, University of Southern California, Los Angeles, California 90089 Victoria Hendrick ( l l l ) Department of Psychiatry, University of California, Los Angeles, School of Medicine, Los Angeles, California 90095 Howard L. Judd (591) Department of Obstetrics and Gynecology, Olive View/ UCLA Medical Center, Sylmar, California 91342; and Department of Obstetrics and Gynecology, University of California, Los Angeles, School of Medicine, Los Angeles, California 90095 Margaret R. Karagas (359) Section of Biostatistics and Epidemiology, Dartmouth Medical School, Lebanon, New Hampshire 03756
CONTRIBUTORS
F. S. J. Keating (509) Endocrinology and Metabolic Medicine, Imperial College School of Medicine, St. Mary's Hospital Medical School, London W2 1PG, United Kingdom Jennifer Kelsey (175, 359, 405) Division of Epidemiology, Department of Health Research & Policy, Stanford University School of Medicine, Stanford, California 94305 Stanley G. Korenman (111) Department of Medicine, University of California, Los Angeles, School of Medicine, Los Angeles, California 90095 Galina Kovalevskaya (61) Irving Center for Clinical Research, Columbia University, College of Physicians and Surgeons, New York, New York 10032 Mark A. Lawson (33) Ligand Pharmaceuticals, Inc., San Diego, California 92121 Annie Lo (175) Westat, Inc., Rockville, Maryland 20850 Leslie Lobel (61) Department of Obstetrics and Gynecology, Columbia University, College of Physicians and Surgeons, New York, New York 10032 Rogerio A. Lobo (429) Department of Obstetrics and Gynecology, Columbia University, College of Physicians and Surgeons, New York, New York 10032 Francisco Jos~ L6pez (33) Ligand Pharmaceuticals, Inc., San Diego, California 92121 Cecilia Magnusson (583) Department of Medical Epidemiology, Karolinska Institutet, S- 171 77 Stockholm, Sweden
N. Manassiev (509) Endocrinology and Metabolic Medicine, Imperial College School of Medicine, St. Mary's Hospital Medical School, London W2 1PG, United Kingdom Robert Marcus (405, 495) Department of Medicine, Stanford University School of Medicine, Geriatrics Research, Education & Clinical Center, Veterans Affairs Medical Center, Palo Alto, California 94304 Karen Matthews (175) Department of Psychiatry, University of Pittsburgh, Pittsburgh, Pennsylvania 15261 Donald P. McDonnell (3) Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, North Carolina 277 l0
CONTRIBUTORS
Valerie McGuire (359) Division of Epidemiology, Department of Health Research & Policy, Stanford University School of Medicine, Stanford, California 94305 Sonja M. McKinlay (203) New England Research Institutes, Watertown, Massachusetts 02172 Arshag D. Mooradian (111) Department of Internal Medicine, Saint Louis University School of Medicine, Saint Louis, Missouri 63104 David Morganstein (175) Westat, Inc., Rockville, Maryland 20850 Robert Neer (175) Division of Endocrinology, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts 02114 AndrOs Negro-Vilar (33) Ligand Pharmaceuticals, Inc., San Diego, California 92121 John O'Connor (61) Department of Pathology and Irving Center for Clinical Research, Columbia University, College of Physicians and Surgeons, New York, New York 10032 Richard L. Prince (287) Department of Medicine, University of Western Australia, and Department of Endocrinology and Diabetes, Sir Charles Gairdner Hospital, Nedlands, Western Australia 6009, Australia Janet H. Prystowsky (309) Department of Surgery, Columbia Presbyterian Medical Center, Columbia University and New York Presbyterian Hospital, New York, New York 10032 Russalind H. Ramos (459) Center for Menopause, Hormonal Disorders, and Women's Health, Sloane Hospital for Women, Columbia Presbyterian Medical Center, New York, New York, 10032 Nancy E. Reame (95) Center for Nursing Research and Reproductive Sciences Program, The University of Michigan, Ann Arbor, Michigan 48109 Robert W. Rebar (135) Department of Obstetrics and Gynecology, University of Cincinnati Medical Center, Cincinnati, Ohio 45267; and the American Society for Reproductive Medicine, Birmingham, Alabama 35216 Clifford J. Rosen (271) St. Joseph Hospital, Maine Center for Osteoporosis Research and Education, Bangor, Maine 04401 Giiran Samsioe (327) Department of Obstetrics and Gynecology, Lund University Hospital, S-221 85 Lund, Sweden Sherry Sherman (175) NIH/NIA, Bethesda, Maryland 20892
XV
Barbara B. Sherwin (617) Department of Psychology and Department of Obstetrics and Gynecology, McGill University, Montreal, H3A 1B 1 Canada Joe Leigh Simpson (77) Departments of Obstetrics and Gynecology and Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas 77030 MaryFran Sowers (175, 245, 535) Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan 48109 Leon Speroff (553) Department of Obstetrics and Gynecology, Oregon Health Sciences University, Portland, Oregon 97201 Margaret G. Spinelli (563) Department of Clinical Psychiatry, Columbia University, College of Physicians and Surgeons; and The New York State Psychiatric Institute, New York, New York 10032 Meir J. Stampfer (543) Departments of Nutrition and Epidemiology, Harvard School of Public Health; and Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115 Frank Z. Stanezyk (421) Department of Obstetrics and Gynecology, University of Southern California School of Medicine, Los Angeles, California 90033 Barbara Sternfeld (175, 495) Department of Epidemiology and Biostatistics, Division of Research, Kaiser Permanente, Oakland, California 94611 J. C. Stevenson (509) Endocrinology and Metabolic Medicine, Imperial College School of Medicine, St. Mary's Hospital Medical School, London W2 1PG, United Kingdom Jennifer Tiseh (245) Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan 48109 Anna N. A. Tosteson (649) Clinical Research Section, Department of Medicine and the Center for the Evaluative Clinical Sciences, Department of Community and Family Medicine, Dartmouth Medical School, Hanover, New Hampshire 03755 Michelle P. Warren (459) Department of Obstetrics and Gynecology and Medicine, Columbia University, College of Physicians and Surgeons, New York, New York 10032; and Center for Menopause, Hormonal Disorders, and Women's Health, Sloane Hospital for Women, Columbia Presbyterian Medical Center, New York, New York, 10032
xvi Gerson Weiss (175) Department of Obstetrics and Gynecology, University of Medicine and Dentistry of New Jersey, Newark, New Jersey 07103 Carolyn Westhoff (607) Columbia University, College of Physicians and Surgeons, New York, New York 10032
CONTRIBUTORS
Preface
lished in recent years, this work represents what we feel will be the first compilation of the entire subject, from basic biology to medical issues and therapeutic considerations. The text has been divided into several areas: basic biology, epidemiology, pathophysiology, and interventions. The chapters in each section are authored by outstanding investigators in the field who are recognized for their expertise. Accordingly, this text will be useful to all who are interested in the field, including basic and clinical investigators, students, residents, fellows, and clinicians. We are indebted to our friends, the contributing authors, and the tireless work of Jenny Wrenn and Jasna Markovac of Academic Press, who have shepherded this project from its inception. Without their help and persistence, this volume would not have been completed so efficiently and professionally, and in a timely manner.
Menopause is defined as the cessation of menstrual flow. Because the age of menopause is largely genetically determined, the average age at which it occurs, approximately 51 years, has not changed over many centuries. However, life expectancy has increased substantially, and the current life expectancy for women is 80 years. If a woman reaches age 54, she can expect to reach the age of 84.3 years. Thus, the years after menopause may account for as much as 40% of a woman's life. Currently in the United States, there are approximately 31 million women over age 55, with estimates of 38 million in 2010 and 46 million in 2020. Similar trends will occur in many parts of the world. Thus, there is a large and ever-increasing population of women in this very important time of life. Many women look forward to this time, even viewing this signal of a change in their reproductive lives as an opportunity for change and for instituting preventive health care. Nevertheless, for many years menopause has not been well understood. Although numerous books on the clinical aspects and the management of menopause have been pub-
Rogerio A. Lobo Jennifer Kelsey Robert Marcus
xvii
2HAPTER
Molecular Pharmacology of Estrogen and
Progesterone Receptors DONALD P.
MCDONNELL
Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710
I. Introduction II. Estrogen and Progesterone Receptors III. Established Models of Estrogen and Progesterone Action IV. Estrogen and Progesterone Receptor Isoforms and Subtypes V. Regulation of Estrogen and Progesterone Receptor Function by Ligands
I. I N T R O D U C T I O N The steroid hormones estrogen and progesterone are small molecular weight lipophilic hormones that, through their action as modulators of distinct signal transduction pathways, are involved in the regulation of reproductive function [1, 2]. These hormones have also been shown to be important regulators in bone, the cardiovascular system, and the central nervous system [3-5]. Despite their different roles in these systems, however, it has become apparent that estrogens and progestins are mechanistically similar [6]. Insights gleaned from the study of each hormone, therefore, have advanced our understanding of this class of molecules as a whole. This review highlights some of the recent mechanistic discoveries that have occurred in the field, and e x -
MENOPAUSE: B I O L O G Y AND PATHOBIOLOGY
VI. Estrogen and Progesterone Receptor Associated Proteins VII. An Updated Model of Estrogen and Progesterone Receptor Action References
plores the subsequent changes in our understanding of the pharmacology of this class of steroid hormones.
II. E S T R O G E N AND PROGESTERONE RECEPTORS The estrogen receptor (ER) and progesterone receptor (PR) cDNAs have been cloned and used to develop specific ligand-responsive transcription systems in heterologous cells, permitting the use of reverse genetic approaches to define the functional domains within each of the receptors [6]. A schematic that outlines the organization of the major functional domains within these two steroid receptors (SR) proteins is shown in Fig. 1. The largest domain (~--300 amino acids) that
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
4
DONALD P. MCDONNELL An additional activation function (AF-1) is located within the amino terminus of each receptor [ 11 ]. The DNA-binding domain (DBD) is a short region (--~70 amino acid residues) located in the center of the receptor protein [12]. Thispermits the receptor to bind as a dimer to target genes. Within the DBD there are nine cysteine residues, eight of which can chelate two zinc atoms, thereby forming two fingerlike structures that allow the receptor to interact with DNA [13]. All of the information required to permit target gene identification by ligand-activated ERs and PRs is contained within this region.
III. ESTABLISHED OF ESTROGEN PROGESTERONE
FIGURE 1 Establishedmodels of estrogen and progesteroneaction. The classic models of estrogen and progesterone action suggested that, in the absence of ligand, the steroid receptor (SR) exists in the nuclei of target cells in an inactive form. On binding an agonist, the receptorwould undergo an activating transformation event that displaces inhibitory heat-shock proteins (HSP) and facilitates the interaction of the receptorwith specific DNA steroid response elements (SRE) within target gene promoters. The activated receptor dimer could then interact with the general transcription machinery and positively or negatively regulate target gene transcription. In this model the role of the agonist is that of a "switch" that merely converts the ER or PR from an inactive to an active form. Thus, when corrected for affinity, all agonists would be qualitatively the same and evoke the same phenotypic response. By inference,antagonists,compounds that oppose the actions of agonists, would competitivelybind to their cognate receptors and freeze them in an inactive form. As with agonists, this model predicted that all antagonists are qualitatively the same. Within the confines of this classic model it was difficult to explain the molecularpharmacologyof the known ER and PR agonists and antagonists. GTA, General transcriptional activity.
is responsible for ligand binding is located at the carboxyl terminus of each receptor. Crystallographic analysis of the agonist-bound forms of ERs and PRs has indicated that this domain consists of 12 short a-helical structures that fold to provide a complex ligand-binding pocket [7, 8]. The ligandbinding domain also contains sequences that facilitate receptor homodimerization and permit the interaction of apo receptors with inhibitory heat-shock proteins. An activation function (AF-2) required for receptor transcriptional activity is also contained within the ligand-binding domain [9, 10].
MODELS
AND ACTION
The steroid hormones estrogen and progesterone are representative members of a larger family of steroid hormones, all of which appear to share a common mechanism of action. It is generally believed that steroid hormones enter cells from the bloodstream by simple passive diffusion, exhibiting activity only in cells in which they encounter a specific high-affinity receptor protein [14]. These receptor proteins are transcriptionally inactive in the absence of ligand, sequestered in a large oligomeric heat-shock protein complex within target cells [15]. On binding ligand, however, the receptors undergo an activating conformational change that promotes the dissociation of inhibitory proteins [16]. This event permits the formation of receptor homodimers that are capable of interacting with specific high-affinity DNAresponse elements located within the regulatory regions of target genes (Fig. 2) [17]. The DNA-bound receptor can then exert a positive or negative influence on target g e n e transcription. In the classic models of steroid hormone action, it was proposed that progestins and estrogens function merely as switches that, on binding to their cognate receptor, permit
FIGURE 2 The domain structures of the estrogen and progesterone receptors are similar.
CHAPTER 1 Estrogen and Progesterone Receptors conversion of the receptor in an all or nothing manner from an inactive to an active state [ 18]. This implied that ER and PR pharmacology was very simple, and that when corrected for affinity all progestins and estrogens were qualitatively the same. Furthermore, it suggested that antihormones (antagonists) function simply as competitive inhibitors of agonist binding, freezing the target receptor in an inactive state within the cell. Under most experimental conditions this simple model was sufficient to explain the observed biology of known PR and ER agonists and antagonists. However, systems were discovered that did not fit this simple model, indicating that the pharmacology of these receptor systems is more complex than originally believed. Studies probing the complex pharmacology of the antiestrogen tamoxifen have been very informative with respect to understanding the inadequacies of the classic model. Tamoxifen is widely used as a breast cancer chemotherapeutic and has been approved for use as a breast cancer chemopreventive in high-risk patients [ 19, 20]. In ER-positive breast cancers, tamoxifen opposes the mitogenic action of estrogen(s) by binding to the receptor and competitively blocking agonist access. However, it has become clear in recent years that tamoxifen is not a pure antagonist, because in some target organs it can exhibit estrogen-like activity. This is most apparent in both the skeletal system, where tamoxifen, like estrogen, increases lumbar spine bone mineral density, and the cardiovascular system, where both tamoxifen and estradiol have been shown to decrease low-density liproprotein (LDL) cholesterol [21, 22]. These in vivo properties of tamoxifen led to its being reclassified as a selective estrogen receptor modulator (SERM) rather than an antagonist. The observation that different ligand-receptor complexes were not recognized in the same manner in all cells was at odds with the established models of ER action. From a clinical perspective this was an important finding, because it suggested for the first time the possibility of developing compounds that, acting through their cognate receptor, could manifest different activities in different cells. From a molecular point of view, however, the observed pharmacology of tamoxifen begged a reevaluation of the classic model of ER action, and initiated the search for the cellular systems that enable ER-ligand complexes to manifest different biologies in different cells. These ongoing investigations have also provided significant insight into PR action, and have demonstrated that, as in the case of ERs, it will be possible to develop compounds that manifest PRagonist activity in a tissue-selective manner.
IV. ESTROGEN AND PROGESTERONE RECEPTOR ISOFORMS AND SUBTYPES One mechanism to explain the cell-selective action of steroid receptor ligands is the likelihood that they may activate different receptor isoforms (derived from the same gene) or
5
Estrogen Receptor Subtypes 1
595
NH21
hERc~
1
530
NH21
hER[3
I
Progesterone Receptor Subtypes 1
933
1
hPR-A
NH~I
769
I~1
~
I
FIGURE 3 At least two distinct forms of the estrogen and progesterone receptors exist in target cells. DBD, DNA-binding domain; LBD, ligandbinding domain.
subtypes (derived from similar genes). This concept has been well established for the ce- and fi-adrenergic systems, where it has been shown that different receptor subtypes have distinct ligand preferences, and that selectivity can be explained by differences in the expression of these subtypes. Until recently, the parallel between this system and that of the nuclear receptors was not obvious. However, the identification and characterization of ER and PR isoforms and subtypes has shed new light on this issue (Fig. 3).
A. ProgesteroneReceptor Isoforms The progesterone receptor was the first receptor for which bona fide isoforms were shown to exist. Human PRs can exist within target cells in either of two distinct forms, hPR-A (94 kDa) or hPR-B (114 kDa) [23]. These proteins, differing only in that the hPR-B isoform contains an additional 164-amino acid extension at its amino terminus, are produced from distinct mRNAs that are derived from different promoters within the same gene [9]. In most progesterone-responsive tissues these two receptor isoforms are expressed in equimolar amounts. This apparent 1:1 relationship is so widespread that until about 10 years ago the hPR-A isoform was thought to be merely an artifact derived from hPR-B by proteolysis during biochemical fractionation. It has now been established that these two proteins are produced in a deliberate manner by the cell, and that they are not functionally equivalent [23-25]. The first evidence in support of this hypothesis came following the cloning and subsequent functional analysis of the chicken progesterone receptor (cPR) cDNA [26]. Specifically, on expression in
6
DONALD P. MCDONNELL
heterologous cells, it was found that although the A and B forms of cPRs display identical ligand binding preferences, they activate different target genes [26]. It was subsequently shown that the amino-terminal sequences, which distinguish cPR-B from cPR-A, are important in determining target gene selectivity. This concept was reaffirmed when the cloned hPR-B and hPR-A were analyzed in a similar manner [24]. In the systems examined thus far, with few exceptions, it has been observed that hPR-B alone functions as a transcriptional activator in response to progesterone, whereas hPR-A displays minimal or no activity. Further analysis has revealed that hPR-A functions primarily as a ligand-dependent transdominant modulator of the transcriptional activity of hPR-B, the ability of hPR-B to activate target gene transcription being influenced by the cellular concentration of hPR-A [24, 27]. Surprisingly, it was also determined that ligandactivated hPR-A can inhibit the transcriptional activity of agonist-activated ERs, androgen receptors (ARs), and mineralocorticoid receptors (MRs) [24]. Thus, by virtue of having two functionally different receptor isoforms, a single hormone such as progesterone can have completely different functions in target cells.
B. E s t r o g e n R e c e p t o r S u b t y p e s The identification of functionally distinct PR isoforms introduced a new dimension to progesterone action, although it was not until a second estrogen receptor was cloned in 1995 that the general significance of isoforms (or subtypes) in steroid receptor signaling was established [28]. Unlike the case of PRs, ERa and ERfl are encoded by different genes, and although they share significant amino acid homology in their ligand-binding domains, they are not pharmacologically equivalent. Both receptors bind the endogenous estrogen, 17fl-estradiol, with equivalent affinity [29]. However, when binding analysis was extended to additional compounds, significant differences in ligand preferences were noted. The biological and pharmacological consequences of these differences remain to be determined. Although the discovery of ERfl has occurred relatively recently, significant progress has been made in elucidating its role in estrogen signaling. It has been determined that the expression pattern of ERfl does not mirror that of ERce [30, 31]. Expression of both isoforms is found in some tissues, whereas ER/3 alone occurs in others, such as the lung, the urogenital tract, and the colon [29]. The distinct roles of these two receptors in the endocrinology of estrogen have been confirmed by the generation of mice whose ER/3 has been genetically disrupted [32]. The phenotype of these mice is different from that of ERa knockout mice [32]. The specific role for ER/3 in tissues in which it is the only estrogen receptor expressed has not yet been identified. The high degree of amino acid homology between ERa and
ERfl within the DNA-binding domain suggests, but does not prove, that these receptors may regulate the same genes. It is also possible that ER/3 may interact with target genes in a manner that does not require direct contact with DNA regulatory elements within target genes. The identification of estrogen and progesterone receptor isoforms and subtypes and the definition of specific functions that they modulate have introduced a new dimension in steroid hormone action. Understanding the regulatory mechanisms that control the expression levels of the individual forms of each receptor is likely to provide novel targets for pharmaceutical intervention.
V. R E G U L A T I O N
OF ESTROGEN
AND PROGESTERONE RECEPTOR FUNCTION
BY LIGANDS
The finding that ERs and PRs could exist in multiple forms within target cells suggests that some of the tissueselective actions of their cognate agonists and antagonists can be explained by their ability to regulate differentially the action of one specific receptor isoform or subtype. Although a specific example of a receptor subtype-selective steroid receptor ligand has not yet emerged, the fact that such ligands for the retinoic acid receptor(s) have been generated makes the discovery of the ER and PR subtype-selective ligands more likely. Regardless, however, it has become apparent from the study of antiestrogens that the identical ligand operating through the same receptor can manifest different biological activities in different target cells [33]. In breast tissue, for instance, where ERa predominates, all of the known antiestrogens oppose the mitogenic actions of estrogen [34]. In the endometrium, however, where ERa also predominates, it has been found that tamoxifen functions as a partial estrogen mimetic [35, 36], whereas compounds such as raloxifene, GW5638, and ICI182,780 function as pure antiestrogens. Thus, the same compounds, acting through ERa, manifest different biological activities in the breast and the endometrium. This finding is not in agreement with the classic models of ER action that indicate that ligands basically fall into two classes, agonists and antagonists. This paradox has been the subject of much investigation leading to the observation that different compounds can induce different alterations in ER structure and that not all structures are functionally identical. It is implied, therefore (discussed in more detail below) that the cell possesses the cellular machinery to distinguish between these dissimilar complexes and that the identification and characterization of the specific components of these systems are the keys to the development of the next generation of tissue-selective ER and PR modulators. Much of what we know about the effect of ligands on ste-
CHAPTER 1 Estrogen and Progesterone Receptors roid receptor structure has come from studies of different ER-ligand complexes. Initially, using differential sensitivity to proteases, it was demonstrated that the hormone-binding domain within the ER adopts different shapes on binding estradiol and tamoxifen, and that these structures are dissimilar to that of the apo receptor [33, 37]. Thus, receptor conformation is affected by the nature of the bound ligand. This relationship between structure and function was later confirmed by the observation that agonists and antagonists induce different alterations in PR structure [38, 39]. Further analysis has revealed that the majority of the structural changes that occur in the PR are located at the extreme carboxyl tail of the receptor, and that removal of the carboxylterminal 42 amino acids of hPR-B permit the antagonist RU486 to function as an agonist [38]. Interestingly, a similarly positioned domain enables the ER to discriminate between different compounds and, not surprisingly, removal of 35 amino acids from the C-terminal tail of the ER abolishes its ability to distinguish between agonists and antagonists [40]. The recent determination of the crystalline structures of the ER-estradiol and ER-tamoxifen complexes confirmed the important role of the carboxyl tail in determining the pharmacology of steroid receptors [7, 41, 42]. This new structural information has also revealed that agonist activation of the ER permits the formation of a unique surface (or pocket) on the receptor that allows it to interact with the general transcription machinery through the mediation of adaptor or coactivator proteins. In the presence of the antagonist tamoxifen, however, the carboxyl tail of the ER is positioned in such a manner that it occludes this coactivator binding pocket, preventing a productive association with the cellular transcription apparatus. In addition to tamoxifen there are several additional SERMs that manifest distinct activities in vivo. One of these compounds, raloxifene, has been approved as a SERM for the treatment of osteoporosis [43]. This compound distinguishes itself from tamoxifen in that it does not exhibit estrogenic action in the postmenopausal endometrium [44, 45]. Although clearly different biologically, the crystal structures of the ER-tamoxifen and ER-raloxifene complexes were shown to be virtually indistinguishable. Although these results appear to be at odds with the hypothesis that links receptor structure to function, some data from our group have reconciled these potential discrepancies. We have used phage display technology to identify small peptides, the ability of which to bind ERs is affected differentially by the nature of the ligand bound to the receptor [46-48]. The rationale behind this approach is that because of the vast complexity of the peptides available in these libraries, it may be possible to find peptides that have the ability to distinguish between two very similar receptorligand complexes. This approach has led to the identification of a series of high-affinity peptide probes that, in addition to
7 being able to distinguish between ER-estradiol and E R tamoxifen complexes, are also able to distinguish among several different E R - S E R M complexes (Fig. 4). This approach has been extended to the study of the PR and it was similarly observed that various PR ligands manifest different
FIGURE 4 Fingerprinting the surfaces of different ER-ligand complexes using conformation-sensitive peptide probes. (A) Random peptide libraries were constructed in an M13 bacteriophage; each of the resulting bacteriophages expressed a unique random peptide on its surface pilus. Screens were subsequently performed to identify specific peptides (bacteriophage) whose interaction with the ER was influenced by the nature of the bound ligand. The bacteriophages identified in this manner were used to develop an enzyme-linked immunoassay to monitor changes that occur in the ER on its interaction with different ligands. Specifically, a biotinylated estrogen response element (ERE) was used to immobilize recombinant ERs on streptavidin-coated plates. After incubation of this complex with the ligand to be tested, to each well was added an aliquot of a different class of ERinteracting bacteriophage. Binding of the bacteriophage was assessed enzymatically using an anti-M 13 antibody coupled to horseradish peroxidase (HRP). (B) Fingerprint analysis of ER conformation in the presence of different ER ligands. Immobilized ER was incubated in the presence of saturating concentrations of the indicated ligands, and the resulting complexes were incubated with aliquots of bacteriophage expressing eight different peptides. Tam, Tamoxifen; DES, diethylstilbestrol; Prog, progesterone. This figure has been published previously in a similar form [47] and is reproduced and presented here with permission (Copyright 1999 National Academy of Sciences, U.S.A.).
8
DONALD P. MCDONNELL
biologies in different cells, allowing the identification of peptide probes whose interaction with the receptor is influenced by the nature of the ligand bound to the PR. All these findings establish a firm relationship between the structure of a receptor-ligand complex and biological activity, and suggest that novel ER and PR ligands with unique pharmaceutical properties may be developed by exploiting this observation.
VI. ESTROGEN AND PROGESTERONE RECEPTOR
ASSOCIATED
PROTEINS
The estrogen and progesterone receptors are liganddependent transcription factors that, on activation by ligands, associate with specific DNA response elements located within the regulatory regions of target genes [ 14]. The DNA-bound receptor can then positively or negatively influence gene transcription by altering RNA polymerase II activity. However, because RNA polymerase does not appear to interact directly with the steroid receptors, there must be additional factors that allow these two proteins to communicate [ 14]. In the past few years it has become clear that there are at least two functional classes of proteins that are involved in recognizing the activated receptor. One class includes components of the basic transcription machinery, the general transcription factors, whose expression levels are generally invariant from cell to cell. The second class of proteins, "cofactors," is not a part of the general transcription machinery, and can exert either a positive or a negative influence on SR transcriptional activity [49]. Those cofactors that interact with agonist-activated SRs have been called coactivators, whereas those that interact with apo receptors or antagonist-activated receptors have been called corepressors. Interestingly, it has become apparent that differences in the relative expression levels of coactivators and corepressors can have a profound effect on the pharmacology of estrogen and progesterone receptor ligands [25, 50, 51].
have revealed that (1) the expression levels of these coactivators vary from cell to cell, (2) coactivators demonstrate specific receptor preferences, (3) a given receptor can interact with more than one type of coactivator, and (4) the conformation of the receptor adopted in the presence of a specific ligand can determine which coactivators are engaged. These findings strongly support the hypothesis that differential cofactor expression is the most important determinant of estrogen and progesterone receptor pharmacology. With the discovery of the nuclear receptor coactivators and the characterization of their biochemical properties has come a new understanding of the mechanism by which differently conformed receptor-ligand complexes are recognized in the cell. The studies that have been performed with ERs are the most informative. As described previously, it has been shown that ERs in the presence of estradiol undergo a conformational change that allows the presentation of surfaces on the receptors, permitting them to interact with coactivators. Because estradiol induces the same conformational change within ERs in all cells, the phenotypic consequence of the exposure of a cell to estradiol will depend on the properties of the coactivators expressed in that cell (Fig. 5). The
A. C o a c t i v a t o r P r o t e i n s One of the most well-characterized coactivator proteins, steroid receptor coactivator 1 (SRC-1), was identified in a yeast two-hybrid screen as a protein that interacted with agonist-activated PR [52]. Subsequently, this protein has been shown to also interact with the estrogen, glucocorticoid, and androgen receptors. It appears that SRC-1 increases target gene transcription by linking the hormone-activated receptor with the general transcription machinery, stabilizing the transcription preinitiation complex, and nucleating a large complex of proteins that together have the ability to acetylate histones and facilitate chromatin decondensation [53, 54]. In addition to SRC-1, over 30 additional coactivators have been identified and characterized [49]. Cumulatively, these studies
FIGURE 5 A molecular explanation for the tissue-selective agonist/antagonist activity of the SERM tamoxifen. The estrogen receptor undergoes different conformational changes on binding the full agonistestrogen or the SERM tamoxifen. The estradiol-induced conformation allowsthe ER to interact with any coactivator protein expressed in target cells, and thus it can activate transcription. The tamoxifen-induced conformational change, on the other hand, is more restrictive and allows the interaction of the ER with only a subset of available coactivators. In those cells in which the tamoxifen-ER complex can engage a coactivator, this compound can manifest agonist activity.In other contexts tamoxifen functions as an antagonist.
CHAPTER1 Estrogen and Progesterone Receptors situation gets more complicated, however, when considering the role of coactivators in mediating the cell-selective action of SERMs such as tamoxifen. It has been shown that the tamoxifen-induced conformational change within the ER does not allow the coactivator binding pocket to form properly, preventing or hindering the interaction of those coactivators that require the coactivator binding pocket in order to interact with the E R [46]. In cells in which this type of coactivator is important, therefore, tamoxifen can function as an antagonist. It is becoming clear, however, that not all coactivators rely on the coactivator binding pocket to the same degree. Thus, the relative agonist-antagonist activity of tamoxifen depends on the ability of the t a m o x i f e n - E R complex to engage a compatible coactivator in target cells [33]. As the repertoire of cofactors increases, we are likely to find that targeting specific c o f a c t o r - r e c e p t o r complexes will yield pharmaceuticals that manifest their activities in a cell- or tissuerestricted manner.
B.
Corepressor
Proteins
Two nuclear corepressor proteins that appear to be important in ER and PR action have thus far been identified. These proteins, N C o R and SMRT, initially found as proteins that interact with D N A - b o u n d thyroid hormone or retinoid X receptors, repress basal transcription in the absence of hormone [55, 56]. However, it has now been shown that these proteins can interact with either PR or ER, either in the absence of h o r m o n e or in the presence of antagonists [50, 51]. Under these conditions, the corepressors nucleate a large protein complex, which represses target gene transcription by deacetylating histones and facilitating chromatin condensation. The physiological importance of corepressors in ER pharmacology was suggested by the studies of Lavinsky and co-workers, who found that passage of breast tumors in mice from a state of tamoxifen sensitivity to an insensitive state was accompanied by a decrease in the expression level of the corepressor [57]. A similar process, if occurring in humans, could explain how cells become resistant to the antiestrogenic actions of tamoxifen.
AN UPDATED M O D E L O F E S T R O G E N AND P R O G E S T E R O N E RECEPTOR ACTION VII.
On ligand binding, the activated receptor (ER or PR) can interact as a dimer with specific D N A response elements within target genes. It is now apparent that the conformation of the resulting receptor is influenced by the nature of the bound ligand and that the shape of the resulting receptor-ligand complex is a critical determinant of whether it can activate transcription. In the presence of a full agonist, the conformation adopted by the receptor facilitates the displacement
9 of corepressor proteins and recruitment of coactivator proteins, permitting the assembly of a histone-acetylating complex on D N A with a concomitant increase in target gene transcription. Pure antagonists, on the other hand, drive their cognate receptor into a conformation that favors corepressor interaction. The activity of mixed agonists/antagonists appears to relate to their ability to alter receptor conformation differentially, and the ability of corepressor and coactivator proteins within a given target cell to recognize these complexes. Clearly, the classic models of estrogen and progesterone action need to be updated to accommodate the insights that have emerged from the study of the genetics and molecular pharmacology of these two receptors.
References 1. Clark, J. H., and Markaverich, B. M. (1988). Actions of ovarian steroid hormones. In "The Physiology of Reproduction," E. Knobil and J. Neill, eds., pp. 675-724. Raven Press, New York. 2. Clarke, C. L., and Sutherland, R. L. (1990). Progestin regulation of cellular proliferation. Endocr. Rev 11, 266-301. 3. Colditz, G. A., Hankinson, S. E., Hunter, D. J., Willett, W. C., Manson, J. E., Stampfer, M. J., Hennekens, C., Rosner, B., and Speizer, E E. (1995). The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. N. Engl. J. Med. 332, 1589-1593. 4. Barzel, U. S. (1988). Estrogens in the prevention and treatment of postmenopausal osteoporosis. Am. J. Med. 85, 847-850. 5. Grodstein, F., Stampfer, M. J., Manson, J. E., Colditz, G. A., Willett, W. C., Rosner, B., Speizer, E E., and Hennekens, C. H. (1996). Postmenopausal estrogen and progestin use and the risk of cardiovascular disease. N. Engl. J. Med. 335, 453-461. 6. McDonnell, D. P., Vegeto, E., and Gleeson, M. A. G. (1993). Nuclear hormone receptors as targets for new drug discovery. Bio/Technology 11, 1256-1260. 7. Brzozowski, A. M., Pike, A. C., Dauter, Z., Hubbard, R. E., Bonn, T., Engstr6m, O., Ohman, L., Greene, G. L., Gustafsson, J. A., and Carlquist, M. (1997). Molecular basis of agonism and antagonism in the oestrogen receptor. Nature (London) 389, 753-758. 8. Williams, S. P., and Sigler, P. B. (1998). Atomic structure of progesterone complexed with its receptor. Nature (London) 393, 392-396. 9. Giangrande, P. H., and McDonnell, D. P. (1999). The A and B isoforms of the human progesterone receptor: Two functionally different transcription factors encoded by a single gene. Recent Prog. Horm. Res. 54, 291-314. 10. Tzukerman, M. T., Esty, A., Santiso-Mere, D., Danielian, P., Parker, M. G., Stein, R. B., Pike, J. W., and McDonnell, D. P. (1994). Human estrogen receptor transactivational capacity is determined by both cellular and promoter context and mediated by two functionally distinct intramolecular regions. Mol. Endocrinol. 8, 21-30. 11. Gronemeyer, H. (1992). Control of transcription activation by steroid hormone. FASEB J. 6, 2524-2529. 12. Kumar, V., Green, S., Stack, G., Berry, M., Jin, J.-R., and Chambon, P. (1987). Functional domains of the human estrogen receptor. Cell (Cambridge, Mass.) 51, 941-951. 13. Schwabe, J. W. R., Chapman, L., Finch, J. T., and Rhodes, D. (1993). The crystal structure of the estrogen receptor DNA-binding domain bound to DNA: How receptors discriminate between their response elements. Cell (Cambridge, Mass.) 75, 567-578. 14. O'Malley, B. W., Tsai, S. Y., Bagchi, M., Weigel, N. L., Schrader, W. T., and Tsai, M.-J. (1991). Molecular mechanism of action of a steroid hormone receptor. Recent Prog. Horm. Res. 47, 1-26.
10 15. Bagchi, M. K., Tsai, S. Y., Tsai, M.-J., and O'Malley, B. W. (1990). Identification of a functional intermediate in receptor activation in progesterone-dependent cell-free transcription. Nature (London) 345, 547-550. 16. Allan, G. E, Tsai, S. Y., Tsai, M.-J., and O'Malley, B. W. (1992). Ligand-dependent conformational changes in the progesterone receptor are necessary for events that follow DNA binding. Proc. Natl. Acad. Sci. US.A. 89, 11750-11754. 17. Bagchi, M. K., Elliston, J. F., Tsai, S. Y., Edwards, D. P., Tsai, M.-J., and O' Malley, B. W. (1988). Steroid hormone-dependent interaction of human progesterone receptor with its target enhancer element. Mol. Endocrinol. 2, 1221-1229. 18. Clark, J. H., and Peck, E. J. (1979). "Female Sex Steroids: Receptors and Function," 1st ed., Vol. 14. Springer-Verlag, New York. 19. Jordan, V. C., ed. (1996). "Tamoxifen: A Guide for Clinicians and Patients." PRR, Huntington, N.Y. 20. Fisher, B., Costantino, J. P., Wickerham, D. L., Redmond, C. K., Kavanah, M., Cronin, W. M., Vogel, V., Robidoux, A., Dimitrov, N., Atkins, J., Daly, M., Wieand, S., Tan-Chiu, E., Ford, I., and Wolmark, N. (1998). Tamoxifen for prevention of breast cancer: Report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study. J. Natl. Cancer Inst. 90, 1371-1388. 21. Love, R. R., Mazess, R. B., Barden, H. S., Epstein, S., Newcomb, P. A., Jordan, V. C., Carbone, P. P., and DeMets, D. L. (1992). Effects of tamoxifen on bone mineral density in postmenopausal women with breast cancer. N. Engl. J. Med. 326, 852-856. 22. Love, R. R., Wiebe, D. A., Newcombe, P. A., Cameron, L., Levanthal, H., Jordan, V. C., Feyzi, J., and DeMets, D. L. (1991). Effects of tamoxifen on cardiovascular risk factors in postmenopausal women. Ann. Intern. Med. 115, 860-864. 23. Lessey, B. A., Alexander, P. S., and Horwitz, K. B. (1983). The subunit characterization of human breast cancer progesterone receptors: Characterization by chromatography and photoaffinity labelling. Endocrinology (Baltimore) 112, 1267-1274. 24. Vegeto, E., Shahbaz, M. M., Wen, D. X., Goldman, M. E., O'Malley, B. W., and McDonnell, D. P. (1993). Human progesterone receptor A form is a cell- and promoter-specific repressor of human progesterone receptor B function. Mol. Endocrinol. 7, ! 244-1255. 25. Jackson, T. A., Richer, J. K., Bain, D. L., Takimoto, G. S., Tung, L., and Horwitz, K. B. (1997). The partial agonist activity of antagonistoccupied steroid receptors is controlled by a novel hinge domainbinding coactivator L7/SPA and the corepressors NCoR or SMRT. Mol. Endocrinol. 11, 693-705. 26. Conneely, O. M., Dobson, A. D. W., Huckaby, C., Carson, M. A., Maxwell, B. L., Toft, D. O., Tsai, M.-J., Schrader, W. T., and O' Malley, B. W. (1988). Structure-function relationships of the chicken progesterone receptor. In "Steroid Hormone Action," G. Ringold, ed., pp. 227-236. Alan R. Liss, New York. 27. Wen, D. X., Xu, Y.-E, Mais, D. E., Goldman, M. E., and McDonnell, D. P. (1994). The A and B isoforms of the human progesterone receptor operate through distinct signaling pathways within target cells. Mol. Cell. Biol. 14, 8356-8364. 28. Kuiper, G. G. J. M., Enmark, E., Pelto-Huikko, M., Nilsson, S., and Gustafsson, J.-A. (1996). Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc. Natl. Acad. Sci. U.S.A. 93, 5925 -5930. 29. Kuiper, G. G. J. M., Carlsson, B., Grandien, K., Enmark, E., H~iggblad, J., Nilsson, S., and Gustafsson, J.-A. (1997). Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors ce and ft. Endocrinology (Baltimore) 138, 863-870. 30. Shughrue, P. J., Komm, B., and Merchenthaler, I. (1996). The distribution of estrogen receptor-fl mRNA in the rat hypothalamus. Steroids 61,678-681. 31. Shughrue, P. J., Lane, M. V., and Merchenthaler, I. (1997). Comparative
DONALD P. MCDONNELL
32. 33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
distribution of estrogen receptor-ce and fi mRNA in the rat central nervous system. J. Comp. Neurol. 388, 507-525. Couse, J. F., and Korach, K. S. (1999). Estrogen receptor null mice: What have we learned and where will they lead us? Endocr. Rev. 20, 358-417. McDonnell, D. P., Clemm, D. L., Hermann, T., Goldman, M. E., and Pike, J. W. (1995). Analysis of estrogen receptor function in vitro reveals three distinct classes of antiestrogens. Mol. Endocrinol. 9, 659-668. Jordan, V. C. (1992). The strategic use of antiestrogen to control the development and growth of breast cancer. Cancer (Philadelphia) 70, 977-982. Kilackey, M. A., Hakes, T. B., and Pierce, V. K. (1985). Endometrial adenocarcinoma in breast cancer patients receiving antiestrogens. Cancer Treat. Rep. 69, 237-238. Gottardis, M. M., Robinson, S. P., Satyaswaroop, P. G., and Jordan, V. C. (1988). Contrasting actions of tamoxifen on endometrial and breast tumor growth in the athymic mouse. Cancer Res. 48, 812-815. Allan, G. F., Leng, X., Tsai, S. Y., Weigel, N. L., Edwards, D. P., Tsai, M.-J., and O'Malley, B. W. (1992). Hormone and antihormone induce distinct conformational changes which are central to steroid receptor activation. J. Biol. Chem. 267, 19513-19520. Vegeto, E., Allan, G. F., Schrader, W. T., Tsai, M.-J., McDonnell, D. P., and O'Malley, B. W. (1992). The mechanism of RU486 antagonism is dependent on the conformation of the carboxy-terminal tail of the human progesterone receptor. Cell (Cambridge, Mass.) 69, 703-713. Wagner, B. L., Pollio, G., Leonhardt, S., Wani, M. C., Lee, D. Y.-W., Imhof, M. O., Edwards, D. P., Cook, C. E., and McDonnell, D. P. (1996). 16a-Substituted anologs of the antiprogestin RU486 induce a unique conformation in the human progesterone receptor resulting in mixed agonist activity. Proc. Natl. Acad. Sci. U.S.A. 93, 87398744. Mahfoudi, A., Roulet, E., Dauvois, S., Parker, M. G., and Wahli, W. (1995). Specific mutations in the estrogen receptor change the properties of antiestrogens to full agonists. Proc. Natl. Acad. Sci. U.S.A. 92, 4206-4210. Tanenbaum, D. M., Wang, Y., Williams, S. P., and Sigler, E B. (1998). Crystallographic comparison of the estrogen and progesterone receptor's ligand binding domains. Proc. Natl. Acad. Sci. U.S.A. 95, 59986003. Shiau, A. K., Barstad, D., Loria, P. M., Cheng, L., Kushner, E J., Agard, D. A., and Greene, G. L. (1998). The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Cell (Cambridge, Mass.) 95, 927-937. Delmas, P. D., Bjarnason, N. H., Mitlak, B. H., Ravoux, A. C., Shah, A. S., Huster, W. J., Draper, M., and Christiansen, C. (1997). Effects of raloxifene on bone mineral density, serum cholesterol concentrations, and uterine endometrium in postmenopausal women. N. Engl. J. Med. 337, 1641 - 1647. Sato, M., Rippy, M. K., and Bryant, H. U. (1996). Raloxifene, tamoxifen, nafoxidine, or estrogen effects on reproductive and nonreproductive tissues in ovariectomized rats. FASEB J. 10, 905-912. Ashby, J., Odum, and Foster, J. R. (1997). Activity of raloxifene in immature and ovariectomized rat uterotrophic assays. Regul. Toxicol. Pharmacol. 25, 226-231. Norris, J. D., Paige, L. A., Christensen, D. J., Chang, C.-Y., Huacani, M. R., Fan, D., Hamilton, P. T., Fowlkes, D. M., and McDonnell, D. E (1999). Peptide antagonists of the human estrogen receptor. Science 285, 744-746. Paige, L. A., Christensen, D. J., Gron, H., Norris, J. D., Gottlin, E. B., Padilla, K. M., Chang, C.-Y., Ballas, L. M., Hamilton, E T., and McDonnell, D. E (1999). Estrogen receptor (ER) modulators each induce distinct conformational changes in ERa and ERfl. Proc. Natl. Acad. Sci. U.S.A. 96, 3999-4004.
11
CHAPTER 1 Estrogen and Progesterone Receptors 48. Wijayaratne, A. L., Nagel, S. C., Paige, L. A., Christensen, D. J., Norris, J. D., Fowlkes, D. M., and McDonnell, D. P. (1999). Comparative analyses of the mechanistic differences among antiestrogens. Endocrinology 140, 5828-5840. 49. Horwitz, K. B., Jackson, T. A., Bain, D. L., Richer, J. K., Takimoto, G. S., and Tung, L. (1996). Nuclear receptor coactivators and corepressors. Mol. Endocrinol. 10, 1167-1177. 50. Smith, C. L., Nawaz, Z., and O'Malley, B. W. (1997). Coactivator and corepressor regulation of the agonist/antagonist activity of the mixed antiestrogen, 4-hydroxytamoxifen. Mol. Endocrinol. 11, 657- 666. 51. Wagner, B. L., Norris, J. D., Knotts, T. A., Weigel, N. L., and McDonnell, D. P. (1998). The nuclear corepressors NCoR and SMRT are key regulators of both ligand- and 8-bromo-cyclic AMP-dependent transcriptional activity of the human progesterone receptor. Mol. Cell. Biol. 18, 1369-1378. 52. Onate, S. A., Tsai, S., Tsai, M.-J., and O'Malley, B. W. (1995). Sequence and characterization of a coactivator for the steroid hormone receptor superfamily. Science 270, 1354-1357. 53. Spencer, T. E., Jenster, G., Burcin, M. M., Allis, C. D., Zhou, J., Mizzen, C. A., McKenna, N. J., Onate, S. A., Tsai, S. Y., Tsai, M.-J., and
54.
55.
56.
57.
O'Malley, B. W. (1997). Steroid receptor coactivator- 1 is a histone acetyltransferase. Nature (London) 389, 194-198. Shibata, H., Spencer, T. E., Onate, S. A., Jenster, G., Tsai, S. Y., Tsai, M.-J., and O'Malley, B. W. (1997). Role of co-activators and corepressors in the mechanism of steroid/thyroid receptor action. Recent Prog. Horm. Res. 52, 141-165. H6rlein, A. J., N~iar, A. M., Heinzel, T., Torchia, J., Gloss, B., Kurokawa, R., Ryan, A., Kamei, Y., S6derstr6m, M., Glass, C. K., and Rosenfeld, M. G. (1995). Ligand-independent repression by the thyroid hormone receptor mediated by a nuclear receptor co-repressor. Nature (London) 377, 397-403. Chen, J. D., and Evans, R. M. (1995). A transcriptional co-repressor that interacts with nuclear hormone receptors. Nature (London) 377, 454-457. Lavinsky, R. M., Jepsen, K., Heinzel, T., Torchia, J., Mullen, T.-M., Schiff, R., Del-Rio, A. L., Ricote, M., Ngo, S., Gemsch, J., Hilsenbeck, S. G., Osborne, C. K., Glass, C. K., Rosenfeld, M. G., and Rose, D. W. (1998). Diverse signaling pathways modulate nuclear receptor recruitment of N-CoR and SMRT complexes. Proc. Natl. Acad. Sci. U.S.A. 95, 2920-2925.
~HAPTER
Ovarian Anatomy and Physiology GREGORY F.
ERICKSON
Department of Obstetrics and Gynecology, University of California, San Diego, La Jolla, California 92093
VI. Accelerated Loss in OR: Activin Hypothesis VII. New Data on the Effects of Activin VIII. Statement of Conclusion References
I. Introduction II. Statement of the P r o b l e m
III. The Primordial Follicle IV. The Adult Ovary V. Ovary Reserve
I. INTRODUCTION
II. STATEMENT OF THE P R O B L E M
A characteristic feature of reproductive capacity in women is its cyclical nature, a feature that is strikingly reflected by the growth and development of oocytes and follicles within the ovaries [1]. Typically, the tissues of the adult ovaries undergo dramatic cyclical changes, which in turn are reflected in cyclical changes in the production of the steroid hormones, estradiol (E 2) and progesterone (p4), as well as key regulatory proteins, such as inhibin. The regulated expression of these ligands is critically important in the mechanisms that underlie fertility in women, including the ovulation of an oocyte with the potential for producing a normal embryo. The process begins at puberty with the first cycle and ends at the menopause with the permanent cessation of menses. The causative event in the menopause is the disappearance of primordial follicles in the ovaries, i.e., the loss of ovary reserve (OR). Consequently, the size of a woman's OR (number of primordial follicles) has great impact on when the menopause begins. Therefore, to understand the menopause from either a physiological or a pathophysiological perspective, we must understand the interactions between OR, folliculogenesis, and the menstrual cycle during aging. This chapter reviews what is currently known about these relationships. MENOPAUSE: BIOLOGY AND PATHOBIOLOGY
Today, the menopause occurs in most women at --~51 years of age. Demographic studies have demonstrated that the mean life expectancy of women in developed countries [1] has increased from - 4 5 years in 1850 to - 8 2 years in 1998 (Fig. 1) [ l a]. This is an important observation because it indicates that most women today will live one-third to onehalf of their lives postmenopausally, i.e., they will live - 30 years after the menopause. Clinicians can therefore expect to extend care increasingly to large numbers of women of advanced reproductive age in whom ovarian dysfunction will be a major cause of infertility and morbidity. If one considers that the vast majority of fertility and gynecologic problems in the aging woman are a direct consequence of the age-related decrease in ovary reserve, it becomes apparent that the disappearance of primordial follicles is one of the critical events in the life of all women. One female function most adversely affected by the agerelated decrease in OR is decreased fecundity. The basis for this age-related change is the failure of dominant follicles to release eggs that can undergo normal embryonic development [2-5]. Results from in vitro fertilization (IVF) [6,7] have demonstrated that this decrease becomes particularly evident in patients after 36 years of age, when pregnancy 13
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
1
4
G
R
E
G
FIGURE 1 Changesin the life expectancy and age of the menopause in women over the past 150 years. From Nachtigall [la], with permission.
rates with self-produced oocytes decline sharply by --~65%, i.e., from ---25% per transfer in women ~<30 years of age to --~9% per transfer in women after the age of 36 years [7]. A similar age-related decrease in female fecundity has been found using artificial insemination with donor semen (AID) [8] and gamete intrafallopian transfer (GIFT) [9]. The low fecundity rate continues through --~44 years, after which viable pregnancies almost never occur [9-11 ]. These facts, together with the delay in childbearing by women in developed countries, have set the stage for an increase in reproductive problems and disorders attributable to female aging, in particular infertility. An important point is that women between 36 and 44 years of age can exhibit regular menstrual cycles [12]. This is important because it argues that the decrease in fecundity in these older women is not the result of the failure of the aged ovaries to produce dominant follicles, presumably the cyclical activity reflects the ability of these dominant follicles to undergo the physiological changes that typically occur during selection, ovulation, luteinization, and luteolysis. One of the main lines of evidence in support of this theory is that nearly normal quantities of androgen [ 13], estrogen, and progesterone [14,15] appear to be secreted from the aged dominant follicles during the menstrual cycle. This evidence supports the idea that the dominant follicles that develop in these older women are fully capable of expressing a nearly normal pattern of steroidogenic activity. By contrast, studies of oocytes from aged dominant follicles have demonstrated the existence of alterations that contribute negatively to pregnancy. For example, investigators have found that aneuploidy increases significantly in embryos that develop from oocytes isolated from the mature follicles of women >35 years of age [16]. Thus, one is led to the conclusion that (1) the endocrine and gametogenic function of dominant follicles can become dissociated in women after --~36 years of age and (2) the aberrant expression of cellular responses in the egg would appear to be the basis for the age-related decrease in fecundity. Understanding how the developmental potential of the
O
R
Y
F. ERICKSON
aged oocyte is altered independently of changes in the granulosa and theca cells is a fundamental question in ovary research. Although relatively little is known about this problem, an interesting role for ovary reserve (the number of primordial follicles) has been suggested from analysis of folliclestimulating hormone (FSH) and inhibin levels and pregnancy rates. An important point to emerge from the clinical studies is that the number of follicles within the ovaries, not oocyte age, is the main determinant predicting pregnancy in older women [9]. That is, the incidence of pregnancy with self-produced oocytes is highest in older women with adequate OR, i.e., those with no significant elevation in plasma FSH levels and whose ovaries contain a comparatively large number of mature Graafian follicles [6,13,17]. Given this relationship, it is not unreasonable to propose that the selective deteriorative changes that occur in aging oocytes are either correlated with or causally connected to a significant decrease in OR, rather than aging itself. There is a fundamental question: How does this occur?
IIl. THE PRIMORDIAL
FOLLICLE
Before addressing this question, we must understand some basic biology of the primordial follicles or OR. The primordial follicles represent a pool of nongrowing follicles from which all dominant preovulatory follicles are selected [1]. Thus, primordial follicles are, in a real sense, the fundamental reproductive units of the ovary. Morphologically, each primordial follicle is composed of an outer single layer of squamous epithelial cells that are termed granulosa or follicle cells, and a small (approximately 15/xm in diameter) immature oocyte arrested in the dictyotene stage of meiosis; both cell types are enveloped by a thin, delicate membrane called the basal lamina (Fig. 2). By virtue of the basal lamina, the granulosa and the oocyte exist in a microenvironment in which direct contact with other cells does not occur. Although small capillaries are occasionally observed in proximity to primordial follicles, these follicles do not have an independent blood supply [1]. The mean diameter of a nongrowing primordial follicle is 29/xm [18]. All the primordial follicles present in a woman's ovaries are formed before birth. Developmentally, the primordial follicles are formed in the cortical cords of the fetal ovaries between the sixth and ninth months of gestation [1]. During this period, all the germ cells are stimulated to initiate meiosis. Because the oocytes in the primordial follicles have entered meiosis, all oocytes that are capable of participating in reproduction during a woman's life are formed at birth, i.e., the human ovaries acquire a lifetime quota of eggs before birth. Soon after primordial follicle formation, some are recruited (activated) to initiate growth. As successive recruitment proceeds over time, the size of the pool of primordial follicles becomes progressively smaller (Fig. 3) [ 18a]. Between the times of birth and menarche the number
CHAPTER2 Ovarian Anatomy and Physiology
15 of primordial follicles (and thus oocytes) decreases from several million to several hundred thousand (Fig. 3). As a woman ages, the number of primordial follicles (OR) continues to decline, until at menopause they are difficult to find (Fig. 4) [ 18b].
IV. THE ADULT OVARY
FIGURE 2 Electron micrograph of a human primordial follicle showing oocyte nucleus (N), Balbiani body (.), and granulosa or follicle cells (arrowheads). From [ 1], Erickson, G. F. (1995). The Ovary: Basic Principles and Concepts. In "Endocrinology and Metabolism" (Felig, E, Baxter, J. D., Broadus, A. E., Frohman, L. A., Eds.), 3rd ed., pp. 973-1015. McGrawHill, with permission.
7.0'
6.0-
_o ; 5.0"
~ U
4.0'
1 ~
:3.0'
N 2.0" ~ Z
1.0" 0.6" 0.3.
g
9i :t 5 ~b
AGE(months pc) ]
2"o
-3"o-
AGE(yeors)
BIRTH
FIGURE 3 Changes in the total number of oocytes in human ovaries during aging. In the embryo at early to midgestation, the number of oocytes increases to almost 7 million. Shortly thereafter, the number falls sharply to about 2 million at birth. The enormous loss (--~70%) of oocytes in the embryo between 6 and 9 months is caused by apoptosis. The store of eggs continues to diminish with age until no oocytes are detected in the ovaries at about 50 years of age. From [18a], Baker, T. G. (1971). Radiosensitivity of mammalian oocytes with particular reference to the human female. Am. J. Obstet. Gynecol. 110, 746-761, with permission.
In this section, we will deal with the anatomy and physiology of folliculogenesis as it occurs in normal women during the reproductive years. We will focus our attention on the manner in which the developmental program is expressed in a recruited primordial follicle as it develops to the ovulatory stage or dies by atresia. An underlying principle of the human ovaries is that of the 2 million primordial follicles, only 400 of so will complete their development and undergo ovulation and corpus luteum formation; all others (99.9%) will die by atresia after recruitment (Fig. 5) [ 18c]. Therefore, the very essence of folliculogenesis is selection.
A. Anatomy The adult human ovary is a mass of follicles, luteal tissue, blood vessels, nerves, and connective tissue elements, all of which form a relatively heterogeneous assemblance of histological units. It is the continuous and progressive change in the follicles and corpora lutea that gives rise to the cyclical changes in the menstrual cycle. During the reproductive years, the normal human ovaries are oval-shaped bodies that each measure 2.5-5.0 cm in length, 1.5-3 cm in width, and 0.6-1.5 cm in thickness [1]. The medial edge of the ovary is attached by the mesovarium to the broad ligament, which extends from the uterus laterally to the wall of the pelvic cavity. The surface of the ovary is covered by an epithelial layer of cuboidal cells resting on a basement membrane. This layer, termed the germinal or serous epithelium, is continuous with the peritoneum. Underlying the serous epithelium is a layer of dense connective tissue termed the tunica albuginea. The ovary is organized into two principal parts: a central zone, the medulla, which is surrounded by a particularly prominent peripheral zone, the cortex (Fig. 6). Embedded in the connective tissue of the cortex are the follicles containing the female gametes, oocytes. The number and size of the follicles vary depending on the age and reproductive state of the female. The existence of follicles of different sizes (primordial, primary, secondary, tertiary, Graafian, and atretic)reflects specific changes associated with their growth, development, and fate. At the end of the follicular phase of the menstrual cycle, the Graafian follicle that reaches maturity secretes its ovum into the peritoneal cavity (Fig. 6). After ovulation, the wall of the ovulating follicle develops into an endocrine structure, the corpus luteum. If implantation does
16
GREGORY F. ERICKSON
FIGURE 4 Photomicrographs of the cortex of human ovaries from birth to 50 years of age. Arrowheads indicate small, nongrowing primordial follicles with a single layer of squamous granulosa cells. From Erickson [ 18b], with permission.
not occur, the corpus l u t e u m deteriorates and eventually bec o m e s a n o d u l e of dense connective tissue t e r m e d the c o r p u s a l b i c a n s . O t h e r cells in the cortex are the s t e r o i d o g e n i c cells t e r m e d s e c o n d a r y interstitial cells. T h e s e cells, which are derived f r o m the theca c o m p a r t m e n t of atretic follicles, are found in nests or cords t h r o u g h o u t the life of the female. At the m e d i a l b o r d e r of the cortex is a mass of loose connective tissue, the medulla. This tissue contains a n e t w o r k of
c o n v o l u t e d b l o o d vessels and associated nerves that pass through the connective tissue toward the cortex (Fig. 6). A distinct group of t e s t o s t e r o n e - p r o d u c i n g cells, the hilus cells, lie in the m e d u l l a at the hilum of the ovary [19]. The arterial supply to the ovary originates from two prin-
Cumulus with egg
Ovulation
2,000,000 Total
........:." ...~
Corpus L-uteum
"i~."..' ': : ": i:' ".: ". :7~.: : ..'! ~..
(~
Primordial Follicles -
ATRESIA
o,oo,,o~
_./
,o,,,.,.
400 1
,F S H
Ovulation--~Corpus Luteum
FIGURE 5 Folliculogenesis is a highly selective process. Of the 2 million primordial follicles at birth, only 400 or so are brought to ovulation and luteinization by FSH and LH. From [18c], Soules, M. R., and Bremmer, W. J. (1982). The menopause and climacteric: Endocrinologic basis and associated symptomatology. J. Am. Geriatr. Soc. 30, 547-561, with permission.
Ear ly Tertiary Follicle
Inter|titiol Cells
\ ry
Follicle
Cortex
Primordial Follicles
FIGURE 6 Diagram summarizing the anatomy and histology of the human ovary during the reproductive years. Development of the follicles and
corpus luteum occurs within the cortex, whereas the spiral arteries, hilus cells, and autonomic nerves are located in the medulla. From [1], Erickson, G. F. (1995). The Ovary: Basic Principles and Concepts. In "Endocrinology and Metabolism" (Felig, P., Baxter, J. D., Broadus, A. E., Frohman, L. A., Eds.), 3rd ed., pp. 973-1015. McGraw-Hill, with permission.
CHAPTER2 Ovarian Anatomy and Physiology cipal sources: one, the ovarian artery, arises from the abdominal aorta; the other is derived from the uterine artery [ 1]. These two vessels, which enter the mesovarium from opposite directions, form an anastomotic trunk and become a common vessel called the r a m u s o v a r i c u s artery. At frequent intervals this artery gives rise to a series of primary branches, which enter the hilum like teeth on a rake. In the hilum, numerous secondary and tertiary branches are given off to supply the medulla (Fig. 6).
17 THECA INTERNA
Luteinizing HormoneI
f
Cholesterol
~lic I..... ! l A"P
,
I Androstenedionel
1I
B. P h y s i o l o g y Typically, the human ovaries produce a single dominant follicle that secretes into the oviduct a mature egg that is ready to be fertilized at the end of the follicular phase of the menstrual cycle. Each dominant follicle begins with the recruitment of a primordial follicle into the pool of growing follicles. It is not known exactly how recruitment occurs, but it appears to be controlled by one or more local ovarian regulatory factors by autocrine/paracrine mechanisms. In a broad sense, all growing follicles can be divided into two groups, healthy and atretic, according to whether apoptosis (programmed cell death) is present in granulosa cells [20,21 ]. As a consequence of successive recruitments, the ovaries appear to always contain a pool of small Graafian follicles from which a prospective dominant follicle can be selected. Once selected, a dominant follicle typically grows and develops to the preovulatory state. Those follicles that are not selected become committed to die by atresia. 1. ENDOCRINOLOGY OF FOLLICULOGENESIS
Regardless of age, follicle growth and development are brought about by the combined action of FSH and luteinizing hormone (LH) on the follicle cells. FSH and LH bind to specific high-affinity receptors on the membranes of the granulosa and theca interstitial cells, respectively. The interaction of these ligands with their receptors activates signal transduction pathways that stimulate mitosis and differentiation responses in the granulosa and theca cells [22,23]. Physiologically, these signaling pathways act in parallel to regulate the expression of specific genes in a precise quantitative and temporal fashion. There are two major endocrine responses associated with folliculogenesis. The first is that the actions of FSH and LH stimulate the production of large amounts of estradiol by the dominant follicle. This important gonadotropin-dependent response is called the twogonadotropin/two-cell concept Fig. 7) [23a]. Because the estradiol response appears to be specific to a dominant follicle, the levels of plasma estradiol can be used as a marker for the status of the dominant follicle. The second endocrine response is the marked increase in the production of inhibin by FSH [24]. With respect to aging, observations support the possibility that localized changes in inhibin production may
1 J~ Androstenedione J~-/ m ~--~"~Basement Lami n
a
~
'
~
j/ ~
~
"
r">! Estrogen l //,~,~ . . . . ', tc, irculatlon/ ~
[Androstenedionel ~Aromatization
Induction
Follicle
Stimulating Hormone
l
( Follicular fluid )
GRANULOSA CELLS
FIGURE 7 The two-gonadotropin/two-cell concept of follicle estrogen production. From Erickson [23a], with permission.
play a role in the accelerated loss of OR. We will return to this issue later. 2. CHRONOLOGY
In women, folliculogenesis is a very long process [22]. In each menstrual cycle, the dominant follicle that is selected to ovulate originates from a primordial follicle that was recruited to grow about 1 year earlier. Whatever the course of development or the final destiny (atresia or ovulation), all follicles undergo various progressive changes (Fig. 8). The very early stages of folliculogenesis (class 1, primary and secondary; class 2, early tertiary) proceed very slowly. Consequently, it requires -->300 days for a recruited primordial follicle to complete the preantral or hormone-independent period. The basis for the slow growth is the very long doubling time (---250 hr) of the granulosa cells. When follicular fluid begins to accumulate at the class 2 stage, the Graafian follicle begins to expand relatively rapidly (Fig. 8). As the antral (hormone-dependent) period proceeds, the Graafian follicle passes through the small, (classes 3, 4, and 5), medium (classes 6 and 7), and large (class 8) stages. A dominant follicle that survives to the ovulatory stage requires about 4 0 - 5 0 days to complete the whole antral period. Selection of the dominant follicle is one of the last steps in
18
GREGORYF. ERICKSON
co~)'-"
'3
c
.LASS 4
L
c Ass 8
ZO
oV/; /
ol ~ ....;I / _ ~
; '--/#l
/ # .~f
~\
-
,~-
'
~
oA -~\+. "~ V-~X
%\
'
\
_.
-
~
,'
\
,1t~ # -o.\ ~
u'24%~ " ATRESIL/ ' 5 0 .... <'( ' ~ ~
~a~,
~
,
~<.,.X~
\
~\~
~
~O,,0~,,k J,' I"~ ,/ OVULATION
":~xoo
-,.'~ -
,
FIGURE 8 The chronology of folliculogenesis in the human ovary. Folliculogenesis is divided into two major periods, preantral (gonadotropin independent) and antral (FSH dependent). In the preantral period, a recruited primordial follicle develops to the primary/secondary (class 1) and early tertiary (class 2) stages, at which time cavitation or antrum formation begins. The antral period includes the small Graafian (0.9-5 mm, classes 4 and 5), medium Graafian 6 - 1 0 mm, class 6), large Graafian (10-15 mm, class 7), and preovulatory (16-20 mm, class 8) follicles. Time required for completion of preantral and antral periods is ---300 and ---40 days, respectively. Number of granulosa cells (gc), follicle diameter (mm), and atresia (%) are indicated. From Gougeon [24a], with permission.
the long process of folliculogenesis. The dominant follicle, which is selected from a cohort of class 5 follicles, requires - 2 0 days to develop to the stage wherein it undergoes ovulation. Those follicles that are not selected become atretic. Atresia can occur at each stage of graafian follicle development, but atresia appears most prominent in follicles at the class 5 stage (Fig. 8) [24a]. 3. SELECTION The dominant follicle that will ovulate its egg in the next cycle is selected from a cohort of healthy, small Graafian follicles (4.7 + 0.7 mm in diameter) at the end of the luteal
MEMBRANAcELLsGRANULOSA ~
~
phase of the menstrual cycle [22]. Morphologically, each cohort follicle contains a fully grown egg, about 1 million granulosa cells, a theca interna containing several layers of theca interstitial cells, and a band of smooth muscle cells in the theca externa (Fig. 9) [24b]. Selection is a quintessential aspect of ovary physiology. It is characterized by a high sustained rate of granulosa mitosis. Shortly after the midluteal phase, the granulosa cells in all the cohort class 4 and class 5 follicles show a sharp increase (---twofold) in the rate of granulosa mitosis [22]. The first indication that the prospective dominant follicle has been selected is that the granulosa cells of the chosen follicle continue dividing at a fast rate while proliferation slows in the nondominant cohort follicles. Because this distinguishing event is seen at the late luteal phase, it has been concluded that selection occurs at this point in the cycle. As mitosis and follicular fluid accumulation continue (Fig. 8), the dominant follicle grows rapidly during the follicular phase, reaching 6.9 + 0.5 mm at days 1-5, 13.7 + 1.2 mm at days 6-10, and 18.8 +_ 0.5 mm at days 11-14. In nondominant follicles, growth and expansion proceed more slowly, and with time, atresia becomes increasingly more evident (Fig. 8). Rarely does an atretic follicle reach ->9 mm in diameter, regardless of the stage in the cycle. FSH is obligatory for follicle selection, and no other ligand by itself can serve in this regulatory capacity. Physiologically, the mechanism of selection is causally connected to the secondary rise in plasma FSH (the primary FSH rise being the midcycle preovulatory surge of FSH and LH). The secondary rise in plasma FSH begins a few days before the progesterone and estradiol concentrations reach baseline values at the end of luteal phase, and it continues through the first week of the follicular phase (Fig. 10) [24c]. The importance of the secondary rise in FSH is demonstrated by the fact that the dominant follicle will undergo atresia if the FSH levels are decreased. Consequently, the secondary rise in FSH is obligatory for the selection of a dominant follicle that will ovulate in the next cycle. One of the major conse-
~
THECA INTERSTITIAL
CELLS
('FOLLICULAR) ANTRUM \ FLUID J
LOOSECONNECTIVE CORONARADIATA GRANULOSACELLS-
~ ~ ~'B '
CAPILLARIES ZONAPELLUCIDA CUMULUSOOPHOROUS GRANULOSACELLS THECAEXTERNA
FIGURE 9 Diagrammatic representation of the histologic architecture of a Graafian follicle. Reprinted from [24b], Mol. Cell. Endocrinol. 29, G. F. Erickson. Primary cultures of ovarian cells in serum-free medium as models of hormone-dependent differentiation, 2149. Copyright 1983 with permission from Elsevier Science.
CHAPTER2 Ovarian Anatomy and Physiology
19
quences of the secondary FSH rise is that a critical threshold level of FSH accumulates in the follicular fluid of the chosen follicle. In normal class 5 to class 8 follicles, the mean concentration of follicular fluid FSH increases from --~1.3 mIU/ ml (--~58 ng/ml) to --~3.2 mIU/ml (--~143 ng/ml) through the follicular phase [25]. In contrast, FSH concentrations are low or undetectable in the microenvironment of the nondominant cohort follicles. Thus, the selection and the continued growth of a dominant follicle involve a progressive increase in the concentration of FSH within its microenvironment. Once activated, the dominant follicle becomes dependent on FSH for its survival. The regulation of FSH levels in follicular fluid is totally obscure. FSH triggers a marked activation of mitosis and differentiation of the granulosa cells, which in turn is reflected in a progressive increase in estradiol and inhibin synthesis and follicular fluid accumulation (Fig. 10). One of the effects of the increased estradiol and inhibin production is that the secondary rise in FSH is suppressed (Fig. 10). When this oc-
curs, the concentration of FSH falls below threshold levels and the development of the nondominant follicles stops. It is noteworthy that mitosis in these atretic follicles can be markedly stimulated by treatment with human menopausal gonadotropin (hMG) during the early follicular phase. Thus, if FSH levels within the microenvironment are increased, the nondominant follicle could perhaps be rescued from atresia.
C. The Role of FSH The major FSH-dependent changes that occur during the development of the dominant follicle are summarized in Fig. 11 25a]. The granulosa cells are the only cell types known to express FSH receptors. It follows, therefore, that FSH-mediated effects in the dominant follicle are at the levels of the granulosa cells. In dominant follicles, the FSHinduced differentiation of the granulosa cells involves three major responses, increased steroidogenic potential, mitosis, and LH receptors. 1. S T E R O I D O G E N I C P O T E N T I A L
a A
<
_=
rmc
80
200
T
,o 40 20
TT
T
TTITT.,.T"-rT.,..r I I I''
T
150
="
100
m "1o
50
3
, -
o
0
b .-.
50
o E
40
-: ='
"
TTT
~,
~
1500
T
rn ." 1 ooo
c
|3 o , . . 20
5o0 10
0 C
~5".~ o
F
,
--
",". . . . . .
40
,
,
,
T
" 10
"r
-,- T T T ' r
20
T.,.
T
-n u~
"
5
14
0
""
10
-;4
-?'
()
7
vAvAvAvAV
vvl ~'r
Days relative to midcycle LH peak
FIGURE 10 The endocrinology of the luteal-follicular transition in women. Data are mean _+SE of daily serum concentrations of FSH, LH, estradiol (E2), progesterone (p4), and immunoreactive inhibin in women with normal cycles. Note the secondary rise in plasma FSH in the late luteal phase (----2 days before menses). From [24c], Groome, N. E, Illingworth, E J., O'Brien, M., Pai, R., Rodger, F. E., Mather, J. E, and McNeilly, A. S. (1996). Measurement of dimeric inhibin B throughout the human menstrual cycle. J. Clin. Endocrinol. Metab. 81, 1401-1405. 9 The Endocrine Society.
THECA-DERIVED L ANDROSTENEDIONE)
ADENYLATECYCLASE
0
- 15
30
0
~FSH)
""
~" "~ -I-
The FSH ligand interacts with its receptor on the granulosa cells, and the binding event is transduced into an intracellular signal via the heterodimeric guanine-nucleotide binding proteins (G proteins) (Fig. 11). The FSH-bound receptor activates t h e aGstimlatory (ceGs) protein, which activates adenylate cyclase to generate increases in production of cyclic adenosine monophosphate (cAMP), which in turn triggers protein kinase A (PKA) to phosphorylate a cAMPresponse element binding (CREB) protein or other related DNA-binding proteins. After phosphorylation, these proteins
GDP CCHOLESTEROO
AT~P
--
--
. . . . . . . . . . . . . . . . . . . . PROTEIN CPREGNENOLONE) KINASE 450SC
3[3-HSD~...... _~....;, ~) A ",
Q
, CPROGESTERONE~
(PROGESTERONE~
"
"~ [LH RECEPTOR]
CANDROSTENEDIONE) II P450AROM .).. '~ (ESTI ONO
---~" (1713-HSD CESTFIADIOL~
L,0o
(ESTR
FIGURE 11 Schematic model of the FSH signal transduction pathway in granulosa cells of a dominant follicle leading to increased steroidogenic potential, mitosis, and LH receptor expression. Redrawn from Erickson [25a], with permission.
20
GREGORY F. ERICKSON
bind to upstream DNA regulatory elements called AMP response elements (CREs), whereby they regulate gene transcription. In this regard the FSH signal mechanisms stimulate the expression of specific genes that control the level of estradiol production by the granulosa cells [26]. The major steroidogenic genes induced by FSH include those for P450 aromatase (P450arom) and 17fl-hydroxysteroid dehydrogenase (17fl-HSD) (Fig. 11). The temporal pattern of expression of these genes has an important role in generating the normal pattern of estradiol production by the dominant follicle during the follicular phase of the cycle (Fig. 10). FSH also acts on the granulosa cells to increase their potential for luteinization as reflected by in vitro experiments with cultured granulosa cells from human follicles at different stages of development [27]. The mechanisms by which this progesterone potential remains suppressed during folliculogenesis in vivo remains unknown, but a putative FSH-dependent luteinizing inhibitor has been proposed [27].
3. INDUCTION OF L H RECEPTOR
The ability of LH and human chorionic gonadotropin (hCG) to activate the ovulatory cascade in the dominant follicle is dependent on the expression of a large number of LH receptors on the granulosa cells [1]. Studies have clearly demonstrated an obligatory role of FSH in the induction of LH receptor (Fig. 11). A key feature of LH receptor expression in the granulosa layer is that it is suppressed throughout most of folliculogenesis. That is, the number of LH receptors remains low in granulosa cells during the early and intermediate stages of dominant follicle growth, but then increases sharply to very high levels at the preovulatory stage. The acquisition of LH receptors implies that when the LH ligand enters the microenvironment of the dominant follicle in the late follicular phase, it can act on the granulosa cells to regulate their function, perhaps even replacing FSH as the principal regulator of granulosa cytodifferentiation.
2. MITOSIS
The granulosa cells in the dominant follicle have the ability to divide at a relatively rapid rate throughout the follicular phase of the cycle, increasing from about 1 • 106 cells at selection to over 50 • 106 cells at ovulation [22,23]. Despite its overall importance to ovarian physiology, it remains unclear how granulosa proliferation is controlled. There is evidence in humans that FSH stimulates the rate of granulosa cell division in vivo and in vitro (Fig. 11), but the mechanism by which FSH stimulates mitosis is not understood.
D. T h e R o l e o f L H There are two hormones, LH and insulin, that regulate steroidogenesis in the interstitial tissue, and both function as stimulators of androgen production [28]. Each hormone interacts with a transmembrane receptor and the binding event is transduced into an intracellular signal that stimulates transcription and translation of specific steroidal genes (Fig. 12). Throughout the life of a woman, LH acts as a critical positive regulator of ovary androgen biosynthesis [28]. The L H -
ADENYLATECYCLASE
^^A
AAAAAA GTP
+? 9 Es-r,=.o.
I
.
.
.
.
.
,,.Oc,,
~~~ATP (cAMP)
I
P
+? . . . . . . . . . . .
..ocl.P " " ANDROSTENEDIONE) J 1 ~FOLLICuLAR CIRCULATION.,~-~==~=Ji~ FLUID
FIGURE 12 Diagramof the LH and insulin signal transduction pathways in ovarian interstitial cells leading to increased androstenedione production. Redrawn from Erickson [28], with permission.
CHAPTER2 Ovarian Anatomy and Physiology receptor interactions in the interstitial cells are critically important in estradiol production by virtue of their ability to promote the production of androstenedione, the P450arom substrate (Fig. 7). The action of the LH-receptor signaling pathway in the ovary interstitial cells results in the expression of a battery of genes leading to increased androgen synthesis (Fig. 12). The role of LH in stimulating androgen production has been intensively studied in women because of its involvement in infertility and hyperandrogenism, such as in polycystic ovary syndrome [29,30]. There are four families of interstitial cells in the human ovaries (Fig. 6), the theca interstitial cells (TICs), secondary interstitial cells (SICs), theca lutein cells (TLCs), and hilus cells (HCs). The TICs, SICs, and TLCs are related to each other by a developmental sequence occurring during folliculogenesis and luteogenesis, a process called thecogenesis [31]. The formation of the TICs, SICs, and TLCs involves a developmental; process that encompasses both proliferation and differentiation [28,32]. Because thecogenesis is accompanied by mitosis, it contributes to total interstitial mass and therefore total androgen potential. LH promotes androgen synthesis through activation of the LH/hCG receptor/AMP-dependent protein kinase A signal transduction pathway (Fig. 12). The heterotrimeric G proteins act as transducers that couple LH/hCG-bound receptors to adenylate cyclase, which forms the second messenger, cAME cAMP activates PKA, which in turn phosphorylates specific serine and threonine residues on substrate proteins. The phosphorylated proteins generate cytoplasmic and nuclear responses that can lead to increased steroidogenesis. Androstenedione is the principal steroid produced by TICs, and treatment with LH increases its production in a time- and dose-dependent manner [28]. This concept explains in part the regulated production of androstenedione in normal women and its overexpression in women with chronically elevated levels of plasma LH. At the molecular level, activation of the LH signaling cascade leads to the stimulation of gene transcription, most notably genes for P450c22 and P450cl 7 [33]. The fact that the level of transcription and translation of these genes increases during folliculogenesis argues that LH-induced differential gene expression plays a physiological role in androstenedione production by human TICs during the menstrual cycle. It has been known for many years that the rate-limiting step in steroidogenesis involves the translocation of cholesterol from the outer mitochondrial membrane to the inner mitochondrial membrane, where it is metabolized to pregnenolone by P450c22. This protein, called steroidogenic acute regulatory (STAR) protein, has been isolated and cloned [34]. An important concept is that StAR is obligatory for LH-induced steroidogenesis. In the human ovary StAR is expressed in the interstitial cells, including the TICs [35]. It is noteworthy that there is convincing evidence that insulin signaling plays a role in regulating interstitial cell function in women [28]. Insulin receptors with protein tyro-
21 sine kinase (PTK) activity have been demonstrated in human ovaries. In situ hybridization and immunohistochemical studies have revealed that insulin receptors are expressed in TICs of Graafian follicles (both dominant and cohort) and in SICs [36]. Insulin stimulates androgen production by isolated TICs and SICs, and the stimulation is believed to be mediated by the insulin receptor. Activation of the insulin receptor signaling pathway can function alone to increase TIC/SIC androgen production and, importantly, the pathway can interact with the LH receptor pathway to further enhance the signals evoked by each receptor (Fig. 12). The cross-talk between the insulin and LH receptor pathways could be clinically relevant because of the development of hyperandrogenism in women with hyperinsulinemia.
E. I n t r a o v a r i a n C o n t r o l As discussed, the development of the dominant follicle is directed by the endocrine hormones FSH and LH. These ligands bind to receptors that are coupled to the AMP/protein kinase A signal transduction pathways, which in turn are coupled to differential gene activity in a quantitative and temporal fashion. An important concept to emerge in the past decade is that growth factors, which are themselves products of the ovary, modulate (either amplify or attenuate) FSH and LH action. All growth factors are ligands that can act in an autocrine/paracrine manner to regulate the timing and degree of hormone-dependent folliculogenesis. This is the autocrine/paracrine or growth factor concept (Fig. 13) [36a]. There are five different classes of growth factors: insulin-like growth factor (IGF), transforming growth factor/3 (TGF-/3), transforming growth factor-a (TGF-c0, fibroblast growth factor (FGF), and cytokines; all five classes have been described within follicles of human ovaries [37]. The principle that arises from all the evidence is that growth factors act by autocrine and paracrine mechanisms to cause plus and minus changes that determine whether a follicle lives or dies. The current challenges are to understand how specific growth factor families exert control of ovary functions and how these modulations are integrated into the overall pattern of physiology and pathophysiology during the life of the female.
V. OVARY R E S E R V E As discussed earlier, the number of ovarian primordial follicles decreases with age from birth through the menopause (Fig. 3). Importantly, studies [38,39] of human ovaries have established the concept that the rate of loss of OR (primordial follicles) is not constant during aging, with a significant accelerated decrease in OR occurring at about 37 years in most women (Fig. 14).
22
GREGORY F. ERICKSON
GROWTH FACTOR
AUTOCRINE
PARACRINE
ENDOCRINE
FIGURE 13 Comparison of the autocrine-paracrine and endocrine concepts. H, Hormone. From [36a], Erickson, G. F. (1994). Non-gonadotropic regulation of ovarian function: Growth hormone and IGFs. In "Ovulation Induction: Basic Science and Clinical Advances (Excerpta Medica Int'l Congress Series)," (Filicori, M.,and Flanigni, C., Eds.), pp. 73-84. With permission from Elsevier Science.
MENARCHE I
I BIRTH I
0
ACCELERATED DECREASE IN OVARY RESERVE
12
37
51
AGE IN Y E A R S FIGURE 14
Major events during a woman's life that impact on fertility and fecundity.
A. Regulation Morphometric analysis of normal ovaries has demonstrated that the rate of recruitment (initiation of primordial follicle growth) accelerates sharply in women at --~38 years of age. Consequently, there is a biexponential decrease in OR in women [38,39]. It can be seen (Fig. 15) that the number of primordial follicles decreases steadily for more than three decades, but when the pool of primordial follicles reaches a critical number o f - 2 5 , 0 0 0 at 37.5 + 1.2 years, the rate of loss of primordial follicles accelerates ---twofold. Consequently, the OR is reduced to 1000 primordial follicles at --~51 years of age [38,39], which corresponds to the median age at natural menopause [40]. If the earlier rate of decrease in primordial follicles persisted, menopause would not be expected until the female reached 71 years of age. An important point in this natural process is that the number of primordial follicles within the ovaries of any given woman who reaches 38 years of age is variable, i.e., important individual differences in OR exist. As seen in Fig. 15, some women reach the critical threshold of 25,000 primordial follicles in their late twenties, whereas others do not reach this threshold until their forties. It seems therefore that age alone has limited predictive value for accurately determining a woman's OR. The significance of this variability is demon-
strated by the fact that women who continue to menstruate regularly after the age of 45 have 10 times more primordial follicles than do those with irregular menses [41 ]. Further, a higher level of primordial follicles is functionally coupled with a higher pregnancy rate in older women [6,13,17]. It can be argued, therefore, that the OR determines the number of maturing Graafian follicles, which in turn determines menstrual activity, which in turn determines fecundity. In a real sense, OR may be of greater importance than a woman's chronological age in predicting fertility. If we accept this argument, then OR, not age, would be the fundamental factor in determining the decrease in fecundity at - 3 8 years. Hence the question: What is the underlying mechanism for the accelerated recruitment at --~38 years of age? Although the answer to this question is not known, it is reasonable to assume that regulatory molecules are involved. In this regard, there are two possibilities: the decrease of one or more necessary inhibitors and/or the increase in one or more stimulators. Despite its physiological importance, very little is known about the mechanisms of recruitment in any species. Evidence from animal studies indicates that the rate of recruitment of primordial follicles can be influenced by several regulatory factors (Table I). One particularly interesting result with aging rats is that the rise in plasma FSH follow-
CHAPTER 2 Ovarian Anatomy and Physiology
23
i06~ §247
§
§
~ r
+ ~+
~ I0
.~ N
10
+
+4-
§
5 -
4
"6 ~ E Z
=
§ %.11o
_ 3
10
+
,. ~~,~~ ' ~
0 o
10
2
0
!
|
i0
|
20
30
|
40
o
!
50
Age (years)
FIGURE 15 The age-related decrease in the total number of primordial follicles within both human ovaries, from birth to the menopause. As a result of recruitment (initiation of primordial follicle growth), the number of primordial follicles decreases progressively from --~1,000,000 at birth to 25,000 at 37 years. At age 37, the rate of recruitment increases sharply, and the number of primary follicles declines to 1000 at the menopause, i.e., at about 51 years. From Faddy et al. [38], with permission.
ing unilateral ovariectomy is associated with a significant reduction in the number of primordial follicles within the ovaries; the most striking feature is that the effect was observed only in old rats [42]. These studies support the potentially important concept that increased levels of FSH might function to accelerate the rate of recruitment during aging. Hence the question: To what extent is the accelerated loss of OR in aging women a consequence of increased circulating FSH? This is an important question because we know that a significant elevation of plasma FSH is observed in women when the loss of OR is accelerated at --~38 years [11,17], and the increased plasma FSH corresponds to the time that fecundity drops. Nonetheless, the question of whether the age-related increase in FSH in
TABLE I Known Modulators of Primordial Follicle Number in Laboratory Animals Regulator Follicle-stimulating hormone Thymus removal Starvation Growth hormone/prolactin Morphine sulfate Epidermal growth factor
Effect on ovary reserve
Ref.
Decrease Decrease Increase Increase Increase Increase
42 43 44 44 46 47
women is causally connected to the stimulation of recruitment remains unanswered. Experiments in mice show that the rate of recruitment can be modulated by several factors, including the thymus, restricted food intake, prolactin (PRL) and/or growth hormone (GH), opiates, and epidermal growth factor (EGF) (Table I). Experiments with neonatal mice indicate that thymectomy leads to a dramatic loss of primordial follicles by apoptosis [43]. Because apoptotic primordial follicles are rarely seen in aging women [38,39], it seems unlikely that a thymus factor plays an important role in the accelerated loss of OR at 38 years. In another study, a 50% reduction in food intake was found to increase the number of primordial follicles, suggesting starvation may increase the OR [44]. This could be potentially important in humans, but the question of whether starvation elicits a similar effect in women needs to be carefully examined. Studies using the sterile Snell dwarf mouse indicate that their ovaries contain significantly more primordial follicles than do those of the wild type. Precisely how this occurs is uncertain, but it has been theorized that the endocrine state resulting from chronically low GH and/ or PRL might be involved [45]. Finally, experiments done in mature mice have shown that the administration of either morphine sulfate [46] or epidermal growth factor [47] leads to a sustained reduction in the rate of primordial follicle recruitment, followed by an increase in OR. These studies, albeit limited, support the concept that the
24
GREGORY F. ERICKSON
rate of recruitment can be modulated, either increased or decreased, by regulatory elements. Although the clinical significance of these animal data is unknown, they raise the intriguing idea that it may be possible to slow down the rate of recruitment. If true, these data could have important implications for increasing the OR, which in turn could have important implications for fecundity in older women.
40'
I.H
35-
---*-----o-.-
Old~rLH (n-I1) YoungerLH (n.11)
30-
~
25-
E 20"
r
~'15" 10-
B. E n d o c r i n e P a r a m e t e r s
5 .
At the present time, there is considerable interest in the hypothesis that increased plasma FSH concentrations consequent to decreased ovary inhibin A and B production may be involved in the mechanisms underlying accelerated recruitment and decreased fecundity in women after 36 years of age (see also Chapter 9). Therefore, to understand the physiological mechanism underlying the age-related increase in FSH, one needs to understand the structure of inhibin and agerelated changes in inhibin production in women. Inhibin is a member of the TGF-fl superfamily [48]. Inhibin molecules are composed of two heterodimeric proteins, a common ce subunit and one of two distinct/3 subunits termed flA or fiB (Fig. 16). The two subunits (a and flA or fiB) are held together by disulfide bonds, producing two different inhibins termed inhibin A and inhibin B. By contrast, the activins are built of two types of the/3 subunits, generating dimeric proteins called activin A, AB, or B. It should be mentioned here that the differential regulation of ce subunit expression might be expected to have profound effects on the levels of inhibin and activin produced by the ovary; that is, a high and low level of a subunit expression would be expected to result in relatively high and low levels of inhibin and activin production, respectively (Fig. 16). It is now clear that a monotropic rise in FSH occurs in women during aging [12]. The rise in FSH, which precedes that of LH by almost 10 years, becomes detectable after 36 years of age [14]. A detailed examination of FSH and LH concentrations in young and old women during the cycle [49] revealed that serum FSH, but not LH, is significantly elevated in older women throughout the menstrual cycle (Fig. 17). It is certainly of interest that the increase in plasma
Inhibin A
Inhibin B
1
s
Inhibins
Activin A
Activin AB
Activin B
Activins
FIGURE 16
A model of the inhibin and activin molecules.
.
.
.
|
.
.
.
.
i
.
.
.
.
9
-
-
- " -
!
.
.
.
.
i
-
-
-
'-
15 FSH
12'
9.
.Ir
e-
0
9
-'~ 5
-"
"-''-
-
i
-10
.
.
.
.
!
-5
.
.
.
.
|
0
.
.
.
.
9
.
.
.
5
.
"
|
10
-"
-
-
-
15
Cycle Days Relative to LH Surge
FIGURE 17 The daily serum FSH and LH levels throughout the menstrual cycle of 11 women in each group (mean +_ SE). Reprinted by permission from [49], The gonadotropin secretion pattern in normal women of advanced reproductive age in relation to the monotropic FSH rise, by Klein, N. A., Battaglia, D. E. Clifton, D. K., Bremner, W. J., and Soules, M. R., Journal of the Society for Gynecologic Investigation, 3, 27-32. Copyright 1996 by the Society for Gynecologic Investigation.
FSH coincides with the accelerated loss in OR. Presumably, some alteration has occurred in the negative feedback mechanism of FSH production in aging women, which is reflected in an increase in plasma FSH levels. The most likely explanation for this observation is that aging in women leads to a significant decrease in inhibin production. Direct evidence that the changing FSH profiles in aging women are accompanied by a concomitant decrease in plasma inhibin during the follicular phase of the cycle has been reported [50]. The strong evidence that inhibin exerts a negative feedback effect on pituitary FSH secretion in animals [51,52] supports the theory that decreased ovary inhibin production might be responsible for the increased FSH levels in women after 36 years, which in turn might be responsible for the decreased fecundity that can occur at this time. Direct evidence to support this theory has come from studies [ 15] showing that women aged 35 years or more produce less inhibin in response to exogenous gonadotropin than do women less than 35 years (Fig. 18). By contrast, no significant differences in plasma estradiol (and progesterone) are detectable in these women (Fig. 18). Studies in normal
CHAPTER 2 Ovarian Anatomy and Physiology
25
j•••••••••••••••••M •••••••••••••••••••••••••••••••••••••M •••••••••M u••u M ••••••M •••••••••••••••••••••••••!
hM G
. . . . . . . . . . . . . . .
<
30
:
30-34 3000
>
hCG
/
INHIBIN
AGE
4000
. CC
3
:
"-
-"
=
=
/
c 2000
1000
i
9
AGE
i
1
1
ESTRADIOL
i
A
. ~
< 30 30 - 3 4 >
t.
z~
35
-5
-4
-3
Day
-2
of
-1
O
+17"
treatment
FIGURE 18 The selective decrease in the inhibin response during ovarian hyperstimulation with human menopause gonadotropin (hMG) in aging women. Inhibin levels, but not estradiol levels, are significantly lower in women 35 years or older, hMG, Human menopausal gonadotropin; CC, clomiphene citrate; hCG, human chorionic gonadotropin. From [15], Hughes, E. G., Robertson, D. M., Handelsman, D. J., Hayward, S., Healy, D. L., DeKretser, D. M. (1990). Inhibin and estradiol responses to ovarian hyperstimulation: Effects of age and predictive value for in vitro fertilization outcome. J. Clin. Endocrinol. Metab. 70, 358-364. 9 The Endocrine Society.
cycling women have revealed a selective decrease in plasma inhibin levels during the follicular phase of the menstrual cycle, beginning at ---36 years of age [15,50]. It should be noted that the fully processed dimeric inhibin A has been shown to be the predominant circulating form in women before and after treatment with menopause gonadotropin [53]. Thus, a functional link between aging in women and decreased expression of ovary inhibin A is suggested (see Fig. 10). However, an important point to emerge from inhibin studies [24c] in normal women during the cycle indicate that important changes in inhibin B can be detected during the luteal-follicular transition (Fig. 19). Indeed, Klein et al. [54] have presented evidence for a role of decreased inhibin B in the monotropic FSH rise in aging women. From all these data, it seems reasonable to propose that a decreased ability of the ovaries to produce inhibin B (and perhaps inhibin A as well) may be the underlying cause of the monotropic rise in FSH in women after ---35 years of age. The question of whether these changes in inhibin and FSH negatively affect the egg has not been resolved. However, the data fit with a prediction that the decrease in inhibin, which begins around the time of the accelerated recruitment
at 37 years of age, may be involved, directly or indirectly, in the mechanisms that cause poor oocyte quality in aging women. What cells in the ovary are responsible for inhibin production? Studies using in situ hybridization and immunohistochemistry have documented the tissue-specific expression of the ce,/3A, and/3B subunits of inhibin in normal human ovaries during the menstrual cycle [55,56]. Yamoto et al. [55] found that the three inhibin subunit proteins are selectively expressed in the granulosa cells of growing follicles; however, there are important differences between the preantral and Graafian follicle stages. Here, the most striking difference is that granulosa cells in preantral follicles express relatively high levels of the/3A and/3B proteins, but the ce subunit proteins appears undetectable [55]. By contrast, the granulosa cells in the healthy antral follicles express all three subunit proteins, and the levels of ce,/3A, and/3B proteins become very high in the dominant follicle, particularly at the preovulatory stage [55]. By virtue of the different pattern of ce subunit expression during folliculogenesis, it would appear that granulosa cells in preantral follicles (primary, secondary, and early tertiary) produce activin, whereas those in healthy Graafian (antral) follicles produce inhibin. The question of whether the dominant follicle produces activin remains to be answered. An in vivo study on inhibin secretion by human ovaries presents a compelling case that the entire pool of healthy
II
4O ~"
T
30
75
/ _
-!
]: 0
|
!
|
'
_
:
:
!
!
|
b 7"
6"
:{ g
U.
-8OO
T
m
5 4 3-
-20o
2" J.
'~-o
1 Days relative to FSH peak
FIGURE 19 Plasma concentrations of inhibin A and B, and estradiol and FSH during the luteal-follicular transition in normal cycling women. Data were aligned with respect to the day of the intercycle FSH peak (mean _ SE). From [24c], Groome, N. P., Illingworth, P. J., O'Brien, M., Pai, R., Rodger, E E., Mather, J. P., and McNeilly, A. S. (1996). Measurement of dimeric inhibin B throughout the human menstrual cycle. J. Clin. Endocrinol. Metab. 81, 1401-1405. 9 The Endocrine Society.
26
GREGORY F. ERICKSON
Graafian follicles synthesize and secrete inhibin [57]. First, the concentration of inhibin is higher in the ovarian veins than in the peripheral circulation during the normal menstrual cycle. This finding indicates that plasma inhibin comes from the ovaries. Second, the amount of inhibin secreted during the late follicular phase is similar in the veins of both ovaries [57]. Therefore, it seems likely that the Graafian follicles in both ovaries (dominant and nondominant) secrete inhibin during the follicular phase. These data implicate all follicles (healthy and atretic) in the production of inhibin during the follicular phase of the menstrual cycle, i.e., peripheral inhibin levels reflect the total number of developing Graafian follicles. It is possible that this concept could have relevance to the lower levels of inhibin seen in older women. It has been shown that women exhibit an age-related decrease in the total number of Graafian follicles in the ovaries. Collectively, these observations offer a plausible explanation for the reduced levels of circulating inhibin after 36 years of age, i.e., fewer Graafian follicles in turn results in lower plasma inhibin levels, leading in turn to increased FSH levels. Another possible explanation is that a decrease in the expression of the a and/or the/3 subunits in the granulosa cells plays a role in the lower level of circulating inhibin in women after 36 years of age. This possibility implies that an age-related defect or alteration occurs in the granulosa cells, which leads to the underexpression of inhibin (but not estradiol) after --~35 years. Indeed, there is evidence from studies with cultured granulosa lutein cells to support this idea [58,59].
VI. ACCELERATED ACTIVIN
L O S S I N OR:
HYPOTHESIS
It is now of interest to discuss the potential causal connection between the monotropic rise in FSH and the mechanisms underlying the accelerated loss of OR at --~38 years of age. The existence of/3A and/3B proteins in the granulosa cells of secondary and early tertiary follicles argues that these proteins serve some function in preantral folliculogenesis in humans. The fact that the a subunit proteins appear undetectable in these follicles suggests that they may dimerize to form activin [55]. Therefore, one could hypothesize that activin may be an autocrine/paracrine regulator of preantral folliculogenesis. This is of interest because activin appears to be a potent inducer of FSH receptors in granulosa cells [60]. Furthermore, it has been shown that activin accelerates folliculogenesis [61,62]. These observations support the possibility that activin may play a part in the accelerated loss of OR through increasing granulosa FSH sensitivity, which could in turn may play a role in the pathogenesis of the egg in an old dominant follicle by causing a premature overexpression of its development. How might this occur? Three different isoforms of activin (Fig. 16), isolated
from porcine follicular fluid [63-65], have been shown to be disulfide-linked homodimers of the inhibin/3A subunit (activin A; M r 24,000) or the /3B subunit (activin B; M r 22,000), or a heterodimer composed of a/3A and/3B subunit protein (activin AB; M r 23,000). The isoforms are present in equimolar concentrations in follicular fluid pooled from all antral follicles [65]. So far, there is no evidence for activin in follicular fluid of dominant follicles. Originally, activin was found to be a stimulator of FSH secretion in vitro [63,64] and in vivo [65-69]; however, subsequent studies demonstrated that activin exerts a wide range of positive and negative effects in many different target cells [70]. Activin achieves these effects by binding to a novel family of transmembrane receptors with protein serine/threonine kinase activity [71]. In women, plasma levels of free activin are low and do not change substantially during the cycle [72]. In women, plasma levels of free activin are low and do not change substantially during the cycle [72]. Thus, it seems likely that activin regulates follicular function physiologically by autocrine/paracrine mechanisms. It has been shown that developing follicles indeed produce and respond to activin. As discussed earlier, the/3A and /3B subunits are selectively expressed in human granulosa cells of healthy follicles between the secondary and preovulatory stages [55,56]. It seems likely that in the absence of the a subunit, activin may function in initiating or maintaining the growth and development of preantral follicles during the gonadotropin-independent stages of folliculogenesis. Studies in the rat have shown that FSH can stimulate activin expression in granulosa cells in vivo [73,74], and convincing evidence that rat granulosa cells from preantral follicles actually secrete dimeric activin has been reported [75]. Further, the mRNAs for activin receptor subtype II (Act RII and Act RIIB) have been identified in rat follicles [76,77], being present in the oocytes and granulosa cells [78]. Moreover, specific binding of radiolabeled activin to these cells has been demonstrated [79-81 ]. Collectively, these results support the hypothesis that human granulosa cells in preantral follicles may produce and respond to activin, and importantly, this process may be amplified by FSH. Much of our understanding of the biological effects of activin in the ovary has come from studies in laboratory animals. There is evidence suggesting that the autosecretion of activin may play a role in regulating follicle growth and development. Most striking is the observation that activin is a potent stimulator of FSH receptor expression in rat granulosa cells [77,82]. Thus far, activin is the only ligand known to induce FSH receptors. This may have relevance to the acquisition of FSH receptors in the granulosa cells, which occurs early in preantral follicle development, e.g., at the primary and secondary stages [83,84]. Another important effect of activin is that it can prevent FSH-induced receptor down-regulation [85]. Therefore, the concept emerging is that the activin produced by granulosa cells might play an
27
CHAPTER 2 Ovarian Anatomy and Physiology
important physiological role in the induction and maintenance of FSH receptors in the granulosa cells during folliculogenesis (Fig. 20) [86]. How might this situation impact the OR and fecundity in aging women? Because FSH stimulates activin production and the FSH levels are elevated in women after 36 years of age, one could postulate that these two elements might act synergistically to accelerate the rate of granulosa cytodifferentiation and folliculogenesis in aging ovaries with respect to the OR, and one could propose the following cascade process. The granulosa cells in preantral follicles synthesize, secrete, and respond to intrinsic activin. One major response to the autosecretion of activin is the expression and maintenance of FSH receptors. The relatively high level of FSH after 36 years has a stimulatory effect on the autocrine activin mechanism. This results in a synergistic interaction between the two signal transduction pathways, which leads to accelerated growth and differentiation responses in the granulosa cells. In this hypothesis, the relatively high amounts of activin could have a strong stimulatory effect on oocyte development in the presence of high FSH. These potent stimulatory effects are then theorized to produce an "overripe" egg lacking a normal meiotic spindle in the aged dominant follicle. Clearly, further work is needed to test the validity of this new hypothesis. Nonetheless, this idea is consistent with the data of Gougeon et al. [39] showing an accelerated loss of developing preantral follicles in women after 37 years. Further, both the fact that activin and FSH interact to in-
crease markedly LH receptor [87] and estradiol production [88] in rat granulosa cells, and the fact that activin can accelerate meiotic maturation [89] and promote antrumlike formation [62], e.g. accelerate FSH-dependent granulosa cytodifferentiation, are consistent with this hypothesis. There is evidence that activin can influence physiological responses in vivo. Doi et al. [90] found that FSH action in vivo can be amplified by exogenous activin. That is, injected activin enhances follicle growth, FSH receptor, number, and estradiol production in intact and hypophysectomized immature rats. These observations are important because they demonstrate that the positive effects of activin on granulosa and other follicle processes observed in vitro will also occur in vivo. Therefore, these results further support the contention that interactions between the autocrine growth factor, activin, and the elevated FSH levels might have a strong stimulatory effect on granulosa cells, which leads to the acceleration of follicle growth and development after 36 years. Interestingly, the fact that the length of the follicular phase of the menstrual cycle is significantly shorter in older women [17] fits with this prediction. It should be mentioned that negative effects of activin on folliculogenesis have also been reported. Foremost is the study by Woodruff et al. [69], who showed that activin injection into the ovary bursa of immature rats caused oocyte degeneration, granulosa pyknosis, and decreased mitosis. Therefore, it is also possible that high levels of activin might induce atresia and trigger oocyte demise in the rat.
.$D~T~LAT! C Y C I A ~
AAAAAA I I I I
GTP GTP MP
ATP
r
r
D~m~
~3ATION
s) + ~ zeeeptm
FIGURE 20 The activin autocrine hypothesis for accelerated granulosa cytodifferentiation and folliculogenesis. From Erickson [86], with permission.
28
GREGORY F. ERICKSON VII.
NEW
DATA
ON THE
EFFECTS
OF ACTIVIN The results of our study of activin action in adult cycling rats are relevant to this new hypothesis [61 ]. We found that the administration of recombinant human activin A to rats produced dramatic structure/function changes in folliculogenesis. The most striking results are as follows. First, activin stimulated a twofold increase in the n u m b e r of large Graafian follicles during the follicular phase of the cycle. The data suggested that activin increased the size of the pool of early tertiary preantral follicles, and their growth and development to the preovulatory stage [61]. Interestingly, nearly all of these large follicles contained apoptotic granulosa cells and therefore they were classified as atretic (Fig. 5). Based on these results, we conclude that activin provides a multifunctional stimulus in v i v o that includes both the stimulation and the inhibition of follicle cell activities. Second, these large atretic follicles ovulated prematurely, i.e., --~24 hr earlier than normal. Histologically, the ovulatory changes evoked by activin paralleled those described for normal physiological o v u l a t i o n m t h e c a l swellings, the initiation of germinal vesicle breakdown, cumulus expansion, stigma formation, release of egg cumulus complexes, and morphological luteinization of the follicle wall [61 ]. These observations provide the first evidence that a ligand, namely activin, can significantly shorten (by 25%) the length of the follicular phase of the normal estrous cycle. This necessarily implies that dominant follicle development and ovulation were accelerated in response to activin administration. Third, we found that the activin-exposed eggs in the oviducts and in the large ovulating follicles were arrested in metaphase I and appeared degenerate (Fig. 6). This finding confirms and extends other studies showing that activin acts in the rat ovary to affect oocyte quality negatively. There is evidence that A C T RII receptors are strongly expressed in the rat oocytes [78] and that activin can accelerate meiotic maturation in isolated rat oocytes [89]. Therefore, this negative action of activin might be mediated by the activin signaling pathway present in the rat egg. The mechanisms and the physiological/pathophysiological implications for the multifunctional actions of activin remain to be elucidated. Nevertheless, our observations support the proposition that the autosecretion of activin may contribute to an acceleration of follicle development that could result in the premature ovulation of overripe eggs in cycling w o m e n by autocrine/paracrine mechanisms.
VIII.
STATEMENT
OF CONCLUSION
F r o m the preceding discussion, it is clear that the primary problem in the dominant follicle that leads to reduced fe-
cundity in older w o m e n is the susceptibility of the egg to meiotic nondisjunction and aneuploidy. A potentially important theory to explain the problem was developed in this discussion. Evidence indicating that an age-related decrease in the production of ovary inhibin leads to a monotropic rise in FSH, which in turn is reflected in the acceleration of the loss of OR by virtue of accelerating the rate of recruitment, was discussed. Further, it was suggested that specific interactions between granulosa-derived activin and increased FSH receptor and ligand may act synergistically to further accelerate the rate of granulosa and oocyte cytodifferentiation: this functional response might then lead to accelerated development of the dominant follicle, which in turn is reflected in the age-related shortening of the follicular phase. At the level of the oocyte, these changes are reflected in an increased potential for aneuploidy.
References 1. Erickson, G. E (1995). The ovary: Basic principles and concepts. In "Endocrinology and Metabolism" (P. Felig, J. D. Baxter, A. E. Broadus and L. A. Frohman, eds.), 3rd ed., pp. 973-1015. McGraw-Hill, New York. l a. Nachtigall, L. E. (1995). The aging woman. In "Gynecology and Obstetrics" (J. J. Sciarra, ed.), Vol. 1, Chap. 28. Lippincott, Philadelphia. 2. Sauer, M. V., Paulson, R. J., and Lobo, R. A. (1990). A preliminary report on oocyte donation extending reproductive potential to women over 40. N. Engl. J. Med. 323, 1157-1160. 3. Navot, D., Bergh, P. A., Williams, M. A., Garrisi, G. J., Guzman, I., Sandier, B., and Grunfeld, L. (1991). Poor oocyte quality rather than implantation failure as a cause of age-related decline in female fertility. Lancet 337, 1375-1377. 4. Sauer, M. V., Paulson, R. J., and Lobo, R. A. (1993). Pregnancy after age 50: Application of oocyte donation to women after natural menopause. Lancet 341, 321-323. 5. Sauer, M. V., Miles, R. A., Dahmoush, L., Paulson, R. J., Press, M., and Moyer, D. (1993). Evaluating the effect of age on endometrial responsiveness to hormone replacement therapy: A histologic, ultrasonographic, and tissue receptor analysis. J. Assist. Reprod. Genet. 10, 47-52. 6. Padilla, S. L., and Garcia, J. E. (1989). Effect of maternal age and number of in vitro fertilization procedures on pregnancy outcome. Fertil. Steril. 52, 270-273. 7. Piette, C., de Mouzon, J., Bachelot, A., and Spira, A. (1990). In-vitro fertilization: Influence of women's age on pregnancy rates. Hum. Reprod. 5, 56-59. 8. Schwartz, D., and Mayaux, M. J. (1982). Female fecundity as a function of age. N. Engl. J. Med. 306, 404-406. 9. Qasim, S. M., Karacan, M., Corsan, G. H., Shelden, R., and Kemmann, E. (1995). High-order oocyte transfer in gamete intrafallopian transfer patients 40 or more years of age. Fertil. Steril. 64, 107-110. 10. Penzias, A. S., Thompson, I. E., Alper, M. M., Oskowitz, S. P., and Berger, M. J. (1991). Successful use of gamete intrafallopian transfer does not reverse the decline in fertility in women over 40 years of age. Obstet. Gynecol. 77, 37-39. 11. Wood, C., Calderon, I., and Crombie, A. (1992). Age and fertility: Results of assisted reproductive technology in women over 40 years. J. Assist. Reprod. Genet. 9, 482-484. 12. Sherman, B. M., and Korenman, S. G. (1975). Hormonal characteris-
29
CHAPTER 2 Ovarian A n a t o m y and P h y s i o l o g y
13. 14.
15.
16.
17.
18.
18a.
18b. 18c.
19.
20.
21. 22. 23.
23a. 24.
24a. 24b.
24c.
25.
25a.
tics of the human menstrual cycle throughout reproductive life. J. Clin. Invest. 55, 699-7067. Navot, D., Rosenwaks, Z., and Margolioth, E. J. (1987). Prognostic assessment of female fecundity. Lancet 2, 645-647. Lee, S. J., Lenton, E. A., Sexton, L., and Cooke, I. D. (1988). The effect of age on the cyclical patterns of plasma LH, FSH, estradiol and progesterone in women with regular menstrual cycles. Hum. Reprod. 3, 851-855. Hughes, E. G., Robertson, D. M., Handelsman, D. J., Hayward, S., Healy, D. L., and DeKretser, D. M. (1990). Inhibin and estradiol responses to ovarian hyperstimulation: Effects of age and predictive value for in vitro fertilization outcome. J. Clin. Endocrinol. Metab. 70, 358-364. Munne, S., Alikani, M., Tomkin, G., Grifo, J., and Cohen, J. (1995). Embryo morphology, developmental rates, and maternal age are correlated with chromosome abnormalities. Fertil. Steril. 64, 382-391. Scott, R. T., Toner, J. P., Muasher, S. J., Oehninger, S., Robinson, S., and Rosenwaks, Z. (1989). Follicle-stimulating hormone levels on cycle day 3 are predictive of in vitro fertilization outcome. Fertil. Steril. 51, 651-654. Gougeon, A., and Chainy, G. B. N. (1987). Morphometric studies of small follicles in ovaries of women at different ages. J. Reprod. Fertil. 81, 433-442. Baker, T. G. (1971). Radio sensitivity of mammalian oocytes with particular reference to the human female. Am. J. Obstet. Gynecol. 110, 746-761. Erickson, G. E (1986). An analysis of follicle development and ovum maturation. Semin. Reprod. Endocrinol. 4, 233-254. Soules, M. R., and Bremner, W. J. (1982). The menopause and climacteric: Endocrinologic basis and associated symptomatology. J. Am. Geriatr. Soc. 30, 547-561. Erickson, G. F., Magoffin, D. A., Dyer, C. A., and Hofeditz, C. (1985). The ovarian androgen producing cells: A review of structure/function relationships. Endocr. Rev. 6, 371-399. Hsueh, A. J., Billig, H., and Tsafriri, A. (1994). Ovarian follicle atresia: A hormonally controlled apoptotic process. Endocr. Rev. 15, 707-724. Erickson, G. F. (1997). Defining apoptosis: Players and systems. J. Soc. Gynecol. Invest. 4, 219-228. Gougeon, A. (1996). Regulation of ovarian follicular development in primates: Facts and hypotheses. Endocr. Rev. 17, 121-155. McNatty, K. P., Moore-Smith, D., Osathanondh, R., and Ryan, K. J. (1979). The human antral follicle: Functional correlates of growth and atresia. Ann. Biol. Anim. Biochim. Biophys. 19, 1547-1558. Erickson, G. F. (1978). Normal ovarian function. Clin. Obstet. Gynecol. 21, 31-52. Groome, N. P., Illingworth, P. J., O'Brien, M., Cooke, I., Ganesan, T. S., Baird, D. T., and McNeilly, A. S. (1994). Detection of dimeric inhibin throughout the human menstrual cycle by two-site enzyme immunoassay. Clin. Endocrinol. (Oxford) 40, 717-723. Gougeon, A. (1986). Dynamics of follicular growth in the human: A model from preliminary results. Hum. Reprod. 1, 81-87. Erickson, G. E (1983). Primary cultures of ovarian cells in serum-free medium as models of hormone-dependent differentiation. Mol. Cell. Endocrinol. 29, 21-49. Groome, N. P., Illingworth, P. J., O'Brien, M., Pai, R., Rodger, E E., Mather, J. P., and McNeilly, A. S. (1996). Measurement of dimeric inhibin B throughout the human menstrual cycle. J. Clin. Endocrinol. Metab. 81, 1401-1405. McNatty, K. P., Hunter, W. M., McNeilly, A. S., and Sawers, R. S. (1975). Changes in the concentration of pituitary and steroid hormones in the follicular fluid of human graafian follicles throughout the menstrual cycle. J. Endocrinol. 64, 555-571. Erickson, G. F. (1994). Polycystic ovary syndrome: Normal and abnor-
26. 27.
28. 29.
30.
31.
32.
33. 34.
35.
36.
36a.
37. 38.
39.
40. 41.
42.
