This Article Abstract Full Text (PDF) All Versions of this Article: 27/6/575    most recent Final Manuscript Author Manuscript Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Turgeon, J. L. Articles by Wise, P. M. Search for Related Content PubMed PubMed Citation Articles by Turgeon, J. L. Articles by Wise, P. M.

Complex Actions of Sex Steroids in Adipose Tissue, the Cardiovascular System, and Brain: Insights from Basic Science and Clinical Studies

Division of Endocrinology, Clinical Nutrition, and Vascular Medicine, Department of Internal Medicine (J.L.T.), University of California, Davis, Davis, California 95616; Division of Endocrinology, Metabolism, and Molecular Medicine (M.C.C.), Northwestern University, Feinberg School of Medicine, Chicago, Illinois 60611; Department of Psychiatry and Psychology (P.M.M.), University of Illinois at Chicago, Chicago, Illinois 60612; Molecular Cardiology Research Institute, Department of Medicine, and Division of Cardiology (M.E.M.), New England Medical Center Hospitals and Tufts University School of Medicine, Boston, Massachusetts 02111; and Departments of Physiology and Biophysics, and Biology (P.M.W.), University of Washington, Seattle, Washington 98195

Correspondence: Address all correspondence and requests for reprints to: Judith Turgeon, Ph.D., Department of Internal Medicine, Division of Endocrinology, Clinical Nutrition, and Vascular Medicine, University of California Davis, Davis, California 95616. E-mail: jlturgeon{at}ucdavis.edu

	Abstract

Top Abstract I. Introduction II. Pharmacology of Estrogens... III. The Inflammatory Process IV. Adipose/Metabolic System and... V. Cardiovascular System and... VI. Central Nervous System... VII. Summary/Conclusions References

 
Recent publications describing the results of the Women’s Health Initiative (WHI) and other studies reporting the impact of hormone therapy on aging women have spurred reexamination of the broad use of estrogens and progestins during the postmenopausal years. Here, we review the complex pharmacology of these hormones, the diverse and sometimes opposite effects that result from the use of different estrogenic and progestinic compounds, given via different delivery routes in different concentrations and treatment sequence, and to women of different ages and health status. We examine our new and growing appreciation of the role of estrogens in the immune system and the inflammatory response, and we pose the concept that estrogen’s interface with this system may be at the core of some of the effects on multiple physiological systems, such as the adipose/metabolic system, the cardiovascular system, and the central nervous system. We compare and contrast clinical and basic science studies as we focus on the actions of estrogens in these systems because the untoward effects of hormone therapy reported in the WHI were not expected. The broad interpretation and publicity of the results of the WHI have resulted in a general condemnation of all hormone replacement in postmenopausal women. In fact, careful review of the extensive literature suggests that data resulting from the WHI and other recent studies should be interpreted within the narrow context of the study design. We argue that these results should encourage us to perform new studies that take advantage of a dialogue between basic scientists and clinician scientists to ensure appropriate design, incorporation of current knowledge, and proper interpretation of results. Only then will we have a better understanding of what hormonal compounds should be used in which populations of women and at what stages of menopausal/postmenopausal life.

I. Introduction

II. Pharmacology of Estrogens and Progestins

A. Menopausal hormone therapy (HT)

B. Estrogen/progestin therapy before menopause

III. The Inflammatory Process

A. Inflammation in brief

B. Estrogen and the inflammatory process

IV. Adipose/Metabolic System and Estrogens

A. Visceral adipose tissue

B. Influence of estrogen on adipose tissue

C. The influence of estrogen on markers of inflammation

V. Cardiovascular System and HT

A. Sex steroid hormone receptor action and cardiovascular physiology

B. Vascular effects of sex steroid hormones, atherosclerosis and ischemic cardiovascular diseases

C. Sex steroid hormones and the myocardium in health and disease

VI. Central Nervous System and Sex Steroids

A. Comparisons between basic science studies and the WHI

B. Studies performed using animal models

C. Mechanisms of estrogen action in the central nervous system

D. Evaluation of the Women’s Health Initiative Memory Study (WHIMS): design and findings

E. WHIMS validity: internal and external

F. Differences with previous clinical and observational studies: age of initiation of treatment

G. Potential of SERMS as HT: efficacy on central nervous system parameters

H. Effect of progestins on cognition

VII. Summary/Conclusions

	I. Introduction

Top Abstract I. Introduction II. Pharmacology of Estrogens... III. The Inflammatory Process IV. Adipose/Metabolic System and... V. Cardiovascular System and... VI. Central Nervous System... VII. Summary/Conclusions References

 
WOMEN TYPICALLY UNDERGO the transition to postmenopause at approximately 51 yr of age. This dramatic physiological transition, which can be abrupt or can occur over several years, is associated with decreases in ovarian hormone secretion and secondary changes in multiple endocrine and metabolic parameters. Many women suffer from menopausal symptoms that have been attributed to the lack of ovarian steroids, including hot flushes, sleep disturbance, vaginal dryness, urinary symptoms, and emotional instability. In addition, postmenopausal women are thought to be at increased risk for several diseases compared with age-matched cycling women such as cardiovascular disease (CVD), cerebrovascular stroke, osteoporosis, hip fracture, dementia, and Alzheimer’s disease (AD).