43. 44.
mal steroidogenesis. In "Ovarian Endocrinopathies" (R. Schats and J. S choemaker, eds.), pp. 103-115. Parthenon, Carnforth, England. Richards, J. S. (1994). Hormonal control of gene expression in the ovary. Endocr. Rev. 15, 725-751. Channing, C. P. (1970). Influences of the in vivo and in vitro hormonal environment upon luteinization of granulosa cells in tissue culture. Recent Prog. Horm. Res. 26, 589-622. Erickson, G. F. (1993). Normal regulation of ovarian androgen production. Seminar. Reprod. Endocrinol. 11, 307-312. Erickson, G. E, and Yen, S. S. C. (1993). The polycystic ovary syndrome. In "The Ovary" (E. Adashi and P. K. C. Leung, eds.), pp. 561579. Raven Press, New York. Erickson, G. E (1995). PCO: The ovarian connection. In "Reproductive Endocrinology, Surgery, and Technology" (E. Y. Adashi, J. A. Rock, and Z. Rosenwaks, eds.), pp. 1141-1160. Raven Press, New York. Magoffin, D. A., and Erickson, G. E (1994). Control systems of thecainterstitial cells. In "Molecular Biology of the Female Reproductive System" (J. K. Findlay, ed.), pp. 39-65. Academic Press, San Diego, CA. Erickson, G. E, Magoffin, D. A., Dyer, C., and Hofeditz, C. (1985). The ovarian androgen producing cells: A review of structure/function relationships. Endocr. Rev. 6, 371-399. Miller, W. L. (1988). Molecular biology of steroid hormone synthesis. Endocr. Rev. 9, 295-318. Clark, B. J., Wells, J., King, S. R., and Stocco, D. M. (1994). The purification, cloning, and expression of a novel luteinizing hormoneinduced mitochondrial protein in MA-10 mouse Leydig tumor cells. J. Biol. Chem. 269, 28314-28322. Kiriakidou, M., McAllister, J. M., Sugawara, T., and Strauss, J. F., Jr. (1996). Expression of steroidogenic acute regulatory protein StAR in the human ovary. J. Clin. Endocrinol. Metab. 81, 4122-4128. E1-Roeiy, A., Chen, X., Roberts, V. J., LeRoith, D., Roberts, C. T., Jr., and Yen, S. S. C. (1993). Expression of insulin-like growth factor-I (IGF-I) and IGF-II and the IGF-I, IGF-II, and insulin receptor genes and localization of the gene products in the human ovary. J. Clin. Endocrinol. Metab. 77, 1411-1418. Erickson, G. F. (1994). Non-gonadotropic regulation of ovarian function: Growth hormone and IGFs. In "Ovulation Induction: Basic Science and Clinical Advances" (M. Filicori and C. Flanigni, eds.), pp. 73-84. Elsevier, Amsterdam. Adashi, E., and Leung, P. C. K. (1993). "The Ovary: Comprehensive Endocrinology." Raven Press, New York. Faddy, M. J., Gosden, R. G., Gougeon, A., Richardson, S. J., and Nelson, J. E (1992). Accelerated disappearance of ovarian follicles in mid-life: Implications for forecasting menopause. Hum. Reprod. 7, 1342-1346. Gougeon, A., Ecochard, R., and Thalabard, J. C. (1994). Age-related changes of the population of human ovarian follicles: Increase in the disappearance rate of non-growing and early-growing follicles in aging women. Biol. Reprod. 50, 653-663. McKinlay, S. M., Brambilla, D. J., and Posner, J. G. (1992). The normal menopause transition. Maturitas 14, 103-115. Richardson, S. J., Senikas, V., and Nelson, J. E (1987). Follicular depletion during the menopausal transition: Evidence for accelerated loss and ultimate exhaustion. J. Clin. Endocrinol. Metab. 65, 12311237. Meredith, S., Dudenhoeffer, G., Butcher, R. L., Lerner, S. E, and Walls, T. (1992). Unilateral ovariectomy increases loss of primordial follicles and is associated with increased metestrous concentration of follicle-stimulating hormone in old rats. Biol. Reprod. 47, 162-168. Lintern-Moore, S. (1977). Effect of athymia on the initiation of follicular growth in the rat ovary. Biol. Reprod. 17, 155-161. Lintern-Moore, S., and Everitt, A. V. (1978). The effect of restricted
30
45.
46.
47.
48. 49.
50.
51. 52.
53.
54.
55.
56.
57.
58.
59. 60.
61. 62.
GRECO~u F. ERICKSON
food intake on the size and composition of the ovarian follicle population in the Wistar rat. Biol. Reprod. 19, 688-691. Howe, E., Lintern-Moore, S., Moore, G. E, and Hawkins, J. (1978). Ovarian development in hypopituitary Snell dwarf mice. Biol. Reprod. 19, 959-964. Lintern-Moore, S., Supasri, Y., Pavasuthipaisit, K., and Sobhon, E (1979). Acute and chronic morphine sulfate treatment alters ovarian development in prepuberal rats. Biol. Reprod. 21, 379-383. Lintern-Moore, S., Moore, G. E M., Panaretto, B. A., and Robertson, D. (1981). Follicular development in the neonatal mouse ovary; effect of epidermal growth factor. Acta endocrinol. (Copenhagen) 96, 123 -126. Massagu6, J., Attisano, L., and Wrana, J. L. (1994). The TGF-/3 family and its composite receptors. Trends Cell Biol. 4, 172-178. Klein, N. A., Battaglia, D. E., Clifton, D. K., Bremner, W. J., and Soules, M. R. (1996). The gonadotropin secretion pattern in normal women of advanced reproductive age in relation to the monotropic FSH rise. J. Soc. Gynecol. Invest. 3, 27-32. Lenton, E. A., DeKretser, D. M., Woodward, A. J., and Robertson, D. M. (1991). Inhibin concentrations throughout the menstrual cycles of normal, infertile, and older women compared with those during spontaneous conception cycles. J. Clin. Endocrinol. Metab. 7 3 , 1 1 8 0 1190. Rivier, C., Vale, W., and Rivier, J. (1987). Studies of the inhibin family of hormones: A review. Horm. Res. 28, 104-118. Rivier, C., and Vale, W. (1989). Immunoneutralization of endogenous inhibin modified hormone secretion and ovulation rate in the rat. Endocrinology (Baltimore) 125, 152-157. Muttukrishna, S., Fowler, E A., Groome, N. E, Mitchell, G. G., Robertson, W. R., and Knight, P. G. (1994). Serum concentrations of dimeric inhibin during the spontaneous human menstrual cycle and after treatment with exogenous gonadotrophin. Hum. Reprod. 9, 16341642. Klein, N. A., Illingworth, E J., Groome, N. E, McNeilly, A. S., Battaglia, D. E., and Soules, M. R. (1996). Decreased inhibin B secretion is associated with the monotropic FSH rise in older, ovulatory women: A study of serum and follicular fluid levels of dimeric inhibin A and B in spontaneous menstrual cycles. J. Clin. Endocrinol. Metab. 81, 2742 -2745. Yamoto, M., Minami, S., Nakano, R., and Kobayashi, M. (1992). Immunohistochemical localization of inhibin/activin subunits in human ovarian follicles during the menstrual cycle. J. Clin. Endocrinol. Metab. 74, 989-993. Roberts, V. J., Barth, S., and EI-Roeiy, A. (1993). Expression of inhibin/activin subunits and follistatin messenger ribonucleic acids and proteins in ovarian follicles and the corpus luteum during the human menstrual cycle. J. Clin. Endocrinol. Metab. 77, 1402-1410. Illingworth, E J., Reddi, K., Smith, K. B., and Baird, D. T. (1991). The source of inhibin secretion during the human menstrual cycle. J. Clin. Endocrinol. Metab. 73, 667-673. Seifer, D. B., Gardiner, A. C., Lambert, Messerlian, G., and Schneyer, A. L. (1996). Differential secretion of dimeric inhibin in cultured luteinized granulosa cells as a function of ovarian reserve. J. Clin. Endocrinol. Metab. 81, 736-739. Seifer, D. B. (1997). In "Perimenopause" (R. Lobo, ed.), Serono Symp., pp. 144-153. Springer-Verlag, Berlin. Dain, L. B., Stein, R, Krimer, A. R. D., Asch, R. H., Polak de Fried, E., Charreau, E. H., and Tesone, M. (1991). Progesterone production in cultured human granulosa cells: Correlation with follicular fluid hormone levels. Fertil. Steril. 58, 1093 - 1098. Erickson, G. E, Kokka, S., and Rivier, C. (1995). Activin causes premature superovulation. Endocrinology (Baltimore) 136, 4804-4813. Li, R., Phillips, D. M., and Mather, J. E (1995). Activin promotes ovarian follicle development in vitro. Endocrinology (Baltimore) 136, 849-856.
63. Vale, W., Rivier, J., Vaughan, J., McClintock, R., Corrigan, A., Woo, W., Karr, D., and Spiess, J. (1986). Purification and characterization of an FSH releasing protein from porcine ovarian follicular fluid. Nature (London) 321,776-779. 64. Ling, N., Ying, S.-Y., Ueno, N., Shimasaki, S., Esch, F., Hotta, M., and Guillemin, R. (1986). Pituitary FSH is released by a heterodimer of the/3-subunits from the two forms of inhibin. Nature (London) 321, 779 -782. 65. Nakamura, T., Asashima, M., Eto, Y., Takio, K., Uchiyama, H., Moriya, N., Ariizumi, T., Yashiro, T., Sugino, K., Titani, K., and Sugino, H. (1992). Isolation and characterization of native Activin B. J. Biol. Chem. 267, 16385-16389. 66. Schwall, R., Schmelzer, C. H., Matsuyama, E., and Mason, A. J. (1989). Multiple actions of recombinant activin-A in vivo. Endocrinology (Baltimore) 125, 1420-1423. 67. Rivier, C., and Vale, W. (1991). Effect of recombinant activin-A on gonadotropin secretion in the female rat. Endocrinology (Baltimore) 129, 2463-2465. 68. Carroll, R. S., Kowash, E M., Lofgren, J. A., Schwall, R. H., and Chin, W. W. (1991). In vivo regulation of FSH synthesis by inhibin and activin. Endocrinology (Baltimore) 129, 3299-3304. 69. Woodruff, T. K., Krummen, L. A., Lyon, R. J., Stocks, D. L., and Mather, J. E (1993). Recombinant human inhibin A and recombinant human activin A regulate pituitary and ovarian function in the adult female rat. Endocrinology (Baltimore) 132, 2332-2341. 70. DePaolo, L. V., Bicsak, T. A., Erickson, G. E, Shimasaki, S., and Ling, N. (1991). Follistatin and activin: A potential intrinsic regulatory system within diverse tissues. Proc. Soc. Exp. Biol. Med. 198, 500-512. 71. Mathews, L. S. (1994). Activin receptors and cellular signaling by the receptor serine kinase family. Endocr. Rev. 15, 310-325. 72. Demura, R., Suzuki, T., Tajima, S., Mitsuhashi, S., Odagiri, E., Demura, H., and Ling, N. (1993). Human plasma free activin and inhibin levels during the menstrual cycle. J. Clin. Endocrinol. Metab. 76, 1080-1082. 73. Meunier, H., Cajander, S. B., Roberts, V. J., Rivier, C., Sawchenko, P. E., Hsueh, A. J. W., and Vale, W. (1988). Rapid changes in the expression of inhibin ce-,/3A-, and/3B-subunits in ovarian cell types during the rat estrous cycle. Mol. Endocrinol. 2, 1352-1363. 74. Meunier, H., Roberts, V. J., Sawchenko, R E., Cajander, S. B., Hsueh, A. J. W., and Vale, W. (1989). Periovulatory changes in the expression of inhibin ce-,/3A-, and j3B-subunits in hormonally induced immature female rats. Mol. Endocrinol. 3, 2062-2069. 75. Miyanaga, K., Erickson, G. F., DePaolo, L. V., Ling, N., and Shimasaki, S. (1993). Differential control of activin, inhibin, and follistatin proteins in cultured rat granulosa cells. Biochem. Biophys. Res. Commun. 194, 253-258. 76. Feng, Z. M., Madigan, M. G., and Chen, C. L. C. (1993). Expression of type II activin receptor genes in the male and female reproductive tissues of the rat. Endocrinology (Baltimore) 132, 2593-2600. 77. Nakamura, M., Minegishi, T., Hasegawa, Y., Nakamura, K., Igarashi, S., Ito, I., Shinozaki, H., Miyamoto, K., Eto, Y., and Ibuki, Y. (1993). Effect of an activin A on follicle-stimulating hormone (FSH) receptor messenger ribonucleic acid levels and FSH receptor expressions in cultured rat granulosa cells. Endocrinology (Baltimore) 133, 538544. 78. Cameron, V. A., Nishimura, E., Mathews, L. S., Lewis, K. A., Sawchenko, R E., and Vale, W. W. (1994). Hybridization histochemical localization of activin receptor subtypes in rat brain, pituitary, ovary, and testis. Endocrinology (Baltimore) 134, 799-808. 79. LaPolt, E S., Soto, D., Su, J. G., Campen, C. A., Vaughan, J., Vale, W., and Hsueh, A. J. (1989). Activin stimulation of inhibin secretion and messenger RNA levels in cultured granulosa cells. Mol. Endocrinol. 3, 1666-1673. 80. Xiao, S., and Findlay, J. K. (1991). Interaction between activin and follicle-stimulating hormone-suppressing protein and their mecha-
31
CHAPTER 2 Ovarian A n a t o m y and Physiology
81.
82.
83.
84.
85.
nisms of action on cultured rat granulosa cells. Mol. Cell. Endocrinol. 79, 99-107. Woodruff, T. K., Krummen, L., McCray, G., and Mather, J. E (1993). In situ ligand binding of recombinant human [125I]activin-A and recombinant human [125I]inhibin-A to the adult rat ovary. Endocrinology (Baltimore) 133, 2998-3006. Xiao, S., Robertson, D. M., and Findlay, J. K. (1992). Effects of activin and follicle-stimulating hormone (FSH)-suppressing protein/follistatin on FSH receptors and differentiation of cultured rat granulosa cells. Endocrinology (Baltimore) 131, 1009-1016. Presl, J., Pospisil, J., Figarov~i, V., and Krabec, Z. (1974). Stagedependent changes in binding of iodinated FSH during ovarian follicle maturation in rats. Endocrinol. Exp. 8, 291-298. Zeleznik, A. J., Schuler, H. M., and Reichert, L. E. (1981). Gonadotropin-binding sites in the Rhesus monkey ovary; Role of the vasculature in the selective distribution of human chorionic gonadotropin to the preovulatory follicle. Endocrinology (Baltimore) 109, 356-362. Nimrod, A., and Lamprecht, S. A. (1980). Hormone-induced desen-
86.
87.
88.
89.
90.
sitization of cultured rat granulosa cells to FSH. Biochem. Biophys. Res. Commun. 92, 905-911. Erickson, G. F. (1997). Dissociation of endocrine and gametogenic ovarian function. In "Perimenopause" (R. Lobo, ed.), Serono Symp., pp. 101-118. Springer-Verlag, Berlin. Nakamura, K., Nakamura, M., Igarashi, I. S., Miyamoto, K., Eto, Y., Ibuki, Y., and Minegishi, T. (1994). Effect of activin on luteinizing hormone-human chorionic gonadotropin receptor messenger ribonucleic acid in granulosa cells. Endocrinology (Baltimore) 134, 2329-2335. Mir6, F., Smyth, C. D., and Hillier, S. G. (1991). Development-related effects of recombinant activin on steroid synthesis in rat granulosa cells. Endocrinology (Baltimore) 129, 3388-3394. Itoh, M., Igarashi, M., Yamada, K., Hasegawa, Y., Seiki, M., Eto, Y., and Shibai, H. (1990). Activin A stimulates meiotic maturation of the rat oocyte in vitro. Biochem. Biophys. Res. Commun. 166, 1479-1484. Doi, M., Igarashi, M., Hasegawa, Y. U., Eto, Y., Shibai, H., Miura, T., and Ibuki, Y. (1992). In vivo action of activin-A on pituitary-gonadal system. Endocrinology (Baltimore) 130, 139-144.
~HAPTER
Regulation of the Hyp oth al ami c -- Pituitary Gonadal Axis: Role of Gonadal Steroids " and lmpl"~cat~ons for the Menopause FRANCISCO
Josr L6PEZ, P A T R I C I A
D. FINN, MARK A. LAWSON,
AND ANDRI~S NEGRO-VILAR Ligand Pharmaceuticals, Inc., San Diego, California 92121
I. Introduction II. Functional Organization of the Hypothalamic-Pituitary- Ovarian Axis III. Anatomical and Biochemical Correlates of Gonadal Steroid Hormone Action in the Central Nervous System
IV. Ovarian Steroid Action in the Central Nervous SystemnControl of Reproduction V. The Menopause: Consequences of Steroid Hormone Loss on the Central Nervous System References
I. I N T R O D U C T I O N
monthly chance of becoming pregnant. This regular ovulatory pattern continues for the life span of females in most species except humans, for whom reproduction ceases in midlife with the onset of menopause. Because of an increasing life expectancy, women spend an increased portion
The window of opportunity for conception to occur in most mammalian species is cyclical. Women produce a viable egg roughly every month, providing, therefore, a
MENOPAUSE: BIOLOGY AND PATHOBIOLOGY
33
Copyright9 2000by AcademicPress. All rightsof reproductionin any formreserved.
34
L 6 P E Z ET AL.
of their life in the postmenopausal state. In fact, 35 million American w o m e n are currently postmenopausal and this n u m b e r increases at a rate of over a million every year [1]. M e n o p a u s e is characterized by a marked decrease in estrogen production from the ovaries. The estrogen decrease that occurs not only results in a loss of reproductive function, but is also associated with a n u m b e r of conditions that reduce the quality of life of this population of women. It has b e c o m e clear over the past 1 0 - 1 5 years that the ubiquitous presence of the estrogen receptor (ER) in body tissues and organ systems provides the substrate for estrogenic regulation or modulation of multiple body functions. That knowledge has run in parallel with ever increasing evidence of the beneficial effects of estrogen replacement therapy after the menopause. What started originally as an approach to treat vasomotor symptoms and vaginal dryness has evolved into a comprehensive therapeutic approach that can span several decades after the m e n o p a u s e and is intended to prevent and/ or treat osteoporosis, cardiovascular disease, cognitive dysfunction, and possibly other neurodegenerative disorders such as Alzheimer disease. Central to this expansive approach to h o r m o n e replacement therapy is not only a more refined understanding of the mechanisms that lead to estrogen's beneficial effects on many tissues and functions, but also to an understanding of the undesired side effects that result from estrogenic activity at sites that are not the intended therapeutic targets (i.e., uterus and breast). The concept of tissue selectivity in estrogen action has e m e r g e d in recent years and provides the cellular and molecular basis to understand the diversity of estrogen activity and the distinct m o d e of action of full versus partial estrogen agonists [ 2 4]. It is hoped that a better understanding of these issues will lead to a better h o r m o n e replacement therapy in postmenopausal women.
II. FUNCTIONAL ORGANIZATION OF THE H Y P O T H A L A M I C PITUITARY-OVARIAN AXIS A. Functional Organization Multiple physiological factors influence reproductive fitness, as illustrated by the alteration of normal menses by factors such as stress and metabolic state. However, the principal components of the reproductive endocrine axis, the luteinizing hormone-releasing h o r m o n e (LHRH) neurons of the forebrain, the pituitary gonadotroph, and the gonads, each have principal and distinct roles in modulating reproductive function. H o r m o n a l feedback regulation appears to occur at all levels of this axis (Fig. 1). Therefore, a complex regulatory system is implicated in the mechanisms that lead to the release of the egg. In spite of the complexity, it is interesting that a small n u m b e r of neurons in the forebrain, par-
FIGURE 1 Schematic representation of the hypothalamic-pituitarygonadal axis in females. Neurons located in the forebrain synthesize the decapeptide luteinizing hormone-releasing hormone (LHRH), also known as gonadotropin-releasing hormone. The activity of LHRH-producing neurons is controlled by multiple afferent terminals, which either stimulate (positive sign on top of the diagram) or inhibit (negative sign on top of the diagram) the LHRH neuronal network. LHRH is secreted at the level of the median eminence into the hypophyseal portal system in a pulsatile manner. Via the portal circulation, LHRH reaches the anterior pituitary (AP), where it stimulates (positive sign close to arrows beneath LHRH) the production of the pituitary gonadotropins, luteinizing hormone (LH) and follicle-stimulating (FSH) hormone. Gonadotropins, in turn, are secreted in a pulsatile manner into the bloodstream to regulate (positive sign near the gonad) gonadal homeostasis in terms of gametogenesis and production of gonadal hormones such as estradiol (E2), progesterone (P), and inhibin. Gonadal hormones exhibit trophic effects on sex accessory tissues and contribute to the regulation of the hypothalamic-pituitary axis by the establishment of the classical long-loop negative feedback (LL-). Under some conditions estradiol is able to stimulate the activity of the LHRH neuronal network, resulting in the ovulatory surge of gonadotropins. This type of control mechanism is known as positive feedback (LL+). Two other feedback loops have been demonstrated in the hypothalamic-pituitary-gonadal axis. These involve the gonadotropin-dependent negative regulation of LHRH production (negative short-loop feedback; SL-) and the LHRHdependent inhibition of LHRH production constituting the ultrashort-loop negative feedback (USL-). NL, Neurointermediate lobe of the pituitary gland.
ticularly the hypothalamus, are responsible for all aspects of reproduction. These neurons, which n u m b e r 2 0 0 0 - 3 0 0 0 in humans, synthesize the decapeptide L H R H , also known as gonadotropin-releasing hormone, which in essence represents the conductor of reproductive function. The absence of L H R H results in an absolute impairment of reproduction as
CHAPTER 3 Role of Gonadal Steroids in Menopause evidenced by the phenotype of the hypogonadal mouse in which there is a mutation on the LHRH gene leading to the synthesis of a nonfunctional prohormone [5]. A similar reproductive impairment is observed in humans with Kallmann's syndrome, in which there is an absence of LHRH neurons in the forebrain [6]. In these abnormal conditions, reproductive function can be recovered by introduction of either a normal gene, in the case of the mice [7], or by exogenous LHRH treatment, in the case of humans with Kallmann's syndrome [8]. Neurons of the LHRH system originate outside of the brain in the olfactory placode and migrate into the ventral medial forebrain [9]. Rather than residing in a discrete anatomical location, LHRH-containing neurons form a loose continuum of scattered cells that span areas such as the diagonal band of Broca, dorsal septum, bed nucleus of the stria terminalis, lateral and medial preoptic areas, and mediobasal hypothalamus [10-16]. In many rodent species, including mice and rats, the majority of LHRH neurons are found in the rostral medial preoptic area, near the organum vasculosum of the lamina terminalis, and very few neurons are found caudal to the medial preoptic area [10,12]. In primates, including humans, many LHRH neurons are also found in the medial preoptic area. However, because LHRH neurons migrate more caudally, many are also located in the mediobasal hypothalamus [ 11,13-15]. Although LHRH neuronal perikarya are scattered throughout many areas of the forebrain, rather than concentrated in specific nuclei, the majority of LHRH axon terminals converge in the median eminence through the preoptico-infundibular tract [17]. At the level of the median eminence, LHRH terminals establish contacts with fenestrated capillaries, forming the hypophyseal portal system. Via this specialized vascular system, secreted LHRH is transported to the anterior pituitary, where it interacts with specific membrane receptors in gonadotrophs to stimulate the production and secretion of the gonadotropins, luteinizing hormone (LH) and follicle-stimulating hormone (FSH). The interactions between the different regulatory sites of the hypothalamic-pituitary-gonadal axis are schematically represented in Fig. 1. The pituitary gonadotropins, LH and FSH as well as thyroid-stimulating hormone, belong to the family of heterodimeric pituitary glycoprotein hormones [18]. All of these glycoproteins contain a common a subunit that is linked noncovalently with a hormone-specific/3 subunit (Fig. 2). Despite their similar structures and sequence homologies, these hormones exhibit only marginal cross-reactivity between their respective receptors. A common structural feature of these proteins is the presence of multiple glycosylation sites on both subunits. For example, FSH contains four such glycosylation sites, two on each of the ce and/3 subunits [19]. Luteinizing hormone and FSH control gonadal function via interactions with specific membrane receptors present in various cell types within the gonad. The recep-
35
FIGURE 2 General structural features of the pituitary glycoprotein hormone family. Luteinizing hormone (LH), follicle-stimulating hormone (FSH), and thyroid-stimulatinghormone (TSH) belong to the family of pituitary heterodimeric glycoprotein hormones. The three proteins share a common ce subunit and have hormone-specific/3 subunits. Chorionic gonadotropin (CG) producedby the placenta is highlyhomologousto pituitary LH and interacts with LH receptors; it is thus included in the diagram for the sake of completeness.
tors transducing gonadotropin signals belong to the seventransmembrane-spanning domains, G protein-coupled receptors. In general, gonadotropin receptors are coupled to the cellular response via activation of adenylate cyclase and the subsequent production of cyclic adenosine monophosphate (cAMP) [20]. Gonadotropin actions on the gonad encompass production of steroid and peptide hormones as well as gametogenesis. Through negative feedback mechanisms gonadal ster o i d s - e s t r o g e n , progesterone, and testosterone--close the regulatory loop in the hypothalamic-pituitary-gonadal axis by inhibiting the release of LHRH, and the gonadotropins do so by acting at the hypothalamus and pituitary (Fig. 1; longloop feedback). Estrogen is the only gonadal steroid that has the ability to also exert positive feedback action on the hypothalamus and pituitary. The positive feedback effects of estrogen play a fundamental role in initiating the series of events that lead to the preovulatory surge of gonadotropins in spontaneous ovulators, including humans (reviewed in Section III,A,2). In spite of extensive work on the central actions of sex steroids in the control of reproductive functions, the molecular mechanisms involved in these actions remain elusive. It is clear, however, that gonadal steroid effects are mediated by discrete subpopulations of neurons that are strategically located to control reproductive functions. Direct effects of sex steroids on LHRH neurons have not yet been conclusively documented in vivo due to the failure to detect ERs or estrogen binding in LHRH neurons [21-24]. However, ER immunoreactivity in LHRH neurons in the medial preoptic area of rats has been demonstrated [25]. More recently, Skynner et al. (1999) have shown the presence of ERce and ERfl mRNA in single preoptic area neurons by nested RT-PCR. Although it is intriguing to speculate that estrogen can act directly on LHRH neurons, the identification of ERs in
36
L6PEZ ET AL.
LHRH neurons needs to be explored further. Whereas the direct action of estrogen on LHRH neurons remains controversial, it is clear that estrogen acts on neurons that lie in close proximity to LHRH neurons, such as GAB A-containing neurons, as well as on other sets of neurons that innervate LHRH neurons, such as fl-endorphin-containing neurons, whose cell bodies reside in the hypothalamus, and norepinephrine-containing neurons, whose cell bodies are located in the hindbrain. Furthermore, there is ample evidence that these neuronal populations mediate at least some of estrogen's action on the LHRH neuronal network [ 16,26]. Additional findings indicate that sex steroids may also affect LHRH neuronal function by nongenomic mechanisms or by actions that do not require the presence of classical nuclear steroid receptors (for a review, see Ref. 27). Glycoprotein hormones such as activin and follistatin may also exert negative feedback actions on the hypothalamic-pituitary-gonadal axis. Release of these hormones from the pituitary may constitute a short-loop feedback to the hypothalarrius. Evidence for this arises from the observation of LH/chorionic gonadotropin receptors on LHRH neurons [28,29]. Follistatin transcription and release are increased by LHRH in primary pituitary cell cultures under conditions that inhibit FSH synthesis [30], and the presence of follistatin receptors on neurons in the hypothalamus closely associated with LHRH neurons suggesting a possible central nervous system (CNS) site of action [31 ]. Additional data suggest the possibility that follistatin may indirectly influence LHRH neuron activity through antagonism of activin action [32].
B. D i s t r i b u t i o n o f L H R H N e u r o n s in the F o r e b r a i n It is generally thought that LHRH neurons are sparsely distributed in the medial forebrain of primates and are most abundant in the medial preoptic area and mediobasal hypothalamus. However, this view has been challenged. In humans, the distribution of LHRH mRNA-expressing neurons has been mapped using a cDNA probe complementary to mammalian LHRH. Many LHRH mRNA-expressing neurons were found both inside and outside the medial forebrain in humans [33]. In fact, more than 12,000 LHRH mRNA-expressing neurons per subject were counted, a number nearly 10-fold more than had been reported for other primate species by other investigators using in situ hybridization or immunocytochemistry [ 11,14,15]. These authors divided LHRH neurons into subtypes, types I through III, based on morphological characteristics and mRNA expression levels. Type I and type II LHRH neurons are small in size, whereas type III is large (Table I)[33a]. In addition, type I neurons display the classical oval-fusiform morphology typical for LHRH neurons in other species [17]. In contrast, type II neurons are small, oval, and have more rounded somata (Table I). Type III neurons exceed 500/xm 2 in profile area, are round to oval in morphology, and are characterized by prominent Nissl staining located in the periphery of the cells. Additionally, type I LHRH neurons express very high levels of LHRH, whereas type II neurons express the least amounts of mRNA for LHRH. Levels of LHRH mRNA in type III neurons are intermediate between types I and II (Table I).
TABLE I Subpopulations of LHRH Neurons in Humans: Differential Regulation by the Postmenopausal Reduction of Estrogen Levels LHRH neuron subtype Parameter
Type I
Type II
Type III
Size a
Small
Small
Large
Level of mRNA expression in LHRH neurons`"
Heavy
Light
Intermediate
Shape"
Oval to fusiform
Round to oval
Round to oval
Major location of the cells`"
Mediobasal hypothalalamus and the medial preoptic area
Scattered in the dorsal preoptic area, the septal area, the amygdala and the substantia innominata
Putamen and the magnocellular basal forebrain complex
50% increase
No change
Not evaluated
LHRH mRNA levels: changes in postmenopausal women b
a Characteristics of the different subsets of LHRH neurons were taken from the paper by Rance et al. [33].
b LHRH mRNA levels were determined by in situ hybridization and reported by Rance's laboratory in 1996 [33a].
CHAPTER3 Role of Gonadal Steroids in Menopause The distribution patterns of the three subtypes of LHRH neurons are also somewhat distinct. Type I cells are located in areas that have been previously shown to contain LHRH neurons in other primate species [11,13-15]: i.e., the medial preoptic area and mediobasal hypothalamus. Moreover, these cells are similar in number to that reported for LHRH neurons found in other primate species [11,14,15,34]. In contrast, type II LHRH neurons were found to be numerous in the medial forebrain, the medial septum, medial preoptic area, and the lateral forebrain. Small numbers of cells are also found in the bed nucleus of the stria terminalis and in striatal areas such as the globus pallidus and ventral pallidum. Type III LHRH neurons were found in high numbers in lateral forebrain areas such as the putamen, sublenticular substantia innominata, and nucleus basalis of Meynert. Table I summarizes the characteristics and distribution of the different types of LHRH neurons. Since the original report on the distribution of LHRH mRNA-expressing neurons in human brain [33], there have been other reports verifying the presence of previously undetected populations of LHRH neurons in the lateral forebrain of both fetal and adult rhesus monkeys [35,36]. Like the traditional well-described LHRH neurons implicated in gonadotropin secretion, these previously undetected populations of LHRH neurons that reside in the lateral forebrain arise from the olfactory placode and begin their migration into forebrain areas slightly earlier than the LHRH neurons destined to reside in the medial forebrain [35]. The distribution of LHRH-immunoreactive neurons in the lateral forebrain of fetal rhesus macaques is similar to the distribution of LHRH mRNA-expressing neurons described in human lateral forebrain [33,35]. Interestingly, these lateral forebrain LHRH neurons are detectable with only one of five antisera made to mammalian LHRH, suggesting that these LHRH neurons may be distinct from those that reside in the medial forebrain [35]. Further analysis of these lateral forebrain LHRH neurons revealed that they also express the enzyme EP24.15, which is capable of cleaving LHRH between the fifth and sixth amino acids [37]. Taken together these data suggest that lateral forebrain LHRH neurons contain an amino-terminal five-amino acid cleavage product (LHRH 15) of the LHRH decapeptide. The physiological significance of LHRH-expressing neurons in the lateral forebrain is unknown. Few, if any, lateral forebrain neurons project to the median eminence, the site where the vast majority of medial forebrain LHRH neurons establish contacts with fenestrated capillaries to release LHRH into the hypophyseal circulation to elicit gonadotropin release from the pituitary [ 16]. Several of the lateral forebrain areas, in which LHRH neurons are found in humans, appear to be sites for estrogen's action in nonhuman primates [38,39], yet few of these areas are directly implicated in reproductive function. Instead, several of these areas undergo changes with aging that are associated with the development
37 of Alzheimer disease [40]. What role, if any, the LHRH gene may play in these areas is still unknown.
III. ANATOMICAL BIOCHEMICAL GONADAL
AND
CORRELATES
STEROID
IN THE CENTRAL
OF
HORMONE
NERVOUS
ACTION
SYSTEM
Steroid hormones exhibit their regulatory functions by modulating the activity of ligand-dependent transcription factors (for reviews on the mechanism of action of steroid hormones, see Refs. 41 and 42; see also Chapter 1). In addition, a growing body of evidence supports the notion that steroid hormones also exert effects that are independent of those mediated via nuclear receptors [27]. This dual genomic and nongenomic action of steroid hormones as well as the identification of multiple nuclear receptors complicate our understanding of the biology of steroid hormones and their regulatory actions on LHRH production and secretion. With menopause, there is a cessation of ovarian function, resulting in the decreased production of sex steroids, progesterone, and, in particular, estrogens. It is necessary to understand the mechanisms by which, and the sites where, ovarian steroids act in the CNS to appreciate the consequences of estrogen withdrawal and to develop strategies for safe and effective estrogen replacement therapies. The following discussion summarizes general concepts of estrogen signal transduction mechanisms as well as the distribution of ERs in the brain.
A. E s t r o g e n R e c e p t o r S i g n a l Transduction Mechanisms Interactions of the ER with transcription factors and subsequent binding to hormone response elements (HREs) provide a mechanistic framework by which the selectivity of estrogen action can be explained. The multiple transcription factors involved in the model of nuclear receptor activation of genes has been integrated into a tripartite pharmacological model [3]. As illustrated in Fig. 3, three modes of action combined with three ligand-dependent conformations of the receptor make up this model. In the first mode of action, initial binding of the hormone in the receptor induces a conformational change to the receptor that results in binding of the receptor/ligand complex to a hormone response element (HRE). In addition, ligand-receptor interaction leads to dimerization and recruitment of the general transcription factor complex. A second mode of action involves an adaptor protein driving final activation of transcription and involving the interaction of the receptor complex with specific HREs in the promoter region of particular genes. Third, recruitment of other transcription factors acting as coactivators with the receptor eliminates the need for specific HREs in
38
FIGURE 3 Schematic depiction of the tripartite receptor pharmacological model to explain pleiotropic actions of steroid hormones. Steroid receptors interact with various ligands (L1 through L3), resulting in a change in the spatial conformation of the ligand receptor complex (represented on the left side of the figure by the diverse shapes of the receptor dimers). As a consequence of the different ligand-dependent conformations, various responses can be achieved in the tissues of interest. The specific responses are conformation dependent and probably use different mechanisms to modulate gene expression. In the first mechanism represented, the receptor uses a hormone response element (HRE) present on the gene of interest. After interacting with the HRE, the receptor recruits general transcription factors (GTFs) to form a complex that activates transcription of the gene. In the second mechanism, interaction with ligand 2 leads to a different conformation of the receptor that uses the HRE, but needs an adaptor protein to interact with the general transcription machinery. The third mechanism is HRE independent, but requires an interaction with other transcription factors to elicit activation of gene transcription. From [3] Katzenellenbogen, J. A., O'Malley, B. W., and Katzenellengoben, B. S. (1996). Tripartite steroid hormone receptor pharmacology: Interaction with multiple effector sites as a basis for the cell- and promoter-specific action of these hormones. Mol. Endocrinol. 10, 119-131. 9 The Endocrine Society. the target gene. Diverse conformations of the receptor may be generated by interaction with different ligands (Fig. 3, L 1 through L3), providing a molecular mechanism for the pleiotropic actions of the steroid receptor complex in response to pharmacological or physiological agents interacting with the receptor. The identification of a second ER confers an additional level of complexity by adding a level of selectivity, which depends on the discrete distribution of the various receptor subtypes. The classical, human ER was cloned and identified in 1985 as a ligand-dependent transcription factor that binds estrogens and antiestrogens [43]. Ten years later, Kuiper et al. [44], using a rat prostate cDNA library, isolated another ER, which has been designated fl (the classical ER has been defined as a). Various functional domains were defined within the ERa coding sequence [45], two of which are involved in DNA- and ligand-binding interactions (Fig. 4). In addition, two other functional domains, AF-1 and AF-2, were defined; the AF-1 domain is involved in hormone-
L 6 P E Z ET AL.
independent activation of the receptor, whereas the AF-2 domain participates in hormone-dependent activation [45]. A comparison of the human ERc~ and ERfl cDNAs illustrates that the DNA-binding domain is almost 100% conserved in both receptors (Fig. 4), suggesting that both receptors interact with similar HREs on DNA. In contrast, the similarity of the ligand-binding domains is poor (59%; Fig. 4). This difference may be exploited by the identification of novel ligands that display differential receptor specificity, allowing the possibility to modulate estrogen action selectively via ERa, ERfl, or both. A variety of molecular entities corresponding to the ERfl sequence have been cloned. Shortly after the initial description of the receptor, a human homolog was cloned from the testis [46], which encodes a protein that is eight amino acids shorter than the rat ERfl. Later, the human sequence was extended to the first rat methionine codon, adding the eight amino acids [47]. At the same time, the mouse homolog was cloned [48], showing a similar eight-amino acid extension. Further efforts have resulted in the identification of extended forms of ERfl. A clone that encodes 45 additional amino acids was recently isolated in humans [49]. In addition, another clone for the rat sequence, which contains an additional 64 amino acids, has been described (GenBank accession number AJ002602). Figure 4 depicts the homology of the short and long forms of the rat ERfl compared to the ERa. These longer forms of the receptor have also been identified in the mouse and are accessible in GenBank. The existence of these molecular entities raises questions about the validity of our current cloning strategies [50], because it is difficult to assess if any or all of the isolated clones represent physiologically relevant ERfl. In addition to extended N terminal receptor subtypes, various reports [50a,b,c] identify
AF-I i
ERo~
DNAHORMONE BINDING BINDING/AF-II 180 263302
A/B
I
C
IDI
E
553 595
I FI
HINGE i
ERI3 Short
96 18%
i
ERI3 Long
166 211
A/B I C I D I 97% 30%
149 A/B
30%
449 477
E
IF I
59%
18%
219 264
I c ]DI 97% 30%
502 530
E
IF I
59%
~8%
FIGURE 4 Functional estrogen receptor (ER) domains (A though F) and identity between the human ERa and ERfl short and long forms. Activation function I (AF-I) resides in the N terminus of the estrogen receptor, whereas activation function II (AF-II) resides in the C terminus. Notice that the DNA-binding domain of the ER is highly similar (97% homologous) among the ce and/3 forms of the receptor. In contrast, the identity in the hormone-binding domain is low (59% homologous), suggesting the potential for subtype-selective ligands. The ERa, ER/3 short form, and ER/3 long form protein sequences were obtained from GenBank accession numbers NM000125, X99101, and AB006590, respectively. They were aligned using the Clustal W algorithm [45a] and Omiga 1.1.3 software (Oxford Molecular Limited).
CHAPTER3 Role of Gonadal Steroids in Menopause isoforms of ERfl with amino acid insertions in the ligand binding domain, and this has been observed in humans as well as rats. It is apparent (Fig. 3) that interaction of estrogen with its receptor causes a receptor-ligand complex dimerization that associates with DNA and subsequently modulates gene transcription. Studies looking at colocalization of ERs have been able to identify cell types that express both subtypes of the ER, particularly in the CNS (see Section II,B). Colocalization of both receptor subtypes in the same neuron would imply that E R a and ERfl could conceivably form heterodimers to control gene transcription (Fig. 5). Pettersson et al. demonstrated, using a mammalian two-hybrid system, that both subtypes of the ER could form functionally active heterodimers [51 ]. These observations have been confirmed using immunoprecipitation [49]. The possibility that ERce and ER/3 form functionally active heterodimers raises the question of whether heterodimers would interact with classical estrogen response elements (EREs) or other DNA response elements (Fig. 5). Heterodimer formation, therefore, could contribute to the pleomorphic action of estrogens in different tissues and in response to different ligands. ERce and ERfl bind estradiol with similar affinity [52], and both act at EREs to regulate gene expression [53]. However, they differ in their ligand-dependent activation of AP-l-regulated gene expression. In the presence of 17/3estradiol, ERce, but not ERfl, activated gene transcription via AP-1, whereas in the presence of antiestrogens both receptors activated gene transcription via AP-1 [53]. This pro-
FIGURE 5 The existence of two estrogen receptor subtypes, ERa and ERfl, that coexist in somepopulations of cells suggestthat ER heterodimers could modulate gene transcription. Estrogen receptor c~and fl homodimers have been shown to modulate transcription via interactions with classical estrogen response elements (EREs). However,ERfl homodimers could also use nonconsensus ERE (other REs) to modulateactivityof additional genes. Likewise, ERcdERfl heterodimers modulate transcription via interaction with classical EREs and possibly other nonconsensus EREs in ERa/ERfl responsive genes. Estradiol is represented in the figure by a gray triangle. Notice the change in receptor conformation after estradiol is bound to the receptor.
39 vides an additional biochemical substrate by which liganddependent conformations of the receptor could induce differential gene activation utilizing diverse signal transduction mechanisms.
B. Distribution of ERc~ and ERfl in the Central Nervous System The gonadal steroid, estradiol, has widespread effects on the brain. The sites of estradiol's action in the CNS have been explored by examining binding sites for estradiol using autoradiography as well as expression of ER mRNA and protein using in situ hybridization and immunocytochemistry, respectively. Since the discovery that there are two subtypes of ER, a and fl (see Section II,A), studies have been conducted in several species to compare the distribution of these different receptor subtypes and gain some insight on the possible role of ERfl based on its specific anatomical location compared to that of ERa. The distribution of estrogen binding sites found in the brain of females of different species exhibits many commonalties [54]. In all vertebrates examined to date, estrogenbinding cells are found in the medial preoptic area, the mediobasal hypothalamus, the limbic forebrain including the septum, bed nucleus of the stria terminalis, and corticomedial amygdala, as well as the midbrain tegmentum. Furthermore, in rodents there is evidence that the cell groups that bind estrogen in the brain participate in hormonally controlled functions and behaviors, such as gonadotropin secretion and mating behavior [54,55]. Estrogen receptor fl mRNA and protein distribution in the brain has been described only in rats, monkeys, and E R a knockout mice (ERaKO) [38,56-59]. Although there are some discrepancies among the reported distribution of both ER subtypes in rats, the data so far show that both E R a and ERfl are abundantly expressed in limbic forebrain areas such as the bed nucleus of the stria terminalis, the medial amygdala, the medial preoptic area, and the periventricular nucleus of the hypothalamus (see Table II) [57,58]. Estrogen receptor c~ mRNA is expressed at much higher levels than is ERfl mRNA in two mediobasal hypothalamic areas, the ventromedial and arcuate nuclei, whereas ERfl mRNA is more highly expressed (and possibly exclusively) in the supraoptic and paraventricular hypothalamic nuclei, the cerebral cortex, and the cerebellum. Less clear is the differential expression of E R a and ERfl mRNA in some other areas of rat brain. Laflamme and colleagues reported higher expression of ERc~ mRNA, compared to ERfl mRNA, in the CA3 area of hippocampus and exclusive expression of ERa in the locus coeruleus [58], one of the major sites of noradrenergic neurons innervating the forebrain. In contrast, Shughrue et al. [57] reported the opposite pattern than that observed by Laflamme et al. [58] in the hippocampus and equal expression
40
L6PEZ
TABLE II
ET AL.
Distribution of ERa and ERfl in the Central Nervous System of Female Rats a H y b r i d i z a t i o n signal r e p r e s e n t i n g m R N A b D e n s i t y of labeled cell bodies r
Relative intensity of labeling d
B r a i n area
ERa
ERfl
ERa
ERfl
Isocortex
-/+
+/+ + +
Hippocampus (CA1-CA3)
-/+
+ + +
Basal n u c l e u s of M e y n e r t Medial amygdala
+ + + + +
+ + + + +
N.L. + + + e
N.L. + +/+ + +
-/+
+ +
-/+
Lateral s e p t u m
+ +
+
+ +/+ + +
+
Medial septum
+
+ +
+ +
-
+ + + + +
+ + + + + +
+ + + + +f
+ + +f
N u c l e u s of the diagonal b a n d B e d n u c l e u s of the stria terminalis (posterior) Substantia innominata M e d i a l preoptic area Periventricular hypothalamic nucleus S u p r a o p t i c nucleus
N.L.
N.L.
-
-
+ + + +
+ + + +
+ + +
+ +/+ + +
+ + +
+
+ +
+
-
+ + + +
-
+ + +/+ + + +
-
+ +/+ + + +
P a r a v e n t r i c u l a r h y p o t h a l a m i c nucleus V e n t r o m e d i a l h y p o t h a l a m i c nucleus
+ .qt_.qt_ e
_ g
.qt_ --I- + + h
A r c u a t e nucleus
+ + +
+
+ + +
+
+ +
+
+ + +
-/+
Dorsal r a p h e Locus coeruleus
+
+ +
N.L. + + +
N.L. -
Cerebellum
-
+ + /+ + +
-
+ /+ +
P e r i a q u e d u c t a l gray
-/+
+
+ +/+ + + + ..]_[..1_ + h
a D a t a a c c o r d i n g to S h u g h r u e et al. [57] and L a f l a m m e et al. [58]. bN.L., Not listed; - , undetectable signal; + , low but positive signal; + + , moderate signal; + + + , strong signal; + + + + , very strong signal. CFrom S h u g h r u e et al. [57]. d F r o m L a f l a m m e et al. [58]. e
Posterior part.
f Principal part. g Ventrolateral part. h C a u d a l part.
of ERa and ERfl mRNA in the locus coeruleus. Table II summarizes the localization of the two receptor subtypes in a variety of brain areas according to both research groups. Shughrue and colleagues [56] found ERfl mRNA to be expressed in ERaKO mice in many of the same areas in which it is expressed in normal rats, including the bed nucleus of the stria terminalis, medial amygdala, medial preoptic area, and paraventricular nucleus of the hypothalamus. Of particular interest was the low level of ERfl mRNA expression in the supraoptic nucleus in ERceKO mice, an area in rats that shows a high degree of both ERfl mRNA and protein expression [57-59]. However, ERceKO mice are exposed to very high levels of estrogen as a consequence of the lack of negative feedback effects of estrogen mediated by ERa. Therefore, these changes in ERfl expression may be the result of the different endocrine milieu to which these animals are exposed. Pau et al. used a reverse transcriptase polymerase chain reaction (RT-PCR) and in situ hybridization to determine the differential expression of ERa and ERfl mRNA in rhesus macaque brain [38]. Because the full-length cDNAs for the
two subtypes of ER have not yet been cloned in the monkey, the authors refer to their findings as putative sites of ER expression. In most areas examined in female monkey brain, both ERce and ERfl mRNA were expressed. These areas include the septum, amygdala, medial preoptic area, medial basal hypothalamus, and paraventricular hypothalamic nucleus. Estrogen receptor ce mRNA alone was found in areas such as the frontal cortex and locus coeruleus. As the question of differential expression of the ER subtypes is investigated in more species and as additional specific antibodies that discriminate ERa and ERfl become available, a clearer picture of their pattern of expression in the brain should emerge and a better understanding of their physiological role should arise. Colocalization of ERa and ERfl mRNAs in the same cells in brain has been studied. In female rats, many cells express ERfl mRNA and are immunoreactive for ERce in the preoptic part of the periventricular nucleus, the posterior part of the bed nucleus of the stria terminalis, and the medial nucleus of the amygdala [60]. A smaller number of double-labeled cells are found in the medial preoptic area, and only a few double-
CHAPTER3 Role of Gonadal Steroids in Menopause labeled cells are found in the ventromedial and arcuate nuclei of the hypothalamus. The consequences of this coexpression of the two ER subtypes are presently unknown; however, it is likely that cells that express one or both of the ER subtypes respond to estrogen differently. In addition, cells in which ERa and ERfl are expressed may express an array of genes that could be activated by the heterodimer a/fl (see Section II,A). Studies in female rats suggest that specific areas of the brain may become less sensitive to estrogen with aging. Comparisons of nuclear ER concentrations in the medial preoptic area and hypothalamus show a decline with aging under several endocrine conditions [61-63]. Under one of these paradigms, Wise and Parsons [64] evaluated female sexual behavior, and the deficit in sexual behavior paralleled the decline in ER concentrations, suggesting that the deficit in behavior resulted from aging-dependent estrogen insensitivity. Whether similar mechanisms are operative in humans is yet to be determined.
C. D i s t r i b u t i o n o f P r o g e s t e r o n e R e c e p t o r in the C e n t r a l N e r v o u s S y s t e m In all vertebrate species studied to date, progesterone receptor (PR) is found in the medial preoptic area and mediobasal hypothalamus [65]. In most species, PR is also found in other regions of the brain, including the limbic forebrain, midbrain tegmentum, cortex, and cerebellum [65,66]. Similar to ER binding, PR binding may decrease with aging in rats. Wise and colleagues showed a decline in the concentration of PR binding in the medial preoptic area and mediobasal hypothalamus of middle-aged rats compared to young rats after 2, but not 4, days of estradiol treatment [64]. Furthermore, this decline in PR in middle-aged rats correlated with deficits in all aspects of female reproductive behavior tested [64]. In contrast, Brown and colleagues did not detect a decrease in progesterone binding in various microdissected regions, including the medial preoptic area and mediobasal hypothalamus in young compared to older rats [63]. However, unlike the paradigm used by Wise and Parsons [64], Brown et al. [63] exposed the animals to estrogens for 3 days and attained higher circulating levels of estrogens, thus it may be that under certain hormonal conditions there is a decline in PR binding with aging. From a strictly functional point of view, PRs can be divided into two distinct anatomical classes; one class is induced by estradiol and the other is not [66]. A variety of studies examining several species, and using either immunocytochemistry for PR binding of radiolabeled progesterone or in situ hybridization for PR mRNA, show that estradiol induces PRs in the medial preoptic area and mediobasal hypothalamus [67-72]. Furthermore, estrogen and progesterone receptors are localized in the same cells in these areas
41 of the guinea pig brain [73]. In rats, the time course of PR induction by estrogen and the subsequent decline of PRs is similar to that observed in the onset and decline of reproductive behavior and the LH surge, demonstrating an association between these events [74-77]. Additionally, studies in ERaKO mice have provided some mechanistic insights into the estrogen-dependent regulation of PRs. In animals lacking ERce, it has been shown that estradiol induces both PR mRNA and protein in the medial preoptic area and mediobasal hypothalamus (albeit at lower levels than in wildtype animals) [78,79]. These observations suggest that both ERce and ERfl are involved in estrogen-dependent induction of PR expression in these areas. In other areas of the brain, including the medial amygdala, thalamus, and cortex, estrogen does not induce PR expression [68,69], even though ERs are present in these areas. Whether both receptors are located in the same cells in these areas remains a question and the absence of such colocalization could explain the lack of effect of estradiol on PRs in these specific areas of the brain.
IV. O V A ~ A N
STEROID
IN T H E C E N T R A L CONTROL
ACTION
NERVOUS
SYSTEM--
OF REPRODUCTION
Although it is clear that ovarian steroids act directly on the CNS to control reproductive function, the mechanisms by which this is accomplished are yet to be elucidated. The role of LHRH as a primary regulator of reproductive function suggests that feedback regulation of LHRH synthesis and release by ovarian steroids is the most direct route for modulating reproductive function. Yet, there is little evidence that ovarian steroids act directly on LHRH neurons [21,23-25]. In contrast, much evidence suggests that other neurotransmitter and neuropeptide systems mediate the effects of ovarian steroids on LHRH neurons [77,80]. A selective description of possible mechanisms by which ovarian steroids act on the brain to control reproductive function follows.
A. Effects o f E s t r o g e n s o n the L H R H N e u r o n a l N e t w o r k 1. NEGATIVE FEEDBACK EFFECTS
A classical action of ovarian steroids, particularly estrogen, on LHRH secretion is the suppression of LHRH release via negative feedback (Fig. 1). Although there is much evidence for this action of estradiol on the hypothalamicpituitary axis, the actual mechanisms underlying this effect are not clear. The paucity of LHRH neurons and their lack of distinct anatomical organization have hampered the analysis of the regulation of LHRH gene expression and secretion in
42 mammalian model systems. Therefore, studies of mRNA synthesis and LHRH secretion have proved difficult, and data are highly dependent on the experimental model used. However, due to the development of in vitro cell models and increased sensitivity of techniques for analysis of mRNA levels, our knowledge of LHRH gene expression and release has increased dramatically. Measurement of changes in LHRH mRNA levels in response to estrogen in rodents by in situ hybridization has provided conflicting results, and these have been reviewed extensively [81-83]. Overall, in studies that find a change in LHRH gene expression with ovariectomy, the change found is small, making it unlikely that the large increase in LHRH secretion observed with estrogen withdrawal is solely subserved by a change in mRNA synthesis. Instead, ovarian steroids may be acting at the level of mRNA translation and/or LHRH secretion to elicit their effects. In 1990, Mellon and colleagues [84] developed an immortalized mouse LHRH neuronal cell line GT1, which provides a model to address mechanistic problems on the regulation of LHRH neurons in vitro. Like LHRH neurons in vivo, GT1 cells secrete LHRH in a pulsatile manner [8587]. Furthermore, several groups have shown that GT1 cells express functional ERs. Poletti et al. [88] determined that ERs are present in a subclone of the GT1 cells, based on whole cell binding experiments. Estrogen receptors in immortalized LHRH neurons are low in abundance, but have high affinity for estradiol [88,89]. The low binding capacity of ERs in GT1 cells (6.2 fmol/mg cytosolic protein) [88,89] compared with MCF-7 cells, a human breast carcinomaderived cell line (60 fmol/mg cytosolic protein) [90,91], may help to explain why the majority of in vivo studies failed to detect ERs in LHRH neurons [21,23-25]. The presence of ERa mRNA transcripts in GT1-7 cells has been demonstrated by RT-PCR followed by Southern hybridization [89]. Identification of ERa mRNA- and estradiol-binding sites in immortalized LHRH neurons does not preclude the existence of ERs that are incapable of mediating transcriptional activation. However, transient transfection studies have shown that endogenous ERs can be activated sufficiently by estradiol to drive transcription of a luciferase reporter gene via an estrogen response element (ERE)-dependent mechanism [89]. Although the construct used in these studies contains a consensus vitellogenin ERE, the fact that estrogen action occurred at physiological concentrations of the steroid indicates that LHRH neurons may respond to estrogens under in vivo conditions by activating or repressing particular genes. The pure antiestrogen, ICI182-780, and other antiestrogens specifically block this response [89], suggesting that this effect is ER dependent. Further evidence for functional ERs in GT1 neurons comes from demonstrations of estradiol-dependent modulation of androgen receptors in the GTI-1 subclone [88] and galanin gene expression in the GT1-7 subclone [89].
L6PEZ ET AL.
Examination of the effects of estrogen on transcription of the LHRH gene in cultured LHRH-expressing neuronal cell lines has not revealed direct regulation by ERs through binding to regulatory sequences. Moreover, evaluation of cloned rat and mouse LHRH promoter sequences indicates that they do not contain recognizable EREs. However, the human LHRH promoter does contain an ERE that binds purified ER from calf uterus in DNase footprinting studies [92]. These observations imply that if LHRH neurons do indeed produce functional ERs as was shown by immunocytochemistry [25] and via nested RT-PCR [92a], the necessary cisacting elements are present in the human LHRH gene for ER-dependent regulation. From a molecular perspective, it is extremely difficult to delineate the mechanism(s) by which estradiol elicits ER-dependent gene expression in vivo in LHRH neurons. However, the functional relevance of this site in LHRH-expressing neurons has yet to be demonstrated in LHRH neurons or cell lines. 2. P O S I T I V E FEEDBACK EFFECTS In spontaneous ovulators, including humans, the rising tide of estradiol during the first phase of the reproductive cycle exerts positive feedback effects both on the brain to induce the hypersecretion of LHRH and on the pituitary to increase its sensitivity to LHRH. These events result in the generation of a preovulatory gonadotropin surge and culminate in ovulation [55,93]. The molecular mechanisms governing the hypersecretion of LHRH and the resulting preovulatory surge of LH are still not well understood, however, it is clear that these physiological processes are dependent on the action of the ovarian steroids, estradiol and progesterone, on the CNS [55,93]. Plasma levels of LHRH in the portal circulation increase at, or very near, the time of the LH surge [94-97]. It remains a subject of some controversy whether this increase in LHRH secretion and the steroid-induced LH surge are associated with alterations in the cellular content of LHRH mRNA. In rats and ferrets, the results of studies examining levels of LHRH mRNA relative to a spontaneous or an ovarian steroid-induced LH surge are contradictory [98108]. Even when the general findings among a set of these studies agree, the reports are in discord with one another on the time course of the rise and fall in LHRH gene expression relative to the LH surge. Several laboratories generally have found either little difference in LHRH mRNA over the estrous cycle or at the time of an LH surge [100102,104,107,108], or have found that LHRH mRNA gene expression increases prior to, or at the time of, a preovulatory LH surge in at least a subset of LHRH neurons [98100,103-106]. The sources of the discrepancies among the many studies that have examined LHRH mRNA relative to the LH surge remain largely unknown. (For a thorough examination of these issues, see Refs. 81 and 82.) Those studies that reported a change in LHRH gene expression prior to
CHAPTER 3 Role of Gonadal Steroids in Menopause or at the time of the LH surge show that the increase is relatively small. This scarce change is unlikely to account for the large increase in LHRH secretion at this time, making it likely that other mechanisms, such as translation, processing, and release, contribute to hypersecretion.
AT
3. ACTIVATION OF L H R H NEURONS THE TIME OF THE L H SURGE
Expression of c-Fos or other immediate early gene products by individual neurons can be used as a marker of cell activation and has been used to gain a better understanding of the molecular mechanisms underlying LHRH secretion. Work in several laboratories has shown increased expression of c-Cos mRNA and protein in LHRH neurons near the time of the LH surge in rats [ 107,109-116], mice [ 117], hamsters [118,119], and sheep [120]. This phenomenon is restricted to the time of the LH surge and depends on synaptic transmission [ 110,114]. This suggests that increased transcription of one or more genes in LHRH neurons accompanies the release of LHRH and may be responsible for initiating events that are crucial for the LH surge mechanism. The time course of c-Cos mRNA and protein expression in LHRH neurons has been examined in the rat. Lee and colleagues found that the number of LHRH neurons expressing c-Fos increases during the ascending phase of the LH surge [113]. In a more detailed study, Finn and colleagues found that levels of c-Cos mRNA in LHRH neurons are significantly elevated only after serum LH levels begin to rise, 2 hours before the peak of the LH surge [107]. Concomitantly with the occurrence of the LH surge, LHRH neurons also express Jun, another immediate early gene product [ 112]. Fos and Jun form a heterodimer, AP-1, which is known to regulate expression of a variety of genes [ 121]. Among the possible target genes for AP-1 regulation in LHRH neurons are the genes for LHRH and galanin (GAL) (this is discussed in Section III,A,4). The pattern of c-Fos expression in rat LHRH neurons changes during aging [ 122-124]. Specifically, fewer LHRH/ c-Fos double-labeled neurons were found in middle-aged compared to young rats at the time of either a preovulatory or a sex steroid-induced LH surge. Furthermore, c-Fos induction was delayed and abbreviated in LHRH neurons of middle-aged rats compared to younger animals. These data suggest differences in the molecular physiology of LHRH neurons during the aging process. Witkin and colleagues evaluated c-Fos in LHRH neurons at the time of the surge in female rhesus macaques and observed little expression of the immediate early genes [ 125]. There were no differences in the number of double-labeled cells in intact and ovariectomized females and among ovariectomized females treated with oil, estradiol, or estradiol and progesterone [ 125]. Luteinizing hormone levels were verified and the animals were sacrificed at the time of unchanging or ascending LH levels. The findings in this study can be interpreted either as showing that LHRH neurons in primates are
43 not activated at the time of the LH surge, that LHRH neurons may be activated, but express immediate early gene products other than c-Fos, or that the animals were sacrificed before c-Fos levels in LHRH neurons increased. In summary, it is clear that in at least some species LHRH neurons increase expression of c-Fos at the time of the LH surge. Although it is known that c-Fos and other immediate early genes increase gene expression, the identity of the genes induced in LHRH neurons by immediate early genes and the consequences of their induction remain unknown. 4. GALANIN AS A MARKER OF SEXUALLY DIMORPHIC EFFECTS OF ESTROGEN
Galanin is widely distributed throughout the CNS [126] and is expressed in rats in a subset of LHRH neurons in the diagonal band of Broca and rostral medial preoptic area [ 127,128]. This observation represented the first example of the existence of another neuropeptide in the LHRH neuronal network, and further studies have corroborated the presence of GAL mRNA [129]. Given the participation of endogenous GAL in the control of gonadotropin secretion in female rats during proestrus (reviewed in Section III,B,1), and the responsiveness of GAL to estrogen in other tissues [130], including brain [131,132], it was posited that GAL within LHRH neurons is sensitive to gonadal steroids and may contribute to the molecular mechanisms underlying the hypersecretion of LHRH that induces the preovulatory surge of LH from the pituitary. The expression of GAL mRNA and peptide in rat LHRH neurons is sexually dimorphic and extremely sensitive to estrogen. The sexual dimorphism favors females, in which the majority of LHRH neurons, about 6 0 - 8 0 % , express GAL mRNA/peptide [ 129,131 ]. In male rats, only 10-15% of LHRH neurons express GAL. This same pattern exists at the level of LHRH axon terminals at the median eminence, where GAL and LHRH are copackaged in the same secretory vesicles [133]. The copackaging of these two peptides provides the anatomical and biochemical bases for our observation of cosecretion of LHRH/GAL into hypophyseal portal blood [ 134]. The sexual dimorphic expression of GAL in LHRH neurons is due to the action of sex steroids acting both in adulthood (activating effects of sex steroids) and during early perinatal development (organizing effects of sex steroids). The number of LHRH neurons coexpressing GAL does not change in adult male rats with orchidectomy and subsequent testosterone or estradiol replacement therapy, whereas in female rats ovariectomy drastically reduces the number of LHRH neurons expressing GAL (from 80% to approximately 15%) [135]. Estradiol (or testosterone) therapy restores the number of double-labeled neurons to control levels [ 135]. However, if male rats are neonatally orchidectomized, their LHRH neuronal system is responsive to estrogen in adulthood, demonstrating that the system is sexually differ-
44 entiated [135]. Thus, colocalization of GAL in a subset of LHRH neurons represents a situation in which a primary phenotype (LHRH expression), which is not sexually dimorphic, is expressed in all neurons and another phenotype (GAL expression), which is sexually dimorphic, is expressed in a subset of neurons. Functionally, the observations imply that at least three subsets of LHRH neurons exist in the forebrain of male and female rats: (1) a subset that does not express GAL under any steroidal condition; (2) a subset that expresses GAL independent of steroidal input; and (3) a subset that expresses GAL only under the appropriate steroidal input (Fig. 6). The neurons of this third subset, although present in the brains of both males and females, are much more numerous in the female brain (Fig. 6). Gene expression studies in rats demonstrate a close association between the induction of GAL mRNA in LHRH neurons and the production of an LH surge. Galanin mRNA levels in LHRH neurons increase nearly two-fold between the morning and afternoon of a preovulatory or steroid-induced LH surge [102,107,136]. A thorough examination of GAL gene expression in LHRH neurons relative to a steroid-induced LH surge reveals that the actual increase in GAL mRNA levels in LHRH neurons occurs after LH levels rise and, once elevated, remain high well after the completion of the surge [107]. This rise in GAL mRNA levels in LHRH neurons does not occur if the LH surge is blocked by disrupting synaptic transmission by any number of treatments [136,137]. These data taken together show that the induction of GAL gene expression in LHRH neurons is closely associated with an LH surge and are consistent with a role
L6PEZ ET AL.
for GAL release from LHRH neurons in the LH surge mechanism. The rise in GAL mRNA levels in LHRH neurons at the time of an LH surge is sex steroid dependent and sexually differentiated. Treatment of ovariectomized rats with estradiol at levels sufficient to elicit a small LH surge results in an increase in GAL mRNA levels in LHRH neurons [138]. When estradiol treatment is followed by progesterone administration, a larger release of LH is elicited and higher levels of GAL mRNA in LHRH neurons are attained. Administration of progesterone alone, however, has no effect on serum LH levels or on GAL mRNA concentrations in LHRH neurons [138]. As perinatal exposure to testosterone precludes the ability of rats primed with ovarian steroids to undergo an LH surge, it also precludes the ability of sex steroid priming to elicit GAL gene expression in LHRH neurons [139]. Thus, the same mechanisms responsible for the sexual differentiation of the LH surge mechanism appear to be responsible for sexual differentiation of the induction of GAL gene expression in LHRH neurons in response to ovarian hormone stimulation. As reviewed in Section III,A,3, LHRH neurons express c-Fos mRNA and protein at the time of the LH surge and there is evidence providing strong support for the hypothesis that GAL gene expression is activated by c-Fos. First, there are elements in the 5' flanking region of the rat GAL gene that are able to bind members of the Fos/Jun family of transcription factors [ 140]. Second, both c-Fos and GAL mRNA are induced in LHRH neurons coincident with a preovulatory or steroid-induced LH surge [102,107,109-116,136139]; the induction of both is dependent on synaptic transmission [110,114,136,137]. Third, GAL mRNA is expressed in the majority of the LHRH neurons that colocalize c-Fos protein at the time of a steroid-induced LH surge [115]. Fourth, in estrogen-primed rats, progesterone increases both the number of LHRH neurons that express c-Fos and the level of GAL gene expression in LHRH neurons [111,138]. Finally, the first detectable rise in GAL mRNA levels in LHRH neurons occurs 2 hr after c-fos mRNA is induced in these neurons [ 107], providing sufficient time for translation of c-fos mRNA into c-Fos protein and the initiation of GAL gene expression by an AP-1-dependent mechanism.
B. N e u r o t r a n s m i t t e r a n d N e u r o p e p t i d e S y s t e m s I m p l i c a t e d in L H R H / L H FIGURE 6 Functional subpopulations of luteinizing hormone-releasing hormone (LHRH) neurons. Expression of galanin in luteinizing hormonereleasing hormone neurons is sexually dimorphic and neonatally determined. In this respect, LHRH neurons can be classified as those that never express galanin, those that express galanin under basal conditions and galanin levels are not regulated by estradiol, and those that express galanin only after estrogen treatment. The latter cell subpopulation occurs primarily in female rats and in neonatally orchidectomized males.
Regulation
1. GALANIN Galanin [141] has emerged as an important modulator of reproductive events (see Ref. 126 for a recent review). Galanin stimulates LH secretion in rats, and these effects may be mediated, at least in part, by increasing LHRH secretion. This conclusion is based on the following observations:
CHAPTER 3 Role of Gonadal Steroids in Menopause (1) GAL stimulates LHRH release from arcuate nucleusmedian eminence fragments obtained from male or female rats [128,142,143]; (2) infusion of GAL into the cerebral ventricles stimulates LH secretion in estradiol-primed, ovariectomized rats [143,144], and this effect is blocked by prior infusion of the GAL receptor antagonist, galantide [143]; and (3) infusion of galantide alone into the brain blunts the LH secretion that occurs with either a steroid-induced or spontaneous LH surge [143], as does passive immunization against GAL [145]. Based on these and other observations of higher GAL expression in LHRH neurons of females compared to males, and the close association between induction of GAL gene expression in LHRH neurons and the LH surge, it has been postulated that the GAL derived from LHRH neurons facilitates LH secretion. In fact, early studies showed direct effects of GAL on basal and LHRHinduced LH secretion from the anterior pituitary gland in vitro [127,134]. 2. NOREPINEPHRINE
Noradrenergic neurons play a role in the release of LHRH and LH, as demonstrated by experiments in rodents, rabbits, and primates [80]. In each of these species, disruption of noradrenergic activity results in a suppression of pulsatile LH secretion. This suppression appears to be mediated by ce-adrenergic receptors [80]. In monkeys, the release of norepinephrine from the stalkmedian eminence is pulsatile and the pulses of norepinephrine occur synchronously with pulses of LHRH [146]. celAdrenergic stimulation at the stalk-median eminence level, either by norepinephrine or methoxamine, increases LHRH secretion, whereas cel-adrenergic blockade by prazosin decreases LHRH release [146-148]. Infusions of other adrenergic compounds that act through ce2- or/3-adrenergic receptors have no effect on LHRH secretion [147,148]. The action of norepinephrine on LHRH secretion appears to be mediated, at least in part by prostaglandin E 2, because norepinephrine-stimulated LHRH secretion at the stalk-median eminence is associated with an increase in extracellular prostaglandin E 2 levels, and infusion of prostaglandin E 2 stimulates LHRH secretion [147,148]. Noradrenergic input also appears to be important for the generation of an LH surge in rats. Noradrenergic turnover (a measure of neuronal activity) increases in the medial preoptic area and median eminence at the time of an estrogeninduced LH surge [ 149]. In addition, the episodic release of norepinephrine increases in the medial preoptic area between the morning and afternoon on the day of the estrogen-induced LH surge [150]. An increase in noradrenergic turnover is also found in these same areas coincident with the LH surge in ovarian steroid-primed rats as well as in regularly cycling rats [149,151]. Additionally, disruption of noradrenergic input by lesions or surgical transection blocks the LH surge in regularly cycling rats [152,153], however
45 these effects are transient. Similarly, inhibition of noradrenergic transmission results in the blockade of the LH surge [ 152-156], and these effects appear to be mediated through a-adrenergic receptors. Noradrenergic input may be of less importance to the generation of an LH surge in primates, because ovulation occurs in monkeys even after complete dennervation of the medial basal hypothalamus [157], which severs ascending noradrenergic inputs from the hindbrain. However, this persistence of function may be more a reflection of redundancy in the neural circuitry controlling ovulation than a demonstration that noradrenergic input does not normally affect LHRH secretion at the time of the LH surge in primates. Noradrenergic neurons are restricted to various cell groups in the hindbrain, of which several accumulate estrogen [ 158]. These same cell groups send noradrenergic projections to the hypothalamus [ 159], but it is unknown whether these inputs are estrogen sensitive. In the rat, neurons containing dopamine fi-hydroxylase (a marker of noradrenergic and adrenergic neurons) synapse near, but not on, LHRH neurons [1601, suggesting an indirect effect of norepinephrine on LHRH secretion. However, the immortalized LHRH-containing cell line GT1 does express fil-adrenergic receptors [161], suggesting that a direct interaction occurs between noradrenergic and LHRH neurons. 3. OPIOIDS
Opioids exert tonic inhibitory effects on LHRH and LH secretion in mammals. These effects appear to be mediated by two opioids, fl-endorphin and dynorphin [77,162]. Estrogen can exert direct effects on fl-endorphin- and dynorphinproducing neurons, because subsets of these neurons bind estrogen [ 163,164]. However, it is currently unknown which ER subtype mediates the effects of estrogens on these subpopulations of neurons. Studies in rats indicate several effects of estrogen on opioid systems. For example, ovariectomy increases and estrogen replacement decreases mRNA levels for the precursor of fl-endorphin, proopiomelanocortin, in the arcuate nucleus [165-167]. On the morning of proestrus, when estrogen levels are high, fl-endorphin levels decrease in the arcuate nucleus, the site where it is synthesized, and fl-endorphin levels increase in the median eminence, the site from which it is released into the hypophyseal portal blood [168,169]. Thus, estrogen also appears to influence the transport and/or release of this peptide. A direct action of fl-endorphin on LHRH secretion is supported by studies showing direct synaptic contacts between LHRH and fl-endorphin-containing neurons at the level of the medial preoptic area in rats [ 1 7 0 172] and the mediobasal hypothalamus in primates [173]. Although a study examining the expression of opioid receptors in rat LHRH neurons reported negative results [174], this may be due to a lack of sensitivity in the technique employed. The immortalized LHRH cell line G T I - 1 does
46
L6PEZ ET AL.
express the 6 opioid receptor and both the G T I - 1 and GT1-7 lines respond to opioids with decreases in LHRH secretion [ 175-177]. Although it is clear that opioids influence LHRH/LH secretion, whether these effects are direct or mediated by other neurons is yet to be elucidated. 4. SUBSTANCE P Tachykinins, or neurokinins, are a class of peptides including tachykininA, tachykininB (also known as substance K), and substance P, which are derived from protachykinin precursors. Tachykinins are widely distributed in the brain and are particularly abundant in the medial preoptic area and mediobasal hypothalamus (see Brownstein et al. [ 178] and Ljungdahl et al. [ 179] for a complete description). Infusion of substance P peripherally or directly into the ventricles of the brain stimulates LH secretion in estrogenprimed ovariectomized rats [ 180], whereas blockade of substance P effects, with intracerebroventricular infusion of antiserum, suppresses LH secretion in ovariectomized rats [ 180]. These effects could be explained by direct effects of substance P on LHRH neurons, because substance P-containing neurons appear to form synaptic contacts with LHRH neurons [181]. Substance P levels are responsive to estrogen and fluctuate in regions of the forebrain during the estrous cycle [ 182,183]. Moreover, estrogen treatment of ovariectomized rats increases substance P levels in the medial preoptic area preceding the daily rise in LH secretion [184] and increases substance P immunoreactivity in the medial amygdala [185]. This rise in substance P levels may be subserved by rises in mRNA levels, because estrogen treatment of ovariectomized rats increases tachykinin mRNA levels in the ventromedial nucleus of the hypothalamus and in the medial amygdala [ 186,187]. The effects of estrogen on a subpopulation of substance P neurons in the mediobasal hypothalamus appear to be direct because those neurons concentrate estrogen [ 188].
rapidly around age 35 [194]. This is illustrated in Fig. 7 [195a,195b], which shows the number of remaining ovarian follicles in relation to age. The rate of follicular loss during the first 30 years progresses steadily. Assuming this decline continues at a constant rate, extrapolation reveals that the age at follicular exhaustion would be 80 years. However, this extrapolation is not correct, because around age 35 years the rate of follicular disappearance increases dramatically [190], and follicular exhaustion actually occurs at a much earlier age. The age-related reduction in the follicular pool and the ensuing infertility are associated with a marked elevation in levels of FSH during the follicular phase of the menstrual cycle [ 196-201 ]. This increase in FSH levels commences at age 35 years [197] and is followed several years later by increases in LH levels [196,197,201]. Elevated levels of gonadotropins are the consequence of a loss in ovarian estradiol secretion [198,199]. In particular, elevated follicular levels of FSH in women over age 45 years are associated with reductions in total inhibin levels [202], contributing, therefore, to a differential elevation of pituitary gonadotropins. Inhibin is a dimeric glycoprotein produced by some cell types of the gonad. It is composed of an ce and one of two/3 subunits,/3A and/3B, resulting in two types of inhibins, A (ce-/3A) and B (ce-/3B) [203,204]. Studies have characterized in detail the levels of both inhibins during female reproductive aging [205]. Inhibin B levels are reduced during the follicular phase of the menstrual cycle in older women (Fig. 8). This decrease in inhibin B concentration is accompanied by an elevation in serum FSH levels during the
rY LU
m
106 ee
E) z 10 s
eee
e ee 9 9 oc] 9 9
9
[]
UJ ._1
r't
eee I ~ 1 ~ ~ o
[:] r~ D
__0 10 V. T H E M E N O P A U S E " OF STEROID CENTRAL
HORMONE
NERVOUS
CONSEQUENCES LOSS ON THE
SYSTEM
Reproductive aging is characterized by a gradual reduction of the ovarian follicular pool, which underlies the associated decline in fertility and endocrine changes observed as menopause progresses [ 189-192]. The endocrinological aspects of the menopausal transition have been extensively reviewed [193]. This issue is further discussed in Chapter 9. Rather, the most evident changes in menopause as they relate to the action (or lack thereof) of estrogen on the hypothalamic-pituitary-gonadal axis will be highlighted. The decline in fertility begins in a moderate, progressive fashion in the third decade of life [194,195], but begins to accelerate
_J ._1
0
10 3
_
10
LL _.1
121 n,' 0 n~ a.
9
&
&
A
10 1
I
I
I
I
9
/
mm
10 20 30 40 50 AGE IN YEARS FIGURE 7 Relationship between age in years and number of ovarian primordial follicles. This represents a compilation of the studies of Block (e) [189,195a], Gougeon (El) [195b], and Richardson et al. ( 0 , regular; A, perimenopausal; I, postmenopausal) [ 190]. From Richardson [ 191], with permission from W. B. Saunders.
CHAPTER
3 Role of Gonadal Steroids in Menopause
47
A
.J
-J
200-10
150 rn c 100 .-9 J:::
50-
--
0
~"
20-
r= tL
10
r
2
5
.E e~ ~h,,1
t~l
I
I
It
I
'
' 1 .......................
I
!
.......... 1 ......................
1
0
1
c
--
150 IOOB m
50
0 20 ...
300
15
E
5
~
I
.J
:=
400
....
''
I
''i'
' ' l ~ V ~
~-
C~
200 r
w 100 0
I
!
!
t
.10
-5
0
!
I
I
i
5 10 -10 -5 Days Centered to Ovulation
I
I
I
0
5
10
FIGURE 8 Mean inhibin B, follicle-stimulating hormone (FSH), estradiol (E2), inhibin A, and progesterone (P4) levels in cycling women 2 0 - 3 4 years old ( 9 and 3 5 - 4 6 years old (e). Hormone levels are depicted as centered to the day of ovulation (., P < 0.04; **, P < 0.02, when comparing the two age groups). From [205], Welt, C. K., McNicholl, D. L, Taylor, A. E., and Hall, J. E. (1999). Female reproductive aging is marked by decreased secretion of dimeric inhibin. J. Clin. Endocrinol. Metab. 84, 105-111. 9 The Endocrine Society.
early follicular phase, which results in sustained elevated concentrations of estradiol (Fig. 8). Furthermore, reduced levels of inhibin B are observed during ovulation, resulting in slightly higher FSH levels in older women after the ovulatory surge of FSH and during the luteal phase. Inhibin A concentrations show a different pattern of secretion, i.e, inhibin A levels do not change during the follicular phase of the menstrual cycle and are significantly reduced during the luteal phase (Fig. 8). Luteinizing hormone and progesterone levels remain unchanged among the younger and older age groups (Fig. 8) [205]. Therefore, it appears that low inhibin B levels in older cycling women may be the earliest marker of accelerated follicular loss and reproductive aging [205] (see Chapter 9). The different subsets of LHRH neurons described in Section I,B may represent distinct functional pools of LHRH neurons. It has been shown that LHRH mRNA levels are increased only in the type I neurons during the menopause [33a]. These changes indicate that estrogens under physiological conditions somehow reduce production of LHRH mRNA. Consequently, the increase in LHRH mRNA would be related to an increase in production of the peptide, which would drive an increase in gonadotropins as observed in the menopausal period. The data described so far present an in, teresting view with respect to the action of estrogens in the gonadal axis. The consequences of reduced ovarian steroid action in the CNS are poorly understood, especially in terms
of regulation of the LHRH neuronal network function. However, it is clear that LHRH levels increase with menopause due to the diminished negative feedback of ovarian steroid hormones. Menopause is not a consequence of inadequate production of LHRH. Estradiol exhibits activational effects that are translated to elevations in the expression of various genes; for example, estrogens stimulate the expression of the GAL gene in LHRH neurons [206]. Conversely, estrogens also exhibit potent repression of various gene products, such as/3-endorphin [165-167], some of which are intimately related to the reproductive aging process. Similarly, LHRH mRNA elevation in a subset of LHRH neurons during the menopause [33a] would represent such a repression by estradiol. Thus, determining the physiological consequences of differential gene regulation by estradiol should increase our understanding of the pathophysiology of the menopausal period. Sex steroid secretion is eventually compromised in postmenopausal women [207], consistent with the follicular depletion. However, the consequence of estrogen withdrawal on brain function is poorly understood. Early studies described neuronal hypertrophy in postmenopausal women in the subventricular nucleus (a subdivision of the arcuate nucleus), suggesting a relationship to steroid depletion. More detailed studies have confirmed the original observations by clearly demonstrating that neurons in the arcuate nucleus of postmenopausal women are 30% larger than those of their
48
L 6 P E Z ET AL.
premenopausal counterparts [208,209]. Evidence for the involvement of estradiol depletion as the hypertrophic driver was provided by the observation of ER transcripts in the arcuate nucleus of pre- and postmenopausal women [208]. This anatomical association (ER mRNA presence and hypertrophy) strongly suggests that estrogen negative feedback occurs by a direct effect on a subpopulation of ER mRNAcontaining neurons in the arcuate nucleus [208]. In addition, this hypertrophic response has also been observed in older men [209], even though this effect is smaller in males (22%) than in postmenopausal females (30%) [208,209], reflecting the diverse endocrine status of males and females during reproductive aging. With respect to the identity of the hypertrophied neurons, Rance and Young III [210] demonstrated that the hypertrophied neurons located in the arcuate nucleus contain the tachykinins, substance P and neurokinin B. In addition, the expression of both of these neuropeptides was elevated in postmenopausal compared to premenopausal women [210]. This increase appears to be due to estrogen withdrawal and not to hypothalamic aging. This is based on evidence in monkey studies that show an increase in tachykinin mRNA levels after ovariectomy and abolition of the effect after estrogen replacement therapy [211 ]. These data, together with the evidence in rodents indicating that substance P has a stimulatory role in the control of gonadotropin secretion [180,212-215], suggest that substance P and tachykinins, in general, may be involved in the increased levels of gonadotropins observed during menopause. There are several possible mechanisms by which estrogen withdrawal at menopause results in increased levels of substance P. One possibility is that estrogen acts directly on tachykinin-producing neurons of the arcuate nucleus to inhibit substance P production, and this suppression is released with estrogen withdrawal at menopause. An alternate possibility is that estrogen is stimulatory to a system that inhibits substance P production. Estrogen is stimulatory (not inhibitory) to substance P production in rats, therefore these models are of little help in elucidating molecular mechanisms operating in humans. Such a task may require a nonhuman primate model whose reproductive physiology is more similar to humans. Thus, resolution of the mechanisms by which substance P levels are increased and the consequences of this increase in postmenopausal women await further investigation.
A. H o t F l u s h One of the most common and troubling physiological manifestations of menopause is the hot flush (hot flash). Hot flushes occur in about 85% of menopausal women whether the menopause occurs naturally or surgically [216,217]. They often first occur within months of the last menses and can persist for years after ovarian quiescence. Hot flushes
have been characterized by a variety of investigators [218222]. They begin by an ascending flush of the upper body, starting at the thorax caused by cutaneous vasodilatation. This results in a feeling of warmth. Blood pressure is stable, but changes in heart rate often occur. The vasodilatation causes a decrease in body core temperature, resulting in a sensation of cold that often elicits shivering. Hot flushes often occur at night (night sweats) and disrupt normal sleep patterns, leading to complaints such as irritability, fatigue, and forgetfulness. Estrogen replacement virtually eliminates hot flushes, demonstrating their origin in the estrogen-withdrawal that occurs with ovarian follicular depletion at menopause [222-224]. Hot flushes appear to result from a dysfunction of thermoregulatory centers in the hypothalamus and are correlated with pulses of circulating estrogen and gonadotropin secretion in menopausal women (see Fig. 9; see also Chapter 13) [225-228]. In studies employing frequent blood sampling, hot flushes have been temporally correlated with pulses of LH (Fig. 9) [229,230], yet it is unlikely that LH release initiates hot flushes. Hot flushes occur in women with disrupted pituitary function [231,232] or who have been treated with an LHRH agonist to suppress LH pulses [233,234]. Moreover, the rise in LH following discontinuation of the latter treatment does not induce hot flushes [235], further evidence against an LH-initiated mechanism for hot flushes. Because LH levels reflect LHRH release from the hypothalamus, hot flushes are also closely associated with LHRH secretion, but like LH, LHRH release appears to be correlated with, not causative of, hot flushes. This conclusion is supported by studies showing that women with abnormal LHRH
FIGURE 9 Schematicrepresentation of the consequences of the ovarian insufficiency occurring during menopausal transition. During menopause, the exhaustion of ovarian follicles results in a decrease in circulating levels of estradiol and inhibin. Under physiological conditions estrogenexhibits a tonic inhibitory influence on both the LHRH neuronal system and the thermoregulatory center in the hypothalamus. The lack of estrogen-dependent inhibition results in increased secretion of gonadotropins and the advent of hot flush episodes that are coincident with secretory episodes of LH release. This coincidence can be dissociated by treatment with LHRH analogs, suggesting an association between these phenomenarather than a cause-effect relationship.
CHAPTER3 Role of Gonadal Steroids in Menopause secretion and associated reproductive impairments experience hot flushes in response to estrogen withdrawal [236]. Thus, hot flushes occur in the absence of LHRH secretion. However, in women in which suppression of LHRH secretion appears to result from perturbations in neurotransmitter/ neuropeptide systems that control LHRH release [237,238], estrogen withdrawal did not result in hot flushes. It is currently thought that estrogen withdrawal influences neuropeptide/transmitter systems that act at the medial preoptic area/ hypothalamus to cause an overall lowering of the thermoregulatory set point. The association of hot flushes with increased LH secretion is due to the associated stimulation of LHRH neurons that also reside in these areas of the brain. In summary, during menopause, the exhaustion of ovarian follicles results in a decrease in circulating levels of estradiol and inhibin. Under physiological conditions estrogen exhib, its a tonic inhibitory influence of both the LHRH neuronal system and the thermoregulatory center in the hypothalamus (Fig. 9). The lack of estrogen-dependent inhibition results in increased secretion of gonadotropins and in the appearance of hot flush episodes that are coincident with secretory episodes of LH release. Noradrenergic neurons are one of several neurotransmitter and neuropeptide systems that have been implicated in the hot flush mechanism. Noradrenergic neurons are responsive to estrogen [239], and a subset of noradrenergic neurons projects to the medial preoptic area [240]. Furthermore, norepinephrine modulates LHRH secretion (reviewed in Section II,B,2) and thermoregulation [241 ]. The strongest evidence for hot flushes being mediated by disruption in noradrenergic systems comes from studies showing that clonidine, an c~-adrenergic agonist, reduces the frequency of hot flushes in postmenopausal women [242-245]. The exact mechanism by which clonidine reduces hot flush frequency is unknown, but possible mechanisms include centrally mediated stabilization of thermoregulatory centers in the medial preoptic area/hypothalamus and/or peripherally mediated blockade of vasomotor symptoms. The latter mechanism is supported by the observation that clonidine treatment in humans reduced the increase in forearm blood flow produced by epinephrine, norepinephrine, and angiotensin [246]. In the CNS, there is evidence suggesting that clonidine's action is dependent on stimulation of a2-adrenergic receptors. In addition, there are data indicating that clonidine may be effective by stimulating imidazoline-preferring binding sites, a potential new family of noradrenergic receptors [247]. The mechanisms by which estrogen alters noradrenergic systems are unclear. Some studies show that estrogen can increase central noradrenergic activity by increasing the activity of the rate-limiting enzyme in its synthesis [248], decreasing levels of an enzyme responsible for its catabolism [249], and inhibiting its reuptake [250]. Other studies show that estrogen lowers noradrenergic activity. In rats, ovariectomy (estrogen withdrawal) increases [251-253] and estro-
49 gen replacement decreases [254] hypothalamic noradrenergic activity. Thus, if hot flushes are caused by changes in noradrenergic input due to estrogen withdrawal, it is unclear whether this is the result in an overall increase or decrease in activity. Norepinephrine applied directly to the medial preoptic area of nonhuman primates stimulates many of the same physiological changes associated with hot flushes, such as peripheral vasodilatation, heat loss, and a decrease in core body temperature [255,256]. This supports the idea that an increase in noradrenergic activity may cause hot flushes. However, unlike women experiencing hot flushes, the animals in these studies also showed bradycardia and hypotension; therefore, this model does not completely mimic the physiological changes occurring in women during a hot flush. This may be due to restricted application of the norepinephrine to the medial preoptic area, or perhaps an increase in noradrenergic activity is not fully responsible for hot flushes. Estrogen withdrawal may result in changes in noradrenergic activity, and such changes could account for disruption in thermoregulatory centers producing hot flushes; however, other systems such as the opioid system are also implicated in the etiology of hot flushes. Opioid-containing neurons richly innervate the medial preoptic area [257] and subsets of opioid-containing neurons bind estrogen [258], similar to noradrenergic neurons. Also, like norepinephrine, opioids influence both reproduction (reviewed in Section III,B,3) and thermoregulation [259,260]. Several observations support the idea that opioids are involved in the etiology of hot flushes. First, placebo treatment in women significantly decreases hot flush frequency [261 ]. This effect could be mediated by an increase in opioid transmission since placebo effects can be blocked by blockade of opioid neural transmission [262]. Second, opiate and estrogen withdrawal share many of the same physiological symptoms, including hot flushes and sleep disturbances; however, opiate withdrawal is associated with additional symptoms that do not occur with estrogen withdrawal. Possible explanations for this dissociation are that opioids interact with other neural systems to induce the common symptomatology. Alternatively, a subset of opioid-producing neurons, those responsive to estrogen, may account for the symptoms present with both opiate and estrogen withdrawal, whereas the remaining opioid-producing neurons, those not sensitive to estrogen, are responsible for the additional symptoms seen with opiate withdrawal. Intriguingly, opiatewithdrawal symptoms can be alleviated in rats by estrogen treatment [263] lending support to the latter hypothesis. It has been shown in women that proopiomelanocortin (POMC; one of the precursor of opioids) mRNA levels in the hypothalamus decrease after menopause [264]. However, using a primate model (cynomolgus monkeys), the same group has shown that POMC mRNA levels are similar in untreated, estrogen-treated, and estrogen/progesteronetreated ovariectomized animals [211], suggesting that the
50
L6PEZ ET AL.
decrease in POMC gene expression observed in postmenopausal women is due to hypothalamic aging, not to ovarian steroid withdrawal [211 ]. Presently, nothing is known about the physiological relevance of increased substance P levels in the arcuate nucleus of postmenopausal women (see Section III,B,4 for a more detailed description). However, like several of the neurotransmitters mentioned above, tachykinins have also been implicated in the etiology of some types of flushing, and might therefore play a role in hot flushes in menopausal women. In healthy men, substance P infusion induces both vasodilatation [265,266] and flushing. Furthermore, in individuals with carcinoid tumors, pentagastrin infusion, food, or alcohol can induce flushing, and this flushing is usually associated with increases in tachykinin levels [267-272]. Concurrent treatment with a somatostatin analog will decrease both hot flushing and tachykinin levels [268-272]. Although there is a strong association between increased tachykinin levels and flushing, the association is not absolute, suggesting that tachykinins are not the only cause of flushing in these individuals. Whether tachykinins, especially of central origin, play a role in the postmenopausal flush is yet to be determined.
B. C o g n i t i v e F u n c t i o n s a n d M o o d Many investigators have examined the effect of estrogen on cognition in postmenopausal women using a variety of tests that assess cognition. Although there are discrepancies in the results of these studies, in general, women receiving estrogen replacement after natural or surgical menopause score higher on a very limited number of memory tests. Thus, estrogen does not appear to enhance cognition globally, but rather appears to enhance only specific aspects of cognition (see Chapter 20). Some early studies of estrogen replacement in postmenopausal women found that estrogen improved memory as assessed by self-report and by objectively administered memory tests [224,273], whereas other studies did not [274,275]. Sherwin and colleagues undertook a series of studies examining the effects of estrogen replacement in women who had undergone surgical menopause. Memory tests were administered to a group of women prior to surgery. Postsurgery, the women were treated with estrogen, androgen, or estrogen and androgen for 3 months, treated with placebo for the fourth month, and then randomly crossed over to a different hormonal treatment group and treated for an additional 3 months. In the Paragraph-Recall Test of short-term verbal memory, scores did not differ significantly pre- and postoperatively in women who received sex steroids, whereas scores were lower post-operatively in women who received placebo [276]. Similarly, women who underwent surgical menopause were in another study in which patients were
treated with estrogens and no differences were observed on a retention test of new material (Paired-Associate Test) preversus postsurgery, whereas there was a decline in scores postsurgery in women treated with placebo [277]. In addition, scores on the Paragraph-Recall Test increased in the estrogen-treated women after surgery, but were merely maintained in the placebo-treated group [277]. Finally, there were no significant group differences pre- and postsurgery for two other memory tests (immediate or delayed visual recall) [277]. In a third study involving a larger number of women who underwent surgical menopause, Sherwin and colleagues found that women treated with estrogen scored higher postsurgery on specific cognitive tests (immediate and delayed recall of the Paired-Associate Test), whereas women treated with placebo had lower scores postsurgery [278]. Less has been done to examine estrogen's effects on cognition in women undergoing natural menopause. In one study, women between 45 and 60 years did not show any differences on two subtests of the Weschler Adult Intelligence Scale when treated with estrogen or with placebo, either when scores were compared within groups pre- versus postsurgically or when compared between groups postsurgically [279]. Other studies have found that postmenopausal women who had or were currently using estrogen scored higher on some cognitive tests involving verbal memory (Category Fluency Test, Mini-Mental State Examination, or immediate and delayed Paragraph Recall), but not on other verbal, visual, or spatial memory tests [280,281 ]. Taken together then, these studies examining the effect of estrogen replacement on cognitive functioning in women undergoing estrogen withdrawal show that estrogen appears to have positive effects on some specific aspects of verbal memory. However, this conclusion should be taken with caution, because most studies finding positive effects of estrogen on cognitive measures did not control for the effect of hot flushes that accompany estrogen withdrawal in a majority of women. This is important because hot flushes occur frequently at night, disrupting sleep, and this effect alone could be responsible for the poorer performance of women after estrogen withdrawal than before. Resolution of this confounding factor awaits the results of current studies examining the effect of estrogen replacement on cognition in postmenopausal women. At the present time, little is known about the neural mechanisms underlying estrogen's effects on memory; however, a likely site of action is the hippocampus. The hippocampus is intimately involved in learning and memory, and ERce and ERfl are both found in the hippocampus [58,282]. Studies in rats demonstrate that estrogen withdrawal after ovariectomy results in a decrease in dendritic spines in specific regions of the hippocampus and that this decrease can be reversed with estrogen treatment [283]. Moreover, the density of dendritic spines on a subtype of hippocampal neurons varies over the estrous cycle, with the highest number found at proestrus, coincident with peak estrogen levels and
51
CHAPTER 3 Role of Gonadal Steroids in Menopause the lowest number at estrus, coincident with basal estrogen levels [284]. Changes in dendritic spines are thought to be one of the mechanisms involved in learning and m e m o r y [285]. Thus the dynamic changes estrogen exerts on dendritic morphology could be a factor explaining at least some of the effects of estrogen on cognition. In addition to cognition, estrogen may influence m o o d in women. Estrogen replacement therapy compared to placebo has been shown to decrease depression scores in postmenopausal w o m e n [279]. In a study of w o m e n over age 50 years, depression scores were found to increase with age in w o m e n that did not receive estrogen replacement, but did not increase in w o m e n who received estrogen replacement [286]. This positive association between m o o d and estrogen is also seen in w o m e n undergoing surgical menopause. Sherwin and Gelfand found, in a prospective, cross-over study of w o m e n who underwent surgical menopause, that depression scores were higher in w o m e n treated with placebo than in w o m e n who were treated with sex steroid hormones (estrogen and/or androgen) [287]. The scores of these w o m e n were also higher than those of w o m e n who underwent a hysterectomy only [287]. Similar results were obtained in another group of w o m e n who, after undergoing a surgical hysterectomy and ovariectomy 4 years earlier, reported more positive moods after 2 years of treatment with either estrogen or estrogen plus testosterone, compared to untreated w o m e n [276]. In all of these studies, estrogen was given in doses used for estrogen replacement therapy in postmenopausal women, and the w o m e n treated were not clinically depressed. A study of depressed w o m e n given estrogen at doses used for estrogen replacement showed a decrease in depression scores in a majority of w o m e n classified as mildly depressed but not in a majority of w o m e n classified as clinically depressed [288]. However, in a study of w o m e n with severe, refractory depression, estrogen enhanced m o o d when given at pharmacological doses [289]. Thus, estrogen treatment at doses used for replacement therapy appears to have a favorable effect on m o o d in women, but does not appear to alleviate symptoms of clinically depressed women. Estrogen may mediate its effects on m o o d through the neurotransmitter, serotonin. A deficit in serotoninergic brain activity is implicated in the etiology of depression, and increasing serotoninergic activity with serotonin reuptake inhibitors presently serves as one of the most effective treatments for depression [290]. Estrogen increases serotoninergic activity in a variety of ways, including decreasing serotonin catabolism [249], decreasing the m R N A for the transporter responsible for its reuptake [291 ], increasing the m R N A levels of the rate-limiting enzyme in its production [292], and reducing serotonin 1A autoreceptors [293]. Thus, with the advent of menopause, it is likely that serotoninergic activity decreases, which could result in an increase in negative moods in at least a population of women. Currently, the ER has not been localized to serotoninergic cells of rats
[294], but estradiol does induce the progestin receptor in serotoninergic neurons in primates [295] suggesting that these neurons also express the ER. The presence of both estrogen and progestin receptors in these neurons would provide a mechanism by which sex steroids could directly modulate serotoninergic activity, thereby influencing mood.
References 1. U.S. Congress, Office of Technology Assessment (1992). The Menopause, Hormone Therapy, and Womens' Health, OTA-BP-BA-88, 3-7. 2. McDonnell, D. E, Clevenger, B., Dana, S., Santiso-Mere, D., Tzukerman, M. T., and Gleeson, M. A. (1993). The mechanism of action of steroid hormones: A new twist to an old tale. J. Clin. Pharmacol. 33, 1165-1172. 3. Katzenellenbogen, J. A., O'Malley, B. W., and Katzenellenbogen, B. S. (1996). Tripartite steroid hormone receptor pharmacology: Interaction with multiple effector sites as a basis for the cell- and promoter-specific action of these hormones. Mol. Endocrinol. 10, 119-131. 4. Yang, N. N., Venugopalan, M., Hardikar, S., and Glasebrook, A. (1996). Identification of an estrogen response element activated by metabolites of 17fl-estradiol and raloxifene. Science 273, 12221225. 5. Mason, A. J., Hayflick, J. S., Zoeller, R. T., Young, W. S., III, Phillips, H. S., Nikolics, K., and Seeburg, E H. (1986). A deletion truncating the GnRH gene is responsible for hypogonadism in the hpg mouse. Science 234, 1366-1371. 6. Kallmann, E J., Schoenfeld, W. A., and Barrera, S. E. (1944). The genetic aspects of primary eunuchoidism. Am. J. Ment. Defic. 48, 203-236. 7. Mason, A. J., Pitts, S. L., Nikolics, K., Szonyi, E., Wilcox, J. N., Seeburg, P. H., and Stewart, T. A. (1986). The hypogonadal mouse: Reproductive functions restored by gene therapy. Science 234, 1372-1378. 8. Crowley, W. F., Jr. and McArthur, J. W. (1980). Simulation of the normal menstrual cycle in Kallman's syndrome by pulsatile administration of luteinizing hormone-releasing hormone (LHRH). J. Clin. Endocrinol. Metab. 51, 173-175. 9. Schwanzel-Fukuda, M., and Pfaff, D. W. (1988). Origin of luteinizing hormone-releasing hormone neurons. Nature (London) 338, 161-164. 10. Silverman, A. J., Krey, L. C., and Zimmerman, E. A. (1979). A comparative study of the luteinizing hormone releasing hormone (LHRH) neuronal networks in mammals. Biol. Reprod. 20, 98-110. 11. Marshall, E E., and Goldsmith, E C. (1980). Neuroregulatory and neuroendocrine GnRH pathways in the hypothalamus and forebrain of the baboon. Brain Res. 193, 353-372. 12. Merchenthaler, I., Gorcs, T., Setalo, G., Petrusz, E, and Flerko, B. (1984). Gonadotropin-releasing hormone (GnRH) neurons and pathways in the rat brain. Cell. Tissue Res. 237, 15-29. 13. King, J. C., Anthony, E. L., Fitzgerald, D. M., and Stopa, E. G. (1985). Luteinizing hormone-releasing hormone neurons in human preoptic/hypothalamus: Differential intraneuronal localization of immunoreactive forms. J. Clin. Endocrinol. Metab. 60, 88-97. 14. Standish, L. J., Adams, L. A., Vician, L., Clifton, D. K., and Steiner, R. A. (1987). Neuroanatomical localization of cells containing gonadotropin-releasing hormone messenger ribonucleic acid in the primate brain by in situ hybridization histochemistry. Mol. Endocrinol. 1,371-376.
5
2
L
6
P
15. Goldsmith, P. C., Thind, K. K., Song, T., Kim, E. J., and Boggan, J. E. (1990). Location of the neuroendocrine gonadotropin-releasing hormone neurons in the monkey hypothalamus by retrograde tracing and immunostaining. J. Neuroendocrinol. 2, 157-168. 16. Silverman, A.-J., Livne, I., and Witkin, J. W. (1994). The gonad atropin-releasing hormone (GnRH), neuronal systems: Immunocytochemistry and in situ hybridization. In "The Physiology of Reproduction" E. Knobil, and J. D. Neill, (eds.), pp. 1683-1709. Raven Press, New York. 17. Merchenthaler, I., Kovacs, G., Lavasz, G., and Setalo, G. (1980). The preoptico-infundibular LH-RH tract of the rat. Brain Res. 198, 63 -74. 18. Boothby, M., Ruddon, R. W., Anderson, C., McWilliams, D., and Boime, I. (1981). A single gonadotropin alpha-subunit gene in normal tissue and tumor-derived cell lines. J. Biol. Chem. 256, 51215127. 19. Sairam, M. R. (1981). Primary structure of the ovine pituitary follitropin alpha-subunit. Biochem. J. 197, 535-539. 20. Dufau, M. L., and Catt, K. J. (1978). Gonadotropin receptors and regulation of steroidogenesis in the testis and ovary. Vitam. Horm. (N. Y.) 36, 461-592. 21. Shivers, B. D., Harlan, R. E., Morrell, J. I., and Pfaff, D. W. (1983). Absence of oestradiol concentration in cell nuclei of LHRH-immunoreactive neurones. Nature (London) 304, 345-347. 22. Leranth, C., MacLusky, N. J., Brown, T. J., Chen, E. C., Redmond, D. E., and Naftolin, F. (1992). Transmitter content and afferent connections of estrogen-sensitive progestin receptor-containing neurons in the primate hypothalamus. Neuroendocrinology 55, 667-682. 23. Watson, R. E., Langub, M. C., and Landis, J. W. (1992). Further evidence that most luteinizing hormone-releasing hormone neurons are not directly estrogen-responsive: Simultaneous localisation of luteinizing hormone-releasing hormone and estrogen receptor immunoreactivity. J. Neuroendocrinol. 4, 311-318. 24. Lehman, M. N., and Karsch, E J. (1993). Do gonadotropin-releasing hormone, tyrosine hydroxylase-, and beta-endorphin-immunoreactive neurons contain estrogen receptors? A double-label immunocytochemical study in the Suffolk ewe. Endocrinology (Baltimore) 133, 887-895. 25. Butler, J. A., Sj6berg, M., and Coen, C. W. (1999). Evidence for oestrogen receptor alpha-immunoreactivity in gonadotrophin-releasing hormone-expressing neurones. J. Neuroendocrinol. 11, 331-335. 26. Pfaff, D. W. (1989). Patterns of steroid hormone effects on electrical and molecular events in hypothalamic neurons. Mol. Neurobiol. 3, 135-154. 27. Baulieu, E. E., and Robel, E (1995). Non-genomic mechanisms of action of steroid hormones. Ciba Found. Symp. 191, 24-37. 28. Lei, Z. M., Rao, C. V., Kornyei, J. L., Licht, E, and Hiatt, E. S. (1993). Novel expression of human chorionic gonadotropin/luteinizing hormone receptor gene in brain. Endocrinology (Baltimore) 132, 2262-2270. 29. Lei, Z. M., and Rao, C. V. (1994). Novel presence of luteinizing hormone/human chorionic gonadotropin (hCG) receptors and the down-regulating action of hCG on gonadotropin-releasing hormone gene expression in immortalized hypothalamic GT1-7 neurons. Mol. Endocrinol. 8, 1111-1121. 30. Besecke, L. M., Guendner, M. J., Schneyer, A. L., Bauer-Dantoin, A. C., Jameson, J. L., and Weiss, J. (1996). Gonadotropin-releasing hormone regulates follicle-stimulating hormone-fl gene expression through an activin-follistatin autocrine or paracrine loop. Endocrinology (Baltimore) 137, 3667-3673. 31. MacConell, L. A., Barth, S., and Roberts, V. J. (1996). Distribution of follistatin messenger ribonucleic acid in the rat brain: Implications for a role in the regulation of central reproductive functions. Endocrinology (Baltimore) 137, 2150-2158.
E
Z
ET AL.
32. MacConell, L. A., Widger, A. E., Barth-Hall, E., and Roberts, V. J. (1998). Expression of activin and follistatin in the rat hypothalamus: Anatomical association with gonadotropin-releasing hormone neurons and possible role of central activin in the regulation of luteinizing hormone release. Endocrine 9, 233-241. 33. Rance, N. E., Young, W. S., III, and McMullen, N. T. (1994). Topography of neurons expressing luteinizing hormone-releasing hormone gene transcripts in the human hypothalamus and basal forebrain. J. Comp. Neurol. 339, 573-586. 33a. Rance, N. E., and Uswandi, S. V. (1996). Gonadotropin-releasing hormone gene expression is increased in the medial basal hypothalamus of postmenopausal women. J. Clin. Endocrinol. Metab. 81, 3540-3546. 34. Goldsmith, P. C., and Song, T. (1987). The gonadotropin-releasing hormone containing ventral hypothalamic tract in the fetal rhesus monkey (Macaca mulatta). J. Comp. Neurol. 257, 130-139. 35. Quanbeck, C., Sherwood, N. M., Millar, R. P., and Terasawa, E. (1997). Two populations of luteinizing hormone-releasing hormone neurons in the forebrain of the rhesus macaque during embryonic development. J. Comp. Neurol. 380, 293-309. 36. Urbanski, H. F., White, R. B., Fernald, R. D., Kohama, S. G., Garyfallou, V. T., and Densmore, V. S. (1999). Regional expression of mRNA encoding a second form of gonadotropin- releasing hormone in the macaque brain. Endocrinology (Baltimore) 140, 1945-1948. 37. Terasawa, E., and Sherwood, N. M. (1997). The presence of a luteinizing hormone-releasing hormone (LHRH) fragment in the fetal monkey forebrain. 79th Annu. Meet. Endocr. Soc., p. 380. 38. Pau, C. Y., Pau, K. Y., and Spies, H. G. (1998). Putative estrogen receptor beta and alpha mRNA expression in male and female rhesus macaques. Mol. Cell. Endocrinol. 146, 59-68. 39. Blurton-Jones, M. M., Roberts, J. A., and Tuszynski, M. H. (1999). Estrogen receptor immunoreactivity in the adult primate brain: neuronal distribution and association with p75, trkA, and choline acetyltransferase. J. Comp. Neurol. 405, 529-542. 40. Gibbs, R. B., and Aggarwal, P. (1998). Estrogen and basal forebrain cholinergic neurons: Implications for brain aging and Alzheimer's disease-related cognitive decline. Horm. Behav. 34, 98-111. 41. Beato, M., Truss, M., and Chavez, S. (1996). Control of transcription by steroid hormones. Ann. N. Y. Acad. Sci. 784, 93-123. 42. McEwen, B. S. (1999). Clinical review 108: The molecular and neuroanatomical basis for estrogen effects in the central nervous system. J. Clin. Endocrinol. Metab. 84, 1790-1797. 43. Green, S., Walter, P., Greene, G., Krust, A., Goffin, C., Jensen, E., Scrace, G., Waterfield, M., and Chambon, P. (1986). Cloning of the human oestrogen receptor cDNA. J. Steroid Biochem. 24, 77-83. 44. Kuiper, G. G., J. M., Enmark, E., Pelto-Huikko, M., Nilsson, S., and Gustafsson, J.-A. (1996). Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc. Natl. Acad. Sci. U.S.A. 93, 5925 -5930. 45. Kumar, V., Green, S., Stack, G., Berry, M., Jin, J. R., and Chambon, P. (1987). Functional domains of the human estrogen receptor. Cell (Cambridge, Mass.) 51,941-951. 45a. Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994). CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673-4680. 46. Mosselman, S., Polman, J. and Dijkema, R. (1996). ER beta: Identification and characterization of a novel human estrogen receptor. FEBS Lett. 392, 49-53. 47. Enmark, E., Pelto-Huikko, M., Grandien, K., Lagercrantz, S., Lagercrantz, J., Fried, G., Nordenskj61d, M., and Gustafsson, J. A. (1997). Human estrogen receptor beta-gene structure, chromosomal localization, and expression pattern. J. Clin. Endocrinol. Metab. 82, 4258-4265.
CHAPTER 3 Role of Gonadal Steroids in Menopause 48. Tremblay, G. B., Tremblay, A., Copeland, N. G., Gilbert, D. J., Jenkins, N. A., Labrie, E, and Giguere, V. (1997). Cloning, chromosomal localization, and functional analysis of the murine estrogen receptor ft. Mol. Endocrinol. 11, 353-365. 49. Ogawa, S., Inoue, S., Watanabe, T., Hiroi, H., Orimo, A., Hosoi, T., Ouchi, Y., and Muramatsu, M. (1998). The complete primary structure of human estrogen receptor beta (hERbeta) and its heterodimerization with ER alpha in vivo and in vitro. Biochem. Biophys. Res. Commun. 243, 122-126. 50. Leygue, E., Dotzlaw, H., Lu, B., Glor, C., Watson, P. H., and Murphy, L. C. (1998), Estrogen receptor beta: Mine is longer than yours? J. Clin. Endocrinol. Metab. 83, 3754-3755. 50a. Fuqua, S. A., Schiff, R., et al. (1999). Expression of wild-type estrogen receptor beta and variant isoforms in human breast cancer. Cancer Res. 59(21), 5425-5428. 50b. Moore, J. T., McKee, D. D., et al. (1998). Cloning and characterization of human estrogen receptor beta isoforms. Biochem. Biophys. Res. Commun. 247(1), 75-78. 50c. Peterson, D. N., Tkalcevic, G. T., et al. (1988). Identification of estrogen receptor beta2, a functional variant of estrogen receptor beta expressed in normal rat tissues. Endocrinology 139(3), 1082-1092. 51. Pettersson, K., Grandien, K., Kuiper, G. G., and Gustafsson, J. A. (1997). Mouse estrogen receptor beta forms estrogen response element-binding heterodimers with estrogen receptor alpha. Mol. Endocrinol. 11, 1486-1496. 52. Kuiper, G. G. J. M., Carlsson, B., Grandien, K., Enmark, E., H~iggblad, J., Nilsson, S., and Gustafsson, J.-,~. (1997). Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptor a and ft. Endocrinology (Baltimore) 138, 863-870. 53. Paech, K., Webb, P., Kuiper, G. G., Nilsson, S., Gustafsson, J., Kushner, P. J., and Scanlan, T. S. (1997). Differential ligand activation of estrogen receptors ERalpha and ERbeta at AP1 sites. Science 277, 1508-1510. 54. Pfaff, D. W., Schwartz-Giblin, S., McCarthy, M. M., and Kow, L.-M. (1994). Cellular and molecular mechanisms of female reproductive behavior. In "The Physiology of Reproduction." E. Knobil and J. D. Neill (eds.), pp. 107-220. Raven Press, New York. 55. Freeman, M. E. (1994). The neuroendocrine control of the ovarian cycle of the rat. In "The Physiology of Reproduction." E. Knobil and J. D. Neill (eds.), pp. 613-658. Raven Press, New York. 56. Shughrue, P., Scrimo, P., Lane, M., Askew, R., and Merchenthaler, I. (1997). The distribution of estrogen receptor-fl mRNA in forebrain regions of the estrogen receptor-a knockout mouse. Endocrinology (Baltimore) 138, 5649-5652. 57. Shughrue, P. J., Lane, M. V., and Merchenthaler, I. (1997). Comparative distribution of estrogen receptor-a and -/3 mRNA in the rat central nervous system. J. Comp. Neurol. 388, 507-525. 58. Laflamme, N., Nappi, R. E., Drolet, G., Labrie, C., and Rivest, S. (1998). Expression and neuropeptidergic characterization of estrogen receptors (ERalpha and ERbeta) throughout the rat brain: Anatomical evidence of distinct roles of each subtype. J. Neurobiol. 36, 357-378. 59. Li, X., Schwartz, P. E., and Rissman, E. F. (1997). Distribution of estrogen receptor-/3-1ike immunoreactivity in rat forebrain. Neuroendocrinology 66, 63-67. 60. Shughrue, P. J., Scrimo, P. J., and Merchenthaler, I. (1998). Evidence for the colocalization of estrogen receptor-beta mRNA and estrogen receptor-alpha immunoreactivity in neurons of the rat forebrain. Endocrinology (Baltimore) 139, 5267-5270. 61. Wise, P. M., and Camp, P. (1984). Changes in concentrations of estradiol nuclear receptors in the preoptic area, medial basal hypothalamus, amygdala, and pituitary gland of middle-aged and old cycling rats. Endocrinology (Baltimore) 114, 92-98. 62. Wise, P. M., McEwen, B. S., Parsons, B., and Rainbow, T. C. (1984).
53 Age-related changes in cytoplasmic estradiol receptor concentrations in microdissected brain nuclei: Correlations with changes in steroid-induced sexual behavior. Brain Res. 321, 119-126. 63. Brown, T. J., MacLusky, N. J., Shanabrough, M., and Naftolin, E (1990). Comparison of age- and sex-related changes in cell nuclear estrogen-binding capacity and progestin receptor induction in the rat brain. Endocrinology (Baltimore) 126, 2965-2972. 64. Wise, P. M., and Parsons, B. (1984). Nuclear estradiol and cytosol progestin receptor concentrations in the brain and the pituitary gland and sexual behavior in ovariectomized estradiol-treated middle-aged rats. Endocrinology (Baltimore) 115, 810-816. 65. Kato, J. (1985). Progesterone receptors in brain and hypophysis. In "Current Topics in Neuroendocrinology," D. Ganten and D. Pfaff, (eds.), pp. 32-81. Springer-Verlag, Berlin. 66. MacLusky, N. J., and McEwen, B. S. (1978). Oestrogen modulates progestin receptor concentrations in some rat brain regions but not in others. Nature (London) 274, 276-278. 67. Parsons, B., Rainbow, T. C., MacLusky, N. J., and McEwen, B. S. (1982). Progestin receptor levels in rat hypothalamic and limbic nuclei. J. Neurosci. 2, 1446-1452. 68. Romano, G. J., Krust, A., and Pfaff, D. W. (1989). Expression and estrogen regulation of progesterone receptor mRNA in neurons of the mediobasal hypothalamus: An in situ hybridization study. Mol. Endocrinol. 3, 1295-1300; erratum: Ibid., p. 1860 (1989). 69. Lauber, A. H., Romano, G. J., and Pfaff, D. W. (1991). Sex difference in estradiol regulation of progestin receptor mRNA in rat mediobasal hypothalamus as demonstrated by in situ hybridization. Neuroendocrinology 53, 608- 613. 70. Bethea, C. L., Fahrenbach, W. H., Sprangers, S. A., and Freesh, E (1992). Immunocytochemical localization of progestin receptors in monkey hypothalamus: Effect of estrogen and progestin. Endocrinology (Baltimore) 130, 895-905. 71. Simerly, R. B., Carr, A. M., Zee, M. C., and Lorang, D. (1996). Ovarian steroid regulation of estrogen and progesterone receptor messenger ribonucleic acid in the anteroventral periventricular nucleus of the rat. J. Neuroendocrinol. 8, 45-56. 72. Brown, T. J., Yu, J., Gagnon, M., Sharma, M., and MacLusky, N. J. (1996). Sex differences in estrogen receptor and progestin receptor induction in the guinea pig hypothalamus and preoptic area. Brain Res. 725, 37-48. 73. Blaustein, J. D., and Turcotte, J. C. (1989). Estradiol-induced progestin receptor immunoreactivity is found only in estrogen receptorimmunoreactive cells in guinea pig brain. Neuroendocrinology 49, 454-461. 74. Blaustein, J. D., and Feder, H. H. (1979). Cytoplasmic progestinreceptors in guinea pig brain: Characteristics and relationship to the induction of sexual behavior. Brain Res. 169, 481-497. 75. Parsons, B., MacLusky, N. J., Krey, L., Pfaff, D. W., and McEwen, B. S. (1980). The temporal relationship between estrogen-inducible progestin receptors in the female rat brain and the time course of estrogen activation of mating behavior. Endocrinology (Baltimore) 107, 774-779. 76. Blaustein, J. D., and Olster, D. H. (1989). Gonadal steroid hormone receptors and social behaviors. In "Molecular and Cellular Basis of Social Behavior in Vertebrates" J. Balthazart (eds.), Advances in Comparative and Environmental Physiology, vol. 3, pp. 31-104. Springer-Verlag, Berlin. 77. Kalra, S. P. (1993). Mandatory neuropeptide-steroid signaling for the preovulatory luteinizing hormone-releasing hormone discharge. Endocr. Rev. 14, 507-538. 78. Shughrue, P. J., Lubahn, D. B., Negro-Vilar, A., Korach, K. S., and Merchenthaler, I. (1997). Responses in the brain of estrogen receptor a-disrupted mice. Proc. Natl. Acad. Sci. U.S.A. 94, 11008-11012. 79. Moffatt, C. A., Rissman, E. F., Shupnik, M. A., and Blaustein, J. D.
54
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
92a.
93.
94.
95.
LdPEZ ET AL.
(1998). Induction of progestin receptors by estradiol in the forebrain of estrogen receptor-alpha gene-disrupted mice. J. Neurosci. 18, 9556-9563. Kordon, C., Drouva, S. V., Martinez de la Escalera, G., and Weiner, R. I. (1994). Role of classic and peptide neuromediators in the neuroendocrine regulation of luteinizing hormone and prolactin. In "The Physiology of Reproduction" E. Knobil and J. D. Neill (eds.), pp. 1621-1681. Raven Press, New York. Sagrillo, C. A., Grattan, D. R., McCarthy, M. M., and Selmanoff, M. (1996). Hormonal and neurotransmitter regulation of GnRH gene expression and related reproductive behaviors. Behav. Genet. 26, 241-277. Gore, A. C., and Roberts, J. L. (1997). Regulation of gonadotropinreleasing hormone gene expression in vivo and in vitro. Front, Neuroendocrinol. 18, 209-245. Nelson, S. B., Eraly, S. A., and Mellon, E L. (1998). The GnRH promoter: Target of transcription factors, hormones, and signaling pathways. Mol. Cell. Endocrinol. 140, 151-155. Mellon, E L., Windle, J. J., Goldsmith, E C., Padula, C. A., Roberts, J. L., and Weiner, R. I. (1990). Immortalization of hypothalamic GnRH neurons by genetically targeted tumorigenesis. Neuron 5, 110. Martfnez de la Escalera, G., Choi, A. L., and Weiner, R. I. (1992). Generation and synchronization of gonadotropin-releasing hormone (GnRH) pulses: Intrinsic properties of the GTI-1 GnRH neuronal cell line. Proc. Natl. Acad. Sci. U.S.A. 89, 1852-1855. Wetsel, W. C., Valen~a, M. M., Merchenthaler, I., Liposits, Z., L6pez. F. J., Weiner, R. I., Mellon, P. L., and Negro-Vilar, A. (1992). Intrinsic pulsatile secretory activity of immortalized LHRH-secreting neurons. Proc. Natl. Acad. Sci. U.S.A. 89, 4149-4153. L6pez, F. J., Moretto, M., Merchenthaler, I., and Negro-Vilar, A. (1997). Nitric oxide is involved in the genesis of pulsatile LHRH secretion from immortalized LHRH neurons. J. Neuroendocrinol. 9, 647-654. Poletti, A., Melcangi Cosimo, R., Negri-Cesi, P., Maggi, R., and Martini, L. (1994). Steroid binding and metabolism in the luteinizing hormone-releasing hormone-producing neuronal cell line GTI-1. Endocrinology (Baltimore) 135, 2623-2628. Shen, E., Meade, E. H., P6rez, M. C., Deecher, D. C., Negro-Vilar, A., and L6pez, F. J. (1998). Expression of functional estrogen receptors and galanin messenger ribonucleic acid in immortalized luteinizing hormone-releasing hormone neurons: Estrogenic control of galanin gene expression. Endocrinology (Baltimore) 139, 939-948. Brooks, S. C., Locke, E. R., and Soule, H. D. (1973). Estrogen receptor in a human cell line (MCF-7) from breast carcinoma. J. Biol. Chem. 248, 6251-6253. Lippman, M., Bolan, G., and Huff, K. (1976). The effects of estrogens and antiestrogens on hormone-responsive human breast cancer in long-term tissue. Cancer Res. 36, 4595-4601. Radovick, S., Ticknor, C. M., Nakayama, Y., Notides, A. C., Rahman, A., Weintraub, S. D., Cutler, G. B., and Wondisford, F. E. (1991). Evidence for direct estrogen regulation of the human gonadotropin-releasing hormone gene. J. Clin. Invest. 88, 1649-1655. Skynner, M. J., Sim, T. A., and Herbison, A. E. (1998). Detection of estrogen receptor ce and /3 messenger ribonucleic acids in adult gonadotropin-releasing hormone neurons. Endocrinology 140, 5195-5201. Hotchkiss, J., and Knobil, E. (1994). The menstrual cycle and its neuroendocrine control. In "The Physiology of Reproduction" E. Knobil andJ. D. Neill (eds.), pp. 711-749. Raven Press, New York. Sarkar, D. K., Chiappa, S. A., Fink, G., and Sherwood, N. M. (1976). Gonadotropin-releasing hormone surge in prooestrous rats. Nature (London) 264, 461-463. Levine, J. E., and Ramirez, V. D. (1982). Luteinizing hormone-releasing hormone release during the rat estrous cycle and after ovari-
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
ectomy, as estimated with push-pull cannulae. Endocrinology (Baltimore) 111, 1439-1448. Levine, J. E., Norman, R. L., Gliessman, E M., Oyama, T. T., Bangsberg, D. R., and Spies, H. G. (1985). In vivo gonadotropin-releasing hormone release and serum luteinizing hormone measurements in ovariectomized, estrogen-treated rhesus macaques. Endocrinology (Baltimore) 117, 711-721. Moenter, S. M., Caraty, A., Locatelli, A., and Karsch, E J. (1991). Pattern of gonadotropin-releasing hormone (GnRH) secretion leading up to ovulation in the ewe: Existence of a preovulatory GnRH surge. Endocrinology (Baltimore) 129, 1175-1182. Zoeller, R. T., and Young, W. S., III (1988). Changes in cellular levels of messenger ribonucleic acid encoding gonadotropin-releasing hormone in the anterior hypothalamus of female rats during the estrous cycle. Endocrinology (Baltimore) 123, 1688-1689. Zoeller, R. T., Seeburg, E A., and Young, W. S., III (1988). In situ hybridization histochemistry for messenger ribonucleic acid (mRNA) encoding gonadotropin-releasing hormone (GnRH): Effect of estrogen on cellular levels of GnRH mRNA in female rat brain. Endocrinology (Baltimore) 122, 2570-2577. Park, O.-K., Gugneja, S., and Mayo, K. E. (1990). Gonadotropinreleasing hormone gene expression during the rat estrous cycle: Effects of pentobarbitol and ovarian steroids. Endocrinology (Baltimore) 127, 365-372. Malik, K. E, Silverman, A.-J., and Morrell, J. I. (1991). Gonadotropin-releasing hormone mRNA in the rat: Distribution and neuronal content over the estrous cycle and after castration in males. Anat. Rec. 231,457-466. Marks, D. L., Smith, M. S., Vrontakis, M., Clifton, D. K., and Steiner, R. A. (1993). Regulation of galanin gene expression in gonadotropin-releasing hormone neurons during the estrous cycle of the rat. Endocrinology (Baltimore) 132, 1836-1844. Porkka-Heiskanen, T., Urban, J. H., Turek, F. W., and Levine, J. E. (1994). Gene expression in a subpopulation of luteinizing hormonereleasing hormone (LHRH) neurons prior to the preovulatory gonadotropin surge. J. Neurosci. 14, 5548-5558. Gore, A. C., and Roberts, J. L. (1995). Regulation of gonadotropinreleasing hormone gene expression in the rat during the luteinizing hormone surge. Endocrinology (Baltimore) 136, 889-896. Petersen, S. L., McCrone, S., Keller, M., and Shores, S. (1995). Effects of estrogen and progesterone on luteinizing hormone-releasing hormone messenger ribonucleic acid levels: Consideration of temporal and neuroanatomical variables. Endocrinology (Baltimore) 136, 3604-3610. Suzuki, M., Nishihara, M., and Takahashi, M. (1995). Hypothalamic gonadotropin-releasing hormone gene expression during the rat estrous cycle. Endocr. J. 42, 789-796. Finn, P. D., Steiner, R. A., and Clifton, D. K. (1998). Temporal patterns of gonadotropin-releasing hormone (GnRH), c-fos, and galanin gene expression in GnRH neurons relative to the luteinizing hormone surge in the rat. J. Neurosci. 18, 713-719. Bakker, J., Rubin, B. S., and Baum, M. J. (1999). Changes in mediobasal hypothalamic gonadotropin-releasing hormone messenger ribonucleic acid levels induced by mating or ovariectomy in a reflex ovulator, the ferret. Endocrinology (Baltimore) 140, 595-602. Wang, H.-J., Hoffman, G. E., and Smith, M. S. (1995). Increased GnRH mRNA in the GnRH neurons expressing cFos during the proestrous LH surge. Endocrinology (Baltimore) 136, 36733676. Lee, W.-S., Smith, M. S., and Hoffman, G. E. (1990). Luteinizing hormone-releasing hormone neurons express Fos protein during the proestrous surge of luteinizing. Proc. Natl. Acad. Sci. U.S.A. 87, 5163-5167. Lee, W.-S., Smith, M. S., and Hoffman, G. E. (1990). Progesterone enhances the surge of luteinizing hormone by increasing the activa-
CHAPTER 3 Role of Gonadal Steroids in M e n o p a u s e
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
tion of luteinizing hormone-releasing hormone neurons. Endocrinology (Baltimore) 127, 2604-2606. Lee, W.-S., Abbun, R., Smith, M. S., and Hoffman, G. E. (1992). LHRH neurons express c-Jun protein during the proestrous surge of luteinizing hormone. Endocrinology (Baltimore) 130, 3101-3103. Lee, W.-S., Smith, M. S., and Hoffman, G. E. (1992). c-Fos activity identifies recruitment of luteinizing hormone releasing hormone neurons during the ascending phase of the proestrous luteinizing hormone surge. J. Neuroendocrinol. 4, 161-166. Lee, W.-S., Abbud, R., Hoffman, G. E., and Smith, M. S. (1993). Effects of N-methyl-D-aspartate receptor activation on cFos expression in luteinizing hormone-releasing hormone neurons in female rats. Endocrinology (Baltimore) 133, 2248-2254. Hrabovszky, E., Vrontakis, M. E., and Petersen, S. L. (1995). Triplelabeling method combining immunocytochemistry and in situ hybridization histochemistry: Demonstration of overlap between Fosimmunoreactive and galanin mRNA-expressing subpopulations of luteinizing hormone-releasing hormone neurons in female rats. J. Histochem. Cytochem. 43, 363-370. Hoffman, G. E., Lee, W. S., Attardi, B., Yann, V., and Fitzsimmons, M. D. (1990). Luteinizing hormone-releasing hormone neurons express c-fos antigen after steroid activation. Endocrinology (Baltimore) 126, 1736-1741. Wu, T. J., Segal, A. Z., Miller, G. M., Gibson, M. J., and Silverman, A.-J. (1992). FOS expression in gonadotropin-releasing hormone neurons: Enhancement by steroid treatment and mating. Endocrinology (Baltimore) 131, 2045-2050. Berriman, S. J., Wade, G. N., and Blaustein, J. D. (1992). Expression of Fos-like proteins in gonadotropin-releasing hormone neurons of Syrian hamsters: Effects of estrous cycles and metabolic fuels. Endocrinology (Baltimore) 131, 2222-2228. Doan, A., and Urbanski, H. E (1994). Diurnal expression of Fos in luteinizing hormone-releasing hormone neurons of Syrian hamsters. Biol. Reprod. 50, 301-308. Moenter, S. M., Karsch, E J., and Lehman, M. N. (1993). Fos expression during the estradiol-induced gonadotropin-releasing hormone (GnRH) surge of the ewe: Induction in GnRH and other neurons. Endocrinology (Baltimore) 133, 896-903. Morgan, J. I., and Curran, T. (1991). Stimulus-transcription coupling in the nervous system: Involvement of the inducible proto-oncogenes fos andjun. Annu. Rev. Neurosci. 14, 421-451. Rubin, B. S., Lee, C. E., and King, J. C. (1994). A reduced proportion of luteinizing hormone (LH)-releasing hormone nuerons express Fos protein during the preovulatory or steroid-induced LH surge in middle-aged rats. Biol. Reprod. 51, 1264-1272. Lloyd, J. M., Hoffman, G. E., and Wise, P. M. (1994). Decline in immediate early gene expression in gonadotropin-releasing hormone neurons during proestrus in regularly cycling, middle-aged rats. Endocrinology (Baltimore) 134, 1800-1805. Rubin, B. S., Mitchell, S., Lee, C. E., and King, J. C. (1995). Reconstructions of populations of luteinizing hormone releasing hormone neurons in young and middle-aged rats reveal progressive increases in subgroups expressing Fos protein on proestrus and age-related deficits. Endocrinology (Baltimore) 136, 3823-3830. Witkin, J. W., Xiao, E., Popilskis, S., Ferin, M., and Silverman, A. J. (1994). FOS expression in the gonadotropin-releasing hormone (GnRH) neuron does not increase during the ovarian steroid-induced GnRH surge in the rhesus monkey. Endocrinology (Baltimore) 135, 956-961. Merchenthaler, I., L6pez, E J., and Negro-Vilar, A. (1993). Anatomy and physiology of central galanin-containing pathways. Prog. Neurobiol. 40, 711-769. Coen, C. W., Montagnese, C., and Opacka-Juffry, J. (1990). Coexistence of gonadotrophin-releasing hormone and galanin: Immunohistochemical and functional studies. J. Neuroendocrinol. 2, 107-111.
55 128. Merchenthaler, I., L6pez, E J., and Negro-Vilar, A. (1990). Colocalization of galanin and luteinizing hormone-releasing hormone in a subset of preoptic hypothalamic neurons: Anatomical and functional correlates. Proc. Natl. Acad. Sci. U.S.A. 87, 6326-6330. 129. Marks, D. L., Wiemann, J. N., Burton, K. A., Lent, K. L., Clifton, D. K., and Steiner, R. A. (1992). Simultaneous visualization of two cellular mRNA species in individual neurons by the use of a new double in situ visualization method. Mol. Cell. Neurosci. 3, 3 9 5 405. 130. Kaplan, L. M., Gabriel, S. M., Koenig, J. I., Sunday, M. E., Spindel, E. R., Martin, J. B., and Chin, W. W. (1988). Galanin is an estrogeninducible, secretory product of the rat anterior pituitary. Proc. Natl. Acad. Sci. U.S.A. 85, 7408-7412. 131. Merchenthaler, I., L6pez, E J., Lennard, D. E., and Negro-Vilar, A. (1991). Sexual differences in the distribution of neurons coexpressing galanin and luteinizing hormone-releasing hormone in the rat brain. Endocrinology (Baltimore) 129, 1977-1986. 132. Miller, M. A. (1998). Regulation of galanin in memory pathways. Ann. N. Y. Acad. Sci. 863; 323-341. 133. Liposits, Z., Reid, J. J., Negro-Vilar, A., and Merchenthaler, I. (1995). Sexual dimorphism in copackaging of luteinizing hormonereleasing hormone and galanin into neurosecretory vesicles of hypophysiotropic neurons: Estrogen dependency. Endocrinology (Baltimore) 136, 1987-1992. 134. L6pez, F. J., Merchenthaler, I., Ching, M., Wisniewski, M. G., and Negro-Vilar, A. (1991). Galanin: A hypothalamic hypophysiotropic hormone modulating reproductive functions. Proc. Natl. Acad. Sci. U.S.A. 88, 4508-4512. 135. Merchenthaler, I., Lennard, D. E., L6pez, F. J., and Negro-Vilar, A. (1993). Neonatal imprinting predetermines the sexually dimorphic, estrogen-dependent expression of galanin in luteinizing hormonereleasing hormone neurons. Proc. Natl. Acad. Sci. U.S.A. 90, 1047910483. 136. Marks, D. L., Lent, K. L., Rossmanith, W. G., Clifton, D. K., and Steiner, R. A. (1994). Activation-dependent regulation of galanin gene expression in gonadotropin-releasing hormone neurons in the female rat. Endocrinology (Baltimore) 134, 1991-1998. 137. Rossmanith, W. G., Marks, D. L., Steiner, R. A., and Clifton, D. K. (1996). Inhibition of steroid-induced galanin mRNA expression in GnRH neurons by specific NMDA-receptor blockade. J. Neuroendocrinol. 8, 179-184. 138. Rossmanith, W. G., Marks, D. L., Clifton, D. K., and Steiner, R. A. (1996). Induction of galanin mRNA in GnRH neurons by estradiol and its facilitation by progesterone. J. Neuroendocrinol. 8, 185191. 139. Finn, P. D., McFall, T. B., Clifton, D. K., and Steiner, R. A. (1996). Sexual differentiation of galanin gene expression in gonadotropinreleasing hormone neurons. Endocrinology (Baltimore) 137, 47674772. 140. Anouar, Y., MacArthur, L., Cohen, J., Iacangelo, A. L., and Eiden, L. E. (1994). Identification of a TPA-responsive element mediating preferential transactivation of the galanin gene promoter in chromaffin cells. J. Biol. Chem. 269, 6823-6931. 141. Tatemoto, K., R6kaeus, A., J6rwall, H., McDonald, T. J., and Mutt, V. (1983). Galanin--A novel biologically active peptide from porcine intestine. FEBS Lett. 164, 124-128. 142. L6pez, F. J., and Negro-Vilar, A. (1990). Galanin stimulates luteinizing hormone-releasing hormone secretion from arcuate nucleusmedian eminence fragments in vitro: Involvement of an alpha-adrenergic mechanism. Endocrinology (Baltimore) 127, 2431-2436. 143. Sahu, A., Xu, B., and Kalra, S. P. (1994). Role of galanin in stimulation of pituitary luteinizing hormone secretion as revealed by a specific receptor antagonist, galantide. Endocrinology (Baltimore) 134, 529-536. 144. Sahu, A., Crowley, W. R., Tatemoto, K., Balasubramanian, A., and
56
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.
158.
159. 160.
L6PEZ ET AL.
Kalra, S. E (1987). Effects of neuropeptide Y, NPY analog (norleucine4-NPY), galanin and neuropeptide K on LH release in ovariectomized (OVX) and OVX estrogens, progesterone-treated rats. Peptides (N. Y.) 8, 291-296. L6pez, F. J., Meade, E. H. J., and Negro-Vilar, A. (1993). Endogenous galanin modulates the gonadotropin and prolactin preovulatory surge in the rat. Endocrinology (Baltimore) 132, 795-800. Terasawa, E., Krook, C., Hei, D. L., Gearing, M., Schultz, N. J., and Davis, G. A. (1988). Norepinephrine is a possible neurotransmitter stimulating pulsatile release of luteinizing hormone-releasing hormone in the rhesus monkey. Endocrinology (Baltimore) 123, 18081816. Gearing, M., and Terasawa, E. (1991). The alpha-1-adrenergic neuronal system is involved in the pulsatile release of luteinizing hormone-releasing hormone in the ovariectomized female rhesus monkey. Neuroendocrinology 53, 373-381. Gearing, M., and Terasawa, E. (1991). Prostaglandin E2 mediates the stimulatory effect of methoxamine on in vivo luteinizing hormonereleasing hormone (LH-RH) release in the ovariectomized female rhesus monkey. Brain Res. 560, 276-281. Wise, E M., Rance, N., and Barraclough, C. A. (1981). Effects of estradiol and progesterone on catecholamine turnover rates in discrete hypothalamic regions in ovariectomized rats. Endocrinology (Baltimore) 108, 2186-2193. Demling, J., Fuchs, E., Baumert, M., and Wuttke, W. (1985). Preoptic catecholamine, GABA, and glutamate release in ovariectomized and ovariectomized estrogen-primed rats utilizing a push-pull cannula technique. Neuroendocrinology 41, 212-218. Rance, N., Wise, E M., Selmanoff, M. K., and Barraclough, C. A. (1981 ). Catecholamine turnover rates in discrete hypothalamic areas and associated changes in median eminence luteinizing hormonereleasing hormone and serum gonadotropins on proestrus and diestrous day 1. Endocrinology (Baltimore) 108, 1795-1802. Clifton, D. K., and Sawyer, C. H. (1980). Positive and negative feedback effects of ovarian steroids on luteinizing hormone release in ovariectomized rats following chronic depletion of hypothalamic norepinephrine. Endocrinology (Baltimore) 106, 1099-1102. Nicholson, G., Greeley, G., Humm, J., Youngblood, W., and Kizer, J. S. (1978). Lack of effect of noradrenergic denervation of the hypothalamus and medial preoptic area on the feedback regulation of gonadotropin secretion and the estrous cycle of the rat. Endocrinology (Baltimore) 103, 559-566. Kalra, S. P. (1974). Role of estrogen in the restoration of LH release following stimulation of partially deafferented hypothalamus in rat. Brain Res. 68, 297-307. Coen, C. W., and Coombs, M. C. (1983). Effects of manipulating catecholamines on the incidence of the preovulatory surge of of luteinizing hormone and ovulation in the rat: Evidence for a necessary involvement of hypothalamic adrenaline in the normal or 'midnight' surge. Neuroscience 10, 187-206. Kalra, S. P. (1985). Catecholamine involvement in preovulatory LH release: Reassessment of the role of epinephrine. Neuroendocrinology 40, 139-144. Krey, L. C., Butler, W. R., and Knobil, E. (1975). Surgical disconnection of the medial basal hypothalamus and pituitary function in the rhesus monkey. I. Gonadotropin secretion. Endocrinology (Baltimore) 96, 1073-1087. Heritage, A. S., Grant, L. D., and Stumpf, W. E. (1977). 3H estradiol in catecholamine neurons of rat brain stem: Combined localization by autoradiography and formaldehyde-induced fluorescence. J. Comp. Neurol. 176, 607-630. Dahlstrom, A., and Fuxe, K. (1964). Localization of monoamines in the lower brain stem. Experientia 20, 398-399. Jennes, L., Beckman, W. C., and Stumpf, W. E. (1982). Anatomical relationships of serotoninergic and noradrenalinergic projections
with the GnRH system in septum and hypothalamus. Exp. Brain Res. 46, 331-338. 161. Martfnez de la Escalera, G., Choi, A. L. H., and Weiner, R. I. (1992). /31-Adrenergic regulation of the GT1 gonadotropin-releasing hormone (GnRH) neuronal cell lines: Stimulation of GnRH release via receptors positively coupled to adenylate cyclase. Endocrinology (Baltimore) 131, 1397-1402. 162. Ferin, M., Van Vugt, D., and Wardlaw, S. (1984). The hypothalamic control of the menstrual cycle and the role of endogenous opioid peptides. Recent Prog. Horm. Res. 40, 441-485. 163. Jirikowski, G. F., Merchenthaler, I., Rieger, G. E., and Stumpf, W. E. (1986). Estradiol target sites immunoreactive for beta-endorphin in the arcuate nucleus of rat and mouse hypothalamus. Neurosci. Lett. 65, 121-126. 164. Simerly, R. B., Young, B. J., and Carr, A. M. (1996). Co-expression of steroid hormone receptors in opioid peptide-containing neurons correlates with patterns of gene expression during the estrous cycle. Brain Res. Mol. Brain Res. 40, 275-284. 165. Wilcox, J. N., and Roberts, J. L. (1985). Estrogen decreases rat hypothalamic proopiomelanocortin messenger ribonucleic acid levels. Endocrinology (Baltimore) 117, 2392-2396. 166. Petersen, S. L., Keller, M. L., Carder, S. A., and McCrone, S. (1993). Differential effects of estrogen and progesterone on levels of POMC mRNA levels in the arcuate nucleus: Relationship to the timing of LH surge release. J. Neuroendocrinol. 5, 643-648. 167. Treiser, S. L., and Wardlaw, S. L. (1992). Estradiol regulation of proopiomelanocortin gene expression and peptide content in the hypothalamus. Neuroendocrinology 55, 167-173. 168. Barden, N., Merand, Y., Rouleau, D., Garon, M., and Dupont, A. (1980). Changes in the beta-endorphin content of discrete hypothalamic nuclei during the estrous cycle of the rat. Brain Res. 204, 441445. 169. Knuth, U. A., Sikand, G. S., Casanueva, E F., Havlicek, V., and Friesen, H. G. (1983). Changes in beta-endorphin content in discrete areas of the hypothalamus throughout proestrus and diestrus of the rat. Life Sci. 33, 1443-1450. 170. Leranth, C., MacLusky, N. J., Shanabrough, M., and Naftolin, E (1988). Immunohistochemical evidence for synaptic connections between pro-opiomelanocortin-immunoreactive axons and LH-RH neurons in the preoptic area of the rat. Brain Res. 449, 167-176. 171. Chen, W. P., Witkin, J. W., and Silverman, A. J. (1989). beta-Endorphin and gonadotropin-releasing hormone synaptic input to gonadotropin-releasing hormone neurosecretory cells in the male rat. J. Comp. Neurol. 286, 85-95. 172. Chen, W. P., Witkin, J. W., and Silverman, A. J. (1990). Sexual dimorphism in the synaptic input to gonadotropin releasing hormone neurons. Endocrinology (Baltimore) 126, 695-702. 173. Thind, K. K., and Goldsmith, P. C. (1988). Infundibular gonadotropin-releasing hormone neurons are inhibited by direct opioid and autoregulatory synapses in juvenile monkeys. Neuroendocrinology 47, 203-216. 174. Sannella, M. I., and Petersen, S. L. (1997). Dual label in situ hybridization studies provide evidence that luteinizing hormone-releasing hormone neurons do not synthesize messenger ribonucleic acid for mu, kappa, or delta opiate receptors. Endocrinology (Baltimore) 138, 1667-1672. 175. Nazian, S. J., Landon, C. S., Muffly, K. E., and Cameron, D. E (1994). Opioid inhibition of adrenergic and dopaminergic but not serotonergic stimulation of luteinizing hormone releasing hormone release from immortalized hypothalamic neurons. Mol. Cell. Neurosci. 5, 642-648. 176. Maggi, R., Pimpinelli, E, Martini, L., and Piva, F. (1995). Characterization of functional opioid delta receptors in a luteinizing hormonereleasing hormone-producing neuronal cell line. Endocrinology (Baltimore) 136, 289-295.
CHAPTER 3 Role of G o n a d a l Steroids in M e n o p a u s e 177. Maggi, R., Pimpinelli, F., Martini, L., and Piva, F. (1995). Inhibition of luteinizing hormone-releasing hormone secretion by delta-opioid agonists in GTI-1 neuronal cells. Endocrinology (Baltimore) 136, 5177-5181. 178. Brownstein, M. J., Mroz, E. A., Kizer, J. S., Palkovits, M., and Leeman, S. E. (1976). Regional distribution of substance P in the brain of the rat. Brain Res. 116, 299-305. 179. Ljungdahl, A., H6kfelt, T., and Nilsson, G. (1978). Distribution of substance P-like immunoreactivity in the central nervous system of the rat. I. Cell bodies and nerve terminals. Neuroscience 3, 861-943. 180. Arisawa, M., De Palatis, L., Ho, R., Snyder, G. D., Yu, W. H., Pan, G., and McCann, S. M. (1990). Stimulatory role of substance P on gonadotropin release in ovariectomized rats. Neuroendocrinology 51,523-529. 181. Tsuruo, Y., Kawano, H., Hisano, S., Kagotani, Y., Daikoku, S., Zhang, T., and Yanaihara, N. (1991). Substance P-containing neurons innervating LHRH-containing neurons in the septo-preoptic area of rats. Neuroendocrinology 53, 236-245. 182. Frankfurt, M., Siegel, R. A., Sim, I., and Wuttke, W. (1986). Estrous cycle variations in cholecystokinin and substance P concentrations in discrete areas of the rat brain. Neuroendocrinology 42, 226-231. 183. Micevych, P. E., Matt, D. W., and Go, V. L. (1988). Concentrations of cholecystokinin, substance P, and bombesin in discrete regions of male and female rat brain: Sex differences and estrogen effects. Exp. Neurol. 100, 416-425. 184. Jarry, H., Perschl, A., Meissner, H., and Wuttke, W. (1988). In vivo release rates of substance P in the preoptic/anterior hypothalamic area of ovariectomized and ovariectomized estrogen-primed rats: Correlation with luteinizing hormone and prolactin levels. Neurosci. Lett. 88, 189-194. 185. Malsbury, C. W., and McKay, K. (1989). Sex difference in the substance P-immunoreactive innervation of the medial nucleus of the amygdala. Brain Res. Bull. 23, 561-567. 186. Priest, C. A., Vink, K. L., and Micevych, P. E. (1995). Temporal regulation by estrogen of beta-preprotachykinin mRNA expression in the rat ventromedial nucleus of the hypothalamus. Brain Res. Mol. Brain Res. 28, 61-71. 187. Eckersell, C. B., and Micevych, P. E. (1997). Opiate receptors modulate estrogen-induced cholecystokinin and tachykinin but not enkephalin messenger RNA levels in the limbic system and hypothalamus. Neuroscience, 80, 473-485. 188. Akesson, T. R., and Micevych, P. E. (1988). Estrogen concentration by substance P-immunoreactive neurons in the medial basal hypothalamus of the female rat. J. Neurosci. Res. 19, 4 1 2 - 4 1 9 , 4 7 0 - 4 1 1 . 189. Block, E. (1952). Quantitative morphological investigations of the follicular system in women. Acta Anat. 14, 108-123. 190. Richardson, S. J., Senikas, V., and Nelson, J. F. (1987). Follicular depletion during the menopausal transition: Evidence for accelerated loss and ultimate exhaustion. J. Clin. Endocrinol. Metab. 65, 12311237. 191. Richardson, S. J. (1993). The biological basis of the menopause. Baillieres Clin. Endocrinol. Metab. 7, 1-16. 192. Gougeon, A., Echochard, R., and Thalabard, J. C. (1994). Age-related changes of the population of human ovarian follicles: Increase in the disappearance rate of non-growing and early-growing follicles in aging women. Biol. Reprod. 50, 653-663. 193. Prior, J. C. (1998). Perimenopause: The complex endocrinology of the menopausal transition. Endocr. Rev. 19, 397-428. 194. Schwartz, D., and Mayaux, M. J. (1982). Female fecundity as a function of age: Results of artificial insemination in 2193 nulliparous women with azoospermic husbands. N. Engl. J. Med. 306, 404-406. 195. Menken, J., Trussell, J., and Larsen, U. (1986). Age and infertility. Science 233, 1389-1394. 195a. Block, E. (1953). A quantitative morphological investigation of the follicular system in women. Acta Anat.17, 108-123.
57 195b. Gougeon, A. (1984). Caract~re qualitative et quantatif de la population folliculaire dans l'ovaire humain adulte. Contraception Fertil. Sex. 12, 527-535. 196. Lenton, E. A., Sexton, L., Lee, S., and Cooke, I. D. (1988). Progressive changes in LH and FSH and LH:FSH ratio in women throughout reproductive life. Maturitas 10, 35-43. 197. Lee, S. J., Lenton, E. A., Sexton, L., and Cooke, I. D. (1988). The effect of age on the cyclical patterns of plasma LH, FSH, oestradiol and progesterone in women with regular menstrual cycles. Hum. Reprod. 3, 851-855. 198. Sherman, B. M., and Korenman, S. G. (1975). Hormonal characteristics of the human menstrual cycle throughout reproductive life. J. Clin. Invest. 55, 699-706. 199. Sherman, B. M., West, J. H., and Korenman, S. G. (1976). The menopausal transition: Analysis of LH, FSH, estradiol, and progesterone concentrations during menstrual cycles of older women. J. Clin. Endocrinol. Metab. 42, 629-636. 200. Chakravarti, S., Collins, W. P., Forecast, J. D., Newton, J. R., Oram, D. H., and Studd, J. W. (1976). Hormonal profiles after the menopause. Br. Med. J. 2, 784-787. 201. Reyes, F. I., Winter, J. S., and Faiman, C. (1977). Pituitary-ovarian relationships preceding the menopause. I. A cross-sectional study of serum follicle-stimulating hormone, luteinizing hormone, prolactin, estradiol, and progesterone levels. Am. J. Obstet. Gynecol. 129, 557564. 202. MacNaughton, J., Banah, M., McCloud, P., Hee, J., and Burger, H. (1992). Age related changes in follicle stimulating hormone, luteinizing hormone, oestradiol and immunoreactive inhibin in women of reproductive age. Clin. Endocrinol. (Oxford) 36, 339-345. 203. Vale, W., Rivier, C., Hsueh, A., Campen, C., Meunier, H., Bicsak, T., Vaughan, J., Corrigan, A., Bardin, W., Sawchenko, P., Petraglia, F., Yu, J., Plotsky, P., Spiess, J., and Rivier, J. (1988). Chemical and biological characterization of the inhibin family of protein hormones. Recent Prog. Horm. Res. 44, 1-34. 204. de Kretser, D. M., and Robertson, D. M. (1989). The isolation and physiology of inhibin and related proteins. Biol. Reprod. 40, 33-47. 205. Welt, C. K., McNicholl, D. J., Taylor, A. E., and Hall, J. E. (1999). Female reproductive aging is marked by decreased secretion of dimeric inhibin. J. Clin. Endocrinol. Metab. 84, 105-111. 206. Rossmanith, W. G., Clifton, D. K., and Steiner, R. A. (1996). Galanin gene expression in hypothalamic GnRH-containing neurons of the rat: A model for autocrine regulation. Horm. Metab. Res. 28, 257266. 207. Baird, D. T., and Guevara, A. (1969). Concentration of unconjugated estrone and estradiol in peripheral plasma in nonpregnant women throughout the menstrual cycle, castrate and postmenopausal women and in men. J. Clin. Endocrinol. Metab. 29, 149-156. 208. Rance, N. E., McMullen, N. T., Smialek, J. E., Price, D. L., and Young, W. S., III (1990). Postmenopausal hypertrophy of neurons expressing the estrogen receptor gene in the human hypothalamus. J. Clin. Endocrinol. Metab. 71, 79-85. 209. Rance, N. E., Uswandi, S. V., and McMullen, N. T. (1993). Neuronal hypertrophy in the hypothalamus of older men. Neurobiol. Aging 14, 337-342. 210. Rance, N. E., and Young, W. S., III (1991). Hypertrophy and increased gene expression of neurons containing neurokinin-B and substance-P messenger ribonucleic acids in the hypothalami of postmenopausal women. Endocrinology (Baltimore) 128, 2239-2247. 211. Abel, T. W., Voytko, M. L., and Rance, N. E. (1999). The effects of hormone replacement therapy on hypothalamic neuropeptide gene expression in a primate model of menopause. J. Clin. Endocrinol. Metab. 84, 2111-2118. 212. Vijayan, E., and McCann, S. M. (1979). In vivo and in vitro effects of substance P and neurotensin on gonadotropin and prolactin release. Endocrinology (Baltimore) 105, 64-68.
58 213. Hoffman, G. E. (1985). Organization of LHRH cells: Differential apposition of neurotensin, substance P and catecholamine axons. Peptides (N. Y.) 6, 439-461. 214. Dees, W. L., Skelley, C. W., and Kozlowski, G. E (1985). Central effects of an antagonist and an antiserum to substance P on serum gonadotropin and prolactin secretion. Life Sci. 37, 1627-1631. 215. Ohtsuka, S., Miyake, A., Nishizaki, T., Tasaka, K., Aono, T., and Tanizawa, O. (1987). Substance P stimulates gonadotropin-releasing hormone release from rat hypothalamus in vitro with involvement of oestrogen. Acta Endocrinol. (Copenhagen) 115, 247-252. 216. Thompson, B., Hart, S. A., and Durno, D. (1973). Menopausal age and symptomatology in general practice. J. Biosoc. Sci. 5, 71-82. 217. McKinlay, S. M., and Jeffreys, M. (1974). The menopausal syndrome. Br. J. Prev. Soc. Med. 28, 108-115. 218. Molnar, G. W. (1975). Body temperatures during menopausal hot flashes. J. Appl. Physiol. 38, 499-503. 219. Sturdee, D. W., Wilson, K. A., Pipili, E., and Crocker, A. D. (1978). Physiological aspects of menopausal hot flush. Br. Med. J. 2, 79-80. 220. Meldrum, D. R., Shamonki, I. M., Frumar, A. M., Tataryn, I. V., Chang, R. J., and Judd, H. L. (1979). Elevations in skin temperature of the finger as an objective index of postmenopausal hot flashes: Standardization of the technique. Am. J. Obstet. Gynecol. 135, 713-717. 221. Tataryn, I. V., Lomax, E, Bajorek, J. G., Chesarek, W., Meldrum, D. R., and Judd, H. L. (1980). Postmenopausal hot flushes: A disorder of thermoregulation. Maturitas 2, 101-107. 222. Tataryn, I. V., Lomax, E, Meldrum, D. R., Bajorek, J. G., Chesarek, W., and Judd, H. L. (1981). Objective techniques for the assessment of postmenopausal hot flashes. Obstet. Gynecol. 57, 340-344. 223. Coope, J., Thomson, J. M., and Poller, L. (1975). Effects of "natural oestrogen" replacement therapy on menopausal symptoms and blood clotting. Br. Med. J. 4, 139-143. 224. Campbell, S., and Whitehead, M. (1977). Oestrogen therapy and the menopausal syndrome. Clin. Obstet. Gynecol. 4, 31-47. 256. Aksel, S., Schomberg, D. W., Tyrey, L., and Hammond, C. B. (1976). Vasomotor symptoms, serum estrogens, and gonadotropin levels in surgical menopause. Am. J. Obstet. Gynecol. 126, 165-169. 226. Hutton, J. D., Jacobs, H. S., Murray, M. A., and James, V. H. (1978). Relation between plasma oestrone and oestradiol and climacteric symptoms. Lancet 1, 678-681. 227. Abe, T., Furuhashi, N., Yamaya, Y., Wada, Y., Hoshiai, A., and Suzuki, M. (1977). Correlation between climacteric symptoms and serum levels of estradiol, progesterone, follicle-stimulating hormone, and luteinizing hormone. Am. J. Obstet. Gynecol. 129, 65-67. 228. Chakravarti, S., Collins, W. E, Newton, J. R., Oram, D. H., and Studd, J. W. (1977). Endocrine changes and symptomatology after oophorectomy in premenopausal women. Br. J. Obstet. Gynaecol. 84, 769-775. 229. Casper, R. E, Yen, S. S., and Wilkes, M. M. (1979). Menopausal flushes: A neuroendocrine link with pulsatile luteninizing hormone secretion. Science 205, 823-825. 230. Tataryn, I. V., Meldrum, D. R., Lu, K. H., Frumar, A. M., and Judd, H. L. (1979). LH, FSH and skin temperaure during the menopausal hot flash. J. Clin. Endocrinol. Metab. 49, 152-154. 231. Mulley, G., Mitchell, J. R., and Tattersall, R. B. (1977). Hot flushes after hypophysectomy. Br. Med. J. 2, 1062. 232. Meldrum, D. R., Erlik, Y., Lu, J. K., and Judd, H. L. (1981). Objectively recorded hot flushes in patients with pituitary insufficiency. J. Clin. Endocrinol. Metab. 52, 684-688. 233. Casper, R. E, and Yen, S. S. (1981). Menopausal flushes: Effect of pituitary gonadotropin desensitization by a potent luteinizing hormone-releasing factor agonist. J. Clin. Endocrinol. Metab. 53, 1056-1058. 234. DeFazio, J., Meldrum, D. R., Laufer, L., Vale, W., Rivier, J., Lu, J. K.,
L6PEZ ET AL.
235.
236.
237.
238.
239. 240.
241.
242.
243.
244.
245.
246.
247.
248.
249.
250.
251.
252. 253.
and Judd, H. L. (1983). Induction of hot flashes in premenopausal women treated with a long-acting GnRH agonist. J. Clin. Endocrinol. Metab. 56, 445-448. Menon, V., Edwards, R. L., Lynch, S. S., and Butt, W. R. (1983). Luteinizing hormone releasing hormone analogue in treatment of hypergonadotrophic amenorrhoea. Br. J. Obstet. Gynaecol. 90, 539542. Gambone, J. C., Pardridge, W. M., Lagasse, L. D., and Judd, H. L. (1982). In vivo availability of circulating estradiol in postmenopausal women with and without endometrial cancer. Obstet. Gynecol. 59, 416 -421. Yen, S. S., Rebar, R., VandenBerg, G., and Judd, H. (1973). Hypothalamic amenorrhea and hypogonadotropinism: Responses to synthetic LRE J. Clin. Endocrinol. Metab. 36, 811-816. Quigley, M. E., Sheehan, K. L., Casper, R. E, and Yen, S. S. (1980). Evidence for increased dopaminergic and opioid activity in patients with hypothalamic hypogonadotropic amenorrhea. J. Clin. Endocrinol. Metab. 50, 949-954. Sar, M., and Stumpf, W. E. (1981). Central noradrenergic neurones concentrate 3H-oestradiol. Nature (London) 289, 500-502. Moore, R. Y., and Bloom, E E. (1979). Central catecholamine neuron systems: Anatomy and physiology of the norepinephrine and epinephrine systems. Annu. Rev. Neurosci. 2, 113-168. Van Tienhoven, A., Scott, N. R., and Hillman, E E. (1979). The hypothalamus and thermoregulation: A review. Poult. Sci. 58, 16331639. Clayden, J. R., Bell, J. W., and Pollard, P. (1974). Menopausal flushing: Double-blind trial of a non-hormonal medication. Br. Med. J. 1, 409-412. Edington, R, F,, Chagnon, J. E, and Steinberg, W. M. (1980). Clonidine (Dixarit) for menopausal flushing. Can. Med. Assoc. J. 123, 2 3 26. Laufer, L. R., Erlik, Y., Meldrum, D. R., and Judd, H. L. (1982). Effect of clonidine on hot flashes in postmenopausal women. Obstet. Gynecol. 60, 583-586. Nagamani, M., Kelver, M. E., and Smith, E. R. (1987). Treatment of menopausal hot flashes with transdermal administration of clonidine. Am. J. Obstet. Gynecol. 156, 561-565. Ginsburg, J., O'Reilly, B., and Swinhoe, J. (1985). Effect of oral clonidine on human cardiovascular responsiveness: A possible explanation of the therapeutic action of the drug in menopausal flushing and migraine. Br. J. Obstet. Gynaecol. 92, 1169-1175. Walsh, B. W., and Schiff, I. (1996). The hot flash: Impact of sex steroid replacement. In "Reproductive Endocrinology, Surgery, and Technology" E. Y. Adashi, J. A. Rock, and Z. Rosenwaks, (eds.). pp. 1841 - 1850. Lippincott-Raven, Philadelphia and New York. Beattie, C. W., Rodgers, C. H., and Soyka, L. F. (1972). Influence of ovariectomy and ovarian steroids on hypothalamic tyrosine hydroxylase activity in the rat. Endocrinology (Baltimore) 91,276-279. Luine, V. N., and McEwen, B. S. (1977). Effect of oestradiol on turnover of type A monoamine oxidase in brain. J. Neurochem. 28, 1221-1227. Nixon, R. L., Janowsky, D. S., and Davis, J. M. (1974). Effects of progesterone, beta-estradiol, and testosterone on the uptake and metabolism of 3H-norepinephrine, 3H-dopamine and 3H-serotonin in rat brain synaptosomes. Res. Commun. Chem. Pathol. Pharmacol. 7, 233-236. Donoso, A. O., Stefano, F. J., Biscardi, A. M., and Cukier, J. (1967). Effects of castration on hypothalamic catecholamines. Am. J. Physiol. 212, 737-739. Anton-Tay, F., and Wurtman, R. J. (1968). Norepinephrine: Turnover in rat brains after gonadectomy. Science 159, 1245. Bapna, J., Neff, N. H., and Costa, E. (1971). A method for studying norepinephrine and serotonin metabolism in small regions of rat
CHAPTER 3 Role of G o n a d a l Steroids in M e n o p a u s e
254.
255.
256.
257.
258.
259.
260.
261. 262. 263.
264.
265.
266.
267.
268.
269.
270.
brain: Effect of ovariectomy on amine metabolism in anterior and posterior hypothalamus. Endocrinology (Baltimore) 89, 13451349. Fuxe, K., Lofstrom, A., Eneroth, P., Gustafsson, J. A., Skett, P., Hokfelt, T., Wiesel, E A., and Agnati, L. (1977). Involvement of central catecholamines in the feedback actions of 17beta-estradiolbenzoate on luteinizing hormone secretion in the ovariectomized female rat. Psychoneuroendocrinology 2, 203-225. Myers, R. D., and Yaksh, T. L. (1969). Control of body temperature in the unanaesthetized monkey by cholinergic and aminergic systems in the hypothalamus. J. Physiol. (London) 202, 483-500. Toivola, P., and Gale, C. C. (1970). Effect on temperature ofbiogenic amine infusion into hypothalamus of baboon. Neuroendocrinology 6, 210-219. Dupont, A., Barden, N., Cusan, L., Merand, Y., Labrie, E, and Vaudry, H. (1980). beta-Endorphin and met-enkephalins: Their distribution, modulation by estrogens and haloperidol, and role in neuroendocrine control. Fed. Proc. Fed. Am. Soc. Exp. Biol. 39, 2 5 4 4 2550. Akesson, T. R., and Micevych, P. E. (1991). Endogenous opioid-immunoreactive neurons of the ventromedial hypothalamic nucleus concentrate estrogen in male and female rats. J. Neurosci. Res. 28, 359-366. Clark, W. G., and Ponder, S. W. (1980). Thermoregulatory effects of (D-ala2)-methionine-enkephalinamide in the cat. Evidence for multiple naloxone-sensitive opioid receptors. Brain Res. Bull. 5, 4 1 5 420. Francesconi, R., and Mager, M. (1981). Thermoregulatory effects of centrally administered bombesin, bradykinin, and methionineenkephalin. Brain Res. Bull. 7, 63-68. Clayden, J. R. (1972). Effect of clonidine on menopausal flushing. Lancet 2, 1361. Levine, J. D., Gordon, N. C., and Fields, H. L. (1978). The mechanism of placebo analgesia. Lancet 2, 6 5 4 - 657. Merchenthaler, I., Funkhouser, J. M., Carver, J. M., Lundeen, S. G., Ghosh, K., and Winneker, R. C. (1998). The effect of estrogens and antiestrogens in a rat model for hot flush. Maturitas 30, 307-316. Abel, T. W., and Rance, N. E. (1999). Proopiomelanocortin gene expression is decreased in the infundibular nucleus of postmenopausal women. Brain Res. Mol. Brain Res. 69, 202-208. Schaffalitzky De Muckadell, O. B., Aggestrup, S., and Stentoft, P. (1986). Flushing and plasma substance P concentration during infusion of synthetic substance P in normal man. Scand. J. Gastroenterol. 21, 498-502. Newby, D. E., Sciberras, D. G., Mendel, C. M., Gertz, B. J., Boon, N. A., and Webb, D. J. (1997). Intra-arterial substance P mediated vasodilatation in the human forearm: Pharmacology, reproducibility and tolerability. Br. J. Clin. Pharmacol. 43, 493-499. Norheim, I., Theodorsson-Norheim, E., Brodin, E., and Oberg, K. (1986). Tachykinins in carcinoid tumors: Their use as a tumor marker and possible role in the carcinoid flush. J. Clin. Endocrinol. Metab. 63, 605-612. Conlon, J. M., Deacon, C. F., Richter, G., Stockmann, F., and Creutzfeldt, W. (1987). Circulating tachykinins (substance P, neurokinin A, neuropeptide K) and the carcinoid flush. Scand. J. Gastroenterol. 22, 97-105. Oberg, K., Norheim, I., Theodorsson, E., Ahlman, H., Lundqvist, G., and Wide, L. (1989). The effects of octreotide on basal and stimulated hormone levels in patients with carcinoid syndrome. J. Clin. Endocrinol. Metab. 68, 796-800. Balks, H. J., Conlon, J. M., Creutzfeldt, W., and Stockmann, F. (1989). Effect of a long-acting somatostatin analogue (octreotide) on circulating tachykinins and the pentagastrin-induced carcinoid flush. Eur. J. Clin. Pharmacol. 36, 133-137.
59 271. Vinik, A. I., Gonin, J., England, B. G., Jackson, T., McLeod, M. K., and Cho, K. (1990). Plasma substance-P in neuroendocrine tumors and idiopathic flushing: The value of pentagastrin stimulation tests and the effects of somatostatin analog. J. Clin. Endocrinol. Metab. 70, 1702-1709. 272. Saslow, S. B., O'Brien, M. D., Camilleri, M., vonder Ohe, M., Homburger, H. A., Klee, G. G., Pitot, H. C., and Rubin, J. (1997). Octreotide inhibition of flushing and colonic motor dysfunction in carcinoid syndrome. Am. J. Gastroenterol. 92, 2250-2256. 273. Hackman, B. W., and Galbraith, D. (1976). Replacement therapy and piperazine oestrone sulphate ('Harmogen') and its effect on memory. Curr. Med. Res. Opin. 4, 303-306. 274. Vanhulle, G., and Demol, P. (1976). A double-blind study into the influence of estriol on a number of psychological tests in postmenopausal women. In "Consensus on Menopause Research" (P. A. van Keep, R. B. Greenblat, and M. Albeaux-Fernet (eds.), pp. 9 4 99. University Park Press, Baltimore, MD. 275. Fedor-Freybergh, P. (1977). The influence of oestrogens on the wellbeing and mental performance in climacteric and postmenopausal women. Acta Obstet. Gynecol. Scand. Suppl. 64, 1-91. 276. Sherwin, B. B. (1988). Estrogen and/or androgen replacement therapy and cognitive functioning in surgically menopausal women. Psychoneuroendocrinology 13, 345-357. 277. Sherwin, B. B., and Phillips, S. (1990). Estrogen and cognitive functioning in surgically menopausal women. Ann. N. Y Acad. Sci. 592, 474 -475. 278. Phillips, S. M., and Sherwin, B. B. (1992). Effects of estrogen on memory function in surgically menopausal women. Psychoneuroendocrinology 17, 485-495. 279. Ditkoff, E. C., Crary, W. G., Cristo, M., and Lobo, R. A. (1991). Estrogen improves psychological function in asymptomatic postmenopausal women. Obstet. Gynecol. 78, 991-995. 280. Barrett-Connor, E., and Kritz-Silverstein, D. (1993). Estrogen replacement therapy and cognitive function in older women. JAMA, J. Am. Med. Assoc. 269, 2637-2641. 281. Kampen, D. L., and Sherwin, B. B. (1994). Estrogen use and verbal memory in healthy postmenopausal women. Obstet. Gynecol. 83, 979-983. 282. Shughrue, P. J., Lane, M. V., Scrimo, P. J., and Merchenthaler, I. (1998). Comparative distribution of estrogen receptor-alpha (ERalpha) and beta (ER-beta) mRNA in the rat pituitary, gonad, and reproductive tract. Steroids 63, 498-504. 283. Gould, E., Woolley, C. S., Frankfurt, M., and McEwen, B. S. (1990). Gonadal steroids regulate dendritic spine density in hippocampal pyramidal cells in adulthood. J. Neurosci. 10, 1286-1291. 284. Woolley, C. S., Gould, E., Frankfurt, M., and McEwen, B. S. (1990). Naturally occurring fluctuation in dendritic spine density on adult hippocampal pyramidal neurons. J. Neurosci. 10, 4035-4039. 285. Woolley, C. S. (1998). Estrogen-mediated structural and functional synaptic plasticity in the female rat hippocampus. Horm. Behav. 34, 140-148. 286. Palinkas, L. A., and Barrett-Connor, E. (1992). Estrogen use and depressive symptoms in postmenopausal women. Obstet. Gynecol. 80, 30-36. 287. Sherwin, B. B., and Gelfand, M. M. (1985). Sex steroids and affect in the surgical menopause: A double-blind, cross-over study. Psychoneuroendocrinology 10, 325-335. 288. Schneider, M. A., Brotherton, P. L., and Hailes, J. (1977). The effect of exogenous oestrogens on depression in menopausal women. Med. J. Aust. 2, 162-163. 289. Klaiber, E. L., Broverman, D. M., Vogel, W., and Kobayashi, Y. (1979). Estrogen therapy for severe persistent depressions in women. Arch. Gen. Psychiatry 36, 550-554. 290. Owens, M. J., and Nemeroff, C. B. (1994). Role of serotonin in the
60 pathophysiology of depression: Focus on the serotonin transporter. Clin. Chem. (Winston-Salem, N.C.) 40, 288-295. 291. Pecins-Thompson, M., Brown, N. A., and Bethea, C. L. (1998). Regulation of serotonin re-uptake transporter mRNA expression by ovarian steroids in rhesus macaques. Brain Res. Mol. Brain Res. 53, 120-129. 292. Pecins-Thompson, M., Brown, N. A., Kohama, S. G., and Bethea, C. L. (1996). Ovarian steroid regulation of tryptophan hydroxylase mRNA expression in rhesus macaques. J. Neurosci. 16, 7021-7029.
L6PEZ ET AL. 293. Pecins-Thompson, M., and Bethea, C. L. (1999). Ovarian steroid regulation of serotonin-1A autoreceptor messenger RNA expression in the dorsal raphe of rhesus macaques. Neuroscience 89, 267-277. 294. Alves, S. E., Weiland, N. G., Hayashi, S., and McEwen, B. S. (1998). Immunocytochemical localization of nuclear estrogen receptors and progestin receptors within the rat dorsal raphe nucleus. J. Comp. Neurol. 391, 322-334. 295. Bethea, C. L. (1994). Regulation of progestin receptors in raphe neurons of steroid-treated monkeys. Neuroendocrinology 60, 50-61.
7HAPTER /
Gonadotropins and Menopause: New Markers S T E V E N B I R K E N 1 Department of Medicine, Columbia University College of Physicians and Surgeons, New York, New York 10032
JOHN O' CONNOR
Department of Pathology and Irving Center for Clinical Research, Columbia University College of Physicians and Surgeons, New York, New York 10032
G A L I N A KOVALEVSKAYA Irving Center for Clinical Research, Columbia University College of Physicians and Surgeons, New York, New York 10032 LESLIE LOBEL
I. II. III. IV.
Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, New York, New York 10032
V. Changes in Gonadotropin Patterns in Menopause VI. Gonadotropin Fragments as Urinary Analytes References
Introduction Gonadotropin Parameters of the Normal Menstrual Cycle Structures of Gonadotropins and Their Fragments Assay of Gonadotropins
I. I N T R O D U C T I O N
from the corpus luteum; and chorionic gonadotropin (hCG), the hormone of pregnancy, which rescues the corpus luteum and continues steroid support for the developing fetus. The fourth glycoprotein hormone, thyroid-stimulating hormone (hTSH), stimulates production of thyroxine to control bodily metabolism [1]. All of the glycoprotein hormones are produced by the pituitary except for hCG, which is primarily a product of trophoblast cells in the placenta. A slight quantity of hCG appears to be of pituitary origin and is released along with hLH in nonpregnant individuals [2-6]. All four human glycoprotein hormones are thought to have evolved from a single ancestral gene [7,8]. Although the a subunits are identical among all four glycoproteins in terms of pri-
The gonadotropins are heterodimeric glycoprotein hormones composed of a common ce subunit and hormone-specific/3 subunits. There are four members of this family of hormones in humans (h), three of which are directly involved in reproduction: follicle stimulating hormone (hFSH), which promotes maturation of the follicle and ovulation; luteinizing hormone (hLH), which provides the final stimulus for ovulation and stimulates steroid secretion
1To whom correspondence should be addressed.
MENOPAUSE: BIOLOGY AND PATHOBIOLOGY
61
Copyright9 2000 by AcademicPress. All rights of reproductionin any form reserved.
62 mary structure, there are some differences in glycosylation patterns, especially in the addition of sulfate to hLH in place of some of the terminal sialic acid sugar residues [9,10]. hCG and hLH have nearly the same fi subunit structures and are the most closely related of the glycoprotein hormones in both structure and function [1,8,11]. hCG differs from hLH mainly in its extension at its carbohydrate-rich fi COOH-terminal region which endows the hormone with a much longer circulating half-life than that of hLH, namely, 24 versus 1 hr [11,12].
BIRKEN ET AL.
Ovulation Panel 1
LH
-20
Panel 2
mIU/mgCr
-40
FSH
II. G O N A D O T R O P I N P A R A M E T E R S OF THE N O R M A L M E N S T R U A L C Y C L E The menstrual cycle is characterized by a distinctive pattern of secretion of two gonadotropins, hLH and hFSH. Both hormones are secreted in a pulsatile manner under control of gonadotropin-releasing hormone (GnRH) [ 13-16]. The pulsatile pattern of circulating gonadotropins is essential for proper gonadotropin activity at its target receptors, i.e., the ovary in females and the testis in males. Approximately hourly gonadotropin pulsatility is found both in young normal cycling women and in women of later reproductive age, although the rate of pulsation may alter somewhat during aging [13-19]. As illustrated by the classical Fig. 1 of the menstrual cycle, the gonadotropins regulate the concentrations of estradiol and progesterone, which are the effector molecules of the reproductive system. An intricate feedback system of the circulating steroids and peptides controls gonadotropin secretion [20]. hFSH concentrations are most critically controlled because their precise regulation ensures development of only one dominant follicle [22,22]. hFSH appears to be controlled by several negative and positive feedback systems in addition to circulating steroids, including inhibins and activins [23,24]. Higher gonadotropin concentrations characteristic of reproductive aging in women are discussed later in this chapter. The temporal profile of the pituitary gonadotropins undergoes a substantial change as a woman proceeds through her reproductive years; with increasing ovarian senescence, a consequent diminution of sex steroid production disrupts the monthly cyclicity and the menopausal period ensues. The changes in gonadotropin secretion concomitant with menopause are best understood under terms of the gradual alterations in the normal menstrual cycle as a woman approaches the menopausal transition. The normal menstrual cycle is defined to commence with the first day of menstrual bleeding and in the human is of 2 4 - 3 2 days duration. Several cytokines, including epidermal growth factor (EGF), insulin-like growth factor-I (IGF-I), transforming growth factor-fi (TGF-fi), activin, and inhibin are involved in follicular growth and maturation [25]. Control of the circulating gonadotropins hLH and hFSH is primarily under the regulation of the hypothalamic GnRH
-40
Follicular ~ Luteal i[~ Phase ._l _. Phase ._
-20
Panel 3
mIU/mgCr
-100
Estrone Conjugates
ng/mgCr -50
-25
Panel 4
PDG
~g/mgCr
I
-15 -10
I
-5
0
5
10
15
Days FIGURE 1 Urinary hormone profiles throughout the normal menstrual cycle. Panel 1: In the follicular phase, the levels of (preovulatory) human luteinizing hormone (hLH) remain relatively constant, with sharply higher hLH values occurring around midcycle LH surge (ovulation defined as day 0). During the luteal phase, hLH decreases to low levels until the late luteal phase, when it again begins to rise and a new cycle begins. Panel 2: Human follicle-stimulating hormone (hFSH) levels are elevated in the follicular phase, reflecting the primary role this hormone plays in the maturation of the dominant follicle, hFSH induces its own receptors and by doing so enables the cellular machinery to produce all of the protein factors necessary for cellular growth and function. At ovulation, there is a secondary spike of hFSH, coincident with hLH; afterward, hFSH levels decline until late in the luteal phase; when the diminishing levels of estradiol and progesterone derepress hFSH production and levels again rise at the beginning of a new cycle. Panel 3: This panel profiles estrone glucuronide and sulfate, the principal excretory metabolites ofestradiol and estrone. In the circulation, there is a rise ofestradiol coincident with the periovulatory gonadotropin surge and a secondary rise in the luteal phase under the influence of low levels of gonadotropins. Panel 4: The profile of pregnandiol-3-glucuronide (PDG), the major urinary metabolite of circulating progesterone, the hormone synthesized by the corpus luteum and peaking in the luteal phase. Progesterone exerts local effects within the corpus luteum and promotes the maturation of the endometrium. The corpus luteum is "rescued" by hCG in a conceptive cycle; otherwise it collapses, forming the corpus albicans prior to initiation of a new cycle.
and under feedback control from the inhibins and steroids produced by the dominant follicle [25]. GnRH modulates the release of both hLH and hFSH from the pituitary [26]. The menstrual cycle is segmented into three phases (see Fig. 1): the follicular phase, during which a follicle (the dominant follicle) destined to ovulate forms; the ovulatory phase, during which the dominant follicle ruptures and re-
63
CHAPTER4 Gonadotropins and Menopause: New Markers leases a mature ovum; and the luteal phase, whereby the ruptured follicle undergoes a process known as luteinization, in which ovarian thecal and granulosa cells enlarge and produce increased quantities of progesterone. During the preovulatory or follicular phase of the menstrual cycle (10-16 days), under the influence of increased hFSH from the pituitary, ovarian tissue surrounding the ovum destined to ovulate (i.e., the dominant follicle) undergoes proliferation and a cavity is formed within the follicle. The events involved in the mechanism of recruitment of one follicle (the dominant follicle) for ovulation are complex, involving the action of hLH and other growth factors. Estradiol production within the follicle is increased. Coincidentally, the tissues of the uterus, under the influence of increased estradiol, undergo proliferation and thickening. Ovulation commences on days 12-15 of the cycle. The egg, with its surrounding zona pellucida, breaks through the wall of the ovary and escapes into the abdominal cavity near the oviduct [27]. The ovulatory event is preceded by an increased release of hLH from the pituitary, which drives the granulosa cells of the dominant follicle to produce progesterone in addition to estradiol. These cells increase in size and lipid content, produce a yellow pigment (lutein), and form the corpus luteum subsequent to the expulsion of the ovum. The corpus luteum is vascularized in response to increased production of angiogenic factors elaborated by the granulosa and theca cells of the ovary. Immediately following ovulation, the uterine endometrial cells undergo a transformation in preparation for the implantation of the fertilized ovum (blastocyst). The luteal phase of the cycle (13-14 days) is characterized by increased production of progesterone, which in turn results in changes in uterine endometrial cell morphology. The endometrial glandular cells form secretory vacuoles, which contain glycogen. By days 2 0 - 2 1 of the cycle, secretion from these cells peaks and the uterine environment is optimized for implantation [27]. Progesterone production by the corpus luteum peaks coincident with maximum vascularization. Unless the corpus luteum is maintained by the hCG produced in a conceptive cycle, the corpus luteum undergoes structural and functional degeneration (luteolysis). Although both estrogen and prostaglandins have been invoked in the mechanism of luteolysis, the exact process remains unresolved. The pattern of urinary hormone metabolite secretion by the corpus luteum is illustrated in Fig. 1. Estrogen concentrations fall subsequent to ovulation, with a secondary rise occurring during the middle of the luteal phase, and again falling toward the end of the cycle. Progesterone and hydroxyprogesterone concentrations rise in parallel with estrogen, whereas inhibin A, which remains relatively depressed and static during the follicular phase, exhibits a surge in the luteal phase with a concomitant decrease in circulating hFSH. As the concentrations of estrogen and progesterone decrease in the late luteal phase, hFSH again begins to increase, initiating a new cycle.
III. STRUCTURES OF GONADOTROPINS AND THEIR FRAGMENTS Although the primary structures of the gonadotropins were determined between 1970 and 1975, more than 20 years elapsed until any three-dimensional structure became known, due to problems with crystallization of hormones with high carbohydrate content. The primary structures of the ce and fl subunits of hCG, hLH, and hFSH are shown in Fig. 2B. The ce subunit (Fig. 2A) is common to all of the glycoprotein hormones, whereas the fl subunit differs in each. The fl subunits of hLH and hCG are very similar. The structure of the hLH/3 subunit in Fig. 2B depicts its form
1
10
Ala-Pro-Asp-Val-Gln-Asp-Cys-Pro-G1u-Cys-Thr-Leu-Gln-Glu-Asn-Pro-Phe20 30 Phe-Ser-Gln-Pro-Gly-Ala-Pro-Ile-Leu-Glx-eys-Met-Gly-Cys-eys-Phe-Ser40
50
Arg-Ala-Tyr-Pro-Thr-Pro-Leu-Arg-Ser-Lys-Lys-Thr-Met-Leu-Val-Gln-LysCHO
60
Asn-Val-Thr-Ser-Glu-Ser-Thr-Cys-Cys-Val-Ala-Lys-Ser-Tyr-Asn-Arg-Val70
CHO
80
Thr-Val-Met-Gly-Gly-Phe-Lys-Val-Glu-Asn-His-Thr-Ala-Cys-His-Cys-Ser90
Thr-Cys-Tyr-Tyr-His-Lys-Ser
1
CHO
10
Asn-Ser-Cys-Glu-Leu-Thr-Asn- I le-Thr- I le-Ala- I le-Glu-Lys-Glu-Glu-eys20
CHO
30
Arg-Phe-Cys- Ile- Ser- Ile-Asn-Thr-Thr-Trp-Cys-Ala-Gly-Tyr-Cys-Tyr-Thr40
50
Arg-Asp-Leu-Val-Tyr-Lys-Asp- Pro-Ala-Arg- Pro-Lys- I le-Gln-Lys- Thr-Cys60
Thr-Phe-Lys-Glu-Leu-Val-Tyr-Glu-Thr-Val-Arg-Val-Pro-Gly-Cys-Ala-His70
80
His-Ala-Asp- Ser-Leu-Tyr- Thr-Tyr-Pro-Val-Ala-Thr-Gln-Cys-His-Cys-Gly90
100
Lys-Cys-Asp-Ser-Asp- Ser-Thr-Asp-Cys-Thr-Val-Arg-Gly-Leu-Gly-Pro-SerII0 Tyr-Cys-Ser-Phe-Gly-Glu-Met-Lys-Glu
FIGURE 2 (A) Primary structure of the ce subunit, common to all of the glycoprotein hormones. (B) Primary structure of the fi subunit of hLH as encoded by its mRNA. The lightly shaded area indicates the amino acids present in the hLHficf; the darkly shaded area indicates additional amino acids present in some of the core molecules. The brackets, both solid and dotted, indicate amino acids encoded by hLHfl mRNAsthat are usually not found in circulating or excreted hLH, but only within the pituitary tissue. (C) The fi subunit of hFSH. N-linked carbohydrate is present on residues 7 and 24 in FSH fi and residue 30 in hLH fl subunits. CHO indicates glycanmoiety.
64
BIRKEN ET AL.
in the circulation as well as its original biosynthetic form within the pituitary. Even when hLH is extracted from pituitary tissue, most of its very hydrophobic, leucine-rich seven-residue fl COOH-terminal region is missing (as indicated in brackets in Fig. 2B). This hydrophobic structure is involved in internal transport of hLH during its synthesis within the pituitary and is thought to be cleaved off prior to release into the circulation and certainly does not exist at all in urinary hLH [28]. A partially deglycosylated form of hCG was finally crystallized and its structure solved in 1994 [29-31]. The structures of both subunits strongly resembled each other (mainly fl sheet with a small a-helical region in ce; see Fig.3) [29,31 ], a finding that could not be deduced from the diverse primary amino acid sequences of each subunit. Both subunits consisted of a large loop and two smaller loops, with the fl subunit having an unusual seatbelt capture portion that held alpha in its embrace (Fig. 3). The two subunits are held together by noncovalent forces, chiefly hydrogen bonding and hydrophobic interactions; further stabilization is achieved by the region of the fl subunit starting at about residue 92, which folds around the ce loop 2 region. The seatbelt loop is held fast by an intradisulfide bridge between residues 26 and 110 [29,31]. This unexpected disulfide bridge surprisingly does not inhibit dissociation of subunits in acid buffers or chaotropic reagents. The manner in which dissociation takes place in relation to the seatbelt interaction is not yet known, but it is assumed that there is either enough structural alteration to widen the seatbelt to allow a to slip out, or that the
FIGURE 3
26-110 disulfide bridge exists in an equilibrium between closed (oxidized) and open (reduced) states. Solution of the gonadotropin structure also pointed out the homology of the gonadotropins to disulfide knot growth factors such as TGF-fl. Recombinant studies support the evolution of heterodimeric gonadotropins from an ancient homodimeric molecule, based on a recombinant construction made from a combination of segments from ce and fl subunits that resulted in a homodimeric bioactive molecule [7]. Although the three-dimensional structures of hLH, hFSH, and hTSH have not been individually solved, it is likely that they all resemble hCG structurally, and each has been modeled based on the hCG structure [29,31 ]. Carbohydrate plays an important role in the bioactivity of glycoprotein hormones. Some carbohydrate groups are important for signal transduction at the receptor [32] whereas other carbohydrate groups are important for long-term survival in the circulation [ 12]. Baenziger and colleagues made important observations concerning sulfation of some carbohydrate groups in hLH. They found that sulfated glycoproteins bind to a sulfate/sugar receptor in the liver and contribute to the pulsatility of hLH, which is necessary for its biological activity [9,33-37]. The pulses of hLH are accentuated not only by pulse release from the pituitary but also by rapid withdrawal of a portion of the circulating hLH by the liver receptor, which increases the downslope of the pulse. Human FSH is only slightly sulfated unlike the other FSH isoforms of the hormone in other mammals [35-37]. Figure 4 illustrates the nature of the carbohydrate moi-
Three-dimensional structures of hLH isoforms. The noncovalently bound heterodimer structure shown for hLH (a) is based on the solved structure for hCG, which is presumed to be very similar. The "seatbelt" region of the fl subunit, which wraps around a is indicated. The a subunit appears as a string in this representation for convenience, although the structures of both a and fl are mostly fl sheet except for a small a-helical structure in the a subunit, which is indicated on the illustration. In vivo, it is presumed that hLH dissociates into free hLHfl (b) and then is degraded into the hLHfl core fragment (c) in a body compartment such as the kidney. The hLHfl core fragment is a stable molecule that is rapidly cleared and excreted into the urine. The structures of hLHfl and hLHfl core fragment are likely different from their structures within the hLH molecule, but for convenience are shown here as the same.
CHAPTER4 Gonadotropins and Menopause: New Markers
FIGURE 4 Carbohydrate chains attached to asparagines of hLH and hFSH represented schematicallybased on the studies of Baenziger [35]. Carbohydrate formsexhibitmicroheterogeneityas indicated. A, Sialicacid; e, galactose; m, N-acetylgalactosamine; O, sulfate; 5, N-acetylglucosamine; O, mannose.
eties in hLH and hFSH, as well as some of the microheterogeneity that results in the very wide range of isoelectric points in the glycoprotein hormones [35-37]. Human LH is partially sulfated whereas hFSH is much less so (Fig. 4). Sialic acid is the most common terminal sugar for the gonadotropins in general and is responsible for their acid isoelectric points. The sulfate groups that replace sialic acid groups on half of the hLH molecules maintain the same acid nature of hLH, giving it a negative charge at physiological pH. Figure 4 has been greatly simplified to indicate the major N-linked carbohydrate moieties present on gonadotropins. There are literally dozens of variations, including molecules lacking both sialic acid and sulfate circulating in the bloodstream [9,35-37]. The various structures and their relative proportions in humans and other mammals have been described in detail by Baenziger [35-37].
IV. A S S A Y O F G O N A D O T R O P I N S A. B i o a s s a y This chapter focuses on measurement of gonadotropins and gonadotropin-derived polypeptides. In this section we provide the background information necessary to understand the complexity involved in measuring these diverse molecules.
65 The measurement of gonadotropins has proceeded from intact animal bioassays, in which administration of the hormone produces a measurable change in an organ weight or an increase in production of a hormone by that organ. A primary example of this technique is the mouse uterine weight assay for hLH (or hCG), because both of these hormones bind to the same receptor. The end point of the assay is an increase in mouse uterine weight after administration of either gonadotropin [38]. In general, intact animal bioassays are not usually performed due to the high cost, substantial interanimal variability, and lack of assurance that the animal model response will be identical to the human response. Despite these shortcomings, intact animal bioassays do have the advantage of taking into consideration both the intrinsic activity of the hormone at the receptor and, additionally, the circulating half-life of the hormone. This latter a component of biological activity is not evaluated in either cell or tissue preparations derived from the test animal or, more recently, in stable cell lines that have been genetically engineered to express the human gonadotropin receptor [39-41 ]. Most bioassays in current use rely on cell preparations from intact animals. Biologically active hFSH is measured by employing the in vitro immature rat Sertoli cell aromatase bioassay, i.e., measuring the amount of estrogen produced under cellular stimulation by hFSH [42]. Similarly, the biological activity hCG or hLH is routinely determined using a cultured rat Leydig cell preparation, the end point being an increase in cellular testosterone production under the influence of hLH [43]. In some instances, the measurement of the increased production of the initial intracellular product of receptor activation, cyclic AMP, rather than the hormonal endproduct, is employed in bioassays. A recent development in bioassay technology is development of cell lines that have been genetically engineered to express the functional human gonadotropin receptor, i.e., either the LH/CG receptor or the FSH receptor (44). This procedure allows for a more precise estimation of biological activity of human hormones than does use of the rodent receptor.
B. R a d i o r e c e p t o r A s s a y An alternative to the bioassay is the radioreceptor assay, in which intact cells or solubilized cell membranes from hormone target tissue containing the appropriate gonadotropin receptor are employed as the binding protein [45,46]. It should be noted that hormonal binding to receptor does not per se signify biological activity. It is well recognized that there exist isoforms of the gonadotropins, presumably with altered carbohydrate structure, with diminished or no biological activity despite their sometimes high affinity for the receptor [47].
66
BIRKEN ET AL.
Unfortunately, technical difficulty in performing these assays, variation in response of different systems/preparations, and difficulty in obtaining a pure reference preparation have limited use of these assays to research, with only modest clinical applications.
C. Immunoassay Attempts to simplify and improve both precision and accuracy of hormone measurements have resulted in the development of hormone-specific binding assays. Early applications of this technique utilized a naturally occurring hormone-binding protein, e.g., thyroxine binding globulin for T4 assay [48]. However, the majority of binding assays utilize specific antibodies developed against the gonadotropins. The first immunologically based assays employed polyclonal antibodies raised in animals. These immunoassays, which generally recognized multiple epitopes on the gonadotropin molecule, had the ability to detect many of the circulating isoforms, subunits, and fragments of the gonadotropin indiscriminately. A further assay refinement came with the development of monoclonal antibody technology. Monoclonal antibodies afford a continuous supply of a binding reagent of defined epitope specificity. This has led to the development of the immunometric assay, the most common formulation of which involves two monoclonal antibodies, each directed to a different epitope on the gonadotropin molecule. One of these antibodies (the capture antibody) is immobilized on a solid support and serves to extract the analyte from the matrix (usually blood or urine). The second antibody (the detection antibody) is labeled, e.g., with ~25I or an enzyme, and binds to a second epitope on the analyte molecule. The assay response is directly proportional to the quantity of analyte present in the specimen. The practical consequences of this development in the gonadotropin field have been an increase in the sensitivity of the measurements, but more importantly, the ability to specifically and independently measure intact hormone isoforms, free subunits, and metabolic fragments, especially in the urine. It has been demonstrated for hCG that urinary measurements can provide more complete and, in many instances, more clinically useful information than can serum determinations. [For example, hCG fl core fragment (hCGflcf), found in measurable quantities only in urine, is a much more effective marker of nontrophoblastic malignancy and some pregnancy disorders than is serum hCG] [49-52]. The core fragments are discussed in detail in Sections IV,D and VI. It should be noted that highly specific assays sometimes confer an important disadvantage compared to the less specific polyclonal-based older assay systems. Monoclonal antibody-based systems are directed to one epitope for each an-
tibody in the system and some such assays tend to detect poorly, or not at all, some of the isoforms of the gonadotropins. Thus, studies have documented cases in which the biological/immunological activity ratio changes throughout pregnancy [43,53] and with pregnancy disorders (hCG) [54], and throughout the human menstrual cycle (hLH or hFSH) [55] and with reproductive aging [56]. In extreme cases, investigators have documented the presence of either only biologically active or immunologically active gonadotropin in a subject. As will be described later, the way to avoid this difficulty may be to focus assays on gonadotropin fragments, which have much less epitope diversity than the holohormone or its subunits and which can be measured more reliably. The substantial variation that exists in the results obtained from both bioassays and immunoassays at different research centers highlights the need to identify stable gonadotropin markers of well-defined structure. The urinary fl core fragments of the parent gonadotropin molecules have the potential to fulfill this need.
D. fl Core Fragments of Gonadotropins as Stable Urinary Analytes Although the structures of the gonadotropins are very important to their biological functions, their structures are equally important to their immunochemical measurement (see Sections IV, A-IV, C). Such measurements can be performed in both blood and urine, although gonadotropinderived additional molecular species not detectable in blood appear in the urine. These molecules are derived from proteolysis of the fl subunits of gonadotropins. Currently, only fl core fragments of hLH and hCG have been reported to appear along with their parent heterodimeric and whole subunit parents in urine [57-60]. Although hFSH fragments should prove of greatest interest for menopausal studies, the fl core fragments of hCG and hLH are first described, because they have been isolated and characterized, and measurement systems for them have been developed; the information they provide should also prove relevant to hFSH fragments. The fl core fragments of hCG and hLH are very similar in structure. The fl core structures consist of approximately half of the fl subunit, comprising two disulfide-bridged chains. Figure 5 shows the subtle structural differences in the hCGflcf and hLHflcf (see Fig. 3 for the likely proteolytic origin of such fragments), hCGflcf is found mainly in urine, from which it has been isolated [57,59,61,62]. It is composed of fl residues 6 - 4 0 covalently linked to residues 55-92 by disulfide bridges [57]. hLHflcf has, thus far, been isolated only from pituitary extracts. Based on immunochemical studies using antibodies developed to pituitary hLHflcf, an
CHAPTER4 Gonadotropins and Menopause: New Markers 1
67
/
S e t Lys C ! - P r e LeuVArg-Pro-Arg-Cys-Arg-Pro-Ile-Asn-Ala-Thr-Leu-Ala-Val-Glu-Lys-Glu-Gly23 .~u~ + Cys-Pro-Val-Cys-Ile-Thr-Val-Asn-Thr-Thr-Ile-Cys-Ala-Gly-Tyr-Cys-Pro-Thr A .rk.. A..g V.~! 9 .~.. ~ . 67
r . u . ~ . ~ 9 .~_ D.... , ~.~ 9 .~..Dr^~t,
al-Val-Cys-Asn-Tyr-Arg-Asp-Val-Arg-Phe-Glu-Ser
••e•Ar••Leu•Pr••G•y•Cys•Pr••Ar••G•y•Va••Asn•Pr••Va••Va]•Ser•Tyr•A•a•Va••A•a•Leu••er•Cys• 89 Gln-CysAla-Leu~YLys+,A.rg .,rgA ~,ere ,ru,. "ru.. Asp P " ~ r,t,, ~_h, o..,,_9 y~ A.~p u ; - D..,, 9 ,~.. -n,,._r, yo /
1
. A. , ,.~ . . A. ,:,~ . . D,-,,~v ,~,.gA ,. DI.,,.~ f " h , ,
ili
1
ili
A ,,~!. G~,a~- G~,a~. G ~ . .
~.Jl~
Gh,~.- 9 J. . . .
o
~llij
A in
~,.1111
1
9 ~
~
d
l l i O
D,.,,~ D,.,,.~ Dw.~,.G~A~- 9 A . ,
i
9 ~
.ll.~lh~l.dl
I
ili
~
1,~
D,.,,~ G~,a,. D,.,,~ G~,~,.
t-3-3
Arg Le'a Pre Cly Pre Ser A~p Thr Pr" "-~ 9 "-" Pro--C!.~ 1 +A 20 Ser Arg Clu Pre:Leu rg-Pro-Trp-Cys-His-Pro-Ile-Asn-Ala-Ile-Leu-Ala-Val-Glu-Lys-Glu-Gly
23
_~
-Cys-Pro-Val-Cys-Ile-Thr-Val-Asn-Thr-Thr-Ile-Cys-Ala-Gly-Tyr-Cys-Pro-Thr
N e t M.et A r g - V a !
Le'a C ! n Ala xra! eu-Pro-Pro-Leu-Pro-Gln-Val-Val-Cys-Thr-Tyr-Arg-Asp-Val-Arg-Phe-Glu-Ser67 ••e-Arg-Leu-Pro-G•y-Cys-Pr•-Arg-G•y-Val-Asp-Pr•-Va•-Va•-Ser-Phe-Pro-Va•-Ala-Leu-Ser-Cys-
89
~
Arg-Cys-Gly-Pro-Cy A r g A r g e,,.. -rk..+ e,,.. , , p Cys C!y C!y P r e Lys Asp His P r e L e u T ~ r t44 Cy3 A s p H i s P r e G ! n Le'-'- Ser Cly~Leu Let: P~e L e u F I G U R E 5 Comparison of the primary structures of the fl core fragments of hCG (top) and h L H (bottom). Areas absent from the cores are indicated by the cross-out lines. Peptide bond cleavages are indicated by vertical arrows. N-linked carbohydrate moieties are present on residues 13 and 30 of h C G / 3 core (top) and on residue 30 of h L H fl core (bottom).
analogous core form exists in urine [59,60,63]. Although hLHflcf (pituitary form) is very similar in primary structure to hCGficf, the second peptide component was found to be heterogeneous, being either residue 4 9 - 9 3 or 5 5 - 9 3 [58,59,63]. Note that hLHficf includes the Cys at residue 93, giving the core an even number of Cys residues, whereas hCGflcf ends at residue 92 and has an odd number of Cys residues (Fig. 5). Another point must be considered in making these comparisons, hCGficf was isolated from urine and hLHflcf was isolated from a pituitary extract. We already have evidence that the urinary form of hLHflcf is different in structure from the pituitary form on the basis of high-performance liquid chromatography (HPLC) properties [59,60,63]. However, the immunochemical system we have developed recognizes both pituitary and urinary forms of hLHflcf very well, and so both forms can be quantified. We have immunochemical evidence of the existence of a urinary hFSH/3 core fragment (hFSHflcf) (but not a pituitary hFSHflcf), although the structure of hFSHflcf has not yet been determined. All of these cores arise from a proteolytic process within a body tissue compartment and then are released into the urine (Fig. 3). What is remarkable about hCGficf is that it is homogeneous in urine in terms of its two peptide constituents. Frequently, proteolytic processes result in considerable peptide heterogeneity due to incomplete proteolytic bond cleavages. In the case of hCGficf, this proteolytic process proceeds to completion within a body compart-
ment, likely the kidney, and has not been duplicated by protease mixtures in vitro despite much effort [64,65]. The two fl core fragments have very desirable properties in terms of urinary analytes: (1) both molecules are present in urine in substantial concentrations relative to the parent hormone, faciliting their measurement; (2) both represent the end point of a proteolytic process within a tissue compartment, usually the kidney but sometimes the pituitary (hLHficf) or placenta (hCGficf), and are highly stable molecules; and (3) the fl core fragments are potent immunogens and are easily measured by most fl-directed antibodies as well as by specific antibodies generated against the cores [66-68] because they display a unique determinant. In contrast, as described earlier, hLH and hFSH are excreted in a variety of isoforms and as such present a diversity of epitopes, which complicates their measurement. For example, hLH may appear as isoforms undetectable by some immunochemical commercial and research assay systems [63,69,70], although they are still a part of the bioactive pool of hLH. These "invisible" hLH isoforms can complicate hLH measurements in some patients [63,71]. Figure 6 depicts the results of one such analysis of first morning void urine samples, which show an absent hLH periovulatory surge [63]. The heterodimeric gonadotropins are not very stable in urine because they tend to dissociate slowly into free subunits during multiple freeze-thaw cycles or simply during
68
BIRKEN ET AL.
120 0
100-
E m
80-
7O0 A
(~
9
//
hLH-1
-
9
o
~
60-
~
40-
e-
-
500 o
-
400 E
-
200 "6
200-
I
0
I
I
B
25-
I
9
I
0
6
E~-3-Glucuronide e
300 ~E
lOO5 t-
-
_1 e-
600
-SE
20-4
.,= -a
~ "1o
15
- 3
0
~, lO
2
-2~ -ld)
ui
"E
0 0
i
i
5
10
i
1
i
15
20
25
-6
0
day
FIGURE 6 Profiles of urinary hLH molecular forms in a normally ovulating subject who did not express measurable intact hLH in either of two hLH assays (A). Note that both hLHfl and hLHflcf surges are clearly apparent. (B) The corresponding urinary steroid metabolite patterns for the cycle. It can be inferred from the steroid profiles that the subjects experienced normal ovulatory cycles, even in the absence of detectable intact hLH. Concentrations were normalized to creatinine. Day 1 is the first day of menses. Reproduced from [63]; O'Connor, J. E, Kovalevskaya, G., Birken, S., Schlatterer, J. P., Schechter, D., McMahon, D. J., and Canfield, R. E. (1998). The expression of the urinary forms of human luteinizing hormone beta fragment in various populations as assessed by a specific immunoradiometric assay. Hum. Reprod. 13, 826-835, with permission of the authors and
Human Reproduction.
prolonged storage. Core molecules are extremely stable for weeks at room temperature in the presence of a bacteriostat [59,60,63,72]. We have found that hLHflcf is stable for 40 or more freeze-thaw cycles [63]. These core molecules are, therefore, candidates for use as new, stable urinary markers of menopause.
V. CHANGES IN GONADOTROPIN PATTERNS IN MENOPAUSE Both hLH and hFSH concentrations rise significantly in menopausal women, having escaped the negative feedback inhibition of the steroid hormones usually produced by the ovary, hFSH pituitary secretion also rises due to the absence of inhibin B in menopausal women. Klein et al., Welt et al., as well as Santoro and colleagues showed a parallel decline of inhibins, especially inhibin B, with the rise of FSH during the follicular phase of perimenopausal women [24,73,74] (see also Chapter 9). Burger also showed that in early perimenopause, serum hFSH rose rapidly as inhibin B declined, with no significant fall in inhibin A [75]. Later in perimenopause inhibin A fell markedly. The escape of hFSH control due to inadequate luteal phase inhibin regulation may lead to
perimenopausal ovarian hyperstimulation with excessive development of a number of follicles [20] (see Fig. 7B). Inhibin B levels may fall precipitously low before the number of hFSH-responsive follicles decline. This situation may be responsible in part for menopausal symptoms such as unpredictable mood swings, headaches, and fluid retention. Inhibin control of hFSH may be more critical than the negative feedback of high estrogen levels at this stage of the menopausal transition. Burger proposed this scenario for the high FSH levels as women approach menopause: "With increasing age, the gradual decline in follicle numbers in the ovary leads to a fall in inhibin B levels, and a consequent rise in FSH sufficient to maintain ovulatory function and continued secretion of E 2 and inhibin A, mainly by the dominant follicle" [75]. Burger goes on to hypothesize that as follicle numbers decline further, leading to menstrual irregularity, inhibin B falls much further leading to a jump again in hFSH concentration in the circulation and continued dominant follicle development. When no more follicles are available at the time of cessation of menses, ovulatory function stops and inhibin A and E 2 levels fall with the final additional rise in hFSH concentration. Figure 7 summarizes these changes in pictorial format. In a review of reproductive aging in the rodent, Wise et al. hypothesized that the origin of the perimenopausal changes in gonadotropins, both their lowered rate of pulsation and their rise in concentration, is due to changes in the GnRH patterns, such as lower frequency of GnRH pulsation favoring increased release of hFSH over hLH [ 15,16]. Wise et al. attribute the change in GnRH pulsatility to deterioration of pacemakers within the brain, chiefly the suprachiasmatic nuclei that send branches to GnRH-generating neurons. A deterioration of this key neural pacemaker may initiate the gradual disintegration of neurotransmitter rhythms that are critical for precise and regular gonadotropin secretion [15,16]. Elderly postmenopausal women, compared to prematurely menopausal women, do appear to have increased asynchrony of GnRH secretion, although circadian rythmicity appears to be preserved [ 19]. Other studies in support of the neuroendocrine hypothesis include those of Matt et al., which indicated alterations in both interpulse intervals and pulse width of hLH in older cycling women as compared to younger women [76]. Studies of Reame et al. likewise found faster midluteal pulses in perimenopausal women as compared to midreproductive age women [56]. Soules, Battaglia, and Klein recently described their evaluation of the two competing hypotheses regarding reproductive aging [77,78]. Both hypotheses consider a rise in hFSH as contributing to exhaustion of the remaining follicles. The neuroendocrine hypothesis, however, considers slowing of the GnRH pulse generator as the cause of hFSH elevation whereas the ovarian hypothesis holds that it is the lower inhibin B levels that prompt the hypothalamus and pituitary to increase hFSH, which accelerates loss of the remaining follicles.
69
CHAPTER 4 G o n a d o t r o p i n s and M e n o p a u s e : N e w Markers
A
B
Premenopausal W o m e n Reproductive A g e < 35 years
~GnRH
Perimenopausal W o m e n Early
C
Perimenop usal W o m e n Late
~ GnRH~, FSH LH
+ effecf,
F~S~. ;effec~~H ?,. Estrogen RRIDIR V~
9?~r
i.
I
F S ~ GnRH~'~[~LH stro
~r
t~o~er do
left ovary
right ovary.
D
E Postmenopausal W o m e n
t,s. @
Postmenopausal W o m e n Elderly > 75
xt
"//,. LH
,,ro~ +
Postmenopausal W o m e n POF
I I
~r
GnRH X
F
~FSH~,~ "~tLH~V~ ~Estrogen
@
9
@
z
GnRH X-
LH
~(Estrogen
FIGURE 7 Summary illustration of hormonal changes during reproductive aging of women. Key: e, Dominant follicle; 1", level elevated in proportion to length of arrow (bold indicates significant elevation); $, level reduced in proportion to length of arrow (bold indicates significant reduction); arrows that are N-shaped indicate that a level is not always elevated; zig-zag lines indicate oscillating levels and pulsations (width and height symbolize frequency and amplitude, respectively); slanted lines crossed by two parallel bars indicate a disruption of an effect. (A) Premenopause: hFSH under control of GnRH (+ effect, bold arrow signifies strong effect), inhibins A and B from developing follicles ( - effect, dashed bold arrow signifies strong negative effect), and estrogen ( - effect). (B) Early perimenopause: hFSH begins to be variable while inhibin declines (although the decline in inhibin is somewhat questionable). Estrogen concentrations rise and hFSH is variably higher resulting in more follicle development and higher estrogen levels. (C) Late perimenopause: hFSH levels high, inhibin levels lower, and HLH variably higher. A dominant follicle still may develop. (D) Postmenopause: hFSH levels high with pulsatile pattern similar to that in a younger woman but with greater amplitude (note: pulsating hormone levels shown only for hFSH and hLH in postmenopausal women); same for hLH. GnRH patterns variable, inhibin absent, and estrogen very low due to absence of dominant follicle; no follicle development. (E) Postmenopausal elderly: hFSH levels high but pulsation rate slower; same for hLH. GnRH low and quite variable, no inhibin, and very low estrogen. (F) Postmenopause, premature ovarian failure (POF): high hFSH levels but pulsation pattern similar to that of a younger woman, as is hLH. No inhibin. GnRH levels and patterns similar to those of younger women. A portion of this illustration is based on Figure 7 from [20], Prior, J. (1998). Perimenopause: the complex endocrinology of the menopausal transition. Endocrine Reviews 19, 397-428. 9 The Endocrine Society.
Wise et al. support the neuroendocrine hypothesis [15,16]. Soules and colleagues have made a number of observations that are not consistent with either the ovarian hypothesis, in terms of inhibin levels in aging women, or the neuroendocrine hypothesis, in terms of similar GnRH pulsation patterns during aging. Nonetheless, they reach the overall conclusion that the ovarian hypothesis can better predict the changes in the gonadotropin levels and the number of follicles that are a precursor to menopause. Burger and others [79,80] have presented interpretations for the different stages during the transition to menopause. Prior has presented a synthesis of these various proposals
[20]. We have attempted to simplify this summary in Fig. 7. In this illustration, we present six panels illustrating the changing gonadotropin patterns from young premenopausal women to early and late perimenopausal, to postmenopausal, and to elderly postmenopausal women. Also included are young women experiencing premature ovarian failure. In many perimenopausal patients hLH and hFSH ovarian receptors were very low and such receptors were absent in postmenopausal patients [81 ]. This led Vihko to suggest that high serum gonadotropin levels act in concert with low or absent ovarian receptors for these hormones. Santoro demonstrated that there is also an age-related
70
BIRKEN ET AL.
alteration in hypothalamic or pituitary function that acts to decrease what would be even higher hLH and hFSH levels. In studies of women with premature ovarian failure compared to normal age menopausal women, it was apparent that hLH secretion was greater in the younger women along with greater pulse amplitudes, although pulse frequency was similar in both groups [19]. Therefore, the changes in gonadotropin secretion as women age demonstrate a decline in capabilities of the pituitary to produce gonadotropins in response to GnRH and/or a decline in GnRH during aging. These capacity differences in gonadotropin secretion in women with premature ovarian failure as compared to normal age menopausal women are also apparent in the gonadotropin fragment analysis patterns, which will be reviewed later. Gonadotropin bioactivity also increases in menopause along with the total immunoreactive concentration ofgonadotropins [82]. Changes in sialylation (increased sialic acid content) with decreasing circulating steroids contribute to the increase in bioactivity of the gonadotropins after menopause. Studies of biological to immunological gonadotropin ratios show that postmenopausal women display nearly twice the ratio value than do normal young cycling women or older perimenopausal women [82]. Although the high concentrations of hFSH and hLH after menopause are apparent, these vary greatly during the transition to menopause, the perimenopause. Santoro provides an excellent illustration of the changes in gonadotropin and steroid levels among 11 midreproductive age women and 11 perimenopausal women during the menstrual cycle [18] (Fig. 8). The rise in overall hFSH concentrations during the
,50
§ .... ii , ,',
peri I
40
LH
FSH
40 =_
a0
C:~ 30
20
2o
10
I0
E
E
150
PDG
;F i (-~
20
100
E
"~
so
__ 0
I
J
-,s -io
I
-s
=
o
~
DAYS
,~
i~
~-;s
-,o
~--"-
-s
-
~
~
I
1o
I
,~
,
~o
I0
DAYS
FIGURE 8 Daily hLH, hFSH, estrone conjugates (El), and PDG excretions (mean _ SEM) corrected for creatinine and standardized to the day of presumed ovulation (day 0) in 11 regularly menstruating perimenopausal women (open circles) compared with 11 younger women (closed circles). E] was higher in the perimenopausal women (P = 0.023) and integrated PDG was lower (P = 0.015). Reproduced from [18], Santoro, N., Brown, J. R., Adel, T., and Skurnick, J. H. (1996). Characterization of reproductive hormonal dynamics in the perimenopause. J. Clin. Endocrinol. Metab. 81, 1495-1501. 9 The Endocrine Society.
early follicular phase and an overall rise in estrogen in perimenopausal women are apparent in this illustration. Although hFSH increases prior to a rise in hLH levels during perimenopause, a great range of possible values makes such gradual increases of little diagnostic value in determining the perimenopausal stage [20,79,80]. Likewise, estrogen values are quite variable during perimenopause, being even higher in some perimenopausal women than in young, normal cycling women. Despite the use of day-3 serum FSH and/or circulating estrogen concentrations as indicative of the perimenopausal state, these single point measurements are far from conclusive; further exploration should provide new and improved markers that can indicate proximity to menopause.
VI. G O N A D O T R O P I N AS URINARY
FRAGMENTS
ANALYTES
The gonadotropins, being heterodimeric hormones whose subunits are noncovalently bound together, are unstable in urine. Although it is possible to decrease their instability by addition of additives such as glycerol and by minimizing freeze-thaw cycles, intact gonadotropins are not ideal molecules for accurate quantification in urine [83-86]. The gonadotropins undergo proteolytic cleavages during transit through the kidney and their stability in urine is likely to be further reduced by such damage [57,59,64,68,87]. Most commercial gonadotropin assays are certified as quantitatively accurate in blood but only qualitatively accurate in urine due to such problems [88,89]. It is easier to quantify glycoprotein hormones in blood, but each sampling requires a technician or physician or even extremes, such as an indwelling catheter for multiple samplings. Large-scale studies using blood specimens cannot be conducted for this reason, but such studies can readily be performed in urine if an appropriate, highly stable analyte can be identified. The/3 core fragments of the gonadotropins are ideal molecules for urinary measurement. The two gonadotropins of significance to menopause are hFSH and hLH, with the former being the most relevant. Although we have evidence for the presence of a urinary hFSHflcf, only/3 core fragments of hCG and hLH have been isolated, characterized, and measured thus far. The hCGflcf has been used extensively in urinary measurement systems for the past decade. It is extremely stable, as attested to by a variety of groups [72], and is generally present in higher concentrations than heterodimeric hCG in urine. Measurement applications include certain cancers and various problem pregnancies, including Down syndrome. The utility of hCGflcf in the cancer marker field (used only informally for this purpose) is limited to monitoring the recurrence of hCGsecreting tumors after therapy. The higher molar concentration of this fragment as compared to hCG in urine leads to greater sensitivity of detection. The hLHflcf epitope is similarly stable in urine. Current assays for the hLHflcf are based
CHAPTER4 Gonadotropins and Menopause" New Markers on the isoform isolated from pituitary tissue, the urinary form of this molecule has not yet been purified [59,60,63]. hLH is more difficult to measure accurately in urine than is hCG because it is less stable than hCG and frequently displays isoforms not recognized by many monoclonal antibody-based immunochemical measuring systems [63,71,9093] (see Fig. 6). In addition, hLH may appear to be absent from urine after having completely dissociated into subunits or after having been completely metabolized to smaller fragments. FSH also presents measuring problems due to dissociation of subunits in urine. This problem has occurred in analyses of FSH in nonhuman primates, which are used as models for studies of human reproduction. One lab has recently proposed boiling all monkey urine samples to cause complete dissociation of FSH into subunits and then measuring the released fl subunit [94]. However, if any heterodimeric hormone is proteolytically cleaved, such boiling may cause irreversible loss of some epitopes and total fl concentration may not represent total FSH originally present in the specimen. In all of these situations,/3 core fragment measurement provides the most stable and consistent reflection of gonadotropins in circulation. These fragments have reached an end point in proteolysis and are not further degraded in urine as long as microbial growth is inhibited. While acknowledging that FSHflcf would most likely provide the best menopausal marker, because we have already developed an immunochemical system for measurement of hLHflcf, we have applied these measurements first to assess their utility in studies of women in menopausal transition. As previously described, we demonstrated that hLHflcf appeared in high concentrations in women during the menstrual cycle starting on the day of the LH surge, peaking 1-2 days after the surge [63]. Figure 9 shows the hormone profiles of 15 normal cycling premenopausal women and the temporal relationship in the appearance of hLH, hLHfl, hCG, and hLHflc. The X axis is normalized to day 0, this being the zenith of the hLH surge appearance in urine. This delayed appearance of hLHflcf after the hLH surge suggests that circulating hLH is sequestered in a body compartment (likely kidney tissue) and is excreted after 2 4 - 4 8 hr of proteolytic processing, hLHflcf can be easily measured in urine but not in serum. This is shown in Figure 10 and implies that the kidney creates the fragment by absorbing hLH from the blood and excreting it later into the urine. It is also possible that some of the hLHcf is directly secreted by the pituitary into the bloodstream but is cleared so rapidly that its circulating concentration remains very low. Our original hypothesis was that postmenopausal women, who do not experience the large midcycle surge of hLH but instead maintain a continuous pulsatile high concentration of hLH, would tend to display a high plateau level of hLHflcf. Essentially, this would result from an integration of the multitude of hLH pulses in the time delay during the proteolytic processing steps. We tested this hypothesis by examining first morning void urines for 10 consecutive days in a series
71
100 -
hLH
~[
9
hCGI3cf
hLH~l
4 -
80(o r
600
~
S Eo
4020-
32-
1 -
0
I
I
I
0 i i i i 1 6 0 - S t e r o i d h o r m o n e ratio
I
600 o
120O m 400 800
E 200
oo
-
-
40-
LU ,,_..
0-
0
I
I
I
I
80-
I
I
I
I
-2
0
2
4
day ~o 6 0 E J)
E
E
4020-
0
I
I
I
I
-2
0
2
4
day
FIGURE 9 Hormone profiles in the urine of normally cycling women (n = 15). Concentrations were presented as mean _+ SE, femtomoles/milligram creatinine (fmol/mg C). hLH concentration has been measured using two different immunoradiometric assays (n = 8 for the hLH-2 assay). Steroid hormone ratios are calculated using estrone-3-glucuronide (E1-3-G) and pregnandiol-3-glucuronide (Pd-3-G) (• 103). Day 0 is the day of hLH surge. Reproduced from [63], O'Connor, J. F., Kovalevskaya, G., Birken, S., Schlatterer, J. E, Schechter, D., McMahon, D. J., and Canfield, R. E. (1998). The expression of the urinary forms of human luteinizing hormone beta fragment in various populations as assessed by a specific immunoradiometric assay. Hum. Reprod. 13, 826-835, with permission of the authors and Human Reproduction.
of postmenopausal women. The resultant data did not agree with our hypothesis. Instead, very large-amplitude pulses of hLHflcf, which could not readily be correlated with hLH surges, appeared in the urine of these women. One such pattern of hLHflcf excretion in a postmenopausal woman is illustrated in Fig. 11. Note the very large fluctuations of hLHflcf (between 0 and 600 fmol/mg creatinine), which do not correlate with urinary hLH, even considering the 24to 48-hr time delay in the appearance of the core after an hLH surge (Fig. 9). Analysis of 10 consecutive first morning void urine samples from cycling women of reproductive age (< 35 yr) during the follicular phase (day 1 was the first day of menses) indicated that much shallower fluctuations occurred in these women, prior to the hLH surge at the time of ovulation. Indeed, by integrating the area under the peaks of graphs of hLHflcf in femtomoles/milligram of creatinine versus day of collection, it was possible to differentiate young cycling women from postmenopausal women, even with the occasional high spikes in some young women
72
BIRKEN ET AL.
120 Serum lOO -
~0 80o
~
I
60-
_J e-
~ "r" _1 '-
40 20 o 200
~
I
l
I
1200 e
-
1000
150 800
0
-~
0
~100 Z
- 600
:f
-
o -1-
~
400
t-
50 - 200
0
i
i
i
1
0
1
2
3
0
day
FIGURE 10 hLH and hLHflcf in serum and urine of the same patient, o, hLH- 1 ; r~, hLH2; A, hLHfl; e, hLHflcf. The serum levels of intact hLH (o) and hLHflcf (e) indicate that there is an insignificant amount of hLHflcf detected in the blood. The lower panel illustrates the urinary values for hLH and hLHflcf for the same days of collection. The surge of hLH (day 0) and the surge of hLHflcf (1-2 days later) are detected in urine, but the peak of hLHflcf lags that of the intact hLH by 1-2 days, suggesting that urinary hLHflcf is a consequence of the peripheral or renal metabolic processing of intact hLH. Reproduced from [63], O'Connor, J. E, Kovalevskaya, G., Birken, S., Schlatterer, J. P., Schechter, D., McMahon, D. J., and Canfield, R. E. (1998). The expression of the urinary forms of human luteinizing hormone beta fragment in various populations as assessed by a specific immunoradiometric assay. Hum. Reprod. 13, 826-835, with permission of the authors and Human Reproduction.
during the follicular phase. We selected 10-day intervals as a convenient research set, and specimens could be easily collected by volunteer subjects and stored frozen until samples were brought to the laboratory. In a future marker assay, the collection protocol is unlikely to consist of such a large sample set. When samples are analyzed from women still experiencing regular menstrual cycles, the 10-day interval collection provides a convenient starting point within the cycle and encompasses the follicular phase that most closely corresponds to the postmenopausal state of relatively low circulating steroids. For regularly cycling women, the mean of the areas under the peaks of 10 subjects was 278 with a median area of 169. The mean of the areas among postmenopausal subjects ranged between approximately 1000-4000 with medians in the ranges of 900-3000. The postmenopausal subjects differed significantly from the population of normal cycling women by the amplitude and area under the peaks of the daily fluctuations of this fragment. We com-
pared the areas under the peak of the lowest hLHflcf levels of postmenopausal women with the highest core levels of premenopausal women and could statistically differentiate the two groups even in this worst-scenario sampling situation. Perimenopausal women fall in between, with some clearly in the postmenopausal pattern and some in a premenopausal pattern. Figures 12 and 13 illustrate analysis of two perimenopausal women. The patterns of hLHflcf shown in Fig. 12 are typical of those seen in normal cycling women of midreproductive age (see Fig. 9) [63]. A peak hLHflcf appears 1-2 days after the hLH surge, after the presumed metabolic breakdown of circulating hLH into hLHflcf within a tissue compartment. Figure 13 depicts a perimenopausal woman with a typical postmenopausal pattern (see Fig. 11), with many very large hLHflcf peaks, not coordinated to particular hLH peaks in the urine. However, we must keep in mind the earlier discussion of the problems of accurately measuring hLH, in urine and the phenomenon of "invisible" urinary hLH, which we have encountered several times in our own laboratories [63,71 ]. We are in the process of developing assays for a similar hFSHflcf. With the combination of the two fragment assays,
600
500
400 (
o C:~
300
0
E 2oo 0 _J t-
0
10
0
1'0
20
30
40
50
60
70
20
30
4'0
50
60
70
20
~" 0
15
g Z) E "l_J t-
10
5
D a y of U r i n e C o l l e c t i o n
FIGURE 11 The 60-day patterns of hLHflcf and hLH in first morning void urine collections from a postmenopausal woman. In the upper panel the hLHflcf is normalized to creatinine, whereas in the lower panel the hLH is measured by the DELFIA assay on glycerol-preserved urines.
73
CHAPTER 4 Gonadotropins and Menopause: New Markers
800
600
5O0
t
t 400 t_ o 300 E
600
1
i,..
o
E 400 O
O
E 200 tO
~- 100 zIZ
~" 200 -.I r
._1
"
0 10
20
30
40
50
I
14 I 12 -~ 10 E -r
._J
.c
10
20
30
40
50
60
60
T
I
_
0
60
20 18 16
,
8
liI
50
T
JI II li
40
I..
oO') 30 E D 20 E -110 ....I
6 4
t-
O
0
10
20
30
40
50
60
Day of Urine Collection FIGURE 12 The 60-day patterns of hLH/3cf and hLH in first morning void urine collections from a perimenopausal woman. In the upper panel the hLH/3cf is normalized to creatinine, whereas in the lower panel the hLH is measured by the DELFIA assay on glycerol-preserved urines. The pattern of hLH/3cf resembles that of midreproductive age women (see Fig. 9).
hLH/3cf and hFSH/3cf, it may be possible to develop urinary assays that will define the stage of m e n o p a u s a l transition. Because changes in h F S H concentrations usually precede those of hLH, it is envisioned that the hFSH/3cf assay, when developed, m a y provide a stronger discriminant function than the hLH/3cf assay. In summary, the gonadotropin patterns in b l o o d and urine undergo significant alterations during the m e n o p a u s a l transition. The d e v e l o p m e n t of sensitive assays for stable proteolytically derived fragments of h L H and h F S H in urine will provide new markers to determine the phases of the menopausal transition. The standardization of these assays with stable h o r m o n a l metabolites should obviate the p r o b l e m s often encountered in c o m p a r i s o n and analysis of data collected from different laboratories and thereby ease collection of data in large-scale epidemiological studies. Furthermore, implementation of assays based on stable h o r m o n a l metabolites will heighten awareness in the general clinical c h e m i s t r y c o m m u n i t y o f the utility of metabolic by-products as markers for other physiologically relevant proteins.
0
10
20
30
40
50
60
70
Day of Urine.Collection FIGURE 13 The 60-day patterns of hLHflcf and hLH in first morning void urine collections from a perimenopausal woman. In the upper panel the hLI-I/3cfis normalized to creatinine, whereas in the lower panel the hLH is measured by the DELFIA assay on glycerol-preserved urines. The pattern of hLHflcf resembles that of a postmenopausal woman. Part of this figure is reproduced from Birken, S., Santoro, I. V., Maydelman, Y., Kovalevskaya, G., Lobo, R., Freeman, E. W., Warren, M., McMahon, D., and O'Connor, J. (1999). Differences in urinary excretion patterns of the hLH beta core fragment in premenopausal, perimenopausal, and postmenopausal women. Menopause 6(4), with permission of the authors and Lippincott Williams & Wilkins.
Acknowledgments This work was supported by NIH grants RO1-AG 13783 and ROlES07589. We wish to express appreciation to Nanette Santoro for thoughtful advice.
References 1. Pierce, J. G., and Parsons, T. E (1981). Glycoprotein hormones: Structure and function. Annu. Rev. Biochem. 50, 465-495. 2. Birken, S., Maydelman, Y., Gawinowicz, M. A., Pound, A., Liu, Y., and Hartree, A. S. (1996). Isolation and characterization of human pituitary chorionic gonadotropin. Endocrinology (Baltimore) 137, 14021411.
74 3. Sawitzke, A. L., Griffin, J., and Odell, W. D. (1991). Purified preparations of human luteinizing hormone are contaminated with small amounts of a chorionic gonadotropin-like material. J. Clin. Endocrinol. Metab. 72, 841-846. 4. Hammond, E., Griffin, J., and Odell, W. D. (1991). A chorionic gonadotropin-secreting human pituitary cell. J. Clin. Endocrinol. Metab. 72, 747-754. 5. Odell, W. D., Griffin, J., Bashey, H. M., and Snyder, E J. (1990). Secretion of chorionic gonadotropin by cultured human pituitary cells. J. Clin. Endocrinol. Metab. 71, 1318-1321. 6. Odell, W. D., and Griffin, J., (1989). Pulsatile secretion of chorionic gonadotropin during the normal menstrual cycle. J. Clin. Endocrinol. Metab. 69, 528-532. 7. Moyle, W. R., Myers, R. V., Wang, Y., Han, Y., Lin, W., Kelley, G. L., Ehrlich, P. H., Rao, S. N., and Bernard, M. P. (1998). Functional homodimeric glycoprotein hormones: Implications for hormone action and evolution. Chem. Biol. 5, 241-254. 8. Talmadge, K., Vamvakopoulos, N. C., and Fiddes, J. C. (1984). Evolution of the genes for the beta subunits of human chorionic gonadotropin and luteinizing hormone. Nature (London) 307, 37-40. 9. Manzella, S. M., Dharmesh, S. M., Beranek, M. C., Swanson, P., and Baenziger, J. U. (1995). Evolutionary conservation of the sulfated oligosaccharides on vertebrate glycoprotein hormones that control circulatory half-life. J. Biol. Chem. 270, 21665-21671. 10. Manzella, S. M., Hooper, L. V., and Baenziger, J. U. (1996). Oligosaccharides containing beta- 1,4-1inked N-acetylgalactosamine, a paradigm for protein-specific glycosylation. J Biol Chem 271, 12117-12120. 11. Fiddes, J. C., and Goodman, H. M. (1980). The cDNA for the betasubunit of human chorionic gonadotropin suggests evolution of a gene by readthrough into the 3'-untranslated region. Nature (London) 286, 684-687. 12. Furuhashi, M., Shikone, T., Fares, F. A., Sugahara, T., Hsueh, A. J., and Boime, I. (1995). Fusing the carboxy-terminal peptide of the chorionic gonadotropin (CG) beta-subunit to the common alpha-subunit: Retention of O-linked glycosylation and enhanced in vivo bioactivity of chimeric human CG. Mol. Endocrinol. 9, 54-63. 13. Knobil, E. (1988). The hypothalamic gonadotrophic hormone releasing hormone (GnRH) pulse generator in the rhesus monkey and its neuroendocrine control. Hum. Reprod. 3, 29-31. 14. Knobil, E. (1989). The electrophysiology of the GnRH pulse generator in the rhesus monkey. J. Steroid Biochem. 33, 669- 671. 15. Wise, P. M., Krajnak, K. M,. and Kashon, M. L. (1996). Menopause: The aging of multiple pacemakers Science 273, 67-70. 16. Wise, P. M., Kashon, M. L., Krajnak, K. M., Rosewell, K. L., Cai, A., Scarbrough, K., Harney, J. P., McShane, T., Lloyd, J. M., and Weiland, N. G. (1997). Aging of the female reproductive system: A window into brain aging. Recent. Prog. Horm. Res. 52, 279-303. 17. Brown, J. R., Skurnick, J. H., Sharma, N., Adel, T., and Santoro, N. (1993). Frequent intermittent ovarian function in women with premature menopause: A longitudinal study. Endocr. J. 1,467-474. 18. Santoro, N., Brown, J. R., Adel, T., and Skurnick, J. H. (1996). Characterization of reproductive hormonal dynamics in the perimenopause. J. Clin. Endocrinol. Metab. 81, 1495-1501. 19. Santoro, N., Banwell, T., Tortoriello, D., Lieman, H., Adel, T., and Skurnick, J. (1998). Effects of aging and gonadal failure on the hypothalamic-pituitary axis in women. Am. J. Obstet. Gynecol. 178, 732741. 20. Prior, J. C. (1998). Perimenopause: The complex endocrinology of the menopausal transition. Endocr. Rev. 19, 397-428. 21. Zeleznik, A. J., and Hillier, S. G. (1984). The role of gonadotropins in the selection of the preovulatory follicle. Clin. Obstet. Gynecol. 27, 927-940. 22. Hillier, S. G., Zeleznik, A. J., Knazek, R. A. and Ross, G. T. (1980). Hormonal regulation of preovulatory follicle maturation in the rat. J. Reprod. Fertil. 60, 219-229.
BIRKEN ET AL. 23. Reame, N. E., Wyman, T. L., Phillips, D. J., de Kretser, D. M., and Padmanabhan, V. (1998). Net increase in stimulatory input resulting from a decrease in inhibin B and an increase in activin A may contribute in part to the rise in follicular phase follicle-stimulating hormone of aging cycling women. J. Clin. Endocrinol. Metab. 83, 3302-3307. 24. Santoro, N., Adel, T., and Skurnick, J. (1999). Decreased inhibin tone and increased activin A secretion characterize reproductive aging in women. Fertil. Steril. 71,658-662. 25. Speroff, L., Glass, R. H., and Kase, N. G. (1994). Regulation of the menstrual cycle. In "Clinical Gynecologic Endocrinology and Infertility" (C. Mitchell, ed.), pp. 183-230. Williams & Wilkins, Baltimore, MD. 26. Carr, B. R. (1998). The ovary. In "Textbook of Reproductive Medicine" (B. R. Carr and R. E. Blackwell, eds.), pp. 207-243. Appleton & Lange, Stanford, CT. 27. Singer, D. B. and Haning, R. V. (1995). The menstrual cycle. In "Human Reproduction, Growth and Development" (D. R. Coustan, R. V. Hanning, and D. B. Singer, eds.), pp. 3-3, Little, Brown, New York. 28. Pantel, J., Robert, E, Troalen, E, Kujas, M., Bellet, D., and Bidart, J. M. (1998). Characterization of human lutropin carboxyl-terminus isoforms. Endocrinology (Baltimore) 139, 527-533. 29. Lapthorn, A. J., Harris, D. C., Littlejohn, A., Lustbader, J. W., Canfield, R. E., Machin, K. J., Morgan, F. J., and Isaacs, N. W. (1994). Crystal structure of human chorionic gonadotropin Nature (London) 369, 455461. 30. Lustbader, J. W., Wu, H., Birken, S., Pollak, S., Gawinowicz, K. M., Pound, A. M., Austen, D., Hendrickson, W. A., and Canfield, R. E. (1995). The expression, characterization, and crystallization of wildtype and selenomethionyl human chorionic gonadotropin. Endocrinology (Baltimore) 136, 640-650. 31. Wu, H., Lustbader, J. W., Liu, Y., Canfield, R. E., and Hendrickson, W. A. (1994). Structure of human chorionic gonadotropin at 2.6 A resolution from MAD analysis of the selenomethionyl protein. Structure 2, 545-558. 32. Matzuk, M. M., Keene, J. L., and Boime, I. (1989). Site specificity of the chorionic gonadotropin N-linked oligosaccharides in signal transduction. J. Biol. Chem. 264, 2409-2414. 33. Baenziger, J. U. (1996). Glycosylation: To what end for the glycoprotein hormones? Endocrinology (Baltimore) 137, 1520-1522. 34. Fiete, D., and Baenziger, J. U. (1997). Isolation of the SO4-4-GalNAcbetal,4GlcNAcbetal,2Manalpha-specific receptor from rat liver. J. Biol. Chem. 272, 14629-14637. 35. Baenziger, J. U., and Green, E. D. (1988). Pituitary glycoprotein hormone oligosaccharides: Structure, synthesis and function of the asparagine-linked oligosaccharides on lutropin, follitropin and thyrotropin. Biochim. Biophys. Acta 947, 287-306. 36. Baenziger, J. U., Kumar, S., Brodbeck, R. M., Smith, P. L., and Beranek, M. C. (1992). Circulatory half-life but not interaction with the lutropin/chorionic gonadotropin receptor is modulated by sulfation of bovine lutropin oligosaccharides. Proc. Natl. Acad. Sci. U.S.A. 89, 334-338. 37. Baenziger, J. U. (1994). Protein-specific glycosyltransferases: How and why they do it! FASEB J. 8, 1019-1025. 38. Wide, L., and Hobson, B. (1987). Some qualitative differences of hCG in serum from early and late pregnancies and trophoblastic diseases. Acta Endocrinol. (Copenhagen) 116, 465-472. 39. Gudermann, T., Brockmann, H., Simoni, M., Gromoll, J., and Nieschlag, E. (1994). In vitro bioassay for human serum follicle-stimulating hormone (FSH) based on L cells transfected with recombinant rat FSH receptor: Validation of a model system. Endocrinology (Baltimore) 135, 2204-2213. 40. Christin-Maitre, S., Taylor, A. E., Khoury, R. H., Hall, J. E., Martin, K. A., Smith, E C., Albanese, C., Jameson, J. L., Crowley, W. F., Jr., and Sluss, E M. (1996). Homologous in vitro bioassay for folliclestimulating hormone (FSH) reveals increased FSH biological signal
CHAPTER 4 Gonadotropins and Menopause: New Markers
41. 42.
43.
44. 45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
during the mid- to late luteal phase of the human menstrual cycle. J. Clin. Endocrinol. Metab. 81, 2080-2088. Christin-Maitre, S., and Bouchard, P. (1996). Bioassays of gonadotropins based on cloned receptors. Mol. Cell. Endocrinol. 125, 151-159. Padmanabhan, V., Chappel, S. C., and Beitins, I. Z. (1987). An improved in vitro bioassay for follicle-stimulating hormone (FSH): Suitable for measurement of FSH in unextracted human serum. Endocrinology (Baltimore) 121, 1089-1098. Diaz-Cueto, L., Mendez, J. P., Barrios-de-Tomasi, J., Lee, J. Y., Wide, L., Veldhuis, J. D., and Ulloa-Aguirre, A. (1994). Amplitude regulation of episodic release, in vitro biological to immunological ratio, and median charge of human chorionic gonadotropin in pregnancy. J. Clin. Endocrinol. Metab. 78, 890-897. Nisula, B. C. (1993). Measurement of human chorion gonadotropin by biological methods. Scand. J. Clin. Lab. Invest., Suppl. 216, 114-117. Schneyer, A. L., Sluss, P. M., Whitcomb, R. W., Hall, J. E., Crowley, W. F., Jr., and Freeman, R. G. (1991). Development of a radioligand receptor assay for measuring follitropin in serum: Application to premature ovarian failure. Clin. Chem. (Winston-Salem, N.C.) 37, 508514. Whitcomb, R. W., and Schneyer, A. L. (1990). Development and validation of a radioligand receptor assay for measurement of luteinizing hormone in human serum. J. Clin. Endocrinol. Metab. 71, 591-595. Ravindranath, N., Srilatha, N. S., Sairam, M. R., and Moudgal, N. R. (1992). Ability of deglycosylated human chorionic gonadotropin (dghCG) to block luteal function and establishment of pregnancy in bonnet monkeys (Macaca radiata). Indian J. Exp. Biol. 30, 982-986. Larsen, P. R. (1976). Quantitation of triiodothyronine and thyroxine in human serum by radioimmunoassay. In "Hormones in Human Blood: Detection and Assay" (H. N. Antonaides, ed.), pp. 679-697. Harvard University Press, Cambridge, MA. O'Connor, J. F., Schlatterer, J. P., Birken, S., Krichevsky, A., Armstrong, E. G., McMahon, D., and Canfield, R. E. (1988). Development of highly sensitive immunoassays to measure human chorionic gonadotropin, its beta-subunit, and beta core fragment in the urine: Application to malignancies. Cancer Res. 48, 1361-1366. Alfthan, H., Haglund, C., Roberts, P., and Stenman, U. H. (1992). Elevation of free beta subunit of human choriogonadotropin and core beta fragment of human choriogonadotropin in the serum and urine of patients with malignant pancreatic and biliary disease. Cancer Res. 52, 4628-4633. Spencer, K., Aitken, D. A., Macri, J. N., and Buchanan, P. D. (1996). Urine free beta hCG and beta core in pregnancies affected by Down's syndrome. Prenat. Diagn. 16, 605-613. Cole, L. A., Wang, Y. X., Elliott, M., Latif, M., Chambers, J. T., Chambers, S. K., and Schwartz, P. E. (1988). Urinary human chorionic gonadotropin free beta-subunit and beta- core fragment: A new marker of gynecological cancers. Cancer Res. 48, 1356-1360. Diaz-Cueto, L., Barrios-de-Tomasi, J., Timossi, C., Mendez, J. P., and Ulloa-Aguirre, A. (1996). More in vitro bioactive, shorter-lived human chorionic gonadotrophin charge isoforms increase at the end of the first and during the third trimesters of gestation. Mol. Hum. Reprod. 2, 643-650. Ho, H.-H., O'Connor, J. F., Tieu, J., Overstreet, J. W., and Lasley, B. L. (1997). Characterization of hCG in normal and abnormal pregnancies. Early Pregnancy: Biol. Med. 3, 213-224. Romani, P., Robertson, D. M., and Diczfalusy, E. D. (1977). Biologically active luteinizing hormone (LH) in plasma: II. Comparison with immunologically active LH levels throughout the human menstrual cycle. Acta Endocrinol. (Copenhagen) 84, 697-712. Reame, N. E., Kelche, R. P., Beitins, I. Z., Yu, M. Y., Zawacki, C. M., and Padmanabhan, V. (1996). Age effects of follicle-stimulating hormone and pulsatile luteinizing hormone secretion across the menstrual cycle of premenopausal women. J. Clin. Endocrinol. Metab. 81, 15121518.
75 57. Birken, S., Armstrong, E. G., Kolks, M. A., Cole, L. A., Agosto, G. M., Krichevsky, A., Vaitukaitis, J. L., and Canfield, R. E. (1988). Structure of the human chorionic gonadotropin beta-subunit fragment from pregnancy urine. Endocrinology (Baltimore) 123, 572-583. 58. Birken, S., Chen, Y., Gawinowicz, M. A., Agosto, G. M., Canfield, R. E., and Hartree, A. S. (1993). Structure and significance of human luteinizing hormone-beta core fragment purified from human pituitary extracts. Endocrinology (Baltimore) 133, 985-989. 59. Birken, S., Kovalevskaya, G., and O'Connor, J. (1996). Metabolism of hCG and hLH to multiple urinary forms. Mol. Cell. Endocrinol. 125, 121-131. 60. Kovalevskaya, G., Birken, S., O'Connor, J., Schlatterer, J., Maydelman, Y., and Canfield, R. (1995). HLH beta core fragment in the urine of ovulating women: A sensitive and specific immunometric assay for its detection. Endocrine 3, 881-887. 61. Blithe, D. L., Akar, A. H., Wehmann, R. E., and Nisula, B. C. (1988). Purification of beta-core fragment from pregnancy urine and demonstration that its carbohydrate moieties differ from those of native human chorionic gonadotropin-beta. Endocrinology (Baltimore) 122, 173-180. 62. Nisula, B. C., Blithe, D. L., Akar, A., Lefort, G., and Wehmann, R. E. (1989). Metabolic fate of human choriogonadotropin. J. Steroid Biochem. 33, 733-737. 63. O'Connor, J. F., Kovalevskaya, G., Birken, S., Schlatterer, J. P., Schechter, D., McMahon, D. J., and Canfield, R. E. (1998). The expression of the urinary forms of human luteinizing hormone beta fragment in various populations as assessed by a specific immunoradiometric assay. Hum. Reprod. 13, 826-835. 64. Birken, S., Gawinowicz, M. A., Kardana, A., and Cole, L. A. (1991). The heterogeneity of human chorionic gonadotropin (hCG). II. Characteristics and origins of nicks in hCG reference standards. Endocrinology (Baltimore) 129, 1551-1558. 65. Cole, L. A., Kardana, A., Ying, F. C., and Birken, S. (1991). The biological and clinical significance of nicks in human chorionic gonadotropin and its free beta-subunit. Yale J. Biol. Meal. 64, 627-637. 66. Krichevsky, A., Birken, S., O'Connor, J., Bikel, K., Schlatterer, J., Yi, C., Agosto, G., and Canfield, R. (1991). Development and characterization of a new, highly specific antibody to the human chorionic gonadotropin-beta fragment. Endocrinology (Baltimore) 128, 12551264. 67. Krichevsky, A., Birken, S., O'Connor, J., Bikel, K., Schlatterer, J., and Canfield, R. (1994). The development of a panel of monoclonal antibodies to human luteinizing hormone and its application to immunological mapping and two-site assays. Endocrine 2, 511-520. 68. O'Connor, J. F., Birken, S., Lustbader, J. W., Krichevsky, A., Chen, Y., and Canfield, R. E. (1994). Recent advances in the chemistry and immunochemistry of human chorionic gonadotropin: Impact on clinical measurements. Endocr. Rev. 15, 650-683. 69. Pettersson, K., Ding, Y. Q., and Huhtaniemi, I. (1991). Monoclonal antibody-based discrepancies between two-site immunometric tests for lutropin. Clin. Chem. (Winston-Salem, N.C.) 37, Pt 1):1745-1748. 70. Pettersson, K., Ding, Y. Q., and Huhtaniemi, I. (1992). An immunologically anomalous luteinizing hormone variant in a healthy woman. J. Clin. Endocrinol. Metab. 74, 164-171. 71. Martin-Du-Pan, R. C., Horak, M., and Bischof, P. (1994). Clinical significance of invisible or partially visible luteinizing hormone. Hum. Reprod. 9, 1987-1990. 72. de Medeiros, S. F., Amato, F., and Norman, R. J. (1991). Stability of immunoreactive beta-core fragment of hCG. Obstet. Gynecol. 77, 53-59. 73. Klein, N. A., Illingworth, P. J., Groome, N. P., McNeilly, A. S., Battaglia, D.E., and Soules, M. R. (1996). Decreased inhibin B secretion is associated with the monotropic FSH rise in older, ovulatory women: A study of serum and follicular fluid levels of dimeric inhibin A and B in spontaneous menstrual cycles. J. Clin. Endocrinol. Metab. 81, 27422745.
76 74. Welt, C. K., McNicholl, D. J., Taylor, A. E., and Hall, J. E. (1999). Female reproductive aging is marked by decreased secretion of dimeric inhibin. J. Clin. Endocrinol. Metab. 84, 105 - 111. 75. Burger, H. G., Cahir, N., Robertson, D. M., Groome, N. P., Dudley, E., Green, A., and Dennerstein, L. (1998). Serum inhibins A and B fall differentially as FSH rises in perimenopausal women. Clin. Endocrinol. (Oxford) 48, 809-813; published erratum: Ibid., 49(4), 550. 76. Matt, D. W., Kauma, S. W., Pincus, S. M., Veldhuis, J. D., and Evans, W. S. (1998). Characteristics of luteinizing hormone secretion in younger versus older premenopausal women. Am. J. Obstet. Gynecol. 178, 504-510. 77. Klein, N. A., and Soules, M. R. (1998). Endocrine changes of the perimenopause. Clin. Obstet. Gynecol. 41,912-920. 78. Soules, M. R., Battaglia, D. E., and Klein, N. A. (1998). Inhibin and reproductive aging in women. Maturitas 30, 193-204. 79. Burger, H. G. (1996). The endocrinology of the menopause. Maturitas 23, 129-136. 80. Burger, H. G. (1996). The menopausal transition. Bailliere's Clin. Obstet. Gynecol. 10, 347-359. 81. Vihko, K. K. (1996). Gonadotropins and ovarian gonadotropin receptors during the perimenopausal transition period. Maturitas 23 (Suppl.), S19-$22. 82. Schmidt, P. J., Gindoff, P. R., Baron, D. A., and Rubinow, D. R. (1996). Basal and stimulated gonadotropin levels in the perimenopause. Am. J. Obstet. Gynecol. 175, 643-650. 83. Saketos, M., Sharma, N., Adel, T., Raghuwanshi, M., and Santoro, N. (1994). Evalution of time-resolved immunofluorometric assay and specimen storage conditions for measuring gonadotropins. Clin. Chem. (Winston-Salem, N.C.) 40, 749-753. 84. Kesner, J. S., Knecht, E. A., and Krieg, E. F., Jr. (1995). Stability of urinary female reproductive hormones stored under various conditions. Reprod. Toxicol. 9, 239-244. 85. Livesey, J. H., Hodgkinson, S. C., Roud, H. R., and Donald, R. A. (1980). Effect of time, temperature and freezing on the stability of im-
BIRKEN ET AL.
86.
87.
88. 89.
90.
91.
92.
93.
94.
munoreactive LH, FSH, TSH, growth hormone, prolactin and insulin in plasma. Clin. Biochem. 13, 151-155. Livesey, J. H., Roud, H. K., Metcalf, M. G., and Donald, R. A. (1983). Glycerol prevents loss of immunoreactive follicle-stimulating hormone and luteinizing hormone from frozen urine. J. Endocrinol. 98, 381-384. Kardana, A., Elliott, M. M., Gawinowicz, M. A., Birken, S., and Cole, L. A. (1991). The heterogeneity of human chorionic gonadotropin (hCG). I. Characterization of peptide heterogeneity in 13 individual preparations ofhCG. Endocrinology (Baltimore) 129, 1541-1550. Sturgeon, C. M., and McAllister, E. J. (1998). Analysis ofhCG: Clinical applications and assay requirements. Ann. Clin. Biochem. 35,460- 491. Cole, L. A. (1997). Immunoassay of human chorionic gonadotropin, its free subunits, and metabolites Clin. Chem. (Winston-Salem, N.C.) 43, 2233-2243. Pettersson, K. S., and Soderholm, J. R. (1991). Individual differences in lutropin immunoreactivity revealed by monoclonal antibodies. Clin. Chem. (Winston-Salem, N.C.) 37, 333-340. Barbe, E, Legagneur, H., Watrin, V., Klein, M., and Badonnel, Y. (1995). Undetectable luteinizing hormone levels using a monoclonal immunometric assay. J. Endocrinol. Invest. 18, 806-808. Mitchell, R., Hollis, S., Crowley, V., McLoughlin, J., Peers, N., and Robertson, W. R. (1995). Immunometric assays of luteinizing hormone (LH): Differences in recognition of plasma LH by anti-intact and betasubunit-specific antibodies in various physiological and pathophysiological situations. Clin. Chem. (Winston-Salem, N.C.) 41, (Pt. 1), 11391145. Costagliola, S., Niccoli, P., and Carayon, P. (1994). Glycoprotein hormone isomorphism and assay discrepancy: The paradigm of luteinizing hormone (LH). J. Endocrinol. Invest. 17, 291-299. Qiu, Q., Kuo, A., Todd, H., Dias, J. A., Gould, J. E., Overstreet, J. W., and Lasley, B. L. (1998). Enzyme immunoassay method for total urinary follicle-stimulating hormone (FSH) beta subunit and its application for measurement of total urinary FSH. Fertil. Steril. 69, 278-285.
~HAPTER
Genetic Programming In Ovarian
Development
and Oogenesis JOE LEIGH SIMPSON
Departments of Obstetrics and Gynecology and Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas 77030
I. II. III. IV.
Ovarian Differentiation Requires Only One X (Constitutive) Polygenic and Stochastic Control over Oocyte Number Monosomy X X Chromosomal Mosaicism: 45,X/46,XX and 45,X/47,XXX V. Pitfalls in Localizing Ovarian Maintenance Genes to Specific Regions of the X
VI. VII. VIII. IX. X. XI.
I. O V A R I A N
Failure of germ cell development is associated with complete ovarian failure, resulting in lack of secondary sexual pubertal development (primary amenorrhea). A decreased number but not a total absence of germ cells is more likely associated with premature ovarian failure, presenting with infertility or secondary amenorrhea (see Chapter 8). Yet complete and premature ovarian failure may be different manifestations of the same underlying pathogenic and etiologic processes. Many different genetic mechanisms are pertinent to the processes m chromosomal abnormalities, Mendelian mutations of autosomal or X-linked genes, and polygenic/ multifactorial factors. In this contribution, we enumerate clinical disorders associated with germ cell abnormalities, deducing etiologic factors responsible for ovarian differentiation and oogenesis in normal females.
MENOPAUSE:
BIOLOGY AND PATHOBIOLOGY
Genes on the X Short Arm Genes on the X Long Arm Nature of X Ovarian Maintenance Determinants Autosomal Chromosomal Abnormalities Autosomal Genes (Mendelian) To What Extent Is Premature Ovarian Failure Genetic? References
REQUIRES
DIFFERENTIATION
ONLY ONE X
(CONSTITUTIVE) In the absence of the Y chromosome, the indifferent embryonic gonad always develops into an ovary. Germ cells exist in 45,X human fetuses [ 1]. Oocyte development initially exists even in 46,XY phenotypic females, such as in infants with XY gonadal dysgenesis [2] or the genito-palatocardiac syndrome [3]. Oocyte development in the presence of a Y chromosome is also well documented in mice [4]. Thus, the pathogenesis of germ cell failure in humans can be deduced to be increased germ cell attrition. If two intact X chromosomes are not present, ovarian follicles in 45,X individuals usually degenerate by birth. Genes on the second
77
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
78
JOE LEIGH SIMPSON
X chromosome are thus responsible for ovarian maintenance, rather than for ovarian differentiation.
II. P O L Y G E N I C CONTROL
OVER
AND STOCHASTIC OOCYTE
NUMBER
It is to be expected that oocyte number (reservoir) will be low in some women simply on statistical (stochastic) grounds. Normal distribution exists for all common anatomic traits (e.g., height), and this principle should apply to oocyte number and reservoir at birth. That a normal distribution of germ cell number exists in ostensibly normal females is well established in animals but difficult to prove in humans. Different rodent strains show characteristic breeding duration, implying genetic control over either the rate of oocyte depletion or the number of oocytes initially present. It follows that some ostensibly normal (menstruating) women may have decreased oocyte reservoir or increased oocyte attrition on a genetic basis, analogous to animal models. In humans a genetic basis for the above can be presumed by analogy to the heritability of age at human menopause, a characteristic that clearly shows familiar tendencies. Assessing heritability of age at menopause is complicated because iatrogenic behavior (e.g., hysterectomy) and other confounding factors (e.g., leiomyomata or uterine cancer) must be taken into account. However, several studies within the past decade have directly addressed the issue. Cramer et al. [5] performed a case control study on 10,606 United States women who were between 45 and 54 years of age. Women with an early menopause ( 4 0 - 4 5 years) were age-matched with controls who were either still menstruating or had experienced menopause after age 45 years. Of 129 early menopause cases (<46 years), 37.5% had a family history of a similarly affected mother, sister, aunt, or grandmother. Only 9% of controls had such a relative (odds ratio after adjustment, 6.1; 95% C.I., 3.9 to 9.4). As predicted on the basis of polygenic expectations, the odds ratio was greatest (9.1) for sisters and for when menopause occurred prior to 40 years. The frequency of galactose1-phosphate uridylyltransferase (GALT) variants (N314D or Q188R) did not deviate from expected in early menopause cases, in contrast to previous studies by the same authors [6]. Torgerson and colleagues [7] reported that if women underwent menopause during the 5-year centile aged 4 5 - 4 9 years, the likelihood was increased that menopause would occur in a similar 5-year centile in their daughter. Twin studies have been used to estimate heritability of age at menopause. Two studies have shown similar results [8]. Snieder et al. [9] studied 27 monozygotic (MZ) and 353 dizygotic (DZ) twin pairs in the United States. For age at menopause correlation (r) was 0.58 in MZ twins and 0.39 in DZ twins (h 2 = 63%). Treolar et al. [8,10] performed a simi-
lar study in 1177 MZ and 711 DZ Australian twin pairs. For age at menopause, correlations (r) were 0.49-0.57 for MZ and 0.31-0.33 for DZ. Heritability (h 2) was 31-53% [8]. Differences between MZ and DZ groups held when iatrogenic menopause (hysterectomy for leiomyomata or endometriosis) was taken into account.
III. M O N O S O M Y X The complement most frequently associated with ovarian dysgenesis is 45,X. The proportion of 45,X individuals in a given sample will depend on the method of ascertainment. Fewer 45,X individuals will be detected if primary amenorrhea is the presenting complaint than if short stature or various somatic anomalies are the presenting complaints. Primary amenorrhea is more likely to be the presenting complaint in women examined by gynecologists, whereas short stature is likely to be evaluated in children examined by pediatricians. Overall, about 50% of all patients with gonadal dysgenesis have a 45,X complement; 25% have sex chromosomal mosaicism with a structural abnormality (e.g., 45,X/46,XX). Far fewer have a structurally abnormal X or Y chromosome or no detectable chromosomal abnormality [11,12]. In 80% of cases the paternally derived X has been lost [13]. With one possible exception to be noted, the phenotypes of 45,X m and 45,XP cases do not differ. (Xm,X of maternal origin; XP,X of paternal origin). That is, no evidence exists in general for imprinting [14,15]. In structurally abnormal X chromosomes, it is also the paternal X that is lost [16,17]. The above suggests that X m and XP chromosomes are lost at random [ 18]. Because 45,Y is lethal, the theoretical percentage of 45,X m cases would be 67%, not greatly different from the 80% actually observed.
A. G o n a d s In most 45,X adults with gonadal dysgenesis, the normal gonad is replaced by a white fibrous streak, 2 to 3 cm long and about 0.5 cm wide, located in the position ordinarily occupied by the ovary. A streak gonad is characterized histologically by interlacing waves of dense fibrous stroma, indistinguishable from normal ovarian stroma (Fig. 1). That germ cells are usually completely absent in adults but present in 45,X embryos is the basis for the belief that the pathogenesis of germ cell failure is increased atresia, not failure of germ cell formation. Speed [19,20] has shown that in monosomy X oogenesis ceases in meiosis I at or before the pachytene meiotic stage. Degenerating pachytene oocytes are observed. A variety of pairing abnormalities and disruptions are seen at later stages of meiosis I, but oocytes in dictyotene are rare.
CHAPTER5 Genetic Programming in Ovarian Development
Ogata and Matsuo [ 18] argue that the ovarian failure found in monosomy X is caused by generalized (nonspecific) meiotic pairing errors, the extent of ovarian failure correlating with extent of pairing failure. Ovarian rete tubules, which probably originate from either mesonephric tubules or medullary sex cords, are present in the median portion of most streak gonads. Hilar cells are usually detected in streak gonads of patients past the age of expected puberty. That 45,X humans manifest streak gonads is not so obvious as one might expect. Relatively normal ovarian development occurs in many other monosomy X mammals (e.g., mice). These observations are, incidentally, at odds with the hypothesis [19,20] that ovarian failure merely reflects meiotic pairing errors. If the hypothesis were true, the monosomy X mouse should not differ from the monosomy X human. The more likely explanation is that in humans not all loci on the normal heterochromatic (inactive) X are inactivated. In addition, X inactivation never exists in oocytes, with X reactivation of germ cells occurring before entry in meiotic oogenesis [21 ]. X inactivation could also occur only after some crucial time of differentiation, beyond which only a single euchromatic (active) X is necessary for continued oogenesis.
79
B. S e c o n d a r y S e x u a l D e v e l o p m e n t Although streak gonads are usually present in 45,X individuals, about 3% of adult cases menstruate spontaneously and 5% show breast development (Table I). Occasionally, the interval between menstrual periods appears normal in 45,X patients, and fertile patients have been reported. Although an undetected 46,XX cell line should always be suspected in menstruating 45,X patients, it is plausible that a few 45,X individuals could be fertile, inasmuch as germ cells are present in 45,X embryos. The rare offspring of 45,X women are probably not at greatly increased risk for chromosomal abnormalities [22,23], although theoretically they should be. Some authors disagree with this statement [24], and work with X-specific fluorescence in situ hybridization (FISH) probes suggests that low-grade 45,X/46,XX mosaicism may be increased in women experiencing repeated abortion [25]. Irrespective, menstruation and fertility occur so rarely that 45,X patients should be counseled to anticipate primary amenorrhea and sterility. After hormone therapy is initiated in such women, uterine size becomes normal. This permits 45,X women to carry pregnancies in their own uterus after receipt of donor embryos or donor oocytes mixed with their husband's sperm.
80
JOE LEIGH SIMPSON TABLE I
Somatic Features Associated with 45,X C h r o m o s o m a l C o m p l e m e n t G r o w t h a
Body size Decreased birth weight Decreased adult height (141-146 cm) Intellectual function Verbal IQ > performance IQ Cognitive deficits (space-form blindness) Craniofacial Premature fusion of sphenooccipital and other sutures, producing brachycephaly Abnormal pinnae Retruded mandible Epicanthal folds (25%) High-arched palate (36%) Abnormal dentition Visual anomalies, usually strabismus (22%) Auditory deficits; sensorineural or secondary to middle ear infections Neck Pterygium coli (46%) Short, broad neck (74%) Low nuchal hair (71%) Chest Rectangular contour (shield chest) (53%) Apparent, widely spaced nipples Tapered lateral ends of clavicle
Cardiovascular Coarctation of aorta or ventricular septal defect (10-16%) Renal (38%) Horseshoe kidneys Unilateral renal aplasia Duplication of ureters Gastrointestinal Telangiectasias Skin and lymphatics Pigmented nevi (63%) Lymphedema (38%) due to hypoplasia of superficial vessels Nails Hypoplasia and malformation (66%) Skeletal Cubitus valgus (54%) Radial tilt of articular surface of trochlear Clinodactyly V Short metacarpals, usually IV (48%) Decreased carpal arch (mean angle 117 ~ Deformities of medcal tibial condyle Dermatoglyphics Increased total digital ridge count Increased distance between palmar triradii a and b Distal axial triradius in position t'
aModified from Simpson [11].
IV. X CHROMOSOMAL MOSAICISM: 45,X/46,XX AND 45,X/47,XXX If nondisjunction or anaphase lag occurs in the zygote and embryo, two or more cell lines may result (mosaicism) (Fig. 2). The final c o m p l e m e n t will depend on the stage at which abnormal cell division occurs and on the types of daughter cells that survive following nondisjunction or anaphase lag. Detection of mosaicism depends on the n u m b e r of
NORMAL MITOSIS
cells analyzed per tissue and on the n u m b e r of tissues analyzed [11,12]. The most c o m m o n form of mosaicism associated with gonadal dysgenesis is 45,X/46,XX. Individuals with a 45,X/ 46,XX c o m p l e m e n t predictably show fewer anomalies than do 45,X individuals. Simpson [11] tabulated that 12% of 45,X/46,XX individuals menstruate c o m p a r e d with only 3% of 45,X individuals. A m o n g 45,X/46,XX individuals, 18% undergo breast development c o m p a r e d with 5% of 45,X in-
MITOTIC NONDISJUNCTION
FIGURE 2 Diagrammatic representation of the products of normal mitosis and mitosis characterized by nondisjunction of a Y chromosome. If all daughter cells survived, the complement would be 45,X/46,XY/47,XYY [ 12].
CHAPTER5 Genetic Programming in Ovarian Development dividuals. Mean adult height is greater with a 45,X/46,XX complement than with 45,X; more mosaic (25%) than nonmosaic (5%) patients reach adult heights greater than 152 cm [11]. Somatic anomalies are less likely to occur in 45,X/ 46,XX than in 45,X individuals. 45,X/47,XXX occurs less often but is phenotypically similar to 45,X/46,XX. Individuals with 45,X/46,XY may also show bilateral streak gonads; however, more often they show a unilateral streak gonad and a contralateral dysgenetic testis (mixed gonadal dysgenesis).
V. P I T F A L L S I N L O C A L I Z I N G O V A R I A N MAINTENANCE GENES TO SPECIFIC REGIONS OF THE X Delineating the region (genes) on the X responsible for ovarian maintenance is the first step in understanding normal ovarian differentiation and in producing gene products of therapeutic benefit. Until the past decade, phenotypickaryotypic correlations to deduce location of gonadal and somatic determinants relied solely on metaphase analysis. Prometaphase karyotypes allow 1200 band analysis (traditional GTG banding 400-500), but each band still contains considerable DNA. More refined analysis is now possible using polymorphic DNA markers that allow precise resolution far beyond the capacity of light microscopy. Progress has, nonetheless, been slow compared to that achieved in delineating the regions of the Y necessary for testicular differentiation (SRY) or spermatogenesis (DAZ). Several impediments have resulted in a relative lack of progress. One is that the incidence of X deletions is very low. Analyzing only cases ascertained by population-based methods is impractical because no individuals with X deletions were recovered among 50,000 consecutively born neonates [26]. Most del(Xp) or del(Xq) individuals have been identified only because they manifested clinical abnormalities, exceptions being familial cases or cases detected in fetuses at the time of prenatal genetic diagnosis undertaken because of their mother's advanced maternal age. Doubtless many less severely affected individuals escape detection. Mode of ascertainment should ideally be considered in phenotypickaryotypic analysis, but in reality this is impractical because sample sizes are too small. Inevitably biases of selection arise. Another pitfall impeding molecular analysis of X ovarian maintenance genes is that analysis is not always derived from individuals who are well-studied cytogenetically. Mosaicism in nonhematogenous tissues has not always been excluded to the extent reasonably possible. Individuals with unstable aberrations (rings, dicentrics) should probably be excluded from phenotypic-karyotypic deductions because monosomy X and other cell lines may arise secondarily, sometimes in
81 tissues (e.g., gonads) relatively inaccessible to study. Utilizing X/autosome translocations for analysis may also be hazardous because of vicissitudes of X inactivation, and because autosomal regions are not devoid of significance for gonadal differentiation.
VI. GENES ON THE X SHORT
ARM
A. 46,X, d e l ( X p ) or 4 5 , X / 4 6 , X , d e l ( X p ) D e l e t i o n s Deletions of the short arm of the X chromosome show variable phenotype, depending on the amount of Xp persisting. The most common breakpoint for terminal deletions is Xpl 1 (Fig. 3). In 46,X, del(X)(pl 1), only proximal Xp remains; the del(Xp) chromosome thus appears acrocentric or telocentric. Chromosomes characterized by progressively more distal breakpoints have been reported: Xp21, 22.1, and 22.3. X/autosomal translocations leading to Xp interstitial deletions have been reported, and are analytically useful albeit subject to caveats noted in the previous section. Availability of polymorphic DNA markers now allow precise determinations of breakpoints in terminal deletions, but still
FIGURE 3 A normalX chromosomeand deletions of the X chromosome derived fromthree differentpersons [32].
82
JOE LEIGH SIMPSON
relatively few cases have been subjected to refined molecular analysis. Approximately half of 46,X, del(Xp)(pll) individuals show primary amenorrhea and gonadal dysgenesis. The others menstruate and usually show breast development. In one tabulation by the author of 27 reported del(X)(p 11.2 11.4) individuals, 12 menstruated spontaneously; however, menstruation was rarely normal [27]. Additional compilations have not materially altered these conclusions [28-30]. Ogata and Matsuo [ 18] estimate that 50% of del(X)p 11 cases show primary amenorrhea, with 45% showing secondary amenorrhea. Ovarian function is thus observed more often in individuals with a del(Xpl 1) chromosome than in 45,X individuals. Women with more distal deletions [del(X)(p21.1 to p22.1.22)] menstruate more often, but many are still infertile or even have secondary amenorrhea (Fig. 4). Thus, Xp [X(pter --->p21)] retains a role in ovarian development [2830]. The distal region of importance must involve Xp21,
22.3 22.2 22.1 21 11.4 11.3 11.2 11.1 11 12
13
B. Isochromosomes for Xq m
~
9
"
II ~
"iii
US
amain m
9
m m
m|m
9
mm9 21
mm
22 23 24
ano
mm
A
9
9
25 26
22.1, or 22.2 because del(X)(p22.3) cases do not show primary amenorrhea. Most women with deletions of Xp are short in stature. Thus, statural determinant(s), i.e., regions with genes, must exist on Xp. Because del(Xp) women may menstruate but still be short, regions on Xp responsible for ovarian and statural determinants must be distinct [28-32]. Clinically it is important to realize that del(Xp) women may be short despite manifesting normal ovarian function. Both mother and daughter may show the same Xp deletions, not only in association with X/autosome translocation but also in association with terminal deletions. In 1977 Fraccaro et al. [31] first emphasized familial distal Xp deletions. Among 10 del(Xp) cases subsequently studied by James et al. [16] were two mother-daughter pairs; only 6 of their 10 cases arose de novo. Familial cases involved deletion at Xp 11 as well as at Xp22-12 are also reported. Xp interstitial deletions involving Xp 11-22 and Xp 11.422.3 [33,34] have been reported.
mm
27
28 FIGURE 4 Ovarian function associated with simple deletions of the X chromosome. All cases are characterized by banding studies and reasonable exclusion of mosaicism [27]. n, 1o amenorrhea; A, 2 ~ amenorrhea oligomenorrhea; o, fertility or regular menses.
Division of the centromere in the transverse rather than in the longitudinal plane results in an isochromosome, a metacentric chromosome consisting of isologous arms. Both arms are structurally identical and contain the same genes. An isochromosome for the X long arm [i(Xq)] differs from a terminal deletion of Xp in that not just the terminal portion but all of the Xp is deleted. Many isochromosomes for Xq are in reality isodicentrics, the clinical significance of which is that a minute portion of Xp is duplicated and retained in addition to duplication of the entire Xq. An isochromosome for the X long arm is the most common X structural abnormality, but coexisting 45,X cell lines (mosaicism) are typical. Nonmosaic cases are relatively uncommon. 46,X,i(Xq) individuals almost always have streak gonads and primary amenorrhea. Occasionally menstruation is observed, but surveys continue to agree with early reports published by the author [ 11 ] in showing rarity of menstruation [18]. The near complete lack of gonadal development in 46,X,i(Xq) contrasts to that in 46,X, del(X)(p 11) individuals, about half of who menstruate or develop breasts. The contrast is in greater with more distal Xp deletions. Phenotypic differences could be explained if gonadal determinants were present at several different locations on Xp, one locus being deficient in i(Xq) yet retained in del(X)(p 11). Alternatively, 46,XX cells may be associated with del(Xp) more often than generally appreciated. Irrespective, duplication of Xq, that is, i(Xq), fails to compensate for deficiency of Xp. One explanation is that gonadal determinants on Xq and Xp have different functions. Another is that all loci on i(Xq) chromosomes are completely inactivated. It seems unlikely that
CHAPTER5 Genetic Programming in Ovarian Development TABLE II
83
Ovarian Functiona Ovarian failure (%) in X deletions Complete (primary amenorrhea or streak gonads)
Partial (secondary amenorrhea or abnormal menses)
Monosomy X (45,X)
88
12
0
Short arm deficiency del(X)(p 11) del(X) (p21 - 22.2 ) del(X)(p22.3) i(Xq) idic(Xq)
50 13 0 91 80
45 25 0 9 20
5 62 100 0 0
Long arm deficiency del(X)(ql3-21) del(X)(q22-25) del(X)(q26-28) idic(Xp)
69 31 8 73
31 56 67 27
0 13 25 0
Deficiency
No failure (presumed normal)
a As tabulated on the basis of cases reviewed in 1995 by Ogato and Matsuo [ 18]. Ogato and Matsuo provided data in the first two columns, with the assumption being that the remainder of cases have normal ovarian function (e.g., 5% in del(X)(pll). Publications surveyed overlap in large part those used for analysis by Simpson [27] (see Fig. 4).
duplication of Xq per se produces abnormalities but is unknown, given that 47,XXX often appears clinically normal. Almost all reported 46,X,i(Xq) patients are short. Their mean height seems to be less than that of 45,X patients (Table II). The mean height of nonmosaic 46,X,i(Xq) patients is 136 cm [11], and many somatic features of the Turner stigmata are observed [ 11 ]. Somatic anomalies occur as frequently in 46,X,i(Xq) individuals as in 45,X individuals, and the spectrum of anomalies associated with the two complements is in general similar. James et al. [16] attempted an extensive molecular analysis of i(Xq), confirming short statue and finding relative deficiency of pterygium coli (webbing of the neck.) This observation suggests a protective effect for Xq with respect to pterygium coli.
VII. GENES ON THE X L O N G A R M
A. 46,X, del(Xq) and 45,X/46,X, del(Xq) Deletions Deletions of the X long arm are well known [27-30] and vary in composition. If the breakpoint leading to a terminal deletion originates at band Xql 3, the derivative chromosome resembles No. 17 or No. 18; a breakpoint at band Xq21 produces a chromosome resembling No. 16 (see Fig. 3).
Almost all deletions originating at Xql3 are associated with primary amenorrhea, lack of breast development, and complete ovarian failure [30]. Xq 13 thus seems to be an important region for ovarian maintenance. Key loci could lie in proximal Xq21, but not more distally, given that del(X) (q21) to (q24) individuals menstruate far more often (Fig. 4). Menstruating del(X)(q21) women might have retained a region that contained an ovarian maintenance gene, whereas del(X)(ql 3 or q21) women with primary amenorrhea might have lost such a locus [30]. Molecular attempts at mapping the region of Xq most integral for ovarian development have been reported. Sala et al. [35] studied seven X/autosome translocations involving Xq21-22. Four cases showed primary amenorrhea; the other three were described as follows: (1) one case with secondary amenorrhea, elevated gonadotropins, and small ovaries; (2) one case with secondary amenorrhea; and (3) one case with amenorrhea, absent breast development, and small ovaries. A region of Xq spanning 15 mb encompassed breakpoints in all seven cases. Breakpoints in four other X/autosome translocations studied by Philippe et al. [36] were also localized to the same region. The YAC contig encompassing these breakpoints spanned most of the Xq21 region and extended between DXS233 and DXS 1171 [37]. An X q - Y p homologous region is contained within this contig [38]. That breakpoints associated with ovarian failure spanned the entire Xq21 region makes it unlikely that a single gene causes ovarian failure, unless in these balanced X/autosome translocations ovarian failure is the result not of disruption of a gene per se, but rather is reflective of generalized cytologic perturbation. To this end several other observations are of note. First, a normal female has been observed having an X/autosome breakpoint in Xq21 (case 5513 of Philippe et al. [36]). If not explained in other ways (e.g., a 46,XX cell line), one must conclude that not all of Xq21 is obligatory for ovarian development. This concept would be at odds both with existence of a critical region between Xql 3 and Xq25 [39-42] as well as with the hypotheses that ovarian failure reflects generalized meiotic breakdown caused by any type of rearrangement [20]. Second, observation of a normal female with a breakpoint-involved Xq21 [36] is consistent with existence of many distinct ovarian genes in the same region. Third, in two but not the other five cases of Sala et al. [35] ovarian failure was accompanied by choroiderma. Again, this favors existence of multiple genes (contiguous gene syndrome). Sala et al. [35] concluded that eight different genes are responsible for X maintenance on Xq 21. In more distal Xq deletions, the more common phenotype is not primary amenorrhea but premature ovarian failure [28,29,43,44]. Although distal Xq seems less important than proximal Xq for ovarian maintenance, the former must still have regions important for ovarian maintenance. Informative cases have included terminal deletions originating at
84
JOE LEIGH SIMPSON
various sites and two interstitial deletions [43,45]. These interstitial deletions point out hazards of interpretation without molecular-based studies. Although there is no clear demarcation into discrete regions, it is heuristically useful to stratify terminal deletions into those occurring in regions Xq 13 --9 21, Xq22-25, and Xq 26-28. Table II shows the extent of ovarian function tabulated by Ogato and Matsuo [18] using such stratification. Figure 4 shows the author's tabulation in different format. Both estimates are based on pooled cases, and both are generally consistent. Distal Xq deletions are not infrequently familial. Some familial Xq deletions are derivative of Xq autosome translocations, but familial terminal or interstitial deletions also exist [45]. Familial Xp terminal or interstitial deletions have been characterized by various breakpoints between Xq25 and Xq28. Breakpoints near or in Xq27 seem most common. Some families have been ascertained for reasons other than premature ovarian failure, a case reported by our group having been ascertained following amniotic fluid analysis in a fetus [46]. This suggests that additional families would be detected if prometaphase analysis or polymorphic molecular studies were more routinely performed in premature ovarian failure. Distal Xq deletions seem to have a less severe effect on stature than do proximal deletions. Somatic anomalies of the Turner stigmata are uncommon and perhaps no more common than in the general population.
VIII. NATURE MAINTENANCE
OF X OVARIAN DETERMINANTS
Clearly the X chromosome is necessary for ovarian maintenance, preventing premature germ cell attribution and permitting progression beyond meiotic pachytene. Moreover, we have deduced that several different regions seem to contain crucial genes. At the least, key regions exist on proximal Xp and proximal Xq; an unknown number of determinants exist on distal Xp and distal Xq. As noted, Sala et al. [35] believed that eight different genes exist Xq21 alone. Ultimately both the number of individual genes and their gene product(s) will be determined. This would have both prognostic and potential therapeutic value, given that recombinant technology allows synthesis of a protein gene product(s) once the DNA sequence is known. At present little is known about the nature of ovarian maintenance gene products. Jones et al. [46] proposed that a key gene product is DFFRX; its locus is located on Xp 11.4 and homologous to a locus on Yql 1.2. Both genes escaped inactivation in two de novo (X)(pl 1.2) deletions. James et al. [16] considered DFFRX an unlikely candidate after observing ovarian function despite haploinsufficiency; however, neither of the two cases of James et al. [16] were completely normal clinically,
for which reason a role for DFFRX in gonadal development is not categorically excluded. An attractive candidate gene for a role in ovarian maintenance is the human homolog of the Drosophila melanogaster gene diaphanous (dia). This gene causes sterility in male and female Drosophila [47]. Sequence comparisons between dia and the relevant human expressed sequence tag (EST) DRE25 show significant homology. DRE25 in turn maps to human Xq22 [47]. As we have already noted, Xq22 is a key region for ovarian maintenance. The product of Drosophila dia is a member of a family of proteins that help establish cell polarity, govern cytokinesis, and reorganize the actin cytoskeleton. Studying familial premature ovarian failure, an Xq21/autosome translocation alluded to earlier [48] was found to be associated with disruption of DRE25 [49]. Perturbation involved the last intron, producing a human DIA characterized by truncated transcripts; the transcript was unstable and could not be translated. This mechanism seems relevant to the phenomenon of mRNAs in oocytes and embryos remaining untranslated for long periods of time, presumably until their proteins become necessary later in differentiation. On the other hand, human DIA is expressed not only in developing ovaries, but also in testis and other tissues. This suggests that in humans the role dia plays in oogenesis is neither primary nor specific. Four Xq 21 ~ Xqter deletions associated with premature ovarian failure recently by the same group [50] also indicated that dia is not necessarily involved in del(Xq) deletions. In one of the four cases, loss of exons at the 3' end of dia was detected; however, in the other three no perturbations of dia were evident. A broader biologic question can be posed concerning X ovarian maintenance determinants. Do the various regions contain gene(s) coding for different gene products? If so, the presumption would be that all these genes are either essential or at least contribute to normal ovarian differentiation. If different genes exist, the prospect of alternative therapeutic options is raised given that various gene products will eventually all be synthesized. However, it would seem hazardous evolutionarily if perpetuation of the species were to depend on transcription and translation of an entire cascade of ovarian differentiation genes, perturbation of any of which would be deleterious if not lethal (genetically). Moreover, ovarian disturbance associated with many X deletions is rarely complete. We have noted that terminal deletions involving either proximal Xp or proximal Xq may be associated with complete ovarian failure, but more distal deletions of Xp or Xq are far more likely to be associated with premature ovarian failure or normal ovarian function. Teleologically, it might be more attractive if all X ovarian maintenance determinants were to produce the same gene product or perhaps products capable of interaction (dimerization, for example). This would seem to be more conservative evolutionarily because mutation or deletion of a single locus would not be singularly catastrophic. If such a scenario
CHAPTER5 Genetic Programming in Ovarian Development were true, an ineluctable corollary would be that the X ovarian genes act in threshold fashion, thus exerting their primary effect through an autosomal gene. One mode of action might involve transcriptional or translational regulation of DNA-binding proteins.
85 ments is important because their offspring are at increased risk for gametes showing unbalanced segregation.
X. A U T O S O M A L
GENES
(MENDELIAN)
A. X X G o n a d a l D y s g e n e s i s IX. A U T O S O M A L ABNORMALITIES
CHROMOSOMAL
A. T r i s o m y Autosomal trisomy has long been known to affect adversely ovarian development. The question has always been whether this effect is mediated by nonspecific meiotic perturbation or by chromosome-specific genes, perhaps acting in double dose. Trisomies 13 and 18 are frequently associated with ovarian failure, as indicated by necropsy observations in stillborns or deceased neonates. Few longitudinal data are available for trisomy 13 or 18 because few affected females survive until menarche. In trisomy 21, however, ovarian function may be normal. There seems to be no objective information on age at menopause. Pregnancies occur in trisomy 21 females [51]. About onethird of offspring are aneuploid (fewer than the theoretically expected 50%). Ostensibly normal ovarian function in trisomy 21 suggests that specific ovarian genes exist on chromosomes 13 and 18. If nonspecific meiotic breakdown is merely secondary to an uneven number of chromosomes, the effect should be the same effect with chromosome 21 as with chromosomes 18 and 13.
B. T r a n s l o c a t i o n s Chromosomal rearrangements, specifically balanced autosomal reciprocal translocations, are not infrequently observed in otherwise normal women with complete or partial ovarian failure. As with autosomal trisomy it is unclear whether this association reflects disruption of autosomal loci integral for ovarian preservation and oogenesis. That no chromosome(s) is consistently involved suggests nonspecific meiotic perturbation. In fact, men who are azoospermic or oligospermic but otherwise normal clinically show balanced autosomal translocations far more often than expected: about 1% of men requiring intracytoplasmic sperm injection (ICSI) show a balanced autosomal rearrangement, typically a balanced translocation [52]. A problem of comparable magnitude probably exists in women, but relative accessibility of ovaries makes cytogenetic studies more difficult. In both sexes the pathogenesis presumably leading to meiotic breakdown involves malalignment or failure of synapis. Recognizing individuals with autosomal rearrange-
Gonadal dysgenesis histologically similar to that occurring in individuals with an abnormal sex chromosomal complement may be present in 46,XX individuals, as shown by the author over 25 years ago [53]. Mosaicism has been reasonably excluded in affected individuals, although mosaicism restricted to the embryo can never be excluded. The general term XX gonadal dysgenesis can be applied to those individuals. 1. PHENOTYPE
Many different forms of 46,XX gonadal dysgenesis exist, but the prototypic form of XX gonadal dysgenesis not associated with somatic anomalies is clearly inherited in autosomal recessive fashion. Affected individuals are normal in stature (mean height, 165cm) [54], and Turner stigmata are usually absent. Frequent reports of consanguinity have long made it clear that autosomal recessive genes are responsible. More recent segregation analysis by the author and colleagues revealed the segregation ratio to be 0.16 for female sibs. Thus, two-thirds of gonadal dysgenesis cases in 46,XX individuals are genetic [55]. The one-third of cases that are nongenetic (phenocopies) could be due to infection, infarction, infiltrative, or autoimmune phenomena. Of considerable clinical interest is the variable expressivity. In some families one sib has streak gonads, whereas another affected individual had primary amenorrhea and extreme ovarian hypoplasia (presence of a few oocytes) [5359]. If the mutant gene responsible for XX gonadal dysgenesis is capable of variable expression, the gene may be responsible for some sporadic cases of premature ovarian failure. 2. MECHANISM OF GENE ACTION
The mechanism underlying failure of germ cell persistence in most forms of XX gonadal dysgenesis is unknown, but several hypotheses seem reasonable. One is perturbation of meiosis, a general mechanism already invoked to explained germ cell breakdown in both monosmy X and balanced chromosomal translocations. In plants and lower mammals meiosis is known to be under genetic control. Surely this is true in humans as well, for which reason one would predict existence of mutations that would be manifested as ovarian failure and infertility in otherwise normal women. Other possibilities include interference with germ cell migration, abnormal connective tissue milieu, or gonadotropin receptor perturbation (see later). Table III lists
86 TABLE III
JOE LEIGH SIMPSON
Mouse Genes Affecting Germ Cell Number or Gametogenes
Gene
Description/function
gcd
Germ cell deficiency
dhh
Desert hedge hog/protein signaling
BMP 8B
Bone morphogenetic protein 8B/TGF-/3 family
Dazla
RNA-binding protein
Igf-I
Insulin-like growth factor-I
Nblb2
Basic helix-loop transcription factor
Ztx
Zinc-finger protein
Hsp 70-2, Hsc 7 O t p
Heat-shock protein
Mhl, Pms2
Mismatch repair
ATM
Cell check point
Dazla
RNA binding (premeiotic)
as 50 to 100 families should identify chromosomal region(s) worthy of sequencing. This method was, in fact, applied successfully in Finland to elucidate the form of XX gonadal dysgenesis due to follide-stimulating hormone (FSH) receptor mutation (see later.)
B. P e r r a u l t S y n d r o m e ~ X X
Gonadal Dysgenesis
with Neurosensory Deafness A distinct variant of XX gonadal dysgenesis is that associated with neurosensory deafness. This condition is called Perraut syndrome. Like XX gonadal dysgenesis without deafness, Perrault syndrome is autosomal recessive [54-64].
C. F S H R e c e p t o r M u t a t i o n some genes in mice and other species that deleteriously affect germ cell development or gametogenesis. Attractiveness of the human homologs of these genes as explanations for XX gonadal dysgenesis is underscored by the phenotype of various murine "knockout" models. Often the only abnormality is ovarian or testicular failure or abnormalities restricted to gametogenesis. This is true even though these murine and Drosophila genes would have been predicted to act in ways seemingly disparate from those expected of genes affecting germ cells or gametogenesis. These genes may affect cell cycle checkpoints, DNA repair genes, or heat-shock proteins (which as chaperone proteins, protect unoccupied steroid receptors). Usually, meiotic breakdown occurs at or before the pachytene stage. That germ cells are affected is not a surprise, but that abnormalities seem restricted to the reproductive system is. The model for the murine gene germ cell deficiency (gcd) [60] is especially attractive. This murine autosomal recessive gene produces decreased numbers of germ cells in both ovary and testes. Its human homolog could be responsible for some cases of XX gonadal dysgenesis. Even more plausibly, gcd could be responsible for the syndrome of germ cell deficiency in both sexes (see later). In the absence of candidate genes such as those cited above, identifying autosomal genes responsible for the various forms of XX gonadal dysgenesis is more difficult. An investigator might await the fortuitous family in which an autosomal translocation cosegregates with XX gonadal dysgenesis. Sporadic cases of gonadal dysgenesis have long been associated with reciprocal autosomal translocations, but for years there seemed to be little consistency in the chromosome involved. The alternate approach is a "brute force" genome-wide search for relevant gene(s), utilizing sib-pair analysis, with the polymorphic DNA markers readily available throughout the genome. Using sib-pair analysis, as few
In Finland Aittomaki et al. [58,59] searched hospitals and cytogenetic labs to identify 75 patients country-wide having XX gonadal dysgenesis, defined in 46,XX women as primary or secondary amenorrhea and serum FSH >- 40 mlU/ ml. These 75 included 57 sporadic cases and 18 cases having affected relatives (seven different families). Most cases were found in north central Finland, a sparsely populated part of the country. The overall frequency of the disorder in Finland was 1 per 8300 liveborn females, a relatively high incidence attributed to a founder effect. Segregation ratio of 0.23 for female sibs was consistent with autosomal recessive inheritance, as was the consanguinity rate of 12%. Sib-pair analysis using polymorphic DNA markers were next used to localize the gene to a specific region. Chromosome 2p, a region that had previously been known to contain genes for both the FSH receptor (FSHR) and the luteinizing hormone (LH) receptor (LHR). One specific mutation (C566T:alanine to valine) in exon 7 was observed in six multiplex families [59,65]. That C566T was not found in all Finnish XX gonadal dysgenesis cases indicates genetic heterogeneity. The C566Tnegative cases could represent the same disorder discussed previously (XX gonadal dysgenesis with no somatic anomalies). Consistent with this is that the C566T mutation is rarely detected in samples from women with 46,XX ovarian failure who reside outside Finland [66]. Layman et al. [66] found no mutations in the FSHR gene in 35 46,XX women having hypergonadotopic hypogonadism. Of the 35, 15 had primary amenorrhea and 20 had secondary amenorrhea. Three of the 35 had one or more affected sisters. Liu et al. [67] found no sequence abnormalities in one multigeneration primary ovarian failure (POF) family, 4 sporadic POF cases and 2 cases of hypergonadotopic hypogonadism cases. Irrespective of apparently being found mostly in Finland,
CHAPTER5 Genetic Programming in Ovarian Development TABLE IV Phenotypes of Finnish XX Gonadal Dysgenesis, with and without the C566T Mutation a Measure
C566T
Wild-type566
Primary amenorrhea Height (cm) Delayed puberty FSH (IU/liter) Follicles (ultrasound)
22/22 160.5 6/16 61 6/12
22/30 165.8 10/20 84 1/11
aData of Aittomakiet al. [65].
it was a surprise to many that at least one form of XX gonadal dysgenesis was caused by mutation involving FSHR. It had been expected that ovaries associated with gonadotropin receptor mutations would be characterized by numerous but undeveloped primordial follicles, not streak gonads. This is the phenotype said to exist in the so-called Savage syndrome, considered to be due to a gonadotropin-resistance syndrome. Aittomaki et al. [65] later contrasted the phenotype of C566T XX gonadal dysgenesis with non-C566T XX gonadal dysgenesis. Subjects with the former, examined by ultrasound, were more likely to have ovarian follicles (Table IV). C566T XX gonadal dysgenesis thus showed some features expected for gonadotropin resistance; however, FSH was clearly elevated and the phenotype was far more reminiscent of the bilateral streak gonads and prototypal XX gonadal dysgenesis.
D. Inactivating L H R e c e p t o r M u t a t i o n s Another trophic hormone receptor gene in which a mutation causes gonadal dysgenesis is the LH receptor. Sultan and Lumbroso [68] tabulated the 46,XY cases, nine having female phenotype and two having micropenis. All 46,XX cases occurred in sibships in which an affected 46,XY male had Leydig cell hypoplasia. Latronico et al. [69] reported primary amenorrhea in a 22-year-old woman. In that family, three males and one female had a homozygous nonsense (stop) mutation at codon 554 (C554X). The newly produced stop codon resulted in a truncated protein having five rather seven transmembrane domains. The affected female had breast development but only a single episode of menstrual bleeding at age 20 years; LH was 37 mIU/ml; FSH, 9 mIU/ml. The mutation reduced signal transduction activity of the LH receptor gene. In a 46,XX case reported by Toledo e t al. [70], secondary amenorrhea occurred; LH and FSH were 10 and 9 mIU/ml, respectively. The mutations was Ala593Pro. Activating LH receptor mutations seem to have little effect in females, although in males precocious puberty occurs
87 [68]. No females with activating LHR have shown precocious puberty.
E. X X G o n a d a l D y s g e n e s i s a n d M u l t i p l e Malformation Syndromes Mutant genes that act on multiple organ systems are called pleiotropic. The pleiotropic combination of XX gonadal dysgenesis and neurosensory deafness, termed Perrault syndrome, has already received comment. Several other syndrome are recognized: XX gonadal dysgenesis and cerebellar ataxia [71]; XX gonadal dysgenesis, microcephaly, and arachnodactyly [72]; XX gonadal dysgenesis and epibulbar dermoid [73]; and XX gonadal dysgenesis, short stature and metabolic acidosis [74]. These four disorders are presumably all autosomal recessive in nature, based on multiple affected sibs. Contiguous gene syndrome (chromosomal deletions) or other non-Mendelian mechanisms are not excluded. No molecular progress has been made toward elucidating genes responsible for the syndromes described above, but the gene responsible for one autosomal dominant syndrome has been localized. The blepharophimosis-ptosis syndrome has long been recognized as associated with ovarian failure [75,76]. Using several large kindreds, sib-pair analysis using polymorphic DNA variants [77] localized the gene for blepharophimosis-ptosis to chromosome 3 (3q21-24). This region contains no obvious candidate genes. Puzzling with respect to the blepharophimosis-ptosis syndrome is the report of Fraser e t al. [78] showing that ovaries in blepharophimoisis-ptosis were unresponsive to gonadatropins. This is reminiscent of the phenotype associated with the Finnish XX gonadal dysgenesis C566T FSHR mutation [65]. In each of the multiple malformation syndromes associated with ovarian failure, the underlying biologic question must be posed. Does the seemingly pleiotropic gene cause both the somatic anomalies and the ovarian failure? Do the somatic and nonsomatic phenotypes involve only closely linked genes, i.e., contiguous gene syndrome? Could an unrecognized parental chromosomal rearrangements exist? In turn, do any of these genes play pivotal roles in normal ovarian differentiation and maintenance? Or, is perturbation of ovarian development merely secondary, perhaps occurring through generalized disturbance of connective tissue?
E F r a g i l e X and E x p a n s i o n o f the T r i p l e t Nucleotide Repeat CGG A special category of pleiotropic genes affecting ovarian development is those involving expansion of triplet nucleotide repeats. The prototype is the fragile X syndrome, caused
88
JOE LEIGH SIMPSON
by mutation of the F M R 1 gene on Xq27. "Fragile" refers to the tendency toward chromosomal breakage when affected Cells are grown in vitro in folic acid-deficient media. Several fragile sites exist, FRAXA and F R A X E in particular. In FRAXA, affected males show mental retardation, characteristic facial features, and large testes. The molecular basis involves presence of 230 or more (CGG) n repeats; the normal number of repeats in males is only 6-50. When heterozygous females show 50-200 repeats, a permutation is said to be present. During female (but not male) meiosis the number of triplet repeats may increase (expand). A woman with a FRAXA premutation may thus have an affected son if her X were to expand during meiosis to more than 230 repeats and if that son were to inherit that expanded rather than the normal X. Females may also be affected, but show a less severe phenotype than males. Females with the FRAXA premutation may show premature ovarian failure. Schwartz et al. [79] reported that fragile X carrier females more often showed oligomenorrhea than do noncarrier female relatives (38 vs. 6%). Premature ovarian failure was increased with FRAXA premutation (25 vs. 8%). Murray et al. [80] reported an extensive analysis of 1268 controls, 50 familial POF, and 244 sporadic POF cases. Table V shows the frequency of FRAXA in each. Of familial cases, 16% of showed FRAXA premutation; among sporadic cases, 1.6% showed POE In the same sample POF was not increased in F R A X E . Consistent with the above data are observations that FRAXA carrier women respond poorly to ovulation-inducing agents, thus producing fewer oocytes and fewer embryos in ART [81 ]. Although the consensus is that FRAXA is indeed associated with POF, Kenneson et al. [82] do not agree. They believe that the phenomenon is explained best by a contiguous gene syndrome m two separate but closely linked loci that may or may not be detected in a given individual or family. That Xq27-28 contains both F M R 1 ( F R A X A ) and an ovarian maintenance gene is consistent with, but does not prove, the hypothesis.
TABLE V Frequency of FRAXA Alleles on the X Chromosomes of 25 Women with Familial and 122 with Sporadic Premature Ovarian Failure a Phenotype
Allele
Number of CGG repeats
Familial POF
Sporadic POF
Normal
Minimal Common Intermediate Premutation Full mutation
0-10 11-40 41-60 61-200 >200
0 44 2 4 0
0 236 6 2 0
1 1237 30 0 0
50
244
1268
Total aData of Murray et al. [80].
The manner by which FRAXA premutation could produce POF remains obscure. That ovulation also seems difficult to induce, at least one other triplet nucleotide expansion disorder--myotonic dystrophymraises the possibility that nucleotide expansion per se is not salutary for ovarian development. However, no such correlation exists in Huntington disease, another triplet nucleotide expansion disorder. More importantly, women with complete FRAXA expansion do not always show POF. G. M y o t o n i c D y s t r o p h y a n d C T G Nucleotides Expansion Myotonic dystrophy is another nucleotide expansion disorder. It is an autosomal dominant disorder characterized by muscle wasting (head, neck, and extremities), frontal balding, cataracts, and male hypogonadism (80%) caused by testicular atrophy. Female hypogonadism is much less common and not well documented; however, it is usually assumed that age of menopause does not seem to be decreased. Pathogenesis involves nucleotide expansion of CTG, with (CTG),, occurring in the 3' untranslated region. Normally there are 5-27 CTG repeats. Heterozygotes usually have at least 50 repeats, and severely affected individuals have 600 or more. As in FRAXA, poor response to ovulation induction is observed. Sermon et al. [83] report fewer embryos per cycle than in standard assisted reproductive technologies (ART) and few pregnancies in preimplantation genetic diagnosis. Neither fragile X nor myotonic dystrophy need ordinarily be considered in the differential diagnosis of primary ovarian failure. However, these disorders occasionally are the explanation for POE H. G a l a c t o s e m i a In galactosemia enzyme deficiencies prevent synthesis of glucose from galactose. Several different enzymes and their mutant alleles, usually autosomal recessive in nature, may be responsible. Especially of interest is the form of galactosemia caused by deficiency of galactose-1-phosphate uridylyltransferase, controlled by a gene on 9p. In addition to renal, hepatic, and ocular damage, ovarian failure may be associated. Initially Kaufman et al. [84] reported 12 of 18 female cases had premature ovarian failure (POF). Waggoner et al. [85] later reported that 8 of 47 (17%) females with galactosemia had ovarian failure. Pathogenesis presumptively involves galactose toxicity after birth, given that maternal enzymes should be protective in utero. Consistent with this idea is that a neonate with galactosemia showed normal ovarian histology [86]. Given the clinical severity of galactosemia and necessity for childhood dietary treatment, it seems highly unlikely that undiagnosed galactosemia would prove to be the cause of
CHAPTER5 Genetic Programming in Ovarian Development ovarian failure in women presenting solely with primary amenorrhea or premature ovarian failure. Of greater general interest, therefore, was the report in 1989 by Cramer et al. [6] that GALT heterozygotes were at increased risk for POF. However, the same author later failed to observe GALT abnormalities in another sample of women with early menopause [5]; Kaufman et al. [87] likewise failed to confirm. Moreover, not all homozygotes for human galactosemia are abnormal, nor are even transgenic mice in which GALT is inactivated [88]. In summary, homozygous but probably not heterozygous galactose-l-phosphate uridylyltransferase deficiency is associated with ovarian failure.
I. D e f i c i e n c y o f 1 7 c e - H y d r o x y l a s e / 17,20-Desmolase Sex steroid synthesis requires intact adrenal and gonadal biosynthetic pathways. Various genes and their products (enzymes) are necessary to convert cholesterol to testosterone or androstenedione and, hence, to estrogens. An enzyme block may have varying consequences, depending on its site of action. The most common adrenal biosynthetic problem is female pseudohermaphroditism due to deficiency of 21- or 1 lfl-hydroxylase. However, ovarian development is normal in both these adrenal biosynthetic defects. If the enzyme 17ce-hydroxylase is not present (Fig. 5), neither androgens nor estrogens are synthesized. Thus, primary amenorrhea occurs in 46,XX cases. In 46,XY cases lack of androgens results in lack of virilization. However, the situation is complex because 17ce-hydroxylase and 17,20desmolase (lyase) are governed by a single gene on 10q. The gene C Y P 1 7 codes for a cytochrome P450 enzyme and is
I BIOSYNTHETIC
Acetate
89 located on 10q. About 150 cases have been reported, and most are male (46,XY) pseudohermaphrodites. Mutations have usually been point mutations rather than deletions or gene conversions. C Y P 1 7 mutations have been described that affect only 17,20-desmolase function [89,90]. In females, deficiency of 17ce-hydroxylase/17,20-desmolase (CYP17) should be considered a rare cause of hypergonadotropic hypogonadism. Ovaries are hypoplastic and sometimes streaklike in appearance. Oocytes appear incapable of reaching diameters greater than 2.5 mm [91 ]. However, stimulation with exogenous gonadotropins can mature follides to produce oocytes capable of fertilization in vitro [92]. Hypertension caused by hypervolemia occurs because of mineralocorticoid excess, an important diagnostic clue. If hypertension is not present, clinical presentation is similar to XX gonadal dysgenesis without somatic anomalies. Diagnosis is based on elevated progesterone, deoxycorticosterone, and corticosterone coupled with decreased testosterone and estrogen.
J. A r o m a t a s e M u t a t i o n s Conversion of androgens (A4-androstenedione) to estrogens (estrone) requires cytochrome P450 aromatase, an enzyme that is the gene product of a single 40-kb gene located on chromosome 15q21.1 [93]. The gene consists of 10 exons. Ito et al. [94] reported a mutation in this C Y P 1 9 (P450arom) gene in a 18-year-old 46,XX woman with primary amenorrhea and cystic ovaries. The patient was a compound heterozygote for two different point mutations in exon 10. The mutant protein had no activity. The above cases notwithstanding, deficiency of the aromatase enzyme is more often associated with genital
PATHWAYS ]
Cholesterol
Pregnenolone Progesterone
c
C
D 17 ~ - O H pregnenolone --~ Dehydroep~androsterone B JrB D E 17 c~ - OH progesterone ---~Androstenedione -l~Testosterone
II - deoxycorticosterone
II _ deoxycortisol
Corticosterone
Cortlsol
Estrone ~ E s t r a d l o l
Aldosterone
FIGURE 5 Importantadrenal and gonadal biosynthetic pathways. Letters designate enzymes required for the appropriate conversions. (A) 20a-Hydroxylase,22a-hydroxylase, and 20,22-desmolase; (B) 38fl-ol-dehydrogenase; (C) 17ce-hydroxylase; (D) 17,20-desmolase;(E) 17-ketosteroidreductase; (F) 21-hydroxylase;(G) 11-hydroxylase.(From Simpson [12].
90
JOE LEIGH SIMPSON
ambiguity. Shozu et al. [95] detected placental aromatase deficiency manifesting as maternal virilization during the third trimester. The 46,XX infant was born with genital ambiguity (female pseudohermaphroditism). Adrenal enzyme defects were not evident. The molecular basis of the mutation was an 87-bp insert in exon 6 of the aromatase gene, altering the splice junction site to produce a novel protein with 29 additional amino acids. Aromatase mutation in 46,XX female infants has been associated with genital ambiguity [94] or clitoromegaly [96]. In latter cases clitoral enlargement occurred at puberty, but breast development did not. Multiple ovarian follicular cysts were evident. FSH was elevated, estrone and estrodial were low. Estrogen and progesterone therapy resulted in a growth spurt, decreased FSH, decreased androstenedione and testosterone, breast development, menarche, and fewer follicular cysts. Molecular studies demonstrated compound heterozygosity for CYP19 point mutations.
K. P o l y g l a n d u l a r A u t o i m m u n e S y n d r o m e and O v a r i a n A n t i b o d i e s Type I polyglandular autoimmune syndrome is an autosomal recessive disorder especially common in the Finnish and Iranian Jewish populations [97]. Affected individuals show moniliasis due to immune deficiencies, and deficiencies of the parathyroids, adrenals, and gonads exist. The frequency of hypergonadotropic ovarian failure is 60%. The gene is localized to 2q22.3 [98]. The murine homolog is the autoimmune regulation gene (AIRE). Type I polyglandular autoimmune syndrome ovarian failure is uncommon outside the Finnish and Iranian Jewish populations. Type II polyglandular autoimmune syndrome, also called Schmidt syndrome, is heterogenesis. Inheritance is autosomal dominant. Gonadal failure occurs, as well as adrenal, thyroid, and pancreatic hypofunction. Systemic defects involve the hematological, gastrointestrial, and ocular systems and the tegument (hair). Immunologic dysfunction is often pronounced and moniliasis common. HLA-B8 is associated, as are, to a lesser extent, DR3 and DR4 [99]. A related issue is the considerable literature concerning a potential relationship between POF and antiovarian antibodies. These antibodies may be either of generalized nature or specific against a single cellular component (e.g., gonadotropin receptor, stromal cells, zona pellucida). Antiovarian antibodies as a cause of ovarian failure were reviewed by Anasti [100]. To this author, a casual relationship seems less likely than a secondary effect (epiphenomenon), arising only after damage has occurred for unrelated reasons. Even if causative, POF and not primary amenorrhea would be expected. Similar reasoning applies as well to oophoritis and ovarian failure.
L. G e r m Cell Failure in B o t h Sexes In several sibships, male and female sibs have each shown germ cell failure. Affected females show streak gonads, whereas males show germ cell aplasia (Sertoli cell-~6nly syndrome, or del Castillo phenotype). In two families, parents were consanguineous, and in each no somatic anomalies coexisted [ 101,102]. In three other families, coexisting patterns of somatic anomalies suggested distinct entities. Hamet et al. [103] reported germ cell failure, hypertension, and deafness; A1-Awadi et al. [104] reported germ cell failure and alopecia; Mikati et al. [105] reported germ cell failure, microcephaly, and short stature. These families demonstrate that the same autosomal gene is capable of deleteriously affecting germ cell development in each sex, presumably acting at a site common to early germ cell development. Elucidating such genes could have profound implications for understanding normal germ cell development and gametogenesis. To this end the noteworthy murine model for gcd (germ cell deficiency) is worth recalling. This autosomal recessive disorder is characterized by germ cell deficiency in both male and female mice.
M. 4 6 , X X A g o n a d i a In agonadia, the gonads are absent, the external genitalia are abnormal, and all but rudimentary Mtillerian or Wolffian derivatives are absent. Almost all affected individuals are 46,XY. External genitalia usually consist of a phallus about the size of a clitoris, underdeveloped labia majora, and nearly complete fusion of labioscrotal folds. A persistent urogenital sinus is often present. By definition, gonads cannot be detected, and neither normal Mtillerian derivatives nor normal Wolffian derivatives are ordinarily present. Rudimentary structures may be present along the lateral pelvic wall. In about one-half the cases, somatic anomalies coexist, i.e., craniofacial and vertebral anomalies and mental retardation [ 106]. Because almost all cases are 46,XY, pathogenic explanations have always focused on loss of testes early in embryogenesis. Any pathogenic explanation for agonadia in 46,XY individuals must take into account not only absence of gonads (usually testes), but also abnormal external genitalia and lack of internal genital ducts. Two general explanations have been invoked: (1) Fetal testes functioned long enough to complete male differentiation, but then disappeared, hence the synonymous designation "testicular regression syndrome." (2) Gonadal, ductal, and genital systems concomitantly developed abnormally as result of either defective anlagen, defective connective tissue, action of a teratogen, or other mechanisms. Given existence of both heritable tendencies [107] and frequent coexistence of somatic anomalies, defective connective tissue has seemed plausible
CHAPTER5 Genetic Programming in Ovarian Development in certain cases. Alternatively, cases with and without somatic anomalies could be etiologically distinct (genetic heterogeneity). Of interest in the present context of causes of primary amenorrhea, are there even rarer reports of 46,XX agonadia? Sporadic cases were reported by Duck et al. [108] and Levinson et al. [ 109]. Mendonca et al. [ 110] reported agonadia without somatic anomalies in phenotypic sibs of unlike chromosomal complements (46,XY and 46,XX). Kennerknecht et al. [111] reported agonadism, hypoplasia of the pulmonary artery and lung, and dextrocardia in XX and XY sibs.
91 pressivity. Thus, the autosomal recessive mutations responsible for certain forms of XX gonadal dysgenesis may be manifested as less severe ovarian pathology. In Finnish cases ascertained by Aittomaki [58,59], POF also not infrequently coexisted in the same kindred as complete ovarian failure (COF). This applied to both the FSH receptor mutation cases (C566T) as well as the non-C566T cases. XX gonadal dysgenesis genes may therefore be responsible for familial premature ovarian failure.
C. A u t o s o m a l D o m i n a n t P O F
XI. TO WHAT EXTENT IS PREMATURE OVARIAN FAILURE G E N E T I C ? As already discussed, premature ovarian failure can result from several different genetic mechanisms. These include (1) X chromosomal abnormalities, (2) autosomal recessive genes causing the various types of XX gonadal dysgenesis, and (3) autosomal dominant genes whose action is restricted to POF. The first two mechanisms have already been considered in detail, so we shall focus here only the third. However, prior to doing so it is useful to recall the role that the former two etiologies play in POE
A. X C h r o m o s o m a l A b n o r m a l i t i e s Premature rather than complete ovarian failure is not rare in the X abnormalities. At least 10 to 15% of 45,X/46,XX individuals menstruate, compared to fewer than 5% of 45,X individuals [ 11]. The former percentage represents a minimum because many mosaic individuals are so mildly affected that they are never detected clinically. Spontaneous menstruation occurs in about half of all 46,X, del(X)(pl 1) and 46,X, del(X)(p21 or 22) individuals, who often present with secondary amenorrhea and premature ovarian failure. Deletions or X/autosomal translocations involving regions Xp22 and Xq26 are more likely to be associated with premature ovarian failure than with complete ovarian failure. Recall also that women with the FRAXA premutation ( F M R 1 ) show an increased frequency of premature ovarian failure, a phenomenon that may or may not be the result of perturbations of the terminal Xq ovarian maintenance genes.
B. A u t o s o m a l R e c e s s i v e P O F In some families we have noted that the propositus may have 46,XX gonadal dysgenesis and streak gonads, with a sib having only ovarian hypoplasia with some oocytes. These sibships suggest that the mutant gene responsible for XX gonadal dysgenesis is capable of exerting variable ex-
An entity different from the above is ostensibly idiopathic premature ovarian failure transmitted in more than one generation [ 112,113]. This suggests autosomal dominant inheritance, although in some of these families subtle X deletions or FRAXA premutations could have existed. This mechanism has already been noted to exist in the blepharophimosisptosis syndrome. Mattison et al. [114] studied five families. That these families were probably ascertained from a very large population base raises the concern that the familial aggregates could have been observed by chance or on the basis of polygenic factors. In none were ovarian antibodies present. Useful information on idiopathic autosomal dominant POF is coming from Italian investigators. Defining POF as cessation of menses for 6 months or longer, Vegetti et al. [115] studied 81 women under age 40 years. After excluding 10 cases of presumptively known etiology (5 abnormal karyotypes, 3 previous ovarian surgery, 1 prior chemotherapy, 1 galactosemia), pedigree analysis was performed. Of the remaining 71 cases, 23 (31%) had an affected female relative. The median age of subjects with a positive family history was older (37.5 yr) than those without such a history (31 yr). Family members were affected in a pattern consistent with autosomal dominant inheritance. Transmission occurred through both male and female members. Neither blepharophimosis-ptosis syndrome nor fragile X were observed. In the Italian study several different genes could be involved, but the familial nature of POF seems firmly established.
References 1. Jir~isek,J. (1976). Disorders of sexual differentiation. In "Principles of Reproductive Embryology" (J. L. Simpson, ed.), pp. 51-111. Academic Press, New York. 2. Cussen, L. K., and McMahon, R. (1979). Germ cells and ova in dysgenetic gonads of a 46,XY female dizygote twin. Arch. Dis. Child. 133, 373-375. 3. Greenberg,F., Greesick, M. V., Carpenter, R. J., Law, S. W., Hoffman, L. E, and Ledbetter, D. H. (1987). The Gardner-Silengo-Wachtel syndrome: Male pseudohermaphroditism with micrognathia, cleftpalate, and conotruncal cardiac defect. Am. J. Hum. Genet. 26, 59-64.
92
JOE LEIGH SIMPSON
4. Evans, E. E, Ford, C. E., and Lyon, M. F. (1977). Direct evidence of the capacity of the XY germ cell in the mouse to become an oocyte. Nature (London) 267, 430-431. 5. Cramer, D. W., Xu, H., and Harlow, B. L. (1995). Family history as a predictor of early menopause. Fertil. Steril. 64, 740-745. 6. Cramer, D. W., Harlow, B. L., Barbieri, R. L., and Ng, W. G. (1989). Galactose-l-phosphate uridyl transferase activity associated with age at menopause and reproducutive history. Fertil. Steril. 51, 609-615. 7. Torgerson, D. J., Thomas, R. E., and Reid, D. M. (1997). Mothers' and daughters' menopausal ages: Is there a link? Eur. J. Obstet. Gynecol. Reprod. Biol. 74, 63-66. 8. Treloar, S. A., Do, K. A., and Martin, N. G. (1998). Genetic influences on the age at menopause. Lancet 352, 1084-1085. 9. Snieder, H., MacGregor, A. J., and Spector, T. D. (1998). Genes control the cessation of a woman's reproductive life: A twin study of hysterectomy and age at menopause. J. Clin. Endocrinol. Metab. 83,
26.
27.
28.
29.
1875-1880.
10. Treloar, S. A., and Martin, N. G. (1990). Age at menarche as a fitness trait: Nonadditive genetic variance detected in a large twin sample. Am. J. Hum. Genet. 47, 137-148. 11. Simpson, J. L. (1975). Gonadal dysgenesis and abnormalities of the human sex chromosomes: Current status of the phenotypic-karyotypic correlations. Birth Defects, Orig. Artic. Ser. 11(4), 23-59. 12. Simpson, J. L. (1976)."Disorders of Sexual Differentiation: Etiology and Clinical Delineation," pp. 259-302. Academic Press, New York. 13. Loughlin, S. A. R., Redha, A., McIver, J., Boyd A, Carothers, A., and Connor, J. M. (1991). Analysis of the origin of Turner's syndrome using polymorphic DNA probes. J. Med. Genet. 28, 156-158. 14. Mathur A., Stekol, L., Schatz, D., MacLaren, N. K., Scott, M. L., and Lippe, B. (1991). The parental origin of the single X chromosome in Turner syndrome: Lack of correlation with prenatal age or clinical phenotype. Am. J. Hum. Genet. 48, 682- 686. 15. Lorda-Sanchez, I., Binkert, E, Maechler, M., and Shinzel, A. (1992). Molecular study of 45,X conceptuses: Correlation with clinical findings. Am. J. Med. Genetic. 42, 487-490. 16. James, R. S., Dalton, P., Gustashaw, K., Wolff, D. J., Willard, H. E, Mitchell, C., and Jacobs, P. A. (1997). Molecular characterization of isochromosomes of Xq. Ann. Hum. Genet. 61,485-490. 17. James, R. S., Coppin, B., Dalton, P., Dennis, N. R., Mitchell, C., Sharp, A. J., Skuse, D. H., Thomas, N. S., and Jacobs, P. A. (1988). A study of females with deletions of the short arm of the X chromosome. Hum. Genet. 102, 507-516. 18. Ogata, T., and Matsuo, N. (1995). Turner syndrome and female sex chromosome aberrations: Deduction of the principal factors involved in the development of clinical features. Hum. Genet. 95, 607-629. 19. Speed, R. M. (1986). Oocyte development in XO foetuses of man and mouse: The possible role of heterologous X-chromosome pairing in germ cell survival. Chromosoma 94, 115-124. 20. Speed, R. M. (1988). The possible role of meiotic pairing anomalies in the atresia of human fetal oocytes. Hum .Genet. 78, 260-266. 21. Migeon, B. R., and Jelalian, K. (1977). Evidence for two active X chromosomes in germ cells of female before meiotic entry. Nature (London) 269, 242-243. 22. Simpson, J. L. (1981). Pregnancies in women with chromosomal abnormalities. In "Genetic Disease in Pregnancy" (J. D. Schulman and J. L. Simpson, eds.), p. 493. Academic Press, New York. 23. Dewhurst, J. (1978). Fertility in 47,XXX and 45,X patients. J. Med. Genet. 15, 132. 24. Singh, D. N., Hara, S., Foster, H. W., and Grimes, E. M. (1980). Reproductive performance in women with sex chromosome mosaicism. Obstet. Gynecol. 55, 608-611. 25. Ishikawa, M., Hidaka, E., Wakui, K., Nakayama, K., Takagi, Y., Katsuyama, T., Fukushima, Y., and Nikawa, N. (1998). Habitual abortion and low frequent X chromosome monosomy mosaicism: Detection by
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40. 41. 42.
43.
interphase FISH analyses of buccal mucosa cells and lymphocytes. 48th Annual Meeting, American Society of Human Genetics. Am. J. Hum. Genet. 63, A 108. Hook, E. B., and Hamerton, J. L. (1977). The frequency of chromosome abnormalities detected by consecutive newborn studies; differences between studies-results by sex and by severity of phenotypic involvement. In "Population Cytogenetic Studies in Human" (E. B. Hook and I. H. Porter, eds.), pp. 63-72, Academic Press, New York. Simpson, J. L. (1987). Genetic control of sexual development. In "Fertility and Sterility Series" (E. S. Teoh, S. S. Ratnam, and V. H. H. Goh, eds.), pp. 165-173, Vol. 3. Parthenon Press, Lancaster, England. Simpson, J. L. (1997). Ovarian maintenance determinants on the X chromosome and on autosomes. In "New Horizons in Reproductive Medicine," (C. Coutifaris and L. Mastroianni, eds.), pp. 439-444. Parthenon Press, Lancaster, England. Simpson, J. L. (1998). Genetics of female infertility. In "Treatment of Infertility: The New Frontiers" (M. Filicori and C. Flamigni, eds.), Communications Media for Education, Boca Raton, FL. pp. 37-52. Simpson, J. L. (1998). Genetics of oocyte depletion. In "Perimenopause" (R. A. Lobo, ed.), Serono Symp., pp. 36-45. Springer, New York. Fraccaro, M., Maraschio, E, Pasquali, F., and Scappaticci, S. (1977). Women heterozygous for deficiency of the (Xpter ---) X21) region of the X chromosome are fertile. Hum. Genet. 39, 283-292. Simpson, J. L., and LeBeau, M. M. (1981). Gonadal and statural determinants on the X chromosome and their relationship to in vitro studies showing prolonged cell cycles in 45,X,46,X, del(X)(pll), 36,X, del(X)(ql3) and q(22) fibroblasts. Am. J. Obstet. Gynecol. 141, 930-939. Wilson, M. G., Modebe, O., Towner, J. W., Frasier, S. D., and Lin, M. S. (1983). Ullrich-Turner syndrome associated with interstitial deletion of Xpl 1.4 leads to p22.31. Am. J. Med. Genet. 14, 567-576. Herva, R. B., Kaluzewski, B., and de la Chapelle, A. (1979). Inherited interstitial del(Xp) with minimal clinical consequences: With a note on the location of genes controlling phenotypic features. Am. J. Med. Genet. 3, 43-58. Sala, C., Arrigo, G., Torri, G., Martinazzi, E, Riva, P., Larizza, L., Philippe, C., Jonveaux, E, Sloan, E, Labella, T., and Toniolo, D. (1997). Eleven X chromosome breakpoints associated with premature ovarian failure (POF) map to a 15-Mb YAC contig spanning Xq21. Genomics 40, 123-131. Philippe, C., Arnould, C., Sloan, E, vanBokhoven, H., van der VeldeVisser, S. D., Chery, M., Ropers, H. H., Gilgenkrantz, S., Monaco, A. P., and Cremers, E M. (1995). A high-resolution interval map of the q21 region of the human X chromosome. Genomics 27, 539-543. Willard, H. F., Cremers, F. P., Mandel, J. L., Monaco, A. E, Nelson, D. L., and Schlessinger, D. (1994). Report of the fifth international workshop on human X chromosome mapping. Cytogenet. Cell Genet. 67, 295-358. Sargent, C. A., Briggs, J., Chalmers, I. J., Lambson, B., Walker, E., and Affara, N. A. (1996). The sequence organization of Yp/proximal Xq homologous regions of the human sex chromosomes is highly conserved. Genomics 32, 200-209. Bottazzo, G. E, McIntosh, C., Stanford, W., and Preece, M. (1980). Growth hormone cell antibodies and partial growth hormone deficiency in a girl with Turner's syndrome. Clin. Endocrinol. 12, 1-9. Sarto, G. E., Therman, E., and Patau, K. (1973). X inactivation in man: A woman with t ( X q - ; 12q+). Am. J. Hum. Genet. 25, 262-270. Therman, E., Laxova, R., and Susman, B. (1990). The critical region on the human Xq. Hum. Genet. 85, 455-461. Therman, E., and Susman, B. (1990). The similarity of phenotypic effects caused by Xp and Xq deletions in the human female: A hypothesis. Hum. Genet. 85, 175-183. Krauss, C. M., Turkray, R. N., Atkins, L., McLaughlin, C., Brown,
93
CHAPTER 5 Genetic Programming in Ovarian Development
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56. 57.
58. 59.
K. G., and Page, D. C. (1987). Familial premature ovarian failure due to interstitial deletion of the long arm of the X chromosome. N. Engl. J. Med. 317, 125-131. Fitch, N., de Saint, V. J., Richer, C. L., Pinsky, L., and Sitahal, S. (1982). Premature menopause due to small deletion in long arm of the X chromosome: A report of three cases and a review. Am. J. Obstet. Gynecol. 142, 968-972. Tharapel, A. T., Anderson, K. E, Simpson, J. L., Martens, E R., Wilroy, R. S., Llerena, J. C., and Schwartz, C. E. (1993). Deletion (X) (q26.1 ~ q28) in a proband and her mother: Molecular characterisation and phenotypic-karyotypic deductions. Am. J. Hum. Genet. 52, 463-471. Jones, M. H., Furlong, R. A., Burkin, H., Chalmers, I. J., Brown, G. M., Khwaja, O., and Affara, N. A. (1996). The Drosophila developmental gene fat facets has a human homologue in Xpl 1.4 which escapes X-inactivation and has related sequences on Ypl 1.2. Hum. Mol. Genet. 5, 1695-1701. Castrillon, D. H., and Wasserman, S. A. (1994). diaphanous is required for cytokinesis in Drosophila and shares domains of similalrity with the products of the limb deformity gene. Development (Cambridge, UK) 120, 3367-3377. Philippe, C., Cremers, F. E M., Chery, M., Bach, I., Abbadi, N., Ropets, H. H., and Gilgenkrantz, S. (1993). Physicial mapping of DNA markers in the q13-q22 region of the human X chromosome. Genomics 17, 147-152. Bione S, Sala C, Manzini C, Arrigo, G., Zuffardi, O., Banff, S., Borsani, G., Jonveaux, E, Philippe, C., Zuccotti, M., Ballabio, A., and Toniolo, D. (1998). A human homologue of the Drosophilia melanogaster diaphanous gene is disrupted in a patient with premature ovarian failure: Evidence for conserved function in oogenesis and implications for human sterility. Am. J. Hum. Genet. 62, 533-541. Marozzi, A., Furlan, D., Manfredini, E., Tibiletti, M. G., Vegetti, W., Testa, G., and Dalpra, L. (1998). Diaphanous gene disruption in a woman with Xq deletion and secondary hypergonadotropic amenorrhea. 48th Annual Meeting, The American Society of Human Genetics. Am. J. Hum. Genet. 63, A l l 4 (Abstr. No. 812). Van de Velde-Staquet, E., Breynaert, R., Walbaum, R., Saint-Aubert, E, Farriaux, J. E, and Fontaine, G. (1973). La descendeance des mere trisomiques 21. Genet. Hum. 21, 187-206. Assche, E. V., Bonduella, M., Tournaye, H., Joris, H., Verheyen, G., Devroey, E, Van Steirteghem, A., and Liebaers, I. (1996). Cytogenetics of infertile men. Hum. Reprod. 11, 1-24. Simpson, J. L., Christakos, A. C., Horwith, M., and Silverman, F. (1971). Gonadal dysgenesis associated with apparently chromosomal complements. Birth Defects, Orig. Artic. Ser. 7, 215-228. Simpson, J. L. (1979). Gonadal dysgenesis and sex chromosome abnormalities. Phenotypic/karyotypic correlations. In "Genetic Mechanisms of Sexual Development" (H. L. Vallet and I. H. Perter, eds.), pp. 365-405. Academic Press, New York. Meyers, C. M., Boughman, J. A., Rivas, M., Wilroy, R. S., and Simpson, J. L. (1996). Gonadal dysgenesis in 46,XX individuals: Frequency of the autosomal recessive form. Am. J. Med. Genet. 63, 518524. Boczkowski, K. (1970). Pure gonadal dysgenesis and ovarian dysplasia in sisters. Am. J. Obstet. Gynecol. 106, 626-628. Portuondo, J. A., Neyro, J. L., Benito, J. A., de la Rios, A., and Barral, A. (1987). Familial 46,XX gonadal dysgenesis. Int. J. Fertil. 32, 5 6 58. Aittomaki, K. (1994). The genetics of XX gonadal dysgenesis. Am. J. Hum. Genet. 54, 844-851. Aittomaki, K., Dieguez Luccena, J. L., Pakarinen, E, Sistonen, E, Tapanainen, J., Gromoll, J., Kaskikari, R., Sankila, E. M., Lehvaslaiho, H., Engel, A. R., Nieschlag, E., Huhtaniemi, I., and de la Chapelle, A. (1995). Mutation in the follicle-stimulating hormone receptor gene
60.
61.
62.
63. 64.
65.
66.
67.
68.
69.
70.
71.
72.
73. 74.
75.
76.
77.
78.
causes hereditary hypergonadotropic ovarian failure. Cell (Cambridge, Mass.) 82, 959-968. Pellas, T. C., Ramachandran, B., Duncan, M., Pan, S. S., Marone, M., and Chada, K. (1991). Germ-cell deficient (gcd), an insertional mutation manifested as infertility in transgenic mice. Proc. Natl. Acad. Sci. U.S.A. 88, 8787-8791. Christakos, A. C., Simpson, J. L., Younger, J. B., and Christian, C. B. (1969). Gonadal dysgenesis as an autosomal recessive condition. Am. J. Obstet. Gynecol. 104, 1027-1030. Pallister, E D., and Opitz, J. M. (1979). The Perrault syndrome: Autosomal recessive ovarian dysgenesis with facultative, non sex-limited sensorineural deafness. Am. J. Med. Genet. 22, 629-631. McCarthy, D. J., and Opitz, J. M. (1985). Perrault syndrome in sisters. Am. J. Med. Genet. 22, 629-631. Nishi, Y., Hamamoto, K., Kajiyama, M., and Kawamura, I. (1988). The Perrault syndrome: Clinical report and review. Am. J. Med. Genet. 31, 623-629. Aittomaki, K., Herva, R., Stenman, U. H., Juntunen, K., Ylostalo, E, Hovatta, O., and de la Chapelle, A. (1996). Clinical features of primary ovarian failure caused by a point mutation in the folliclestimulating hormone receptor gene. J. Clin. Endocrinol. Metab. 81, 3722 -3726. Layman, L. C., Amede, S., Cohen, D. E, Jin, M., and Xie, J. (1998). The Finnish follicle-stimulating hormone receptor gene mutation is rare in North American women with 46,XX ovarian failure. Fertil. Steril. 69, 300-302. Liu, J. Y., Gromoll, J., Cedars, M. I., and La Barbera, A. R. (1998). Identification of allelic variants in the follicle-stimulating hormone receptor genes of females with or without hypergonadotropic amenorrhea. Fertil. Steril. 70, 326-31. Sultan, L. H. and Lumbroso, S. (1998). LH receptor defects. In "Fertility and Reproductive Medicine" (R. D. Kempers, J. Cohen, A. E Haney, and J. B. Younger, eds.), pp. 769-797. Elsevier, Amsterdam. Latronico, A. C., Anasti, J., Arnhold, I. J., Rapaport, R., Mendonca, B. B., Bloise, W., Castro, M., Tsigos, C., and Chrousos, G. E (1996). Brief report: Testicular and ovarian resistance to luteinizing hormone caused by inactivating mutations of the luteinizing hormone-receptor gene. N. Engl. J. Med. 334, 507-512. Toledo, S. E, Brunner, H. G., Kraaij, R., Post, M., Dahia, E L., Hayashida, C. Y., Kremer, H., and Themmen, A. E (1996). An inactivating mutation of the luteinizing hormone receptor causes amenorrhea in a 46,XX female. J. Clin. Endocrinol. Metab. 81, 3850-3854. Skre, H., Bassoe, H. H., Berg, K., and Frovig, A. G. (1976). Cerebella ataxia and hypergonadotropic hypogonadism in two kindreds. Chance concurrence, pleiotropism or linkage? Clin. Genet. 9, 234-244. Maximilian, C., Ionescu, B., and Bucur, A. (1970). Deux soeurs avec dysgenesie gonadique majeure, hypotrophic staturale, microcephalie, arachondactylie et caryotype 46,XX. J. Genet. Hum. 10, 261-280. Quayle, S. A., and Copeland, K. C. (1991) 46,XX gonadal dysgenesis with epibulbar dermoid. Am. J. Med. Genet. 40, 75-76. Pober, B. R., Zemel, S., and Hisama, F. M. (1998). 46,XX gonadal dysgenesis, short stature and recurrent metabolic acidosis in two sisters. 48th Annual Meeting, The American Society of Human Genetics, Am. J. Hum. Genet. 63, 652, A 117. Zlotogora J., Sagi, M., and Cohen, T. (1983). The blepharophimosis, ptosis, and epicanthus inversus syndrome: Delineation of two types. Am. J. Hum. Genet. 35, 1020-1027. Panidis, D., Rousso, D., Vavilis, D., Skiadopoulos, S., and Kalogeropoulos, A. (1994). Familial blepharophimosis with ovarian failure. Hum. Reprod. 9, 2034-2037. Harrar, H. S., Jeffrey, S., and Patton, M. A. (1995). Linkage analysis in blepharophimosis-ptosis syndrome confirms localisation to 3q2124. J. Med. Genet. 32, 774-777. Fraser, I. S., Shearman, R. E, Smith, A., and Russel, E (1988). An
94
79.
80.
81.
82.
83.
84.
85.
86. 87.
88. 89.
90.
91.
92.
93.
94.
95.
96.
JOE LEIGH SIMPSON association among blepharophimosis, resistant ovary syndrome, and true premature menopause. Fertil. Steril. 50, 747-751. Schwartz, C. E., Howard-Peebles, P. N., Bugge, M., Mikkelsen, M., Tommerup, N., Hull, C., Hagerman, R., Holden, J. J., and Stevenson, R. E. (1994). Obstetricial and gynecological complications in fragile X carriers: Multicenter study. Am. J. Med. Genet. 51, 400-402. Murray, A., Webb, J., Grimley, S., Conway, G., and Jacobs, P. (1998). Studies of FRAXA and FRAXE in women with premature ovarian failure. J. Med. Genet. 35, 637-640. Black, S. H., Levinson, G., Harton, G. L. Palmer, F. T., Sisson, M. E., Schoener, C., Nance, C., Fugger, E. F., and Fields, R. A. (1995). Preimplantation genetic testing (PGT) for fragile X (fraX). Am. J. Hum. Genet. 57, A31. Kenneson, A., Cramer, D. W., and Warren, S. T. (1997). Fragile X premutations are not a major cause of early menopause. Am. J. Hum. Genet. 61, 1362-1369. Sermon, K., De Vos, A., Van de Velde, H., Seneca, S., Lissens, W., Joris, H., Vandervorst, M., Van Steirteghem, A., and Liebaers, I. (1998). Fluorescent PCR and automated fragment analysis for the clinical application of preimplantation genetic diagnosis of myotonic dystrophy. Mol. Hum. Reprod. 4, 791-796. Kaufman, F. R., Kogut, M. D., Donnell, G. N., Goebelsmann, U., March, C., and Koch, R. (1981). Hypergonadotropic hypogonadism in female patients with galactosemia. N. Engl. J. Med. 304, 994-998. Waggoner, D. D., Buist, N. R., and Donnell, G. N. (1990). Long-term prognosis in galactosaemia: Results of a survey of 350 cases. J. Inherit. Metab. Dis. 13, 802-818. Levy, H. L., Driscoll, S. G., Porensky, R. S., and Wender D. F. (1984). Ovarian failure in galactosemia N. Engl. J. Med. 310, 50. Kaufman, F. R., Devgan, S., and Donnell, G. N. (1993). Results of a survey of carrier women for the galactosemia gene. Fertil. Steril. 60, 727-728. Leslie, N.D., Yager, K., and Bai, S. (1995). A mouse model for transferase deficiency galactosemia. Am. J. Hum. Genet. 57, 191, A38. Geller, D. H., Auchus, R. J., Mendonca, B. B., and Miller, W. L. (1997). The genetic and functional basis of isolated 17,20-1yase deficiency. Nat .Genet. 17, 201-205. Biason-Lauber, A., Leiberman, E., and Zachmann, M. (1997). A single amino acid substitution in the putative redox partner-binding site of P450c17 as cause of isolated 17,20-1yase deficiency. J. Clin. Endocrinol. Metab. 82, 3807-3812. Araki, S., Chikazawa, K., Sekiguchi, I., Yamauchi, H., Motoyama, M., and Tamada, T. (1987). Arrest of follicular development in a patient with 17 alpha-hydroxylase deficiency: Folliculogenesis in association with a lack of estrogen synthesis in the ovaries. Fertil. Steril. 47, 169172. Rabinovici, J., Blankstein, J., Goldman, B., Rudak, E., Dor, Y., Pariente, C., Geier, A., Lunenfeld, B., and Mashiach, S. (1989). In vitro fertilization and primary embryonic cleavage are possible in 17ahydroxylase deficiency despite extremely low intrafollicular 17/3estradiol. J. Clin. Endocrinol. Metab. 68, 693-697. Simpson, E. R., Michael, M. D., Agarwal, V. R., Hinshelwood, M. M., Bulun, S. E., and Zhao, Y. (1997). Cytochromes P450 11: Expression of the CYPI9 (aromatase) gene: An unusual case of alternative promoter usage. FASEBJ. 11, 29-36. Ito, Y., Fisher, C. R., Conte, E A., Grumbach, M. M., and Simpson, E. R. (1993). Molecular basis of aromatase deficiency in an adult female with sexual infantilism and polycystic ovaries. Proc. Natl. Acad. Sci. U.S.A. 90, 11673-11677. Shozu, M., Akasofu, K., Harada, T., and Kubota, Y. (1991). A new cause of female pseudohermaphroditism: Placental aromatase deficiency. J. Clin. Endocrinol. Metab. 72, 560-566. Mullis, P. E., Yoshimura, N., Kuhlmann, B., Lippuner, K., Jaeger, P., and Harada, H. J. (1997). Aromatase deficiency in a female who is compound heterozygote for two new point mutations in the P450
arom gene: Impact of estrogens on hypergonadotropic hypogonadism, multicystic ovaries, and bone densitometry in childhood. Clin. Endocrinol. Metab. 82, 1739-1745. 97. Ahonen, P., Myllarniemi, S., Sipila, I., and Perheentupa, J. (1990). Clinical variation of autoimmune polyendocrinopathy-candidiasisectodermal dystrophy (APECED) in a series of 68 patients. N. Engl. J. Med. 322, 1829-1836. 98. Aaltonen, J., Bjorses, P., Sandkuiji, L., Perheentupa, J., and Peltonen, L. (1994). An autosomal locus causing autoimmune disease: Autoimmune polyglandular disease type I assigned to chromosome 21. Nat. Genet. 8, 83-87. 99. Butler, M. G., Hodes, M. E., Conneally, P. M., Biegel, A. A., and Wright, J. C. (1984). Linkage analysis in a large kindred with autosomal dominant transmission of polyglandular autoimmune disease type II (Schmidt syndrome). Am. J. Med. Genet. 18, 61-65. 100. Anasti, J. N. (1998). Premature ovarian failure: An update. Fertil. Steril. 70, 1-15. 101. Smith, A., Fraser, I. S., and Noel, M. (1979). Three siblings with premature gonadal failure. Fertil. Steril. 32, 528-530. 102. Granat, M., Amar, A., Mor-Yosef, S., Brautbar, C., and Schenker, J. G. (1983). Familial gonadal germinative failure: Endocrine and human leukocyte antigen studies. Fertil. Steril. 40, 215 -219. 103. Hamet, P., Kuchel, O., Nowacynski, J. M., Rojo Ortega, J. M., Sasaki, C., and Genest, J. (1973). Hypertension with adrenal, genital, renal defects, and deafness. Arch. Intern. Med. 131, 563-569. 104. AI-Awadi, S. A., Farag, W. T. I., Teebie, A. S., Naguib, K., el-Khalifa, M. Y., Kelani, Y., AI-Ansari, A., and Schimke, R. N. (1985). Primary hypergonadotropic hypogonadism, short stature and minor anomalies. A new syndrome. Am. J. Med. Genet. 22, 619-622. 105. Mikati, M. A., Samir, S. N., and Sahil, I. F. (1985). Microcephaly, hypergonadotropic hypogonadism, short stature and minor anomalies. A new syndrome. Am. J. Med. Genet. 22, 599-608. 106. Sarto, G. E., and Opitz. J. M. (1973). The XY gonadal agenesis syndrome. J. Med. Genet. 10, 288-293. 107. de Gronchy, J., Gompel, A., Salomon-Bernard, Y., Juttenn, F., Yaneva, H., Paniel, J. B., Le Merrer, M., Roubin, M., Doussau de Bazignan, M., and Turleau, C. (1985). Embryonic testicular regression syndrome and severe mental retardation in sibs. Ann. Genet. 28, 154-160. 108. Duck, S. C., Sekkan, G. S., Wilbois R., Pagliara, A. S., and Weldon, V. V. (1975). Pseudohermaphroditism with testes and 46,XX karyotypes. J. Pediatr. 87, 58-62. 109. Levinson, G., Zarate, A., Guzman-Toledano, R., Canales, E. S., and Jimenez, M. (1976). An XX female with sexual infantilism, absent gonads, and lack of Mullerian ducts. J. Med. Genet. 13, 68-69. 110. Mendonca, B. B., Barbosa, A. S., Arnhold, I. J., McElreavey, K., Fellous, M., and Moreira-Filho, C. A. (1994). Gonadal agenesis in XX and XY sisters: Evidence for the involvement of an autosomal gene. Am. J. Med. Genet. 52, 39-43. 111. Kennerknecht, I., Sorgo, W., Oberhoffer, R., Teller, W. M., Mattfeldt, T., Negri, G., and Vogel, W. (1993). Familial occurrence of agonadism and multiple internal malformations in phenotypically normal girls with 46,XY and 46,XX karyotypes, respectively: A new autosomal recessive syndrome. Am. J. Med. Genet. 47, 1166-1170. 112. Starup, J., and Sele, V. (1972). Premature ovarian failure. Acta Obstet. Gynecol. Scand. 52, 259-268. 113. Austin, G. E., Coulam, C. B., and Ryan, J. R. (1979). A search for antibodies to luteinizing hormone receptors in premature ovarian failure. Mayo Clin. Proc. 54, 3945-4000. 114. Mattison, D. R., Evans, M. I., Schwinner, W. B., White, B. J., Jensen, B., and Schulman, J. D. (1984). Familial ovarian failure. Am. J. Hum. Genet. 36, 1341-1348. 115. Vegetti, W., Grazia Tibiletti, M., Testa, G., de Lauretis Yankowski, Alagna, F., Castoldi, E., Taborelli, M., Motta, T., Bolis, P. F., Dalpra, L., and Crosignani, P. G. (1998). Inheritance in idiopathic premature ovarian failure: Analysis of 71 cases. Hum. Reprod. 13, 1796-1800.
SHAPTER (
Neuroendocrlne " Reg ulat "on o f the Perimenopause Transition NANCY E.
I. II. III. IV.
REAME
Center for Nursing Research and Reproductive Sciences Program, The University of Michigan, Ann Arbor, Michigan 48109
V. The Neuroreproductive Axis in Postmenopause VI. Brain Aging and Reproductive Senescence VII. Future Studies References
Definition of Perimenopause Ovarian Determinants of Reproductive Aging Dynamic Gonadotropin Secretion in Young Women Gonadotropin Changes during Perimenopause
I. D E F I N I T I O N
year 2025, the number of postmenopausal women in the United States is expected to double [5]. Given the increasing life span of women in the United States, the number of years spent in the postmenopausal state is significant. Thus, the cessation of menses and the resulting hypoestrogenism may have important health consequences for the quality of life of a large and growing proportion of the population. Urinary continence, bone metabolism, cardiovascular function, memory and cognition, the synchrony of daily biorhythms, and the aging process itself have all been shown to be influenced by estrogen. Moreover, compared to other organ systems, the female reproductive system is unique in that it undergoes spontaneous failure at a relatively young age, thus making it an excellent model for the study of the aging process free of chronic disease. The concept of the perimenopause was first introduced by Treolar and colleagues in 1967 [6] when they conducted a cross-sectional analysis of several hundred menstrual cycle calendars obtained from women across the reproductive life span. In that study, the critical marker of aging was menstrual irregularity, defined as a change in genital bleeding to
OF PERIMENOPAUSE Reproductive aging is a continuum that begins with a steep decline in fertility by age 35 years, long before the final menstruation (menopause) occurs at age 51 years on average [ 1]. The climacteric, which is that period of time when reproductive function declines, is associated with a discordant rise in follicle-stimulating hormone (FSH); this rise in FSH is viewed as monotropic, because luteinizing hormone (LH) levels remain normal. Progressive loss of regular menstrual cyclicity and depletion of responsive follicles from the ovary are also associated with the climacteric [2]. According to the definition of the World Health Organization [3], the perimenopause is the period immediately prior to menopause when endocrinological, biological, and clinical features of approaching menopause commence, continuing for at least the first year after menopause. As the year 2000 approaches, there will be some 35 million perimenopausal women in the United States, with half a million women added annually to the midlife population [4]. By the
MENOPAUSE: BIOLOGY AND PATHOBIOLOGY
95
Copyright9 2000by AcademicPress. Allrightsof reproductionin anyformreserved.
96
NANCY E. REAME
either longer or shorter flow intervals. However, menstrual cycle irregularity is now known to be a late feature of the perimenopause transition, brought about by neuroendocrine events that occur well before cyclic ovulation is disrupted. Premature menopause (premature ovarian failure) is a condition causing amenorrhea, hypoestrogenism, and elevated gonadotropins in women less than age 40 years [7]. In a survey of 2000 women [8], the age-specific incidence was 1 in 100 by age 40 years, and 1 in 1000 by age 30 years. Although most cases have an autoimmune etiology, some cases are idiopathic. In addition to the premature loss of fertility, health consequences for these young women appear to be similar to consequences for older hypoestrogenic women, and include higher cardiac risk and bone loss acceleration. The clinical picture of this disordered folliculogenesis has been likened to the natural perimenopausal stage of diminished ovarian reserve when transient fluctuations of follicular development and atresia occur prior to complete cessation of ovarian function. However, compared to the perimenopause, women with premature menopause show much higher FSH concentrations, perhaps due to the younger reproductive axis or, alternatively, due to less restraint from differences in inhibin [9]. Despite the inevitability of menopause for all women who live long enough, the events leading up to and following the last menses are highly variable in duration and magnitude from woman to woman, thus contributing to the wide range in age at menopause (ages 45-55 years). In a study of ovarian function in 17 women in this age range, the number of follicles present in the ovary varied from more than 1000 in those still cycling to a few hundred in perimenopause, and to none in the postmenopausal state [ 10]. Moreover, although the mean duration of the perimenopausal transition is approximately 3.5 years, nearly 10% of women will have no perimenopause transition whatsoever, but rather the persistence of regular cycles until an abrupt cessation of menstruation [ 11 ]. Such variability in the onset and regulation of the menopause experience suggests that its underlying etiology is multifactoral and dependent on a variety of external and internal influences. This chapter describes the evidence to date that aging effects at all three levels of the hypothalamic -pituitary- ovarian (HPO) axis may contribute to the pronounced hypersecretion of pituitary gonadotropins that occurs at the close of the reproductive years.
II. O V A R I A N
DETERMINANTS
OF REPRODUCTIVE
AGING
A. R i s i n g F S H Studies in the past decade have clarified that the hallmark sign of reproductive aging is the disproportionate increase in circulating FSH vs. LH, especially prominent in the early
follicular phase. This discordant increase in FSH occurs gradually across the midreproductive years, becoming pronounced in women over 40 years old [12,13]. The etiology of this rise of one gonadotropin but not the other has been the topic of investigation. It is now generally accepted that gonadotropin-releasing hormone (GnRH) is the stimulating factor for both LH and FSH and any divergence in their secretion can be explained by (1) differential sensitivities of LH and FSH to variations in the dose or frequency of pulsatile GnRH secretion, (2) the gonadal hormonal milieu, including sex steroid and nonsteroidal factors, and/or (3) alterations in the pituitary tone of activin and/or follistatin. In keeping with these assumptions, the prevailing view is that the monotropic increase in FSH results from a greater sensitivity to declining ovarian feedback on the hypothalamicpituitary-ovarian axis compared to LH. Because the increase in basal FSH occurs in conjunction with normal or even elevated levels of estradiol, it is presumed that diminished input from nonsteroidal factors, notably inhibin from the declining pool of follicles, is responsible for the dampened inhibitory tone of ovarian feedback [14,15].
B. A g e C h a n g e s in L H Evidence for a clear age-related increase in LH secretion in regular-cycling women has been more equivocal than for FSH. In a study of 94 subjects ages 24 to 50 years, there was a significant increase in mean FSH secretion detected by age 35 years, with no aging effects observed in LH secretion until age 45 years [ 12]. In another study by the same authors of 500 regularly cycling, infertile women, the rise in FSH levels was observed to begin as early as age 28 years [13]. In addition, a statistically significant increase in mean LH levels during the follicular phase could be detected by age 35 years, followed by a further increase in women older than age 40 years. Unlike prior investigations, the authors concluded that an increase in both FSH and LH concentrations occurred in women with regular ovulatory cycles quite early in reproductive life and could be used as the earliest endocrine markers of reproductive aging. The previously undetected rise in LH prior to age 40 years may have been uncovered in this cohort due to the very large sample size. The clinical relevance and mechanisms for such a subtle age-related change in LH concentrations await further study. The current view is that the regulation of LH secretion is relatively resistant to the incipient decline in ovarian reserve, with perhaps only a subtle rise in concentrations until just prior to menopause, when the ovarian pool of responsive follicles is dramatically reduced along with a steep decline in estradiol and gonadal inhibin [ 10]. The fact that ovariectomy of premenopausal women is associated with a dramatic increase (four- to sixfold) in LH levels within 1 - 4 weeks of surgery [ 16] has served as evidence that ovarian inhibition is
CHAPTER6 Perimenopause Neuroendocrine Regulation the major input to the GnRH-mediated regulation of LH secretion. It should be noted that in the absence of negative feedback signals (estradiol and inhibin) after ovariectomy, the discordance in gonadotropin secretion is particularly evident, with FSH levels rising earlier and remaining persistently higher than LH, suggesting that the unrestrained, endogenous GnRH pulse generator may directly or indirectly favor FSH secretion.
97
II []
Young Older
9
9
6
_J D 6 ~SIS
._1 -~
"r _1
3
3
0
~
0
C. G o n a d a l P e p t i d e s
ao ~
--~'l::~300 t
Reports now suggest that an age-related decline in ovarian inhibin-B is the primary trigger for the selective FSH rise in the menstrual cycles of older women [ 17] (see Chapter 2). However, to presume that a singular deficiency in inhibin activity could account for age-related elevations in FSH fails to consider the other component gonadal peptides known to mediate FSH regulation. The inhibin a, fiA, and fiB protein subunits are members of the transforming growth factor-fi family of peptides. Produced in the ovary and pituitary, they are encoded by distinct genes and dimerize to give rise to inhibin A (ce,fiA), inhibin B (ce,fiB), activin A (flA, fiA), activin AB (flA, fiB), and activin B (fiB, fiB). Inhibins and activins have opposing effects on FSH secretion: inhibins suppress FSH, and activins stimulate FSH production [ 18]. Follistatins, monomeric proteins distinct from both inhibins and activins, have functional overlap with inhibins in suppressing FSH release [19], through their action as binding proteins for both activins and inhibins [20]. Inhibin A (from the mature follicle and corpus luteum) and inhibin B (from small antral follicles) are secreted in reciprocal, mirror-image patterns of each other across the ovulatory menstrual cycle, acting in tandem to regulate the cyclic profile of FSH secretion [21 ]. Studies have now documented the negative influence of estradiol, inhibin, and follistatin, and the positive roles of activin and GnRH in regulating FSH release. With the development of highly specific assays for most of these gonadal peptides, investigations of the combined steroidal and nonsteroidal ovarian milieu during reproductive aging are now possible. Noteworthy is the finding that in postmenopausal women, circulating levels of both follistatin [22] and activin [23] are elevated. We and others have begun to test the hypothesis that changes in the overall tone of ovarian feedback may contribute to the age-related rise in FSH. When measured simultaneously, the secretory profiles of the inhibins and activin A have been demonstrated to be altered across the menstrual cycle of women over age 40 years. Compared to youngeraged controls, activin A demonstrates higher concentrations that do not change across the cycle, whereasthe overall inhibin tone is reduced [24,25]. The pool of developing follicles may be more sensitive to the age-related effects on in-
12
e"
.--_o200 I-
20
o~ tU 100
10
ID t2 ffl
0
0 G.
0
Follicular
Luteal
Follicular
Luteal
FIGURE 1 Mean concentrations of gonadotropins and sex steroids in young (mean age = 27.9 + 2 years) and older (mean age = 43.7 + 1 years) cycling women.., Group differences; LH, p < 0.05; FSH,p < 0.001. From Ref. 24, N. E. Reame, T. L. Wyman,D. J. Phillips, D. M. de Kretser, and V. Padmanabhan (1998). Net increase in stimulatory input resulting from a decrease in inhibin B and an increase in activin A may contribute in part to the rise in follicular phase follicle-stimulating hormone of aging cyclic women. J. Clin. Endocrinol. Metab. 83, 3302-3307. 9 The Endocrine Society.
hibin production than the corpus luteum, because the decline in inhibin B is already apparent in women as young as age 35 years, prior to a decline in FSH or luteal phase inhibin A [26]. Both total and and free follistatin concentrations in the periphery do not appear to be influenced by age or cycle phase [24] (Figs. 1 and 2). This altered proportion of dimers in older women has been explained by a reduction in inhibin a subunit available from the depleted ovarian pool to combine with the/3 subunits to form inhibins, thus favoring activin formation and, in turn, enhanced FSH secretion [25]. In a group of perimenopausal women with irregular cycles, the suppression of FSH during hormone replacement therapy was associated with a suppression of activin A in the face of unchanged inhibin B and follistatin [27,28]. Because activins appear capable of direct pituitary stimulation of FSH secretion and are produced in the pituitary [29], this finding is compatible with the view that estradiol's negative feedback action on FSH release may involve indirect, paracrine as well as direct, endocrine mechanisms. However, this hypothesis must await further investigation until more sensitive assays are developed for activin B and activin AB, as well as for measurement of the free (biologically active) form of activins. Moreover, it is not known whether the low levels at which these FSH regulators circulate in the peripheral blood are of sufficient magnitude to play an endocrine role.
98
NANCY E. REAME
InhibinA
Total inhibin a
Inhibin B
m YoungI ['~ Older___l
400 ._1
E O. v t'-
25 200 t-t'n
Free Follistatin
Total Follistatin ~-, 600 E <
.J
6~
U~3 c"
.c_
400
4
..~
co
._ ,,.., 0
<
O tt.
200
2
Follicular
Luteal
Follicular
Luteal
Follicular
Luteal
FIGURE 2
Mean concentrations of gonadal proteins from the same subjects as in Fig. 1. Total inhibin is a derived number from the sum of inhibin A and inhibin B . . , Group differences; p < 0.05. From Ref. 24, N. E. Reame, T. L. Wyman, D. J. Phillips, D. M. de Kretser, and V. Padmanabhan (1998). Net increase in stimulatory input resulting from a decrease in inhibin B and an increase in activin A may contribute in part to the rise in follicular phase follicle-stimulating hormone of aging cyclic women. J. Clin. Endocrinol. Metab. 83, 3302-3307. 9 The Endocrine Society.
III. D Y N A M I C G O N A D O T R O P I N S E C R E T I O N IN Y O U N G W O M E N The classic studies of Knobil and colleagues conducted in the late 1970s clearly demonstrated that a pulsatile pattern of gonadotropin-releasing hormone was essential for physiological gonadotrope function. Since then, much has been learned about the precise nature of the pulsatile rhythms in both reproductive health and illness through the assessment of pulsatile luteinizing hormone as a surrogate marker of GnRH pulse generator activity. LH episodes originate from periodic activation of the pituitary gonadotroph by intermittent hypothalamic GnRH stimulation. The release magnitude of episodic gonadotropin secretion is defined, among other determinants, by the pituitary responsiveness and the capacity of GnRH to prime the gonadotroph. Serial measurement of FSH concentrations is a less accurate estimate of GnRH secretion due to a reduced and delayed release by the pituitary as well as a slower metabolic clearance rate when compared to LH. A. T h e M e n s t r u a l C y c l e Numerous cross-sectional [30-32] as well as longitudinal studies [33] of pulsatile LH characteristics during the ovulatory menstrual cycle have demonstrated that LH pulse frequency increases during the follicular phase from approxi-
mately one pulse every 90-100 min to one pulse per hour at the time of the LH surge. LH pulse frequency is maintained during the LH surge, but slows during the luteal phase under the influence of the corpus luteal steroids (and central endogenous opioids), with one pulse every 2 - 6 hr varying in amplitude. Thus, the ability to change GnRH amplitude and slow frequency appears to be a critical requirement for the maintenance of cyclic ovulatory function. The dynamic gonadotropin secretory patterns in relation to HPO function have now been well documented for many of the reproductive endocrine disorders and shown to differ from that of the
TABLE I Dynamic Gonadotropin Activity" Hormonal component Central opioid tone FSH (mlU/mil) LH (mlU/ml) LH pulse interval (min) LH pulse amplitude Estradiol Androgens
Postmenopause
Cycle day 6
HA h
PCO c
Absent >30 > 30
Absent <- 10 <--10
Present < 10 < 10
Absent --<10 > 10
60-100 High Very low Low
60-100 Low Low Very low
> 100 Variable Very low Low
60 High Elevated Elevated
a Modified from Ref. 33a, with permission from Springer-Verlag. b HA, Hypothalamic amenorrhea. c p c o , Polycystic ovarian syndrome.
CHAPTER6 Perimenopause Neuroendocrine Regulation ovulatory menstrual cycle (Table 1) [33a]. This body of knowledge provides important context for examining dynamic changes in the perimenopause years.
B. G o n a d o t r o p i n C h r o n o b i o l o g y and Sleep Effects Superimposed on the cyclic changes in pulsatile gonadotropin secretion that occur as a function of the menstrual cycle are the effects of circadian rhythms and sleep. As reviewed by Van Cauter [34], the accumulated evidence suggests that all hormones of the hypothalamo-pituitary axis are influenced by both sleep (irrespective of the time of day) and circadian rhythmicity (regardless of the sleep-wake status). In terms of GnRH activity, circadian effects appear to influence pulse amplitude, whereas sleep affects pulse frequency. LH exhibits a nocturnal, sleep-dependent decline in basal concentrations that is restricted to the early follicular phase [35]. Using a 20-min sampling frequency, Soules et al. [36] demonstrated that this nocturnal inhibition was due to a slowing of LH pulse frequency during the early morning between 1 a.m. and 5 a.m., with a corresponding increase in LH pulse amplitude. The sleep-related decline in LH secretion that is unique to the early follicular phase is believed to be mediated by hypothalamic opioid activity, because pulse frequency can be enhanced after naloxone administration [37], whereas dopaminergic blockade has no effect [38]. Rossmanith and colleagues [39] showed that during the early follicular phase, LH and FSH responses to 25 ~g of GnRH are markedly blunted when assessed in awake subjects at night; this blunting could be completely prevented for LH but not FSH when GnRH was administered during sleep. They concluded that the increased LH pulse amplitude observed during sleep in the early follicular phase was not due to increased pituitary responsiveness, given that there was no circadian effect on the priming effect of GnRH dosing. The physiologic relevance of sleep-related changes in LH in the early follicular phase is not known, but it has been suggested to be important for the maintenance of adequate cyclicity and normal folliculogenesis [40].
C. S y n c h r o n i c i t y of O t h e r H o r m o n e s w i t h Nighttime Gonadotropin Secretion Circadian and sleep-entrained variabilities in the release of other pituitary tropic hormones, such as TSH [41] and prolactin [42], have also been reported. In addition, estradiol concentrations have been observed to exhibit sleep and nighttime enhancement [39]. Changes in reproductive hormone release coincident with sleep may represent mani-
99 festations of entrained links between CNS regulation and endocrine function. In rodents, the suprachiasmatic nucleus (SCN) has been identified as the pacemaker for circadian rhythmicity [43]. Several important physiological events presumed to be controlled by the SCN pacemaker occur during the night and include secretion of melatonin, suppression of vasopressin secretion, increased sensitivity to the phaseshifting effects of light pulses, and high electrical activity and glucose utilization in the SCN. How these various phenomena interface with diurnal features of the reproductive axis requires further definition. It has been demonstrated that the ultradian fluctuations in leptin, the obesity gene product, show pattern synchrony with those of both LH and estradiol in young women during the follicular phase, a time when nocturnal slowing of LH pulses was evident [44]. The investigators postulated that in addition to its trophic effects, leptin may contribute to reproductive axis organization by regulating the minute-to-minute oscillations in the levels of LH and estradiol in the critical period before ovulation. Thus, at night, as leptin levels rise, the pulsatility profile of LH changes from high frequencylow amplitude to low frequency-high amplitude, becoming synchronous with leptin pulsatility. Such a view would help explain the disruption of HPO function that is characteristic of states of low leptin synthesis, such as anorexia nervosa. Although melatonin controls seasonal reproductive cyclicity in some mammalian species, its role as a pacemaker of the human reproductive axis is controversial. Cagnacci et al. [45,46] speculated that melatonin may play a role in timing diurnal LH modifications, because melatonin administration enhanced basal daytime LH secretion, LH pulse amplitude, and responses to a low-dose GnRH challenge during the follicular phase, but not in the luteal phase. However, others [47] have not found a consistent melatonin effect on follicular-phase LH levels. In addition, the circadian rhythm of melatonin secretion does not change significantly across the menstrual cycle and supraphysiologic melatonin concentrations did not decrease the midcycle LH surge response [48-50].
D. S e a s o n a l V a r i a t i o n Seasonal variability in pulsatile LH secretion was suggested by the findings Of Martikainen et al. [51 ], who studied normal volunteers in Finland for 6 daytime hours of the midfollicular phase during peak differences in seasonal daylight (December vs. May). Although mean concentrations, pulse frequency, and amplitude of both LH and FSH did not differ by season, the area under the curve was significantly higher during the winter. A seasonal effect has also been observed for the timing of the LH surge, with ovulation more likely to occur in the morning during the spring and in the evening during autumn and winter [52]. Levels of nighttime LH have
100
NANCY E. REAME
been shown to be higher in the summer at midcycle, at a time when the nocturnal rise in melatonin was reduced [53].
E. FSH Chronobiology Whether FSH secretion changes over the 24-hr period remains controversial. When measured every 15 min, a robust circadian rhythm has been described in young women but is diminished after menopause [54]. In that study, a cosine rhythm with a nighttime decline in transverse mean FSH was observed in the early and late follicular phase despite no evidence of circadian rhythmicity in LH or estradiol. Although diminished compared to the early follicular phase, the comparable timing of the FSH acrophase in the late follicular and midluteal phases, and its presence in the postmenopause group, prompted the investigators to propose a central, rather than peripheral, feedback mechanism for the circadian rhythmicity. The authors concluded that their findings provided further evidence for a dissociation in the hypothalamic regulation of pituitary LH and FSH secretion in women. Other studies have failed to confirm a circadian rhythm in FSH [55]. Using an analytical technique to define discrete secretory bursts measured at 10-min intervals over 24 hr, FSH secretion was maximal during the late follicular phase (high estradiol) and in postmenopausal women unrestrained by estrogen [56]. Although FSH and LH secretory bursts demonstrated a significant concordance rate of 25%, the relatively high rate of nonconcordance prompted the investigators to propose that distinct mechanisms other than a single releasing hormone probably operate to regulate differentially the secretion of each gonadotropin. Diurnal variations in LH and FSH were not described.
IV. GONADOTROPIN CHANGES DURING PERIMENOPAUSE
A. Older Cyling Women The fact that FSH is elevated in normal cycling women over age 40 without concomitant decreases in ovarian steroids would also support the possibility of an aging change in the GnRH signal. Attention has turned from the mechanisms that give rise to the early increase in FSH to the causes of the more subtle alterations in pulsatile LH secretion that lead to the magnified secretory profile of the postreproductive years. Collectively, the data are conflicting. Our studies [57] of daytime pulsatile LH secretion where blood was sampled every 10 min across three phases of the same menstrual cycle showed a gradual rise in pulse frequency with advancing age (Fig. 3). In contrast, in other studies, LH pulsatile secretion in older cycling women was observed to be similar [58] or reduced [59] compared to younger controls. The use of less frequent sampling (every 20 min) in one
FIGURE 3 Effectsof age on pulsatile LH secretory characteristics. Age groups are in years. All subjects were studied across the same menstrual cycle. FOLL, Early follicular phase; ML, midluteal phase; LL, late luteal phase.., p < 0.05; ***,p < 0.001. From Ref. 55, N. E. Reame,R. P. Kelch, I. Z. Beitins, M. Y. Yu, C. Zawacki, and V. Padmanabhan (1996). Age effects on FSH and pulsatile LH secretion across the menstrual cycle of premenopausal women. J. Clin. Endocrinol. Metab. 81, 1512-1518. 9 The Endocrine Society. study and the preliminary nature of the other report may in part account for the differences in findings. In our studies, the enhancement of both LH pulse frequency and amplitude occurred prior to any overt reductions in cyclic estradiol or progesterone concentrations and was phase dependent. Using an intensive sampling protocol (every 10 min for 8 daytime hr), subjects of ages 4 0 - 5 0 years demonstrated shorter cycles, higher mean FSH across all 3 study days, and higher mean LH in the follicular and late luteal phase compared to the youngest age group ( 2 0 - 3 4 years). In keeping with earlier findings of a gradual rise in basal LH levels with age [13], we observed a strong
CHAPTER6 Perimenopause Neuroendocrine Regulation 41.3
39.7
32.3
31.2
38.4
37.1
101 34.1
22.8
17.7
67.3
43.0
62.2 9
40
~
3O
.Ira
20
I.i. z .J
49.5 9
9
55.1 9
9
52.2 9
9
10
Age 45 yrs
Age 43 yrs 0
800
10;0
12;0
,14'00
v, 16~)0
16;0
800
1;oo
12'oo
Age 46 yrs
1;oo ''~ 16~o - 18~o
7r 800
-'
~ 1 000
''1200
1400
" ' 1600
18()0
FIGURE 4 Variationsin dynamicgonadotropinsecretionon cycle day 6 in three ovulatorysubjects over age 40 years with similarestradiolconcentrations (pg/ml). ,, LH pulse. See text for discussion.
correlation between increasing age and higher transverse mean LH in the follicular phase (r = 0.42, p = 0.008) and late luteal studies (r = 0.60, p = 0.0001). Unlike previous studies, our ability to document this age-related increase in a sample of only 32 women is probably related to the large number of data points per study (n = 49 over 8 hr) used to calculate mean concentrations. Additionally, the age effects we observed were most evident in the late luteal phase, a time that is typically less represented in daily sampling studies due to the high variability of cycle length in older women. In the women over age 40 years, individual secretory patterns of LH and FSH across the menstrual cycle were highly variable. Figure 4 depicts examples of individual secretory patterns for LH and FSH (open circles) from three older subjects to highlight the variability present on cycle day 6 of a presumed ovulatory cycle. (Ovulation was presumed based on the presence of a midcycle urinary LH surge, and progesterone values ranging between 5 and 10 ng/ml when measured every 30 min during the midluteal phase study.) In five subjects, elevated FSH secretion persisted across the cycle in the presence of normal changes in pulsatile LH secretion; five others exhibited a failure to slow LH pulse frequency and increase amplitude in the luteal phase with or without enhanced FSH secretion. The remaining six subjects exhibited gonadotropin profiles similar to those observed in the youngest age group. In addition to the cross-sectional cohort, we had the rare opportunity to restudy an individual at multiple time points across the climacteric. Figure 5 presents gonadotropin patterns across two ovulatory menstrual cycles of the same subject studied a decade apart. In this particular woman, despite a strikingly similar gonadotropin profile in the follicular phase at both time points, there is a shorter luteal phase by age 45 years as evidenced by an earlier decline in progesterone concentrations. The markedly lower basal estradiol values in the later study may account for the failure of FSH to suppress in the luteal phase. The absence of large-amplitude LH pulses in the midluteal phase by age 45 years was a common finding in the cross-sectional study.
Taken together, these data suggest that (1) the age-related increase in FSH concentrations in ovulatory women, although more pronounced, is associated with phase-dependent enhancement of pulsatile LH secretion; (2) the higher LH concentrations are brought about by changes in both pulse frequency and amplitude; and (3) these age effects preempt overt reductions in cyclic estradiol or progesterone concentrations.
B.
Age-RelatedOligomenorrhea
We have begun a series of studies in perimenopausal women to assess the effect of intermittent follicular activity on the HPO axis. Eligibility criteria include a change in menstrual cycle regularity in the past year, 45 years of age or older, the onset of hot flashes or other estrogen-deficiency symptoms, and a basal serum FSH value of 20 mIU/ml or greater. Figure 6 presents data from three 8-hr studies of dynamic LH secretion conducted in the same subject on cycle day 6 (left panel), day 26 (middle panel), and day 33 (right panel) of a 43-day menstrual cycle. The typical gonadotropin profile of a postmenopausal woman observed in the first study is dramatically altered during an episode of abnormal follicular development, as evidenced by the ultrasound detection of an ovarian cyst. Under the influence of the 20-fold increase in estradiol concentrations, there is a marked suppression of FSH to nearnormal concentrations. The reversal of the LH:FSH ratio in the presence of reduced LH pulse amplitude and frequency is reminiscent of the normal luteal phase despite the absence of significant progesterone exposure. These data serve to highlight the range of HPO secretory activity present during the perimenopause years and the need for caution when attempting to document menopausal status from a single hormone determination at any given point in time. A crosssectional, epidemiologic study of perimenopausal women in Australia revealed much overlap and high variability in single FSH and estradiol determinations for age and body mass index-matched women with regular cycles versus those with irregular menstrual function [60].
102
NANCY E. REAME
0.7 80
0.3 70
10.0 250
14.0
9.0 Prog (ng/ml) 167 E2 (pg/ml)
14.0
225
174
Age 35 28 dsy cycle
A J
9 :3 15
i
z
X _1
5
.
. I
. 9
. 9
I
. !
-
I
"
i
i
~'
I
Day 4
9
I
9
I
9
Day 19 0.2 43
02 62
14.7 30
Day 2 5 14.0 44
2.2 18
0.4 Prog (ng/ml) 10 E2 (pg/ml) ~
!
FSH
Age 45
:c u._
LH
26 dey cycle
10
Ill Z .J
S
0 900
,,oo
,;oo
,;oo
,;oo;oo
,,'oo
,;oo
Day 6
,Lo
Day 20
'
1700
~
.
, ,.o o .
,. ; o o .
" 1500
I ;00
Day 2 6
FIGURE 5 Longitudinalassessment of gonadotropin secretory patterns in a regularly cycling woman studied at a 10year interval. The x axis is clock hours. Day refers to menstrual cycle day (day 1 = first day of menses). From Ref. 33a, with permission from Springer-Verlag.
C. GnRH Stimulation Test as a Probe of Pituitary Aging Conflicting evidence exists about age effects on pituitary reserve in perimenopausal women when assessed with a GnRH challenge test. An early study demonstrated that the gonadotropin responses to high-dose (100 ~g) and low-dose (10 ~g) GnRH are similar in hypogonadal women and in women during the early follicular phase, but are maximally different during the late follicular phase [61 ]. This difference is presumed to reflect the effect of rising estradiol levels in the young women, and in turn, increasing pituitary reserve relative to pituitary sensitivity to GnRH [62]. To determine whether menstrual cycle irregularity during the perimenopause may be related to increased gonadotropin bioactivity, Schmidt et al. [63] performed GnRH challenge tests using a 100-~g dose in the early to midfollicular phase. The perimenopausal group was compared to young, regularcycling women, a group of older, cycling women, and a postmenopausal group. Although the perimenopause was associated with magnified gonadotropin levels, similar to those of the postmenopausal group at both base line and after stimulation, only postmenopausal women demonstrated increased LH bioactivity. Because estradiol in the earlymidfollicular phase in the perimenopausal group was not lower than in young cycling women and there were no group
differences in androgens, the authors concluded that steroidal feedback differences did not explain the enhanced LH bio/immuno ratio in the postmenopausal years. Gonadal peptides were not assessed. To what extent the 5- to 10-year age difference between the peri/postmenopause groups and the older cycling women may have contributed to the magnified GnRH response and bioactive secretion was not explored. Using a GnRH dose of 25 /zg as a measure of nearmaximal pituitary release capacity (sensitivity), Fujimoto et al. [64] compared gonadotropin responses in young and older cycling women in the early follicular phase (Fig. 7). They showed that the percentage change in both FSH and LH was higher in the young versus older group despite similar levels of estradiol and inhibin, suggesting a diminished pituitary responsiveness in the older cohort.
V. THE N E U R O R E P R O D U C T I V E AXIS IN P O S T M E N O P A U S E After menopause, disproportionately higher levels of FSH versus LH are the norm, with only gradual declines in both gonadotropins occurring after the seventh decade [65]. The enhanced LH secretion in postmenopausal women is associated with a pattern of relatively high pulse amplitude and
CHAPTER 6
Perimenopause Neuroendocrine Regulation
50
9.4 0.02
.i
13.5 0.02
103 15.3 0.02
GnRH Administration
5oo
u :
P<0.05
400 ..J
40 i
30
~
.__~
20 -11111
. ....
-2o 10
I
"
800
"
1000
I
I
9
1200
"
l 1800
I
1400
1600
EF
50"
76.7 0.02
83.6 0.02
104.9 0.02
~0b
-. ~2o
FIGURE 7 Percentage change in mean LH after a GnRH challenge in
A. Changes in Dynamic Gonadotropin Secretion
20
10 I 800
9
|
9
,ooo
9
,
oo
9
9
,4oo
,
oo
,
oo
ML 216.9 1.37
265.8 1.00
244.7
E2 (PG/ML)
1.10 PROG ( N ~ )
40
30
20
80 0
~
increase in gene expression is secondary to the ovarian failure and is not the result of nonspecific aging.
30
lVIt't
............... .............. . , 20 40 80 Time (minutes)
older and younger women during the early follicular phase of a normal menstrual cycle. From Ref. 64. Reprinted with permission from the American Society for Reproductive Medicine (Fertility and Sterility, 1996, Vol. 65, pp. 539-544).
40
50
'
o
.
I
1000
"
I
*
1200
I
1400
"
I
1600
*
I
1800
LL
FIGURE 6 Secretory patterns of LH (e) and FSH (o) in a 52-year-old subject on cycle day 6 (top panel), day 26 (middle panel), and day 33 (bottom panel) of a 43-day menstrual cycle. EF, Early follicular; ML, midluteal; LL, late luteal.
a regular frequency similar to that of the follicular phase [66]. Although the hypothalamic content of GnRH decreases in postmenopausal women [67], GnRH gene expression in the medial basal hypothalamus is increased after menopause without any reductions in area or neuron number compared to that of younger controls [68]. Such data suggest that the
Pulsatile LH secretion of postmenopausal women has been compared with that of young women with premature ovarian failure as a way to distinguish age effects on the neuroreproductive axis. Although LH pulse frequency is similar in the two age groups, young hypogonadal women exhibit higher basal concentrations and greater pulse amplitude compared to women with age-appropriate menopause [66]. Moreover, there is less suppression of LH and FSH by estradiol replacement in the younger group [69]. The lower concentrations of gonadotropins secreted by the older group have been cited as evidence for an age-related effect on hypothalamic-pituitary function. To what extent possible differences in gonadal peptide activity may influence the differential gonadotropin secretion was not explored. Because its half-life has been estimated to be two- to fourfold shorter, the glycoprotein free ce subunit (FAS), although tightly correlated with LH secretion, has been proposed as a more sensitive marker of GnRH secretion at fast pulse frequencies, such as in postmenopausal women [70]. FAS may also be more resistant than LH to down-regulation. GnRH agonist administration to women after menopause results in persistent elevation of FAS despite suppression of LH levels [71]. More importantly, perhaps, Hall and colleagues [72] have proposed that the quality of the LH architecture may change as a function of menopause. Disappearance of endogenous LH after GnRH receptor blockade is prolonged in postmenopausal women, compared to young, cycling controis. The disappearance of FAS was not altered, suggesting that age differences in LH relate to LH microheterogeneity rather than to renal clearance factors [72]. The bulk of neuroendocrine studies of postmenopausal
104
NANCY E. REAME
women have been undertaken not for the purpose of assessing aging mechanisms per se, but rather to simulate aspects of the underlying LH pulse signal free of the influence of endogenous ovarian hormones, i.e., a "castrate" model. Despite the widely held view that LH and FSH secretions after menopause change to a uniform picture of high-amplitude, high-frequency secretion (approximately hourly), estimates of pulsatile LH secretion have varied markedly from study to study, with reported mean pulse frequencies ranging from 60 to 120 min [30,66,73-76] and mean LH concentrations ranging from a low of 19 mlU/ml [39] to values exceeding 75 mlU/ml in the two subjects studied by Yen and colleagues [30]. Within the same study, individual variability of pulsatility patterns is high. In the nine subjects studied by Couzinet [76], pulse frequency ranged from 4 pulses/8 hr to 8 pulses/8 hr, with the range spread equally across the sample. Mean LH concentrations averaged 28.8 IU/liter. High inter-group variability in LH pulse patterns was also noted by Rossmanith et al. [66], who reported a mean pulse frequency of 4 pulses/8 hr, a frequency similar to the luteal phase, for their group of seven subjects. Thus, these data when examined closely, demonstrate a level of variability in LH pulsatile secretion similar to that reported across the highly variable menstrual cycle of young women. A number of differences in methodologies and sample selection may contribute to the variable findings in LH pulse profiles after menopause. Previous studies have used assay methods with varying sensitivities, different criteria to define LH pulses, and heterogeneous study groups. To ensure an adequately depleted ovarian milieu, studies have required subjects to be at least 2 years postmenopausal and without a current history of estrogen replacement therapy (ERT). However, the minimum selection criteria have been interpreted broadly by different investigators: studies have included heterogeneous samples of women ranging in gynecologic age from 2 to 20 years beyond menopause and histories of past use of ERT as recent as 6 - 8 weeks before study. In addition, the type of menopause (surgical vs. spontaneous) has not been controlled for, so that samples ranging in size from 5 to 15 subjects have included both oophorectomized, premenopausal women in their early 40s and spontaneously menopausal women in their 60s, adding further to the heterogeneity of populations. Although it has been assumed that these differences in subject traits and characteristics are benign with respect to the study paradigm, the applicability of the findings to the understanding of the normal climacteric may be limited.
B. Gonadotropin Chronobiology after Menopause Rossmanith and Lauritzen [40] studied 24-hr pulsatile patterns of LH in postmenopausal women compared to
EFP
SLEEP
30
~,
20
-r ..t
10
0 0800
1300
1700
21 O0
01 O0
LFP 9 1
0500
SLEEP ...
,
.. 9 ~
~, 2o ~
0800
"
e
9
~o |~
e
"
0
-r 9J
10
o
~
,
0800
1300
9
MLP
21 O0
L,
01 O0
~,
0500
0800
SLEEP
1
,o
"-"
9
1700
lO
-r .J
4
I
0800
1300
II 1700
9 2100
9 0100
9
9
0500
0800
SLEEP
PMW 140 120
lOO
80 .J 4O
L 0800
A 1300
A
A
9
1700
2100
0100
9 0500
9 0800
FIGURE 8 Diurnal variation in LH pulsatile secretion. Studies of women in the early follicular phase (EFP), late follicular phase (LFP), and midluteal phase (MLP), and of a postmenopausal woman (PMW) are presented. Dashed line, cosine function; arrow, acrophase time. Asterisks denote significant pulses. From Ref. 40, with permission from Parthenon Press.
women at three different menstrual cycle phases (Fig. 8). In addition to the sleep-entrained rise in LH in the early follicular phase, when secretory profiles were fitted to cosinor functions, diurnal excursions in LH secretion were observed in 14 of the 16 early follicular-phase women, 11 of the 14 late follicular-phase women, in 12 of the 15 midluteal-phase women, and in 7 of the 8 postmenopausal women. However, the postmenopausal women demonstrated the smallest deviation from mesor levels (the value around which the os-
CHAPTER6 Perimenopause Neuroendocrine Regulation
105
EARLY FOLLICULAR
20
Postmenopause
10 Z < UJ
::i
U,I 09
0:
0.3
o
3O
0 -lO
0.2
E D l
-20
,I ,'
9 z
20
r
10
20 r-
E "1-
UJ
LATE FOLLICULAR
..,.I
0.1
lO
p-
2 0 r U. UJ
0 lj," -10
(9
Z
< T" 0
-20
L
6'
~
I" v
l'
E r"
'1
O
LH
9
T I
0 20
ID ffi
o
I--. 0.0
0.3
20
I,Z
POS T-MENOPAUSAL
o.
E
15
I0
0.2 ,5
0
0
.-~ 1::: 0
-10
0 I,... r
10
o
0.1
-20
oU) I-
0900
A
CLOCK HOURS
FIGURE 9 Mean change (%) from base line in dynamic FSH secretion in young, cycling women and after menopause. The circadian variation in each study group is represented by the best-fit cosine curve. From Ref. 54, J. F. Mortola, G. A. Laughlin, and S. S. C. Yen (1992). A circadian rhythm of serum follicle stimulating hormone in women. J. Clin. Endocrinol. Metab. 75, 861-864. 9 The Endocrine Soceity.
cillation occurs) and a marked shift was observed in the acrophase time from early afternoon in the cycle to early morning hours after menopause. This dampened rhythmicity in pulsatile LH secretion after menopause was attributed to the loss of sex steroid sensitivity of opioid action [40]. FSH secretion was not assessed. Another study using similar clinical protocol and analytical methods uncovered a blunted circadian rhythm in FSH secretion in postmenopausal women compared to young, cycling controls [54]. In that study, a cosine rhythm with a nighttime decline in transverse mean FSH was observed in the early and late follicular phase and was markedly attenuated after menopause (Fig. 9). In contrast to the findings of Rossmanith and Lauritzen [40], these investigators observed no evidence of circadian rhythmicity in LH or estradiol under similar conditions. Such disparate findings suggest additional studies are warranted to clarify the exact nature of the underlying gonadotropin rhythmicities.
C. Pulsatile Testosterone Secretion in Hypogonadal Women Thanks to the elegant studies of Judd and colleagues [77], who measured ovarian vein and peripheral plasma hormone concentrations before and after ovariectomy, the follicular-
09
'
11
'
13
'
15
F '
17
0.0
Clock Time
FIGURE 10 Episodic secretion of testosterone (e) in a postmenopausal woman and its relationship to LH (o) and cortisol (F; A). Pulse concordance with LH, p < 0.01. Cross-correlation with cortisol, r = 0.58, p < 0.01.
depleted ovary has been shown to be the major source of testosterone after menopause. These studies as well as others showing dramatic testosterone suppression after GnRH agonist treatment led Adashi [78] to conclude that the postmenopausal ovary, rather than being a defunct endocrine gland in "end-organ failure," is gonadotropin dependent and responsive to LH. With the advent of the more sensitive immunofluorometric assays (IFMAs), the low concentrations of testosterone normally present in perimenopausal women are now within the range of assay sensitivity, thus making it feasible for the first time to characterize reliably its dynamic secretion. As a further test of the hypothesis that the climacteric ovary is a gonadotropin-responsive, androgenproducing gland, we assessed the relationship between pulsatile LH secretion and episodic release of testosterone (T) in comparison to cortisol secretion in hypogonadal females [79]. Figure 10 presents plasma concentrations of LH, testosterone, and cortisol sampled over 8 daytime hours from a 50-year-old postmenopausal woman with an FSH value > 5 0 mIU/ml and an estradiol level of 5 pg/ml. In this individual, testosterone secretion was episodic and the concordance between T and LH pulses was 71% (p < 0.01), but no significant cross-correlation of secretory patterns was observed. In contrast, there was no significant relationship between pulses of T and cortisol, although a strong positive cross-correlation was observed at a 0 time lag (r = 0.58; p < 0.01). These preliminary data suggest that although the adrenal gland may
106 serve as the rhythm generator for basal testosterone secretion, pituitary LH contributes to the pulsatile release of T in postmenopausal women.
VI. BRAIN AGING AND REPRODUCTIVE SENESCENCE There is currently renewed debate over the relative contribution of the ovaries and the hypothalamic-pituitary unit to the initiation of the human menopause [80]. As reviewed by Wise and colleagues [81,82], heterochronic ovarian transplant studies clearly implicate the hypothalamic-pituitary axis as a primary mediator of both the monotropic FSH rise and the reproductive aging in the rat [83]. Ovulation can be restored in senescent ovaries when transplanted to the kidney capsule of young females [84], and CNS-acting agents can reinitiate estrous and ovulation in aged animals [85]. Conversely, although ovarian transplants from young donors to old recipients in the mouse will double the number of cyclic ovulations, they fail to prevent cycle lengthening [86]. Taken together, such data support a clear influence of the hypothalamic-pituitary system in the onset of reproductive senescence in rodents. Drawing on data from their work, Wise and colleagues [80] have proposed a competing hypothesis on the trigger for menopause. It is their view that preemptive aging changes in the brain lead to the alterations in folliculogenesis, gonadal peptide activity, and gonadotropin augmentation. Several lines of evidence from her laboratory suggest that changes in a variety of neurotransmitter systems that regulate GnRH secretion and possibly circadian, diurnal, and ultradian oscillation are altered with age and may contribute to reproductive senescence. The observation that changes in pulsatile LH release can be detected in middle-aged rats that showed no deterioration in the regularity of their estrous cycles suggested that aging of the hypothalamic pulse generator occurs early during the transition to acyclicity and may play a causative role in age-related transition. The change in LH pulsatility has been shown to correlate with changes in the diurnal pattern of activity or gene expression of norepinephrine, serotonin, and fl-endorphin. Based on these findings, Wise et al. propose that multiple pacemakers in the brain are likely to govern the orchestration of complex neurochemical events that give rise to reproductive cyclicity [80]. Moreover, they hypothesize that the accelerated loss of follicles in women after age 35 years may reflect an age-related desynchronization in the rhythmicity of pulsatile GnRH secretion [82]. Specifically, they postulate that a progressive deterioration of the 24-hr rhythmicity of the biological clock located in the suprachiasmatic nucleus of the hypothalamus leads to an uncoupling of the coordinated neurosecretory inputs that govern the activity of the GnRH pulse generator. Such insults, they propose, would
NANCY E. REAME
lead to a dampening of the GnRH pulse frequency, and in turn the preferential increase in the release of FSH over LH. Some additional lines of evidence would support such a central cause for gonadotropin disturbances. In the rhesus monkey, the frequency of GnRH pulses has been shown to influence the LH:FSH ratio: slow pulses produce mainly FSH, fast pulses produce LH [87]. Moreover, disturbances in diurnal LH secretion are key features of the HPO axis in patients with reproductive endocrine disorders [88]. Although contrary to our findings of enhanced (rather than diminished) LH pulsatility in older ovulatory women, the idea of derangements in the circadian controls of GnRH secretion clearly merits further examination given the growing body of evidence for age-related declines in other neuroendocrine systems mediated by hypothalamic function. For example, studies have demonstrated diminished function of the somatotropic axis of premenopausal women [89,90], and menopause-related effects on the nighttime suppression of cortisol [91]. Prolactin is pulsatile and magnified with sleep in postmenopausal women but is dampened overall due to a lower pulse frequency compared to normal cycling, younger women [92]. These lines of evidence would suggest that subtle aging deficits in hypothalamic function may exist much earlier than previously believed, but to what extent these alterations are relevant to the initiation of menopause remains to be determined. Studies underway in our laboratory are addressing whether the perimenopausal changes in gonadotropin secretion occur first at night, as in puberty, and if menopause is heralded by a suppression of the sleepinduced changes in LH pulsatility. In summary, there is heightened interest in the role of central aging deficits in the etiology of the menopause. A fundamental question is whether GnRH secretion increases at the time of menopause, and if so, whether this is mediated by declines in the integrity of central circadian pacemakers. This hypothesis provides important new directions for studies of the HPO axis at menopause. An understanding of the factors that interact and initiate the process of hypoestrogenism in aging women is needed to develop strategies for alleviating the negative aspects of the menopause and better comprehend the process of biologic aging.
VII. FUTURE
STUDIES
With the advent of specific assays for the bioactive forms of the gonadal peptides and their subunits, it will be possible in the near future to assess systematically the component roles of the inhibins, activins, and follistatin as local and central mediators of aging changes in gonadotropin secretion (Fig. 11). Such information should greatly add to our understanding of the ovulatory process and the abnormalities associated with premature menopause and infertility. Cross-sectional studies of hormonal profiles have for the
CHAPTER 6 Perimenopause Neuroendocrine Regulation
107
FIGURE 11 A model for the changes in gonadotropin and estrogen secretion observed in older, cycling women: proposed aging effects on the HPO axis during the early follicular phase. Age effects at both the level of central pacemakers and the ovary may act together to promote the rise in FSH, enhanced LH pulsatility, and hyperestrogenism.
most part relied on daily or weekly blood measures of LH, FSH, and sex steroids collected from healthy volunteers or wives of infertility patients. Although only women with regular menstrual cycles were studied, few additional clinical data, such as smoking history, parity, body weight characteristics, or diet, were reported. Such factors have been linked to ovarian function [2] and may help explain the controversy about whether estradiol levels increase, decrease, or remain unchanged with age in women who continue to experience regular menstrual cycles. Although seldom reported, it is presumed that the majority of neuroendocrine studies of the menopause transition have been limited to predominantly white, middle-class samples of patients or volunteers. How race, socioeconomic status, and lifestyle factors (e.g., smoking history, diet, body fat
characteristics, exercise) may collectively or independently govern the nature and timing of perimenopause events is unknown. Thus, studies are needed of the influence of extragonadal factors on the aging of the HPO axis. In 1992, the National Institute on Aging launched a multisite, longitudinal study of the hormonal and systemic effects of the natural menopause transition in African-American, Hispanic, Asian-American, and white women as a way to address the limited information available about determinants of the menopause experience, especially for women of color. Early findings suggest that there are significant differences across racial and ethnic groups in menopause symptomatology, such as hot flash severity, which are not accounted for by body mass index, smoking, or socioeconomic factors [93].
108
NANCY E. REAME
Acknowledgments This work was supported by NIH grants RO1 NR01373, 5M01RR00042, U54 HD29184, and RO 1 AG15083. The author is indebted to her colleague Vasantha Padmanabhan, Ph.D., for her thoughtful review of an earlier version of this manuscript.
18.
19.
References 20. 1. Menken, J., Trussell, J., and Larsen, U. (1986). Age and infertility. Science 26, 1389-1394. 2. U.S. Congress, Office of Technology Assessment (1992). "The Menopause, Hormone Therapy, and Women's Health," OTA-BP-BA-88. U.S. Govt. Printing Office, Washington, DC. 3. World Health Organization (1981). "Research on the Menopause," Tech. Rep. Ser. 670. W.H.O., Geneva. 4. Bachmann, G. A. (1993). The changes before "the change": Strategies for the transition to the menopause. Postgrad. Med. 95, 113-124. 5. Gist, Y. J., and Velkoff, V. A. (1998). "Gender and Aging. Demographic Dimensions," Census Bureau Report. U.S. Govt. Printing Ofrice, Washington, DC. 6. Treolar, A. E., Boynton, R. E., Behn, B. G., and Brown, B. W. (1967). Variation of the human menstrual cycle through reproductive life. Int. J. Fertil. 12 (1-2), 77-126. 7. Nelson, L. M., Anasti, J. N., Flack, M. R. (1997). Premature ovarian failure. In "Reproductive Endocrinology, Surgery and Technology" (E. Y. Adashi, J. A. Rock, and Z. Rosenwaks, eds) Vol. 2, pp. 13931410. Raven Press, New York. 8. Coulam, C. B., Adamson, S. C., and Annegers, J. E (1986). Incidence of premature ovarian failure. Obstet. Gynecol. 67, 604-606. 9. Santoro, N. (1996). Characterization of reproductive hormonal dynamics in the perimenopause. J. Clin. Endocrinol. Metab. 81, 14951501. 10. Richardson, S. J., Senikas, V., and Nelson, J. E (1987). Follicular depletion during the menopausal transition: Evidence for accelerated loss and ultimate exhaustion at menopause. J. Clin. Endocrinol. Metab. 65, 1231-1237. 11. McKinlay, S. M., Brambilla, D. J., and Posner, J. G. (1992). The normal menopause transition. Am. J. Hum. Biol. 4, 37-46. 12. Lee, S. J., Lenton, E. A., Sexton, L., and Cooke, I. D. (1988). The effect of age on the cyclical patterns of plasma LH, FSH, estradiol and progesterone in women with regular menstrual cycles. Hum. Reprod. 3, 851-855. 13. Ebbiary, A., Lenton, E. A., and Cooke, I. D. (1994). Hypothalamicpituitary aging: Progressive increase in FSH and LH concentrations throughout the reproductive life in regularly menstruating women. Clin. Endocrinol. (Oxford)41, 199-206. 14. MacNaughton, J., Bangah, M., McCloud, P., Hee, J., and Burger, H. (1992). Age-related changes in follicle stimulating hormone, luteinizing hormone, estradiol and immunoreactive inhibin in women of reproductive age. Clin. Endocrinol. (Oxford) 36, 339-345. 15. Lenton, E. A., de Kretser, D. M., Woodward, A. J., and Robertson, D. M. (1991). Inhibin concentrations throughout the menstrual cycles of normal, infertile, and older women compared with those during spontaneous conception cycles. J. Clin. Endocrinol. Metab. 7 3 , 1 1 8 0 1190. 16. Monroe, S. E., Jaffe, R. B., and Midgley, A. R. (1972). Regulation of human gonadotropins. XIII. Changes in serum gonadotropins in menstruating women in response to oophorectomy. J. Clin. Endocrinol. Metab. 34, 420-422. 17. Klein, N., Illingworth, P. J., Groome, N. P., McNeilly, A. S., Battaglia, D. E., and Soules, M. R. (1996). Decreased inhibin B secretion is as-
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31. 32.
33.
33a.
sociated with the monotropic FSH rise in older, ovulatory women: A study of serum and follicular fluid levels of dimeric inhibin A and B in spontaneous menstrual cycles. J. Clin. Endocrinol. Metab. 81, 2742 -2745. Ying, S. Y. (1988). Inhibins, activins and follistatins: Gonadal proteins modulating the secretion of follicle-stimulating hormone. Endocr. Rev. 9, 267-293. Shimasaki, S., Koga, M., Esch, F., Mercado, M., Cooksey, K., Koba, A., and Ling, N. (1988). Porcine follistatin gene structure supports two forms of mature follistatin produced by alternate splicing. Biochem. Biophys. Res. Commun. 152, 717-723. Nakamura, T., Takio, K., Eto, Y., Shibai, H., Titani, K., and Sugino, H. (1990). Activin-binding protein from rat ovary is follistatin. Science 247, 836-838. DePaulo, L. V., Bicsak, T. A., Erickson, G. E, Shimasaki, S., and Ling, N. (1991). Follistatin and activin: A potential intrinsic regulatory system within diverse tissues. Proc. Soc. Exp. Biol. Med. 198, 500-512. Wakatsuki, M., Shintani, Y., Abe, M., Liu, Z.-H., Shitsukawa, K., and Saito, S. (1996). Immunoradiometric assay for follistatin: Serum immunoreactive follistatin levels in normal adults and pregnant women. J. Clin. Endocrinol. Metab. 81,630-634. Harada, K., Shintani, Y., Sakkamoto, Y., Wakatsuki, M., Shitsukawa, K., and Saito, S. (1996). Serum immunoreactive activin A levels in normal subjects and patients with various diseases. J. Clin. Endocrinol. Metab. 81, 2125-2130. Reame, N. E., Wyman, T. L., Phillips, D. J., de Kretser, D. M., and Padmanabhan, V. (1998). Net increase in stimulatory input resulting from a decrease in inhibin B and an increase in activin A may contribute in part to the rise in follicular phase follicle-stimulating hormone of aging cyclic women. J. Clin. Endocrinol. Metab. 83, 3302-3307. Santoro, N., Adel, T., and Skurnick, J. H. (1999). Decreased inhibin tone and increased activin A secretion characterize reproductive aging in women. Fertil. Steril. 71,658-662. Welt, C. K., McNicholl, D. J., Taylor, A. E., and Hall, J. E. (1999). Female reproductive aging is marked by decreased secretion of dimeric inhibin. J. Clin. Endocrinol. Metab. 84, 105-111. Reame, N. E., Zuliani, G. C., Lukacs, J., Lukacs, N., Rolfes-Curl, A., and Padmanabhan, V. (1998). Circulating levels of activin decrease in response to hormone replacement therapy. Menopause 5(4), 267 (Abstr. No. P45). Reame, N., Lukacs, J., Olton, P., and Padmanabhan, V. (1999). Differential effects of ovulation and hormone replacement therapy on circulating levels of inhibin A, inhibin B, activin A and follistatin in perimenopausal women. 81st Annu. Meet. Endocr. Soc., Abstract P2-55, p. 292. San Diego. Mather, J. P., Woodruff, T. K., and Krummen, L. A. (1992). Paracrine regulation of reproductive function by inhibin and activin. Proc. Soc. Exp. Biol. Med. 201, 1-15. Yen, S. S. C., Tsai, C. C., Naftolin, E, Vandenberg, G., and Ajabor, L. (1972). Pulsatile patterns of gonadotropin release in subjects with and without ovarian function. J. Clin. Endocrinol. Metab. 34, 671-675. Santen, R. J., Bardin, C. W. (1973). Episodic luteinizing hormone secretion in man. J. Clin. Invest. 52, 2617-2628. Backstrom, C. T., McNeilly, A. S., Leask, R. M., and Baird, D. T. (1982). Pulsatile secretion of LH, FSH, prolactin, oestradiol and progesterone during the human menstrual cycle. Clin. Endocrinol. (Oxford) 17, 29-42. Reame, N. E., Sauder, S. E., Kelch, R. P., and Marshall, J. C. (1984). Pulsatile gonadotropin secretion during the human menstrual cycle: Evidence for altered frequency of gonadotropin-releasing hormone secretion. J. Clin. Endocrinol. Metab. 59, 328-337. Reame, N. E. (1997). Gonadotropin changes in the perimenopause. In "Perimenopause" (R. Lobo, ed.), p. 161. Springer-Verlag, New York.
109
CHAPTER 6 Perimenopause Neuroendocrine Regulation 34. Van Cauter, E. (1990). Diurnal and ultradian rhythms in human endocrine function: A minireview. Horm. Res. 34(2), 45-53. 35. Kapen, S., Boyar, R., Hellman, L., and Weitzman, E. D. (1976). The relationship of luteinizing hormone to sleep in women during the early follicular phase: Effects of sleep reversal and a prolonged 3 hour sleep-wake schedule. J. Clin. Endocrinol. Metab. 42, 1031-1040. 36. Soules, M. R., Steiner, R. A., Cohen, N. L., Bremner, W. J., and Clifton, D. K. (1985). Nocturnal slowing of pulsatile luteinizing hormone secretion in women during the follicular phase of the menstrual cycle. J. Clin. Endocrinol. Metab. 61, 43-49. 37. Rossmanith, W. G., and Yen, S. S. C. (1987). Sleep-associated decrease in LH pulse frequency during the early follicullar phase of the menstrual cycle: Evidence for an opioidergic mechanism. J. Clin. Endocrinol. Metab. 67, 713 -718. 38. Rossmanith, W. G., Mortola, J. E, and Yen, S. C. (1988). Effects of dopaminergic blockade on the sleep-associated changes of LH pulsatility in early follicular phase women. Neuroendocrinology 48, 634-639. 39. Rossmanith, W. G., Boscher, S., Kern, W., and Fehm, H. L. (1993). Impact of sleep on the circadian excursions in the pituitary gonadotropin responsiveness of early follicular phase women. J. Clin. Endocrinol. Metab. 76, 330-336. 40. Rossmanith, W. G., and Lauritzen, C. (1991). The luteinizing hormone pulsatile secretion: Diurnal excursions in normally cycling and postmenopausal women. Gynecol. Endocrinol. 5, 249-265. 41. Greenspan, S. L., Klibanski, A., Schoenfeld, D., and Ridgway, E. C. (1986). Pulsatile secretion of thyrotropin in man. J. Clin. Endocrinol. Metab. 63, 661- 668. 42. Sassin, J. E, Frantz, A. G., Weitzman, E. D., and Kapen, S. (1972). Human prolactin: 24-hour pattern with increased release during sleep. Science 177, 1205-1207. 43. Moore, R. Y. (1983). Organization and function of a CNS circadian oscillator: The suprachiasmatic nucleus. Fed. Proc. Fed. Am. Soc. Exp. Biol. 42, 2783-2788. 44. Licinio, J., Negrao, A. B., Mantzoros, C., Kaklamani, V., Wong, M. I., Bongiorno, P. B., Mulla, A., Cearnal, L., Veldhuis, J. D., Flier, J. S., McCann, S. M., and Gold, P. W. (1998). Synchronicity of frequentlysampled 24hour concentrations of circulating leptin, luteinizing hormone and estradiol in healthy women. Proc. Natl. Acad. Sci. U.S.A. 95(5), 2541-2546. 45. Cagnacci, A., Paoletti, A. M., Soldani, R., Orru, M., Maschio, E., and Melis, G. B. (1995). Melatonin enhances the luteinizing hormone and follicle-stimulating hormone responses to gonadotropin-releasing hormone in the follicular, but not in the luteal menstrual phase. J. Clin. Endocrinol. Metab. 80, 1095-1099. 46. Cagnacci, A., Soldani, R., and Yen, S. S. C. (1995). Exogenous melatonin enhances luteinizing hormone levels of women in the follicular but not in the luteal menstrual phase. Fertil. Steril. 63, 996-999. 47. Terzolo, M., Revelli, A., Guidetti, D., Piovesan, A., Cassoni, P., Paccotti, P., Angeli, A., and Massobrio, M. (1993). Evening administration of melatonin enhances the pulsatile secretion of prolactin but not of LH and TSH in normally cycling women. Clin. Endocrinol. (Oxford) 39, 185-189. 48. Brzezinski, A., Lynch, H. J., Seibel, M. M., Deng, M. H., Nuder, T. M., and Wurtman, R. J. (1988). The circadian rhythm of plasma melatonin during the normal menstrual cycle and in amenorrheic women. J. Clin. Endocrinol. Metab. 66, 891-895. 49. Zimmermann, R.C., Schroder, S., Baars, S., Schumacher, M., and Weise, H. C. (1990). Melatonin and the ovulatory luteinizing hormone surge. Fertil. Steril. 54, 612- 618. 50. Berga, S. L., and Yen, S. S. (1990). Circadian pattern of plasma melatonin concentrations during four phases of the human menstrual cycle. Neuroendocrinology 51,606-612. 51. Martikainen, H., Ruokonen, A. I., Tomas, C., and Kauppila, A. (1996). Seasonal changes in pituitary function: Amplification of mid-
follicular luteinizing hormone secretion during the dark season. Fertil. Steril. 65, 7188-7220.
52. Testart, J., Frydman, R., and Roger, M. (1982). Seasonal influence of diurnal rhythms in the onset of the plasma luteinizing hormone surge in women. J. Clin. Endocrinol. Metab. 55, 374-377. 53. Kivela, A., Kauppila, A., Ylostalo, P., Vakkuri, O., and Leppaluoto, J. (1988). Seasonal, menstrual and circadian secretions of melatonin, gonadotropins and prolactin in women. Acta physiol. Scand. 132, 321327. 54. Mortola, J. E, Laughlin, G. A., and Yen, S. S. C. (1992). A circadian rhythm of serum follicle stimulating hormone in women. J. Clin. Endocrinol. Metab. 75, 861-864. 55. Veldhuis, J. D., Iranmanesh, A., Johnson, M. L., and Lizarralde, G. (1990). Twenty-four hour rhythms in plasma concentrations of adrenohypophyseal hormones are generated by distinct amplitude and/or frequency modulation of underlying pituitary secretory bursts. J. Clin. Endocrinol. Metab. 71, 1616-1623. 56. Booth, R. A., Weltman, J. Y., Yankov, V. I., Murray, J., Davison, T. S., Rogol, A. D., Asplin, C. M., Johnson, M. L., Veldhuis, J. D., and Evans, W. S. (1996). Mode of pulsatile FSH secretion in gonadal hormone-sufficient and deficient women. J. Clin. Endocrinol. Metab. 81, 3208-3214. 57. Reame, N. E., Kelch, R. P., Beitins, I. Z., Yu, M. Y., Zawacki, C., and Padmanabhan, V. (1996). Age effects on FSH and pulsatile LH secretion across the menstrual cycle of premenopausal women. J. Clin. Endocrinol. Metab. 81, 1512-1518. 58. Klein, N. A., Battaglia, D. E., Clifton, D. K., Bremner, W. J., and Soules, M. R. (1996). The gonadotropin secretion pattern in normal women of advanced reproductive age in relation to the monotropic FSH rise. J. Soc. Gynecol. Invest. 3, 27-32. 59. Matt, D. W., Weltman, J. Y., Evans, W. S., Kauma, S. W., Veldhuis, J. D., James, C. A., and Board, J. A. (1994). Follicular phase LH pulse frequency is decreased in middle-aged regularly cyclic women. 76th Annu. Meet. Endocr. Soc., Anaheim, CA, 1994, Abstr. No. 710. 60. Burger, H. G., Dudley, E. C., Hopper, J. L., Shelley, J. M., Green, A., Smith, A., Dennerstein, L., and Morse, C. (1995). The endocrinology of the menopausal transition: A cross-sectional study of a populationbased sample. J. Clin. Endocrinol. Metab. 80, 3537-3545. 61. Wang, C. F., Lasley, B. L., Lein, A., and Yen, S. S. C. (1976). The functional changes of the pituitary gonadotrophs during the menstrual cycle. J. Clin. Endocrinol. Metab. 42, 718-728. 62. Hoff, J. D., Lasley, B. L., Wang, C. F., and Yen, S. S. C. (1977). The two pools of pituitary gonadotropin: Regulation during the menstrual cycle. J. Clin. Endocrinol. Metab. 44, 302-312. 63. Schmidt, P. H., Gindoff, P. R., Baron, D. A., and Rubinow, D. R. (1996). Basal and stimulated gonadotropin levels in the perimenopause. Am. J. Obstet. Gynecol. 175, 643-650. 64. Fujimoto, V. Y., Klein, N. A., Battaglia, D. E., Bremner, W. J., and Soules, M. R. (1996). The anterior pituitary response to a gonadotropinreleasing hormone challenge test in normal older reproductive-age women. Fertil. Steril. 65, 539-544. 65. Rossmanith, W. G., Liu, C. H., Laughlin, G. A., Mortola, J. E, Suh, B. Y., and Yen, S. S. C. (1990). Relative changes in LH pulsatility during the menstrual cycle: Using data from hypogonadal women as a reference point. Clin. Endocrinol. (Oxford) 32, 647-660. 66. Rossmanith, W. G., Handke-Vesely, A., Wirth, U., and Scherbaum, W. A. (1994). Does the gonadotropin pulsatility of postmenopausal women represent the unrestrained hypothalamic-pituitary activity? Eur. J. Endocrinol. 130, 485-493. 67. Parker, C. R., and Porter, J. C. (1984). Luteinizing hormone-releasing hormone and thyrotropin-releasing hormone in the hypothalamus of women: Effects of age and reproductive status. J. Clin. Endocrinol. Metab. 58, 488-491. 68. Rance, N. E., and Uswandi, S. V. (1996). Gonadotropin-releasing hormone gene expression is increased in the medial basal hypothalamus
110
69.
70.
71.
72.
73.
74.
75.
76.
77.
78. 79.
NANCY E. REAME
of postmenopausal women. J. Clin. Endocrinol. Metab. 81, 35403546. Santoro, N., Banwell, T., Tortoriello, D., Lieman, H., Adel, T., and Skemick, J. (1998). Effects of aging and gonadal failure on the hypothalamic-pituitary axis in women. Am. J. Obstet. Gynecol. 178, 732-741. Hayes, F. J., McNicholl, D. J., Schoenfeld, D., Marsh, E. E., and Hall, J. E. (1999). Free alpha subunit is superior to luteinizing hormone as a marker of gonadotropin-releasing hormone despite desensitization at fast pulse frequencies. J. Clin. Endocrinol. Metab. 84, 1028-1036. Oppenheim, D. S., Bikkai, H., Crowley, W. F., Jr., and Klibanski, A. (1992). Effects of chronic GnRH analogue administration on gonadotrophin and alpha-subunit secretion in post-menopausal women. Clin. Endocrinol. (Oxford) 36, 559-564. Sharpless, J. L., Supko, J. G., Martin, K. A., and Hall, J. E. (1999). Disappearance of endogenous luteinizing hormone is prolonged in postmenopausal women. J. Clin. Endocrinol. Metab. 84, 688-694. Veldhuis, J. D., Evans, W. S., Rogol, A. D., Thorner, M. O., and Stumpf, P. (1987). Actions of estradiol on discrete attributes of the luteinizing hormone pulse signal in man. J. Clin. Invest. 79, 769-776. Cagnacci, A., Melis, G. B., Paoletti, A. M., Gambacciani, M., Soldani, R., Spinetti, A., and Fioretti, P. (1989). Influence of oestradiol and progesterone on pulsatile LH secretion in postmenopausal women. Clin. Endocrinol. (Oxford) 31, 541-550. D'Amico, J. F., Greendale, G. A., Lu, J. K., and Judd, H. L. (1991). Induction of hypothalamic opioid activity with transdermal estradiol administration in postmenopausal women. Fertil. Steril. 55, 754-758. Couzinet, B., Lahlou, N., Thomas, G., Thalabard, J. C., Bouchard, P., Roger, M., and Schaison, G. (1991). Effects of gonadotrophin releasing hormone antagonist and agonist on the pulsatile release of gonadotrophins and alpha-subunit in postmenopausal women. Clin. Endocrinol. (Oxford) 34, 477-483. Judd, H. L., Lucas, W. E., and Yen, S. S. C. (1974). Endocrine function of the postmenopausal ovary: Concentrations of androgens and estrogens in ovarian and peripheral vein blood. J. Clin. Endocrinol. Metab. 39, 1020-1024. Adashi, E. Y. (1994). The climacteric ovary as a functional gonadotropin-driven androgen-producing gland. Fertil. Steril. 62, 20-27. Reame, N. E., Rolfes-Curl, A., Foster, C., and Padmanabhan, V. (1995). Testosterone secretion during the perimenopause: Evidence for episodic secretion and responsiveness to pulsatile LH. Menopause 2(4), 247 (S-9).
80. Wise, R M., Krajnak, K. M., and Kashon, M. L. (1996). Menopause: The aging of multiple pacemakers. Science 273, 67-70. 81. Wise, E M., Scarbrough, K., Lloyd, J., Cai, A., Harney, J., Chiu, S., Hinkle, D., and McShane, T. (1994). Neuroendocrine concomitants of reproductive aging. Exp. Gerontol. 29, 275-283. 82. Wise, E M. (1998). Menopause and the brain. Sci. Am. (Special Issue Women's Health: A Lifelong Guide) 9 (Summer), 79-81. 83. Sopelak, V. M., and Butcher, R. L. (1982). Contribution of the ovary versus hypothalamus-pituitary to termination of estrous cycles in aging rats using ovarian transplants. Biol. Reprod. 27, 29-37. 84. Ascheim, E (1983). Relation of neuroendocrine system to reproductive decline in female rats. In "Neuroendocrinology of Aging" (J. Meittes, ed.), pp. 73-101, Plenum, New York. 85. Quadri, S. K., Kledzik, G. S. and Meites, J. (1973). Reinitiation of estrous cycles in old constant-estrous rats by central-acting drugs. Neuroendocrinology 11,248-255. 86. Nelson, J. E, and Felicio, L. S. (1990). Hormonal influences on reproductive aging in mice. Ann. N.Y. Acad. Sci. 592, 8-12. 87. Wildt, L., Hausler, A., Marshall, G., Hutchison, J. S., Plant, T. M., Belchetz, E E., and Knobil, E. (1981). Frequency and amplitude of gonadotropin-releasing hormone stimulation and gonadotropin secretion in the rhesus monkey. Endocrinology (Baltimore) 109, 376-385. 88. Khoury, S., Reame, N. E., Kelch, R. E, and Marshall, J. C. (1991). Diurnal patterns of pulsatile LH secretion in hypothalamic amenorrhea: Reproducibility and response to opiate blockade and an alpha-2 andrenergic agonist. J. Clin. Endocrinol. Metab. 64, 755-762. 89. Wilshire, G. B., Loughlin, J. S., Brown, J. S., Adel, T. E., and Santoro, N. (1995). Diminished function of the somatotropic axis in older reproductive-aged women. J. Clin. Endocrinol. Metab. 80, 608- 613. 90. Cano, A., Catelo-Branco, C., and Tarin, J. (1999). Effect of menopause and different combined estradiol-progestin regimens on basal and growth hormone-releasing hormone-stimulated serum growth hormone, insulin-like growth factor-l, insulin-like growth factor binding protein (IGFBP)-I, and IGFBP-3 levels. Fertil. Steril. 71, 261-267. 91. Van Coevorden, A., Mockel, J., and Laurent, E. (1991). Neuroendocrine rhythms and sleep in aging men.Am. J. Physiol. 260, E851-E861. 92. Katznelson, L., Riskind, R N., Saxe, V. C., and Klibanski, A. (1998). Prolactin pulsatile characteristics in postmenopausal women. J. Clin. Endocrinol. Metab. 83, 761-764. 93. Gold, E. B., Sternfeld, B., Kelsey, J. L., Brown, C., Mouton, C., Reame, N., Salamone, L., and Stellato, R. (2000). The relation of demographic and lifestyle factors to symptoms in a multi-racial/ethnic population of women 40-55 years of age. Am. J. Epidemiol., in press.
_~HAPTER
Changes in Men as They Age: The Manopause STANLEY G. ARSHAG D.
KORENMAN
MOORADIAN
VICTORIA HENDRICK
I. II. III. IV. V. VI.
Department of Medicine, University of California, Los Angeles, School of Medicine, Los Angeles, California 90095 Department of Internal Medicine, Saint Louis University School of Medicine, Saint Louis, Missouri 63104 Department of Psychiatry, University of California, Los Angeles, School of Medicine, Los Angeles, California 90095
Introduction Age-Related Changes in the Male Reproductive System Age-Related Changes in Other Hormonal Systems Other Changes with Aging Related to Hormonal Factors Androgen Effects on Sexual and Reproductive Function Nonerectile Sexual Dysfunctions
VII. VIII. IX. X.
I. I N T R O D U C T I O N
a "viropause" remains controversial. A syndrome is usually defined as a set of symptoms, signs, and diagnostic features that are all due to a single etiology, e.g., Cushing's syndrome is due to the overproduction of cortisol or administration of excess glucocorticoids. By that definition the manopause is not a syndrome, but neither are some features of the menopause. Perhaps whether a syndrome even exists is not an interesting question. If the manopause is defined to include concomitantly appearing changes that may influence e a c h other and that become apparent in middle age, then it may indeed exist. Clinical features associated with aging in men include sexual dysfunction, hypogonadism, and psychological changes. In Fig. 1 the pie chart shows a separation of these three clinical features although they are really connected and
Although the menopause may be defined by the final occurrence of menses at a defined time in each woman, many biological changes occur in women prior to that last ovulation. Among these are loss of bone mineral density, lipoprotein alterations, increased body weight and altered composition, development of hot flashes, and alterations in mood and perceptions of well-being, in addition to the endocrine system changes. In men, although there is no comparable signal event associated with aging, analogous changes in bone density, lipoproteins, and psychological factors as well as alterations in endocrine status take place. The question of whether these and other changes represent a syndrome, a "manopause," or
MENOPAUSE:
BIOLOGY AND PATHOBIOLOGY
Erectile Dysfunction Manopause and Mental Health Psychological State and Sexual Function Conclusions References
111
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
112
KORENMAN ET AL.
A. A n a t o m i c C h a n g e s o f H y p o t h a l a m i c Pituitary Testicular System
FIGURE 1 Problems of the aging man. The parallel lines between the three pie sections are to indicate that the components interact with one another even though they may be independent. do influence each other. Hypogonadism affects mood, depression influences sexual function, anxiety affects androgen production, and erectile dysfunction affects mood and selfesteem. These will be discussed later in more detail. There are other changes that occur in men as they age, and to cover them fully would require an entire volume, rather than a chapter. Among the issues of concern are benign prostatic hyperplasia, prostate carcinoma, atherosclerosis and its sequelae, lipid alterations, changes in body mass and composition, altered glucose sensitivity, changes in muscle strength and composition, declining bone density, and the changes in endocrine systems other than the reproductive system. To cover all of these thoroughly is beyond the scope of this chapter. Therefore, we will deal primarily with androgens, sexual function, and psychological findings and introduce other elements more in passing. As the reader will see, the information available is inadequate in many areas. Connections between findings are hard to establish. Another difficulty is the emphasis of studies on men over the age of 60 years. However, in our analysis, we will emphasize investigations that include men from ages 40 to 60 years, the age range of interest. One obvious conclusion from this analysis is the need for much more research into aging of the male.
II. AGE-RELATED MALE
REPRODUCTIVE
CHANGES
I N THE
SYSTEM
Aging is often associated with significant histological and functional changes of the male reproductive system [1-3]. These changes are aggravated by a variety of chronic organ system diseases that are commonly found in elderly men [4]. Thus, the distinction between an age-related change and a disease state is sometimes blurred. Nevertheless, the preponderance of currently available data suggest that aging, independent of disease, is associated with altered reproductive function in men.
Striking structural changes occur in the testis with age [5]. These include thickening of the basement membrane of seminiferous tubules, peritubular fibrosis, thinning of the germinal epithelium, germ cell arrest, and narrowing and collapse of tubules with reduction in the number of Sertoli cells. However, areas of spermatogenesis are usually preserved until late in life. The number of Leydig cells in the testes may be reduced with age, although some studies have failed to find an age-related decrease in Leydig cell number [3,5]. The presence of inflammatory cells near degenerating seminiferous tubules suggests that autoimmune damage of testes occurs with age. The finding of increased serum sperm agglutinating antibodies with age is of questionable significance [3,5]. These antibodies may well be another nonspecific manifestation of age-related dysregulation of the immune system. The pituitary gland also undergoes histological changes with age. The incidence of pituitary adenomas and empty sellae increases with age [6], and numbers of eosinophilic cells decrease and numbers of basophilic and chromophobic cells proportionately increase [7,8]. Aging, particularly during the transition from youth to middle age, is associated with a significant decline in the number and size of cells producing growth hormone, with hypertrophy and relative hyperplasia of thyrotrophs [7,8]. The number of prolactinsecreting cells is not altered in aging men. Because of the large reserve capacity of the endocrine glands, the correlation between histological changes and changes in endocrine function is generally poor. However, some of the structural changes, such as loss of Leydig or Sertoli cells, undoubtedly contribute to the age-related alterations in hypophyseal-gonadal function.
B. N o r m a l P i t u i t a r y - T e s t i c u l a r P h y s i o l o g y The hypothalamic-pituitary testicular axis is outlined in Fig. 2. Testosterone, the principal circulating androgen, is secreted by testicular Leydig cells under luteinizing hormone (LH) stimulation. Sperm production and inhibin [a Sertoli cell hormone that inhibits follicle-stimulating hormone (FSH) secretion] are stimulated by FSH. The inhibitory effect of testosterone on LH secretion and of inhibin on FSH secretion establishes the negative feedback loop of the axis. Testosterone mostly slows the hypothalamic gonadotropin-releasing hormone (GnRH) pulse generator whereas estrogens derived from testosterone seem primarily to affect gonadotropin pulse height at the pituitary level [9]. The gonadotropin-secreting cells and Leydig cells are adapted to
CHAPTER 7 Changes in Aging Men
113 The biological effects of T are due to its interaction with androgen receptors or the interaction of its metabolites, such as DHT or estradiol (E2), with the androgen or estrogen receptor. In addition, DHT may act as antiestrogen by competing for the estrogen receptor [15]. These receptors then mediate the effects of those hormones on specific genes involved in the phenotypic expression of various tissue functions. Androgens can also act on various enzymes, especially in the liver through receptor-independent mechanisms. The biological effects of androgens and their active metabolites are manifest in essentially every organ system [ 16]. These biological effects, as well as the androgen economy, are altered with age.
FIGURE 2 The hypothalamic-pituitary-testicular axis and relation to peripheral responses. A simplified version of the interactions related to testosterone secretion and physiological actions. SHBH, Sex hormone binding globulin; FSH, follicle-stimulating hormone; GNRH, gonadotropin-releasing hormone; LH, luteinizing hormone.
the pulsatile release of GnRH. Thus, continuous LH or human chorionic gonadotropin stimulation of the testis results in down-regulation of LH receptors and loss of steroidogenic response [10]. Testosterone (T) secretion shows diurnal and seasonal variation [ 11 ]. The peak plasma concentration of T occurs at 6 : 0 0 - 8 : 0 0 a.m., falls slowly by about 35% during the day, and begins to rise again at about midnight. The plasma concentrations of T also tend to be higher in spring and summer months. In plasma, up to 54% of T is bound to sex h o r m o n e binding globulin (SHBG) and 45% is weakly bound to albumin. For most tissues, the available T is the sum of the albumin-bound fraction and the dialyzable fraction This sum of T fractions is referred to as the "bioavailable" testosterone (BT) [12]. SHBG also binds other 17fi-hydroxylated steroids, including estradiol, and its secretion by the liver is enhanced by estrogen and suppressed by androgens. Secreted T is metabolized to active or inactive compounds in various target tissues. In the liver, T is metabolized to inactive reduced metabolites, which are then conjugated to yield glucuronides or sulfates for excretion into the urine [13]. In adults, reproductive tissue T is converted to 5cedihydrotestosterone (DHT) via the enzyme 5a-reductase. In adipose tissue, in Leydig and Sertoli cells, and in certain brain nuclei, T is converted to estradiol by the aromatase enzyme, whereas muscle responds directly to T as the active hormone. In reproductive tissue, DHT is further metabolized to 3ce,17fi-androstanediol and to 3/3 metabolites. These metabolites undergo glucuronide conjugation in the prostate and may play a pivotal role in androgen removal from that organ [13,14]. Thus, specific metabolism of T in various tissues allows generation and clearance of different active compounds.
C. A g e - R e l a t e d
C h a n g e s in t h e H y p o t h a l a m i c -
Pituitary-Gonadal
Axis
In men, age-related changes in the reproductive system occur at multiple levels (Table I). Circulating testosterone declines longitudinally with age at the rate of 100 ng/dl/ decade of life [17-19]. This is accompanied by an increase in SHBG such that free and bioavailable testosterone concentrations drop to an even greater extent. The increase in SHBG may be due in part to the decreased mean concentration of T and maintenance of the E 2 level commonly seen with age. The age-related fall in unbound (free) T and bioavailable T (free and albumin bound fraction) with age
TABLE I Age-Related Changes in the Male Reproductive System Reproductive tissue structural changes
Testes: thickening of basement membrane of seminiferous tubules; peritubular fibrosis; germ cell arrest Prostate: hyperplasia, cancer Epididymus: regression of secretory epithelium Seminal vesticles: modest changes in epithelial components Hormonal changes
Decreased serum total testosterone level Decreased bioavailable testosterone Decreased clearance of testosterone Decreased accumulation of 5a-reduced steroids in reproductive tissues Increased plasma binding of testosterone Increased mean LH and FSH levels Decreased LH pulse frequency Increased incidence of both hypergonadotrophic and hypogonadotrophic hypogonadism Changes in sexuality
Decreased libido Increased latency Reduced frequency and rigidity of erections Decreased ejaculatory volume Reduced orgasmic contractions
1
1
4
K
O
R
E
[ 18,20] has been documented, although not in every study. As a result of declining T production and increased binding, mean serum gonadotropin concentrations rise slowly with age, although inappropriately low serum gonadotropin values are reported, suggesting that pituitary hypothalamic dysfunction is also common in elderly subjects [20-22]. The normal diurnal fluctuation in plasma T is attenuated in the elderly so that the morning peak is much lower [ 11 ]. Compared to young men, healthy older men have a reduced LH pulse frequency but the pulse amplitude is maintained. The age at which this begins is unknown. The reduced diurnal fluctuations of T with age are possibly due to the decreased frequency of LH pulses [23]. However, some studies have failed to find an age-related change in LH pulse frequency [24]. In addition to changes in LH pulse frequency, there is also an age-related decay in the 24-hr rhythm of LH and FSH, and the circannual rhythm of FSH is lost while the circannual rhythm of LH seems to be preserved [25,26]. The gonadotropin-suppressing activity of T or DHT is enhanced in healthy elderly men [27]. This may partly account for reduced FSH/LH response to GnRH in aging men [28]. It is noteworthy that Muta et al. [29] found that LH suppressibilty by T or estradiol is reduced in older men. This study, unlike that of Winters et al. [28], included subjects with relatively more severe primary testicular failure. In healthy aging men, treatment with clomiphene citrate stimulates LH and FSH secretion similarly to that in young men [30]. However, older men showed a smaller rise in serum testosterone, bioavailable T, and estradiol, suggesting a reduced testicular response to LH. Primary alterations in testicular function are supported by the observed histological changes in Leydig cell or Sertoli cell number and reduced spermatogenesis. Sertoli cell inhibin production declines significantly by the fifth decade [31 ]. This accounts in part for the increase in mean FSH levels found in aging men. Recent studies have indicated that inhibin B, which is inducible by exogenous FSH, is the only inhibin detectable in adult men and appears to be the physiologically relevant inhibin [32]. The Leydig cell response to stimulation with exogenous human chorionic gonadotropin (hCG) or endogenous LH is also reduced with age [33,34], at least partially due to a decrease in pregnenolone production, which limits the availability of substrate for conversion to testosterone. Aging is often associated with alterations in androgen action and metabolism. The metabolic clearance rate of both testosterone and DHT is reduced [35]. T reduction to 5/3 metabolites is increased. There is also an increase in testosterone conversion to estradiol [36]. This, together with the reduction in E 2 clearance with age, accounts for the increased Ez/T ratio in aging men [36]. It is believed that the enzyme responsible for converting T to DHT, namely 5ce-reductase, is inducible by androgens [37]. This concept is also sup-
N
M
A
N
ET AL.
ported by the studies of scrotal testosterone patches showing that increased conversion of T to DHT locally was possibly due to induction of 5ce-reductase in scrotal skin [38]. This observation is relevant to changes found in aging in which the concentration of 5ce-reduced androgens in reproductive tissues is diminished [39]. Thus, it is possible that the lower circulating androgen levels in older men permits diminished 5a-reductase activity and therefore reduced responsiveness of the reproductive tissues to androgens. However, the capacity of DHT-degrading enzymes in prostatic epithelium, such as 3ce- and 3fl-hydroxysteroid reductase, is also reduced with age [40]. The effect of those changes on overall androgen metabolism is small. It is noteworthy that inherent genetic differences determine sensitivity to T effects. The number of CAG trinucleotide repeats in the androgen receptor gene correlates with age of onset of prostate cancer [41]. Although many factors, including coexisting diseases and nutritional changes, have been linked to the age-related decline in T production, careful studies suggest that chronic illness plays a minor role and that age is the more powerful risk factor. However, there is a wide individual variation in the rate of decline in reproductive function and some men can maintain both normal T production and spermatogenesis well into advanced age [ 17,18].
D. A n d r o g e n E f f e c t s With age there is a loss of lean body mass, primarily muscle mass. Androgens can cause muscle hypertrophy without altering muscle cell number [16]. They also potentiate muscular growth produced by GH [16]. Thus, it is possible that some of the age-related loss of muscle mass is related to T and GH deficiency. Certain muscle groups, such as cardiac muscle or diaphragm, that are not highly sensitive to androgens or are independent of lifestyle changes do not show a significant loss of mass with age. This suggests that regular exercise would prevent some of the age-related losses of muscle mass. However, other muscle groups, notably temporalis, levator pubii, ischiocavernosus, and bulbospongiosus muscle, are androgen dependent. These muscle groups atrophy as androgen availability to tissues decreases with aging.
E. O t h e r E f f e c t s o f A n d r o g e n s Androgens affect the hematopoietic system, particularly erythropoiesis, and may have some antiinflammatory effects. Androgens stimulate erythropoietin production and also have direct stimulatory effects on the stem cells of the bone marrow [42]. Thus, in the elderly, reduced T may contribute to reduced hematopoiesis. Although androgens alter
CHAPTER7 Changes in Aging Men
115
the expression or activity of various hepatic and renal enzymes, these effects do not seem to be of any relevance to the changes seen in these organs with age. Similarly, the relevance of the effect of T on feeding behavior or body fat distribution with aging is not clear. In general, low doses of T increase adipose tissue, especially in the upper body area, whereas high physiological doses of T suppress adipose tissue mass and lipoprotein lipase as a result of aromatization to estrogens [43-45]. It is possible that the changes in adipose tissue distribution with age are related to altered T secretion. Testosterone therapy may also reduce adiposity and improve muscular strength [46,47]. The role of T or DHEA in preventing osteoporosis in men is modest compared to the known role of estrogens in women. Nevertheless, in aging men there is a direct relationship between bone mass and plasma T levels [48,49]. Mild to moderate androgen insufficiency occurs frequently in older men. Considering the diversity of the biological actions of androgens, it is likely that some of the common physiological changes with age are related to androgen deficiency. Therefore, it is possible that some older men with androgen deficiency would benefit from androgen replacement therapy. Multicenter, placebo-controlled trials are needed to establish the long-term safety and efficacy of T treatment in men.
III. A G E - R E L A T E D IN OTHER
HORMONAL
CHANGES SYSTEMS
The biological consequences of altered androgen status with age are often modulated by the other hormonal changes commonly found in aging men (Table II). These additional hormonal systems will be discussed briefly.
A. G r o w t h H o r m o n e The changes of growth hormone (GH) physiology with age and the known biological correlates of these changes constitute the rationale for the preliminary attempts at "rejuvenating" older subjects with GH therapy. Although pituitary GH content may not change with age, sleep-induced GH secretion, the peak amplitude, and the 24-hr plasma levels of GH are reduced in older subjects [50-53]. The response to a variety of GH stimulators is also altered with age in some but not all studies. The response to GH releasing hormone (GHRH) [54,55] or insulin-induced hypoglycemia [56] may be reduced. The reduction in GH production could be partly attributed to reduced growth hormone releasing hormone receptors [57] and more importantly to increased pituitary sensitivity to inhibition by somatostatin [58].
TABLE II Hormonal Changes Commonly Found in Aging Men System/activity
Change
Hypothalamic-pituitary unit Growth hormone Basal Nocturnal peak IGF-I IGFBP- 1 IGFBP-3 ACTH Basal Response to CRH TSH Basal Response to TRH LH/FSH
N or decreased Decreased in males Decreased or N
Thyroid
See Table III
Adrenal Cortisol basal concentrations Sensitivity to dexamethasone supression DHEA + DHEA sulfate concentrations Aldosterone concentrations
N Decreased Decreased Decreased
Testis
See Table I
Calcium/bone PTH 1,25-(OH)2D
Decreased Decreased
Water metabolism Arginine vasopressin in response to Osmotic stimuli Baroreceptor stimuli Atrial natriuretic factor
Decreased N Decreased
Carbohydrate metabolism Insulin Glucagon
Increased Increased
N in males a Decreased Decreased Decreased Decreased N N
aN, No change.
In addition to the changes in GH secretion, there are changes in insulin-like growth factor-I (IGF-I) and its binding proteins. Thus, with age, IGF-I is decreased and IGF binding protein-3 (IGFBP-3) is also decreased [59]. In contrast, IGF binding protein- 1 (IGFB P- 1), an inhibitor of IGF-I bioactivity, is increased with age [60]. The changes in GH secretion, along with reduction in bioactivity of IGF-I, may contribute to the age-related increase in bone loss and muscle wasting. However, so far GH replacement has failed to demonstrate clear anabolic effects in elderly subjects. Of note is that the age-related changes in GH secretion vary greatly among populations. Nutritional factors and coexisting diseases often influence GH secretion. There appears to be significant gender-related differences; in women, basal GH levels decline with advancing age [61] whereas in men they remain unchanged [62]. The more marked fall in basal GH that occurs in older women may be related to estrogen
116
KORENMAN ET AL.
deficiency. Although testosterone, through its conversion to estrogen, has a positive modulating influence on GH secretion, the age-related reduction in T is not sufficient to cause a decrease in GH levels; perhaps, because of an increased conversion rate, estrogen levels do not fall in men as they age and may not be reduced in the GH cells.
B. T h y r o i d H o r m o n e E c o n o m y Aging is commonly associated with an altered thyroid hormone (TH) economy (Table III) [62a], although the feedback control system maintains the same plasma concentrations of total or free T 4 and T 3. The production and clearance rates of thyroid hormones are decreased in parallel [63-65]. The TH binding to serum proteins is usually not altered [63]. The pituitary thyroid-stimulating hormone (TSH) response to thyrotropin-releasing hormone (TRH) may be reduced in aging men but is normal in elderly women [60,63]. Thyroid gland response to endogenous or exogenous TSH may also be reduced. The incidence of primary thyroidal failure, as evidenced by increased plasma TSH levels, increases with age, especially in women. It has been estimated that 4.4% of the population over the age of 60 years has primary thyroidal failure [66]. In addition to the changes in thyroid hormone economy, the tissue responsiveness to thyroid hormones is reduced
T A B L E II I
Age-Related Changes in Thyroid Hormone Economy a
Parameter h Radioactive iodine uptake
Changes Decreased
T 4 production
Decreased or unaltered
T 3 production
Decreased or unaltered
T 4 degradation
Decreased or unaltered
T 3 degradation
Decreased or unaltered
Serum T 4 concentration (total or free)
Unaltered
Serum T 3 concentration (total or free)
Unaltered or decreased
Serum thyroid hormone binding capacity (T 3 resin uptake)
Unaltered
Serum TSH
Unaltered or increased
Circadian TSH variation
Decreased
TSH response to TRH
Unaltered or decreased, especially in men
TSH rise in response to thyroid hormone deficiency
Decreased
Thyroid response to TSH
Decreased or unaltered
a Reprinted from Mooradian and Wong [62a], with permission of the publisher. b T4, Thyroxine; T 3, triiodothyronine; TSH, thyroid-stimulating hormone, TRH, thyrotropin-releasing hormone.
with age [65]. This is partly related to reduced TH transport across the cellular plasma membrane [67,68] and the altered biology of transcription factors involved in the expression of genes responsive to TH [69]. The TH receptor affinity or capacity is not significantly altered with age [68]. The clinical correlates of the age-related changes in tissue responsiveness to TH are the finding of classical features of hypothyroidism in biochemically euthyroid elderly individuals. In addition, postprandial thermogenesis is also blunted partly because of loss of fl-adrenergic receptor activity in brown adipose tissue [70]. It is also possible that the age-related reduction in thyroid hormone responsiveness contributes to reduced metabolic rate and blunted postprandial thermogenesis. These changes may partially account for the cold intolerance and increased risk of hypothermia in the elderly.
C. A d r e n a l P h y s i o l o g y l.
GLUCOCORTICOIDS
In humans, the hypothalamic-pituitary adrenal axis remains intact with age. The plasma cortisol level remains stable [71], and the pituitary adrenocorticotropic hormone (ACTH) response to exogenous corticotropin releasing hormone (CRH) or to intravenous metyrapone does not change [72,73]. Cortisol production and clearance rates are both decreased proportionately with aging. The circadian rhythm of cortisol secretion remains intact, although in men over 40 years, a phase advance in cortisol secretion is observed [74]. In this group, the peak and nadir of cortisol secretion occur approximately 3 hr earlier than in younger controls [74]. The cortisol response to ACTH or to insulin-stimulated hypoglycemia is usually unaltered in healthy elderly subjects [56]. An increased cortisol response to CRH, despite the normal ACTH response, suggests a diminished sensitivity of ACTH to negative feedback by glucocorticoids in older men. The early phase of ACTH inhibition by hydrocortisone is blunted in the elderly (>age 65 years), although the late phase remains intact. This could be related to changes in blood-brain transport of cortisol [72]. The elderly have a higher serum cortisol response to perioperative stress or during depression and the dexamethasone suppression test often fails to yield the expected drop in plasma cortisol levels [75]. It is tempting to speculate that this relative increase in cortisol level with age may contribute to some age-related changes in body composition, notably osteopenia. 2.
MINERALOCORTICOIDS
Aldosterone secretion decreases with age, probably because of reduced plasma renin activity [76]. This reduction is evident both at basal conditions and during stimulation with salt restriction, upright posture, or ACTH [76]. An
CHAPTER7 Changes in Aging Men age-related reduction in aldosterone clearance partially offsets the decreased aldosterone production rate. The reduced plasma renin activity and aldosterone production may contribute to the orthostatic hypotension commonly found in the elderly and may expose them to a higher risk of developing hyperkalemia following administration of angiotensinconverting enzyme inhibitors. 3. ADRENAL ANDROGENS
One of the most dramatic age-related changes in the hormonal system is adrenal androgen production. In men, dehydroepiandrosterone (DHEA) secretion declines progressively between 20 and 96 years of age [77]. The serum level of DHEA in older men is approximately 5 - 3 0 % of that seen in young men [77]. This "adrenopause" is probably the result of reduced 17,20-desmolase activity with age. Weight loss in overweight men but not women is associated with a 125% rise in serum DHEA sulfate, which suggests that agerelated increase in adiposity or insulin resistance may contribute to reduced DHEA levels in aging men [78]. The biological implications of the decline of DHEA in aging humans are still not clear. Experiments in animals who do not secret DHEA suggest that DHEA may be implicated in longevity and have protective effects in tumorigenesis, atherosclerosis, and age-related memory disturbances [79]. In one epidemiological study, death from cardiovascular disease in men over the age of 50 years was inversely related to DHEA sulfate levels [80]. DHEA administration reduces plasma plasminogen activator inhibitor type 1 and tissue plasminogen activator concentrations in men [81]. DHEA also inhibits platelet activity [82]. These effects may help prevent heart disease in men. More interventional studies are needed to establish the clinical relevance of DHEA in the biology of aging. 4. ADRENAL MEDULLA
Elevated plasma levels of epinephrine and norepinephrine (NE) have been found in healthy octogenarians compared to younger subjects [83,84]. Plasma dopamine levels do not change with age. The increased NE levels are due to an increased production and decreased clearance rate. This is accompanied by a decrease in platelet az-adrenergic receptors [85] and cardiac fl-adrenergic transmission [86]. The NE response to upright posture, during the cold pressor test, following glucose ingestion, and during insulin tolerance testing is increased in the elderly, whereas the NE and epinephrine response to exercise may be reduced in healthy elderly men [84]. The clinical consequences of these changes are not apparent but they may contribute to orthostatic or postprandial hypotension [87]. They could also be related to increased vascular resistance and therefore contribute to hypertension and the need for after-load reduction, especially in those with congestive heart failure.
117 D. C a l c i u m a n d B o n e M e t a b o l i s m Whereas age-related bone loss is a common phenomenon in both men and women, the process is accelerated by coexisting hormonal deficiencies, notably estrogen deficiency during the menopause and androgen deficiency in men. In healthy men, radial bone mineral content decreases by 1% per year whereas vertebral bone mineral content decreases by 2.3% per year [88]. Parathyroid hormone (PTH) secretion increases with age, as production of 1,25-dihydroxycholecalciferol (calcitriol) and intestinal calcium absorption are reduced [89]. Nutritional deficiency vitamin D and limited exposure to sunlight, coupled with reduced conversion of 25a-hydroxycholecalciferol to calcitriol, contribute to the reduced calcitriol levels seen in elderly men [89,90] Age-related osteopenia is the result of multiple factors, including altered dynamics of bone cell populations inherent to aging p e r s e aggravated by multiple nutritional and hormonal changes including deficiencies of sex steroids, GH, IGF-I, and calcitriol.
E. C a r b o h y d r a t e M e t a b o l i s m One of the major consequences of the age-related hormonal changes is the emergence of type 2 diabetes. This is the result of both altered insulin secretion and action with age [91]. These changes may be partly due to decreased physical activity and altered body composition favoring accumulation of central adiposity. The incidence of type 2 diabetes increases progressively with age starting at about age 40. Approximately 20% of the population in the United States over the age of 65 years has type 2 diabetes mellitus and at least 40% have glucose intolerance [91,92]. A decline in insulin secretory capacity with age along with reduced insulin sensitivity is common [91]. The plasma levels of glucagon and its clearance rate remain unaltered. Clinical diabetes, especially when poorly managed, causes a variety of complications that are associated with a poor quality of life. Increased glycation of various proteins and enhanced lipid peroxidation accelerate the age-related deterioration of various organ systems. In particular, body composition and vigor are adversely affected. Older subjects with diabetes are at increased risk of dehydration and malnutrition. Optimization of blood glucose control reverses most of the short-term and possibly long-term complications of diabetes.
E Water Metabolism The age-related changes in water and electrolyte homeostasis are summarized in Table IV. The total body water and
118
KORENMAN
TABLE IV
Biological Changes with Age
Body composition Increased: body weight until the sixth decade, thereafter declines; central adiposity Normal:extracellular fluid volume Decreased: lean body mass, bone mass, muscle mass; intracellular and total body water. Cardiovascular system Increased: stroke volume, end-diastolic volume, systolic blood pressure Normal: cardiac output, myocardial contractility Decreased: heart rate, ejection fraction Pulmonary function Increased: residual volume, closing volume Normal: total lung capacity Decreased: vital capacity, arterial PaO2, elastic recoil, maximum expiratory flow rate, maximum voluntary ventilation Digestive system Increased: frequency of teritary contractions in esophagus Normal: motility of stomach and intestine Decreased: salivation, taste, peristalsis of esophagus, acid/pepsin production; colonic motility Hormonal system See Tables I-III Central nervous system Increased: incidence of neurodegenerative diseases and depression, difficulty in learning new tasks Normal: overall intelligence Decreased: speed of cognitive processing, memory I-Iematopoietic/immune system Increased: incidence of anemia, certain hematological malignancies Normal: complete blood count, B cells, macrophages Decreased: progenitor cells, certain components of complement, T cells, intracellular bactericidal activity
intracellular fluid volume are decreased with age while extracellular blood volume is maintained. Elderly men have reduced thirst perception [93]. Basal arginine vasopressin (AVP) secretion may increase with age [94]. The osmolar threshold (the level of plasma osmolarity that will initiate AVP secretion) is lower in the elderly. However, AVP responses to volume and pressure changes are reduced, and the renal response to AVP is also blunted with age [94]. This results in a reduced capacity to conserve water that predisposes the elderly to dehydration, especially when water access is limited or during excess water losses as a result of intercurrent illness. A reduced capacity to generate angiotensin, a potent stimulator of AVP and thirst, also limits the ability of older subjects to maintain water homeostasis. The ability to maintain salt and water balance is further compromised by changes in atrial natriuretic peptide (ANP) secretion [95]. Baseline ANP is increased in elderly subjects and the expected reduction in ANP following dehydration is blunted. The natriuretic effect of ANP is probably preserved, although hemodynamic responses to ANP may be reduced
ET AL.
[95]. The increased ANP secretion with age may contribute to the suppression of plasma renin activity and aldosterone secretion.
IV. O T H E R RELATED
CHANGES
WITH
TO HORMONAL
AGING
FACTORS
A variety of other physiological and structural changes occur with age [96] (Table IV). Some of these are probably related to processes inherent to aging per se, whereas others are secondary to lifestyle changes or nutritional and hormonal alterations. Consequent to the loss of muscle mass, the basal metabolic rate is reduced with age [96]. The reduced skeletal muscle mass with age, along with changes in cardiovascular and pulmonary physiology, results in reduced exercise capacity and low maximum oxygen consumption (VO 2 max). It appears that the changes in pulmonary function are more important than the changes in the cardiovascular system in limiting exercise capacity in the elderly [97]. The partial pressure of arterial oxygen (PaO2) declines steadily with age while P a C O 2 is not significantly altered [97]. Weak respiratory muscles, decreased lung compliance, and increased chest wall stiffness account for most of the age-related changes in pulmonary functions, some of which may be related to cigarette smoking. The other components of age-related loss of lean body mass are bone loss and altered body water content. Hormonal factors, such as loss of androgens, GH, and IGF-I and nutritional factors such as calcium and vitamin D deficiency, along with genetic factors, account for the bone loss (Table V) [98].
TABLE V
Epidemiological Correlates with Erectile Dysfunction a
Positively associated with ED Age Cigarette smoking Depression, inward looking or with expressed anger Diabetes, treated with medications (more severe) Cardiovascular disease, treated Use of vasodilators (however they were defined) Not grossly associated with ED Hypertension Alcohol intake over a wide range Allergies Serum cortisol or DHT level Inversely associated with ED Dominance DHEA level in blood HDL cholesterol level in blood a Adapted from Feldman et al. [98].
CHAPTER7 Changes in Aging Men V. A N D R O G E N
EFFECTS
AND REPRODUCTIVE
119 ON SEXUAL
FUNCTION
The biological actions of androgens are far reaching, and reduced androgen availability in aging men contributes to a host of biological changes. Testosterone acts on most body tissues. However, the biological effects in aging tissue do not always correlate with plasma concentrations, because local tissue factors such as conversion to DHT or E 2, or metabolism to glucuronides, modulate its activity. Thus, although T concentrations may decrease with age, some T-sensitive organs, especially the prostate, commonly undergo hyperplasia. Although androgens have a permissive role for prostatic tissue growth and development, their precise role in benign prostatic hyperplasia (BPH) is not clear. This may be related to altered T metabolism in aging prostate such that in B PH the ratio of DHT to 3ce-androstanediol is increased compared to normal prostate [16]. The androgen dependency of the other accessory sex organs is also well established. Secretory epithelium of epididymus regresses following castration. Exogenous androgen treatment restores some but not complete secretory function [16]. The seminal vesicles, in particular the epithelial component, are androgen dependent and so is spermatogenesis. However, these organs change modestly with age, suggesting that the age-related reduction in T availability is not sufficient to result in a clinically relevant change in these organs. Androgens do not appear to alter erections in response to erotic films [99,100]. To what extent do androgens play a role in erectile function in the adult male and what are the effects of declining androgen availability with aging on sexual function? Of course androgens are necessary both in utero and after birth for the proper development of the male external and internal genitalia. Growth of the penis and testes during puberty is also totally dependent on adequate androgen availability. It has already been noted that adequate T concentrations are necessary for normal libido [ 101 ]. Pharmacological reduction of circulating T in young men reduced sex drive and responses although the erectile response to erotic stimuli was unchanged [ 102]. It has been reported that severe hypogonadism is responsible for less than 7% of cases of erectile dysfunction (ED) [103,104]. An important role for T is suggested by studies demonstrating in other species that castration reduces erectile capacity, which can be preserved with dihydrotestosterone (DHT) [105]. In rats, T facilitates centrally mediated erections and yawning (a sexual response) [106]. Androgens stimulate the sexually dimorphic brain nuclei and increase the size and dendritic spread of the spinal cord motor neurons innervating the bulbospongiosus and ischiocavernosus muscles [107]. They may also affect the penile vascular re-
sponse mechanism [ 108]. In men, androgens are responsible for normal seminal fluid and prostatic secretions and the frequency of nonerotic or nocturnal erections [99]. The frequency of nocturnal penile erections correlates with circulating testosterone [109]. The reduction in testosterone status after age 45 is generally paralleled by a gradual decline in sexual desire, arousal, and activity [110]. The magnitude of this decline varies widely [ 111] and is frequently the result not of aging but of other factors, including medications, depressed mood, alcohol use, obesity, and chronic illnesses (e.g., diabetes, vascular disease). Despite reduced libido, sexual enjoyment and satisfaction do not decline with age [112]. An important predictive factor for sexual enjoyment in aging men, as in younger men, is the quality of the marital relationship [ 112]. As distinguished from libido, the role of a decline in T in ED in men as they age has not been well characterized, but in one study there was no relationship between the mild hypogonadism of aging and ED. Both conditions were common, but independently segregated [21 ]. Many men seek androgen replacement to improve erectile function, but the use of testosterone for this purpose in eugonadal men is usually unsuccessful [20,113]. Testosterone supplementation does, however, increase sexual interest [ 114].
VI. NONERECTILE SEXUAL DYSFUNCTIONS Sexual dysfunction in men consists of a small group of problems including early or premature ejaculation, retarded or lack of ejaculation, loss of libido, and erectile d y s function.
A. E j a c u l a t o r y D y s f u n c t i o n The mechanism of ejaculation encompasses seminal emission, ejaculation, and bladder neck closure. Afferent stimuli include activation of higher centers of sexual response to reach a threshold. Sympathetic nerves then cause smooth muscle contraction in the epididymis, vas, seminal vesicle, and prostate to produce filling of the prostatic urethra with the seminal emission. Finally, contraction of the bulbospongiosus and ischiocavernosus muscles and contraction of the bladder neck lead to propulsion of the semen out of the penis while the sensation of orgasm is experienced. 1. PREMATURE EJACULATION Early or premature ejaculation is perhaps the most common disorder of sexual function in men, affecting at least a third. Ejaculation usually occurs very close to the time of
120
K O R E N M A N ET AL.
vaginal penetration, in the most severe cases before vaginal penetration. Milder cases are associated with ejaculation after a few seconds of thrusting. This condition improves with sexual experience and age but persists in a substantial number of men well into the fourth and fifth decade. It is thought to be due to anxiety associated with sexual activity. For men with premature ejaculation, sex therapy behavioral techniques are beneficial but usually do not suffice [115]. Pharmacological interventions targeted at augmentation of serotonin function have been reported in numerous studies to be highly effective for this condition. In particular, the use of standard doses of the serotoninergic agents fluoxetine, paroxetine, sertraline, and clomipramine have been found to prolong significantly latency to ejaculation [116-121 ], with improvement noted as early as 1 week following initiation of medication [120]. One study found that clomipramine produced the greatest increase in latency time, although it was associated with more side effects than the serotonin reuptake inhibitors (fluoxetine, paroxetine, and sertraline). Following discontinuation of the serotonin reuptake inhibitors, premature ejaculation has been observed to recur in 90% of treated men [ 117]. 2. RETARDED EJACULATION Retarded ejaculation, which is unusual, is sometimes found in association with the use of antipsychotic drugs and with certain antidepressives [ 122,123]. Sometimes removal or a change of medication will reverse the condition. Retrograde ejaculation is one of the consequences of prostate surgery. Most commonly this is due to surgical damage to the vesical-urethral sphincters, making it easier to pass the ejaculate into the bladder than through the urethra. This may also occur with diabetic autonomic neuropathy, in which the same sphincters become dysfunctional. Retrograde ejaculation is usually treated with reassurance but many men complain that their sensation of orgasm and release is substantially reduced in the absence of an ejaculate.
VII. ERECTILE
DYSFUNCTION
Erectile dysfunction, on the other hand, is progressively common with age and has many etiologies and risk factors. There have been extensive publications and reviews of the field [ 124-127] and we will not try to recapitulate here the history of research in the area. In the past year the problem of ED has mutated from an underdiagnosed disorder managed by a few physicians to a public phenomenon characterized by media frenzy, numerous jokes, and an intense debate featuring patients, health care providers, and government, regarding whether treatment of ED with sildenafil or other oral agents should remain covered by insurance. What kind of problem is ED and how often does it occur?
A. E p i d e m i o l o g y At the 1993 National Institutes of Health Consensus Development conference on ED [ 128], erectile dysfunction was formally defined as "an inability of the male to achieve an erect penis as part of the overall multifaceted process of male sexual function." Although the adopted definition seemed refreshingly simple and useful at the time, in practice it has become too vague. For example, how erect? Does a nonrigid but usable erection count as ED or normal erectile function? How is a full erection for masturbation and on awakening, but no erection in the presence of a partner, to count? What should we call variable erectile response? These questions pertain not only to research determining the prevalence of the condition but also to difficult questions as to support of the treatment of ED by health insurers. Feldman and colleagues in the Massachusetts Male Aging Study (MMAS) [98] developed a nine-point questionnaire and divided ED into three levels of dysfunction, with the most severe being a complete absence of sexual response and "mild" being an occasional failure of certain aspects of response. By these criteria, in a community-based group of men from ages 40 to 70 years, 9.6% had complete ED, 25.2% had moderate ED, 17.2% had minimal Ed, and 48% had no ED. Although this approach engendered a degree of criticism, until the advent of sildenafil this partition was effective in the selection of patients for treatment, because men usually wished to be treated only if they were seriously affected by the problem. With the advent of sildenafil, the target population could conceivably become 52% of men ages 4 0 - 7 0 years and much higher percentages of older men [129]. If, of the 110 million American males over age 40 years, 50 million had ED, and they wanted to have sex once weekly, that would require 52 • 50 million pills at $8.00/pill or nearly $21 billion/year for this single indication. Obviously, a more precise, medically determined objective diagnosis of ED is required. Table VI lists factors found in the MMAS [98] to be associated with an increase of ED and factors decreasing ED. Factors inhibiting sexual function include coronary artery disease and diabetes, especially if treated (more severe) and if associated with smoking cigarettes. Depression and anger, whether internalized or expressed, were highly associated with ED. This is of particular importance because depression is very strongly associated with loss of libido (see above) and is greatly underdiagnosed in men, especially in middleaged men (see below). Studies of populations in the medical system, however, although biased because of the patterns of referral and the expertise of the practitioner, demonstrate a high prevalence of hypertension, coronary heart disease, and diabetes, as well as treatment for each, associated with ED [103,104]. Gener-
CHAPTER 7 Changes in Aging Men
121
TABLE VI Condition
Criteria for Depressive C o n d i t i o n s Symptoms
Duration
Major depressive episode
Five or more of the following symptoms present for the same 2-week period (must include symptom 1 or 2): 1. Depressed mood most of the day nearly every day 2. Markedly diminished interest or pleasure in almost all activities 3. Significant weight loss when not dieting, or weight gain 4. Insomnia or hypersomnia 5. Psychomotor agitation or retardation 6. Fatigue or loss of energy 7. Feelings of worthlessness or excessive/inappropriate guilt 8. Diminished ability to think/concentrate, or indecisiveness 9. Recurrent thoughts of death
2 weeks
Dysthymic disorder
Depressed mood most of the day more days than not for at least 2 years (two or more of the following symptoms while depressed): 1. Poor appetite or overeating 2. Insomnia or hypersomnia 3. Low energy or fatigue 4. Low self-esteem 5. Poor concentration/indecisiveness 6. Feelings of hopelessness
At least 2 years
Adjustment disorder with depressed mooda
Development of emotional or behavioral symptoms in response to an identifiable stressor occurring within 3 months of the onset of the stressor; predominant manifestations are depressed mood, tearfulness, or feelings of hopelessness (these symptoms are clinically significant as evidenced by either of the following criteria): 1. Marked distress in excess of what would be expected from exposure to the stressor 2. Significant impairment in social/occupational functioning
Subclinical depression
Depressive symptoms that do not meet criteria for major depression, dysthymia, or adjustment disorder with depressed mood
a
Occurs within 3 months of stressor and does not persist beyond 6 months after stressor terminates
a During the 2-year disturbance, the person has never had a major depressive episode and has never been without the above symptoms for more than 2 months. Symptoms must cause significant distress or impairment in functioning and are not due to the effects of a substance or general medical condition. b The disturbance should not meet criteria for another psychiatric disorder and does not represent bereavement. Once the stressor has terminated, the symptoms do not persist more than 6 months.
ally, in those studies, the definition was limited to those with a c o m p l e t e inability to c o m p l e t e sexual i n t e r c o u r s e for at least 3 m o n t h s . Other clinical associations with E D i n c l u d e d pelvic surgical p r o c e d u r e s and m a j o r p e r i p h e r a l vascular disease as well as n e u r o l o g i c a l disorders s u c h as m u l t i p l e sclerosis. T h e drugs associated with E D i n c l u d e d p r i m a r i l y vasodilators, t r e a t m e n t s for d e p r e s s i o n and psychosis, and h o r m o n e s or drugs affecting the reproductive e n d o c r i n e syst e m [98,122,123].
B. Erectile Mechanism W h e n at rest, the penis m a i n t a i n s a state of flaccidity t h r o u g h a - a d r e n e r g i c a l l y m e d i a t e d c o n t r a c t i o n of c a v e r n o s a l and vascular s m o o t h m u s c l e , inhibiting b l o o d flow into the organ [130,131 ]. As the result of an erotic stimulus, received t h r o u g h one or m o r e of the five senses or via m e m o r y (fan-
tasy), inhibition of the s y m p a t h e t i c d i s c h a r g e takes place and a p a r a s y m p a t h e t i c d i s c h a r g e is initiated, with p r e s y n a p t i c t e r m i n a l s in the pelvic p l e x u s [ 1 3 2 - 1 3 4 ] . Postsynaptically, the signals travel by n o n a d r e n e r g i c , n o n c h o l i n e r g i c ( N A N C ) nitric oxide (NO) nerves to t e r m i n a t e in the s m o o t h m u s c l e of the c a v e r n o s a l arteries and t r a b e c u l a r sinusoids [135]. (Fig. 3). T h e s e m u s c l e s relax w h e n N O stimulates g u a n y l y l cyclase to c o n v e r t G T P to cyclic GMP. In s m o o t h m u s c l e , cyclic G M P inhibits Ca entry and facilitates Ca loss [136, 137]. In the a b s e n c e of sufficient C a 2+ s m o o t h m u s c l e relaxes, allowing the heart to p u m p m u c h m o r e b l o o d into the corpora, i n d u c i n g penile swelling. O t h e r n e u r o t r a n s m i t t e r s that have b e e n related to erectile f u n c t i o n include prostaglandin E 1 (PGE1) and other stimulators of a d e n y l y l cyclase, vasoactive intestinal p e p t i d e (VIP), endothelin, calcitoninrelated peptide, and histamine. T h e y p r o b a b l y play a m i n o r role in the h u m a n u n d e r p h y s i o l o g i c a l conditions. I n c r e a s e d inflow of b l o o d alone will not result in an
122
KORENMAN ET AL.
FIGURE 3 Neurogenic mediation of penile vasodilatation via smooth muscle relaxation. On the left the postganglionic autonomic nerve is seen stimulating a smooth muscle cell, on the right. The group of cells in the left lower section represent endothelial cells. ARG, Arginine; CIT, citrulline, GC, guanylyl cyclase; DHT, 5a-dihydrotestosterone EFS, electric field stimulation (nerve discharge); ACH, acetylcholine; PGE~, prostaglandin E I.
erection; it requires nearly complete inhibition of venous return. That is accomplished passively by the unique anatomy of the penis, in which the expanding corporal sinusoids compress the subtunical plexus of veins draining the corpora cavernosa against the unyielding tunica albuginea [138]. The subtunical plexus, in turn, is drained through veins penetrating the tunica albuginea that are also compressed during stimulation, so that at maximum erection, penile blood flow is nearly zero [ 139,131 ]. Penile blood pressure may exceed systemic pressure at that time as a result of contraction of the ischiocavernosus muscle, which acts as a constriction ring at the base of the penis [ 140]. Ejaculation is neurally mediated in response to the filling of the prostatic urethra with semen and achievement of the "orgasmic plateau" of sexual stimulation, via contraction of the bulbospongiosus muscle [ 140].
C. Inadequate Erectile Function What happens to cause ED? First, there can be inhibition of the central nervous system centers mediating the response to erotic stimuli. Both testosterone deficiency (see above) and depression (see below) reduce libido substantially and together they are responsible for the majority of cases of reduced sexual interests in adult men. Second, the integrity of the neural pathways mediating an erection can be interrupted by spinal cord injury, pelvic surgery (usually due to resection of a prostate or colon cancer),
or autonomic neuropathy such as in Type I diabetes mellitus, or by primary neurological diseases such as multiple sclerosis [ 141 ]. A number of medications and recreational drugs affect the neural response at the periphery so as to contribute to ED [98,122,123,142,143]. Third, there may be a failure of response to the neural signals as a result of diminished NO synthesis, which has some relation to intact neural pathways, adequate androgen availability, and cavernosal smooth muscle integrity. This is commonly found in association with diabetic ED [105,144,145]. There is also evidence of enhanced contractility due to increased ce-adrenergic sensitivity in ED [146]. Fourth, ED is very commonly associated with abnormalities of the intrinsic tissues of the corpora cavernosa, including disrupted muscle fibers, an increase of dense connective tissue in the perisinusoidal area, and a reduction of tunical elastic fibers, which probably prevent adequate compression of the subtunical venous plexus [144,147-149]. This is commonly associated with atherosclerotic disease and diabetes and is thought to be due in part to ischemia. Fifth, the blood supply to the penis may be compromised by arterial atheromatous disease, a very low cardiac output, or arteriolar disease [ 150]. These conditions, once thought to be common and irreversible, probably account for a small proportion (less than 20%) of cases of ED. Failure of venous occlusion is very common in ED [ 151, 152], and although once it was considered to be a common etiologic factor, it is now believed to be largely a conse-
CHAPXER7 Changes in Aging Men quence of an inadequate filling rate and a degree of scarring in the perisinusoidal tissues except in cases of penile trauma, in which damage to large vessels is not uncommon, and in Peyronie's disease, in which peripheral fibrous plaque formation often inhibits venous compression [ 148]. Penile vein ablative surgery has been generally unsuccessful in the management of ED. How conditions associated with a high incidence of ED [98,124] produce the effects listed above is not fully understood, and until a minimally invasive acceptable method of biopsying the corpora cavernosa is developed we have no good way to correlate penile structure and ultrastructure with function and disease.
D. D i a g n o s t i c A s s e s s m e n t With the advent of sildenafil and other oral medications to come, the role of the health system in the diagnosis and treatment of ED has changed. No longer must physicians carefully elicit information about sexual function from reluctant patients. Rather, they often have to ensure that the patient requesting treatment indeed has an erectile problem as opposed to other sexual disorders or a desire for a somewhat enhanced lifestyle. Prior to initiating therapy, physicians must take a detailed sexual history that includes the nature of the dysfunction-weak or absent erection, erection of short duration, curved or distorted erection; the duration and progression of the condition; prior level of sexual activity, including repertoire and frequency as well as partner's interest, availability, and satisfaction. A careful review of the presence of nocturnal and especially morning erections gives considerable information about the erectile potential of the individual with simple therapies [153], although the long-standing attempt to differentiate psychogenic from organic ED by these means seems irrelevant. However, psychological elements, which should be elucidated, are present in virtually every case of ED. In this context, a simple questionnaire designed to evaluate male sexual dysfunction, e.g., the International Index of Erectile Function, can be used by clinicians [154]. Honest answers to this line of inquiry will provide an understanding of the degree of ED, relationship issues, and, in concert with the remainder of the medical assessment, medical and life style risk factors that contribute to the problem. Associated factors including the patient's endocrine status also need to be assessed during the history and physical examination. For laboratory testing, in patients who are regularly followed, we simply measure a TSH and bioavailable T, and if both are normal, we proceed. A low bioavailable T will precipitate measurement of prolactin and LH and a very low bioavailable T will precipitate an MRI of the pituitary gland, especially in younger men [155]. In the vast majority of men over the age of 40, and more particularly over the age
123 of 50, ED will be multifactorial in origin and susceptible to simple therapies (see below). For those with unusual problems, such as penile trauma or severe Peyronie's disease, evaluation by a skilled urologist is required.
E. T r e a t m e n t For men with bone fide ED, it is difficult to withhold initial therapy when a simple and effective agent such as sildenafil is available. The drug, a pill, is taken 1-2 hr prior to anticipated sexual activity [156]. It acts as an inhibitor of phosphodiesterase V. Phosphodiesterase V is responsible for degrading cyclic GMP to GMP, eliminating its biological effect (Fig. 3). As noted previously, cyclic GMP is the second messenger stimulated by NO in the corpus cavernosum. It is responsible for inhibiting Ca e+ intake and increasing Ca e+ egress from the smooth muscle of the corpora cavernosa, relaxing the arteries and sinusoids. Inhibition of phosphodiesterase V (PDE V) maintains the level of cyclic GMP for a much longer time, facilitating the erection. To be effective, the drug requires a substantial innervation of the penis. The drug does not affect libido. The side effects of sildenafil are attributed to its lack of perfect specificity. Patients may experience headaches, flushes, gastrointestinal distress, visual blurring, or a bluish haze during the 5 hr or so that the drug is active but they willingly accept the side effects if the primary effect is delivered (Fig. 4). In this study, note that with increased dosage, which produced increased efficacy, the discontinuation rate declined despite increased side effects. There are no reports of long-term consequences of sildenafil, which is taken only when intercourse is anticipated. It has been quite successful in restoring erectile function in over two-thirds of men with moderate ED [ 156]. The most significant issues are an absolute contraindication of sildenafil use
ADVERSE EVENTS Dose Response Study
20
Discontinuation Rate
10
7 o
7 c:~
0 4030-
Percent
20-
m~ 0
P
25
50
100
Dose of Sildenafil
H=headache
F=flushing
D=dyspepsia
V=visual disturbance
FIGURE 4 Adverse effects and discontinuation of sildenafil therapy. Derived from the data of Goldstein et al. [ 156].
124 in anyone taking nitrates in any form and an absolute contraindication of the use of nitrates in anyone who has recently used sildenafil. How recently? No one knows for sure. The drugs, in combination, can and have produced vascular collapse. A few deaths have been reported in association with the use of sildenafil in older men, by and large either using nitrates or with preexisting heart disease. It is likely that the death rate is not in excess of what is expected for this population. The next line of therapy in ED is based on the introduction of PGE 1 and or other agents, mainly papaverine and phentolamine, into the corpora cavernosa, by direct injection. Prostaglandin E l (alprostadil) is available in an injectable form (Caverject, Upjohn; EDEX, Schwarz Pharma) [157], and in a form that allows introduction through the urethra [158]. PGE 1 stimulates adenylyl cyclase to produce cyclic AMP. This second messenger inhibits Ca 2§ entry into smooth muscle cells, causing their relaxation. These agents affect the penis directly. They require neither an erotic stimulus nor an intact neural system. Thus, they are useful in spinal cord injury or after radical prostatectomy or colectomy. Their main problems are the necessity to introduce them directly into the penis and their propensity to produce hypotension in about 1% of patients, especially those with severe cardiovascular disease. This is particularly significant with the medicated urethral system for erection (MUSE), which requires up to 1 mg of PGE 1, whereas the injection therapies require only up to 20/zg to be introduced. Intracavernosal injection produces a satisfactory result for about two-thirds of men tested, whereas MUSE is effective for only about one-third. Vacuum tumescence devices can produce an erection in about 90% of the men who attempt them [ 159]. They consist of a plastic cylinder in which the pendulous penis is placed. The cylinder is connected to a vacuum pump, and on evacuation, blood is drawn into the penis. To keep the blood there, an obstructing band is slipped over the base of the penis when a full erection is achieved. These devices are efficient and inexpensive over time. They produce a somewhat abnormal erection in that all of the tissues, not only the corpora, are engorged, and obstruction of blood flow sometimes leads to cooling of the penis. This procedure does not interfere with a patient's medications, nor does it produce hypotension. However, there is a loss of spontaneity with sex, the erection is sometimes "on a hinge," sometimes ejaculation through the obstructing band is a problem, and, in some instances, application of excessive negative pressure produces petechiae. Also a degree of manual dexterity and skill is required to get the device to work. Penile prostheses were once the primary therapy for ED. The recent advent of much less costly and invasive approaches had made them more of a last resort. They produce a satisfactory erection in about 80% of the men who have them. Unfortunately, the rigid rod versions have a tendency
KORENMAN ET AL.
to explant and the versions including a pump mechanism have a tendency to undergo mechanical failure [ 160]. Vascular surgery is employed in specialized urology centers to establish appropriate blood flow to the penis. It is indicated only for specialized conditions such as penile or pelvic trauma to a young person. Considering that sildenafil is a small molecule, it is highly likely that other oral agents with the appropriate specificity will become available for therapy. With the cloning of penile inducible nitric oxide synthase [ 161 ], it may be possible to develop gene therapy or very specific small molecules that can enhance or preserve NO synthesis or concentrations. Numerous other components of this now well-understood system are susceptible to pharmacological attack as well.
VIII. MANOPAUSE AND MENTAL HEALTH A. D e p r e s s i o n in M i d d l e - A g e d and Elderly M e n ~ E p i d e m i o l o g y and R i s k Factors Although a number of epidemiological studies have assessed the relationship between mood and aging in women [ 162-164], the psychological changes accompanying aging in men have received little attention. Studies of mood in middle-aged men are particularly scarce. However, several authors have reported high rates of depressed mood, insomnia, mood swings, irritability, impotence, decreased libido, weakness, and lethargy in this population [165,166]. Because these symptoms may not meet criteria for major depression, as defined by the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (see Table VII), they are likely to be missed in epidemiological studies of psychiatric disorders. Among the elderly (older than 65 years), epidemiological surveys report lower rates of major depression as compared to younger populations [ 167,168]. However, studies exploring the prevalence of subsyndromal depression, i.e., depressive symptoms not meeting criteria for major depression, have consistently found a high rate among elderly persons [169-171]. Despite the high prevalence of depressive symptomatology among older persons, the symptoms are seldom recognized or treated [ 169,172]. This undertreatment may reflect clinicians' attribution of the symptoms to physical illnesses or to understandable responses to adversity [ 173,174]. Also, the greater tendency among elderly patients to express psychological distress through somatic symptoms contributes to oversights in the diagnosis of depressive disorders [ 174]. Factors associated with depressive symptoms in older men include limited economic resources, poor health, Caucasian race, and impaired sexual functioning [165-177].
125
CHAPTER 7 C h a n g e s in A g i n g M e n
TABLE VII
Criteria for Anxiety and Panic Syndromes
Condition Panic disorder without agoraphobia
Symptoms Recurrent 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
unexpected panic attacks consisting of the following symptoms a: Palpitations, pounding heart, or accelerated heart rate Sweating, trembling, or shaking Sensations of shortness of breath or smothering Feeling of choking, chest pain, or discomfort Nausea or abdominal stress Feeling dizzy, unsteady, lightheaded, or faint Derealization (feelings of unreality) or depersonalization (being detached from oneself) Fear of losing control or going crazy Fear of dying Numbness or tingling sensations Chills or hot flashes
At least one of the attacks has been followed by the following symptoms: 1. Persistent concern about having additional attacks 2. Worry about the implications of the attack or its consequences (e.g., losing control, having a heart attack, or going crazy) 3. A significant change in behavior related to the attacks Panic disorder with agoraphobia
Same as above, but with anxiety about being in places or situations from which escape might be difficult or embarrassing or in which help may be unavailable b
Generalized anxiety disorder
Excessive anxiety and worry occurring more days than not, about a number of events and activities. Difficulty in controlling the worry, and the anxiety and worry are associated with three or more of the following symptoms (with at least some symptoms present for more days than not for the past 6 months)c: 1. Restlessness or feeling keyed up or on edge 2. Being easily fatigued 3. Difficulty concentrating, mind going blank 4. Irritability 5. Muscle tension 6. Sleep disturbance (difficulty falling or staying asleep, or restless, unsatisfying sleep)
aThe panic attacks are not due to the direct physiological effects of a substance (e.g., a drug of abuse or a medication) or a general medical condition (e.g., hyperthyroidism). bThe situations are avoided or endured with distress or anxiety about having a panic attack or paniclike symptoms, or require the presence of a companion. CThe anxiety, worry, or physical symptoms cause clinically significant distress or impairment in social, occupational, or other important areas of functioning, and are not due to the direct physiological effects of a substance (e.g., a drug of abuse or a medication) or a general medical condition (e.g., hyperthyroidism), and do not occur exclusively during a mood disorder, a psychotic disorder, or a pervasive developmental disorder.
Poor physical function is also a risk factor for depressive symptoms and, conversely, depressive symptoms are associated with subsequent physical decline in elderly persons [169,178]. Being widowed, divorced, or separated are additional risk factors for depressive symptoms in this population [176,177]. Negative stereotypes of aging also are likely to impact a man's mood and self-image [179]. For many men, retirement produces a sense of letdown and can contribute to depressive symptoms [ 180]. Retirement frequently represents a loss of prestige, income, status, purpose, and workrelated friendships [ 181]. Depressive symptoms related to retirement occur most frequently in men whose lives and self-identity centered around their work or in men who have had to retire because of poor health or inability to maintain
their jobs [ 181 ]. Men who are healthy, active, and have adequate financial resources and extended social networks are least likely to experience difficulty with retirement [181].
B. E v a l u a t i o n and T r e a t m e n t o f D e p r e s s e d M o o d in A g i n g M e n Depressive symptoms in aging individuals are often missed by hospital physicians [ 182]. A careful assessment of mood is important in this population, particularly because somatic complaints may mask symptoms of depressed mood [183]. Any patient presenting with fatigue, changes in
126 appetite or sleep, and reduced libido should be evaluated for depressed mood. Depression rating scales such as the Geriatric Depression Scale [184] help screen for major depression. Unusual thought content should also be explored, because approximately 50% of depressed men over age 60 years experience delusional depression. Common presentations include delusions of being ill (somatic delusions) and of being followed or spied on (persecutory delusions) [183]. Criteria for depressive illnesses and anxiety syndromes are presented in Tables VI and VII. Patients who endorse depressive symptoms but do not meet full criteria for major depression or dysthymia may still benefit from treatment. The relationship between the depressive symptoms and psychosocial stressors (e.g., death of a family member, loss of a job, onset of an illness, or relocation) should be explored. Knowledge of factors that may have triggered the depressive mood changes is important in the choice of intervention. Interpersonal psychotherapy is particularly helpful for men who have undergone recent life transitions, because it focuses on strategies to cope with role changes or grief or to modify unrealistic expectations about relatives and other people in one's life. Cognitive-behavior therapy is an alternative approach based on training people to identify and challenge self-defeating thoughts, such as "I'm no longer working, therefore people will find me boring." Certain medications (e.g., antihypertensive agents) have been linked with depressive mood changes, thus a complete determination of the patient's medication usage should be obtained. Laboratory studies should include thyroid function testing and a bioavailable T level to rule out hypothyroidism and hypogonadism as contributing factors to the depressive symptoms. Antidepressant medications can be very helpful in promoting the recovery from major depression. Currently the most commonly used antidepressant medications are the serotonin reuptake inhibitors (fluoxetine, sertraline, and paroxetine), and they are generally well tolerated and are relatively safe in overdose. Typical side effects from these medications include gastrointestinal symptoms and impairment of sexual function, mostly libido. The tricyclic antidepressants (e.g., nortriptyline, desipramine, doxepin, and amitriptyline) have less effect on sexual function but can produce sedation, orthostatic hypotension, blurry vision, constipation, and EKG changes. Other available antidepressant medications include venlafaxine, nefazodone, bupropion, and the monoamine oxidase inhibitors (tranylcypromine and phenelzine). The antidepressants that are least likely to affect sexual function are bupropion and nefazodone. Monoamine oxidase inhibitors have the disadvantage of requiring very close attention to dietary guidelines and drug-drug interactions, to avoid the possibility of a "tyramine reaction," which can produce an abrupt and dangerous rise in blood pressure. When using antidepressant medications with men aged 65 years or older, the starting dose should be approximately half
K O R E N M A N ET AL.
that used for younger populations. Patients should be reminded that beneficial effects may not become fully apparent until after 4 - 6 weeks of treatment. Once a patient has experienced a positive response, he should be maintained on the same dose for a minimum of an additional 6 months. Long-term follow-up should continue after resolution of the depression, because the likelihood that a depressive episode will recur exceeds 50% for individuals aged 60 years or older [ 185]. Ideally, medications should be used in combination with psychotherapy and life style changes. Depressive symptoms may discourage men from maintaining healthy habits, such as exercising and not smoking, and following healthy diets. Depressed mood has also been significantly associated with a lower likelihood of engaging in walking, gardening, and exercise [186]. Unhealthy aspects of the patient's life style should therefore be explored, such as being sedentary, using alcohol and nicotine, and tendencies toward isolation, and the patient should be encouraged to exercise regularly and to be involved in stimulating activities. For men with limited social support networks, referrals to group therapy is beneficial. Screening and treatment of depressed mood in older people is a cost-effective intervention in terms of health and well being per dollar spent [ 187]. Appropriate treatment also appears to increase the number of years during which older persons are free of disability [ 169]. More research is needed, however, on specific screening and treatment interventions for middle-aged and elderly men, particularly because these populations are growing with the aging of the baby-boom generation.
C. A n d r o g e n s and the Central N e r v o u s S y s t e m Declines in bioavailable T may well account for the reduced libido with age. The latter is believed to be a central nervous system (CNS)-related effect of T that may be modulated through E 2 produced locally [18]. Androgens are necessary but not sufficient for maintaining normal libido. In older men, unlike young men, higher plasma T levels are associated with greater sexual activity [97,99] Also, latency to erection stimulated by erotic material correlates with T levels. In hypogonadal men, T replacement restores sexual interest and improves the latency, frequency, and magnitude of the nocturnal penile tumescence and the frequency of early morning erections [ 100,115]. The effect of T on the CNS extends beyond sexual behavior. T has been shown to alter mood, memory, ability to concentrate, and the overall sense of vigor and well being [ 117119]. A number of studies have examined the relationship between mood and levels of testosterone in men. However, most have included wide age ranges rather than focusing on middle-aged or elderly men. Some of these studies have
CHAPTER7 Changes in Aging Men
127
found testosterone levels in men with major depression to be lower [88], whereas others have found no significant difference from controls [189,190]. Methodological problems may explain the discrepant findings, including a lack of control for time of day of blood-drawing, total T versus free or bioavailable T, age distribution, body mass index (BMI; high B MI is associated with decreased T binding and thus lower total T values), cortisol levels (which may affect the hypothalamic-pituitary-gonadal axis), and medication use. One study that did control for medical illness, age, alcohol use, weight, and use of medications found no significant differences in free or total testosterone among 12 patients with major depression compared with 12 controls. It did identify a trend for lower testosterone levels (10% lower total testosterone and 20% lower free testosterone) in the depressed group [190]. Replication of this study in a population of middle-aged men and with a larger sample size would be of immense interest. In the only study of T levels in middleaged men, a high level of psychosocial stress was inversely related to free T levels in a sample of 439 men aged 51 years [191]. The authors concluded that psychosocial stress may be associated with premature aging in middle-aged men.
IX. P S Y C H O L O G I C A L
STATE
AND SEXUAL FUNCTION A. P s y c h o l o g i c a l C a u s e s o f E D Negative expectations of changes of sexual functioning with age may contribute to erectile difficulties [192]. Other potential causes include marital conflict, employment-related problems, family illnesses, boredom, poor communication of sexual needs, and lack of interest from one's spouse. An important and underrecognized etiology for sexual dysfunction is depressed mood. In a study of 1709 men, moderate to complete ED was found 1.82 times more in those men with depressive symptoms (as assessed by Center for Epidemiological Studies--Depression Scale) compared to those without symptoms, after controlling for age, health, medication use, demographic factors, and hormone levels [ 193]. Depressed mood has also been associated with reduced penile rigidity and nocturnal penile tumescence (NPT) time [194,195]. In the Massachusetts Male Aging Study [98], measures of depressed mood and anger were strongly correlated with ED, and were postulated to result from elevations in blood catecholamines, producing vasoconstriction and thereby inhibiting the physiological events necessary for normal sexual function. Because ED can dramatically affect mood and self-confidence [165], a vicious cycle may develop in which depressed mood and anxiety concerning sexual performance exacerbates erectile difficulties.
B. P s y c h o l o g i c a l E v a l u a t i o n a n d T r e a t m e n t of Sexual Dysfunction Premature ejaculation is the most common male sexual dysfunction, occurring in approximately 36-38% of men [ 196]. Psychological factors, including high levels of anxiety [197] and lack of intimacy with one's partner [198], are linked with this condition. Although psychogenic factors are also common in young men with ED, in men over age 50 years, most have organic etiologies for the sexual dysfunction [ 199] and often psychological issues exacerbated by ED [200]. An evaluation of sexual dysfunction, therefore, should include a psychological assessment. A review of situational factors associated with the dysfunction is essential in evaluating the extent to which the problem may have a psychogenic origin. For example, a man's sexual problems may arise only when he feels criticized or rejected, or only when he is under pressure at work. When present, alcohol and substance abuse will impair sexual function. Partners should be present during the evaluation, because they can provide useful observations and the relationship between the two can be explored. A tense or conflictual relationship is a major impediment to successful restoration of sexual function, and couples' therapy should be recommended as part of the treatment strategy. A woman's reaction to her partner's sexual difficulties should also be assessed, because she may view the man's sexual problems as a reflection on herself and feel hurt or angry. An open discussion, in which common reactions are described and normalized, can help bolster the couple's mutual trust and support. Even if an organic etiology for the sexual dysfunction is identified, psychological evaluation is still beneficial because the couple's emotional reactions may exacerbate the problem. Marital therapy may be necessary in cases in which either partner experiences persistent frustration or hostility toward the other. Sex therapy techniques can also be of great benefit, and include structured sexual exercises, psychodynamic exploration of emotional conflicts, and cognitivebehavioral strategies [ 199].
X. C O N C L U S I O N S In the 1960s, the claim was that we all began to go downhill after the age of 30, and we should never trust anyone over 30. Well, that crowd is all in its 50s now. What does happen to men as they age and what can we learn to make that inevitable process healthier and more enjoyable? Must the acquisition of wisdom invariably be associated with "settling" of the body? We really do not know and the information presented here provides only an antipasto to what should be a rich scientific repast. Only by much more intensive investigation of men as
128
KORENMAN ET AL.
t h e y p a s s t h r o u g h t h e i r 4 0 s a n d 5 0 s w i l l w e b e a b l e to r e c o m mend soundly behaviors and evaluations that will not only i m p r o v e h e a l t h a n d w e l l b e i n g , b u t , in t h e l o n g r u n , p e r h a p s
19.
r e d u c e h e a l t h c a r e c o s t s as w e l l .
References 1. Kaufman, J. M., and Vermeulen, A. (1997). Declining gonadal function in elderly men. BailliOres Clin. Endocrinol. Metab. 11,289-309. 2. Nowak, E V., and Mooradian, A. D. (1996). Endocrine function and dysfunction. Encycl. Gerontol. 1,477-491. 3. Mooradian, A. D. (1993). Mechanisms of age-related endocrine alterations. Part II. Drugs Aging 3, 131-146. 4. Mooradian, A. D. (1991). Geriatric sexuality and chronic diseases. Clin. Geriatr. Med. 7, 113-131. 5. Mooradian, A. (1993). Mechanisms of age-related endocrine alterations. Part I. Drugs Aging 3, 81-97. 6. Kovacs, K., Ryan, N., Horvath, E., Singer, W., and Ezrin, C. (1980). Pituitary adenomas in old age. J. Gerontol. 35, 16-22. 7. Sun, Y., Xi, Y. P., Fenoglio, C. M., Pushparaj, N., O'Toole, K. M., Kledizik, G. S., Nette, E. G., and King, D. W. (1984). The effect of age on the number of pituitary cells immunoreactive to growth hormone and prolactin. Hum. Pathol. 15(2), 169-180. 8. Zegarelli-Schmidt, E., Yu, X. R., Fenoglio-Preiser, C. M., O'Toole, K., Pushparaj, N., Kledzik G., and King, D. W. (1985). Endocrine changes associated with the human aging process: II. Effect of age on the number and size of thyrotropin immunoreactive cells in the human pituitary. Hum. Pathol. la(3), 277-286. 9. Matsumoto, A. M., and Brenner, W. J. (1984). Modulation of pulsatile gonadotropin secretion by testosterone in man. J. Clin. Endocrinol. Metab. 58, 609-614. 10. Tsuruhara, T., Dufau, M. L., and Cigorraga, S. (1977). Hormonal regulation of testicular luteinizing hormone receptors. Effects on cyclic AMP and testosterone responses in isolated Leydig cells. J. Biol. Chem. 252, 9002-9009. 11. Bremner, W. J., Vitiello, M. V., and Prinz, P. N. (1983). Loss of circadian rhythmicity in blood testosterone levels with aging in normal men. J. Clin. Endocrinol. Metab. 56, 1278-1281. 12. Manni, A., Pardridge, W. M., Cefalu, W., Nisula, B. C., Bardin, C. W., Santner, S. S., and Santen, R. J. (1985). Bioavailability of albuminbound testosterone. J. Clin. Endocrinol. Metab. 61,705-710. 13. Rittmaster, R. S., Zwicker, H., Thompson, D. L., Konok, G., and Norman, R. W. (1993). Androstandediol glucuronide production in human liver, prostate and skin. Evidence for the importance of the liver in 5alpha-reduced androgen metabolism. J. Clin. Endocrinol. Metab. 76, 977-982. 14. Moghissi, E., Ablan, E, and Horton, R. (1984). Origin of plasma androstanediol glucuronide in men. J. Clin. Endocrinol. Metab. 59, 417421. 15. Casey, R. W., and Wilson, J. D. (1984). Antiestrogenic action of dihydrotestosterone in mouse breast; competition with estradiol for binding to the estrogen receptor. J. Clin. Invest. 74, 2272-2278. 16. Mooradian, A. D., Morley, J. E., and Korenman, S. G. (1987). Biological actions of androgens. Endocr. Rev. 8(1), 1-28. 17. Morley, J. E., Kaiser, E E., Perry, H. M., III, and Patrick, P. (1996). Longitudinal changes in testosterone, luteinizing hormone and folliclestimulating hormone in healthy older men. Metab., Clin. Exp. 46, 410-413. 18. Morley, J. E., Kaiser, E E., Raum, W. J., Perry, H. M., III, Flood, J. E, Jensen, J., Silver, A. J., and Roberts, E. (1997). Potentially predictive and manipulable blood serum correlates of aging in the healthy human
20.
21.
22. 23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
male: Progressive decreases in bioavailable testosterone, dehydroepiandrosterone sulfate and the ratio of insulin-like growth factor 1 to growth hormone. Proc. Natl. Acad. Sci. U.S.A. 94, 7537-7542. Gray, A., Nunez, A. A., Siegel, L. I., and Wade, G. N. (1991). Age, disease, and changing sex hormone levels in middle-aged men: Resuits of the Massachusetts Male Aging Study. J. Clin. Endocrinol. Metab. 73, 1016-1025. Kaiser, F., Viosca, S. P., Morley, J. E., Mooradian, A. D., Davis, S. S., and Korenman, S. G. (1988). Impotence and Aging: Clinical and hormonal factors. J. Am. Geriatr. Soc. 36(6), 511-519. Korenman, S., Morley, J. E., Mooradian, A. D., Stanik-Davis, S., Kaiser, F. E., Silver, A. J., Viiosca, S. P., and Garza, D. (1990). Secondary hypogonadism in older men: Its relation to impotence. J. Clin. Endocrinol. Metab. 71,963-969. Vermeulen, A., and Kaufman, J. M. (1995). Aging of the hypothalamo-pituitary-testicular axis in men. Horm. Res. 43, 25-28. Deslypere, J., Kaufman, J. M., Vermeulen, T., Vogalaers, D., Vandalem, J. L., and Vermeulen, A. (1987). Influence of age on pulsatile luteinizing hormone release and responsiveness of gonadotrophs to sex hormone feedback in men. J. Clin. Endocrinol. Metab. 64, 68-73. Tenover, J. S., Matsumoto, A. M., Clifton, D. K., and Bremner, W. J. (1988). Age-related alterations in the circadian rhythms of pulsatile luteinizing hormone and testosterone secretion in healthy men. J. Gerontol. 43, M 163-M 169. Touitou, Y., F~vre, M., Lagoguey, M., Carayon, A., Bogdan, A., Reinberg, A., Beck, H., Cesselin, E, and Touitou, C. (1981). Age- and mental health-related circadian rhythms of plasma levels of melatonin, prolactin, luteinizing hormone and follicle-stimulating hormone in man. J. Endocrinol. 91(3), 467-475. Touitou, Y., Lagoguey, M., Bogdan, A., Reinberg, A., and Beck, H. (1983). Seasonal rhythms of plasma gonadotrophins: Their persistence in elderly men and women. J. Endocrinol. 96, 15-21. Winters, S. J., Sherins, R. J., and Troen, P. (1986). The gonadotropinsuppressive activity of androgen is increased in elderly men. Metab. Clin. Exp. 33, 1052-1059. Winters, J. J., and Troen, P. (1982). Episodic luteinizing hormone (LH) secretion and the response of LH and follicle-stimulating hormone to LH-releasing hormone in aged men. Evidence for co-existent primary testicular insufficiency and an impairment in gonadotropin secretion. J. Clin. Endocrinol. Metab. 55, 560-565. Muta, K., Kato, K. I., Akamine, Y., and Ibayashi, H. (1981). Agerelated changes in the feedback regulation of gonadotropin secretion by sex steroids in men. Acta Endocrinol. (Copenhagen)96, 154-162. Tenover, J. S., Matsumoto, A. M., Plymate, S. R., and Bremner, W. J. (1987). The effects of aging in normal men on bioavailable testosterone and luteinizing hormone secretion response to clomiphine citrate. J. Clin. Endocrinol. Metab. 65, 118-1126. Tenover, J. S., McLachlan, R. I., Dahl, K. D., Burger, H. G., de Kretser, D. M., and Bremner, W. J. (1988). Decreased serum inhibin levels in normal elderly men: Evidence for a decline in Sertoli cell function with aging. J. Clin. Endocrinol. Metab. 67(3), 455-459. Burger, H. G., and Robertson, D. M. (1997). Editorial: Inhibin in the male--Progress at last. Endocrinology (Baltimore) 138(4), 13611362. Harman, S. M., Tsitouras, P. D., Costa, P. T., and Blackman, M. R. (1982). Reproductive hormones in aging men. II. Basal pituitary gonadotropins and gonadotropin responses to luteinizing hormonereleasing hormone. J. Clin. Endocrinol. Metab. 54, 547-551. Harman, S. M., and Tsitouras, P. D. (1980). Reproductive hormones in aging men. I. Measurement of sex steroids, basal luteinizing hormone and Leydig cell response to human chorionic gonadotropin. J. Clin. Endocrinol. Metab. 51, 35-40.
129
CHAPTER 7 Changes in Aging Men 35. Ishimura, T., Pages, L., and Horton, R. (1977). Altered metabolism of androgens in elderly men with benign prostatic hyperplasia. J. Clin. Endocrinol. Metab. 45, 695-701. 36. Baker, H. W., Burger, H. G., de Krester, D. M., Hudson, B., O'Connor, S., Wang, C., Mirovics, A., Court, J., Dunlop, M., and Rennie, G. C. (1976). Changes in the pituitary-testicular system with age. Clin. Endocrinol. 5(4), 349-372. 37. Moore, R. J., and Wilson, J. D. (1973). The effect of androgenic hormones on the reduced nicotinamide dinucleotide phosphate 4-3-ketosteroid 5-alpha-oxidoreductase of rat ventral prostate. J. Clin. Endocrinol. Metab. 95, 581-592. 38. Korenman, S. G., Viosca, S., Garza, D., Guralnik, M., Place, V., Campbell, P., and Davis, S. S. (1987). Androgen therapy of hypogonadal men with transscrotal testosterone systems. Am. J. Med. 83(3), 471-478. 39. Tunn, S., Hochstrate, H., Grunwald, I., Fluchter, St. H., and Krieg, M. (1988). Effect of aging on kinetic parameters of 5c~-reductase in epithelium and stroma of normal and hyperplastic human prostate. J. Clin. Endocrinol. Metab. 67, 979-985. 40. Tunn, S., Haumann, R., Hey, J., Fluchter, St. H., and Krieg, M. (1990). Effect of aging on kinetic parameters of 3a (fl)-hydroxysteroid oxidoreductases in epithelium and stroma of human normal and hyperplastic prostate. J. Clin. Endocrinol. Metab. 71, 732-739. 41. Hardy, D. O., Scher, H. I., Bogenreider, T., Sabbatini, P., Zhang, Z. F., Nanus, D. M., and Catterall, J. F. (1996). Androgen receptor CAG repeat lengths in prostate cancer: Correlation with age of onset. J. Clin. Endocrinol. Metab. 81, 4400-4405. 42. Gardner, E H., and Besa, E. C. (1983). Physiologic mechanisms and the hematopoietic effects of the androstanes and their derivatives. Curr. Top Hematol. 4, 123-195. 43. Nunez, A. A. (1982). Brief report. Dose-dependent effects of testosterone on feeding and body weight in male rats. Behav. Neural Biol. 34, 445-449. 44. Gray, J. M., Nunez, A. A., Siegel, L. I., and Wade, G. N. (1979). Effects of testosterone on body weight and adipose tissue: Role of aromatization. Physiol. Behav. 23, 465-469. 45. Krotkiewsky, M., Kral, J. G., and Karlsson, J. (1980). Effects of castration and testosterone substitution on body composition and muscle metabolism in rats. Acta Physiol. Scand. 109, 233-237. 46. Urban, R. J., Bodenbury, Y. H., Gilkison, C., Foxworth, J., Coggan, A. K., Wolfe, R. R., and Ferrando, A. (1995). Testosterone administration to elderly men increases skeletal muscle strength and protein synthesis. Am. J. Physiol. 269, E820-E826. 47. Sih, R., Morley, J. E., Kaiser, F. E., Perry, H. M., 3rd, Patrick, P., and Ross, C. (1997). Testosterone replacement in older hypogonadal men: A 12-month randomized controlled trial. J. Clin. Endocrinol. Metab. 82(6), 1661-1667. 48. Morita, R., Yamamoto, I., Fukunaga, M., Dokoh, S., Konishi, J., Kousaka, T., Nakajima, K., Toizuka, K., Aso, T., and Motahashi, T. (1979). Changes in sex hormones and calcium regulating hormones with reference to bone mass associated with aging. Endocrinol. Jpn. Res. Soc. 1, 15-22. 49. Foresta, C., Busnardo, B., Ruzza, G., Zanatta, G., and Mioni, R. (1983). Lower calcitonin levels in young hypogonadal men with osteoporosis. Horm. Metab. Res. 15, 206-207. 50. Rudman, D., Kutner, M. H., Rogers, C. M., Lubin, M. F., Fleming, G. A., and Bain, R. P. (1981). Impaired growth hormone secretion in the adult population: Relation to age and adiposity. J. Clin. Invest. 67(5), 1361-1369. 51. Florini, J. R., Prinz, P. N., Vitiello, M. V., and Hintz, R. L. (1985). Somatomedin-C levels in healthy young and old men: Relationship to peak and 24-hour integrated levels of growth hormone. J. Gerontol. 40(1 ), 2-7. 52. Vermeulen, A. (1987). Nyctohemeral growth hormone profiles in
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
62a. 63. 64. 65. 66.
67. 68.
young and aged men: Correlation with somatomedin-C levels. J. Clin. Endocrinol. Metab. 64, 884-888. Ho, K. Y., Evans, W. S., Blizzard, R. M., Veldhuis, J. D., Merriam, G. R., Samojlik, E., Furlanetto, R., Rogol, A. D., Kaiser, D. L., and Thorner, M. O. (1987). Effects of sex and age on the 24-hour profile of growth hormone secretion in man: Importance of endogenous estradiol concentrations. J. Clin. Endocrinol. Metab. 64(1), 51-58. Shibasaki, T., Shizume, K., Nakahara, M., Masuda, A., Jibiki, K., Demura, H., Wakabayashi, I., and Ling, N. (1984). Age-related changes in plasma growth hormone response to growth hormone-releasing factor in man. J. Clin. Endocrinol. Metab. 58(1), 212-214. Pavlov, E. P., Harman, S. M., Merriam, G. R., Gelato, M. C., and Blackman, M. R. (1956). Responses of growth hormone (GH) and somatomedin-C to GH-releasing hormone in healthy aging men. J. Clin. Endocrinol. Metab. 62(3), 595-600. Muggeo, M., Fedele, D., Tiengo, A., Molinari, M., and Crepaldi, G. (1975). Human growth hormone and cortisol response to insulin stimulation in aging. J. Gerontol. 30(5), 546-551. Abribat, T., Deslauriers, N., Brazcau, P., and Gaudreau, P. (1991). Alterations of pituitary growth hormone-releasing factor binding sites in aging rats. Endocrinology (Baltimore) 128, 633-635. Sonntag, W. E., Forman, L. J., Miki, N., Steger, R. W., Ramos, T., Arimura, A., and Meites, J. (1981). Effects of CNS active drugs and somatostatin antiserum on growth hormone release in young and old male rats. Neuroendocrinology 33(2), 73-78. Blum, W. F., and Ranke, M. B. (1991). Plasma IGFBP-3 levels as clinical indicators. In "Modern Concepts of Insulin-Like Growth Factors" (M. Spencer, ed.), pp. 383-383. Elsevier, Amsterdam. Rutanen, E.-M., Karkkainen, T., Stenman, U.-H., and Yki-Jarvinen, H., (1993). Aging is associated with decreased suppression of insulinlike growth factor binding protein-1 by insulin. J. Clin. Endocrinol. Metab. 77, 1152-1155. Vidalon, C., Khurana, R. C., Chae, S., Gegick, C. G., Stephan, T., Nolan, S., and Danowski, T. S. (1973). Age-related changes in growth hormone in non-diabetic women. J. Amer. Geriat. Soc. 21(6), 235-255. Prinz, P. N., Weitzman, E. D., Cunningham, G. R., and Karacan, I. (1983). Plasma growth hormone during sleep in young and aged men. J. Gerontol. 38(5), 519-524. Mooradian, A., and Wong, N. C. (1994). Age-related changes in thyroid hormone action. Eur. J. Endocrinol. 131, 451. Mooradian, A. D. (1995). Normal age-related changes in thyroid hormone economy. Clin. Geriatr. Med. 11, 159-169. Robuschi, G., Safran, M., Braverman, L. E., Gnudi, A., and Roti, E. (1987). Hypothyroidism in the elderly. Endocr. Rev. 8, 142-153. Mooradian, A. D., Wong, N. C. W. (1995). Age-related changes in thyroid hormone action. Eur. J. Endocrinol. 131, 451-461. Sawin, C. T., Castelli, W. P., Hershamn, J. P., McNamara, P., and Bacharach, P. (1985). The aging thyroid. Thyroid deficiency in the Framingham Study. Arch. Intern. Med. 145, 1386-1388. Mooradian, A. D., (1990). Blood-brain barrier transport of triiodothyronine is reduced in aged rats. Mech. Ageing Dev. 52, 141-147. Mooradian, A. D. (1990). The hepatic transcellular transport of 3,5,3'triiodothyronine is reduced in aged rats. Biochem. Biophys. Acta
1054, 1-7. 69. Mooradian, A., Fox-Robichaud, A., Meijer, M. E., and Wong, N. C. W. (1994). Relationship between transcription factors and S 14 gene expression in response to thyroid hormone with age. Proc. Soc. Exp. Biol. Med. 207, 97-101. 70. Scarpace, P. J., Mooradian, A. D., and Morley, J. E. (1988). Ageassociated decrease in beta-adrenergic receptors and adenylate cyclase activity in rat brown adipose tissue. J. Gerontol. 43, B65-B70. 71. Wolfsen, A. R. (1982). Aging and the adrenals. In "Endocrine As-
130
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
KORENMAN ET AL. pects of Aging" (S. G. Korenman, ed.), pp. 55-74. Elsevier NorthHolland, New York. Boscaro, M., Paoletta, A., Scarpa, E. Barzon, L., Fusaro, P., Fallo, F., and Sonino, N. (1998). Age-related changes in glucocorticoid fast feedback inhibition of adrenocorticotropin in man. J. Clin. Endocrinol. Metab. 83, 1380-1383. Pavlov, E. P., Harman, S. M., Chrousos, G. P., Loriaux, D. L., and Blackman, M. R. (1986). Responses of plasma adrenocorticotropin, cortisol, and dehydroepiandrosterone to ovine corticotropin-releasing hormone in healthy aging men. J. Clin. Endocrinol. Metab. 62(4), 767-772. Sherman, B., Wysham, C., and Pfohl, B. (1985). Age-related changes in the circadian rhythm of plasma cortisol in man. J. Clin. Endocrinol. Metab. 61,439-443. Davis, K. L., Davis, B. M. Math6, A. A., Mohs, R. C., Rothpearl, A. B., Levy, M. I., Gorman, L. K., and Berger, P. (1984). Age and the dexamethasone suppression test in depression. Am. J. Psychiatry 141(7), 872-874. Gergerman, R. I., and Bierman, E. L. (1981). Aging and hormones. In "Textbook of Endocrinology" (R. M. Williams, ed.), pp. 1192-1212. Saunders, Philadelphia. Parker, L. N., and Odell, W. D. (1978). Decline of adrenal androgen production as measured by radioimmunoassay of urinary conjugated dehydroepiandrosterone. J. Clin. Endocrinol. Metab. 47, 600-602. Jakubowicz, D. J., Beer, N. A., Beer, R. M., and Nestler, J. E. (1995). Disparate effects of weight reduction by diet on serum dehydroepiandrosterone-sulfate levels in obese men and women. J. Clin Endocrinol. Metab. 80, 3373-3376. Coleman, D. L., Leiter, E. M., and Applezweig, N. (1984). Therapeutic effects of dehydroepiandrosterone metabolites in diabetic mutant mice (C57 G L / K s - d b / d b ) . Endocrinology (Baltimore) 115, 239-243. Barrett-Connor, E., Khaw, K. T., and Yen, S. S. C. (1986). A prospective study of dehydroepiandrosterone sulfate, mortality, and cardiovascular disease. N. Engl. J. Med. 315, 1519-1524. Beer, N. A., Jakubowicz, D. J., Matt, D. W., Beer, R. M., and Nestler, J. E. (1996). Dehydroepiandroterone reduces plasma plasminogen activator inhibitory type 1 and tissue plasminogen activator antigen in men. Am. J. Med. Sci. 311,205-210. Jesse, R. L., Loesser, K., Eich, D. M., Qian, Y. A., Hess, M. L., and Nestler, J. E. (1995). Dehydroepiandrosterone inhibits human platelet aggregation in vitro and in vivo. Ann. N. Y. Acad. Sci. 774, 281-290. Schwartz, R. S., Jaeger, L. E, and Veith, R. C. (1987). The importance of body composition to the increase in plasma norepinephrine appearance rate in elderly men. J. Gerontol. 421, 546-551. Sowers, J. R., Rubenstein, L. Z., and Stern, N. (1983). Plasma norepinephrine responses to posture and isometric exercise increase with age in the absence of obesity. J. Gerontol. 38, 315-317. Brodde, O. E., Anluf, M., Graben, N., and Bock, K. D. (1982). Agedependent decrease of alpha 2-adrenergic receptor number in human platelets. Eur. J. Pharmacol. 81(2), 345-347. Vestal, R. E., Wood, A. J., and Shand, D. G. (1979). Reduced betaadrenoreceptor sensitivity in the elderly. Clin. Pharmacol. Ther. 26, 181-186. Lipsitz, L. A., Nyquist, R. P., Jr., Wei, J. Y., and Rowe, J. W. (1983). Postprandial reduction in blood pressure in the elderly. N. Engl. J. Med. 309(2), 81-83. Riggs, B. L., Wahner, H. W., Dunn, W. L., Mazess, R. B., Offord, K. P., and Melton, L. J., 3rd (1981). Differential changes in bone mineral density of the appendicular and axial skeleton with aging: Relationship to spinal osteoporosis. J. Clin. Invest. 67(2), 328-335. Avioli, L. V. (1982). Aging, bone and osteoporosis. In "Endocrine Aspects of Aging" (S. G. Korenman, ed.), pp. 199-229. Elsevier North-Holland, New York. Dandona, P., Menon, R. K., Shenoy, R., Houlder, S. Thomas, M., and
91. 92.
93.
94.
95.
96.
97. 98.
99.
100. 101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
Mallinson, W. J. (1986). Low 1,25-dihydroxyvitamin D, secondary hyperparathyroidism, and normal osteocalcin in elderly subjects. J. Clin. Endocrinol. Metab. 63(2), 459-462. Mooradian, A. D. (1996). Drug therapy of non-insulin-dependent diabetes mellitus in the elderly. Drugs 51, 931-941. Campbell, S., and Mooradian, A. D. (1993). Diabetes mellitus. In "Geriatric Pharmacology" (R. Bressler and M. K. Katz, eds.), pp. 409-423. McGraw-hill, New York. Phillips, P. A., Rolls, B. J., Ledingham, J. G., Forsling, M. L., Morton, J. J., Crowe, M. J., and Wollner, L. (1984). Reduced thirst after water deprivation in healthy elderly men. N. Engl. J. Med. 311(12), 753-759. Helderman, J. H. (1982). The impact of normal aging on the hypothalamic-neurohypophyseal-renal axis. In "Endocrine Aspects of Aging" (S. G. Korenman, ed.), pp. 9-32. Elsevier North-Holland, New York. Ohashi, M., Fumio, N., Nawata, H. Kato, K., Ibayashi, H., Kangawa, K., and Matsuo, H. (1987). High plasma concentrations of human atrial natriuretic polypeptide in aged men. J. Clin. Endocrinol. Metab. 64(1), 81-85. Mooradian, A. (1994). Biology of aging. In "Rehabilitation of the Aging and the Elderly Patient" (S. J. G. G. Felsenthal and F. U. Steinberg, eds.), pp. 3-10. Williams & Wilkins, Baltimore, MD. Knudson, R. J. (1989). Aging of the respiratory system. Curr. Pulmonol. 10, 1-24. Feldman, H., Goldstein, I., Hatzichristou, D. G., Krane, R. I., and McKinlay, J. B. (1994). Impotence and its medical and psychosocial correlates: Results of the Massachusetts Male Aging Study." J. Urol. 151,54-61. Kwan, M., Greenleaf, W. J., Mann, J., Crapo, L., and Davidson, J. M. (1983). The nature of androgen action on male sexuality: A combined laboratory-self-report study on hypogonadal men. J. Clin. Endocrinol. Metab. 57(3), 557-562. Bancroft, J., and Wu, F. C. W. (1983). Changes in erectile responsiveness during androgen therapy. Arch. Sex. Behav. 12, 59-66. Davidson, J., Camargo, C., and Smith, E. (1979). Effects of androgen on sexual behavior in hypogonadal men. J. Clin. Endocrinol. Metab. 48, 955-958. Bagatell, C., Heiman, J., Rivier, J., and Bremner, W. Effects of endogenous testosterone and estradiol on sexual behavior in normal young men. J. Clin. Endocrinol. Metab. 78, 711-716. Slag, M. F., Morley, J. E., Elson, M. K., Trence, D. L., Nelson, C. J., Nelson, A. E., Kinlaw, W. B., Beyer, H. S. Nuttall, F. Q., and Shafer, R. B. (1983). Impotence in medical clinic outpatients. JAMA, J. Am. Med. Assoc. 249(13), 1736-1740. Davis, S. S., Viosca, S. P., Guralnik, M., Windsor, C., Buttiglieri, M. W., Baker, J. D., Mehta, A. J., and Korenman, S. G. (1985). Evaluation of impotence in older men. West. J. Med. 142(4), 499-505. Lugg, J., Rajfer, J., and Gonzalez-Cadavid, N. F. (1995). Dihydrotestosterone is the active androgen in the maintenance of nitric oxidemediated penile erection in the rat. Endocrinology (Baltimore) 136, 1495-1501. Heaton, J., and Varrin, S. (1994). Effects of castration and exogenous testosterone supplementation in an animal model of penile erection. J. Urol. 151 797-800. Kurz, E., Sengelaub, D., and Arnold, A. (1986). Androgens regulate the dendritic length of mammalian motoneurons in adulthood. Science 232, 395-396. Leipheimer, R., and Sachs, B. Relative androgen-sensitivity of the vascular and striated-muscle systems regulating penile erection in rats. Physiol. Behav. 54, 1085-1090. Shiavi, R., White, D., Mandeli, J., and Schreiner-Engel, P. (1992). Hormones and nocturnal penile tumescence in healthy aging men. Arch. Sex. Behav. 22, 207-215. Schiavi, R.C., Schreiner-Engel, P., White, D., and Mandeli, J. The re-
CHAPTER 7 Changes in Aging Men
111. 112. 113.
114.
115.
116.
117.
118.
119.
120.
121.
122. 123. 124.
125.
126. 127. 128. 129.
130. 131.
132.
133.
lationship between pituitary-gonadal function and sexual behavior in healthy aging men. Psychosom. Med. 53, 363-374. de Lignieres, B. (1993). Transdermal dihyrotestosterone treatment of "andropause." Ann. Med. 25, 235-241. Schiavi, R. C., Mandeli, J., and Schreiner-Engel, P. (1994). Sexual satisfaction in healthy aging men. J. Sex Marital Ther. 20, 3-13. Davidson, J., Chen, J. J., Crapo, L., Gray, G. D., Greenleaf, W. J., and Catania, J. A. (1983). Hormonal changes and sexual function in aging men. J. Clin. Endocrinol. Metab. 57, 71-77. O'Carroll, R., and Bancroft, J. (1984). Testosterone therapy for low sexual interest and erectile dysfunction in men: A controlled study. Br. J. Psychiatry 145, 146-151. Metz, M. E., P. J., Nesvacil, L. J., Abuzzahab, E, Sr., and Koznar, J. (1997). Premature ejaculation: A psychophysiological review. J. Sex Marital Ther. 23, 3-23. Waldinger, M. D., Hengeveld, M. W., and Zwinderman, A. H. Ejaculation-retarding properties of paroxetine in patients with primary premature ejaculation: A double-blind, randomized, dose-response study. Br. J. Urol. 79(4), 592-595. Ludovico, G. M., C. A., Pagliarulo, G., Cirillo-Marucco, E., Marano, A., and Pagliarulo, A. (1996). Paroxetine in the treatment of premature ejaculation. Br. J. Urol. 77, 881-882. Haensel, S. M., Klem, T. M., Hop, W. C., and Slob, A. K. (1998). Fluoxetine and premature ejaculation: A double-blind, crossover, placebo-controlled study. J. Clin. Psychopharmacol. 18(1), 72-77. Haensel, S. M., Rowland, D. L., and Kallan, K. T. (1996). Clomipramine and sexual function in men with premature ejaculation and controis. J. Urol. 156(4), 1310-1315. Balbay, M. D., Yildiz, M., Salvarci, A., Ozsan, O., and Ozbek, E. (1998). Treatment of premature ejaculation with sertralin. Int. Urol. Nephrol. 30(1), 81-83. Lee, H. S., Song, D. H., Kim, C. H., and Choi, H. K. (1996). An open clinical trial of fluoxetine in the treatment of premature ejaculation. J. Clin. Psychopharmacol. 16(5), 379-382. Buffum, J. (1982). Pharmacosexology: The effects of drugs on sexual function--A review. J Psychoactive Drugs 14, 5-44. Wein, A., and Van Arsdale, K. (1988). Drug-induced male sexual dysfunction. Urol. Clin. North Am. 15, 23-31. Korenman, S. (1998). Sexual function and dysfunction. In "Williams Textbook of Endocrinology" (J. Wilson and D. W. Foster, eds.), pp. 927-937. Saunders, Philadelphia. Korenman, S. (1995). Advances in the understanding and management of erectile dysfunction. J. Clin. Endocrinol. Metab. 80, 19851988. Korenman, S. (1998). New insights into erectile dysfunction: A practical approach. Am. J. Med. (in press). Krane, R., Goldstein, I., and Saenz de Tejada, I. (1989). Impotence. N. Engl. J. Med. 321, 1648-1659. National Institutes of Health Consensus Statement (1992). Impotence 10(1). JCnler, M., Moon, T., Brannan, W., Stone, N. N., Heisey, D., and Bruskewitz, R. C. (1995). The effect of age, ethnicity and geographical location on impotence and quality of life. Br. J. Urol. 75(5), 651-655. Anderson, K., and Wagner G. (1995). Physiology of penile erection. Physiol. Rev. 75, 191. Azadzoi, K. M., Vlachiotis, J., Pontari, M., and Siroky, M. B. (1995). Hemodynamics of penile erection: III. Measurement of deep intracavernosal and subtunical blood flow and oxygen tension. J. Urol. 153(2), 521-526. Lue, T., Zeineh, S. L., Schmidt, R. A., and Tanagho, E. A. (1984). Neuroanatomy of penile erection: Its relevance to iatrogenic impotence. J. Urol. 131, 273-280. De Groat, W., and Booth, A. Neural control of penile erection. In "The Autonomic Nervous System. Nervous Control of the Urogenital System" (C. Maggi, ed.), pp. 465-513. Harwood, London.
131 134. Steers, W. (1993). Neural control of penile erection. Semin. Urol. 8, 66 -79. 135. Raifer, J., Aronson, W. J., Bush, P. A., Dorey, E J., and Ignarro, L. J. (1992). Nitric oxide as a mediator of relaxation of the corpus cavernosum in response to nonadrenergic, noncholinergic neurotransmission. N. Engl. J. Med. 326(2), 90-94. 136. Saenz de Tejada, I., Goldstein, I., Azadzoi, K., Krane, R. J., and Cohen, R. A. (1989). Impaired neurogenic and endothelium-mediated relaxation of penile smooth muscle from diabetic men with impotence. N. Engl. J. Med. 320(16), 1025-1030. 137. Ignarro, L., Bush, P. A., Buga, G. M., Woods, K. S., Fukuto, J. M., and Raifer, J. (1990). Nitric oxide and cyclic-GMP formation upon electric field stimulation cause relaxation of corpus cavernosum smooth muscle. Biochem. Biophys. Res. Commun. 170, 843-850. 138. Saenz de Tejada, I., Moroukian, P., Tessier, I., Kim, J. L., Goldstein, I., and Frohrib, D. (1991). Trabecular smooth muscle modulates the capacitor function of the penis. Studies on a rabbit model. Am. J. Physiol. 260, H1590-H1595. 139. Fournier, G., Juenemann, K.-P., Lue, T. E, and Tanagho, E. A. (1987). Mechanisms of venous occlusion during canine penile erection: An anatomic demonstration. J. Urol. 137, 163-167. 140. Schmidt, M., and Schmidt, H. S. (1993). The ischiocavernosus and bulbospongiousus muscles in mammalian penile rigidity. Sleep 16, 171-183. 141. Mattson, D., Petrie, M., Srivastava, D. K., and McDermott, M. (1995). Multiple sclerosis. Sexual dysfunction and its response to medications. Arch. Neurol. (Chicago) 52(9), 862-868. 142. Jensen, S. (1984). Sexual function and dysfunction in younger married alcoholics. Acta Psychiatr. Scand. 69, 543-549. 143. Mannuino, D., Keuvens, R. M., and Flanders, W. D. (1994). Cigarette smoking: An independent risk factor for impotence? Am. J. Epidemiol. 140, 1003 - 1008. 144. Mersdorf, A., Goldsmith, P. C., Diederichs, W., Padula, C. A., Lue, T. E, Fishman, I. J., and Tanagho, E. A. (1991). Ultrastructural changes in impotent penile tissue: A comparison of 65 patients. J. Urol. 145(4), 749-758. 145. Pickard, R., Powell, P., and Zar, M. (1995). Nitric oxide and cyclic GMP formation following relaxant nerve stimulation in isolated human corpus cavernosum. Br. J. Urol. 75, 516-522. 146. Christ, G., Schwartz, C. B., Stone, B. A., Parker, M., Janis, M., Gondre, M., Valcic, M., and Melman, A. (1992). Kinetic characteristics of al-adrenergic contractions in human corpus cavernosum smooth muscle. Am. J. Physiol. 263, H15-H19. 147. Persson, C., Diederichs, W., Lue, T. E, Yen, T. S., Fishman, I. J., McLin, P. H., and Tanagho, E. A. (1989). Correlation of altered penile ultrastructure with clinical arterial evaluation. J. Urol. 142(6), 14621468. 148. Akkus, E., Carrier, S., Baba, K., Hus, G.-L., Padma-Nathan, H., Nunes, L., and Lue, T. E (1997). Structural alterations in the tunica albuginea of the penis: Impact of Peyronie's disease, ageing and impotence. Br. J. Urol. 79, 47-53. 149. Sattar, A., Wespes, E., and Schulman, C. (1994). Computerized measurement of penile elastic fibers in potent and impotent men. Eur. Urol. 25, 142-44. 150. Bookstein, J., and Valji, K. The arteriolar component in impotence: A possible paradigm shift. Am. J. Radiol. 157, 932-934. 151. Lue, T., Hricak, H., Schmidt, R. A., and Tanagho, E. A. (1986). Functional evaluation of penile veins by cavernosography in papaverineinduced erection. J. Urol. 135, 479-482. 152. Kaufman, J., Borges, F. D., Fitch, W. P., 3rd, Geller, R. A., Grubner, M. B., Hubbard, J. G., McKay, D. L., Jr., Tuttle, J. P., and Witten, E R. (1993). Evaluation of erectile dysfunction by dynamic infusion cavernosometry and cavernosography (DICC). Multi-institutional Study. Urology 41(5), 445-451. 153. Ackerman, M., D'Attilio, J. P., Antoni, M. H., Weinstein, D., Rhamy,
132
154.
155.
156.
157. 158.
159.
160. 161.
162.
163.
164.
165. 166.
167.
168.
169.
170.
171.
KORENMAN ET AL. R. K., and Politano, V. A. The predictive significance of patientreported sexual functioning in rigiscan sleep evaluations. J. Urol. 146, 1559-1563. Rosen, R. C., Riley, A., Wagner, G., Osterloh, I. H., Kirkpatrick, J., and Mishra, A. (1997). The international index of erectile function (IIEF): A multidimensional scale for assessment of erectile dysfunction. Urology 49(6), 822-830. Buvat, J., and Lemaire, A. (1997). Endocrine screening in 1,022 men with erectile dysfunction: Clinical significance and cost-effective strategy. J. Urol. 158(5), 1764-1767. Goldstein, I., Lue, T. F., Padma-Nathan, H., Rosen R. C., Steers, W. D., and Wicker, P. A. (1998). Oral sildenafil in the treatment of erectile dysfunction, Sildenafil Study Group. N. Engl. J. Med. 338(20), 1397-1404. Linet, O., and Neff, L. (1994). Intracavernous prostaglandin E1 in erectile dysfunction. Clin. Invest. 72, 139-149. Padma-Nathan, H., Hellstrom, W. J., Kaiser, F. E., Labasky R. E., Lue, T. E, Nolten, W. E., Norwood, P. C., Peterson, C. A. Shabsigh, R., and Tam, P. Y. (1997). Treatment of men with erectile dysfunction with transurethral alprostadil. Medicated Urethral System for Erection (MUSE) Study Group. N. Engl. J. Med. 336, 1-7. Korenman, S., Viosca, S. P., Kaiser, F. E., Mooradian, A. D., and Morley, J. E. (1990). Use of a vacuum tumescence device in the management of impotence. J. Am. Geriatr. Soc. 38, 217-220. Petrou, S., and Barrett, D. (1990). The use of penile prostheses in erectile dysfunction. Semin. Urol. 8, 138-152. Garban, H., Marquez, D., Magee, T., Moody, J., Rajavashisth, T., Ropdriguez, A., Hung, A., Vernet, D., Rajfer, J., and Gonzalez-Cadavid, N. E. (1997). Cloning of rat and human inducible penile nitric oxide synthase. Application for gene therapy of erectile dysfunction. Biol. Reprod. 56, 954-963. Kaufert, P., Gilblert, P., and Tate, R. (1992). The Manitoba Project: A reexamination of the link between menopause and depression. Maturitas 14, 143-155. Avis, N., and McKinlay, S. M. (1995). The Massachusetts Women's Health Study: An epidemiologic investigation of the menopause. J. Am. Med. Women's Assoc. 50, 45-49. Matthews, K., Wing, R. A., and Kuller, L. H. (1990). Influences of natural menopause on psychological characteristics and symptoms of middle-aged healthy women. J. Consult. Clin. Psychol. 58, 345-351. Schow, D., Redmon, B., and Pryor, J. L. (1997). Male menopause: How to define it, how to treat it. Postgrad. Med. 101, 62-79. McKinlay, J. (1989). Is there an epidemiologic basis for a male climacteric syndrome? The Massachusetts Male Aging Study. In "Menopause: Evaluation, Treatment, and Health Concerns" (Charles B. Hammond and Florence P. Haseltire, eds.), pp. 163-192. Alan R. Liss, New York. Gatz, M., and Hurwicz, M. L. (1990). Are old people more depressed? Cross-sectional data on Center for Epidemiological Studies Depression Scale Factors. Psychol. Aging 5, 284-290. Robins, L., and Regier, D. A. (1991). "Psychiatric Disorders in America: The Epidemiologic Catchment Area Study." Free Press, New York. Penninx, B., Guralnik, J. M., Ferrucci, L., Simonsick, E. M., Deeg, D. J. H., and Wallace, R. B. (1998). Depressive symptoms and physical decline in community-dwelling older persons. JAMA, J. Am. Med. Assoc. 279, 1720-1726. Beekman, A., Deeg, D. J., van Tilburg, T., Smit, J. H., Hooijer, C., and van Tilburg, W. (1995). Major and minor depression in later life: A study of prevalence and risk factors. J. Affective Disord. 36, 65-75. Unutzer, J., Patrick, D. L., Simon, G., Grembowski, D., Walker, E., Rutter, C., and Katon, W. (1997). Depressive symptoms and the cost of health services in HMO patients aged 65 years and older. JAMA, J. Am. Med. Assoc. 277, 1618-1623.
172. Hirschfeld, R., Keller, M. B., Panico, S., Arons, B. S., Barlow, D., Davidoff, F., Endicott, J., Froom, J., Goldstein, M., Gorman, J. M., Marek, R. G., Maurer, T. A., Meyer, R., Phillips, K., Ross, J., Schwenk, T. L., Sharfstein, S. S., Thase, M. E., and Wyatt, R. J. (1997). The National Depressive and Manic-Depressive Association consensus statement on the undertreatment of depression. JAMA, J. Am. Med. Assoc. 277, 333-340. 173. Knaupper, B., and Wittchen, H. U. (1994). Diagnosing major depression in the elderly: Evidence for response bias in standardized diagnostic interviews? J. Psychiat. Res. 28, 147-164. 174. Zisook, S. (1996). Depression in late life. Diagnosis, course, and consequences. Postrgrad. Med. 100, 143-148. 175. Tannock, C., and Katona, C. (1995). Minor depression in the aged. Concepts, prevalence and optimal management. Drugs Aging 6, 278-292. 176. Blazer, D., and Williams, C. D. (1980). Epidemiology of dysphoria and depression in an elderly population. Am. J. Psychiatry. 137, 439-444. 177. Murrell, S., Himmelfarb, S., and Wright, K. (1983). Prevalence of depression and its correlates in older adults. Am. J. Epidemiol. 117, 173-185. 178. Turner, R. J., and Noh, S. (1988). Physical disability and depression: A longitudinal analysis. J. Health Soc. Behav. 29(1), 23-37. 179. Featherstone, M., and Hepworth, M. (1985). The male menopause: Lifestyle and sexuality. Maturitas 7, 235-246. 180. Atchley, R. (1996). Retirement, In "Encyclopedia of Gerontology" (J. Birren, ed.), pp. 437-449. Academic Press, San Diego, CA. 181. Rybash, J., Roodin, P. A., and Hoyer, W. J., eds. (1995). "Adult Development and Aging," 3rd ed., pp. 276-290. Brown & Benchmark, Madison, WI. 182. Finch, E., Ramsay, R., and Katona, C. L. (1992). Depression and physical illness in the elderly. Clin. Geriatr. Med. 8, 275-287. 183. Baker, F. (1996). An overview of depression in the elderly: A U.S. perspective. J. Natl. Med. Assoc. 88, 178-184. 184. Yesavage, J. A. (1983). Development and validation of a Geriatric Depression Screening Scale: A preliminary report. J. Psychiat. Res. 17, 37-49. 185. Hirschfeld, R. (1994). Guidelines for the long-term treatment of depression. J. Clin. Psychiatry 55(12, Suppl.), 61-69. 186. Cornoni-Huntley, J., Ostfeld, A. M., Taylor, J. O., Wallace, R. B., Blazer, D., Berkman, L. F., Evans, D. A., Kohout, F. J., Lemke, J. H., Scherr, P. A. et al. (1993). Established populations for epidemiologic studies in the elderly: Study design and methodology. Aging 5, 27-37. 187. Sturm, R., and Wells, K. (1995). How can care for depression be more cost-effective? JAMA, J. Am. Med. Assoc. 273, 51-58. 188. Yesavage, J. A., Davidson, J., Widlow, L., and Berger, P. A. (1985). Plasma testosterone levels, depression, sexuality, and age. Biol. Psychiatry, 20, 222-225. 189. Davies, R., Harris, B., Thomas, D. R., Cook, N., Read, G., and RiadFahmy, D. (1992). Salivary testosterone levels and major depressive illness in men. Br. J. Psychiatry 161, 629-632. 190. Levitt, A. J., and Joffe, R. T. (1988). Total and free testosterone in depressed men. Acta Psychiatr. Scand. 77(3), 346-348. 191. Nilsson, P., Moiler, L., and Solstad, K. (1995). Adverse effects of psychosocial stress on gonadal function and insulin levels in middle-aged males. J. Intern. Med. 237, 479-486. 192. Schiavi, R., and Rehman, J. (1995). Sexuality and aging. Urol. Clin. North Amer. 4, 711-726. 193. Araujo, A., Durante, R. Feldman, H. A., Goldstein, I., and McKinlay, J. B. (1998). The relationship between depressive symptoms and male erectile dysfunction: Cross-sectional results from the Massachusetts Male Aging Study. Psychosom. Med. 194. Thase, M., Reynolds, C. F., Jennings, J. R., Frank, E., Howell, J. R., Houck, P. R., Berman, S., and Kupfer, D. I. (1988). Nocturnal penile
CHAPTER 7 Changes in Aging M e n tumescence is diminished in depressed men. Biol. Psychiatry 24(1), 33-46. 195. Thase, M., Reynolds, C. F., Glanz, L. M., Jennings, J. R., Sewitch, D. E., Kupfer, D. J., and Frank, E. (1987). Nocturnal penile tumescence in depressed men. Am. J. Psychiatry 144(1), 89-92. 196. Spector, I., and Carey, M. P. (1990). Incidence and prevalence of the sexual dysfunctions: A critical review of the empirical literature. Arch. Sex. Behav. 19, 389-408. 197. Cooper, A., Cernovsky, Z. A., and Colussi, K. (1993). Some clinical
133 and psychometric characteristics of primary and secondary premature ejaculators. J. Sex and Marital Ther. 19, 276-288. 198. McCabe, M. (1997). Intimacy and quality of life among sexually dysfunctional men and women. J. Sex Marital Ther. 23, 276-290. 199. Kaplan, H. (1994). Psychogenic impotence. Curr. Ther. Endocrinol. Metab. 5, 323-328. 200. Wylie, K. (1997). Male erectile disorder: Characteristics and treatment choice of a longitudinal cohort study of men. Int. J. Impotence Res. 9, 217-224.
2 H A P T E R {~
Premature Ovarian Failure ROBERT W. REBAR
I. II. III. IV.
Department of Obstetrics and Gynecology, University of Cincinnati Medical Center, Cincinnati, Ohio 45267; and the American Society for Reproductive Medicine, Birmingham, Alabama 35216
V. Evaluation of Patients with Hypergonadotropic Amenorrhea VI. Therapy References
Early Reports of Premature Ovarian Failure Clinical Features of Premature Ovarian Failure Prevalence of Premature Ovarian Failure Etiology of Premature Ovarian Failure
genes are important in controlling the number of oocytes ovulated and hence presumably the timing of the cessation of reproductive function [3]. Although these data are difficult to extrapolate to humans, given what is known about the control of ovarian function by the X chromosome [4], it is not difficult to believe that inherited tendencies are important. Any role for ovarian inhibin and its feedback action on pituitary FSH secretion also remains to be explored. Also potentially important in the regulation of the onset of menopause is the hypothalamic-pituitary axis. Although oocyte depletion may provide the major reason for the occurrence of menopause in humans, numerous animal studies document changes in neurotransmitter and in central nervous system (CNS) feedback responses to estrogen with aging. Of particular note is the observation that aging ovaries transplanted to young rodents cycle normally whereas young ovaries transplanted to aged animals do not function well [5]. Once more, however, extrapolation of such data to humans is most difficult. The concept that young women under the age of 40 with "hypergonadotropic" amenorrhea by definition should have depletion of their oocytes and premature ovarian failure was supported by the findings of Goldenberg and colleagues [6]. They reported in 1973 that women who had basal FSH concentrations greater than 40 mIU/ml [Second International Reference Preparationmhuman menopausal gonadotropin, (2nd IRP-hMG] without exception had no viable oocytes on ovarian biopsy.
I. EARLY REPORTS OF PREMATURE OVARIAN FAILURE Menopause, defined strictly as the last episode of menstrual bleeding, typically occurs around age 51 and is generally considered premature if it occurs before the age of 40 years. In fact, de Moraes and Jones [1] first defined premature menopause, or premature ovarian failure, as consisting of the triad of amenorrhea, hypergonadotropinism, and hypoestrogenism in women under the age of 40 years. Why the cessation of reproductive life should occur prematurely has been of great interest to clinicians and remains enigmatic in the majority of cases. How little is known about premature menopause is less surprising in view of how little is known about normal menopause. The events that signal menopause are unclear. Depletion of oocytes is obviously an important factor, and it has been documented that follicle depletion accelerates just prior to menopause [2]. Although a few follicles may be present at menopause, they do not respond to folliclestimulating hormone (FSH) and luteinizing hormone (LH). In an unsuccessful effort tastimulate follicular development and estradiol secretion, the hypothalamus signals the pituitary gland to secrete still more FSH and LH. Thus, an increase in serum FSH concentrations is an early sign heralding the cessation of ovarian function. Preliminary studies in strains of mice indicate that specific MENOPAUSE:
BIOLOGY AND PATHOBIOLOGY
135
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
136
ROBERT W. REBAR
The belief that the ovarian "failure" observed in such young women was permanent was first questioned by a number of isolated case reports documenting the initiation or resumption of cyclic menses and/or pregnancy in affected women. Several large series have now confirmed these case reports [7]. In one of those reports, we documented pubertal progression in two young girls with elevated circulating FSH and multiple endocrine deficiencies (i.e., hypoparathyroidism and hypoadrenalism) and suggested that waxing and waning autoimmune dysfunction might account for the transient nature of the ovarian failure [8]. O'Herlihy and coworkers [9] reported that up to one-fourth of younger women with FSH values in the menopausal range will resume ovulation spontaneously and a few will even conceive. In 1982 we reported that 9 of 18 young women presumed to have ovarian failure had circulating estradiol typical of women with functioning ovarian follicles and that 4 of the 9 women who had ovarian biopsies had viable oocytes [10]. In addition, circulating concentrations of serum progesterone typical of ovulation were noted in 5 women, and a spontaneous pregnancy occurred in one. Aiman and Smentek [11] reported that 18% of 157 women reported in the literature who had ovarian biopsies had specimens containing apparently viable oocytes. They also noted that 14 of the women had conceived after the ovarian failure had been diagnosed. A number of recent series have confirmed ovarian follicular activity in many women with ovarian failure. Hague et al. [12] reported evidence of ovarian follicular activity in 12 (17.1%) of 93 women with amenorrhea and elevated FSH concentrations. By pelvic ultrasound Conway and colleagues [ 13] identified follicular activity in 65 of 109 women (60%) with "idiopathic" premature ovarian failure. Bone mineral density was lower in women in whom ovaries were not identified on ultrasound (n -- 26) than in those in whom TABLE I
follicles > 4 m m were identified (n = 57). Similarly, Nelson and colleagues [14] documented ovarian follicular activity by serum estradiol levels greater than 50 pg/ml in nearly half of 65 women with karyotypically normal spontaneous premature ovarian failure and imaged an antral follicle in over 40% [ 15] of the women. These observations led us to suggest that this disorder involved more than just the premature cessation of ovarian function and might more appropriately be termed "hypergonadotropic amenorrhea" [ 13 ] - - at least until such time as it was apparent that the premature loss of ovarian function was permanent.
II. CLINICAL
FEATURES
OF PREMATURE
OVARIAN
FAILURE
To define the clinical spectrum of women with hypergonadotropic amenorrhea, Rebar and Connolly [ 16] compiled data from 115 affected women seen sequentially between 1978 and 1988. Initial inclusion criteria were (1) amenorrhea of 3 or more months' duration, (2) age under 40 years at the onset of the amenorrhea, and (3) circulating FSH of more than 40 mlU/ml on at least two occasions. A number of interesting differences and similarities existed between those with primary and those with secondary amenorrhea and are summarized in Table I. In over 75% of the patients, symptoms of estrogen deficiency, most commonly hot flushes and/or dyspareunia, were evident, but these symptoms were far more common in those with secondary amenorrhea. Chromosomal abnormalities and failure to develop mature secondary sex characteristics were far more common in those with primary amenorrhea. Chromosomal abnormalities were present in over half
Features of Women with Primary and Secondary Amenorrheaa
Feature
Primary amenorrhea
Number of patients Symptoms of estrogen deficiency Incomplete sexual development Karyotypic abnormalities Immune abnormalities Spinal bone density <90% of controls Progestin-induced withdrawal bleeding Pregnancies before diagnosis Evidence of ovulation after diagnosis Pregnancies after diagnosis
18 (15.7) 4 (22.2) 16 (88.9) 10 (55.6) 4 (22.2) 3/4 (75) 2/9 (22.2) 0 0 0
Secondary amenorrhea 97 (84.3) 83 (85.6) 8 (8.2) 6/45 tested (13.3) 16 (16.5) 13/22 (59.1) 36/70 (51.4) 33 (34.0) 23 (23.7) 8(8.2)
Significant differenceb p < 0.001 p < 0.001 p < 0.001 p < 0.01 NS NS NS p < 0.025 P < 0.05 NS
aAdapted from Rebar and Connolly [16] with permission from the American Society for Reproductive Medicine (Fertility and Sterility, 1990, Vol. 53, pp. 804-810). Values in parentheses are percentages of women diagnosed with a type of amenorrheaand associated features. bResults using X2 test. NS, not significant.
CHAPTER 8 Premature Ovarian Failure the women with primary amenorrhea, who tended to have deletions of all or a part of one X chromosome, whereas those with secondary amenorrhea more commonly had an additional X chromosome. Easily detected immune disturbances were present in approximately 20% of the patients. Thyroid abnormalities were most common, with five women having Hashimoto's thyroiditis, two developing primary hypothyroidism, one developing subacute thyroiditis, and one having Graves' disease. Three asymptomatic patients had high titers of antimicrosomal antibodies. One of the women had vitiligo and hypoparathyroidism, one had Addison's disease, and one additional woman had insulin-dependent diabetes mellitus. A relatively small number of the women in this series, all with secondary amenorrhea, had received chemotherapy with alkylating agents and in some cases radiation therapy as well before developing hypergonadotropic amenorrhea. The effects of alkylating agents and radiation therapy on ovarian function have been recognized for several years [ 17]. As more young women with childhood malignancies, especially the various leukemias and lymphomas, are treated and cured, the incidence of patients with induced amenorrhea will no doubt increase. Four of the patients with secondary amenorrhea and normal karyotypes had a family history of early menopause prior to age 40. Four others reported a temporal relationship between the onset of amenorrhea and various infections, including chicken pox, shigellosis, malaria, and an undefined viral syndrome. Spinal bone density, as evaluated by dual-photon absorptiometry, was less than 90% (range 62-105%; mean 85.7%) of the mean value observed in age-matched controls in 16 of the 26 women who underwent such testing. Progestin-induced withdrawal bleeding, presumably an indication of endogenous estrogen activity, occurred in just under 50% of the women tested. Withdrawal bleeding even occurred in 2 of the 9 individuals with primary amenorrhea who were challenged. There was, however, no correlation between the response to exogenous progestin and subsequent ovulation. None of the women with primary amenorrhea ever ovulated or conceived with her own oocytes. In contrast, over one-third of the women with secondary amenorrhea were pregnant at least once before developing hypergonadotropic amenorrhea and almost one-quarter had evidence of ovulation after the diagnosis was established. Yet only 8% of those with secondary amenorrhea later conceived. Twenty-five of the patients with secondary amenorrhea were treated with clomiphene citrate to induce ovulation, but only four (16%) ovulated as determined by serial ultrasound and serum progesterone levels. Because each of the four who ovulated had evidence of spontaneous episodic ovulation before therapy, it is unclear if the clomiphene actually induced ovulation or if ovulation occurred in association with clo-
137 miphene on the basis of chance alone. Fourteen women were suppressed for 1 to 3 months with large doses of exogenous estrogen and then were administered human menopausal gonadotropins (from 50 to 100 ampules, with each ampule containing 75 IU of FSH and 75 IU of LH). We subsequently administered menotropins to five additional women suppressed previously for 1 to 3 months with a gonadotropinreleasing hormone agonist. Two of the patients suppressed with the agonist had evidence of significant follicular activity and ovulation, and one conceived. Thus, ovulation induction is unlikely to be successful in these women. Twelve women with secondary amenorrhea had ovarian biopsies, with apparently viable oocytes noted in seven of the specimens. Yet two of the eight subsequent pregnancies occurred in women with no follicles observed on biopsy. Fully seven of these eight pregnancies occurred while the patients were taking exogenous estrogen; the remaining pregnancy in this series occurred in response to clomiphene. Five of the eight pregnancies resulted in live term births, two ended in spontaneous abortion, and one ended in elective abortion. Only three patients with primary amenorrhea underwent gonadal biopsy: the two with 46,XY karyotypes had dysgerminomas. The one additional patient had fibrous streaks. These observations lead to the obvious conclusion that hypergonadotropic amenorrhea is a heterogeneous disorder. No doubt many of these young women have premature ovarian failure, but clearly others do not, as documented by subsequent ovulations and pregnancies. It would also seem logical to conclude that premature ovarian failure might be the end result of several varied disorders. Ovarian biopsy cannot be recommended in view of documented pregnancies in women who had no follicles found on biopsy. Because of the low incidence of ovulation and pregnancy among women undergoing ovulation induction, it is likewise difficult to recommend such efforts. Moreover, these clinical observations stress the importance for subsequent management of measuring basal FSH concentrations in all amenorrheic women. It is clear that progestin-induced withdrawal cannot be used to distinguish women with chronic anovulation from those with impending ovarian failure.
III. PREVALENCE OF PREMATURE OVARIAN FAILURE Estimation of the prevalence of premature ovarian failure in the general population is difficult, de Moraes-Ruehsen and Jones [ 1] found that 7% of 300 consecutive women presenting with amenorrhea had premature ovarian failure. Aiman and Smentek [ 11 ] combined the observations of several investigators to estimate the frequency of the disorder among American women. Based on the assumptions that 43 million
138
ROBERT W. REBAR
American women were of reproductive age in 1985 and that the incidence of amenorrhea was 3%, they concluded that the frequency of premature ovarian failure is approximately 0.3%, with 129,000 American women being affected in that year. Alper and colleagues [18] estimated that 5 to 10% of women with secondary amenorrhea have this disorder. Coulam and co-workers [19] examined the medical records of 1858 women living in Rochester, Minnesota, in 1950 and calculated that the risk of experiencing menopause prior to age 40 was 0.9%.
IV. ETIOLOGY OF PREMATURE OVARIAN FAILURE de Moraes-Ruehsen and Jones [1] suggested three possible explanations for the early completion of atresia that they believed to exist in women with hypergonadotropic amenorrhea and premature ovarian failure: (1) decreased germ cell endowment, (2) accelerated atresia, and (3) postnatal germ cell destruction. Because these possibilities cannot apply to individuals in whom many follicles still remain, some block to gonadotropin action in ovarian follicles must exist. In view of data that even postmenopausal women may have a few remaining follicles [20,21] and previously cited information that follicle number decreases rapidly in the last several months before menopause [2], it is possible that a few women who are truly "perimenopausal" will ovulate and even conceive. Among the various causes of hypergonadotropic amenorrhea that can be identified, it is clear that some are present only in those who have no oocytes whereas others may have the potential for ovulation and spontaneous pregnancy. Possible causes of hypergonadotropic amenorrhea and premature ovarian failure are listed in Table II.
A. Genetic and Cytogenetic Causes 1. ~ F A M I L I A L ~ P R E M A T U R E OVARIAN F A I L U R E Several reports have described individual families with vertical transmission of premature ovarian failure, implying autosomal dominant, sex-linked inheritance [22-24]. In such families the etiology might well be due to one of the three reasons postulated by de Moraes-Ruehsen and Jones [ 1]. It certainly is well recognized that the number of oocytes differs widely among various strains of mice [25]. Moreover, both individual mice [3,25] and humans [15,26,27] have markedly different rates of follicular atresia. It is possible that the etiology of premature ovarian failure that occurs in some individuals with the neurological disorder myotonia dystrophica also is due to a decreased endowment in germ cell number or to accelerated atresia [28].
TABLE II Causes of Hypergonadotropic Amenorrhea Genetic and cytogenetic etiologies "Familial" premature ovarian failure FSH receptor mutations Fragile X premutations Structural alterations or absence of an X chromosome Trisomy X with or without mosaicism In association with myotonia dystrophica Enzymatic defects 17ce-Hydroxylase deficiency Galactosemia Physical insults Ionizing radiation Chemotherapeutic agents Viral infection Cigarette smoking Surgical extirpation Immune disturbances In association with other autoimmune disorders Isolated Congenital thymic aplasia Defects in gonadotropin structure or actions (genetic?) Secretion of biologically inactive gonadotropin ce or fl subunit defects Gonadotropin receptor or postreceptor defects Circulating FSH-binding inhibitors Idiopathic
Molecular biology has provided explanations for some familial cases of hypergonadotropic ovarian failure. Aittom~iki and colleagues [29] demonstrated a mutation in exon 7 of the FSH receptor gene located on chromosome 2p in which an Ala to Val substitution at residue 189 is present in six Finnish families with multiple affected women. The histological appearance of the ovaries of women with this mutation showed hypoplasia with few primordial follicles [30]. None had the appearance of complete ovarian dysgenesis with streak ovaries. That such FSH receptor abnormalities are rare causes of ovarian failure is suggested by a study in Great Britain failing to identify any such mutations in 30 women with sporadic premature ovarian failure and in 18 women with familial premature ovarian failure [31 ]. It is now clear that women carrying one X chromosome with a fragile X premutation have an increased prevalence of premature menopause [32]. Fragile X syndrome is the most common inherited form of male retardation and is caused by an expansion of a trinucleotide repeat sequence in the first exon of the FMR1 gene (Xq27.3). The full fragile X mutation occurs when the number of trinucleotide repeats exceeds 200, when gene transcription fails and the FMR1 protein is not expressed [33]. In normal individuals there are less than 50 trinuleotide repeats at this fragile site and a fragile X premutation is said to occur when between 50 and 200 tri-
CHAPTER8 Premature Ovarian Failure nucleotide repeats are present. At least one other group has reported that fragile X premutations occur at least 10 times more frequently in women with premature ovarian failure than in the general population [34]. However, no causal link has yet been shown. Because the FMR1 gene is expressed both in the brain and in the gonad, the fragile X premutation may not be as "innocent" as presumed [35].
139 moieties on gonadotropin molecules are altered such that they are biologically inactive or their metabolism is changed. Unfortunately, this postulate does not coincide with experimental data suggesting a direct effect of galactose on the oocyte. Pregnant rats fed a 50% galactose diet delivered pups with significantly reduced numbers of oocytes, apparently due to decreased germ cell migration to the genital ridges [48].
2. GONADAL DYSGENESIS
Studies of individuals with gonadal dysgenesis have documented that two intact X chromosomes are needed for normal maintenance of oocytes. It is known that the gonads of 45,X fetuses contain the normal complement of oocytes at 20 to 24 weeks of fetal age, but that those oocytes rapidly undergo atresia so that virtually none remains at birth [36]. Structural abnormalities of the X chromosome also can affect ovarian function and have been found in women with premature ovarian failure [7,11,37]. Even documented submicroscopic deletions of a portion of X chromosome can apparently lead to premature ovarian failure [38]. 3. TRISOMY X WITH OR WITHOUT MOSAICISM
An excess of X chromosomes also may be found in some women who develop premature menopause [39]. Patients identified to date have developed normal secondary sex characteristics and only later presented with secondary hypergonadotropic amenorrhea. Reports of the triple-X syndrome associated with immunoglobulin deficiency [40] and Marfan syndrome [41], together with the observation that control of T cell function may be related to the X chromosome [42], suggest a possible association between immunological abnormalities and triple-X females with premature ovarian failure.
C. P h y s i c a l Insults Destruction of oocytes by any of several environmental insults, including ionizing radiation, various chemotherapeutic agents, certain viral infections, and even cigarette smoking, may occur [49]. 1. IONIZING RADIATION
Approximately 50% of individuals who receive 400 to 500 rads to the ovaries over 4 to 6 weeks, as may occur in treatment for Hodgkin's disease, will develop permanent hypergonadotropic amenorrhea [ 17,50,51 ]. For any given dose of radiation, the older the woman, the greater her likelihood of developing amenorrhea. It appears that a total of 800 rads is sufficient to result in permanent sterility in all women [50,51]. That amenorrhea following radiation therapy is not always permanent was reported by Jacox in 1939 [52]. The transient nature of amenorrhea in some women suggests that some follicles may be damaged but not destroyed by relatively low doses of radiation. Although surgical transposition of the ovaries outside the field of irradiation is now common practice, it is not clear just how fertile such women ultimately are. 2. CHEMOTHERAPEUTIC AGENTS
B. E n z y m a t i c D e f e c t s 1. 17a-HYDROXYLASE DEFICIENCY The rare women with deficiency of the 17a-hydroxylase enzyme are identified easily because of the associated findings of primary amenorrhea, sexual infantilism, hypergonadotropinism, hypertension, hypokalemic alkalosis, and increased circulating levels of deoxycorticosterone and progesterone [43-45]. Ovarian biopsies have revealed numerous large cysts and follicular cysts, with complete failure of follicular maturation [45]. 2. GALACTOSEMIA
Women with galactosemia develop amenorrhea with elevated gonadotropin levels even when treatment with a galactose-restricted diet begins at an early age [46,47]. Although the etiology of premature ovarian failure in galactosemia is unknown, it is tempting to speculate that the carbohydrate
As more and more young women treated for childhood malignancies, especially leukemias and lymphomas, survive long-term, it has become obvious that chemotherapeutic agents may produce temporary or permanent ovarian failure [17,53-57]. Alkylating agents, particularly cyclophosphamide, are most likely to affect reproductive function. In general, the younger the woman at the time of therapy, the more likely it is that ovarian function will not be compromised by chemotherapy. It may well be that it is the number of oocytes present at the time of therapy that determines if ovarian function will be compromised: the greater the number of oocytes, the more likely it is that normal ovarian function will persist. The frequency of congenital anomalies does not appear to be increased in the children of women previously treated with chemotherapy [58]. It has been suggested, however, that one agent, dactinomycin, may be associated with an increased risk of congenital heart disease, and further studies in this area are clearly needed.
1
4
0
R
O
B
E
3. VIRAL AND OTHER AGENTS
Although several viruses are believed to have the potential to cause ovarian destruction, confirming that such is the case in humans is difficult. Morrison and colleagues [59] reported three presumptive cases in which "mumps oophoritis" preceded premature ovarian failure, including a mother and daughter pair, in which the mother experienced mumps parotitis and abdominal pain during pregnancy just prior to delivery of the daughter. Although there is no evidence that cigarette smoking will lead to premature menopause, data do exist documenting that cigarette smokers experience menopause several months earlier than do nonsmokers [60].
D. I m m u n e D i s t u r b a n c e s Any role for immune disturbances in the etiology of hypergonadotropic amenorrhea remains controversial. It is clear that several autoimmune abnormalities may occur in association with hypergonadotropic amenorrhea (Table III). However, the prevalence of autoimmune abnormalities in normal women is unknown, and it may be that it is not increased in ovarian failure. Moreover, it is not clear if autoimmune disturbances play any role in the development of hypergonadotropic amenorrhea. As is characteristic for other autoimmune disturbances, the ovarian "failure" in affected women may wax and wane, and pregnancies may occur, at least early in the disease process. In a literature review tabulating 380 cases of premature ovarian failure, LaBarbera and colleagues [61] noted that 17.5% had a definite associated autoimmune disorder. Additional evidence that hypergonadotropic amenorrhea may have an autoimmune etiology in at least some cases has been provided by sporadic case reports documenting return of ovarian function following either immunosuppressive therapy or recovery from an autoimmune disease [62-64].
TABLE III
T
Adapted from Rebar et
al.
[7] and LaBarbera et
Juvenile rheumatoidarthritis Keratoconjunctivitisand Sj6gren's syndrome Malabsorption syndrome Myasthenia gravis Polyendocrinopathies(type I, type II, and unspecified) Primary biliary cirrhosis Quantitative immunoglobulinabnormalities Rheumatoid arthritis Systemic lupus erythematosus Thyroid disorders, including Graves' disease and thyroiditis Vitiligo al.
W. REBAR
In a few cases lymphocytic infiltrates suggesting autoimmune dysfunction have been observed in ovarian biopsy specimens [65]. Still other immune abnormalities have been identified in some patients with premature ovarian failure. Enhanced release of leukocyte migration inhibition factor (MIF) by peripheral lymphocytes has been observed following exposure of the lymphocytes to crude ovarian proteins [66,67]. A significant association of early ovarian failure with HLA-DR3 has been noted [68], perhaps suggesting a genetic susceptibility to autoimmunity in some individuals. Several years ago McNatty and colleagues [69] reported complementdependent cytotoxic effects on cultured granulosa cells, as documented by inhibition of progesterone production and cell lysis, in sera from 9 of 23 patients with hypergonadotropic amenorrhea and Addison's disease. Cellular immune abnormalities involving numbers and/or function of peripheral monocytes and of subsets of T cells and B cells have also been noted in women with premature ovarian failure [70]. Indirect immunofluorescence of ovarian biopsy specimens has revealed antibodies reacting with various ovarian components in some patients [71]. Circulating immunoglobulins to ovarian proteins have been detected by immunocytochemical techniques by several investigators [61]. Utilizing a solid-phase, enzyme-linked immunosorbent assay, we have detected antibodies to ovarian tissue in 22% of karyotypically normal women with premature ovarian failure [72,73]. The most thoroughly documented study to the present time remains that of Chiauzzi and colleagues [74], who documented that two patients with ovarian failure and myasthenia gravis had circulating immunoglobulin G that blocked binding of FSH to its receptor. However, it is important to reiterate that ovarian autoantibodies may not be the cause of ovarian failure. Rather, the ovarian failure may result from cell-mediated autoimmunity, and autoantibodies may appear only because of the resultant cell death. However, Anasti and colleagues [75] failed to demonstrate the
Possible Autoimmune Disorders Associated with Premature Ovarian Failure a
Alopecia Anemia, both acquired hemolyticand pernicious Asthma Chronic active hepatitis Crohn's disease Diabetes mellitus Glomerulonephritis Hypoadrenalism (Addison's disease) Hypoparathyroidism Hypophysitis Idiopathic thrombocytopeniapurpura a
R
[61].
CHAPTER8 Premature Ovarian Failure presence of blocking antibodies to LH or FSH receptors in any of 38 premature ovarian failure patients studied. In recent years there has been increasing interest in the relationship between the immune and reproductive systems. Miller and Chatten [76] documented that congenitally athymic girls dying before puberty had ovaries devoid of oocytes on autopsy. Data from our laboratory suggest that the thymus gland may be necessary early in development for normal gonadotropin function. Congenitally athymic mice, well known to develop premature ovarian failure, have lower gonadotropin concentrations prepubertally than do their normal heterozygous littermates [77]. These hormonal alterations, as well as the accelerated loss of oocytes, can be prevented by thymic transplantation at birth [78]. In comparing ovarian development in the rodent to that of the primate, it is essential to recognize that development occurring during the first few weeks of life in the mouse occurs in utero in the human female. Thus, thymic ablation in fetal rhesus monkeys in late gestation is associated with a marked reduction in oocyte number at birth [79]. One possible explanation for the association of thymic aplasia and ovarian failure may be found in our observation that peptides produced by the thymus can stimulate release of gonadotropin-releasing hormone (GnRH) and consequently LH [80]. In recent years it has become evident that organ-specific autoimmunity may be directed against intracellular enzymes, particularly those involved in hormone synthesis [70,81 ]. For example, thyroid peroxidase is a major thyroid autoantigen for autoimmune hypothyroidism, and the 21-hydroxylase enzyme is the foremost autoantigen in Addison's disease. One group has identified 3fi-hydroxysteroid dehydrogenase as an autoantigen in 20% of women with premature ovarian failure [82]. This, too, may merely be an epiphenomenon of ovarian inflammation rather than causal for the development of ovarian failure. From a theoretical viewpoint, identifying patients with an autoimmune etiology for their hypergonadotropic amenorrhea is important because affected patients might be treated effectively early in the disease process before all viable oocytes are destroyed.
141 cases of male pseudohermaphroditism with immunologically active but biologically inactive LH are well documented [84,85]. Altered forms of immunoreactive LH and FSH have been reported in urinary extracts from women with premature ovarian failure compared to those from oophorectomized and postmenopausal women [86], suggesting that metabolism and/or excretion of gonadotropins is altered in some cases of this disorder. However, using two different probes for the fl subunit of FSH (as well as two probes for the FSH receptor gene), one group failed to find any mutations in a small group of patients [87]. These findings do not rule out mutations in other patients or in different portions of the molecule. Interference with F S H action at the ovarian level also might lead to early ovarian failure. Defects in FSH receptor structure (as reported in the Finnish study [29]), FSH receptor antibodies (as noted), competitive inhibitors to FSH binding, or defects in postreceptor systems that mediate hormone action are each theoretically possible. Sluss and Schneyer [88] reported identifying two individuals out of a group of 27 with hypergonadotropic amenorrhea (and intermittent evidence of ovarian function) whose sera had lowmolecular-weight FSH receptor-binding activity that was an antagonist of FSH action. Even when this inhibitor was removed from the serum, however, FSH levels were elevated in both patients. These studies cannot eliminate the possibility that this FSH binding inhibitor is merely produced secondarily to development of ovarian FSH insensitivity. Other possible defects in gonadotropin action remain to be identified. Clearly all of these disorders might well be genetic abnormalities.
E
Idiopathic
The diagnosis of idiopathic causes of premature ovarian failure is one of exclusion, but presently no definitive etiology is identified in most cases of hypergonadotropic amenorrhea. It is likely that additional causes of this entity will be recognized as more is learned about premature ovarian failure.
E. D e f e c t s in G o n a d o t r o p i n S t r u c t u r e or A c t i o n It is possible to envision that abnormal structure, secretion, or metabolism of gonadotropins in some women forms the basis for early ovarian failure. The concept of secretion of altered molecular forms of FSH with reduced or absent biological activity leading to accelerated follicular atresia even as a rare cause for premature ovarian failure is appealing. This is especially true given evidence that normal concentrations of gonadotropins are required early in development: fetal hypophysectomy in rhesus monkeys leads to the newborn having no oocytes in their ovaries [83]. Moreover,
G. R e s i s t a n t O v a r y S y n d r o m e : A Term No Longer Useful As originally defined, the resistant ovary, or "Savage," syndrome was found in young amenorrheic women with (1) elevated peripheral gonadotropin levels, (2) normal but immature follicles in the ovaries, (3) a 46,XX karyotype, (4) mature secondary sex characteristics, and (5) decreased sensitivity to stimulation with exogenous gonadotropin [89]. Individuals fulfilling these criteria might easily have any
142
ROBERT W. REBAR
of several different etiologies for their hypergonadotropic amenorrhea. Moreover, regardless of the etiology, these features may be common to all individuals with hypergonadotropic amenorrhea at some time during the disease process prior to final loss of all oocytes. As a consequence, use of the term resistant ovary will become less and less useful as understanding of ovarian failure increases, and the terminology is already of severely restricted value.
V. EVALUATION OF PATIENTS WITH HYPERGONADOTROPIC AMENORRHEA Young women with hypergonadotropic amenorrhea should be evaluated to identify (1) specific, potentially treatable causes and (2) other potentially dangerous associated disorders. A thorough history and physical examination are always warranted. A maturation index and evaluation of the cervical mucus may help determine if any endogenous estrogen is present. Simple laboratory tests should be performed to exclude thyroid disease, hypoparathyroidism, adrenal insufficiency, diabetes mellitus, and other forms of immune dysfunction. The extent of such testing is unclear, but a reasonable set of tests is listed in Table IV. In addition to the clinical evaluation of estrogen status, measurement of circulating LH, FSH, and estradiol concentrations on more than one occasion may help determine if any functional follicles are present. If the estradiol concentration is greater than 50 pg/ml or if the LH level is greater than the FSH (in terms of mIU/ml) in any sample, then a few viable oocytes still must be present. Irregular uterine bleeding, indicative of continuing estrogen production, also suggests the presence of remaining functional oocytes. Identifiable follicles on transvaginal ultrasonography also can be used to identify women with remaining TABLE IV Evaluation of Hypergonadotropic Amenorrhea in Young Women a Complete history and physical examination Maturation index Karyotype (? limited to women with onset before age 30) Complete blood count with differential, sedimentation rate, total serum protein and albumin/globulin ratio, rheumatoid factor, antinuclear antibody Fasting blood glucose, serum calcium and phosphorus, evaluation of adrenal status T 4, thyroid-stimulating hormone, antithyroglobulin, and antimicrosomal antibodies or antithyroid-stimulating immunoglobulins Serum LH, FSH, and estradiol on at least two occasions Evaluation of bone mineral density a Adapted from Rebar et al. [7].
oocytes [90] and are present in a large percentage of affected women [ 12,13,90]. If available, testing of the patient's serum for antibodies to endocrine tissues, including ovary, may be of some value. The difficulty with this recommendation is the fact that there are no readily available tests for antibodies to any specific antigens. In addition, as noted previously, antibodies may develop because of cytotoxicity in the ovary and may not cause ovarian failure. In which patients chromosomal studies should be conducted is also unclear. It would seem prudent to obtain a karyotype in women with the onset of hypergonadotropic amenorrhea prior to age 30 to identify those with various forms of gonadal dysgenesis, individuals with mosaicism, those with trisomy X, and those with a portion of a Y chromosome. If a Y chromosome is present, gonadal extirpation is warranted because of the increased risk of malignancy [91-93]. Chromosomal evaluation also may be warranted to rule out familial transmission in women who develop hypergonadotropic amenorrhea after the birth of daughters. Although controversial, ovarian biopsy does not appear justified in women with hypergonadotropic amenorrhea and a normal karyotype. It is not clear how the results would alter therapy. Aiman and Smentek [ 11 ] reported one of their two patients who eventually conceived had no oocytes present on biopsy. Similarly, Rebar and Connolly [ 16] noted that two of eight subsequent pregnancies among 97 women with secondary hypergonadotropic amenorrhea occurred in women with no follicles present in ovarian tissue obtained by laparotomy. As also noted by Aiman and Smentek [11], if five sections of an ovarian biopsy are examined and each is 6/zm thick, then the presence of follicles is sought from a sample representing less than 0.15% of an ovary measuring 2 • 3 x 4 cm. Thus, the absence of follicles on biopsy may not be representative of the remainder of the ovary. Moreover, affected individuals almost always require estrogen replacement regardless of the results of the biopsy. Evaluation of bone density appears warranted in women with hypergonadotropic amenorrhea because of the high prevalence of osteopenia [13,16,90]. Periodic assessment may be warranted, regardless of therapy, to assess the rapidity of bone loss. Similarly, monitoring patients for the development of autoimmune endocrinopathies may be warranted even if all tests are normal when the patient is first evaluated; development of other disorders after diagnosis of hypergonadotropic amenorrhea does occur [16].
VI. THERAPY It is reasonable to treat all young women with hypergonadotropic amenorrhea with exogenous estrogen regardless of whether they are interested in childbearing. The accelerated bone loss frequently accompanying this disorder can be
CHAPTER 8 Premature Ovarian Failure prevented by administration of exogenous estrogens [90]. So, too, may the increased risk of cardiovascular disease present in women with estrogen deficiency [94]. Although it also appears that women with premature ovarian failure are at reduced risk of breast cancer [95] (and probably venous thrombosis), the hope is that administration of exogenous estrogen merely returns these relative risks to those found in normal premenopausal women. In addition, almost all spontaneous pregnancies in this disorder occur during or following estrogen administration [16,96]. Even with exogenous estrogens, however, the probability of spontaneous pregnancy appears to be less than 10%. The pregnancy rate is low despite the fact that one-fourth or more of women ovulate after the diagnosis of hypergonadotropic amenorrhea is made. Because of the possibility of pregnancy, women taking exogenous estrogens in any form, even as part of oral contraceptive agents, should be advised to contact their physician if they develop any signs or symptoms of pregnancy or do not have withdrawal bleeding. Although it may not be necessary to advise the use of barrier forms of contraception, the possibility of pregnancy must be discussed. Either oral contraceptives or sequential estrogen-progestin therapy may be utilized, but sequential therapy is more physiologic. It is important to remember that these young women may require twice as much estrogen as do menopausal women to alleviate signs and symptoms of hypoestrogenism. Several isolated case reports have suggested that ovarian suppression with estrogen or a GnRH agonist followed by stimulation with human menopausal gonadotropin may be efficacious in inducing ovulation and allowing conception [97-100]. Most of these reports emanate from one group of investigators. Larger studies suggest that the possibility of successful ovulation induction and pregnancy is small indeed and may be no greater than what concurs spontaneously in these patients [ 14,16,101 ]. I n v i t r o fertilization involving oocyte donation clearly provides individuals with hypergonadotropic amenorrhea the greatest likelihood of bearing children. The first successful case of oocyte donation in humans was reported in 1984. A young woman with ovarian failure was given oral estradiol valerate and progesterone pessaries to prepare the endometrium for transfer of a single donated oocyte following fertilization with her husband's sperm [102]. Since then, several programs utilizing oocyte donation have been successful because of (1) improvements in transvaginal ultrasonography, allowing follicular aspiration and oocyte collection without surgery, (2) improvements in success with embryo cryopreservation and subsequent embryo transfer to the donor at a later time, and (3) improved ability to synchronize artificial cycles in the recipient with the hyperstimulation cycles in the donor, generally by use of GnRH agonists [103-105]. A number of replacement protocols have been developed for the donor with hypergonadotropic amenorrhea, including use of oral, transvaginal, and transdermal administration of
143 estradiol and oral, and transvaginal and intramuscular administration of progesterone. If pregnancy develops from the transferred embryo, given the absence of functional gonads in the recipient, exogenous supplementation with estradiol and progesterone must be continued until placental production of progesterone is well established. Success rates generally have exceeded those observed in standard in v i t r o fertilization programs [103-106]. Thus, oocyte donation offers the possibility of pregnancy to all women with premature ovarian failure so long as a normal uterus is present.
References 1. de Moraes, M., and Jones, G. S. (1967). Premature ovarian failure. Fertil. Steril. 18, 440. 2. Richardson, S. J., Senikas, V., and Nelson, J. E (1987). Follicular depletion during the menopausal transition: Evidence for accelerated loss and ultimate exhaustion. J. Clin. Endocrinol. Metab. 65, 12311237. 3. Spearow, J. L., and Barkley, M. (1992). Mapping genes controlling induced ovulation rate in mice. Abstr., 74th Annu. Meet. Endocr. Soc. San Antonio, TX. 1992, Abstr. No. 1462, 417. 4. Simpson,J. L., and Rebar, R. W. (1999). Normal and abnormal sexual differentiation and development. In "Principles and Practice of Endocrinology and Metabolism" (K. K. L. Becker et al., eds.), 2nd ed. Lippincott, Philadelphia. 5. Parkening, T. A., Collins, T. J., and Elder, E F. B. (1985). Orthotopic ovarian transplantations in young and aged C57BL/6J mice. Biol. Reprod. 32, 989-997. 6. Goldenberg,R. L., Grodin, J. M., Rodbard, D., and Ross, G. T. (1973). Gonadotropins in women with amenorrhea. Am. J. Obstet. Gynecol. 116, 1003-1012. 7. Rebar, R. W., Erickson, G. E, and Coulam, C. B. (1987). Premature ovarian failure. In "Pathology of Infertility" (B. Gondos and D. Riddick, eds.), pp. 123-141. Thieme Medical Publishers, New York. 8. Lucky,A. W., Rebar, R. W., Blizzard, R. M., and Goren, E. M. (1977). Pubertal progression in the presence of elevated serum gonadotropins in girls with multiple endocrine deficiencies. J. Clin. Endocrinol. Metab. 45, 673-678. 9. O'Herlihy, C., Pepperell, R. J., and Evans, J. H. (1980). The significance of FSH elevation in young women with disorders of ovulation. Br. Med. J. 281, 1447. 10. Rebar, R. W., Erickson, G. F., and Yen, S. S. C. (1982). "Idiopathic premature ovarian failure": Clinical and endocrine characteristics. Fertil. Steril. 37, 35-41. 11. Aiman, J., and Smentek, C. (1985). Premature ovarian failure. Obstet. Gynecol. 66, 9-14. 12. Hague, W. M., Tan, S. L., Adams, J., and Jacobs, H. S. (1987). Hypergonadotropic amenorrhea--Etiology and outcome in 93 young women. Int. J. Gynaecol. Obstet. 25, 121-125. 13. Conway,G. S., Kaltsas, G., Patel, A., Davies, M. C., and Jacobs, H. S. (1996). Characterization of idiopathic premature ovarian failure. Fertil. Steril. 65, 337-341. 14. Nelson, L. M., Kimzey, L. M., and Merriam, G. R. (1992). Gonadotropin suppression for the treatment of karyotypically normal spontaneous premature ovarian failure: A controlled trial. Fertil. Steril. 57, 50-55. 15. Block, E. (1953). A quantitative morphological investigation of the follicular system in newborn female infants. Acta Anat. 17, 201-206.
144 16. Rebar, R. W., and Connolly, H. V. (1990). Clinical feature of young women with hypergonadotropic amenorrhea. Fertil. Steril. 53, 804810. 17. Siris, E. S., Leventhal, B. G., and Vaitukaitis, J. L. (1976). Effects of childhood leukemia and chemotherapy on puberty and reproductive function in girls. N. Engl. J. Med. 294. 1143-1146. 18. Alper, M. M., Garner, P. R., Cher, B., and Seibel, M. M. (1986). Premature ovarian failure. J. Reprod. Med. 31,699-708. 19. Coulam, C. B., Adamson, S. C., and Annegers, J. E (1986). Incidence of premature ovarian failure. Obstet. Gynecol. 67, 604-606. 20. Costoff, A., and Mahesh, V. B. (1975). Primordial follicles with normal oocytes in the ovaries of postmenopausal women. J. Am. Geriatr. Soc. 23, 193-196. 21. Hertig, A. T. (1944). The aging ovary--A preliminary note. J. Clin. Endocrinol. Metab. 4, 581-582. 22. Coulam, C. B., Stringfellow, S. S., and Hoefnagel, D. (1983). Evidence for a genetic factor in the etiology of premature ovarian failure. Fertil. Steril. 40, 693-695. 23. Mattison, D. R., Evans, M. I., Schwimmer, W. B., White, B. J., Jensen, B., and Schulman, J. D. (1984). Familial premature ovarian failure. Am. J. Hum. Genet. 36, 1341-1348. 24. Starup, J., Philip, J., and Sele, V. (1978). Oestrogen treatment and subsequent pregnancy in two patients with severe hypergonadotropic ovarian failure. Acta Endocrinol. (Copenhagen) 89, 149-157. 25. Jones, E. C., and Krohn, P. L. (1961). The relationship between age, numbers of oocytes and fertility in virgin and multiparous mice. J. Endocrinol. 21,469-495. 26. Baker, T. G. (1963). A quantitative and cytological study of germ cells in human ovaries. Proc. Soc. London, Ser. B. 158, 417-433. 27. Block, E. (1952). Quantitative morphological investigations of the follicular system in women. Variations at different ages. Acta Anat. 14, 108-123. 28. Harper, P. S., and Dyken, P. R. (1972). Early onset dystrophia myotonia. Lancet 2, 53-55. 29. Aittom~iki, K., Lucena, J. L. D., Pakarinen, P., Sistonen, P., Tapanalnen, J., Gromoll, J., Kaskikari, R., Sankila, E. M., Lehvaslaiho, H., and Engel, A. R. (1995). Mutation in the follicle-stimulating hormone receptor gene causes hereditary hypergonadotropic ovarian failure. Cell (Cambridge, Mass.) 82, 959-968. 30. Aittom~iki, K., Herva, U.-H., Juntunen, K., Ylostalo, P., Horvatta, O., and de la Chapelle, A. (1996). Clinical features of primary ovarian failure caused by a point mutation in the follicle-stimulating hormone receptor gene. J. C. Clin. Endocrinol. Metab. 81, 3722-3726. 31. Conway, E., Hoppner, W., Gromoll, J., Simoni, M., and Conway, G. S. (1997). Mutations of the FSH receptor gene are rate in familial and sporadic premature ovarian failure. J. Endocrinol. 152 (Suppl.), P257, (abstr.). 32. Turner, G., Robinson, H., Wake, S., and Martin, S. (1994). Dizygous twinning and premature menopause in fragile X syndrome. Lancet 344, 1500. 33. Jacobs, P. A. (1991 ). Fragile X syndrome. J. Med. Genet. 28, 809-810. 34. Conway, G. S., Hettiarachchi, S., Murray, A., and Jacobs, P. A. (1995). Fragile X premutations in familial premature ovarian failure. Lancet 346, 309-310. 35. Hinds, H. L., Ashley, C. L., Sutcliffe, J. S., Nelson, D. L., Warren, S. T., H~risman, D. E., and Schalling, M. (1993). Tissue specific expression of FMR-1 provides evidence for a functional role in fragile X syndrome. Nat. Genet. 2, 197-200. 36. Singh, R. P., and Carr, D. H. (1966). The anatomy and histology of XO human embryos and fetuses. Anat. Rec. 155, 369-381. 37. Kinch, R. A. H., Plunkett, E. R., Smout, M. S., and Carr, D. H. (1965). Primary ovarian failure: A clinicopathological and cytogenetic study. Am. J. Obstet. Gynecol. 91,630-644. 38. Krauss, C. M., Turksoy, R. N., Atkins, L., Laughlin, C., Brown, L. G.,
ROBERT W. REBAR
39. 40.
41.
42.
43. 44.
45.
46. 47.
48.
49. 50. 51.
52. 53.
54.
55.
56.
57.
58.
and Page, D. C. (1987). Familial premature ovarian failure due to an interstitial deletion of the long arm of the X chromosome. N. Engl. J. Med. 317, 125-131. Villanueva, A. L., and Rebar, R. W. (1983). The triple X syndrome and premature ovarian failure. Obstet. Gynecol. 62, 705-735. Sills, J. A., Brown, J. K., Grace, S., Wood, S. M., Barclay, G. R., and Urbaniak, S. J. (1978). XXX syndrome associated with immunoglobulin deficiency and epilepsy. J. Pediatr. 93, 469-471. Smith, T. F., and Engel, E. (1981). Marfan's syndrome with 47,XXX genotype and possible immunologic abnormality. South. Med. J. 74, 630-632. Purtillo, D. T., DeFlorio, D., Jr., Hutt, L. H., Bhawan, J., Yang, J. P., Otto, R., and Edwards, W. (1977). Variable phenotypic expression of an X-linked recessive lymphoproliferative syndrome. N. Engl. J. Med. 279, 1077-1082. Bigleri, E. G., Herron, M. A., and Brust, N. (1966). 17-Hydroxylation deficiency in man. J. Clin. Invest. 45, 1946-1954. Goldsmith, O., Solomon, D. H., and Horton, R. (1967). Hypogonadism and mineralocorticoid excess. The 17-hydroxylase deficiency syndrome. N. Engl. J. Med. 277, 673-677. Mallin, S. R. (1969). Congenital adrenal hyperplasia secondary to 17-hydroxylase deficiency. Two sisters with amenorrhea, hypokalemia, hypertension, and cystic ovaries. Ann. Intern. Med. 70, 69-75. Hoefnagel, D., Wurster-Hili, D., and Child, E. L. (1979). Ovarian failure in galactosemia. Lancet 2, 1197. Kaufman, F. R., Kogut, M. D., Donnell, G. N., Goebelsmann, U., March, C., and Koch, R. (1981). Hypergonadotropic hypogonadism in female patients with galactosemia. N. Engl. J. Med. 304, 994-998. Chen, Y.-T., Mattison, D. R., Feigenbaum, L., Fukui, H., and Schulman, J. D. (1981 ). Reduction in oocyte number following prenatal exposure to a diet high in galactose. Science 214, 1145-1147. Verp, M. S. (1983). Environmental causes of ovarian failure. Semin. Reprod. Endocrinol. 1, 101 - 11. Ash, P. (1980). The influence of radiation on fertility in man. Br. J. Radiol. 53, 271-278. Ray, G. R., Trueblood, H. W., Enright, L. P., Kaplan, H. S., and Nelsen, T. S. (1970). Oophoropexy: A means of preserving ovarian function following pelvic megavoltage radiotherapy for Hodgkin's disease. Radiology 96, 175 - 180. Jacox, H. W. (1939). Recovery following human ovarian irradiation. Radiology 32, 538-545. Damewood, M. D., and Grochow, L. B. (1986). Prospects for fertility after chemotherapy or radiation for neoplastic disease. Fertil. Steril. 45, 443-459. Horning, S. J., Hoppe, R. T., Kaplan, H. S., and Rosenberg, S. A. (1981). Female reproductive potential after treatment for Hodgkin's disease. N. Engl. J. Med. 304, 1377-1382. Koyama, H., Wada, T., Nishizawa, Y., Iwanaga, T., Aoki, Y., Terasawa, T., Kosaki, G., Yamamoto, T., and Wada, A. (1977). Cyclophosphamide-induced ovarian failure and its therapeutic significance in patients with breast cancer. Cancer (Philadelphia) 39, 1403-1409. Stillman, R. J., Schiff, I., and Schinfeld, J. (1982). Reproductive and gonadal function in the female after therapy for childhood malignancy. Obstet. Gynecol. Surv. 37, 385-393. Whitheead, E., Shalet, S. M., Blackledge, G., Crowther, D., and Beardwell, C. G. (1983). The effect of combination chemotherapy on ovarian function in women treated for Hodgkin's disease. Cancer (Philadelphia) 52, 988-993. Green, D. M., Zevon, M. A., Lowrie, G., Seigelstein, N., and Hall, B. (1991). Congenital anomalies in children of patients who received chemotherapy for cancer in childhood and adolescence. N. Engl. J. Med. 325, 141-146.
145
CHAPTER 8 Premature Ovarian Failure 59. Morrison, J. C., Givens, J. R., Wiser, W. L., and Fisk, S. A. (1975). Mumps oophoritis: A cause of premature menopause. Fertil. Steril. 26, 655-659. 60. Jick, H., Porte, J., and Morrison, A. S. (1977). Relation between smoking and age of natural menopause. Lancet 1, 1354-1355. 61. LaBarbera, A. R., Miller, M. M., Ober, C., and Rebar, R. W. (1988). Autoimmune etiology in premature ovarian failure. Am. J. Reprod. Immunol. Microbiol. 16, 115-122. 62. Bateman, B. G., Nunley, W. C., and Kitchin, J. D., III (1983). Reversal of apparent premature ovarian failure in a patient with myasthenia gravis. Fertil. Steril. 39, 108-110. 63. Coulam, C. B., Kempers, R. D., and Randall, R. V. (1981). Premature ovarian failure: Evidence for the autoimmune mechanism. Fertil. Steril. 36, 238-240. 64. Lucky, A. W., Rebar, R. W., Blizzard, R. M., and Goren, E. M. (1977). Pubertal progression in the presence of elevated serum gonadotropins in girls with multiple endocrine deficiencies. J. Clin. Endocrinol. Metab. 45, 673-678. 65. Rabinowe, S. L., Berger, M. J., Welch, W. R. et al. (1986). Lymphocyte dysfunction in autoimmune oophoritis. Resumption of menses with corticosteroids. Am. J. Obstet. Gynecol. 81, 348. 66. Edmonds, M., Lamki, L., Killinger, D. W., and Volp6, R. (1973). Autoimmune thyroiditis, adrenalitis, and oophoritis. Am. J. Med. 54, 782-787. 67. Pekonen, E, Siegberg, R., M~ikinen, T., Miettinen, A., and YliKorkala, O. (1986). Immunological disturbances in patients with premature ovarian failure. Clin. Endocrinol. (Oxford) 25, 1-6. 68. Walfish, P. G., Gottesman, I. S., Shewchuk, A. B., Bain, J., Hawe, B. S., and Fared, N. R. (1983). Association of premature ovarian failure with HLA antigens. Tissue Antigens 21, 168-169. 69. McNatty, K. P., Short, R. V., Barnes, E. W., and Irvine, W. J. (1975). The cytotoxic effect of serum from patients with Addison's disease and autoimmune ovarian failure on human granulosa cells in culture. Clin. Exp. Immunol. 22, 378-384. 70. Hoek, A., Shoemaker, J., and Drexhage, H. A. (1997). Premature ovarian failure and ovarian autoimmunity. Endocr. Rev. 18, 107134. 71. Muechler, E. K., Huang, K.-E., and Schenk, E. (1991). Autoimmunity in premature ovarian failure. Int. J. Fertil. 36, 99-103. 72. Kim, J. G., Anderson, B. E., Rebar, R. W., and LaBarbera, A. R. (1989). Determination by ELISA of antiovarian antibodies in premature ovarian failure. Program 45th Annu. Meet. Am. Fertil. Soc. San Francisco, 1989, Abst. O-100, p. 542. 73. Kim, M. H. (1974). "Gonadotropin-resistant ovaries" syndrome in association with secondary amenorrhea. Am. J. Obstet. Gynecol. 120, 257. 74. Chiauzzi, V., Cigorraga, S., Escobar, M. E., Rivarola, M. A., and Charreau, E. H. (1982). Inhibition of follicle-stimulating hormone receptor binding by circulating immunoglobulins. J. Clin. Endocrinol. 54, 1221-1228. 75. Anasti, J. N., Flack, M. R., Froelich, J., and Nelson, L. M. (1995). The use of human recombinant gonadotropin receptors to search for immunoglobulin G-mediated premature ovarian failure. J. Clin. Endocrinol. Metab. 80, 824-828. 76. Miller, M. E., and Chatten, J. (1967). Ovarian changes in ataxia telangiectasia. Acta Paediatr. Scand. 56, 559-561. 77. Rebar, R. W., Morandini, I. C., Erickson, G. E, and Petze, J. E. (1981). The hormonal basis of reproductive defects in athymic mice. I. Diminished gonadotropin concentrations in prepubertal females. Endocrinology (Baltimore) 108, 120-126. 78. Rebar, R. W., Morandini, I. C., Benirschke, K., and Petze, J. E. (1980). Reduced gonadotropins in athymic mice: Prevention by thymic transplantation. Endocrinology (Baltimore) 107, 2130-2132. 79. Healy, D. L., Bacher, J., and Hodgen, G. D. (1985). Thymic regulation
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92. 93. 94.
95.
96. 97.
of primate fetal ovarian-adrenal differentiation. Biol. Reprod. 32, 1127-1133. Rebar, R. W., Miyake, A., Low, T. L. K., and Goldstein, A. L. (1981). Thymosin stimulates secretion of luteinizing hormone-releasing factor. Science 214, 669-671. Weetman, A. P. (1995). Autoimmunity to steroid-producing cells and familial polyendocrine autoimmunity. Bailliere's Clin. Endocrinol. Metab. 9, 157-174. Arif, S., Vallian, S., Farazneh, F., Zanone, M. M., James, S. L., Pietropaolo, M., Hettiarcichihi, S., Vergani, D., Comway, G. S., and Peakman, M. (1996). Identification of 3 beta-hydroxysteroid dehydrogenase as novel target of steroid-producing cell autoantibodies: Association of autoantibodies with endocrine autoimmune disease. J. Clin. Endocrinol. Metab. 81, 4439-4445. Gulyas, B. J., Hodgen, G. D., Tullner, W. W., and Ross, G. T. (1977). Effects of fetal or maternal hypophysectomy on endocrine organs and body weight in infant rhesus monkeys (Macaca mulatta): With particular emphasis on oogenesis. Biol. Reprod. 16, 216-217. Axelrod, L., Neer, R. M., and Kliman, B. (1979). Hypogonadism in a male with immunologically active biologically inactive luteinizing hormone: An exception to a venerable rule. J. Clin. Endocrinol. Metab. 48, 279-287. Park, I. J., Burnett, L. S., Jones, H. W., Jr., Migeon, C. J., and Blizzard, R. M. (1976). A case of male pseudohermaphroditism associated with elevated LH, normal FSH and low testosterone possibly due to the secretion of an abnormal LH molecule. Acta Endocrinol. (Copenhagen) 83, 173-181. Silva de Sa, M. E, Matthews, M. J., and Rebar, R. W. (1988). Altered forms of immunoreactive urinary FSH and LH in premature ovarian failure. Infertility 11, 1-11. Whitney, E. A., Layman, L. C., Lanclos, K. D., Wall, S. W., and McDonough, E G. (1992). Polymerase chain reaction and southern analysis of the follicle-stimulating hormone receptor gene in women with 46,XX premature ovarian failure. Abstr., 74th Annu. Meet. Endocr. Soc., San Antonio, TX, 1992, Abstr. No. 903, p. 277. Sluss, E M., and Schneyer, A. L. (1992). Low molecular weight folliclestimulating hormone receptor binding inhibitor in sera from premature ovarian failure patients. J. Clin. Endocrinol. Metab. 74,1242-1246. Jones, G. S., and de Moraes-Ruehsen, M. (1969). A new syndrome of amenorrhea in association with hypergonadotropinism and apparently normal ovarian follicular apparatus. Am. J. Obstet. Gynecol. 104, 597-600. Metka, M., Holzer, G., Heytmanek, G., and Huber, J. (1992). Hypergonadotropic hypogonadic amenorrhea (World Health Organization III) and osteopenia. Fertil. Seril. 57, 37-41. Manuel, M., Katayama, K. E, and Jones, H. W., Jr. (1976). The age of occurrence of gonadal tumors in intersex patients with a Y chromosome. Am. J. Obstet. Gynecol. 124, 293-300. Schellhas, H. F. (1974). Malignant potential of the dysgenetic gonad. Part I. Obstet. Gynecol. 44, 298-309. Schellhas, H. F. (1974). Malignant potential of the dysgenetic gonad. Part II. Obstet. Gynecol. 44, 455-462. van der Schouw, Y. T., van der Graaf, Y., Steyerberg, E. W., Eijkemans, M. J., and Banga, J. D. (1996). Age at menopause as a risk factor for cardiovascular mortality. Lancet 347, 714-718. Snowden, D. A., Kane, R. L., Beeson, W. L., Burke, G. L., Sprafka, J. M., Potter, J., Iso, H., Jacobs, D. R., Jr., and Phillips, R. L. (1989). Is early menopause a biological marker of health and aging? Am. J. Public Health 79, 709-714. Alper, M. M., Jolly, E. E., and Garner, E B. (1986). Pregnancies after premature ovarian failure. Obstet. Gynecol. 67, 595-625. Check, J. H., and Chase, J. S. (1984). Ovulation induction in hypergonadotropic amenorrhea with estrogen and human menopausal gonadotropin therapy. Fertil. Steril. 42, 919-922.
1
4
6
R
O
B
98. Check, J. H., Chase, J. S., and Spence, M. (1989). Pregnancy in premature ovarian failure after therapy with oral contraceptives despite resistance to previous human menopausal gonadotropin therapy. Am. J. Obstet. Gynecol. 160, 114 - 115. 99. Check, J. H., Chase, J. S., Wu, C. H., and Adelson, H. G. (1989). Ovulation induction and pregnancy with an estrogen-gonadotropin stimulation technique in a menopausal woman with marked hypoplastic ovaries. Am. J. Obstet. Gynecol. 160, 405-406. 100. Check, J. H., Wu, C. H., and Check, M. (1988). The effect of leuprolide acetate in aiding induction of ovulation in hypergonadotropic hypogonadism: A case report. Fertil. Steril. 49, 542-543. 101. Ledger, W. L., Thomas, E. J., Browning, D., Lenton, E. A., and Cooke, I. D. (1989). Suppression of gonadotropin secretion does not reverse premature ovarian failure. J. Obstet. Gynecol. 96, 196-199. 102. Lutjen, P., Trounson, A., Leeton, J., Findlay, J., Wood, C., and Renou, P. (1984). The establishment and maintenance of pregnancy us-
E
103.
104.
105.
106.
R
T
W. REBAR
ing in vitro fertilization and embryo donation in a patient with primary ovarian failure. Nature (London) 307, 174-175. Chan, C. L. K., Cameron, I. T., Findlay, J. K., Healy, D., Lecton, J. F., Lutjen, P. J., Renou, P. M., Rogers, P. A., Trounsen, A. O., and Wood, E. C. (1987). Oocyte donation and in vitro fertilization: Clinical state of the art. Obstet. Gynecol. 42, 350-362. Sauer, M. V., and Paulson, R. J. (1989). Oocyte donation for women who have ovarian failure. Contemp. Obstet. Gynecol. November, pp. 125-135. Lydic, M. L., Liu, J. H., Rebar, R. W., Thomas, M. A., and Cedars, M. I. (1996). Success of donor oocyte IVF-ET in recipients with and without premature ovarian failure. Fertil. Steril. 65, 98-102. Rebar, R. W., and Cedars, M. I. (1994). Hypergonadotropic amenorrhea. In "Ovulation Induction: Basic Science and Clinical Advances" (M. Filicori and C. Flamigni, eds.), pp. 115-121. Elsevier, Amsterdam.
~HAPTER
Perimenopausal Changes in FSH, the lnhibins, and the Circulating Steroid Hormone Milieu HENRY G.
BURGER
Prince Henry's Institute of Medical Research, Clayton, Victoria 3168, Australia
I. Introduction and Definitions II. Current Concepts of Ovarian Physiology as a Function of Age III. Major Longitudinal Studies of Steroids in Relation to the Final Menstrual Period
IV. Hormonal Studies during the Menopausal Transition V. Conclusions References
strual cycles and have no symptoms of approaching menopause. Studies of the hormonal changes occurring during the perimenopause have been based on various experimental designs and definitions. In some instances hormonal changes have been recorded as a function of age with little attention paid to menstrual cycle status [4,5]. In the few longitudinal studies reported, the FMP has been used as a reference point with hormonal changes described in terms of time intervals before and after that point [6-9]. Very few studies have reported on hormone changes in relation to changes in menstrual cycle characteristics, such as the first self-reported change in the amount of menstrual flow, in the frequency of menstruation, or in the combination of changes in flow and frequency. This approach has been adopted in the Melbourne Women's Midlife Health Project [10,11 ], for which data will be provided below. The changes reported, particularly from longitudinal studies, in circulating concentrations of estradiol (E2) and
I. INTRODUCTION
AND DEFINITIONS The World Health Organization has defined the menopause as the permanent cessation of menstruation resulting from loss of ovarian follicular activity [1 ]. The perimenopause is defined as that period which commences when the first features of approaching menopause begin until at least 1 year after the final menstrual period (FMP). The term menopausal transition has been applied to that portion of the perimenopause which ends with the FMP [2]. The menopausal transition as studied in a group of North American women had a duration of approximately 4 years [3]. Thus the overall average duration of the perimenopause is 5 years. It is strongly recommended that the term perimenopause be used in this way, and not applied loosely to women in their forties and early fifties who continue to have regular menM E N O P A U S E : B I O L O G Y AND PATHOBIOLOGY
147
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
1
4
8
H
E
N
in relation to the occurrence of the FMP are examined in this chapter. The sparse data available on the circulating inhibins are reviewed and data from a longitudinal community-based study of women in the menopausal transition are presented. Menopausal status has been defined in terms of menstrual cycle status. Changes in progesterone and in androgens are reviewed briefly. estrone (El),
50-
R
Y
G. BURGER
II. CURRENT CONCEPTS OF OVARIAN PHYSIOLOGY AS A FUNCTION OF AGE Circulating concentrations of follicle-stimulating hormone (FSH), luteinizing hormone (LH), E2, progesterone (P) and the inhibins during the normal menstrual cycle have been described [12,13] (Figs. 1 and 2). The menstrual cycle
B
A
40~" 30:)
,,...,,
:z: 20..I 10-
' -8'
' -4
' -8
' ..~1 '
0
' -8
' -~1 '
0
. . 0. .
4
'
8'
' l~Z '
' ' ~I
'
-4' - '
~I
'
~I
'
II
' 1 '2
' ''
'
-4'-
'
0' 8 '
4'
0
4
'
II
' 1'2
302520:O "T 1 5 G)
u-10-
50i
i
i
'
,4
II'
12
;I'
12 '
i
' 41''' I
I ( 1- ' ' ' 44
' 1'2 '
-8
-4
12501000A
,.I
=0 E
750'
O.
500uJ
2500'
4
'
'-~'-~'+',;'
;,'1~'
'-~'~'
~)' ;,' ;3';2'
'-~'~',~',~'
a'1'2'
'-~'-~'
+' J,' +'1'2'
6045ii o
E 30c ,,.,,
o.
15O-
'.~'.~'+',~'
~ '1+'
Days from LH surge
FIGURE 1 Comparison of the daily geometric mean concentrations of LH, F S H , E 2 and P (with 68% confidence intervals) in 41 women aged 24-35 years (e, the control group, profiles repeated in each section) with (A) 19 women aged 3 6 - 4 0 years, (B) 18 women aged 41-45 years, and (C), 16 women aged 46-50 years (From Lee et al. [12], with permission of Oxford University Press and the authors.)
CHAPTER9 Perimenopausal Hormone Changes
149
a
r200
807 60 <
40
.c::
9~
20
1
0--
150
:::r
I +i lOO
T m
50
I
I
I
I
- - = ' - =I" 0
b
~" o E
40-
9
20-
r9
10-
13
0
o
c
._1
50-
1500
I~
500
I
----
T . . . . . .
I
I
I
40q 30"
0
-15
-10
i
E~
-+1 ~.
~~ "1"1
20
"1-
"
10 0
I
14
women were compared with a young control group showed increased levels of urinary estrogens and a fall in urinary pregnanediol excretion in the luteal phase of the cycle, despite the continuance of regular cyclicity [14]. The occurrence of unchanged or even elevated E 2 concentrations despite rising FSH is most plausibly explained by a decline in circulating INH concentration with age, as has been demonstrated in one study [15]. It is probable that the decline is particularly in INH-B [16]. Luteal-phase P concentrations showed no change with age in some studies and, as indicated, a decline in others. Follicular-phase LH concentrations rise only in the oldest group of regularly cycling women. Measurement of E 2 and INH in both ovarian veins of regularly cycling women has shown that E 2 concentrations are significantly higher in the vein draining the ovary containing the dominant follicle as compared to the contralateral ovarian vein, whereas INH concentrations are equivalent in both ovarian veins regardless of the side of the dominant follicle [ 17]. Although both E 2 and inhibins have been shown to be products of the ovarian granulosa cell, differential mechanisms appear to govern the secretion of the two inhibins. INH-B appears to be a product of the cohort of developing follicles, whereas INH-A together with E 2 is derived particularly from the dominant follicle [13,18]. It is postulated that the circulating concentration of INH-B may reflect the number of follicles recruited from the primordial pool, a number that decreases with increasing age [19]. The ability of the dominant follicles of older women to produce E 2 and INH-A appears to remain intact while regular cycles continue [ 12,16]. This has also been demonstrated in the E 2 response to ovarian hyperstimulation for the purposes of in vitro fertilization, with the response being age invariant [20]. This contrasts with the situation for production of total INH, which declines with increasing age in response to ovarian hyperstimulation [20]. In vitro studies of granulosa cells obtained at oocyte aspiration for the purposes of in vitro fertilization have also shown a diminished ability of granulosa cells from older women to produce INH-A in comparison with those from younger women [21]. Estradiol plays a central role in female reproductive function and has been described previously as the physiological basis of the fertile period [22]. Teleologically, it could be hypothesized that preservation of E 2 secretion would be desirable for as long as possible in the human female. Consequences of the loss of E 2 include the development of estrogen deficiency symptoms and undesirable health outcomes such as loss of bone and increased susceptibility to atherosclerosis and myocardial infarction. The decreased INH-B production in older women with elevated FSH concentrations in the follicular phase of the cycle can be postulated to increase secretion of FSH, which in turn provides increased drive to maintain E 2 secretion as the overall ovarian follicle number decreases.
-7 0 7 D a y s relative to midcycle LH p e a k
lz
FIGURE 2 Plasma concentrations of (a) inhibin A and inhibin B, (b) estradiol and progesterone, and (c) LH and FSH, during the female menstrual cycle. Data displayed with respect to the day of midcycle LH peak. Mean concentrations are shown ___SE. From [13], Groome, N. P., Illingworth, P. J., O'Brien, M., Rodger, P. A. L., Rodger, F. E., Mather, J. F., and McNeilly, A. S. (1996). Measurementof dimeric inhibin B throughout the human menstrual cycle. J. Clin. Endocrinol. Metab. 81, 1401-1405.
is characterized by relatively stable and low values for E 2 and inhibin A (INH-A) during the first half of the follicular phase, with a subsequent rise to a midcycle peak 24 hr prior to the midcycle ovulatory luteinizing hormone surge. After abrupt falls, there are secondary peaks of E 2 and INH-A secretion, parallel to that of P, during the luteal phase, with a subsequent fall leading to the onset of menses. In contrast, inhibin B (INH-B) concentrations rise and fall closely related to those of FSH during the early follicular phase, show a midcycle peak, and subsequently decline to their lowest points during the luteal phase (Fig. 2). The pattern of E 2 and INH-A is preserved when levels are examined as a function of increasing age in regularly cycling women. Thus in a large cross-sectional study of women ranging from 24 to 50 years of age [ 12], all cycling regularly, early follicular-phase concentrations of E 2 w e r e unchanged and in fact slightly higher in the oldest group of subjects, despite a progressive rise in circulating FSH (Fig.l). Another study in which older
150
HENRY G. BURGER
III. MAJOR LONGITUDINAL STUDIES OF STEROIDS IN RELATION TO THE FINAL MENSTRUAL PERIOD A small number of studies have examined the changes in circulating E 2 and E 1, and in some cases in androgens, in relation to FMP [6-9]. Such changes do not appear to be due to changes in steroid metabolism [23]. Most data relate to follicular-phase steroid hormone concentrations, but several have examined circulatory steroids, including P, during the luteal phase. The data reported by Rannevik et al. [9] are typical of the few published investigations of gonadotropin and estrogen around the FMR The study group consisted of 160 women from the Malm6 Perimenopausal Project. It had a 12-year duration and 152 of the 160 women were included over that time period. The mean age at the onset of menopause was 52.1 years. Estradiol concentrations were measured for 7 years prior to the FMP and remained relatively constant (means of 461-515 pmol/liter) until 6 months prior to the FMR In 154 observations 1 to 6 months prior to the FMP, mean E 2 was still 383 pmol/liter but had fallen to 182 pmol/liter 1 to 6 months after the FME A gradual further decline ensued to 171 pmol/liter in the 7 to 12 months after the FMP, with levels reaching 72 pmol/liter 97 to 108 months after that reference point. Concentrations of E~ behaved similarly, with a very small decrease in the 1 to 6 months before the FMP, a moderate decrease 1 to 6 months afterward (from 299 to 216 pmol/ liter), and a gradual decrease to 133 pmol/liter, 97 to 108 months after the FMP (Fig. 3). No data were provided on geometric mean levels in these subjects nor on the relationship between hormonal and self-reported menstrual cycle
1000-
i
-25
I I I
800-
i
I [
~~",~.~_~,~A
-20
...J 6 0 0 o
E
-15
-
_.1 :k
400-
-10
200_
0
'
-8
I'
-6
I ' l ' l '
-4
-2
0
MP
I'
i,
i,
i , l - 0
2
4
6
8
10
Years
FIGURE 3 Mean serum levels during the perimenopausal transition: e , estradiol (pmol/liter); 9 estrone (pmol/liter); &, FSH (/xg/liter); A, LH (/zg/liter). MR menopause. Reprinted from [9] Maturitas 21, G. Rannevik, S. Jeppsson, O. Johnell, B. Bjerre, Y. Laurell-Boruli, and L. Svanberg; A longitudinal Study of the perimenopausal transition: Altered profiles of steroid and pituitary hormones, SHBG and bone mineral density, 103-113, Copyright 1995, with permission from Elsevier Science.
status other than FMR It was, however, noted that during the premenopausal period there was an increasing frequency of inadequate luteal function, with low P concentrations. Similar data for the period after FMP were reported by Longcope et al. [7]. Data based on samples collected during the luteal phase were also broadly similar [24], with E 2 preserved until a few months before the onset of amenorrhea and then gradually falling thereafter, and with P levels falling progressively from 4 years prior to the onset of amenorrhea. From these data it appears that the FMP occurs during a time of relatively steeply falling E 2 concentrations, though substantial further falls occur for many months after that major selfreported endpoint.
IV. HORMONAL STUDIES DURING THE MENOPAUSAL TRANSITION A. FSH and Estradiol Relatively few studies have focused specifically on the endocrinology of the menopausal transition, when the most noteworthy characteristic is significant hormonal variability. A landmark study was that of Sherman and Korenman [25], who reported on 50 complete menstrual cycles in 37 women. Ten women aged 18 to 30 years, with a history of regular cycles, served as a control group. Six cycles were examined in regularly cycling women aged 46 to 51 years, in which it was noted that the follicular phase of the cycle was shorter than in the younger women and that E 2 concentrations during the early follicular phase were significantly lower than those observed in the younger women, but that serum FSH was strikingly increased throughout the cycle, despite the occurrence of E 2 concentrations that might have been expected to suppress its secretion. Daily hormone concentrations were also measured in 2 women, one aged 49 and one aged 50 who were clearly in the menopausal transition. Two of the cycles studied were anovulatory, but were nevertheless characterized by increasing E 2, and initially postmenopausal values of LH and FSH, which subsequently fell with the rise of E 2. An anovulatory cycle was followed by a cycle that demonstrated evidence of follicular maturation. Metcalf and colleagues [6,26] examined the excretion of FSH, LH, estrogens, and pregnanediol in weekly urine samples collected for 14 to 87 weeks from 31 perimenopausal women aged 36 to 55 years. Their study concentrated particularly on the gonadotropin changes, but wide fluctuations in estrogen excretion were noted. These authors stated "about the only conclusion that can be made with confidence concerning pituitary-ovarian function in individual perimenopausal women is that it is unsafe to generalize." In a more recent paper, Metcalf [27] concluded "in older women, a good menstrual history is probably the single most useful measure of ovarian status." Hee et al. [28] confirmed the
CHAPTER 9 Perimenopausal Hormone Changes
151
variability of perimenopausal E 2 concentrations and added data on INH in a small longitudinal study of three volunteer women who had developed irregular cycles at age 4 5 - 4 6 years. Abrupt decreases in E 2 and INH into the postmenopausal range were followed by values characteristic of reproductive-aged women.
100.
! ~ ~-~
~/~-,
U---
))
80o~" 60v OJ
uJ 4020_
0
B. Studies Including Inhibins The Melbourne Women's Midlife Health Project is based on a cross-sectional survey of a randomly selected population sample of 2001 Melbourne women, all Australian born and between 45 and 55 years at the time of their interview in May, 1991 [29]. A longitudinal study of 437 women was undertaken to examine many aspects of the menopausal transition. The data from the first year of this longitudinal study have been subjected to cross-sectional analysis in terms of menstrual cycle history. Of the subjects, 27% reported no change in menstrual frequency or flow, 23% reported a change in flow with no change in frequency, 9% reported a change in frequency without change in flow, 28% reported a change in both frequency and flow, and by the time of blood sampling 13% reported that at least 3 months had elapsed since their last menstrual period. Mean age increased from 48.5 years in the first group to 51.4 years in the last group. The data are shown in Fig. 4. Although unadjusted E 2 values were slightly lower in the groups experiencing a change in frequency or a change in frequency and flow (88 and 82% of those without any change), the only statistically significant decline in E 2 occurred in those who had no menses for at least 3 months, when the geometric m e a n E 2 concentration was 42% of that observed in the first group. When the E 2 data were adjusted for age and body mass index, the only significant change was again in the group with 3 months or more of amenorrhea, with the E 2 geometric mean being 54% of that of the Group 1 women. It must be emphasised that there was a broad spread of E 2 values, with some being > 1500 pmol/liter. Such high levels may reflect hyperstimulation of granulosa cells by elevated FSH levels, and could give rise to symptoms of breast fullness and fluid retention. Immunoreactive inhibin concentrations were significantly lower (71% of those in the first group) in those experiencing a change in frequency and flow and had fallen to 38% in those with 3 months or more of amenorrhea. Following adjustment for age and body mass index, only the change in the final group was significant, with the adjusted geometric mean being 53 % of that in Group 1. These data, particularly when examined without adjustment for age and body mass index, suggested that the decreases in inhibin levels occurred before decreases in E2, consistent with the hypothesis that declining concentration inhibin provides a mechanism for allowing FSH to rise, so as to maintain early follicular-phase
mm
100
ri~-~D
80-
g 6o -rzm
40-
U
200
-
0
~o 500E 400 "r" _o~300 t" :3
200
.E 1 0 0 or) _.o 0 ._o O IJ_
mm nn No change (n=104)
MH
C h a n g e Change Change flow freq. freq. & (n=91) (n=33) flow (n=108)
No period >3mths (n=52)
F I G U R E 4 Changes in E 2 (top), INH (middle), and FSH (bottom), as a percentage of Group 1 women from the Melbourne Midlife Project who had experienced no change in their cycles. In each pair, the left column is unadjusted and the right is adjusted for age and body mass index.
levels relatively intact. When thie data were analyzed without reference to menstrual cycle status and purely as a function of age, a marked decline in INH was observed, inverse to rising FSH levels, whereas E 2 w a s relatively constant until the age of approximately 51 or 52 years, when it too declined steeply. A further cross-sectional analysis of data was undertaken on 110 subjects aged 4 8 - 5 9 years in the third year of the longitudinal-phase Melbourne Women's Midlife Health Project [ 11 ] (Fig.5). Subjects were divided into those calledpremenopausal, with no reported change in menstrual cycle pattern; early perimenopausal, with a reported change in cycle frequency in the preceding year but a bleed in the preceding 3 months; late perimenopausal, with no menses in the preceding 3-11 months; and postmenopausal, with no menses for more than 12 months. The hormone concentrations in the premenopausal subjects were used as reference points for the other groups. Early perimenopausal subjects had significantly lower levels of INH-B (13.5 ng/liter compared with 48 ng/liter) in the presence of a small, statistically nonsignificant rise in FSH (21.4 compared with 13.5 E2
1
5
2
H
E
N
-
!
60
7"
_F-L_F_,. . . . . . . . . . . . . . . . .~ 240q
a ....
b
b
-
0
i
~,, 10
I
a
I
III
a
<
b
~ t_1_1
,~-r
.
.
.
.
.
.
.
.
.
.
i
.
c
!
60-]
!
.L
!
a
='5" 40 e-
~_ 20 _~
=
b
0
. . . .
~_~360-]
a
I
t:+t
b
b
. . . I. t 4t. . . .t. I t
t-
b
b
3
4
;
a
E 2401 o
~5 120 1 0
, . . . . . . . . . . . . .
1
G. BURGER
Interpretation of Estradiol and Inhibin Changes
a
~ 160 "~ 80
Y
fluctuate widely in individual women during the menopausal transition. Grouped data show that mean changes in hormone levels become significant around the FMP, with a decrease in INH-B concentration in early perimenopausal women being the most important and significant initial endocrine event at that time.
9O
~
R
,---l
2 Category
F I G U R E 5 Geometric mean levels (with lower 95% confidence intervals) of FSH, IR-INH, INH-A, INH-B, and E 2 as a function of menopausal status in a group of 110 women from the Melbourne Women's Midlife Health Project. Values with the same superscript (a, b, or c) not statistically different: values with differing superscripts are different, P < 0.05. IR, Immunoreactive. From Burger et al. [ 11 ], with permission of Blackwell Science Ltd.
IU/liter). There was no significant change in E 2 or INH-A. In late perimenopausal subjects, INH-A had fallen (geometric mean 4.2 ng/liter compared with 9.6 ng/liter), whereas INH-B was unchanged and FSH had risen signficantly to 72.2 IU/liter. E 2 also fell significantly to 89 pmol/liter (compared with 306 pmol/liter in the premenopausal group). The postmenopausal subjects showed no further significant changes in the peptide hormones or in FSH but E 2 fell further to 41 pmol/liter. A significant inverse correlation was noted between FSH and E 2, FSH and INH-A, and FSH and INH-B. The data were interpreted as consistent with negative feedback roles for both the dimeric inhibins and E 2 as contributors to the regulation of FSH secretion during the menopausal transition. Overall, circulating E 2 and inhibin concentrations may
Faddy and Gosden [30] have developed a mathematical model to describe the rates of growth and death of follicles in human ovaries in women between 19 and 50 years of age. Their study was based on the number of follicles at three successive stages of development, counted in histological sections of ovaries from 52 regularly cycling normal women. Their model predicts that the number of follicles growing from a stage at which there are at least two layers of granulosa cells surrounding an oocyte that has increased in size will decrease from 51 per day at age 2 4 - 2 5 years to only 1 per day at age 4 9 - 5 0 years. It could be hypothesized that circulating INH-B concentrations may provide an index of the numbers of those follicles progressing from that relatively early stage of development and that INH-B levels decline when there has been a substantial decline in numbers proceeding to maturation, late in reproductive life. In contrast, providing that a competent follicle is able to develop to dominance, or to a size sufficiently large to maintain production of steroids, E 2 and INH-A secretion would be expected to be relatively preserved. It has also been shown that the ability of granulosa cells from the follicles of older women to respond to human chorionic gonadotropin (hCG) stimulation by secretion of INH declines markedly, as does their secretion of P [31 ]. In addition, as noted above, the granulosa cells from older women in fact secrete lower amounts of INH-A into a culture medium than do those from younger women, though whether such secretory properties of granulosa cells obtained after ovarian hyperstimulation reflect the secretory abilities of single dominant follicles in older women could be questioned [21]. Thus it is hypothesized that declining numbers of follicles with declining INH-B levels lead to a reciprocal rise in FSH, which in turn stimulates the rate of follicular development, and maintains the capacity of the ovary to develop a dominant follicle until late in reproductive life. Preservation of that capacity results in preservation of circulating E 2 and INH-A within the normal range. The preserved E 2 is postulated to maintain quality of life and bony and vascular health. Validation of this hypothesis will require measurements of dimeric inhibin as a function of aging during the menopause transition, and comparison of changes in its levels with those of E 2, studies that have been reported [ 11 ] or are in progress. Ultrasound monitoring of the ovary would be necessary to assess follicular growth. Differential measurements of INH-A and INH-B have shed further light on pituitary-ovarian relationships during this phase of waning reproductive function.
CHAPTER 9 Perimenopausal Hormone Changes
153
C. Progesterone
75 e-v
It is well known, from studies in which basal body temperature has been used as a marker of ovulatory function, that anovulatory cycles become more prevalent as a function of increasing a g e m 3 - 7 % of cycles were found to be anovulatory between ages 26 and 40 years, compared to 1215% between 41 and 50 years [32]. A large study of lutealphase P concentrations [24] noted that the frequency of nondetectable P gradually increased as the FMP approached. Interestingly, these authors noted that in 11.5 % of their cases the endometrium was found to be secretory, whereas P concentrations were below 3 nmol. This discordance was seen most frequently in the period from 3 to 1 years before the FMP. The lack of a luteal-phase rise in P is a striking feature of the postmenopause compared with the reproductive period. Rannevik et al. [9] reported that the frequency of cycles with P values indicative of ovulation (concentrations > 10 nmol/liter) decreased from about 60% to less than 10% during the 6 years preceding the FMP. Ovulatory P concentrations were found in 62.2% of women 72 to 61 months premenopausal, and in 4.8% who were 6 to 0 months premenopausal, whereas all serum P measurements were less than 2 nmol/liter postmenopausally. There is some controversy regarding the maintenance of P secretion during the luteal phase in older regularly cycling women. Lee et al. [ 12] showed that P secretion was well preserved in a group of regularly cycling women aged 46 to 50 years, whereas Santoro et al. [14] showed decreased urinary pregnanediol excretion in a group of regularly cycling women aged 43 to 52 years, compared with women aged 19 to 38 years.
D. A n d r o g e n s Variable findings have been reported in regard to the changes in circulating androgens in relation to the FMP. Rannevik et al. [9] reported a small but significant decline in testosterone (T), androstenedione (A), and sex hormone binding globulin (SHBG) during the 2 years around the menopause. Thus T fell from 1.7 nmol 1 to 6 months before the FMP to 1.4 nmol 13 to 24 months afterward and 1.2 nmol 85 to 96 months afterward. SHBG fell from 4.0 mg/liter 1 to 6 months before the FMP to 3.5 mg/liter 85 to 96 months afterward but the ratio T/SHBG was unchanged over that period. The data for A were not specifically listed. Longcope et al. [7] did not see any change in T and A over 80 months from the FMP but noted that the mean concentrations of T in all their subjects, including those still having cyclic menses, were significantly less than those of a group of normal young women sampled on days 5 to 7 of the cycle, and suggested that there is a decrease in the ovarian secretion of T prior to the menopause. It is noteworthy that a recent report
tO L
9
9
~0 5 0 E 25t-"
9
E
|
9
L_
~1"
/
0
20
25
I
I
I
I
I
30
35
40
45
50
Age (years)
FIGURE 6 Plasma total testosterone (24-hr mean) plotted against age, in normal females. The regression equation was testosterone (nanomoles/ liter) = 37.8 • age (years) - 1.12 (r = -0.54; P < 0.003). From [33], Zumoff, B., Strain, G. W., Miller, L. K., and Rosner, W. (1995). Twentyfour hour mean plasma testosterone concentration declines with age in normal premenopausal women. J. Clin. Endocrinol. Metab. 80,1429-1430. 9 The Endocrine Society.
[33] found that there was a steep decline in total serum T concentration with age, such that the levels in a woman aged 40 were approximately 50% of those in a woman aged 21 (0.61 nmol compared with 1.3 nmol) (Fig. 6). Percentage of free T did not vary significantly with age but free T concentration clearly showed a steep decline. The ratio of dehydroepiandrosterone (DHEA) to T and dehydroepiandrosterone sulfate (DHEAS) to T were age invariant because of the declines of DHEA and DHEAS with age. Other studies have suggested that total T levels decrease by approximately 20% and A decreases by approximately 50% with natural menopause [34]. Vermeulen [35] showed that postmenopausal women aged 51 to 65 years had lower mean levels of T (1.03 nmol), A (3.45 nmol), and dihydrotestosterone (DHT) (0.33 nmol) in comparison with women aged 18 to 25 years, i.e., with T (1.53 nmol), A (5.80 nmol), and DHT (1.04 nmol). The effects of ovariectomy on androgen profiles were reported by Judd et al. [36] and Hughes et al. [37]. Before the menopause, oophorectomy results in a decrease of circulating A and T by about 50%, the decrease in the latter being due in large part to the decrease in A. Postmenopausally, removal of the ovaries results in a 50% decline in T and a much lesser decline in A. The postmenopausal ovary secretes more T but less A than its premenopausal counterpart. [34]. In light of the recent report of Zumoff et al. [33], and the difficulty in demonstrating a significant decline in T around the FMP, it may be that the apparent decline in T at the menopause is related as much to aging as to decreased ovarian function in those women with intact ovaries. In the Melbourne Women's Midlife Health Project, there was no significant change seen in total T or in the T/SHBG ratio as a function of changing menopausal status [ 10].
154
HENRY G. BURGER
V. C O N C L U S I O N S 7.
The perimenopause is a time of markedly fluctuating hormone levels. Attempts to define menopausal status purely on the basis of single measurements of FSH or E 2 are unlikely to yield useful information. Though E 2 concentrations appear to be preserved in regularly cycling women at least until the age of 50, INH-B declines and FSH rises. The establishment of menstrual irregularity is marked by a decrease in the follicular-phase concentrations of INH-B, an increase in FSH, but relative preservation of E 2 and INH-A until the time of the FMP. The frequency of anovulatory cycles increases markedly as the FMP approaches. It is difficult to demonstrate substantial changes in androgen concentrations in the immediate perimenopausal period, though levels postmenopausally appear to be lower than those of young regularly cycling women, perhaps as a function of increasing age rather than menopausal status. Hormonal measurements are of little diagnostic value during the perimenopause other than for the purposes of physiological study. The issue of the appropriate reference points for the study of the perimenopause remains unclear.
8.
9.
10.
11.
12.
13.
14.
Acknowledgments 15. The collaboration of my colleagues in the Melbourne Women's Midlife Health Project (Lorraine Dennerstein, Emma Dudley, John Hopper, Adele Green, John Wark, Peter Ebeling, and Janet Guthrie) is acknowledged. David Robertson and his staff at Prince Henry's Institute of Medical Research provided the INH-A and INH-B assays for which Nigel Groome, Oxford Brookes University, Oxford, UK, provided the reagents. Mr. N. Balazs and his staff in the Department of Chemical Pathology, Monash Medical Centre, provided the FSH and estradiol measurements. The Melbourne Women's Midlife Health Project is supported by grants from the Victorian Health Promotion Foundation and the Public Health Research and Development Committee of the Australian National Health and Medical Research Council. Support for the hormone assays has also been provided by Organon Australia Pty Ltd.
References 1. World Health Organization (1981). "Research on the Menopause. Report of a WHO Scientific Group," Tech. Rep. Ser. 670. WHO, Geneva. 2. World Health Organization (1996). "Research on the Menopause in the 1990's," Tech. Rep. Ser. 866. WHO, Geneva. 3. McKinlay, S. M., Brambilla, D. J., and Posner, J. G. (1992). The normal menopause transition. Maturitas 14, 103-115. 4. Sherman, B. M., West, J. H., and Korenman, S. G. (1976). The menopausal transition: Analysis of LH, FSH, estradiol, and progesterone concentrations during menstrual cycles of older women. J. Clin. Endocrinol. Metab. 42, 629-636. 5. Velasco, E., Malacara, J. M., Cervantes, E, Diaz de Le6n, J., Divalos, G., and Castillo, J. (1990). Gonadotropins and prolactin serum levels during the perimenopausal period: Correlation with diverse factors. Fertil. Steril. 53, 56-60. 6. Metcalf, M. G., Donald, R. A., and Livesey, J. H. (1981). Pituitary-
16.
17.
18.
19.
20.
21.
22.
23.
ovarian function in normal women during the menopause transition. Clin. Endocrinol. 14, 245-255. Longcope, C., Franz, C., Morello, C., Baker, R., and Conrad-Johnston, C., Jr. (1986). Steroid and gonadotropin levels in women during the perimenopausal years. Maturitas 8, 189-196. Rannevik, G., Caristr6m, K., Jeppsson, S., Bjerre, B., and Svanberg, L. (1986). A prospective long-term study in women from premenopause to postmenopause: Changing profiles of gonadotrophins, oestrogens and androgens. Maturitas 8, 297-307. Rannevik, G., Jeppsson, S., Johnell, 0., Bjerre, B., Laurell-Boruli, Y., and Svanberg, L. (1995). A longitudinal study of the perimenopausal transition: Altered profiles of steroid and pituitary hormones, SHBG and bone mineral density. Maturitas 21, 103-113. Burger, H. G., Dudley, E. C., Hopper, J. L., Shelley, J. M., Green, A., Smith, A., Dennerstein, L., and Morse, C. (1995). The endocrinology of the menopausal transition: A cross-sectional study of a populationbased sample. J. Clin. Endocrinol. Metab. 80, 3537-3545. Burger, H. G., Cahir, N., Robertson, D. M., Groome, N. P., Green, A., and Dennerstein, L. (1998). Serum inhibins A and B fall differentially as FSH rises in perimenopausal women. Clin. Endocrinol. 48, 809-813. Lee, S. J., Lenton, E. A., Sexton, L., and Cooke, I. D. (1988). The effect of age on the cyclical patterns of plasma LH, FSH, oestradiol and progesterone in women with regular menstrual cycles. Hum. Reprod. 3, 851-855. Groome, N. P., Illingworth, P. J., O' Brien, M., Rodger, P. A. L., Rodger, E E., Mather, J. E, and McNeilly, A. S. (1996). Measurement of dimeric inhibin B throughout the human menstrual cycle. J. Clin. Endocrinol. Metab. 81, 1401-1405. Santoro, N., Brown, J. R., Adel, T., and Skurnick, J. H. (1996). Characterization of reproductive hormonal dynamics in the perimenopause. J. Clin. Endocrinol. Metab. 81, 1495-1501. MacNaughton, J., Bangah, M., McCloud, P., Hee, J., and Burger, H. (1992). Age related changes in follicle stimulating hormone, luteinizing hormone, oestradiol and immunoreactive inhibin in women of reproductive age. Clin. Endocrinol. 36, 339-345. Klein, N. A., Illingworth, P. J., Groome, N. P., McNeilly, A. S., Battaglia, D. E., and Soules, M. R.(1996). Decreased inhibin B secretion is associated with the monotropic rise of FSH in older, ovulatory women: A study of serum and follicular fluid levels of dimeric inhibin A and B in spontaneous menstrual cycles. J. Clin. Endocrinol. Metab. 81, 27422745. Illingworth, P. J., Reddi, K., Smith, K. B., and Baird, D. T. (1991). The source of inhibin secretion during the human menstrual cycle. J. Clin. Endocrinol. Metab. 73, 667-673. Roberts, V. J., Barth, S., EI-Roeiy, A., and Yen, S. S. C. (1993). Expression of inhibin/activin subunits and follistatin messenger ribonucleic acids and proteins in ovarian follicles and the corpus luteum during the human menstrual cycle. J. Clin. Endocrinol. Metab. 77, 1402-1410. Richardson, S. J., Senikas, V., and Nelson, J. E (1987). Follicular depletion during the menopausal transition: Evidence for accelerated loss and ultimate exhaustion. J. Clin. Endocrinol. Metab. 65, 1231-1237. Hughes, E. G., Robertson, D. M., Handelsman, D. J., Haywood, S., Healy, D. I., and de Kretser, D. M. (1990). lnhibin and estradiol responses to ovarian hyperstimulation: Effects of age and predictive value for in vitro fertilization outcome. J. Clin. Endocrinol. Metab. 70, 358 -364. Seifer, D. B., Gardiner, A. C., Lambert-Messerlian, G., and Schneyer, A. L. (1996). Differential secretion of dimeric inhibin in cultured luteinized granulosa cells as a function of ovarian reserve. J. Clin. Endocrinol. Metab. 81,736-739. Burger, H. G. (1984). The physiological basis of the fertile period. In "Fertility and Sterility," R. F. Harrison and B. W. Thompson, eds., pp. 51-8. MTP Press, Lancaster, England. Longcope, C. (1990). Hormone dynamics at the menopause. Ann. N.Y. Acad. Sci. 592, 21-30.
155
CHAPTER 9 P e r i m e n o p a u s a l H o r m o n e Changes 24. Trevoux, R., De Brux, J., Castanier, M., Nahoul, K, Soule, J.-R, and Scholler, R. (1986). Endometrium and plasma hormone profile in the peri-menopause and postmenopause. Maturitas 8, 309-26. 25. Sherman, B. M., and Korenman, S. G. (1975). Hormonal characteristics of the human menstrual cycle throughout reproductive life. J. Clin. Invest. 55, 699-706. 26. Metcalf, M. G., and Donald, R. A. (1979). Fluctuating ovarian function in a perimenopausal woman. N.Z. Med. J. 89, 45-47. 27. Metcalf, M. G. (1988). The approach of menopause: A New Zealand study. N.Z. Med. J. 101, 103-106. 28. Hee, J., MacNaughton, J., Bangah, M., and Burger, H. G. (1993). Perimenopausal patterns of gonadotrophins, immunoreative inhibin, oestradiol and progesterone. Maturitas 18, 9-20. 29. Dennerstein, L., Smith, A.M., Morse, C., Burger, H. G., Green, A., Hopper, J., and Ryan, M. (1993). Menopausal symptoms in Australian women. Med. J. Aust. 259, 232-236. 30. Faddy, M. J., and Gosden, R. G. (1995). A mathematical model of follicle dynamics in the human ovary. Hum. Reprod. 10, 770-775. 31. Pellicer, A., Mari, M., de los Santos, M. J., Sim6n, C., Remohi, J., and Tarin, J. J. (1994). Effects of aging on the human ovary: The secre-
32. 33.
34. 35. 36.
37.
tion of immunoreactive (a-inhibin and progesterone. Fertil. Steril. 61, 663-668. Doring, G. K. (1969). The incidence of anovular cycles in women. J. Reprod. Fertil. 6, 77-81. Zumoff, B., Strain, G. W., Miller, L. K., and Rosner, W. (1995). Twenty-four hour mean plasma testosterone concentration declines with age in normal premenopausal women. J. Clin. Endocrinol. Metab. 80, 1429-30. Judd, H. L. (1976). Hormonal dynamics associated with the menopause. Clin. Obstet. Gynecol. 19, 775-788. Vermeulen, A. (1976). The hormonal activity of the postmenopausal ovary. J. Clin. Endocrinol. Metab. 42, 247-253. Judd, H. L., Lucas, W. E., and Yen, S. S. (1974). Effect of oophorectomy on circulating testosterone and androstenedione levels in patients with endometrial cancer. Am. J. Obstet. Gynecol. 118, 7 9 3 798. Hughes, C. L. Jr., Wall, L. L., and Creasman, W. R. (1991). Reproductive hormone levels in gynaecologic oncology patients undergoing surgical castration after spontaneous menopause. Gynecol. Oncol. 40, 42-45.
7 H A P T E R 1(
Epidemiology." Methodologic Challenges in the Study of Menopause SYBIL L.
I. II. III. IV.
CRAWFORD
New England Research Institutes, Watertown, Massachusetts 02472
Introduction Study Design Data Collection Methods Measurement Issues
V. Analytic Considerations
VI. Conclusion References
I. I N T R O D U C T I O N
care patterns, and a subject's own perceptions of menopause. Statistical data analyses in both cross-sectional and longitudinal studies often require handling issues such as biases of recall and selection and censored and missing data, and complicated data structures such as daily symptoms or reproductive hormone levels. Methodologic challenges in the study of menopause are discussed in this chapter. Topics covered include study design issues for both observational studies and clinical trials of hormone replacement therapy; types of data collection instruments; measurement issues, particularly ascertainment of menopause status; and analytic concerns, including choice of appropriate statistical techniques for different types of data.
The menopause is a complex and multifaceted phenomenon, one that is very challenging to study for a number of reasons. First, study design can be logistically and scientifically demanding. Community-based samples are critical for an examination of menopause in the general population, as opposed to nonrepresentative clinic-based samples, but the former are much more costly to obtain. Moreover, choice of the length of followup and eligibility criteria are complicated by variation within and across women in experiences of menopause transitions. In addition, the menopause involves multiple domains, including physiologic as well as psychologic and lifestyle changes. This has implications for selection of data collection methods, in order to obtain a complete picture. A number of measurement issues also arise in a study of menopause, including how to define menopause status. Many measures, particularly those derived from self-report, are affected by physiologic changes as well as by sociodemographic characteristics, psychologic factors, culture, healthMENOPAUSE: BIOLOGY AND PATHOBIOLOGY
II. STUDY DESIGN A number of issues involved in the design of studies of menopause are summarized in this section. The first topic considered is the choice of a sampling flame and the importance of a population-based sample. Issues relevant to 159
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
160
SYBIL L. CRAWFORD
observational studies include cross-sectional versus longitudinal design, choice of age range, and eligibility criteria for menopause status. For examining the role of hormone replacement therapy, the use of observational studies versus clinical trials is discussed, and eligibility criteria used in several large clinical trials are presented.
A. Choice of Sampling Frame Many early studies of menopause relied on clinic-based samples [1], such as patients at menopause clinics. Such samples are relatively cost-efficient to obtain and response rates are likely to be relatively high. These advantages are greatly offset, however, by the atypical nature of such groups. Women seeking health care, particularly those seeking treatment for menopausal symptoms, differ in a number of ways from other menopausal-aged women, with respect to characteristics likely to be related to menopausal and healthrelated factors under study. For example, they have greater access to health care and are more worried about their health, are more likely to report menopausal symptoms, are more likely to undergo a surgical menopause, are more likely to receive psychiatric treatment, and tend to experience more long-term ill health preceding menopause [2-9]. In short, a sample of patients, particularly those from a menopause clinic, is unlikely to provide adequate representative data on the experience of menopause in the general population. Moreover, the range of variability in characteristics of interest may be restricted, affecting the ability to detect associations with outcomes [10]. Thus, it is critical to sample from a frame including women in the general community. A related consideration is sufficient representation from traditionally understudied subgroups, particularly women of lower socioeconomic status (SES) and races and ethnicities other than non-Hispanic Caucasians [4,11-14]. Compared with non-Hispanic Caucasian women, particularly those with high levels of SES indicators such as education, we know relatively little about the menopausal experience for these women. Sampling frames used in past and current communitybased studies include census lists [11,15], driver's license lists [ 16], health maintenance organization patient lists [ 11 ], utility lists [ 11 ], telephone directories [ 17], and random-digit dialing [ 11 ]. Formal statistical generalization to the population of interest requires use of probability sampling from a sampling frame in which all members of the target population have a known nonzero probability of being sampled [18]. In practice, this may be costly or logistically difficult to achieve, particularly when attempting to sample hard-tofind subgroups less likely to be included on sampling frames. Studies may need to combine information from multiple sampling frames, as was done by several field sites involved in the Study of women's Health Across the Nation (SWAN) [ 11].
Obtaining sufficient numbers of racial/ethnic minorities or low-SES women may be particularly challenging, and may require techniques such as household enumeration, or "snowball" sampling in which potential subjects are referred to the study by already-identified subjects, both of which were employed by SWAN field sites [11]. Note that computation of sampling probabilities under this latter approach can be difficult or impossible, because the probability that a woman is "sampled" as a snowball is typically not known [ 11 ]. In turn, this affects the ability to use the sample to generalize to the population of interest. Thus, there can be trade offs between obtaining a probability sample and sampling sufficient numbers from hard-to-find subgroups such as low-SES subjects.
B. Issues in Observational Studies of Menopause This section considers study design issues relevant for observational studies of menopause, where the natural history of menopausal transitions m including surgical menopausem will be observed. I. CROSS-SECTIONAL VERSUS LONGITUDINAL DESIGN
This issue is particularly important for studies of menopause, due to the relatively lengthy duration of the entire menopausal experience for an individual woman. In a crosssectional study of menopause, one cannot observe withinsubject changes in health occurring concurrently with withinsubject menopause transitions. Thus, inferences regarding the role of menopause are based on between-woman comparisons, e.g., comparing age-matched pre- and postmenopausal women [ 19]. Such analyses assume that cross-sectional estimates of menopause status or reproductive age are equivalent to those estimated from longitudinal observations, an assumption that is not always correct [12,20-25]. In addition, retrospective data are more subject to recall error [26]. Accurate assessment of temporal sequences also is much more difficult with cross-sectional data, e.g., determining whether attitudes prior to the menopause transition predict subsequent menopausal experiences [2,24,27], and again requires assumptions regarding the applicability of between-woman differences. Choice of cross-sectional or longitudinal design, however, often is determined in large part by considerations of cost. The followup time necessary to capture the full menopausal period, from pre- to peri- to postmenopausal (definitions of these terms are presented in Section III), is relatively long. In Caucasians, the median time elapsed between the onset of perimenopause and the final menstrual period (FMP) has been estimated as 3.8 years [27]. Moreover, note that this interval does not include the length of time that a woman is in the study prior to reaching perimenopause. Thus a cross-sectional study can be a cost-efficient way to exam-
CHAPTER 10 Methodologic Challenges in Menopause Studies ine both menopausal transitions, pre- to perimenopause and peri- to postmenopause, in a short time. However, interpretation of results is subject to the caveat noted above regarding use of between-woman comparisons to make inferences regarding within-woman changes. Note that a median length of perimenopause of 3.8 years implies that a large percentage of women will transition from peri- to postmenopause in under 4 years. Length of this transition, however, is not entirely random but is associated with a number of characteristics, including smoking [27], which affects a number of important variables (e.g., cardiovascular disease risk factors), as well as the age at onset of perimenopause [27]. Women with a shorter perimenopause also are less likely to report menopausal symptoms or to be characterized as depressed [2,23]. Thus, care must be used when generalizing from women with a short perimenopause to the full population. An important advantage of following the same individuals through the entire menopausal periodmfrom pre- to peri- to postmenopause--is the enhanced ability to assess the presence of curvilinear (nonlinear) relationships between health outcomes and menopause status or reproductive age, measured as time before/after the final menstrual period [4,22,28]. For example, we can examine whether within-subject bone loss accelerates around the onset of perimenopause, and decelerates after FMP. Such analyses are more complicated and require assumptions when performed on cross-sectional data. 2. CHOICE OF AGE RANGE
A common goal of menopause studies is to distinguish the roles of menopause or reproductive age and chronologic age. Hence, the age range needs to be selected accordingly, with women observed to experience menopause transitions at different chronologic ages in order to reduce confounding between reproductive and chronologic age. The appropriate age range depends in part on whether the study is crosssectional or longitudinal; in general, the age range should be broader for a cross-sectional than for a longitudinal study, since the latter involves following subjects as they age, thereby widening the effective observed age range. For a cross-sectional study of natural menopause, the age range should include the full set of ages at which women typically become menopausal [26], approximately 40 years through 55 years, as was done in SWAN's cross-sectional phase [ 11]. A study of surgical menopause may need to include women younger than 40 years, as many women have a hysterectomy prior to age 40, particularly for diagnoses such as endometriosis [29]. African-American women also tend to undergo hysterectomy at a younger age than do Caucasians [29]. The age range should not be so large, however, as to involve cohort effects, where younger women are different from older women, e.g., with respect to characteristics such as use of oral contraceptives [28]. Another consideration for choice of age range is the
161 prevalence of smoking in the population under study. Smokers tend to experience menopause earlier than do nonsmokers, by 1-2 years on average [10,15,30,31]. Investigators may want to consider a lower age range for smokers than for nonsmokers, in order to capture the pre- to perimenopause transition in both groups. For a longitudinal study of initially premenopausal women, it is important to consider an upper age limit for recruitment. Later menopause has been linked to a number of observed characteristics, most importantly lower rates of smoking, and possibly higher body mass index [30,32,33], both of which are associated with key outcomes such as cardiovascular disease risk and bone density [34,35]. Thus, older women who are still premenopausal are atypical of premenopausal women in general, exhibiting "survivorship" bias [4]. In addition, these women tend to have a shorter perimenopause [27] and as a consequence may experience fewer menopausal symptoms [2,23]. In short, data from older premenopausal women may not be generalizable to the population at large. Note that this group's menopausal experience is important scientifically. A longitudinal study, however, will collect relevant data as women age into this group. For a longitudinal study with followup of subjects of under 4 - 5 years, where the full menopausal period may not be observed, choice of age range is determined in part by the stage of menopause of greatest interest. For example, a study focusing on the transition from pre- to perimenopause should sample primarily younger women, e.g., an age range centered around 4 7 - 4 8 years, the estimated median age for onset of perimenopause in Caucasians [27]. In contrast, a study of postmenopausal women should include primarily older women, e.g., aged 5 0 - 7 9 years, as in the Women's Health Initiative [36]. 3. MENOPAUSE STATUS ELIGIBILITY CRITERIA As with age range, the appropriate choice of eligibility criteria involving menopause status depends in part on whether the study is cross-sectional or longitudinal. For a cross-sectional design, one may want to be more inclusive, sampling women in a variety of menopause stages~including surgical menopause--in order to include data from all phases of menopause in the study. In fact, including only older postmenopausal women can be problematic when the outcome is age at menopause, due to recall bias [4,26]. For longitudinal studies, within-woman changes in menopause status or reproductive aging are observed directly, so that eligibility criteria regarding initial menopause status can be more restrictive than in a cross-sectional study. The aims of a study also affect the choice of status categories or reproductive age. Choice of status eligibility criteria which are appropriate for one set of scientific goals may not be adequate for others. For example, the Postmenopausal Estrogen-Progestin Intervention Study (PEPI) had as its primary goal to compare the impact of different hormone
162
SYBIL L. CRAWVORD
replacement therapy regimes on measures of cardiovascular disease risk [37,38]. To this end, the design employed sampling restrictions in terms of chronologic age and years since menopause (natural or surgical). A consequence of this design was an artificial collinearity between these two variables, i.e., a collinearity that does not exist in the general population, where age and years since menopause are not restricted. Thus, analyses attempting to identify predictors of age at menopause (not an original study goal) were hampered by the inability to distinguish covariates of age at menopausemthe outcome of interestmfrom correlates of chronologic age [38]. As this example demonstrates, careful thought should be given to potentially competing study aims when selecting eligibility criteria. Past studies of menopause vary with respect to inclusion or exclusion of surgically menopausal subjects. These women differ in a number of ways from women experiencing natural menopause, including better access to health care, and lower prior levels of health [4,24,39]. Their menopausal experiences and subsequent disease risk also are likely to diverge from those of naturally menopausal women, in part because of different characteristics prior to surgery, but also because their reproductive hormonal profile~e.g., rapidity and timing of changes in hormone levels w differs as well [24,40]. Thus, one cannot generalize results regarding menopause and health outcomes, such as cardiovascular disease risk factors or menopausal symptoms, from surgically menopausal women to naturally menopausal women. Because surgically menopausal women are not a random subsample of all women going through menopause, however, omitting them from analyses implies that the resulting description of the menopausal experience for the general population is in some sense incomplete. A common recommendation is to include these w o m e n ~ i n cross-sectional studies, include women with a prevalent hysterectomy or bilateral oophorectomy, and in longitudinal studies, continue to follow women with an incident surgical menopausembut study them separately from other women. It may be useful to consider surgically menopausal women as a separate stratum, both for sampling and for data analyses [4,22,36,37].
C. Issues in S t u d i e s o f H o r m o n e Replacement Therapy This section compares the use of observational studies and clinical trials in the study of hormone replacement therapy (HRT), and summarizes eligibility criteria used in several recent trials. 1. OBSERVATIONAL STUDIES VERSUS CLINICAL TRIALS
Clinical trials are preferable to observational studies of this topic due to selection bias [4,41-43]. Many studies have
found that prior to initiation of HRT, users are more likely to exhibit characteristics associated with better health, including higher SES (as indicated by income and education), higher use of health care, more exercise, lower body mass index, and a better risk profile for cardiovascular disease [41,44-49]. Thus part of the difference between users and nonusers in outcomes such as cardiovascular disease risk factors in observational studies is likely due to preexisting differences in health and related characteristics [50]. Comparing women prescribed and not prescribed HRT, many past studies took place when physicians were reluctant to prescribe HRT for women at high risk for cardiovascular disease (CVD) [47], reflected in higher observed CVD risk among nonusers. Even considering only ever-users, women who continue to use HRT differ from those prescribed HRT but who discontinue use. The former group is more tolerant of HRT and is less likely to experience adverse effects [4,51]. In fact, short-term users have been found to have greater subsequent cardiovascular disease risk compared to long-term users [52]. Sturgeon and colleagues also noted a "healthy user survivor effect," whereby women who developed an illness discontinued use of HRT [53]. In addition, long-term HRT users are by definition "compliant." In clinical trials, compliers on both study arms have been found to fare better than noncompliers, and the magnitude of the effect of compliance with placebo was similar to that found for HRT in two meta-analyses [54,55]. Thus, part of the difference in health outcomes or disease risk between users and nonusers may be due to a compliance effect. Moreover, it is difficult to control completely for all such biases. For example, it is not straightforward to adjust appropriately for education and SES when modeling health outcomes [4,43,45,46]. Even in a population that was relatively homogeneous with respect to SES, users differed from nonusers regarding behaviors affecting health promotion and disease prevention [41 ]. In addition, some biases involved in prescribing and compliance may not be observed or known [41,43,56]. In summary, estimates of the benefit of HRT obtained from observational studies are likely to be somewhat overstated [43,53,56]. To assess the impact of HRT on outcomes such as cardiovascular disease risk or events, it is important to look to results from clinical trials such as PEPI, the Women's Health Initiative, or the Heart and Estrogen/Progestin Replacement Study (HERS) [36,42,57]. Such trials avoid selection bias by employing random assignment to treatment arm, including a placebo arm [57].
2. ELIGIBILITY CRITERIA IN CLINICAL TRIALS o r H R T In the final stage of sample selection, both PEPI and HERS included only subjects with a high likelihood of complying with treatment arm, by requiring 80+ % compli-
CHAPTER 10 Methodologic Challenges in Menopause Studies ance during a run-in phase [57,58]. Current HRT users recruited to PEPI were required to stop treatment [58]. The trials also excluded women with contraindications for use of HRT [57,58]. Both restrictions are justified on logistical and ethical grounds. They may, however, limit generalizability of results somewhat to compliant subjects who are potentially able to be long-term HRT users. Moreover, short-term (3 months) cessation of HRT may be insufficient to achieve "wash-out" of its effects, e.g., on rate of bone turnover. McKinlay suggests that the most appropriate subjects for a clinical trial of HRT are those with no prior use [4]. Current users are less likely to report adverse effects for the HRT arm (because they have been taking HRT, and hence they can tolerate it), and are more likely to be unblinded on the placebo arm, whereas past u s e r s m w h o may have discontinued use due to adverse effects m may be less likely to be able to tolerate the HRT arm. Residual effects of HRT also may be an issue with current users. The availability of women with no prior HRT use varies by geographic region, however [59,60].
III. DATA COLLECTION
METHODS
As in all studies, there is typically a trade off between retrospective and prospective data collection in menopause studies, with a lower cost but also potentially lower accuracy for retrospective data collection. Moreover, as in studies of other topics, subject burden increases with the level of detail of data collection, and hence more detailed data collection regimes such as: daily hormone measurements tend to be employed with a selected subset (either by design, or by default due to subject nonresponse); consequently, the resulting data are less generalizable [4,24,61 ]. Thus, menopausal studies, which can involve relatively demanding data collection methods such as daily menstrual calendars, need to balance scientific rigor with participant burden. In addition, the menopause is a highly multifaceted phenomenon, involving changes in physiologic, epidemiologic, and psychosocial factors. Thus, investigators should consider collection of data in a number of domains [ 12]. These multidisciplinary data will provide better, more complete information for key study goals and will make efficient use of the large effort needed to recruit the participant sample. As an example, a recent study noted a link between depression and low bone density [62], possibly related to low estrogen concentrations, and another study found an association between bone density levels and breast cancer [63]; such findings are useful in a number of disciplines, including endocrinology, psychology, and oncology. This section presents a summary of data collection methods and types of instruments used in studies of menopause to collect information in various domains, moving from least to most detailed or demanding.
163 A. M a i l e d S u r v e y s This type of instrument has been used in past large epidemiologic cross-sectional surveys, such as the first phase of the Massachusetts Women's Health Study [15]. Such an instrument is fairly unobtrusive for subjects. It is also relatively inexpensive compared with other modes of data collection, although this advantage is offset somewhat by the need to conduct telephone followup of nonrespondents. For example, in the cross-sectional phase of the Massachusetts Women's Health Study, the initial response rate to the mailed survey was 57%; telephone followup of initial nonrespondents raised the combined final response rate to 77%. Moreover, the initial respondents to the mailed survey differed from nonrespondents to the mailed survey who subsequently responded by telephone, with higher levels of education and access to health care among the former [64]. Thus, it may be necessary to employ a mixed mode in order to reduce nonrespondent bias. By necessity, all data collected on a mailed pen-and-paper survey are self-reported. Thus, the investigators cannot verify data such as prescription medications, or anthropometric measures such as height or weight. Separate data coding and data entry also are required, unless scannable forms are used.
B. T e l e p h o n e S u r v e y s This mode has been used in a number of menopause studies, including the cross-sectional survey in the Melbourne Women's Midlife Health Project [ 17], the Healthy Women's Study [16], the longitudinal phase of the Massachusetts Women's Health Study [15] and in SWAN's cross-sectional phase [ 11 ]. Telephone surveys are useful not only for primary data collection, but for initial screening to identify and recruit specific cohorts. SWAN's cross-sectional telephone survey, for example, included questions used to assess cohort eligibility, particularly menopause status--which is not available on sampling frames; many field sites recruited eligible participants into the cohort at the conclusion of the telephone interview [11]. Depending on available technology, computeraided telephone interviewing can be used, which eliminates extra steps in data coding and data entry and leads to more accurate data collection. As with mailed surveys, however, there may be some bias, in this case due to subjects without telephone or with unlisted telephone numbers. Thus, a mixedmode approach may be needed, combining home visits with telephone interviews, as in SWAN [ 11 ].
C. I n - H o m e or I n - C l i n i c Visits Data collection has been carried out in subjects' homes or in local clinics in a number of recent menopause studies,
164
SYBILL. CRAWFORD
including the Healthy Women's Study, the Melbourne Women's Midlife Health Project, the second phase of the Massachusetts Women's Health Study, and SWAN's longitudinal phase. Many anthropometric measurements have been taken in a subject's home, including height, weight, blood pressure, and girth; blood and urine specimens also can be collected [65]. For longitudinal studies, it is important to employ similar collection methods for an individual subject at each visit in order to estimate within-woman changes, because measures such as blood pressure can vary by setting. Other physiologic data collected in clinics in studies of menopause include bone density and carotid ultrasound [ 11 ]. Collection of blood or urine specimens also can be done, as in the Melbourne Women's Midlife Health Project [17], the Massachusetts Women's Health Study [65], and SWAN [ 11 ]. This may be logistically difficult, however, depending on what is being measured. Specimens may need to be taken on a particular day of the menstrual cycle (e.g., days 2 - 5 in the early follicular phase for regularly cycling women), or at a certain time of day, for assessing concentrations of reproductive hormones. Fasting samples may be required for accurate measurements, e.g., of glucose. Studies examining the relationship of reproductive hormones to cardiovascular disease risk factors may impose multiple conditions on the specimen collection protocol. Consequently, the study should consider accommodation of subjects' schedules by allowing them to "drop in" for specimen collection on a different day from the rest of the data collection, as in SWAN [ 11 ]. This type of data collection, particularly when done at a clinic, involves considerably more participant burden than a mailed or telephone survey, and response rates are correspondingly lower. Comparing response rates for the crosssectional and longitudinal phases of SWAN, for example, the latter were substantially lower [ 11 ].
D. D a i l y C a l e n d a r s Daily collection of self-reported data is useful for measuring a series of similar, recurrent events such as menstrual bleeding or premenstrual or menopausal symptoms. Retrospective recall of these events is poor [22,66-69]. Several large menopause studies have employed calendars, including SWAN and the Massachusetts Women's Health Study. Response rates tend to be lower than for mailed or telephone surveys [67], with most of the dropouts occurring at the beginning of data collection. Depending on the length of the data collection period and on the amount of data women are asked to record, there may be fairly substantial data coding and data entry requirements. For a study of menopausal transitions or the perimenopausal period, a longer period of data collection may be required than for studying premenopausal women, e.g., 2 + years, in order to capture perimenopausal changes in bleeding over time.
The calendars are self-administered in a woman's home, which means that no review by study personnel is possible until after the calendar is sent back to the study site. Retrospective data recording is also an issue; data quality is less accurate when data are collected retrospectively rather than prospectively. Johannes and colleagues pilot-tested an electronic calendar for daily collection of menstrual bleeding and symptoms. The device recorded the time and date of data entry by the participant. In the pilot study comprising a month's data collection, all 23 subjects entered at least one day's data late, i.e.,after the day on which bleeding or symptoms occurred [70]. These results indicate that it is critical to take measures to ensure high-data quality, particularly when using traditional paper instruments. Calendar instruments should be very simple to understand, in order to minimize mistakes and respondent burden. Subjects should be asked to return completed calendars frequently, e.g., monthly, so that participation and data quality can be monitored, and to limit the amount of retrospective data recording. Completed calendarsmparticularly the first several calendars--should be examined in order to identify errors. Researchers may want to send a letter to participants noting commonly made errors, or even to make retraining calls to respondents whose cal' endars demonstrate a large number of problems.
E. D a i l y S p e c i m e n C o l l e c t i o n Daily collection of specimens such as serum or urine can be very informative, particularly in the study of perimenopause, during which hormone concentrations fluctuate widely even within an individual woman [71-76]. Thus such measures are highly superior to annual serum or urine samples, with respect to capturing within-woman variability. Such collection is expensive to conduct and requires a great deal of subject cooperation, however; consequently, sample sizes often are relatively small.
IV. M E A S U R E M E N T
ISSUES
A number of measurement issues arise in the study of menopause, particularly the determination of a woman's menopause status. Various researchers also have noted methodological difficulties in the measurement of menopausal symptoms, as well as cultural or ethnic differences in reporting of variables related to menopause.
A. D e f i n i t i o n s o f M e n o p a u s e Status Indicators used to define menopause status have included age, menstrual bleeding, levels of reproductive hormones, and a woman's self-report.
CHAPTER 10 Methodologic Challenges in Menopause Studies 1. CHRONOLOGIC AGE
Early studies of menopause used chronologic age as a proxy for postmenopause [12,24,77]. This is a very poor measure of menopause status, however, because the final menstrual period occurs over a wide age range [27]. A comparison of menses-based and age-based definitions using data from a case-control study of breast cancer [78] indicated that--using a menses-based definition as the "gold standard"msensitivity and specificity for the age-based definitions differed for cases and controls. In particular, there were more premenopausal women classified incorrectly as postmenopausal among cases than among controls, because age at menopause (by the menses-based definition) was later in cases than in controls. Thus studies of breast cancer employing age-based definitions of menopause status may suffer from this differential misclassification, which affects estimation of the association between menopause status and disease. 2. MENSTRUAL BLEEDING Past World Health Organization Working Group meetings [79,80] have recommended use of the following definitions for menopause status categories, based largely on observed menstrual cycle patterns, which are assumed to reflect underlying endocrinological changes or levels [81 ]: a. Natural menopause: the permanent cessation of menstruation, determined retrospectively after 12 consecutive months of amenorrhea without any other pathological or physiological cause. b. Perimenopause: the period just prior to the final menstrual period through the first year after the final menstrual period, beginning at the onset of endocrinologic and menstrual changes. c. Premenopause: the entire reproductive period prior to the FME d. Induced menopause: the cessation of menses due to removal of both ovaries with or without removal of the uterus, or iatrogenic ablation of ovarian function. e. Premature menopause: natural menopause occurring before age 40. Also known as premature ovarian failure. f. Postmenopause: dating from the FMP, including both natural and surgical menopause.
Note that the perimenopausal period as defined above overlaps with both the first 12 months of postmenopause after the FMP, and the premenopause. Metcalf [75] distinguishes premenopause as menstruating at regular intervals, whereas perimenopause begins with the onset of irregular cycling and continues after the FMP until hormone levels stabilize. Other uses of these terms in the literature [82-85] separate pre-, peri-, and postmenopause, with premenopause ending at the onset of endocrinologic or menstrual changes, and perimenopause concluding with the FMP.
165 3. ENDOCRINE MEASURES The decade prior to the FMP is characterized by an increase in variability in reproductive hormone concentrations, even though a woman may continue to have her normal menstrual bleeding. Abrupt changes in these hormones may occur, with values typical of postmenopause followed by those seen in younger reproductive-aged women [71-75,86]. Although hormone concentrations stabilize 1-2 years after the FMP [87-89], no sharp changes occur at the time of the FMP [74]. To categorize women regarding menopausal status, a cutoff of follicle-stimulating hormone (FSH) of 3 5 - 4 0 IU/liter is commonly used in clinical practice and in research studies [90,91 ]. Some studies have employed cutoffs of FSH greater than 10-20 IU/liter to indicate perimenopause [84,92]. The above discussion indicates, however, that endocrine variables cannot be considered reliable indicators of menopausal status [71,72,87], particularly based on a single serum or urine sample, because within-woman values fluctuate widely during perimenopause, and patterns are variable across women [76]. Hormone concentrations also vary by chronological age as well as by time before the FMP [71,76,93-97]. Although average values within and across women demonstrate general trends during this period, no single cutoff value is likely to be accurate as a predictor of status [98]. 4. SELF-DEFINITIONS
Women's perceptions of their own menopause status vary by culture and race/ethnicity [81], and do not agree completely with categorizations based on bleeding patterns [99]. Self-reported menopause status may not be appropriate for epidemiological purposes [81 ], but may be of interest in its own right or in studying women's experiences during the transition through menopause [99]. Assessment of years since the FMP from self-report on a cross-sectional survey can be inaccurate, because recall becomes increasingly unreliable with greater time elapsed since the FMP and there is evidence for digit preference [33,100-102] (see Section IV, C).
B. C h a r a c t e r i z i n g P e r i m e n o p a u s e The definition of natural menopause presented in Section IV,A,2 has become an accepted standard [82], and investigators have turned their attention to better characterizing perimenopause, for which no standard definition exists [5,82]. Treloar [103,104] defined the onset of perimenopause as the start of an increase in the variability in cycle length. Two later studies also identified menstrual irregularity as a perimenopausal indicator, using as a "gold standard" either an FSH level of at least 15 IU/liter [84] or subsequent transition to postmenopause within 3 years [82]. Menopausal symptoms, particularly hot flashes and night
166
SYBIL L. CRAWFORD
sweats, also were indicative of perimenopause. In the latter study, changes in menstrual flow were associated with status before controlling for irregularity, but were not independently related to status [82]. Further refinement of the characterization of perimenopause is needed [5,105]. Two studies have suggested a distinction between different stages of perimenopause, based on menstrual bleeding regularity [83,106]. Early perimenopause corresponds to self-reported changes in frequency of menstrual bleeding, whereas late perimenopause is defined by prolonged (more than 6 months) amenorrhea. Differentiation of these two stages appears to be informative in terms of summarizing the sequence of bleeding patterns, and for prediction of subsequent transition to postmenopause [83,106]. In addition, the simultaneous incorporation of multiple sources of data, e.g., symptoms, bleeding patterns, and reproductive hormone profiles, has been suggested as an area of future study [ 107].
C. R e l i a b i l i t y o f S e l f - R e p o r t e d D a t a Many epidemiologic studies of menopause ask a woman to report the date of her last menstrual period or the date of surgical menopause--used to estimate her age at menop a u s e m o r patterns of use of hormone replacement therapy. Several studies have investigated the reliability and reproducibility of such self-reported information, by comparison with medical records or by repeated interviewing of subjects over time. Considering menopause status, reliability and reproducibility tend to be relatively high for the type of menopause, i.e., natural versus surgical [108-110]. Reliability and reproducibility also are better for age at surgical menopause than for age at natural menopause [68,69,108-112]. In one study [108], women tended to underestimate age at menopause. This error has implications for estimating the association between age at menopause and disease. For example, breast cancer has been found to be positively related to later age at menopause, whereas osteoporosis and cardiovascular disease have been linked to an earlier menopause. If selfreported age at menopause is misclassified as compared with true age at menopause, the association of age at menopause with disease will be underestimated for breast cancer [ 108, 112] and overestimated for cardiovascular disease or osteoporosis [ 108]. Digit preference in reporting of age at menopause also was observed in several studies [102,111], particularly for ages ending in "0" and "5." For self-reported age at menopause, reliability tended to decrease as the time since the final menstrual period increased [ 109,111 ]. One study found, however, that reproducibility was higher as time since menopause increased, for women with an earlier menopause; thus it may be that women with a relatively young age at menopause (under 40) recall this event better [108].
Regarding recollections of patterns of HRT use, reliability was highest for ever-use [110,113-115]. For epidemiological studies, a single self-report question may be sufficiently accurate for this piece of information [ 113]. Selfreported details of use, including length of use, dose, and reasons for starting/stopping, however, were less reliable [110,113-114]. Lower reliability was related to subject characteristics such as higher age of the subject [114-115] and longer time elapsed since last use [ 113-115]. In summary, self-report may be adequate for basic data such as type of menopause or ever-use of HRT, but less appropriate for more detailed data such as the age at final menstrual period or length of HRT use. Thus, researchers may need to consider prospective designs or other sources of information, e.g., medical records abstraction, for these data.
D. M a s k i n g o f " N a t u r a l " M e n o p a u s e T r a n s i t i o n s Medical interventions, particularly exogenous reproductive hormone use and removal of the uterus and/or ovaries, affect characteristics commonly used to define menopause status [28]. Depending on the regime, use of HRT or oral contraceptives can alter menstrual bleeding, so that their use essentially masks menopausal status defined in terms of menstrual cycling [26,31,116]. Hormone use also can complicate classification based on endocrinological criteria, because it may affect reproductive hormone levels [116119]. Self-defined menopause status also varies by HRT use [99]. Surgical menopause, either from a hysterectomy or oophorectomy, obviously affects menstrual bleeding patterns. Moreover, women with a simple hysterectomy (ovaries not removed) may experience ovarian failure earlier than other women [ 120]. In short, straightforward determination of natural menopause status and timing of natural transitions that would have occurred in the absence of medical interventions is not possible. Implications for data analyses are discussed in Section V, and potential analytic strategies are presented.
E. M e a s u r e m e n t o f M e n o p a u s a l S y m p t o m s A number of methodological problems in past studies of menopausal symptoms have been identified. First, in order to avoid influencing a subject to associate certain symptoms with menopause, it is important to ask about general health and symptoms, including symptoms suspected of being related to menopause [69]; researchers should not identify symptoms a priori as being menopausal, or ask women what symptoms they experienced during menopause. Questions should include both positive and negative symptoms [3,5]. The reference period should be fairly short, e.g., 1 to 4 weeks, in order to minimize inaccuracies in recall [22,69]. Symptom reporting also is affected by cultural norms, as discussed in
CHAPTER 10 Methodologic Challenges in Menopause Studies the following subsection. In addition, researches should employ a standard scale, so that results across different studies can be compared [5,121 ]. A commonly used scale, the B lattKupperman index, is widely used but has been shown to be inadequate. Problems with this scale include development on a possibly unrepresentative sample of women and arbitrary weighting of items [122].
F. C u l t u r a l D i f f e r e n c e s in R e p o r t i n g Ethnicity and culture have been found to affect experiences and perceptions of menopause; this in turn leads to ethnic or cultural differences in reporting of menopauserelated data. For example, symptom reporting varies by culture and geographic region, with lower rates found in Asian and Central American populations than in the United States and Western Europe [ 123-125]. In Mayan Indians, one study [123] found no self-reported hot flashes, despite the occurrence of endocrine changes that were similar to those seen in women in the United States. Japanese women tend to report headaches, shoulder stiffness, and joint pain, whereas vasomotor symptoms appear to be rare and tend not to be associated with menopause status [126]. Ethnicity and culture also are related to sexual attitudes, values, and behavior [22], as well as to perceptions and reporting of menstrual bleeding [66,67]. These differences need to be accounted for in studies of menopause, by using culturally appropriate instruments [66,67,125].
V. A N A L Y T I C
CONSIDERATIONS
This section summarizes a variety of issues involved in analyses of data collected in studies of menopause.
A. M e t h o d s for E s t i m a t i n g A g e at N a t u r a l M e n o p a u s e Data from either cross-sectional (i.e., prevalence) or longitudinal (incidence) studies may be used for estimation of age at natural menopause. 1. PREVALENCE DATA Distributions of recalled age at menopause from crosssectional data can be analyzed using techniques such as Kaplan-Meier plots and Cox proportional hazards modeling [26]. Typically, prevalent cases of natural menopause are asked when their periods stopped, and data for pre- or perimenopausal women are censored at their current age. Techniques that do not account for censoring, e.g., histograms of age at menopause in the subset of prevalent naturally menopausal women, lead to estimates of age at menopause that are biased downward [26]. As noted in Section III, use of retrospective recall of age
167 at the FMP may be problematic; consequently, self-reporting of 12 + months amenorrhea (yes~no) at the time of the interview may be more accurate. Use of this outcome variable suggests estimation of a binary logistic regression of 12+ months of amenorrhea on chronologic age. Median age at menopause (or other percentiles) then can be estimated from the logistic regression model as a function of the intercept and slope [15,127]. Median age at menopause can be estimated for various subgroups, e.g., smokers and nonsmokers, by stratification on the characteristic of interest [15]. Note that a logistic regression analysis excludes women with a prevalent surgical menopause. For Caucasians, several studies suggest that this exclusion is appropriate, and that a competing risks model is not necessary [128,129]. It is unclear whether this holds for other racial/ethnic groups, however, particularly for African-Americans, who have a much higher rate of hysterectomy [29]. 2. INCIDENCE DATA As noted earlier, data from a prospective design, where information on menopause transitions is collected as they occur, is preferable to a retrospective design in terms of accuracy. Covariates also can be assessed prior to transitions, with less recall bias [ 10]. Longitudinal analyses of incidence data often employ hazards modeling, for example, estimating the probability that an event--such as the F M P - - occurs during the study, given that it has not occurred earlier. If the age at the FMP can be measured precisely, e.g., using menstrual calendars, then one can use techniques such as Cox proportional hazards modeling. If menopause status is ascertained only at intervals, e.g., at an annual interview, one can employ an approach used by Brambilla and McKinlay [ 10], which involves multinomial modeling of conditional probabilities of menopause status categories at each interview, including natural menopause, surgical menopause, and not yet menopausal. Analyses of age at natural menopause may need to take into account the competing risk of surgical menopause, although longitudinal analyses by Brambilla and McKinlay in non-Hispanic Caucasians suggest that this is not necessary for this racial/ethnic group [ 10], consistent with cross-sectional findings. B. A n a l y t i c M e t h o d s for M e n s t r u a l C a l e n d a r D a t a Goals of analyses of data from menstrual calendars typically include characterization of the distribution and patterns of menstrual segment lengths. Because bleeding may or may not correspond to a menstrual cycle, the term "segment" is sometimes used rather than "cycle" [67]. Many studies focus on segment length or bleeding length rather than on heaviness of menstrual flow [66,67,130,131 ]. The latter has been found to be less informative, e.g., in defining menopause status [82,83]. The majority of past studies utilizing menstrual calendars have been done in premenopausal women, with the exception of Treloar [103,104], who
168 followed women from menarche to the FME This subsection summarizes issues involved in the analysis of calendar data from premenopausal--regularly cycling--women, as well as analytic complications arising from the study of perimenopausal, i.e., irregular, bleeding. 1. ISSUES IN ANALYSIS OF PREMENOPAUSAL CALENDARS
Data can be analyzed using either a menstrual segment or an individual woman as the unit of analysis. Both approaches are useful, and the appropriate choice depends on the question of interest [132]. If the goal is inference regarding the distribution of segment length for an individual in the population of interest, then the woman should be used as the unit of analysis. An example is the reference period method, where each woman's bleeding patterns are summarized for a standard unit of time, typically 90 days [67]; this provides a cross-sectional summary for each subject. A related issue is length bias. In general, the observation period is fixed and the number of observations (segments) per woman varies. Consequently, women with shorter-and hence more--segments observed are overrepresented in analyses using the segment as the unit of analysis. Thus, using the segment as the unit of analysis can give estimates of segment length that are biased downward. In contrast, bywoman analyses give each woman the same weight. Because the observation period is usually defined in terms of calendar time rather than in terms of completion of a menstrual segment, the last segment is only partially observed. This censoring can cause bias in estimates, because the probability that its length is unknown is related to the length of the segment, with longer segments more likely to be censored [133]. The resulting bias may be small if the data collection period is relatively long. Belsey and Farley [67] summarize a number of analytic methods proposed to handle this censoring, including omitting the last segment from analyses; including it only in estimation of variability but not mean length; truncation, which affects variability estimates; and methods for handling right-censored survival data [ 133]. A number of statistical techniques have been employed in the analysis of menstrual segment lengths, all of which handle the correlation between multiple segments observed in the same subject. Methods also should account for length bias. Techniques such as growth curve modeling may not be particularly useful, because the number of observations (segments) is inherently part of the data to be observed. Methods that explicitly examine within-woman correlation between segments include estimation of segment-to-segment probabilities [106,130], e.g., whether long segments tend to be followed by shorter segments, as well as estimation of autocorrelation to assess the dependence between segment lengths as a function of the lag between segments [130]. Techniques used to model segment length include randomeffects modeling with a random intercept, or equivalently, a
SYBIL L. CRAWFORD
generalized estimating equation (GEE) approach with exchangeable correlation; this assumes that a woman's segment lengths vary randomly about her own mean [130]. Methods that incorporate covariates for segment length include GEE techniques [134,135], Poisson modeling [136], and autoregressive modeling [133]. Harlow and Zeger also employed a mixture model approach to characterize the distribution of segment lengths, whereby one component consisted of "normal" length segments and the other component included very long segments [130]. 2. ANALYTIC COMPLICATIONS FROM PERIMENOPAUSAL DATA Additional analytic issues arise in the study of perimenopausal menstrual segment lengths, due to increased menstrual irregularity. A key question is what constitutes a menstrual segment. World Health Organization definitions require at least one bleeding ~ not spotting~day followed by at least one bleed-free day [67]. Some analysts of premenopausal data omit spotting episodes from analyses entirely [67,137,138]. Spotting episodes, however, may be quite informative in the study of perimenopausal bleeding patterns. Johannes and colleagues, for example, found spotting episodes to be more common in the early perimenopause, indicating the utility of spotting episodes in distinguishing perimenopausal stages [106]. Thus a "conservative" definition of menstrual episodes or segments, whereby any spotting or bleeding is considered separately, may be in order. As noted in Section IV, perimenopause is characterized by within-woman changes in bleeding patterns, particularly an increase in irregularity. Hence the autocorrelation structure for segment lengths within an individual woman may be very different from the exchangeable correlation observed in studies of regularly cycling premenopausal women. To capture this, we may need a more complicated autocorrelation structure. Moreover, irregularity itself is not a welldefined concept. Bleeding patterns during the perimenopause may vary not only across women, but within women as well, and may depend on proximity to the final menstrual period [82,83,106]. Thus models of perimenopausal segment lengths need to allow autocorrelation structures to vary both within and across women. Finally, the complete interval of perimenopausal menstrual bleeding may not be observed during the period of study. Data may be subject to either left or right censoring, or both, depending on the woman's initial menopause status, the length of her perimenopause, and the length of the calendar data collection. Also, as just noted, bleeding patterns may change for an individual over time, depending on proximity to the final menstrual period [82,83,106]. Thus the menstrual segments observed during the study may not be representative of a woman's entire perimenopausal period, unlike segments observed in regularly cycling women. Analyses of perimenopausal segment length should account for this censoring.
CHAPTER 10 Methodologic Challenges in Menopause Studies C. M e t h o d s for C o m b i n i n g D a t a C o l l e c t e d at D i f f e r e n t F r e q u e n c i e s Menopause studies often involve different types of instruments, collecting data at varying frequencies, e.g., annual clinic visits, monthly symptom reporting collected via calendars, and daily urine specimens. Scientific questions of interest may require combining these data, as in assessment of the relationship between reproductive hormone levels measured annually and symptom patterns observed in monthly menstrual calendars, or in a comparison of daily menstrual calendar data to self-reported data on bleeding patterns from an annual interview. Possible approaches include "collapsing" the data measured at a higher frequency of measurement. For example, Johannes and colleagues summarized each woman's 12 months of bleeding patterns in terms of within-woman mean and variance of segment length [ 106]. The correlations of these summaries with annual reproductive hormone values then was computed. The time scale of one of the measures also can be adapted, as was done in several analyses of predictors of menstrual segment length [131,135]. Time-varying covariates such as weight were measured less frequently than monthly, and the schedule of measurement did not correspond to a woman's menstrual segments. To include these variables as predictors, the investigators defined the value of a covariate corresponding to a particular menstrual segment as the average of that covariate during the first 14 days of that segment or of the preceding segment. Another approach is to analyze only data measured concurrently, e.g., estimating the correlation between reproductive hormone concentrations measured at an annual clinic visit and characteristics of the menstrual segment occurring during that annual clinic visit. This has the disadvantage of ignoring other, possibly relevant, data, however.
D. M a s k i n g o f M e n o p a u s e S t a t u s D u e to U s e o f H R T As noted in Section IV, use of HRT prior to observation of 12 + months of amenorrhea results in an inability to assess a woman's "true" menopause status in the absence of HRT, or to determine her age at "natural" menopause. A variety of analytic techniques have been employed or proposed to handle these women in analyses where natural menopause status or transitions are variables of interest. A common approach is to omit ever-users or concurrent users from analyses [28,139-141]. However, as discussed in Section II, HRT users are not a random sample of all women traversing the menopause. Thus analyses that omit these women completely do not describe experiences in the overall
169 population. In longitudinal analyses in which baseline data are available prior to initiation of HRT, one can include some data from these women by censoring their observations at the time of HRT initiation, or omitting observations concurrent with HRT use [ 19]. Another technique is to analyze menopause status for non-HRT users, then compare users and nonusers, omitting menopause status as a variable [142]. This method has the advantage of including users in analyses. HRT users may be a mixture of "natural" menopause status categories, however, so that putting them in a single category may not be appropriate. For the same reason, treating HRT users as a separate menopause status category may distort the estimated association between menopause status and other variables. Other analyses have combined HRT users with postmenopausal women [85]. Many women, however, begin HRT use prior to permanent cessation of menses [49], and thus are likely to be dissimilar to naturally postmenopausal women. Inclusion of HRT users with this latter group could weaken or bias estimated associations of postmenopausal status with other characteristics. HRT use may be included in analyses as a covariate, and menopause status defined in terms of observed bleeding patterns or estrogen levels regardless of HRT use status [23]. For some outcomes, such as depression or sexual activity, the source of estrogen may be relatively unimportant to the question of interest, so that the outcome can be modeled as a function of a marker of the total estrogen exposure (endogenous and exogenous combined), and an indicator for HRT use (yes/no). For other outcomes, however, the distinction between endogenous and exogenous estrogen may be of greater relevance. For example, endogenous estrogen has little liver exposure compared with oral preparations of exogenous estrogen [40,143], and thus using total estrogen as a predictor may not be applicable for outcomes such as circulating lipids. Perhaps the most appropriate general approach is to consider "natural" menopause status as missing for HRT users, and employ techniques developed to handle missing data. Menopause status is unlikely to be missing completely randomly, so that analyses would require techniques that assume either data missing at random (related only to observed data) or nonignorable missingness (related to the unobserved "true" menopause status) [144,145]. Note that all approaches used in this situation necessarily rely on assumptions that are untestable, because "true" menopause status is not observable.
E. H a n d l i n g S u r g i c a l M e n o p a u s e in A n a l y s e s As with HRT use, surgical menopausemhysterectomy and/or bilateral oophorectomy--effectively masks or censors a woman's natural menopause status or transitions that
170 would have occurred in the absence of medical intervention. Often surgically menopausal women are omitted from data anlayses [28]. Similar to HRT users, however, surgically menopausal women are not a random subsample of all women, nor are they a small subsamplemover one-third of women in the United States undergo a hysterectomy by age 60 [ 146]. Thus results from analyses omitting these women will not be generalizable to the entire population [ 129]. Analyses of age at menopause as a potential risk factor for diseases such as breast cancer sometimes assign a "mean" or "typical" age at the FMP to surgically menopausal women. This imputation process, however, does not reflect the underlying distribution of the age at natural menopause, and thus distorts the associations of age at the FMP with disease outcomes [147]. Hysterectomy or oophorectomy status also can be included as a covariate or stratifying factor in analyses. Type of menopause can be included as a predictor, comparing naturally postmenopausal women to surgically menopausal women, as in PEPI [37]. Given the many differences between these two groups, it may be necessary to stratify data analyses on type of menopause [4,81 ]. Data may also be treated as censored for surgically menopausal women. For example, in survival analyses of age at natural menopause using prevalence data, data from surgically menopausal women may be censored at the time of surgery [31 ]. This approach assumes that a woman's experience in the absence of surgery is similar to that of women observed to have a natural menopause [129]. This may be accurate after controlling for predictors of type of menopause, such as access to medical care; that is, natural status may be missing at random. This assumption is essentially untestable, however. Finally, analyses may employ competing risks modeling. Such techniques can be used either with prevalence data, to estimate models for 12+ months of amenorrhea (yes~no) [128], or self-reported age at the FMP or surgery [129], or longitudinal incidence data [ 10].
E Assessing Associations between Menopause Status or Reproductive Age and Health Outcomes For cross-sectional data, any inferences regarding the role of menopause transition within a woman are done using between-women comparisons. Depending on the available data, analyses may employ categorized menopause status (pre, peri, post, surgical), reproductive age defined as time before/after the FMP, or levels of reproductive hormones. Analyses need to control for important confounders such as age, by including them as model predictors or by age matching [19]. For longitudinal data, one can directly examine withinwoman changes in outcomes concurrent with within-woman
SYBIL L. CRAWVORD
changes in menopause status. For example, one can model successive differences in the outcome, e.g., change in serum cholesterol or bone density from one annual visit to the next, as a function of corresponding successive changes in menopause status [23,142]. Longitudinal data also permit better assessment of temporal sequences of events, such as whether fluctuations in reproductive hormone levels precede or follow more overt signs of perimenopause such as increased menstrual irregularity. For cross-sectional data, statistical methods include linear and logistic regression, depending on the outcome variable, including concurrent menopause status as a covariate. Longitudinal analytic techniques must account for within-subject correlation of multiple observations. Approaches that consider each observation separatelymdata are not collapsed within a woman m include generalized estimating equation methods, repeated measures modeling, and random-effects modeling. This last method can be used, for example, to identify women with an extreme menopausal trajectory, e.g., "fast" losers of bone density. Other methods collapse data within a woman, e.g., a paired t-test of a subject's mean outcome level prior to the FMP versus the corresponding mean after the FMP [140]. Alternatively, one can apply spline analysis, fitting within-woman slopes before and after the FMP for a piecewise linear model, and compare pre-FMP and post-FMP slopes [ 139-141 ]. Choice of functional forms is key. For example, analysts of bone density data often use log of years since menopause as a predictor of bone density [92]; this functional form implies that bone loss is more rapid for women in early postmenopause. It is important to determine the presence or absence of curvilinear relationships, e.g., whether bone loss accelerates in perimenopause and levels off after the FME Depending on the outcome of interest, it is also critical to distinguish pre- from perimenopause rather than combining all observations prior to the FMP, because acceleration of changes due to menopause may occur well before the FMP [28]. One should also consider the amount of change in reproductive hormone concentrations in addition to absolute concentrations; rapid hormonal changes may be associated with outcomes such as symptoms [5,148-150].
G. Confounding, Effect Modification, and Stratification A key confounding factor related to both status and many health outcomes of interest is smoking. Smokers have an earlier menopause [10,15,30,31], higher levels of risk factors for cardiovascular disease [34], and a higher risk of low bone density and osteoporosis [35]. Without controlling for smoking status, the role of the menopause transition in changes in levels of disease risk is overstated [4,30,150]. Other potential
171
CHAPTER 10 Methodologic Challenges in Menopause Studies
confounders to consider include body mass index, possibly parity, and oral contraceptive use [30,31-33]. Relationships between menopause status and outcomes such as cardiovascular disease risk also may differ across subpopulations, e.g., smokers versus nonsmokers. In analyses of blood pressure and lipids, for example, menopause status played a larger role for nonsmokers [4]. Thus, analyses may need to include appropriate interaction terms between menopause status and smoking status, or even to stratify on smoking status. Type of menopause~surgical or naturalmis another potential stratification factor. As noted in Section II, surgically menopausal women have very different experiences before, during, and after the menopausal transitions. Consequently, including the type of menopause as a covariate may not be sufficiently complex to assess the role of surgical versus natural menopause; analyses may need to be stratified on type of menopause [4].
VI. CONCLUSION In closing, it is important to note comments by Lock [ 125]. The emphasis of much of the studies of menopause to datemparticularly in the United States and Europemhas been on its negative health consequences, e.g., experience of menopausal symptoms, loss of bone density, and increase in cardiovascular disease risk. Cross-cultural studies suggest, however, that the menopause is not universally a time of decline in health, and that influences other than biology, such as culture, are involved in women's menopausal transitions. Lock proposes that investigators identify factors that are associated with a positive menopausal experience, and that research take into account variables from a variety of domains including lifestyle and psychosocial as well as physiological.
9.
10.
11.
12.
13.
14.
15. 16.
17.
18. 19.
20. 21.
22.
References 1. McKinlay, S. M., and McKinlay, J. B. (1973). Selected studies of the menopause--A methodological critique. J. Biosoc. Sci. 5, 533-555. 2. Avis, N. E., Crawford, S. L., and McKinlay, S. M. (1997). Psychosocial, behavioral, and health factors related to menopause symptomatology. Women's Health: Res. Gender, Behav. Policy 3(2), 103-120. 3. McKinlay, J. B., McKinlay, S. M., and Brambilla, D. J. (1987). Health status and utilization behavior associated with menopause. Am. J. Epidemiol. 125, 110-121. 4. McKinlay, S. M. (1994). Issues in design, measurement, and analysis for menopause research. Exp. Gerontol. 29, 479-493. 5. Greendale, G. A., and Sowers, M. (1997). The menopause transition. Endocrinol. Metab. Clin. North Am. 26(2), 261-277. 6. Gath, D. (1998). The assessment of depression in peri-menopausal women. Maturitas 29, 33-39. 7. Morse, C. A., Smith, A., Dennerstein, L., Green, A., Hopper, J., and Burger, H. (1994). The treatment-seeking woman at menopause. Maturitas 29, 161-173. 8. Groeneveld, F. P. M. J., Bareman, F. P., Barensten, R., Dokter, H. J.,
23.
24.
25.
26. 27. 28. 29.
Drogendijk, A. C., and Hoes, A. W. (1993). Relationships between attitude towards menopause, well-being and medical attention among women aged 45-60 years. Maturitas 17, 77-88. Montero, I., Ruiz, I., and Hernandez, I. (1993). Social functioning as a significant factor in women's help-seeking behaviour during the climacteric period. Soc. Psychiatry Psychiatr. Epidemiol. 28, 178-183. Brambilla, D. J., and McKinlay, S. M. (1989). A prospective study of factors affecting age at menopause. J. Clin. Epidemiol. 42, 10311039. Sowers, M., Crawford, S., Sternfeld, B., Morganstein, D., Gold, E., Greendale, G., Evans, D., Neer, R., Matthews, K., Sherman, S., Lo, A., Weiss, G., and Kelsey, J. (2000). In "Menopause: Biology and Pathobiology" (R. A. Lobo, J. Kelsey, R. Marcus, eds.). Academic Press, San Diego. Rostosky, S. S., and Travis, C. B. (1996). Menopause research and the dominance of the biomedical model 1984-1994. Psychol. Woman Q. 20, 285-312. Miles, M. P., and Malik, K. C. (1994). Menopause and AfricanAmerican women: Clinical and research issues. Exp. Gerontol. 29, 511-518. Harlow, S. D., and Ephross, S. A. (1995). Epidemiology of menstruation and its relevance to women's health. Epidemiol. Rev. 17, 265286. McKinlay, S. M., Bifano, N. L., and McKinlay, J. B. (1985). Smoking and age at menopause in women. Ann. Intern. Med. 103, 350-356. Meilahn, E. N., Matthews, K. A., Egeland, G., and Kelsey, S. E (1989). Characteristics of women with hysterectomy. Maturitas 11, 319-329. Dennerstein, L., Smith, A. M. A., Morse, C., Burger, H., Green, A., Hopper, J., and Ryan, M. (1993). Menopausal symptoms in Australian women. Med. J. Aust. 159, 232-236. Kish, L. (1965). "Survey Sampling." Wiley, New York. Shelley, J. M., Green, A., Smith, A. M. A., Dudley, E., Dennerstein, L., Hopper, J., and Burger, H. (1998). Relationship of endogenous sex hormones to lipids and blood pressure in mid-aged women. Ann. Epidemiol. 8, 39-45. Ware, J. H. (1985). Linear models for the analysis of longitudinal studies. Am. Stat. 39, 95-101. Ware, J. H., Dockery, D., Louis, T. A., Xu, X. P., Ferris, B. G., Jr., and Speizer, F. E. (1990). Longitudinal and cross-sectional estimates of pulmonary function decline in never-smoking adults. Am. J. Epidemiol. 32, 685-700. McCoy, N. L. (1998). Methodological problems in the study of sexuality and the menopause. Maturitas 29, 51-60. Avis, N. E., Brambilla, D., McKinlay, S. M., and Vass, K. (1994). A longitudinal analysis of the association between menopause and depression: Results from the Massachusetts Women's Health Study. Ann. Epidemiol. 4, 214-220. Sowers, M. R., and La Pietra, M. T. (1995). Menopause: Its epidemiology and potential association with chronic diseases. Epidemiol. Rev. 17, 287-302. Davis, J. W., Ross, P. D., Wasnich, R. D., Maclean, C. J., and Vogel, J. M. (1989). Comparison of cross-sectional and longitudinal measurements of age-related changes in bone mineral content. J. Bone Miner. Res. 4, 351-357. Cramer, D. W., and Xu, H. (1996). Predicting age at menopause. Maturitas 23, 319-326. McKinlay, S. M., Brambilla, D. J., and Posner, J. G. (1992). The normal menopause transition. Am. J. Hum. Biol. 4, 37-46. Holte, A. (1998). Menopause, mood and hormone replacement therapy: Methodological issues. Maturitas 29, 5-18. Kjerulff, K. H, Guzinski, G. M., Langenberg, P. W., Stolley, P. D., Moye, N. E., and Kazandjian, V. A. (1993). Hysterectomy and race. Obstet. Gynecol. 82, 757-764.
172 30. Willett, W., Stampfer, M. J., Bain, C., Lipnick, R., Speizer, F. E., Rosner, B., Cramer, D., and Hennekens, C. H. (1983). Cigarette smoking, relative weight, and menopause. Am. J. Epidemiol. 117, 651-658. 31. Bromberger, J. T., Matthews, K. A., Kuller, L., Wing, R. R., Meilahn, E. N., and Plantinga, E (1997). Prospective study of the determinants of age at menopause. Am. J. Epidemiol. 145, 124-133. 32. Stanford, J. L., Hartge, E, Brinton, L. A., Hoover, R. N., and Brookmeyer, R. (1987). Factors influencing the age at natural menopause. J. Chronic. Dis. 409, 995-1002. 33. MacMahon, B., and Worcester, J. (1966). "Age at menopause: United States 1960-1962," Vital Health Stat. Ser. 11, No. 19, PHS Publ. no. 1000, 1-19. U.S. Govt. Printing Office, Washington, D.C. 34. Barrett-Connor, E. (1994). Heart disease in women. Fertil. Steril. 62, 127S-132S. 35. Cummings, S. R., Kelsey, J. L., Nevitt, M. C., and O'Dowd, K. J. (1985). Epidemiology of osteoporosis and osteoporotic fractures. Epidemiol. Rev. 7, 178-208. 36. Women's Health Initiative Study Group (1998). Design of the Women's Health Initiative clinical trial and observational study. Controlled Clin. Trials 19, 61 - 109. 37. Espeland, M. A., Bush, T. L., Mebane-Sims, I., Stefanick, M. L., Johnson, S., Sherwin, R., and Waclawiw, M. (1995). Rationale, design, and conduct of the PEPI trial. Controlled Clin. Trials 16, 3S-19S. 38. Greendale, G. A., Hogan, P., Kritz-Silverstein, D., Langer, R., Johnson, S. R., and Bush, T. (for the PEPI trial investigators) (1995). Age at menopause in women participating in the Postmenopausal Estrogen/Progestins Interventions (PEPI) Trial: An example of bias introduced by selection criteria. Menopause 2, 27-34. 39. Johannes, C. B., and Avis, N. E. (1996). The short-term health consequences of hysterectomy. J. Women's Health 5, 278. 40. Bush, T. L. (1990). The epidemiology of cardiovascular disease in postmenopausal women. Ann. N. Y. Acad. Sci. 592, 263-271. 41. Barrett-Connor, E. (1991). Postmenopausal estrogen and prevention bias. Ann. Intern. Med. 115, 455-456. 42. Barrett-Connor, E., and Bush, T. L. (1991). Estrogen and coronary heart disease in women. J. Am. Med. Assoc. 265, 1861-1867. 43. Barrett-Connor, E. (1996). The menopause, hormone replacement, and cardiovascular disease: The epidemiologic evidence. Maturitas 23, 227-234. 44. Hemminki, E., Malin, M., and Topo, P. (1993). Selection to postmenopausal therapy by women's characteristics. J. Clin. Epidemiol. 46, 211-219. 45. Matthews, K. A., Kuller, L. H., Wing, R. R., Meilahn, E. N., and Plantinga, P. (1996). Prior to use of estrogen replacement therapy, are users healthier than nonusers? Am. J. Epidemiol. 143, 971-978. 46. Vandenbroucke, J. P. (1995). How much of the cardioprotective effect of postmenopausal estrogens is real? Epidemiology 6, 207-208. 47. Posthuma, W. E M., Westendorp, R. G. J., and Vandenbroucke, J. P. (1994). Cardioprotective effect of hormone replacement therapy in postmenopausal women: Is the evidence biased? Br. Med. J. 308, 1268-1269. 48. Derby, C. A., Hume, A. L., Barbour, M. M., McPhillips, J. B., Lasater, T. M., and Carleton, R. A. (1993). Correlates ofpostmenopausal estrogen use and trends through the 1980s in two southeastern New England communities. Am. J. Epidemiol. 137, 1125-1135. 49. Johannes, C. B., Crawford, S. L., Posner, J. G., and McKinlay, S. M. (1994). Longitudinal patterns and correlates of hormone replacement therapy use in middle-aged women. Am. J. Epidemiol. 140, 439-452. 50. Rosenberg, L. (1993). Hormone replacement therapy; The need for reconsideration. Am. J. Public Health 83, 1670-1673. 51. Nachtigall, L. E., Nachtigall, R. H., Nachtigall, R. D., and Beckman, E. M. (1979). Estrogen replacement therapy I: A 10-year prospective study in the relationship to osteoporosis. J. Am. Coil. Obstet. Gynecol. 53, 277-281. 52. Criqui, M. H., Suarez, L. Barrett-Connor, E., McPhillips, J., Wingard,
SYBIL L. CRAWFORD
53.
54.
55.
56. 57.
58.
59.
60.
61.
62.
63.
64.
65. 66. 67.
D. L., and Garland, C. (1988). Postmenopausal estrogen use and mortality: Results from a prospective study in a defined, homogeneous community. Am. J. Epidemiol. 182, 606-614. Sturgeon, S. R., Schairer, C., Brinton, L. A., Pearson, T., and Hoover, R. N. (1995). Evidence of a healthy estrogen user survivor effect. Epidemiology 6, 227-231. Petitti, D. B. (1994). Coronary heart disease and estrogen replacement therapy: Can compliance bias explain the results of observational studies? Ann. Epidemiol. 4, 115-118. Gallagher, E. J., Viscoli, C. M., and Horwitz, R. I. (1993). The relationship of treatment adherence to the risk of death after myocardial infarction in women. J. Am. Med. Assoc. 270, 742-744. Vandenbroucke, J. E (1991). Postmenopausal oestrogen and cardioprotection. Lancet 337, 833- 834. Grady, D., Applegate, W., Bush, T., Furberg, C., Riggs, B., and Hulley, S. B. (1998). Heart and Estrogen/progestin Replacement Study (HERS): Design, methods, and baseline characteristics. Controlled Clin. Trials 19, 314-335. Johnson, S., Mebane-Sims, I., Hogan, P. E., and Stoy, D. B. (1995). Recruitment of postmenopausal women in the PEPI trial. Controlled Clin. Trials 16, 20S-35S. Bastian, L. A., Couchman, G. M., Rimer, B. K., McBride, C. M., Feaganes, J. R., and Siegler, I. C. (1997). Perceptions of menopausal stage and patterns of hormone replacement therapy use. J. Women's Health 6, 467-475. Brett, K. M. and Madans, J. H. (1997). Use of postmenopausal hormone replacement therapy: Estimates from a nationally representative cohort study. Am. J. Epidemiol. 145, 536-545. Hernandez-Avila, M., Stampfer, M. J., Ravnikar, V. A., Willett, W. C., Schiff, I., Francis, M., Longcope, C., and McKinlay, S. M. (1993). Caffeine and other predictors of bone density among pre- and perimenopausal women. Epidemiology 4, 128-134. Michelson, D., Stratakis, C., Hill, L., Reynolds, J., Galliven, E., Chrousos, G., and Gold, P. (1996). Bone mineral density in women with depression. N. Engl. J. Med. 335, 1176-1181. Zhang, Y., Kiel, D. P., Kreger, B. E., Cupples, L. A., Ellison, R. C., Dorgan, J. F., Schatzkin, A., Levy, D., and Felson, D. T. (1997). Bone mass and the risk of breast cancer among postmenopausal women. N. Engl. J. Med. 336, 611 - 617. Brambilla, D. J., and McKinlay, S. M. (1987). A comparison of responses to mailed questionnaires and telephone interviews in a mixed mode health survey. Am. J. Epidemiol. 126, 962-971. McKinlay, S. M., and McKinlay, J. B. (1986). Aging in a "healthy" population. Soc. Sci. Med. 236, 531-535. Snowden, R. (1977). The statistical analysis of menstrual bleeding patterns. J. Biosoc. Sci. 9, 107-120. Belsey, E. M., and Farley, T. M. M. (1987). The analysis of menstrual bleeding patterns: A review. Appl. Stochastic Models Data Analy. 3, 125-150.
68. Bean, J. A., Leper, J. D., Wallace, R. B., Sherman, B. M., and Jagger, H. (1979). Variation in the reporting of menstrual histories. Am. J. Epidemiol. 109, 181-185. 69. Kaufert, P. A., Gilbert, P., and Hassard, T. (1988). Researching the symptoms of menopause: An exercise in methodology. Maturitas 10, 117-131. 70. Johannes, C. B., Crawford, S. L., Woods, J., Goldstein, R. B., Tran, D., Mehrotra, S., Johnson, K. B., and Santoro, N. (2000). An electronic menstrual cycle calendar: Comparison of data quality with a paper version. Menopause, in press. 71. Burger, H. C. (1994). The menopause: When is it all over oris it? Aust. N.Z. J. Obstet. Gynaecol. 34, 293-295. 72. Burger, H. G. (1994). Diagnostic role of follicle-stimulating hormone (FSH) measurements during the menopausal transitionmAn analysis of FSH, oestradiol and inhibin. Eur. J. Endocrinol. 130, 38-42. 73. Hee, J., MacNaughton, J., Banagh, M., and Burger, H. G. (1993).
173
CHAPTER 10 M e t h o d o l o g i c Challenges in M e n o p a u s e Studies
74.
75. 76.
77.
78.
79. 80.
81.
82.
83.
84.
85.
86.
87. 88.
89.
90. 91.
92.
93.
94.
Perimenopausal patterns of gonadotrophins, immunoreactive inhibin, oestradiol and progesterone. Maturitas 18, 9-20. Metcalf, M. G., Donald, R. A., and Livesey, J. H. (1982). Pituitaryovarian function before, during, and after the menopause: A longitudinal study. Clin. Endocrinol. 17, 489-494. Metcalf, M. G. (1988). The approach of menopause: A New Zealand study. N.Z. Med. J. 101, 103-106. Reame, N. E., Kelche, R. P., Beitins, I. Z., Yu, M. Y., Zawacki, C. M., and Padmanabhan, V. (1996). Age effects of follicle-stimulating hormone and pulsatile luteinizing hormone secretion across the menstrual cycle of women. J. Clin. Endocrinol. Metab. 81, 1512-1518. Bungay, G. T., Vessey, M. P., and McPherson, C. K. (1980). Study of symptoms in middle life with special reference to the menopause. Br. Med. J. 281, 181-183. Morabia, A., and Flandre, P. (1992). Misclassification bias related to definition of menopausal status in case-control studies of breast cancer. Int. J. Epidemiol. 21,222-228. World Health Organization Scientific Group (1981). "Research on the Menopause," WHO Tech. Serv. Rep. Ser. 670. WHO, Geneva. World Health Organization Scientific Group (1996). "Research on the Menopause in the 1990s," WHO Tech. Serv. Rep. Ser. No. 866. WHO, Geneva. Kaufert, P., Lock, M., McKinlay, S., Beyenne, Y., Coope, J., Davis, D., Eliasson, M., Gognalons-Nicolet, M., Goodman, M., and Holte, A. (1986). Menopause research: The Korpilampi workshop. Soc. Sci. Med. 22, 1285-1289. Brambilla, D. J., McKinlay, S. M., and Johannes, C. B. (1994). Defining the perimenopause for application in epidemiologic investigations. Am. J. Epidemiol. 140, 1091-1095. Dudley, E. C., Hopper, J. L., Taffe, J., Guthrie, J. R., Burger, H. G., and Dennerstein, L. (1998). Using longitudinal data to define the perimenopause by menstrual cycle characteristics. Climacteric 1, 18-25. Cooper, G. S., and Baird, D. D. (1995). The use of questionnaire data to classify peri- and premenopausal status. Epidemiology 6, 6 2 5 628. Wing, R. R., Matthews, K. A., Kuller, L. H., Meilahn, E. N., and Plantinga, P. (1991). Weight gain at the time of menopause. Arch. Intern. Med. 151, 97-102. Shideler, S. E., DeVane, G. W., Kalra, P. S., Benirschke, K., and Lasley, B. L. (1989). Ovarian-pituitary hormone interactions during the perimenopause. Maturitas 11, 331-339. Burger, H. G. (1996). The endocrinology of the menopause. Maturitas 23, 129-136. Wide, L., Nillis, S. J., Gemzell, C., and Roos, P. (1973). Radioimmunosorbent assay of follicle-stimulating hormone and luteinizing hormone in serum and urine from men and women. Acta Endocrinol. (Copenhagen) 174, 3-58. Chakravarti, S., Collins, W. P., Forecast, J. D., Newton, J. R., Oram, D. H., and Studd, J. W. (1976). Hormonal profiles after the menopause. Br. Med. J. 2, 784-786. Wilson, J. D., and Foster, D. W., eds. (1992). "William's Textbook of Endocrinology, 8th Ed." Saunders, Philadelphia. Goldenberg, R. L., Grodin, J. M., Rodbard, D., and Ross, G. T. (1973). Gonadotropins in women with amenorrhea. Am. J. Obstet. Gynecol. 116, 1003-1012. Nordin, B. E. C., Morris, H. A., Need, A. G., Horowitz, M., and Robertson, W. G. (1990). Relationship between plasma calcium fractions, other bone-related variables, and serum follicle-stimulating hormone levels in premenopausal, perimenopausal, and postmenopausal women. Am. J. Obstet. Gynecol. 163, 140-145. Lenton, E. A., Sexton, L., Lee, S., and Cooke, I. D. (1988). Progressive changes in LH and FSH and LH:FSH ratio in women throughout reproductive life. Maturitas 10, 35-43. MacNaughton, J., Banah, M., McCloud, P., Hee, J., and Burger, H.
95.
96.
97.
98.
99. 100.
101. 102.
103. 104. 105. 106.
107.
108.
109.
110.
(1992). Age related changes in follicle stimulating hormone, luteinizing hormone, oestradiol and immunoreactive inhibin in women of reproductive age. Clin. Endocrinol. 36, 339-345. Reyes, F. I., Winter, J. S. D., and Faiman, C. (1977). Pituitary-ovarian relationship preceding the menopause. Am. J. Obstet. Gynecol. 129, 557-564. Metcalf, M. G., and Livesey, J. H. (1985). Gonadotrophic excretion in fertile women: Effect of age and the onset of the menopause. J. Endocrinol. 105, 357-362. Sherman, B. M., and Korenman, S. G. (1975). Hormonal characteristics of the human menstrual cycle throughout reproductive life. J. Clin. Invest. 55, 699-706. Stellato, R., Crawford, S., McKinlay, S., and Longcope, C. (1998). Can follicular stimulating hormone be used to define menopause? Endocr. Pract. 4, 137-141. Taffe, J., Garamszegi, C., Dudley, E., and Dennerstein, L. (1997). Determinants of self rated menopause status. Maturitas 27, 223-229. Paganini-Hill, A., Krailo, M. D., and Pike, M. C. (1984). Age at natural menopause and breast cancer: The effect of errors in recall. Am. J. Epidemiol. 119, 81-85. McKinlay, S. M., Jefferys, M., and Thompson, B. (1972). An investigation of the age at menopause. J. Biosoc. Sci. 4, 161-173. Benjamin, F. (1960). The age of the menarche and of the menopause in white South African women and certain factors influencing these times. S. Afr. Med. J. 34, 316-320. Treloar, A. E. (1974). Menarche, menopause and intervening fecundability. Hum. Biol. 46, 89-107. Treloar, A. E. (1981). Menstrual cyclicity and the pre-menopause. Maturitas 3, 249-264. Kaufert, P. A., Gilbert, P., and Tate, R. (1987). Defining menopausal status: The impact of longitudinal data. Maturitas 9, 217-226. Johannes, C. B., Crawford, S. L., Longcope, C., and McKinlay, S. M. (1996). Bleeding patterns and changes in the perimenopause: A longitudinal characterization of menstrual cycles. Clin. Consult. Obset. Gynecol. 8, 9-20. Kuller, L. H., Meilahn, E. N., Cauley, J. A., Gutai, J. P., and Matthews, K. A. (1994). Epidemiologic studies of menopause: Changes in risk factors and disease. Exp. Gerontol. 29, 495-509. den Tonkelaar, I. (1997). Validity and reproducibility of self-reported age at menopause in women participating in the DOM-project. Maturitas 27, 117-123. Colditz, G. A., Stampfer, M. J., Willett, W. C., Stason, W. B., Rosner, B., Hennekens, C. H., and Speizer, F. E. (1987). Reproducibility and validity of self-reported menopausal status in a prospective cohort study. Am. J. Epidemiol. 126, 319-325. Paganini-Hill, A., and Ross, R. K. (1982). Reliability of recall of drug usage and other health-related information. Am. J. Epidemiol. 116, 114-122.
111. Hahn, R. A., Eaker, E., and Rolka, H. (1997). Reliability of reported age at menopause. Am. J. Epidemiol. 146, 771-775. 112. Horwitz, R. I., and Yu, E. C. (1985). Problems and proposals for interview data in epidemiological research. Int. J. Epidemiol. 14, 4 6 3 467. 113. Greendale, G. A., James, M. K., Espeland, M. A., and Barrett-Connor, E. (1997). Can we measure prior postmenopausal estrogen/progestin use? Am. J. Epidemiol. 146, 763-770. 114. Goodman, M. T., Nomura, A. M. Y., Wilkens, L. R., and Kolonel, L. N. (1990). Agreement between interview information and physician records on history of menopausal estrogen use. Am. J. Epidemiol. 131, 815-825. 115. Jannausch, M. L. and Sowers, M. R. (1992). Consistency ofperimenopausal estrogen use reporting by women in a population-based prospective study. Maturitas 14, 161-169. 116. Creinin, M. D. (1996). Laboratory criteria for menopause in women using oral contraceptives. Fertil. Steril. 66, 101-104.
174 117. Schiff, I. (1980). The effects of conjugated estrogens on gonadotropins. Fertil. Steril. 33, 333-334. 118. Mathur, R. S., Landgrebe, S. C., Moody, L. O., Semmens, J. E, and Williamson, H. O. (1985). The effect of estrogen treatment on plasma concentrations of steroid hormones, gonadotropins, prolactin and sex hormone-binding globulin in post-menopausal women. Maturitas 7, 129-133. 119. Powers, M. S., Schenkel, L., Darley, E E., Good, W. R., Balestra, J. C., and Place, V. A. (1985). Pharmacokinetics and pharmacodynamics of transdermal dosage forms of 17fi-estradiol: Comparison with conventional oral estrogens used for estrogen replacement. Am. J. Obstet. Gynecol. 152, 1099-1106. 120. Siddle, N., Sarrel, E, and Whitehead, M. (1987). The effect of hysterectomy on the age at ovarian failure: Identification of a subgroup of women with premature loss of ovarian function and literature review. Fertil. Steril. 47, 94-100. 121. Greene, J. G. (1998). Constructing a standard climacteric scale. Maturitas 29, 25-31. 122. Alder, E. (1998). The Blatt-Kupperman menopausal index: A critique. Maturitas 29, 19-24. 123. Martin, M. C., Block, J. E., Sanchez, S. D., Arnaud, C. D., and Beyene, Y. (1993). Menopause without symptoms: The endocrinology of menopause among rural Mayan Indians. Am. J. Obstet. Gynecol. 168, 1839-1845. 124. Beyene, Y. (1986). The cultural significance and physiological manifestation of menopause: A biocultural analysis. Cult. Med. Psychiatry 10, 47-71. 125. Lock, M. (1998). Menopause: Lessons from anthropology. Psychosom. Med. 60, 410-419. 126. Lock, M., Kaufert, E, and Gilbert, E (1988). Cultural construction of the menopausal syndrome: The Japanese case. Maturitas 10, 317-332. 127. Morgan, B. J. T. (1992). "Analysis of Quantal Response Data." Chapman & Hall, London. 128. Krailo, M. D., and Pike, M. C. (1983). Estimation of the distribution of age at natural menopause from prevalence data. Am. J. Epidemiol. 117,356-361. 129. Shinberg, D. S. (1998). An event history analysis of age at least menstrual period: Correlates of natural and surgical menopause among midlife Wisconsin women. Soc. Sci. Med. 46, 1381-1396. 130. Harlow, S. D., and Zeger, S. L. (1991). An application of longitudinal methods to the analysis of menstrual diary data. J. Clin. Epidemiol. 44, 1015-1025. 131. Harlow, S. D., and Matanoski, G. M. (1991 ). The association between weight, physical activity, and stress and variation in the length of the menstrual cycle. Am. J. Epidemiol. 133, 38-49. 132. Harlow, S. D., and Zeger, S. L. (1991). An application of longitudinal methods to the analysis of menstrual diary data. J. Clin. Epidemiol. 44, 1015-1025. 133. Murphy, S. A., Bentley, G. R., and O'Hanesian, M. A. (1995). An analysis for time-varying covariates. Stat. Med. 14, 1843-1857.
SYBIL L. CRAWFORD 134. Harlow, S. D., and Campbell, B. (1996). Ethnic differences in the duration and amount of menstrual bleeding during the postmenarcheal period. Am. J. Epidemiol. 144, 980-989. 135. Harlow, S. D., and Campbell, B. C. (1994). Host factors that influence the duration of menstrual bleeding. Epidemiology 5, 352-355. 136. Collett, D., and Weerasooriya, N. (1993). A modelling approach to the analysis of menstrual diary data. Stat. Med. 12, 955-965. 137. Treloar, A. E., Boynton, R. E., Behn, B. G., and Brown, B. W. (1967). Variation of the human menstrual cycle through reproductive life. Int. J. Fertil. 12, 77-126. 138. Rodriguez, G., Faundes-Latham, A., and Atkinson, L. E. (1976). An approach to the analysis of menstrual patterns in the critical evaluation of contraceptives. Stud. Fam. Plann. 7, 42-51. 139. van Beresteijn, E. C. H., Korevaar, J. C., Huijbregts, P. C. W., Schouten, E. G., Burema, J., and Kok, F. J. (1993). Perimenopausal increase in serum cholesterol: A 10-year longitudinal study. Am. J. Epidemiol. 137, 383-392. 140. Jensen, J., Nilas, L., and Christiansen, C. (1990). Influence of menopause on serum lipids and lipoproteins. Maturitas 12, 321-331. 141. Falch, J. A., and Sandvik, L. (1990). Perimenopausal appendicular bone loss: A 10-year prospective study. Bone 11,425-428. 142. Crawford, S. L., Casey, V. A., Avis, N. E., and McKinlay, S. M. (2000). A longitudinal study of weight and the menopause transition: Results from the Massachusetts Women's Health Study. Menopause, in press. 143. Longcope, C., Herbert, E N., McKinlay, S. M., and Goldfield, S. R. (1990). The relationship of total and free estrogens and sex hormonebinding globulin with lipoproteins in women. J. Clin. Endocrinol. Metab. 71, 76-72. 144. Little, R. J. A., and Rubin, D. B. (1987). "Statistical Analysis with Missing Data." Wiley, New York. 145. Rubin, D. B. (1987). "Multiple Imputation for Nonresponse in Surveys." Wiley, New York. 146. National Center for Health Statistics, Pokras, R., and Hufnagel, V. G. (1987). "Hysterectomies in the United States, 1965-84," Vital Health Stat., Ser. 13, No. 92, DHHS Publ. no. (PHS) 88-1753. U.S. Gov. Printing Office, Washington, DC. 147. Pike, M. C., Ross, R. K., and Spicer, D. V. (1998). Problems involved in including women with simple hysterectomy in epidemiologic studies measuring the effects of hormone replacement therapy on breast cancer risk. Am. J. Epidemiol. 147, 718-721. 148. Schmidt, E J., and Rubinow, D. R. (1991). Menopause-related affective disorders: A justification for further study. Am. J. Psychiatry 148, 844-852. 149. Brincat, M., Magos, A., Studd, J. W., Cardozo, L. D., O'Dowd, T., Wardle, P. J., and Cooper, D. (1984). Subcutaneous hormone implants for the control of climacteric symptoms: A prospective study. Lancet 1, 16-18. 150. Stampfer, M. J., Colditz, G. A., and Willett, W. C. (1990). Menopause and heart disease: A review. Ann. N. Y Acad. Sci. 592, 286-294.
~HAPTER 1
SWAN: A Multi c enter,
Multiethnic, CommunityBased Cohort Study
of Women and the Menopausal Transition MARYFRAN SOWERS,* SYBIL L. CRAWFORD, t BARBARA STERNFELD,; DAVID M O R G A N S T E I N , wE L L E N B . G O L D , G A I L A . G R E E N D A L E , # D E N I S E V A N S , * * R O B E R T N E E R , tt K A R E N M A T T H E W S , ~ S H E R R Y S H E R M A N , w167 ANNIE LO, wGERSON WEISS, A N D J E N N I F E R K E L S E Y ## *Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan 48109; tNew England Research Institutes, Watertown, Massachusetts 02472; *Department of Epidemiology and Biostatistics, Division of Research, Kaiser Permanente, Oakland, California 94611; ~Westat, Inc., Rockville, Maryland 20850; IIDepartmentof Epidemiology and Preventive Medicine, School of Medicine, University of California, Davis, Davis, California 95616; #Departments of Medicine and Obstetrics and Gynecology, University of California, Los Angeles, School of Medicine, Los Angeles, California 90024; **Rush Institute on Aging, Chicago, Illinois 60612; ttDivision of Endocrinology, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts 02114; **Departmentof Psychiatry, University of Pittsburgh, Pittsburgh, Pennsylvania 15261; ~NIH/NIA, Bethesda, Maryland 20892; IIIIDepartmentof Obstetrics and Gynecology, University of Medicine and Dentistry of New Jersey, Newark, New Jersey 07103; and ##Division of Epidemiology, Department of Health Research & Policy, Stanford University, School of Medicine, Stanford, California 94305
I. II. III. IV. V. VI.
Appendix A. SWAN Investigators Appendix B. Specific Sampling and Recruiting Strategies by Sites with List-Based Primary Sampling Frames Appendix C. Specific Sampling and Recruiting Strategies by Sites with RDD-Based Primary Sampling Frames References
Introduction Overview of the Study Design Data Collection Sampling and Recruitment Strengths and Limitations of SWAN Summary
I. I N T R O D U C T I O N
completely understood [1,2]. Furthermore, much of what is known is based on data from Caucasian women, from women who are self-referred to menopause clinics, or from convenience samples of women seen in the clinical setting
Menopause is a universal phenomenon of women, yet, as discussed in other chapters in this book, it is inMENOPAUSE:
BIOLOGY AND PATHOBIOLOGY
175
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
176 for other health problems. In the next two decades, approximately 40 million American women will experience the menopause [3] and by the year 2005, it is estimated that $3-5 billion will be spent annually for hormone replacement therapy (HRT) and the physician monitoring of that HRT use [4]. Additionally, study of the menopause poses special methodological challenges because of its transitional nature, the potential for involving multiple organ systems (i.e., bone, lipids, mental health), and the potential contribution of varied social, behavioral, and cultural factors (see Chapter 10). Thus, study of the menopausal transition is both important and complex. To address many of the knowledge deficits about the menopausal transition identified in chapters of this book, the Study of Women's Health Across the Nation (SWAN), a multisite, longitudinal study of the natural history of menopause, was funded by the National Institute on Aging, the National Institute of Nursing Research, and the Office of Research on Women's Health. The overall goal of SWAN is to describe the chronology of the biological and psychosocial characteristics of the menopausal transition and the effect of this transition on subsequent health and risk factors for agerelated chronic diseases. Because investigation of the menopausal experience in minority women has been neglected, SWAN placed special emphasis on including minority populations. This would allow SWAN to describe the sociocultural, lifestyle, psychological, and biological characteristics of these groups in relation to the menopausal transition [5]. In addition, emphasis was placed on recruiting a sample of women that was community or population based, rather than volunteer or clinically based, so that the sample would be representative of women from the full spectrum of socioeconomic status and cultural experiences. The specific aims of SWAN, shown below, are being addressed in a representative cohort of initially premenopausal women who are socially and culturally diverse. The aims are as follows: 1. To characterize the symptomatology, hormonal, and bleeding pattern characteristics related to the menopausal transition. 2. To investigate the hormonal and menstrual bleeding pattern characteristics related to change in bone mineral density, cardiovascular status markers, measures of carbohydrate metabolism, and body composition during the menopausal transition. 3. To examine the relations of psychosocial factors, personality characteristics, and behaviors, including lifestyle behaviors, as they may relate to age at onset, symptoms, and physiological changes of the menopausal transition. 4. To discern what changes observed over time are related to the menopausal transition as compared to age-related changes, including those changes that appear to accelerate the aging process.
SOWERS ET AL.
5. to describe and quantify cultural and ethnic differences among women with respect to midlife aging and the menopausal transition among the four race/ethnic groups of the cohort, in addition to non-Hispanic Caucasians. This chapter is an overview of the SWAN study design and includes a brief description of SWAN's comprehensive data collection. The data being collected mirrors the specific aims, reflecting the belief that the biologic process of menopause occurs within the context of diverse personality characteristics, psychosocial factors, and behavioral attributes as well as an ethnic and cultural context. Consequently, the methods used to recruit this important sample of multiethnic women are described and the strengths and limitations of those methods are discussed.
II. O V E R V I E W
OF THE STUDY DESIGN
SWAN is organized as a prospective, multicenter, multiethnic, multidisciplinary study of the natural history of the menopausal transition, under the auspices of a cooperative agreement between the National Institutes of Health and seven sites with clinical examination facilities, a data coordinating center, and two laboratories. Appendix 1 shows the locations, investigators, and roles of those investigators. SWAN includes a large and representative sample of African-American, non-Hispanic Caucasian, Chinese, Hispanic, and Japanese women. The study design, developed in a collaborative process, consists of a cross-sectional study and a longitudinal cohort study, both of which employ common protocols across the seven sites with clinical examination facilities. Focus groups were conducted to inform the development of the study design and the protocols and to ensure the relevance and the appropriateness of the protocols to the multiethnic cohort. The SWAN Cross-sectional Study consisted of a 15- to a 20-minute telephone interview (or face-to face interview in those instances in which no telephone number could be associated with the sampled respondent). The interview was administered to 16,065 women aged 40-55 years who were randomly selected from sampling frames established at each site with clinical examination facilities (described more fully in Section IV). The two purposes of the SWAN Cross-sectional Study were to identify women eligible for study longitudinally and assess, cross-sectionally, those factors associated with the age at natural menopause, the prevalence of surgical menopause, symptoms of menopause, health status, and health care use. Additional information about the eligibility criteria, sampling frames, and characteristics of participants are discussed in Section IV. On completion of the interview, eligibility for the longitudinal study was determined and women meeting the eligibility criteria were invited to join that cohort. The annual examinations of the SWAN Longitudinal Study include
CHAPTER 11 SWAN Cohort Study TABLE I
177
The Breadth of Measures and Their Frequency of Ascertainment in the SWAN Longitudinal Study
Type of measurement
Frequency a
Questionnaire
Type of measurement
Frequency
a
Specimen data b
Socioeconomic status Medical history Psychosocial environment Lifestyle behaviors Menstrual status Natural/surgical menopause Symptoms Use of medical services Use of medications Quality of life Sexual activity Food frequency
Annual Annual Annual Annual Annual Annual Annual Annual Annual Annual Annual Base line, F/U-04
Clinic measurements
Blood (serum) E 2, FSH, SHBG, DHEAS, testosterone TSH Glucose and insulin Fibrinogen, factor VII, PAI-1, TPA antigen Lipid profile, HDL subfractions, lipoprotein (a) Biochemical bone turnover markers (at five sites) Serum repository specimens
Annual Base line Annual Base line Base line, F/U-01, 02, 03, 05 Base line, F/U-01, 02, 03 Annual
Urine N-Teleopeptides (at five sites) Urine repository specimens
Annual Annual
Other data collection beyond annual evaluation
Anthropometry Blood pressure Bone density (at five sites)
Annual Annual Annual
Abstract medical records for hysterectomy Menstrual calendars Daily Hormone Study (subsample of 900)
Monthly One cycle, annually
a Note: F/U denotes a follow-up examination; the number denotes which follow-up examination. bAbbreviations: E2, estradiol; FSH, follicle-stimulating hormone; SHBG, sex hormone binding globulin; DHEAS, dehydroepiandrosterone sulfate; TSH, thyroid-stimulating hormone; PAI-1, plasminogen activator inhibitor-1; TPA, tissue plasminogen activator; HDL, high-density lipoprotein.
questionnaires, blood and urine specimen collection, and physical measures (Table I). Because the SWAN Longitudinal Study is focused on the menopausal transition, unique data collection activities are required. For example, the annual examinations are scheduled for days 2 - 5 after bleeding c o m m e n c e s to standardize serum h o r m o n e measures to the early phase of the menstrual cycle. In addition, the cohort is followed with monthly menstrual cycle calendars and a subset of the cohort participates in daily urine collection as well as maintaining a daily symptom diary for one complete menstrual cycle on an annual basis. The following section describes the data collection more fully.
III.
DATA A.
COLLECTION Theoretical
Approaches
As a multidisciplinary study, the SWAN data collection instruments and approaches were developed to address the potential contribution of the multiple theories surrounding the study of the menopausal transition [6]. For example, the biological approach ascribes the experience of the menopause particularly within the framework of alterations in metabolism and endocrine status. The psychological~psychosocial approach maintains the importance of stressors and losses as catalysts for symptoms. The sociocultural/environmental approach indicates that cultural constructs and lifestyle factors define our responses toward the menopause and the presentation of potential symptoms. Finally, the feminist
theory views the menopause as a normal developmental stage with its own unique challenges. The instruments and data collection activities of SWAN have reflected an inclusive approach that acknowledges the need for and value of each of these perspectives, while minimizing the reductionist approach to studying and interpreting the characteristics of the menopause transition.
B. Types of Data The types of data collected from SWAN participants in the annual examinations are shown below in examples that include the type of construct and contributing variables. Construct
Variable
Acculturation
Language used, cultural and religious practices, dietary practices Weight changes associated with each pregnancy, weight cycling Use of contraceptive methodologies Use of hormone preparations, past use of oral contraceptives, and current contraception methodology Smoking history and current passive smoke exposure; current caffeine and alcohol consumption; diet and dietary practices, including use of supplements; amount and frequency of physical activity practices, including planned exercise
Body size history Contraception Hormone use practices
Lifestyle behaviors
178
SOWERS ET AL.
Construct
Variable
Construct
Variable
Menstrual status
Current menstrual bleeding characteristics and their variation according to timing, duration, and intensity; usual premenstrual symptoms, if any
Bone status and its turnover
Psychological status
Depression, hostility, and stress
Bone mineral density of the femoral head, lumbar spine, and total body (from five sites with bone densitometry facilities); biochemical measures of bone formation and resorption
Recent medical care utilization
Frequency of prevention behaviors, including Pap smear, physical breast exam, and physician visit for health problem or routine check-up; use of complementary and alternative health approaches; health insurance
Carbohydrate and energy metabolism
Glucose, insulin and thyroid-stimulating hormone concentrations (the latter at base line)
Clotting factors
Fibrinogen, factor VIIc, plasminogen activator inhibitor-1, tissue plasminogen activator antigen
Relationships
Number, nature, and satisfaction with relationships; life satisfaction
Lipid metabolism
Reproductive history
Age at menarche, gravidity, parity, pregnancy losses, infertility, lactation practices
Total cholesterol, triglycerides, highand low-density lipoprotein cholesterol, high-density lipoprotein cholesterol subfractions, lipoprotein (a)
Reproductive hormones
Self-perception
Quality of life, health status, degree of physical activity
Estradiol, follicle-stimulating hormone, sex hormone binding globulin, progesterone, and testosterone
Sexuality
Types of practices, satisfaction, and attitudes toward sex
Significant life events
Marriage, divorce, death or birth in family, change in or loss of job, illness, social support, occupational stress (autonomy)
Significant medical history
Diagnosis by a physician of hypertension, cardiovascular disease, malignancies, or thyroid disease; fractures, pelvic surgery, urinary incontinence, current medications, family history of health events
Sociodemographic status
Age, birth date, birthplace, marital status, level of education, income of household, occupation and the physicality of one's work, household composition
Social environment
Discrimination, religiosity, and spirituality
The interview data will be linked with other information being collected annually that describes the physical and hormone status of enrollees. The general areas of interest and the variables that contribute to the constructs are shown below.
Construct
Variable
Blood pressure
Resting systolic and diastolic blood pressure, resting heart rate
Body composition and body topology
Weight, height, waist and hip circumference, body composition (the latter from five sites with bone densitometry facilities)
Two additional data collection elements, monthly menstrual cycle calendars and daily specimen/diary collection, are important in more precisely characterizing the transitional process. The monthly menstrual calendars provide a record of the changing characteristics of menstrual bleeding from month to month. These monthly calendars also include a record of the use of oral contraceptives or other hormones, symptoms, and the occurrence of any gynecological surgery or procedures. A more extensive calendar is in use at three clinical sites to ascertain lifestyle factors including dieting, shift work, exercise practice, and smoking behavior, as well as stress. A Daily H o r m o n e Study (DHS) contributes to the SWAN specific aims by providing a more complete understanding of the variation in hormone concentrations throughout menstrual cycles (or equivalent time periods) of the perimenopausal transition and characterizing changes in the nature and frequency of within-cycle events, such as ovulation. Blood and urine specimens are being collected from a subsample of 900 women, with participation at each of the seven sites and from each of the race/ethnic groups as well as the non-Hispanic Caucasian women. Participants collect daily urine specimens for one complete menstrual cycle each year. These urine specimens are assayed for excretion products of the pituitary (the gonadotropins, luteinizing hormone, and follicle-stimulating hormone) and the ovary (estrone conjugate and pregnanediol glucuronide). The goal is to describe the changes in the hormone concentrations at important points during the menopausal transition and prior to the final menstrual period. The DHS also includes a daily diary to characterize symptoms and social dimensions of each day during the cycle of the daily urine collection.
CHAPTER 11 SWAN Cohort Study
179
TABLE II The Site-Specific Recruiting Goals for Race/Ethnicity Percentage in the SWAN Longitudinal Study of the Menopausal Transition in Seven Geographic Locales a Primary race/ethnic self-identification (%) AfricanAmerican
Geographic locale Detroit, Michigan (Ypsilanti/Inkster) Chicago, Illinois (Morgan Park/Beverly) Boston/Cambridge, Massachusetts Pittsburgh, Pennsylvania (Allegheny County) Oakland, California (plus Hayward and Richmond) Newark, New Jersey (Hudson County) Los Angeles, California (South Bay/Sawtelle)
Chinese Hispanic Japanese Caucasian
66 55 45 33
33 45 55 66 45 33 45
55 66 55
a Each site was to recruit at least 450 women to the SWAN Longitudinal Study with the proportion of primary race/ ethnicity among women described in the table.
Collectively, the monthly menstrual calendars and the Daily Hormone Study help to refine the definition of the menopause by providing more frequent and supplemental data during the transitional period. It is anticipated that an outgrowth will be the provision of more comprehensive understanding of the bleeding and hormone markers of the onset of perimenopause and the stages within the transition process.
recruitment strategy that they considered optimal for the Study's scientific questions, the characteristics of the local site (including access to clinical facilities), and the specific minority population to be evaluated. The result was the use of multiple sampling frames and multiple sampling approaches implemented in a coordinated manner. SWAN thus also provides the opportunity to describe and evaluate the various sampling frames, the sampling approaches to recruiting women from those frames, and the relative impact of using the various sampling frames and approaches.
IV. SAMPLING AND RECRUITMENT A. Overview
B.
The SWAN sampling and recruiting was implemented in seven locations in the United States: Boston, Chicago, the Detroit area, Los Angeles, Newark, Pittsburgh, and Oakland, California. The recruitment goal for each of the seven sites was to obtain representative samples of at least 450 women [non-Hispanic Caucasian women and one designated minority group (African-American, Chinese, Hispanic, and Japanese)] in a proportion specific for each site (see Table II). To achieve this goal, each site developed a sampling and
SWAN Recruitment
As indicated previously, recruitment for SWAN was undertaken as a two-step process (Table III). The first recruitment step involved a cross-sectional study to act as a sampiing frame for the SWAN Longitudinal Study. The second recruitment step was the development of a longitudinal study cohort from among the SWAN Cross-sectional Study enrollees. To be eligible for participation in SWAN Cross-sectional Study, women had to meet the following criteria:
TABLE III Summary of Sampling Units Contacted to Determine Eligibility in the Two-Step Process to Identify SWAN Longitudinal Study Enrollees Recruitment step Cross-sectional study recruitment, sampling units contacted Longitudinal study recruitment, units from the cross-sectional study
Sampling units
No. eligible
No. recruited
Response rate (%)
202,985
34,985
16,065
46.6
3,306
50.7
16,065
6,521 a
aThere were 36 Caucasian women included in this table who were "filtered out" (i.e., eligible to enter the cohort, but not recruited because target recruitment had been met).
180
S O W E R S ET AL.
1. Primary residence in designated geographic area 2. Ablility to speak English or designated other language m Spanish, Cantonese, or Japanese 3. Age 4 0 - 5 5 years at time of contact 4. Cognitive ability to provide verbal informed consent 5. Membership in a specific site's targeted ethnic groups To identify women eligible for the cross-sectional study, sites screened the constituent sampling units from the sampiing frames. Depending on the site, the sampling units were the households, telephone numbers, or individual names of women; the sampling frames were the listings of the sampiing units. Study-wide, 202,485 sampling units from sampiing frames were evaluated, leading to the identification of 34,446 eligible women. Of these, 16,065 women completed the SWAN Cross-sectional Study. The eligibility criteria for the SWAN Longitudinal Study wereas follows: 1. Aged 4 2 - 5 2 years 2. No surgical removal of the uterus and/or both ovaries 3. Not currently using exogenous hormone preparations affecting ovarian function 4. At least one menstrual period in the previous 3 months 5. Self-identification with one of each site's designated race/ ethnic group From the SWAN Cross-sectional Study, 6557 women were identified as eligible for longitudinal study. Of these women, 36 Caucasian women were "filtered out," i.e., they were not asked to participate in the longitudinal study because the site had met its target sample size. Of the remaining 6521 women, 3306 were recruited for the SWAN Longitudinal Study (see Table IV). This is the cohort currently being followed.
TABLE IV
Geographic Primary locale frame type" Boston Chicago Detroit
List List List
Frames
To identify the cohort for the longitudinal study, sites had to address successfully three competing requirements. These requirements were to (1) identify populations representative of a defined and diverse community, (2) recruit women from a specified race/ethnic minority group in a proportion significantly greater than the groups' proportion in the general United States population, and (3) implement the recruitment in a defined and circumscribed geographic locale so that relatively intense longitudinal clinical studies could be sustained. To meet these requirements and to be cost-efficient, the seven sites selected study communities that had a relatively higher density of the particular racial/ethnic minority group designated for recruitment. Then, individual sites utilized a variety of sampling frames from which the sample(s) would be drawn. In general, these sampling frames included telephone numbers randomly generated from random digit dialing (RDD)-based and list-based frames (Table IV). The following sections describe both types of frames in the context of the SWAN geographic locations and racial/ethnic group requirements. Appendices 2 and 3 provide specific information about the sampling approach at each clinical site. 1. RANDOM DIGIT DIALING-BASED FRAMES
The sampling unit for RDD frames was a telephone number and the only eligibility information available from an RDD frame consisted of the telephone number itself, i.e., the geographic location associated with the telephone number's exchange. Three sites [Newark area, Pittsburgh area, and Los Angeles (Table IV)] use an RDD-based sampling frame as the major frame. Two of these sites (Newark and Los Angeles) used list-assisted RDD-based sampling, and the Pitts-
The Primary Sampling Frame, Supplemental Frames, and Type of Supplemental Information Provided to the Primary Frame According to Geographic Locales Primary frameb
Oakland List Los Angeles RDD
City census listing Enrollmentlist from earlier study Electricalutility company customer listing for communitycensus HMO enrollment list RDD 3+ approach
Pittsburgh Newark
RDD RDD 3+ approach
RDD RDD
C. T h e S a m p l i n g
Supplemental frames" None None None None VRL, telephone directory list, ethnic organization membershiplists, snowball VRL Snowball
Supplemental information added to frames Telephone numbers, face-to-facecontact None Telephone directory, race from organization lists and VRL, face-to-facecontact None
Telephone directory None
RDD, Random digit dialing. is a variation in random digit dialing that increases the likelihood that telephone numbers are households and not commercialfirms. cVRL, Voter registration list.
a
b 3+
CHAPTER 11 SWAN Cohort Study burgh site used voter registration lists as their important secondary frame. Sites with a primary RDD sampling frame implemented the following steps: 1. Each telephone number was screened to determine if it represented a household. 2. The household was then screened to verify that the household was in the target geographic area and to determine if any woman age-eligible for the cross-sectional study was in residence. 3. Personnel then determined whether the household included at least one age-eligible woman who was Caucasian or was from the site's designated racial/ethnic minority group. All three of the sites that used the RDD sampling frame supplemented the RDD frame with list-based or "snowball" (referral by other participants) sampling frames. 2. LIST-BASED FRAMES At four sites, lists representing households (Detroit area) or individual women (Boston, Chicago, and Oakland areas) comprised the primary sampling frames. The list-based frames were varied and included a state-mandated census in Boston, an electrical utility customer listing in the Detroit area, a census from a previous study in the Chicago area, and a health maintenance organization (HMO) enrollment list in the Oakland, California area. Although each of these sites recruited its entire sample using a list sampling frame, only one site had a single frame that included all the information necessary to determine eligibility a priori (age, address, telephone, geographic area, race/ethnicity, gender) for recruiting to the SWAN Cross-sectional Study. That single list-based frame had been developed in a previous research study.
D.
Sampling Strategies
Specific sampling procedures varied across sites and were a function of the sampling frame(s) used and the level of information available with the frame(s). Sampling procedures included conducting a census, implementing an area probability sampling, and identifying acquaintance networks with snowballing. For example, the Detroit site conducted a census in which every household in the selected geographic area was enumerated and contacted, with the probability of selection being 1. Area probability samples were developed and implemented in the Chicago, Oakland, and Pittsburgh areas, where women were sampled with a known probability of selection that was <1. In addition to using their R D D samples, the sites at Los Angeles and Newark also used snowball sampling. With snowball procedures, selection probabilities could not be determined. To minimize the likelihood of selection bias from factors such as seasonal variation or systematic sampling, sites
181 organized their contacts with the sampling units into "batches." Each batch was a random sample within the overall sample from the particular site. The approximate size of each batch was derived from the number of sampling units that the site projected could be contacted in a 2-month time period. For those households with more than one eligible woman, a single eligible woman was sampled by one of two approaches. Five sites sampled by selecting the woman in the household with the most recent birthday (month and day only), an approach referred to as the "birthday" method. Two sites (Los Angeles and New Jersey) sampled the first woman contacted who was willing to provide information. Only one household member was sampled to avoid clustering of women within households for study variables such as health status and health care use. All sites used computer-assisted telephone interviewing (CATI) with standardized interviews and scripts to facilitate contact in a consistent manner, thereby minimizing the opportunity for information bias. When telephone numbers were unavailable for a sizable proportion of the population, as was the case for the Detroit area census, interviewers directly contacted those households without listed telephone numbers, face to face.
E. C o m p u t a t i o n
of the Response Rates
C o m m o n disposition codes were used to define the status of each sampling unit eligible for contact and to facilitate the computation of the response rates (see Table V). Response rates for the SWAN Cross-sectional Study were calculated as follows:
TABLE V Disposition Codes for Units of Telephone Numbers, Households, or Individual Women in SWAN 1. Unusablesampling unit (i.e., business or fax telephone number, deceased woman) 2. No contact, unknown usability (i.e., busy signals, answering machine, never home, moved and cannot be traced) 3. Contactmade, unknown cross-sectional eligibility (i.e., hang-up, refusal to be screened for the cross-sectional screening interview) 4. Contactmade, ineligible for the cross-sectional interview 5. Contactmade, cross-sectional eligible, unknown cohort eligibility (incomplete cross-sectional screening interview) 6. Completedcross-sectional screening interview; cohort ineligible or cohort eligible and an attempt was made to recruit into the cohort 7. Completedcross-sectional screening interview; cohort eligible but no attempt was made to recruit into the cohort (i.e., "filtered" at the point of cohort recruitment because of adequate recruitment) 8. Cross-sectional eligible but no attempt was made to recruit for crosssectional interview (i.e., "filtered" at point of cross-sectional interview)
182
S O W E R S ET AL.
1. The proportion of eligible women was calculated among women with known eligibility (disposition codes 4 through 8):Pe = (5 + 6 + 7 + 8)/(4 + 5 + 6 + 7 + 8). 2. This proportion then was applied to the total number of sampling units with unknown eligibility (disposition codes 2 and 3), to estimate the potential number of eligible women among those with unknown eligibility (assuming it was the same as those with known eligibility): pe (2 + 3) = Eu. 3. The total number of women whom sites attempted to recruit for the Cross-sectional Study was computed as the sum of known eligible women recruited (5 + 6 + 7) and the estimated potential number of women among those with unknown eligibility (Eu): (5 + 6 + 7) + E u = R. 4. The participation rate was estimated as the number of women participating in and completing the interview divided by the estimated total number of eligible women whom the site attempted to recruit: (6 + 7)/R. Women known to be eligible but not asked to participate in the cross-sectional interview [because a sufficient number had been recruited (disposition 8)] are included in the calculation of percentage eligible. They were excluded from the calculation of the cross-sectional participation rate because they were not actually recruited to complete the crosssectional study. Table VI gives the distribution of sampling units for the SWAN Cross-sectional Study disposition codes by site and indicates the greater efficiency from those frames containing more eligibility criteria information. The percentage of unusable units sampled (disposition code 1) was lower for list-based sites ( 0 - 1 4 % ) as compared with the sites using primarily RDD-based sampling ( 1 9 - 3 5 % ) . Sites primarily using a list-based frame also had a lower ( < 11%) percentage of sampled units with no contact due to busy signals, answering machines, never being home or moved and inability to trace (disposition code 2) as compared with the RDD-
TABLE VI
based sampling (11-25%). The two list-based sites whose frames had substantial eligibility and recruiting information (an H M O enrollment list and a list from a previous research study) required the fewest number of sampling units (<3500) to reach their cohort targets. In contrast, those sites whose frames did not include information about the eligibility criteria information sampled 24,283 to 78,914 sampling units to meet their specific cohort targets. The percentage of sampled units with contact made but unknown cross-sectional eligibility (i.e., hang-ups and disposition code 3) varied widely across sites, but was not related to RDD-based versus list-based frame status. The percentage of known ineligible units (disposition code 4) did not differ greatly across sites, with the exception of the Detroit area, where the investigators were conducting a census of all households and were not allowed into three apartment dwellings. The percentage of eligible women who began but did not complete an interview (disposition code 5) was uniformly low across all sites (less than 3%).
E
Response Rates
A total of 202,985 sampling units were screened for potential participation in the SWAN Cross-sectional Study. Of these, 34,985 were defined as eligible and 16,065 completed the interview, for an overall response rate of 46.6%. Of these, 6521 women were cohort-eligible and were asked to participate in the SWAN Longitudinal Study; a total of 3306 women entered the cohort, for an overall response rate of 50.7% (Table III). Response rates did not vary statistically by age, marital status, parity, or menopausal status (see Table VII) whether considering the overall group or just non-Hispanic Caucasian women. However, women with a high school education or less (response rates of 37.3 and 40.8%, respectively) were
The Number and Percentage of Sampling Units Used in the SWAN Cross-Sectional Study by Site and Common Disposition Codes Number (percentage) per site"
Disposition code 1 2 3 4 5 6 7 8 Site total a
Boston (L)
Chicago(L)
Detroit(L)
Los Angeles (R)
Newark(R)
Pittsburgh(R)
Oakland(L)
Totalacross sites
2517 (13.6) 9524 (51.3) 1188 (6.4) 2778 (15.0) 328 (1.8) 2233(12.0) 0 (0.0) 0 (0.0)
26 (1.1) 43 (1.8) 203 (8.3) 742 (30.3) 43 (1.8) 1393(56.8) 0 (0.0) 0 (0.0)
1006 (4.1) 2587 (10.6) 2504 (10.3) 14,937(61.5) 662 (2.7) 2551 (10.5) 36 (0.1) 0 (0.0)
10,804(24.8) 7909 (18.1) 4749 (10.9) 17,360(39.8) 104 (0.2) 2242 (5.1) 0 (0.0) 446 (1.0)
15,504(19.6) 19,636(24.9) 21,249(26.9) 18,774(23.8) 261 (0.3) 3490 (4.4) 0 (0.0) 0 (0.0)
11,239(35.3) 3540 (11.1) 1552 (4.9) 12,027(37.8) 596 (1.9) 2604 (8.2) 0 (0.0) 278 (0.9)
8 (0.2) 308 ( 9 . 3 ) 434 (13.1) 1025 (30.9) 25 (0.8) 1516(45.7) 0 (0.0) 4 (0.1)
41,100(20.2) 43,547(21.4) 31,879(15.7) 67,643(33.3) 2019 (1.0) 16,029 (7.9) 36(0.02) 728 (0.4)
2450
24,283
43,614
78,914
31,836
3320
18,568
L, List based; R, RDD based. Numbers in parentheses are percentages.
202,985
183
CHAPTER 11 SWAN Cohort Study
TABLE V I I
Longitudinal Cohort Percentage Participation among Cohort-Eligible Women a Overall sample participating in cohort (%)
Caucasians participating in cohort (%)
Overall
50.1 (48.9-51.4)
48.0 (46.2-49.7)
Age (years) 42-45 46-49 50-52
49.4 (47.7-51.1) 51.9 (49.9-53.8) 47.4 (43.7-51.1)
47.3 (44.9-49.8) 49.5 (46.7-52.4) 45.1 (39.8-50.4)
Education Less than high school High school degree Some college College degree Postcollege
40.8 37.3 52.4 56.1 62.1
24.7 29.7 51.0 52.5 60.8
Smoking Never smoked Past smoking Current smoking
51.6 (49.9-53.2) 52.2 (49.8-54.8) 42.3 (39.6-45.0)
49.8 (47.3-52.3) 51.7 (48.5-54.9) 38.1 (34.3-41.9)
Difficulty in paying for basics Very hard Somewhat hard Not hard at all
42.5 (38.9-46.2) 47.0 (44.8-49.1) 53.5 (51.9-55.2)
38.6 (32.2-45.0) 44.2 (41.0-47.5) 50.6 (48.4-52.8)
Marital status Never married Married/living as married Separated/widowed/divorced
50.1 (46.8-53.4) 51.0 (49.5-52.5) 47.7 (45.1-50.3)
47.3 (42.3-52.3) 48.7 (46.6-50.8) 45.6 (41.5-49.8)
Parity 1 + children No children
49.8 (48.4-51.1) 52.0 (49.0-55.0)
47.1 (45.2-49.1) 50.6 (46.9-54.2)
Race/ethnicity African-American Caucasian Chinese Hispanic Japanese
54.2 48.0 69.2 34.1 63.1
(51.9-56.6) (46.2-49.7) (64.4-73.9) (30.8-37.5) (58.6-67.6)
Site Boston Chicago Detroit Los Angeles Newark Oakland Pittsburgh
48.9 73.7 58.9 53.1 29.6 67.3 42.8
(45.7-52.2) (70.1-77.2) (55.7-62.1) (49.9-56.3) (27.2-32.1) (63.8-70.8) (39.8-45.7)
54.9 76.5 59.7 44.1 23.4 65.2 41.4
(50.2-59.6) (71.4-81.6) (54.7-64.8) (39.6-48.5) (19.9-26.9) (60.0-70.4) (37.8-45.0)
Self-reported health Excellent Very good Good Fair Poor
52.5 52.9 48.8 44.1 40.5
(49.8-55.2) (50.9-55.0) (46.5-51.0) (40.7-47.5) (32.7-48.3)
50.8 48.5 45.4 43.6 38.1
(47.5-54.1) (45.8-51.2) (41.8-48.9) (36.5-50.7) (23.2-53.0)
Menopause status Premenopausal Early perimenopausal
49.4 (47.7-51.0) 51.1 (49.3-52.9)
Subject characteristic
(36.8-44.9) (34.9-39.7) (50.2-54.6) (53.2-59.0) (59.3-64.9)
(15.9-33.5) (26.4-33.0) (47.7-54.2) (48.6-56.4) (57.4-64.2)
47.4 (45.0-49.8) 48.5 (45.9-51.0)
apercentages are +95% CI. Participation is according to sociodemographic and lifestyle characteristics in the full sample (n = 6445) and restricted to Caucasians only (n = 3170). The 95% confidence intervals allow the comparison of frequency within each sociodemographic or lifestyle variable; statistically significant differences are shown in bold.
less likely to participate than women with some college education (response rate of 52.4%). Women with a postcollege educational experience were the most likely to participate and had a response rate of 62.1%. Likewise, those women who reported that it was somewhat hard or very hard to pay for basics (i.e., food, shelter and health care) were less likely to participate. Their response rates were 47.0 and 42.5%, respectively, as compared to the response rate of 53.5% for women who reported no difficulty in paying for basics. As reported in Table VII, women who currently smoked were significantly less likely to participate than women who had never smoked or who had quit smoking. Participation varied according to self-reported race/ethnicity status, with Chinese (69.2%) and Japanese (63.1%) being the most likely to participate followed by AfricanAmerican (54.2%), non-Hispanic Caucasian (48.0%), and Hispanic women (34.1%). Because Chinese, Japanese, and Hispanic women were recruited at single sites, the degree of participation might be representative of other site characteristics rather than race or ethnicity. The site differences in response rates, ranging from 29.6% at Newark to 73.7% in Chicago, were believed to reflect, in part, differences in site recruitment strategies. Therefore, we evaluated response rates based on whether the site's recruitment was primarily a list-based or an RDD-based approach (see Table VIII). The response rate for sites whose primary frames were list based was 60.7% compared with a 40.3% response rate for RDD sites. Table VIII shows that although sites that relied primarily on the RDD-based approach had lower percentage participation as compared to participation at list-based sites, similar characteristics were likely to be associated with nonparticipation. Irrespective of recruitment strategy, women with less education and women who smoked or had difficulty in paying for basics were less likely to participate. Importantly for this study, response rates according to menopause status did not vary by recruitment strategy.
V. STRENGTHS AND LIMITATIONS OF SWAN SWAN has successfully recruited and enrolled a large community-based sample of women with substantial representation from five racial/ethnic groups. To do so, SWAN incorporated a wide range of sampling frames and diverse sampling approaches, including the extensive use of listbased frames and supplemental frames or supplemental information to the primary frames. Theoretically, it would have been desirable to select participants as a national probability sample so that the study findings could be generalized to the national population of midlife women, similar to the National Health and Nutrition Examination Surveys (NHANES) [7]. However, the demands of efficiently implementing a longitudinal study that requires intensive and ongoing (annual and
184
SOWERS ET AL. TABLE VIII Longitudinal Cohort Percentage Participation According to Sociodemographic and Lifestyle Characteristics a
Subject Characteristics Overall Age (years) 42-45 46-48 50-52 Education Less than high school High school degree Some college College degree Postcollege Smoking Never smoked Past smoking Current smoking Difficulty in paying for basics Very hard Somewhat hard Not hard at all Marital status Never married Married/living as married Separated/widowed/divorced Parity 1+ children No children Race/ethnicity African-American Caucasian Chinese Hispanic Japanese Self-reported health Excellent Very good Good Fair Poor Menopause status Premenopausal Early perimenopausal
At list-based sites
At RDD-based sites
60.7 (59.0-62.4)
40.3 (38.7-42.0)
59.9 (57.4-62.3) 63.4 (60.7-66.1) 55.3 (50.2-60.5)
40.1 (37.8-42.4) 41.0 (38.3-43.7) 39.1 (33.9-44.3)
48.7 (41.8-55.7) 44.8 (41.0-48.6) 62.8 (59.7-65.9) 69.0 (65.1-72.8) 71.6 (68.1-75.0)
36.6 (31.7-41.5) 31.5 (28.4-34.6) 42.7 (39.7-45.8) 43.5 (39.5-47.6) 49.6 (45.2-54.0)
63.3 (60.9-65.7) 63.1 (59.6-66.5) 51.8 (47.9-55.7)
42.2 (39.9-44.4) 41.6 (38.1-45.1) 33.3 (29.7-36.8)
58.5 (52.6-64.4) 56.5 (53.3-59.6) 63.5 (61.3-65.7)
32.4 (27.9-36.8) 38.8 (35.9-41.6) 43.3 (41.0-45.6)
57.2 (53.0-61.4) 63.4 (61.2-65.6) 56.9 (53.3-60.6)
38.5 (33.2-43.7) 41.4 (39.4-43.4) 37.5 (33.8-41.2)
60.4 (58.5-62.3) 62.1 (58.0-66.2)
40.0 (38.3-41.9) 41.5 (37.2-45.8)
56.4 (53.8-59.1) 62.7 (60.2-65.2) 69.2 (64.4-73.9) -m
45.7 (40.4-50.9) 36.4 (34.1-38.6) 34.1 (30.8-37.5) 63.1 (58.6-67.6)
64.7 (61.0-68.4) 65.4 (62.5-68.3) 56.3 (53.1-59.4) 55.6 (50.7-60.5) 46.4 (34.5-58.2)
41.1 (37.4-44.8) 41.9 (39.1-44.7) 41.4 (38.3-44.5) 33.6 (29.1-38.0) 35.7 (25.4-46.0)
61.4 (59.0-63.8) 60.0 (57.5-62.4)
39.3 (37.1-41.5) 41.8 (39.2-44.3)
apercentages are +_95%CI. Data are from cohort-eligible women with sites categorized as using list-based sampling (n = 3099) or using RDDbased sampling (n = 3346). RDD-based sites were Los Angeles, Newark, and Pittsburgh locales; list-based sites were Boston, Chicago, Detroit, and Oakland locales. The 95% confidence intervals allow the comparison of frequency within each sociodemographicor lifestyle variable, with statistically significant differences shown in bold.
monthly) clinical data collection could not be met with a national probability sample. Typically, exclusive reliance on list-based frames has been viewed with caution by survey researchers because of concerns about inadequate coverage (and attendant nonrepresentativeness) or the inadequacy of information related to eligibility (and attendant inefficiency). However, in SWAN, the response rate was higher at those sites using lists as their primary sampling frame and the response characteristics (i.e., education level, menopause status) were similar for both list-based and RDD-based sites. Potentially, this is attributable to the combining of several lists prior to sample selection or the use of multiple lists to provide supplemental information to facilitate recruitment after samples were selected. Nonetheless, the response rates suggest that bias from inadequate coverage by list-based frames was no different than the bias that might have resulted from the RDD-based frames. SWAN identified supplemental approaches to use with random digit dialing sampling in small geographic areas because the technique is less efficient in small areas than in large areas [8]. These supplemental approaches helped overcome the widely recognized disadvantages of telephone sampling frames, i.e., approximately 8% of the population in the United States does not have current telephone service, and this rate varies widely depending on the socioeconomic status [9,10]. The degree to which using list-based RDD ameliorated this issue is unknown; however, future studies may need to add four to six lists and use list-assisted RDD. These SWAN recruiting efforts demonstrate the need to have a carefully considered sampling strategy that incorporates flexible approaches and numerous frames. No methodology is clearly right or wrong; nonetheless, it is obvious that frames with the most information related to eligibility criteria were the most efficient. This study adds to our understanding of the complexities of sampling and the need for multiple methods of recruiting to identify information applicable to a broadly diverse population. The importance of SWAN's recruitment strategies can be most appreciated within the context of ethnicity and culture. Almost no studies have simultaneously and comparatively investigated the prevalence of menopausal symptoms in multiethnic/multicultural populations, although there are reported cultural/ethnic differences in symptoms, age at menopause, and bleeding characteristics [ 11 ]. It is unclear if these differences (if there are true differences) are due to hormonal differences (i.e., lower serum estradiol and testosterone levels) or other physiological differences (i.e., differences in immune responses). The differences could readily be associated with cultural perceptions of the menopause and aging, dietary practices, physical activity, body composition (more fat mass or less lean mass), or reproductive practices reflected in parity or age at first conception. Furthermore, the
CHAPTER 11 SWAN Cohort Study role of culture and acculturation (the process of incorporating the customs, norms, identification, and social and working activities from different societies in shaping health status, health belief, and health behaviors) has not been widely applied to the menopausal transition. SWAN has also incorporated data-gathering approaches consistent with major theories and approaches to the menopause [6]. Thus, the biological theory ascribes the experience of the menopause to alterations in metabolism and endocrine status. Typically, with this theory, there is a pronounced focus on ovarian function and, in some, attendant focus on hormone replacement or alternative hormone sources. The psychological/psychosocial approach considers the importance of stressors and losses, particularly as catalysts for symptoms. The approach then suggests the need for social supports, well-developed relationships, and coping skills for a "successful" transition. The sociocultural/environmental approach suggests that culture frames our behaviors and attitudes toward the menopause. The external environment (i.e., passive smoking, occupational exposures, or workplace demands) modifies our biological assets and, in doing so, frames our responsiveness to the transitional events. Finally, feminist theory, with its views that the menopause is a normal developmental stage, urges an understanding of how women achieve control of the experience and become active participants, addressing the challenges of symptoms. The information-gathering activities of SWAN have reflected an inclusive approach that acknowledges the value of each of these to a complex process.
VI. SUMMARY The of SWAN study employs a prospective design that includes sufficient pre- and postmenopausal observations to ensure the separation of menopause-related and age-related physiological changes. Other attributes include the comprehensive standardized data collection related to biological, behavioral, physiological, social, environmental, and cultural factors; specialized data collection methodologies suitable to address the monthly and yearly variation in behavioral and biological patterns; generalizability to community-dwelling populations recruited from major United States population centers; sufficiently large sample sizes and numbers of data points to ensure reliable estimates of associations and relevant effect sizes; and inclusion of sufficient numbers of racial/ethnic minorities to provide comparative information with the non-Hispanic Caucasian population. Because of these attributes, SWAN can contribute new and substantive knowledge about women's health in general and the menopause transition in particular. SWAN is the first national study to describe women at midlife, an understudied age group. Its multidisciplinary approach provides the opportu-
185 nity to consider the contributions of both culture and biology so that we may better understand health in American women.
APPENDIX A. SWAN INVESTIGATORS Clinical Sites Boston Principal Investigator: Robert Neer, M.D. Coprincipal Investigator: Joel Finkelstein, M.D. Coinvestigators: Josh Alexander, Ph.D.; Andrew Arnold, M.D.; David MacLaughlin, Ph.D.; Richard Pasternak, M.D. Biostatistician: David Schoenfeld, Ph.D. Project Manager: Tracy Thomas, B.A. Chicago Principal Investigator: Lynda Powell, Ph.D. Coprincipal Investigator: Denis Evans, M.D. Biostatistician: Peter Meyer, Ph.D. Project Manager: Diedre Wesley Data Manager: Gerard Kaszubowski Detroit Principal Investigator: MaryFran Sowers, Ph.D. Coprincipal Investigators: Sioban Harlow, Ph.D.; John Randolph, M.D. Coinvestigators: Carolyn Sampselle, Ph.D.; Nancy Reame, Ph.D. Biostatisticians: Roderick Little, Ph.D.; M. Anthony Schork, Ph.D. Project Manager: Vanessa Harris, M.P.H. Data Managers: Ruth Sanchez-Pena, M.S.; Gavin Welch, M.P.H. Los Angeles Principal Investigator: Gail Greendale, M.D. Coprincipal Investigator: Stanley Korneman, M.D. Project Manager: Miriam Schocken, Ph.D.
Newark Principal Investigator: Gerson Weiss, M.D. Coprincipal Investigator: Nanette Santoro, M.D. Biostatistician: Joan Skurnick, Ph.D. Project Manager: Ann Reinert Data Manager: Pat McTerrell Oakland Principal Investigator: Ellen Gold, Ph.D. Coprincipal Investigator: Barbara Sternfeld, Ph.D. Coinvestigators: Barbara Abrams, Ph.D.; Shelley Adler, Ph.D.; Gladys Block, Ph.D.; Maradee Davis, Ph.D.; Bruce Ettinger, M.D., William Lasley, Ph.D.; Marion Lee, Ph.D.; Helen Schauffler, Ph.D.; Barbara Sommer, Ph.D. Biostatistician: Steven Samuels, Ph.D.
1
8
6
S
O
W
Project Manager: Sarah Rowell, M.S. Data Manager: Marianne O'Neill Rasor, M.A.
Pittsburgh Principal Investigator: Karen Matthews, Ph.D. Coprincipal Investigator: Jane Cauley, Ph.D. Coinvestigators: Joyce Bromberger, Ph.D.; Charlotte Brown, Ph.D.; Kim Sutton-Tyrell, Ph.D.; Sidney Wolfson, M.D. Data Manager: Nancy Remaley, M.S.I.S.
Coordinating Center Principal Investigator: Sonja McKinlay, Ph.D. Coprincipal Investigator: Sybil Crawford, Ph.D. Project Directors: Juli Bradsher, Ph.D.; Kay Johannes, Ph.D. Project Manager: Patricia McGaffigan Data Coordinator: Beth Willis
Laboratory~University of Michigan Principal Investigator: Rees Midgley, M.D. Coprincipal Investigator: Daniel McConnell, Ph.D. Coinvestigator: Barry England, Ph.D. Laboratory Manager: Kimberly Gonzalez, M.T. Systems Analyst: Mark Davis, B.S.
Laboratory~Medical Research Laboratory (MRL) Principal Investigator: Evan Stein, M.D. Coinvestigator: Paula Steiner
Steering Committee Chair
E
R
S
E T AL.
phone numbers were identified through individual contact. The sampling plan was implemented initially by New England Research Institutes, Inc. and subsequently by the California Survey Research Services.
B. Chicago Site The Chicago Health and Aging Project (CHAP) database, a census of all residents initiated prior to SWAN, served as the sampling frame for the Morgan Park and Beverly neighborhoods in Chicago. The CHAP frame ultimately included the complete name, address, gender, age, and race of individuals. About 1% of records with missing age, gender, or race data on the CHAP frame was excluded from SWAN sampling. The sampling approach involved stratification by race (African-American and Caucasian). A random number generated from a uniform distribution between 0 and 1 was assigned to each woman in the sampling frame. Subjects were paired with random numbers based on their position in the sampling group. A sampling fraction was computed as the required race-specific sample size divided by the racial/ethnicspecific frame size. Women whose random numbers were less than or equal to the sampling fraction were included in the sample, i.e., simple random sampling was employed within each race stratum.
Jennifer Kelsey, Ph.D.
Project Officers National Institute on Aging (NIA): Sherry Sherman, Ph.D.; Marcia Ory, Ph.D. National Institute of Nursing Research (NINR): Carole Hudgings, Ph.D.
A P P E N D I X B. SPECIFIC S A M P L I N G AND R E C R U I T I N G STRATEGIES BY SITES WITH LIST-BASED P R I M A R Y SAMPLING FRAMES A. Boston Site The list-based frame used by the Boston SWAN site used the Spring, 1995 Massachusetts Census from all 22 wards in Boston. This census is updated annually and contains the name, address, gender, and age of the residents, but not their race/ethnicity or telephone numbers. The listed age, however, was not always accurate because this census arbitrarily assigns an age if the actual age is missing. Telephone numbers were obtained for some women from Survey Sampling, Inc., the white pages, and directory assistance. Selected tele-
C. Detroit Area Site A census was conducted of all households in the 40 target Census Block Groups in the Ypsilanti community and the 46 Census Block Groups in the Inkster community (located in the Detroit area). The sampling frame for the Ypsilanti and Inkster communities was based on a household list from the commercial electric utility company. The list contained every household name and address in the geographical area of interest (100% coverage) but did not include gender, telephone number, age, or ethnicity. The probability of selection for each household was 1. Prior to contacting the household sampling units within a given Block Group, the U.S. Census Block Groups were randomly assigned to sampling batches to minimize selection bias. In order to contact households, telephone numbers were obtained from cross-matching with the local telephone listing, a reverse telephone directory, and a commercial listing. About 45% of households were matched with a telephone number. Interviewers contacted those households (face to face) without a telephone number to determine if there was an eligible woman in residence, and, if appropriate, conducted the cross-sectional interview. There were more Caucasian women than African-American women in the census
CHAPTER 11 SWAN Cohort Study
187
area. Therefore, Caucasian women were subsampled from the Cross-sectional Study for the Longitudinal Study at a rate of 25% using an 8-sided die for the last 8 months of recruitment. The sampling plan was developed and implemented by site investigators and staff.
D. O a k l a n d A r e a Site The membership list of the Kaiser Permanente Medical Care Program (KPMCP), which insures approximately 30% of the population in the San Francisco Bay area, acted as the sampling frame for the Oakland, California area [12]. This frame included name, age, gender, address, and telephone number, but not race/ethnicity. From the membership roles, two lists were assembled, one of female members with Chinese surnames and one of female members with non-Chinese surnames [ 13]. All of the Chinese-surnamed women whose home zip codes mapped to either the Richmond, Oakland, or Hayward Kaiser facilities and who were in the appropriate age range were included (n = 2446). The list was randomly ordered using a random number generator and then divided sequentially into batches of 100 and sampled until the cohort recruitment goal of 250 Chinese women was achieved (1400 sampling units required). A similar approach was used for the Caucasian women. In that instance, the list consisted of 4418 women randomly selected from the approximately 47000 nonChinese-surnamed women members of the appropriate age and residing in the same geographic area. Ultimately, 1650 sampling units were sampled to achieve the cohort recruitment goal of 200 Caucasian women. The sampling plan was developed and implemented by site investigators and staff.
APPENDIX
C. S P E C I F I C
SAMPLING
AND R E C R U I T I N G S T R A T E G I E S BY SITES WITH RDD-BASED PRIMARY SAMPLING
a sampling frame that contains all listed households and a significant portion of the unlisted households. Thus, the sampling frame consists of all phone numbers found in any eight-digit sequence that contains at least one listed telephone number. To reduce the number of unprofitable (nonhousehold) calls, but at the cost of some bias caused by removing some telephone households from the frame, the expanded list was developed in an abbreviated fashion in the following way. For each eight-digit sequence in the master list, there was a count of the number of listed numbers. The sampler eliminated any sequence with fewer than a specified number of listed numbers. Some market research houses have used a 2 + or a 3 + standard in which sequences with at least two or at least three listed telephone numbers are retained. The Los Angeles site used a 3+ selection method that eliminated residential household blocks with zero, one, or two listed telephone numbers before initiating random sampling. The screening of households was conducted by California Survey Research Services. Although the Los Angeles SWAN site recruited from census tracts with a higher density of Japanese persons ( 6 - 2 0 % of the population) to meet its recruiting goal, the RDD-based sample required supplementation with lists that were devoted entirely to the recruitment of Japanese women. Thus, the Los Angeles site identified all women aged 4 0 - 5 5 years with a first, middle, or family Japanese name from voter registration lists and sampled 100% of these women. The Los Angeles site also used a frame containing listed telephone numbers with Japanese surnames and sampled 100% of these listings. Additionally, the University of California, Los Angeles (UCLA) site also used snowball sampling. The snowball sample for UCLA consisted of referral by Japanese participants of up to five women without regard to the eligibility of that participant.
FRAMES A. L o s A n g e l e s Site
The RDD samples from South Bay and Sawtelle in the Los Angeles area were created from the RDD frame maintained by Survey Sampling Institute. The list-assisted random digit dialing method combines a number of available phone lists (including white pages, drivers' licenses, and vehicle registrations) into a Master List of the first eight digits of a phone number (the area code, the exchange, and two more digits). This Master List is expanded by a factor of 100 by adding all possible two-digit sequences (00-99) to each Master List entry. This expanded list can be used directly as
B. N e w a r k Site The RDD samples for the Newark area were created from the RDD frame maintained by the Survey Sampling Institute, and the screening of households was conducted by California Survey Research Services. Hudson County in New Jersey was stratified into five areas: Hoboken City, Union City, West New York Township, Jersey City, and the remainder of Hudson County so that census tracts containing higher than average densities of Hispanic households could be oversampled. Random digit dialing was then applied to telephone households in those census tracts. A 3 + selection method was also applied at this site in the same manner as the methodology used at the Los Angeles site. The New Jersey site also used snowball sampling. In New Jersey, snowball sampling involved asking women who
188
SOWERS ET AL.
completed the cohort base line but were ineligible for the cohort to provide the names of up to five w o m e n who were cohort age-eligible and who lived in the target areas in Hudson County.
and samples from the 22 zip codes in which a substantial number of African-American w o m e n were known to reside.
References C. P i t t s b u r g h S i t e The major sampling approach implemented at the Pittsburgh site was RDD. Samples of random telephone numbers for households were generated with probability proportional to size across all nonbusiness telephone exchanges (central office codes, or COCs) and working blocks according to the density of the listed residential telephone numbers in the exchange. Area c o d e - C O C - w o r k i n g block combinations (including the first eight digits of the area code) and exchange were selected systematically whereas the last two numbers of the 10-digit telephone number were randomly generated. This systematic selection of exchanges (COCs) and working blocks provided a self-weighting, equal-probability sample. The first-stage selection of a telephone number represented the selection of a household. If a household was found to contain more than one age-eligible female, a second-stage randomized selection of a female was made using the birthday method. The Pittsburgh sites supplemented their R D D sampling with voter registration lists (VRLs) to improve their capacity to oversample their designated ethnic groups and/or target the age group of interest. The VRLs for Pittsburgh included information on gender, birth date, and address for all registered voters and ethnic identification for about 85% of registered voters in Allegheny County. Telephone numbers for the V R L sample were obtained from the Cole Directory for Pittsburgh and Allegheny County, the white pages, and directory assistance. Several different strategies were used; systematic samples were drawn from the voter registration list, including samples reflecting all of Allegheny County
1. Diczfalusy, E. (1986). Menopause, developing countries and the 21st century. Acta Obstet. Gynecol. Scand., Suppl. 134, 45. 2. U.S. Congress, Office of Technology Assessment (1986). "The Menopause, Hormone Therapy and Women's Health," OTA- Bp-BA-88. U.S. Govt. Printing Office, Washington, DC. 3. Skolnick, A. A. (1992). At third meeting, menopause experts make the most of insufficient data. JAMA, J. Am. Med. Assoc., 268, 2483-2485. 4. Weinstein, M. C., and Tosteson, A. N. A. (1990). Cost-effectiveness of hormone replacement. Ann. N.Y. Acad. Sci. 592, 162-72. 5. NIH Guide (1993). "Menopause and Health in Aging Women," Vol. 22, No. 32. National Institutes of Health, Washington, DC. 6. Barile, L.A. (1997). Theories of menopause. Brief comparative synopsis. J. Psychosoc. Nurs. 35,(2), 36-39. 7. Montaquila, J. M., Mohadjer, L., and Khare, M. (1998). The enhanced sample design of the future National Health and Nutrition Examination Survey (HANES). Proc. Am. Star. Assoc: Sect. Surv. Res. Methods. 8. Giesbrecht, L.H. (1996). Estimating coverage bias in RDD samples with current population survey data. Proc. Am. Stat. Assoc. Sect. Surv. Res. Methods 1,503-508. 9. Mohadjer, L. (1988). Stratification of prefix areas for sampling rare populations. In "Telephone Survey Methodology" (R. M. Groves, P. P. Biemer, L. E. Lyberg, J. T. Massey, W. L. Nicholls, and J. Waksberg, eds.), pp. 161-173. Wiley, New York. 10. Thornberry, O. T., Jr., and Massey, J. T. (1988). Trends in U.S. telephone coverage across time and subgroups. In "Telephone Survey Methodology" (R. M. Groves, P. P. Biemer, L. E. Lyberg, J. T. Massey, W. L. Nicholls, and J. Waksberg, eds.), pp. 25-49. Wiley, New York. 11. Sowers, M. E, and LaPietra, M. (1995). Menopause: Its epidemiology and potential association with chronic diseases. Epidemiol. Rev. 17, 287-302. 12. Krieger, N. (1992). Overcoming the absence of socioeconomic data in medical records: Validation and application of a census-based methodology. Am. J. Public Health 82, 703-710. 13. Choi, B. C. K., Hanley, A. J. G., Holowaty, E. J., and Dale, D. (1993). Use of surnames to identify individuals of Chinese ancestry. Am. J. Epidemiol. 138, 723-734.
~HAPTER 1
Demogr aphics, Environmental Influences, and Ethnic and International Differences in thc Menopausal Experience ELLEN B.
GOLD
Department of Epidemiology and Preventive Medicine, School of Medicine, University of California, Davis, Davis, California 95616
I. Introduction II. Demographic Characteristics III. Ethnic and International Differences
IV. Environmental Influences V. Conclusions References
I. I N T R O D U C T I O N
sition. Therefore, this chapter begins with a discussion of the methodologic issues, and is followed by a review of factors studied to date that have been suspected or shown to affect the nature of the transition.
Although menopause is a universal phenomenon among women, the timing of the onset and the signs and symptoms of the perimenopause, menopausal transition, and final menstrual period are not [1]. Most of our knowledge and perceptions of the experience of menopause are derived from studies largely of white women, and many have been studies of clinic-based, rather than population-based, samples of women. Thus, until recently, much of the picture of the menopause experience may have been affected by the nature of the samples of women studied. In addition, a number of methodologic issues arise, which must be considered in conducting and comparing sutdies of the menopausal tranMENOPAUSE: BIOLOGY AND PATHOBIOLOGY
A. Methodologic Concerns Most studies of the menopausal transition have been cross-sectional, rather than longitudinal, in design, providing opportunity for distortion of the true picture of the menopausal experience, particularly for understanding risk factors that precede, rather than accompany or follow, the menopause transition. Further, definitions of menopause have 189
Copyright 9 2000 by Academic Press. All rights of reproduction in any form reserved.
190 varied depending on number of months of amenorrhea, and scales and time frames for assessing symptoms have varied from study to study. Studies have also varied in whether and which factors have been included in controlling simultaneously for the effects of multiple confounding variables. 1. AGE AT MENOPAUSE
The analysis of age at natural menopause in a number of studies has been calculated only as a simple mean, rather than using the less biased survival or multivariable regression analysis approaches, which include more information and observations, because women are included but withdrawn if they have experienced surgical menopause or are still premenopausal [2]. Also, accuracy of reporting of age at menopause can vary by whether menopause was natural and by duration from time of menopause to time of the interview about menopause [3]. Further, in some studies reporting age at menopause, it is unclear if the age at the final menstrual period is being reported, which appears to be more frequent, or if the age at cessation of menses plus one year of amenorrhea, apparently a more rare occurrence, is what is reported [4]. Finally, in future studies, an accurate picture of the true age at menopause may become even more difficult to discern as more women are prescribed and take oral contraceptives or other hormone replacement medication prior to the final menstrual period. 2. PRESENTATION AND SYMPTOMS OF MENOPAUSE A number of methodologic issues may influence reporting and thus the prevalence of menopausal symptoms observed in different studies. First, a lack in consistency of symptoms included, in scales used to assess them, or in the time frame for assessment may lead to differences in ascertainment. Second, a failure to recognize colloquial or culturally specific expressions for certain symptoms, even for hot flashes (which can be considered a Western colloquialism), may account for some of the differences in prevalence between different populations. An extension of this problem is failure to recognize and address cultural sensitivities in reporting symptoms, which may also be affected by who asks the questions, and may result in underreporting of specific or even most symptoms. Third, most studies do not incorporate hormonal measures and may use inadequate questions about menstrual bleeding so that perimenopausal status cannot be well established. Further, a lack of comparability among studies with regard to the age group of women studied may also be reflected in the resulting differences in prevalence rates of symptoms. Finally, in some studies, age is used as a surrogate for menopausal status, so that status is presumed based on age rather than on menstrual function, thus introducing misclassification of menopause-related symptoms and possible lack of comparability with studies that have asked about menstrual function.
ELLEN B. GOLD
In recent years, more information has begun to appear regarding differences in the timing and presentation of menopause experienced by samples of women of different socioeconomic, ethnic, and lifestyle backgrounds, resulting in a fuller and more varied picture and also greater insights (and questions) regarding the physiology underlying the menopausal experience. This field of investigation will benefit and greater understanding will result if this trend continues, along with increased standardization of instruments and methods so that a clearer picture of the menopausal experience will emerge.
B. S u m m a r y o f U n d e r l y i n g P h y s i o l o g y Although the physiologic changes that accompany the menopause are described in detail elsewhere in this volume (Chapters 1-9), a brief summary is given here to provide relevant context to the signs and symptoms of the menopause experience and to the factors that affect them that are discussed in this chapter. The cessation of menstruation that defines menopause is believed to be due to a cessation of ovulation due to a loss of ovarian follicles, which produce estrogen. This loss results in a number of endocrine changes, particularly the decline in ovarian production of estradiol, the most biologically active form of estrogen [5,6], as well as increased circulating concentrations of follicle stimulating hormone (FSH) and decreased concentrations of inhibin, which normally inhibits the release of FSH [5]. Age at menopause is thought to be more sensitive to varying rates of atresia of ovarian follicles [7] than to the absolute number of oocytes depleted [8]. It has been widely held that as circulating estrogen concentrations decline in the perimenopause, variations in the timing of menstrual bleeding and in the nature of bleeding may occur. Menstruation may occur at irregular intervals due to irregular maturation of residual follicles, with diminished responsiveness to gonadotropin stimulation, or to anovulatory uterine bleeding after estrogen withdrawal without evidence of corpus luteum function [9]. As menstrual cycles become increasingly irregular, uterine bleeding may occur after an inadequate luteal phase or without ovulation or evidence of corpus luteum function [9], usually indicated by a short luteal phase, characteristic of women over age 40 [ 10,11 ]. Such cycles may be associated with insufficient FSH in the follicular phase, in turn resulting in lower luteal-phase estrogen and progesterone secretion. The absence of the corpus luteum, resulting in estrogen secretion (even hyperestrogenicity [10,12]) unopposed by progesterone, may lead to profuse blood loss. Conversely, relatively low concentrations of estrogen that may accompany the menopause transition also may lead to intermittent spotting. Thus, the nature and timing of bleeding may vary both within and between
CHAPTER 12 Demographics, Environment, Race, and Nationality women, but little is known about host, environmental, or lifestyle factors that may affect such variation. Extraovarian sources, such as adipose tissue, convert androstenedione to testosterone and estrone postmenopausally [ 13-16], although postmenopausal estrone production is only about one-third that of premenopausal production [7,17] but may be elevated in perimenopausal women in both the follicular and the luteal phases [10]. This peripheral production of estrone postmenopausally has been related to the amount of body fat [14,18] but has not been consistently related to age, height, or years since menopause. Also, in the perimenopause, FSH concentrations increase [6] without a concomitant increase in luteinizing hormone (LH) [19]. Reduced concentrations of estrogen and progesterone and increases in FSH [ 19] affect the central nervous system [2023] and result in vasomotor instability, leading to the characteristic hot flashes or flushes in some women. Two longitudinal studies have shown that vasomotor symptoms in the peri- and postmenopause are related to serum estrogen levels [20,24]. Hot flashes may also be more prevalent in women who experience irregular menses prior to menopause than in women who experience an abrupt cessation of menses [25]. One other longitudinal study has reported that the prevalence of symptoms appears to peak during the perimenopause transition and to decrease after menopause [26]. In the international literature, some have reported that about 4 0 - 6 0 % of perimenopausal and 6 0 - 8 0 % of menopausal white women experience hot flashes [27], and a substantial majority of these report that they are moderate or severe [28,29]. Discrepancies in the prevalence of hot flashes partially reflect inconsistencies in research methods and study populations
[30]. Other symptoms in hormone-receptive tissues may also occur in the perimenopause and postmenopause [22]. Changes in the vagina and vulva may result in atrophy, pruritus, dryness, bleeding, and dyspareunia. Estimates of the prevalence of vaginal dryness range from 12-34%, depending on the age group of women studied [20,28,31-34]. The embryology of the urinary and genital systems is shared, and the urethral epithelium and submucosa are affected by estrogen [35,36]. Although about a quarter of midlife women report some form of incontinence, its frequency does not appear to be related to menopausal status as determined from menstrual changes or serum FSH or estradiol concentrations, even though rates of incontinence increase during the time of menopause and then decline thereafter [37]. Whereas mood changes and sleep disturbances may also occur at this time, the causal time sequence of vasomotor symptoms, mood changes, and sleep disturbances and the factors that influence their occurrence and/or perception have not been clarified. Thus, for example, it is not known if hormonal changes affect mood and sleep independent of their effect on vasomotor instability.
191 A variety of physical symptoms, such as headaches, joint pain, aches in the back of the neck and shoulders, constipation, and dizzy spells have been thought to increase during the peri- and postmenopausal years [38]. However, the empirical data are inconclusive in identifying which, if any, are more prevalent during different stages of the menopause transition. For example, some investigators have found that women report significantly more frequent joint pain and dizzy spells when perimenopausal than they did when premenopausal [24], and others have found a greater proportion of peri- and postmenopausal women reporting aches and pains, but not dizzy spells, compared to premenopausal women [39]. Still others have found no association between any somatic symptoms and menopausal status [40]. Although some factors are known to be associated with early age at menopause and risk of experiencing symptoms, the relation of many have not been examined, most have not been examined in relation to duration of the perimenopause, and the endocrinologic effects of known risk factors in perimenopausal women still remain to be adequately explored.
II. D E M O G R A P H I C CHARACTERISTICS A. A g e at N a t u r a l M e n o p a u s e a n d O n s e t o f the P e r i m e n o p a u s e Age at natural menopause has traditionally been defined as the age at the final menstrual bleeding, which is followed by at least 12 months of amenorrhea. Some researchers have suggested that the age at which natural menopause occurs may be a marker of aging and health [41-43]. Crosssectional data indicate that endocrine changes characteristic of the onset of the perimenopause begin at around 45 years [44]. The median age at menopause in white women from industrialized countries is between 50 and 52 and at perimenopause is 47.5 years [26,45-48], with slight evidence of increasing age over time [48-51 ]; these onsets may vary by race/ethnicity (see Section III,A) and may be affected by lifestyle factors (discussed in Section IV,A). 1. SOCIOECONOMIC STATUS
Lower social class, as measured by a woman's level of education completed or by her own or her husband's occupation, has been associated in more than one study with an earlier age at menopause [46-48,52]. One study found that education was more strongly associated than occupation [47]. Most studies that have examined the relation of marital status have found that single women have menopause at an earlier age and that this association cannot be explained by nulliparity [47,53,54].
192 2. MENSTRUAL AND REPRODUCTIVE HISTORY Age at menopause may be a marker for hormonal status or changes earlier in life [55]. In a landmark longitudinal study of largely white, well-educated women, those whose median menstrual cycle length between the ages of 20 and 35 years was less than 26 days were reported to have menopause 1.4 years earlier than women with cycle lengths between 26 and 32 days, whereas a later natural menopause (mean = 0.8 year later) was observed in women with cycle lengths of 33 days or longer [56]. In addition, variability in cycle length of 9 or more days was also associated with a later age at menopause in this and other studies [47,57], although one early study reported an earlier menopause in women with irregular menses [48]. Increasing parity, particularly in women of higher socioeconomic status (SES), has also been associated with later age at menopause [45-47,50,52,54,55,58], consistent with the theory that menopause occurs after sufficient depletion of oocytes [58]. Although some studies report no familial relationship, one study has reported that age at menopause is positively associated with maternal age at menopause [52], and one study has shown genetic control of age at menopause in a study of twins [59]. Age at menarche has been fairly consistently observed not to be associated with age at menopause, after adjusting for parity and cycle length [47,48,50,53,60-62], as has prior spontaneous abortion, age at first birth, or history of breastfeeding [47,61,62]. Women who have used oral contraceptives (OCs) have also been reported to have a later age at menopause [47,52,62,63], an observation that is also consistent with the theory that OCs delay depletion of oocytes. However, the finding is not wholly consistent across studies, because one study reported that this delay became nonsignificant after a time-dependent adjustment for when OCs were used [47], and another study reported that OC users had a significantly earlier natural menopause than did nonusers, although this association was not consistent across 5-year age groups [45].
B. P r e s e n t a t i o n a n d S y m p t o m s o f M e n o p a u s e 1. SOCIOECONOMIC STATUS Although the majority of menopausal white women report vasomotor symptoms (hot flushes or flashes or night sweats), the prevalence varies greatly by socioeconomic status. Estimates of the incidence of hot flushes in menopausal white women range in population studies in the United States and worldwide from 24 to 93% [27]. Less educated women report more hot flashes and irritability compared to more educated women [39,64-68]. One large cross-sectional study reported increased prevalence of all symptoms associated with difficulty in paying for
ELLEN B. GOLD
basics [49]. One relatively small cross-sectional study found that women who reported mood changes or irritability that they believed were related to the menopausal transition were significantly more likely to report more other everyday complaints [69]. This was in contrast to women who reported hot flashes, sweating, or headaches associated with the menopausal transition who did not report more other everyday complaints. Homemakers have been shown to report hot flashes for a longer period compared to employed women [70], although working women of lower SES report more stress and tension during menopause [70,71 ], and worsening work stress has been associated with increased reporting of vasomotor symptoms, general health symptoms, and sexual difficulties [39]. 2. MENSTRUAL AND REPRODUCTIVE HISTORY The relationship of menstrual characteristics to the probability of experiencing menopausal symptoms largely remains unexplored, and reproductive history has been examined in a few studies but with somewhat inconsistent results. One cross-sectional study reported that women having natural menopause before age 52 years had a significantly greater reporting of hot flashes [66]. In another study, multiparous women had a lower prevalence of hot flashes compared to nulliparous women or women with abrupt cessation of menses [25]. However, another study showed menopausal symptoms to be associated positively with increasing parity [72], and some studies report no differences in symptom reporting frequency by parity [67,73,74]. Two longitudinal studies and one retrospective study have shown that reporting of menopausal vasomotor symptoms was more frequent among women who reported experiencing premenstrual tension before menopause [64,75,76], a finding that may be related to higher FSH levels in menstruating women with hot flushes [77]. One of these studies also reported that women with vasomotor symptoms were significantly more likely to report that their mothers also had vasomotor symptoms than were women without symptoms [75]. Most studies show symptoms to be more prevalent in hysterectomized women [39] and among women who experience an early menopause. In summary, later age at menopause may be a marker of health and longevity [41-43]. Studies have fairly consistently shown that lower socioeconomic status [46-48, 50,52], single marital status [47,53,54], regular menstrual cycles [47,57], nonuse of oral contraceptives [47,52,62,63], and lower parity [45-47,50,51,54,55,58] are associated with earlier menopause. Age at menarche [47,48,50,53,60-62], prior spontaneous abortion, age at first birth, and prior breastfeeding [47,61,62] are not associated with age at menopause. Lower socioeconomic status [39,64-68] and history of premenstrual tension [64,75,76] are associated with greater menopausal symptom reporting. However, the relation of parity to prevalence of symptoms has been inconsistent across studies.
CHAPTER 12 Demographics, Environment, Race, and Nationality III. ETHNIC
AND
INTERNATIONAL
DIFFERENCES
A. A g e at N a t u r a l M e n o p a u s e a n d at O n s e t o f P e r i m e n o p a u s e 1. ETHNIC DIFFERENCES
African-American [57] and Latina [78] women have been observed to have natural menopause about 2 years earlier than white women, despite their increased average body mass relative to white women (see Section IV,A,2). However, one small study in Nigeria reported the average age at menopause to be 52.8 years [79], nearly 2 years higher than that generally reported for white women in industrialized nations. Mayan women, despite their high parity, have been reported to experience menopause at about age 45 years [80]. Further, Mexican-American women may have shorter bleeding periods and follicular phase lengths [81]. In contrast, Asian and Caucasian women tend to be of similar age at menopause [82], although Thai women have been reported to have a lower median age at menopause (49.5 years), despite their high parity (see Section II,A,2) [60], and Filipino Malay women have been reported to have an average age at menopause of 47-48 years [83]. 2. INTERNATIONALDIFFERENCES A number of reports tend to indicate that women living in developing countries (including Indonesia, Singapore, Pakistan, Chile, and Peru) experience menopause several years earlier than do those in developed countries [63,84-87]. Some work has also indicated that women living in urban areas have a later menopause than do women in rural areas [88]. Women living at high altitude in the Himalayas or in the Andes of Peru have been shown to undergo menopause 1-1.5 years earlier than those living at lower altitudes or in less rural areas [63,89-91 ]. It is unclear if these geographic differences in the age at natural menopause reflect socioeconomic, environmental, racial/ethnic, or lifestyle differences, and whether and how these affect physiology.
B. P r e s e n t a t i o n a n d S y m p t o m s o f M e n o p a u s e
193 women. Among Filipino Malay women aged 4 0 - 5 5 years, reporting of vasomotor or circulatory symptoms occurred in 63%, and nervous or psychological symptoms (particularly headache or irritability) were reported by 79%, although only 31% consulted a physician, a rate that was higher among women with a vocational or college education [83]. It is unclear whether these ethnic differences in symptom frequencies are due to differences in cultural perceptions of menopause and reporting symptoms [ 100], diet [ 101 ], physical activity or body mass [102], differences in use of herbs or plant-estrogen-containing products [103], or in the use of acupuncture (which lowers excretion of the vasodilating neuropeptide calcitonin gene-related peptidelike immunoreactivity) [ 104] between Asian and Caucasian women, or to differences in serum estradiol levels (lower in Asian women in relatively nonsystematic studies that did not indicate adequate control of the day of the menstrual cycle on which blood was drawn for estrogen assays) [ 102,105]. Mayan women report no hot flashes [80], despite hormone profiles similar to those of Western women [ 106]. On the other hand, African-American and Hispanic women have been reported to have a higher prevalence of vaginal dryness compared to Caucasian women [20,49,67]. Some researchers believe that differences in the prevalences of symptom reporting reflect negative cultural stereotypes of aging and of the menopause experience [ 107,108] and are related to mental health [32]. However, others believe that because some studies report a frequency of hot flashes, night sweats, and vaginal dryness in countries such as Indonesia and Southeast Asia, for example, similar to that seen in Western countries, the latter view may be too simplistic [109]. Rather, cultural values of menopause as well as climate, dietary habits, and lifestyle may also be related. In summary, ethnicity appears to be related to both age at menopause and symptom reporting. African-American [57] and Latina [78] women have an earlier menopause than do Caucasian or some Asian [83] women, although not all Asian women [60,83]. Women in less developed countries also experience menopause earlier [63,84-87]. Additionally, Mayan [80] and Asian [60] women report fewer hot flashes, whereas African-American and Hispanic women have a higher prevalence of vaginal dryness than do Caucasian women [20,49,67].
1. ETHNIC DIFFERENCES Although the majority of Caucasian women experience menopausal symptoms [27,29], the reported frequency is much lower in most Asian women that have been studied [29,49,92-97], although one retrospective study reported no difference in symptom prevalence between Japanese and Caucasian menopausal women in Hawaii [98]. Further, some estimates of the prevalence of hot flashes have varied in similar Asian populations, e.g., from 23 [60] to 69% [99] in Thai
IV. E N V I R O N M E N T A L
INFLUENCES
A. A g e at N a t u r a l M e n o p a u s e a n d O n s e t o f the P e r i m e n o p a u s e 1. SMOKING
Perhaps the single most consistently shown (micro) environmental effect on menopause is that women who
194 smoke stop menstruating 1 to 2 years earlier than comparable nonsmokers [45,46,50,52,57,110-114] and have a shorter perimenopause [26]. In some studies heavy smokers have been observed to have an earlier menopause than light smokers, suggesting a dose-response effect of smoking on atrophy of ovarian follicles [52,113-117]. However, former smokers have only a slightly earlier age at menopause than never smokers, and increased time since quitting diminishes the difference [ 115,118], suggesting a reversible effect. The polycyclic aromatic hydrocarbons in cigarette smoke are known to be toxic to ovarian follicles [ 119,120] and thus could result in premature loss of ovarian follicles and early menopause in smokers, although the fact that former smokers have only a slightly earlier menopause than nonsmokers is not wholly consistent with this, even though the latter could reflect a duration and thus a dose effect. Because drug metabolism is enhanced in smokers [ 121 ], estrogen also may be more rapidly metabolized in the livers of smokers, possibly leading to an earlier decline in estrogen levels [122]. Smoking has also been observed to have antiestrogenic effects [123]. Greater prevalence of hysterectomy among premenopausal smokers than nonsmokers [115,124] does not appear to account for the earlier menopause in smokers [ 125]. Although one study reported that nonsmoking women whose spouses smoked had an age at menopause resembling that of smokers [ 126], very little is known about the effect of passive smoke exposure on age at menopause. 2. B o o r MASS AND COMPOSITION
A number of studies have examined the relation of body mass to age at menopause, and the findings have been rather inconsistent. Some studies have reported both increased body mass [indicated by weight for height] and upper body fat distribution [indicated by waist-to-hip ratio] to be positively associated with later age at menopause [45,122,124] and increased sex hormone concentrations [127], although other studies report no significant association of body mass with age at menopause [46,47,57,128,129]. Some studies have found a relationship between weight [ 117] or increased upper body fat distribution [ 128] and earlier age at menopause, particularly in smokers. One study reported earlier menopause in women on weight reduction programs or who had gained more than 26 pounds between the ages of 20 and 45 years [57]. Some of these discrepant findings may be explained by differences in study design (cross-sectional or retrospective vs. prospective) and/or analysis (e.g., inadequate or varying control of confounding variables and/or survival analysis vs. comparison of crude means). In general, the better designed and analyzed studies show no relationship. Although body mass and composition may be related to age at menopause and risk of developing symptoms, they are also related inversely to physical activity, alcohol consumption, and education and positively related to infertility and parity [130]. Further research is needed to examine the independent con-
ELLEN B. GOLD
tribution or interactive effect of body mass and composition and these other factors on the age at and course of menopause, using appropriate longitudinal study design and data analysis techniques that control for the effects of multiple confounding variables simultaneously. 3. PHYSICAL ACTIVITY
Exercise results in changes in a number of endocrine parameters [estradiol, progesterone, prolactin, luteinizing hormone, and follicle-stimulating hormone], both during and after intense physical activity [131-133], with concentrations of these hormones tending to be lower at rest [131,132,134]. Athletes experience a later age at menarche and increased incidence of anovulation [135] and amenorrhea [136] and, in those who menstruate, a shortened luteal phase and reduced mean and peak progesterone levels [130,134]. Although exercise is associated with decreased concentrations of reproductive hormones and frequency of ovulation, few studies have examined the effect of exercise on age at menopause, although one study of modest size has reported no relationship [57]. 4. OCCUPATIONAL]ENVIRONMENTAL FACTORS
Almost nothing is known about the effects of occupational or other environmental factors on age at and course of menopause, although occupational exposures and stressors [such as shift, hours worked, and hours spent standing and heavy lifting] have been shown to increase risk of adverse pregnancy outcomes [137-140] and to affect menstrual cycle length and variability and fecundability [141-144]. In addition, a number of environmental exposures, such as to DDT and polychlorinated biphenyls, have been shown to have estrogenic activity and may be associated with an increased risk of breast cancer [145,146], although this association has not been consistently observed [147,148]. Thus, it is reasonable to expect that occupational and environmental exposures may be related to endocrine disruption that is reflected in altered age at menopause. It is estimated that 40 million women in the United States alone, and several hundred million worldwide [149], will experience the menopausal transition in the next two decades, due to the aging of the "baby boomer" generation [150]. Approximately 70% of American women have worked outside the home [151 ]. Thus, this period in reproductive epidemiologic research presents a prime opportunity to learn more about the effects of occupational and environmental exposures on the menopause transition in these women. 5. DIET
A study from Papua New Guinea has suggested that malnourished women have cessation of menses about 4 years earlier compared to well-nourished women [152], consistent with other studies showing that women with greater weight [88,117] and height [53] may have a later age at menopause.
CHAPTER 12 Demographics, Environment, Race, and Nationality Vegetarians have also been observed to have an earlier age at menopause in one report [ 153]. Inclusion of meat in the diet of vegetarians has been observed to increase the episodic releases of LH and FSH and the length of the menstrual cycle [154]. Thus, meat may modify the interaction of hormones along the hypothalamic-pituitary-ovarian axis. At least one study has reported that increased meat or alcohol consumption is significantly associated with later age at menopause after adjusting for age and smoking [52]. Dietary fiber (whose intake tends to be inversely related to meat intake) may interrupt enterohepatic circulation of sex hormones, leading to the lower estrogen concentrations observed in vegetarian women [ 155]. Premenopausal women administered soy have shown increased plasma estradiol concentrations and follicular phase length, delayed menstruation, and/or suppressed midcycle surges of LH and FSH [ 156]. In postmenopausal women fed soy, FSH and LH did not decrease significantly, nor did sex hormone binding globulin (SHBG) increase, and little change occurred in endogenous estradiol or body weight, although a small estrogenic effect on vaginal cytology was observed [157]. The role of dietary phytoestrogens, fat, protein, and other nutrients in affecting age at menopause and/ or risk and severity of menopausal symptoms in perimenopausal and postmenopausal women remains to be studied systematically.
B. P r e s e n t a t i o n a n d S y m p t o m s o f M e n o p a u s e 1. SMOKING
Smokers have been shown to have lower serum estradiol and estrone concentrations among postmenopausal women [158] and lower urinary estrogen among premenopausal women [159]. These hormonal effects may be related to the findings in the few studies undertaken that smokers report more hot flashes and irritability than do nonsmokers [39,65], as well as more change in sexual desire [66]. Current smoking has been associated with increased reporting of symptoms during the menopausal transition in a number of studies [49,66,68,160]. However, one study reported no significant increase in hot flashes in smokers [67] but showed that thin women who smoked premenopausally had the greatest increase in hot flashes. Further, virtually no information is available about the relation of passive smoke exposure or whether smoking affects severity or frequency of symptoms. 2. BODY MASS AND COMPOSITION Much of the early clinical literature suggested that higher body weight might reduce the probability of experiencing symptoms, particularly hot flashes [73,161,162], due to higher circulating estrogen levels in heavier women due to peripheral production of estrone in adipose tissue. However,
195 most subsequent studies have not reported this [49]. One study showed no relation of body mass index (BMI) to reporting of hot flashes in nonsmokers [67]. Another study reported that women with lower body fat reported more hot flashes [163], and one small study reported significantly higher BMI in women reporting hot flashes, pins and needles, backaches, aches/stiffness in joints, shortness of breath, and fluid retention [20]. Additionally, two population-based studies have reported no significant increase in weight at the menopause [164,165], and one large study reported no increase in waist-to-hip ratio with menopause [ 128]. 3. PHYSICAL ACTIVITY
The effects of physical activity on symptoms reported during the menopause transition are covered in detail elsewhere in this volume (Chapter 34) and thus will only be summarized briefly here for completeness. Serum concentrations of estradiol, progesterone, prolactin, LH, and FSH all tend to increase during and after intense exercise [131-133], whereas resting values tend to be lower in athletes [ 132,134]. The findings from various studies regarding the effect of physical activity on reporting of symptoms, particularly vasomotor symptoms, have been inconsistent, perhaps due to differences in techniques in assessing physical activity and in sample sizes. Midlife women who participate in an exercise program have been reported in some studies to experience less frequent and less severe vasomotor symptoms, despite the fact that lower estrogen concentrations are associated with higher levels of physical activity [166,167]. However, this has not been consistent in other crosssectional or case-control studies, some of which have found no association of physical activity with symptoms [66,168171], and others have found a protective effect [49,172]. Because the onset of hot flashes is accompanied by lower circulating concentrations of plasma fl-endorphins [173], and physical activity increases secretion of endogenous opioid peptides, particularly fl-endorphins [ 174], exercise may prevent symptoms. Exercise also appears to have antidepressant effects [ 175,176] and thus may also be associated with wellbeing and with fewer midlife psychological symptoms, including negative mood and change in sexual desire [66]. 4. DIET
A number of dietary factors are considered to play a role in production, metabolism, and excretion of estrogen, in phases of the menstrual cycle, and in severity of menopausal symptoms. Vegetarian women have been shown to have lower plasma estrone and estradiol concentrations, perhaps due to lower saturated fat intake [177]. Further, Asian women, who consume less fat, excrete two to four times as much estrogen and have substantially lower plasma estrone and estradiol concentrations than do Caucasian women [102,105]. The relation of fat, alcohol, protein, or other nutrient
196 [such as antioxidant] intake to risk of experiencing menopausal symptoms has not been well studied. Nonetheless, some reports have indicated that alcohol may be estrogenic and may contain phytoestrogens [178], and that alcohol intake is inversely associated with levels of SHBG [127, 179,180]. Plant sterols have also been under study with regard to their effects on circulating hormones, menstrual cycles, and menopausal symptoms. Phytoestrogen is a term that includes classes of compounds that are nonsteroidal and either of plant origin or derived from metabolism of precursors in plants eaten by humans [181] (see also Chapter 33). The main classes of compounds are isoflavones and lignans. They structurally resemble estradiol and have been shown to have weak estrogenic activity, compete with estradiol for binding to estrogen receptors in tissues [182,183], and when ingested have estrogenic and antiestrogenic effects, depending on the concentrations of circulating endogenous estrogens and estrogen receptors [ 184,185]. In rats, the most potent of these, coumestrol, suppressed estrous cycles but did not behave as a typical antiestrogen [ 186]. Soy products are rich in phytoestrogens, which have been detected in high concentrations in the plasma or urine of individuals who consumed soy or other phytoestrogens [187]. Other less concentrated dietary sources of phytoestrogens include rice, corn, alcohol, cereal bran, whole wheat, and beans [188]. In Japanese women, phytoestrogen excretion is 100 times higher and endogenous estrogen excretion is 100 to 1000 times higher than in American and Finnish women [189]. Differences in phytoestrogen intake may be a (partial) explanation for the differences in frequencies of menopausal symptoms observed in Asian and Caucasian women, although this is not currently known. Urinary excretion of phytoestrogens and the concentration of plasma sex hormone binding globulin have been positively associated with dietary intake of fiber, which has been inversely related to plasma percentage free estradiol [190]. In postmenopausal women supplemented with soy or wheat flour (which contains less potent enterolactones), statistically significant (40 and 25%, respectively) reductions in hot flashes were observed, whereas vaginal cell maturation was unchanged and FSH was decreased [191 ]. In addition, in a small randomized trial of a 12-week phytoestrogen-rich diet, postmenopausal women on the diet showed significantly increased SHBG, significant reduction in hot flashes and vaginal dryness, and significant increases in serum concentrations of phytoestrogens, though no significant change in estradiol [ 181 ]. In summary, environmental factors do influence the menopausal transition. Active smoking has been consistently associated with a 1- to 2-year earlier menopause [45, 46,50,52,57,110-114], in a dose-response relationship, although the role of passive smoke exposure is uncertain. Findings regarding the relations of body weight and body
ELLEN B. GOLD
composition to age at menopause and symptom reporting have been inconsistent. The relations of diet, physical activity, and occupational or other environmental factors to age at menopause have largely not been investigated. Active smoking has also been associated with increased symptom reporting during the menopause transition [39,49,65,66, 68,160]. However, findings that relate physical activity to symptom reporting have been inconsistent. Phytoestrogen intake has been related to reduced hot flashes [ 181,191 ], but the role of other dietary factors is only beginning to be explored.
V. C O N C L U S I O N S Despite important methodologic differences and the limitations in the study designs used and the populations studied in the accumulating literature on the menopausal experience, an interesting and complex picture is emerging. A number of demographic (e.g., education, employment, race/ethnicity), menstrual and reproductive, and lifestyle (e.g., smoking and diet) factors appear to be important determinants of the age at which menopause occurs and to have meaningful relationships to the varied symptom experience of women. AfricanAmerican and Latina ethnicity, smoking, lower parity, vegetarian diet and undernutrition, and lower socioeconomic status have been found fairly consistently to be associated with earlier menopause. Symptom reporting varies by ethnicity, with less reporting of vasomotor symptoms in most Asian populations and increased reporting of vaginal dryness in African-American and Hispanic women. History of premenstrual tension, smoking and lower socioeconomic status have been associated with increased symptom reporting, whereas dietary phytoestrogen intake appears to reduce hot flashes. However, a number of the relationships are inconsistent (e.g., the role of body mass and composition and physical activity), possibly due to varying methodologic approaches and limitations, and others remain largely unexplored (e.g., passive smoke exposure and occupational and other environmental exposures). Thus, much remains to be learned about how these factors affect hormones at the physiologic level and thus determine the onset of the perimenopause, the timing of the final menstrual period, and the occurrence of the constellation of symptoms that are associated with the menopause transition. Furthermore, increased understanding of the underlying physiologic bases of these influences needs to include potential racial/ethnic differences in physiologic responses to lifestyle factors and other environmental exposures, as well as increased understanding of the cultural contexts, cultural differences, and cultural sensitivities that affect the presentation and experience of the menopausal transition. Increasing knowledge about these relationships ultimately offers women and their health care providers
197
CHAPTER 12 Demographics, Environment, Race, and Nationality
choices based on deeper understanding as to the variety of alternatives available to deal with the individual presentations of menopause. Acknowledgments The author is indebted to the following collaborators for their contributions to this study of the natural history of the menopause: Drs. Barbara Abrams, Shelley Adler, Gladys Block, Maradee Davis, Bruce Ettinger, Bill Lasley, Marion Lee, Marianne O'Neill Rasor, Steven Samuels, Helen Schauffler, Barbara Sommer, and Barbara Sternfeld.
17.
18. 19.
20.
21.
References 1. Avis, N. E., Kaufert, R A., Lock, M., McKinlay, S. M. and Vass, K. (1993). The evolution of menopausal symptoms. Baillikre's Clin. Endocrinol. Metab. 7, 17-32. 2. Cramer, D. W., and Xu, H. (1996). Predicting age at menopause. Maturitas 23, 319-326. 3. Hahn, R. A., Eaker, E., and Rolka, H. (1997). Reliability of reported age at menopause. Am. J. Epidemiol. 146, 771-775. 4. Sowers, M. E, and LaPietra, M. T. (1995). Menopause: Its epidemiology and potential association with chronic diseases. Epidemiol. Rev. 17, 287-302. 5. Gosden, R. G. (1985). "Biology of the Menopause: The Causes and Consequences of Ovarian Ageing." Academic Press, London. 6. Burger, H. G., Dudley, E. C., Hopper, J. L., Shelley, J. M., Green, A., Smith, A., Dennerstein, L., and Morse, C. (1995). The endocrinology of the menopausal transition: A cross-sectional study of a populationbased sample. J. Clin. Endocrinol. Metab. 80, 3537-3545. 7. Soule, M. R., and Bremner, W. J. (1982). The menopause and climacteric: Endocrinologic basis and associated symptomatology. J. Am. Geriatr. Soc. 30, 547. 8. Thomford, E J., Jelovsek, F. R., and Mattison, D. R. (1987). Effect of oocyte number and rate of atresia on the age of menopause. Reprod. Toxicol. 1, 41-51. 9. Sherman, B. M., West, J. H., and Korenman, S. G. (1976). The menopausal transition: Analysis of LH, FSH, estradiol and progesterone concentrations during menstrual cycles of older women. J. Clin. Endocrinol. Metab. 42, 629. 10. Santoro, N., Rosenberg-Brown, J., Adel, T., and Skurnick, J. H. (1996). Characterization of reproductive hormonal dynamics in the perimenopause. J. Clin. Endocrinol. Metab. 81, 1495-1501. 11. Upton, G. V. (1982). The perimenopause: Physiologic correlates and clinical management. J. Reprod. Med. 27, 1. 12. Shideler, S. E., DeVane, G. W., Kalra, E S., Benirschke, K., and Lasley, B. L. (1989). Ovarian pituitary hormone interactions during the menopause. Maturitas 11, 331-339. 13. Grodin, J. M., Siiteri, E K., and MacDonald, E C. (1973). Extraglandular estrogen in the postmenopause. In "The Menopause and Aging" (K. J. Ryan and D. C. Gibson, eds.), p. 15. U.S. Govt Printing Office, Washington, DC. 14. Nimrod, A., and Ryan, K. J. (1975). Aromatization of androgens by human abdominal and breast fat tissue. J. Clin. Endocrinol. Metab. 40, 367. 15. Kase, N. (1978). Estrogen replacement therapy for the menopause. In "Hormonal Contraceptives, Estrogens and Human Welfare" (M. C. Diamond, and C. C. Korenbrot, eds.), p. 75. Academic Press, New York. 16. Hershcopf, R., and Bradlow, H. L. (1987). Obesity, diet, endogenous
22. 23. 24.
25.
26. 27. 28. 29.
30. 31.
32.
33.
34.
35. 36. 37. 38. 39.
estrogens and the risk of hormone-sensitive cancer. Am. J. Clin. Nutr. 45, 283-289. Johanson, A. G., Forslund, A., Sjodin, A., Mallmin, H., Hambraeus, L. and Ljunghall, S. (1984). Determination of body composition: A comparison of dual energy X-ray absorptiometry and hydrodensitometry. Am. J. Clin. Nutr. 57, 373-376. Cutler, W. B., and Garcia, C. R. (1984). "The Medical Management of Menopause and Perimenopause." Lippincott, Philadelphia. Metcalf, M. G., Donald, R. A., and Livesey, J. H. (1982). Pituitaryovarian function before, during and after the menopause: A longitudinal study. Clin. Endocrinol. 17, 489-494. Wilbur, J. E., Miller, A. M., Montgomery, A., and Chandler, P. (1998). Sociodemographic characteristics, biological factors, and symptom reporting in midlife women. Menopause 5, 43-51. Greene, J. G. (1992). The cross-sectional legacy: An introduction to longitudinal studies of the climacteric. Maturitas 14, 95-101. Barbo, D. (1987). The physiology of the menopause. Med. Clin. North Am. 71, 11. World Health Organization (1981). "Research on the Menopause." WHO, Geneva. Matthews, K. A., Wing, R. R., Kuller, L. H., Meilahn, E. N. and Plantinga, P. (1994). Influence of the perimenopause on cardiovascular risk factors and symptoms of middle-aged healthy women. Arch. Intern. Med. 154, 2349-2355. Batrinos, M. L., Panitsa-Faflia, C., Pitoulis, S., Pavlou, S., Piaditis, G., Alexandridis, T., and Liappi, C. (1979). The clinical features of the menopause and its relation to the length of pregnancies and lactation. Maturitas 1, 261-268. McKinlay, S. M., Brambilla, D. J., and Posner, J. G. (1992). The normal menopause transition. Am. J. Human Biol. 4, 37-46. Kronenberg, F. (1990). Hot flashes: Epidemiology and physiology. Ann. N.Y. Acad. Sci. 592, 52-86. Hagstad, A., and Janson, P. O. (1986). The epidemiology of climacteric symptoms. Acta Obstet. Gynecol. Scand. Suppl. 134, 159. Samsioe, G., Bryman, I., and Ivarsson, E. (1985). Some anthropological aspects of the climacteric syndrome. Acta Obstet. Gynecol. Scand. Suppl. 130, 5-7. Dennerstein, L. (1996). Well-being, symptoms and the menopausal transition. Maturitas 23, 147-157. Stadberg, E., Mattsson, L. A., and Milsom, I. (1997). The prevalence and severity of climacteric symptoms and the use of different treatment regimens in a Swedish population. Acta Obstet. Gynecol. Scand. 76, 442-448. O'Connor, V. M., DelMar, C. B., Sheehan, M., Siskund, V., FoxYoung, S., and Cragg, C. (1995). Do psychosocial factors contribute more to symptom reporting by middle-aged women than hormonal status? Maturitas 20, 6 3 - 69. Oldenhave, A., Jaszman, L. J., and Haspels, A. A. (1993). Impact of climacteric on well-being: A survey based on 5213 women, 39 to 60 years old. Am. J. Obstet. Gynecol. 168, 772-780. Hammar, M., Berg, G., Fahraeus, L., and Larsson-Cohn, U. (1984). Climacteric symptoms in an unselected sample of Swedish women. Maturitas 6, 345-350. Everett, H.S. (1991). Urology in the female. Am. J. Surg. 52, 521659. Asmussen, M., and Miller, A. (1983). "Clinical Gynaecological Urology." p. 21. Blackwell, Oxford. Versi, E., Cardozo, L., Studd, J., Brincat, M., and Cooper, D. (1995). Urinary disorders and the menopause. Menopause 2, 89-95. Neugarten, B. L., and Kraines, R. J. (1965). "Menopausal Symptoms" in women of various ages. Psychosom. Med. 27, 266-273. Kuh, D. L., Wadworth, M., and Hardy, R. (1997). Women's health in midlife: The influence of the menopause, social factors and health in earlier life. Br. J. Obstet. Gynaecol. 104, 923-933.
198 40. Mikkelsen, A., and Holte, A. (1982). A factor-analytic study of climacteric symptoms. Psychiatry Soc. Sci. 2, 35-39. 41. Cooper, G. S., and Sandler, D. R (1998). Age at natural menopause and mortality. Ann. Epidemiol. 8, 229-235. 42. Wise, A. M., Krajnak, K. M., and Kashon, M. L. (1996). Menopause: The aging of multiple pacemakers. Science 273, 67-70. 43. Snowdon, D. A., Kane, R. L., Beeson, W. L., Burke, G. L., Sprafka, J. M., Potter, J., Iso, H., Jacobs, D. R., Jr., and Phillips, R. L. (1989). Is early natural menopause a biologic marker of health and aging? Am. J. Public Health 79, 709-714. 44. Trevoux, R., DeBrux, J., Castaneir, M., Nahool, K., Soule, J. R, and Scholler, R. (1986). Endometrium and plasma hormone profile in the peri-menopause and post-menopause. Maturitas 8, 309. 45. Greendale, G., Hogan, R, Kritz-Silverstain, D., Langer, R., Johnson, S., and Bush, T., for the PEPI trial investigators (1995). Age at menopause in women participating in the postmenopausal estrogen/progestins interventions (PEPI) trial: An example of bias introduced by selection criteria. Menopause 2, 27-34. 46. Luoto, R., Laprio, J., and Uutela, A. (1994). Age at natural menopause and sociodemographic status in Finland. Am. J. Epidemiol. 139, 64-76. 47. Stanford, J. L., Hartge, R, Brinton, L. A., Hoover, R. N., and Brookmeyer, R. (1987). Factors influencing the age at natural menopause. J. Chronic Dis. 40, 995-1002. 48. Magursky, V., Mesko, M., and Sokolik, L. (1975). Age at the menopause and onset of the climacteric in women of Martin district, Czechoslovakia. Int. J. Fertil. 20, 17-23. 49. Gold, E. B., Sternfeld, B., Kelsey, J., Brown, C., Mouton, C., Reame, N., Salamone, L., and Stellato, R. (2000). The relation of demographic and lifestyle variables to symptoms in a multiracial/ethnic population of women aged 40-55 years. Am. J. Epidemiol. In press. 50. vanNoord, R A. H., Dubas, J. S., Dorland, M., Boersma, H., and teVelde, E. (1997). Age at natural menopause in a population-based screening cohort: The role of menarche, fecundity, and lifestyle factors. Fertil. Steril. 68, 95-102. 51. Flint, M. (1978). Is there a secular trend in age of menopause? Maturitas 1, 133-139. 52. Torgerson, D. J., Avenell, A., Russell, I. T., and Reid, D. M. (1994). Factors associated with onset of menopause in women aged 45-49. Maturitas 19, 83-92. 53. Brand, R C., and Lehert, P. H. (1978). A new way of looking at environmental variables that may affect the age at menopause. Maturitas 1, 121-132. 54. McKinlay, S., Jeffreys, M., and Thompson, B. (1972). An investigation of the age at menopause. J. Biosoc. Sci. 4, 161-173. 55. Whelan, E. A., Sandler, D. P., McConnaughey, D. R., and Weinberg, C. R. (1990). Menstrual and reproductive characteristics and age at natural menopause. Am. J. Epidemiol. 131,625-632. 56. Treloar, A. R., Boynton, R. E., Behn, B. G., and Brown, B. W. (1966). Variation of the human menstrual cycle through reproductive life. Int. J. Fertil. 12, (Part 2) 77-126. 57. Bromberger, J. T., Matthews, K. A., Kuller, L. H., Wing, R. R., Meilahn, E. N., and Plantinga, P. (1997). Prospective study of the determinants of age at menopause. Am. J. Epidemiol. 145, 124-133. 58. Soberon, J., Calderon, J. J., and Goldzieher, J. W. (1966). Relation of parity to age at menopause. Am. J. Obstet. Gynecol. 96, 96-100. 59. Snieder, H., MacGregor, A. J., and Spector, I. D. (1998). Genes control cessation of a woman's reproductive life: A twin study of hysterectomy and age at menopause. J. Clin. Endocrinol. Metab. 83, 18751880. 60. Chompootweep, S., Tankeyoon, M., Yamarat, K., Poomsuwan, R, and Dusitsin, N. (1993). The menopausal age and climacteric complaints in Thai women in Bangkok. Maturitas 17, 63-71. 61. Parazzini, F., Negri, E., and LaVecchia, C. (1992). Reproductive and
ELLEN B. GOLD
62. 63. 64.
65. 66.
67. 68.
69.
70.
71.
72. 73.
74. 75.
76. 77.
78.
79.
80.
81.
82.
83.
general lifestyle determinants of age at menopause. Maturitas 15, 141-149. van Keep, R A., Brand, R C., and Lehert, E H. (1979). Factors affecting the age at menopause. J. Biosoc. Sci., Suppl. 6 37-55. Gonzales, G. F., and Villena, A. (1997). Age at menopause in central Andean Peruvian women. Menopause 4, 32-38. Avis, N., Crawford, S., and McKinlay, S. (1997). Psychosocial, behavioral, and health factors related to menopause symptomatology. J. Women's Health 3, 103-120. Leidy, L. E. (1997). Menopausal symptoms and everyday complaints. Menopause 4, 154-160. Collins, A., and Landgren, B.-M. (1995). Reproductive health, use of estrogen and experience of symptoms in peri-menopausal women, a population-based study. Maturitas 20, 101-111. Schwingl, R J., Hulka, B. S., and Harlow, S. D. (1994). Risk factors for menopausal hot flashes. Obstet. Gynecol. 84, 29-34. Dennerstein, L., Smith, A. M., Morse, C. A., Burger, H., Green, A. Hopper, J. and Ryan, M. (1993). Menopausal symptomatology: The experience of Australian women. Med. J. Aust. 159, 232-236. Hunter, M. R., Battersby, R., and Whitehead, M. I. (1986). Relationships between psychological symptoms, somatic complaints and menopausal status. Maturitas 8, 217-228. Scaglia, H., Medina, M., Pinto-Ferreira, A. L., Vazques, G., Gual, C., and Perez-Palacios, G. (1976). Pituitary LH and FSH secretion and responsiveness in women of old age. Acta Endocrinol. (Copenhagen) 81,673-679. vanKeep, P. A., and Kellerhals, J. M. (1974). The impact of sociocultural factors on symptom formation. Some results of a study on aging women in Switzerland. Psychother. Psychosom. 23, 251-263. Kirchengast, S. (1992). Relations between fertility, body shape and menopause in Austrian women. J. Biosoc. Sci. 24, 555-559. Sherman, B. M., Wallace, R. B., Bean, J. A., Chang, Y., and Schlabaugh, L. ( 1981 ). The relationship of menopausal hot flushes to medical and reproductive experience. J. Gerontol. 36, 306-309. McKinlay, S. M., and Jeffreys, M. (1974). The menopausal syndrome. Br. J. Prev. Med. 28, 108-115. Skarsgard, C., Bjors, E., Nedstrand, E., Wyon, Y., and Hammar, M. (1996). Do women's premenstrual symptoms and their mother's climacteric history predispose to their own vasomotor symptoms? Menopause 3, 133-139. Hunter, M. (1992). The south-east England longitudinal study of the climacteric and postmenopause. Maturitas 14, 117-126. Studd, J., Chakravarti, S., and Oram, D. (1977). The climacteric. In "Clinics in Obstetrics and Gynecology" (R. G. Greenblatt, and J. Studd., eds.), Vol. 1. Saunders, Philadelphia. Alvarado, G., Rivera, R., Ruiz, R., Arenas, M., and Rueda, D. (1988). Characteristicas del patron de sangrado menstrual en un grupo de mujeres normales de Durango. Ginecol. Obstet. Mex. 56, 127-133. Otolorino, E. O., Adeyefa, I., Osotimehin, B. O., Fatinikun, T., Ojengbede, O., Otubu, J. O., and Ladipo, O. A. (1989). Clinical, hormonal and biochemical features of menopausal women in Ibadan, Nigeria. Afr. J. Med. Sci. 18, 251-255. Beyene, Y. (1986). Cultural significance and physiological manifestations of menopause, a bicultural analysis. Cult. Med. Psychiatry 10, 47-71. World Health Organization (1983). A prospective multicenter trial of the ovulation method of natural family planning. III. Characteristics of the menstrual cycle and of the fertile phase. Fertil. Steril. 40, 7 7 3 778. Boulet, M. (1990). "The Menopause and the Climacteric in Seven Asian Countries," 6th Int. Congr. Menopause. Parthenon, Park Ridge, N.J. Ramoso-Jalbuena, J. (1994). Climacteric Filipino women: A preliminary survey in the Philippines. Maturitas 19, 183-190.
CHAPTER 12 Demographics, Environment, Race, and Nationality 84. McCarthy, T. (1994). The prevalence of symptoms in menopausal women in the Far East: Singapore segment. Maturitas 19, 199-204. 85. Samil, R. S., and Wishnuwardhani, S. D. (1994). Health of Indonesian women, city-dwellers of perimenopausal age. Maturitas 19,191-197. 86. Wasti, R., Robinson, S. C., Akhtar, Y., Khan, S., and Badaruddin, N. (1993). Characteristics of menopause in three groups in Karachi, Pakistan. Maturitas 16, 61-69. 87. Blumel, J., Cubillos, M., Brandt, A., Munoz, L., and Ayarza, E. (1988). Some clinical aspects of menopause. Rev. Chil. Obstet. Ginecol. 53, 278-282. 88. MacMahon, B., and Worcester, J. (1966). "Age at Menopause, United States 1960-1962," Vital Health Stat., Ser. 11, No. 19, pp. 1-19. National Center for Health Statistics, Rockville, MD. 89. Kapoor, A. K., and Kapoor, S. (1986). The effects of high altitude on age at menarche and menopause. J. Biometeorol. 30, 21-26. 90. Beall, C. M. (1983). Ages at menopause and menarche in a high altitude Himalayan population. Ann. Hum. Biol. 10, 365-370. 91. Flint, M. E (1974). Menarche and menopause in Rajput women. Ph.D. Dissertation, City University of New York. 92. Boule