Over the past 20 yr, numerous observational, retrospective, interventional, and meta-analytic studies (reviewed in Refs. 1, 2, 3) as well as studies using animal models (reviewed in Refs. 4, 5, 6, 7, 8, 9) have supported the hypothesis that ovarian steroids exert important protective actions in women and the absence of these hormones after the menopause makes postmenopausal women more vulnerable than younger premenopausal women to CVD, central nervous system imbalances, neurodegenerative diseases, osteoporosis, and immune dysfunction. However, the recent release of several studies (10, 11, 12, 13, 14, 15, 16, 17) has forced us to reassess these conclusions and reevaluate the benefits and risks of hormone therapy (HT) in older women. The study of the Women’s Health Initiative (WHI) (11) (see http://www.nih.gov/PHTindex.htm for a complete list of publications) has caught the attention of researchers, physicians, and the lay public. The WHI was comprised of two large, randomized, placebo-controlled clinical trials for HT in postmenopausal women and had three significant design characteristics: 1) only one type of hormone regimen was studied, i.e., oral conjugated equine estrogens (CEE) at a single concentration without or with continuous oral medroxyprogesterone acetate (MPA); 2) the average age of the participants at entry was 63 yr, and, for a majority of the women, estrogen deficiency had been present for more than a decade; and 3) nearly 70% of the subjects were either overweight or obese. These studies were expected to settle the question of whether postmenopausal women would benefit from HT and would solve the dilemma of what women should do after the menopause once and for all. Instead, they have raised many questions and caused much reanalysis and reinterpretation of the results of previous studies.

Fallout from the WHI has been an indictment of all menopausal HT. The unfortunate consequence is that women and their physicians are left with, at best, unsettling choices. In attempting to reconcile the disparate conclusions derived from the WHI and observational studies/basic research, three points emerge that may account for the disparities: 1) different estrogens (or progestins) are not recognized in the same way in all cells and do not have equivalent functions; 2) hormone delivery regimens have a major impact on outcome measures; and 3) age is a critical interactive factor in hormone action, particularly in the timing of the initiation of HT relative to the onset of menopause. An additional and perhaps crucial element in accounting for the study disparities is our emergent understanding of estrogen’s interface with inflammation and with the adipose/metabolic system. There are direct effects of estrogen on the inflammatory process and on adipose tissue distribution and function, and the question is how does this contribute to the phenotypic shifts in dyslipidemia, diabetes, and cardiovascular and neurodegenerative diseases in women through their life span. Recent discoveries in these areas allow us to consider novel concepts regarding the varied roles of estrogen and the design of future hormone therapies.

Here, we review the basic pharmacology of estrogens and progestins and provide an overview of the inflammatory response. On this background, we examine the points raised above with a focus in three major areas: adipose/metabolic, cardiovascular, and central nervous systems.

	II. Pharmacology of Estrogens and Progestins

Top Abstract I. Introduction II. Pharmacology of Estrogens... III. The Inflammatory Process IV. Adipose/Metabolic System and... V. Cardiovascular System and... VI. Central Nervous System... VII. Summary/Conclusions References

 
A. Menopausal hormone therapy (HT)
Publication of results from clinical trials, such as the WHI and the Heart and Estrogen/Progestin Replacement Study (HERS) (10), has prompted an extremely useful reexamination of the roles of ovarian hormones outside of the reproductive axis. An equally constructive result has been an increased awareness of lapses that can occur in the dialogue between clinical and basic scientists on estrogen and progestin actions.

In comparing clinical studies and animal models of HT, the focus most often is on subject characteristics and main outcome measurements and less frequently on the specifics of the intervention. Here, we address the principle that not all estrogens or progestins are equal and briefly review the consequences of differences in receptor activation, delivery modes, and timing of hormone treatment. The reader is also directed to reviews by Shoham and Kopernik (18, 19) that provide a wonderfully lucid and insightful perspective on the history, pharmacology, and bioactivity of HT formulations.

1. Are all estrogens or progestins equal in action?
Clearly, the answer is no. "Estrogen" and "progestin" are generic terms covering an array of endogenous hormones and synthetic compounds. Although useful as shorthand, these nonspecific terms obscure the importance of the unique actions of individual estrogens or progestins. As reviewed in Refs. 3, 20 , and 21 , binding of ligand to a steroid receptor induces distinct structural alterations in the receptor. Instead of functioning as simple switches for receptor activation, ligands have a more complex role in that different ligands induce unique conformations in a receptor. This is crucial to the ultimate cellular response because recruitment of appropriate coregulators to the ligand-receptor complex is conformation-dependent. The coregulator complement, in turn, confers cell specificity. The final conformation of the ligand-receptor-coregulator complex affects how its activity is modified (by phosphorylation, etc.) and ultimately determines target-gene promoter specificity or, in the case of extranuclear receptors, signaling targets (21, 22, 23, 24, 25, 26). This versatile system is made more so by at least two different forms of the estrogen receptor (ER and ERß) and the progesterone receptor (PR-A and PR-B). Within their ligand binding domains, the ER isoforms differ by more than 40% in amino acid sequence, thus providing further opportunity for different estrogens to induce distinct conformational changes in an ER-coregulator complex to elicit discrimination in function (25). The isoforms within a receptor type are differentially expressed and regulated across cell types and have both common and distinct target genes (27, 28, 29, 30, 31, 32, 33, 34).

The significance of ligand-dependent receptor conformation cannot be overstated in comparing the actions of different types of estrogens or progestins. Examples of the influence that structural differences in ligand have on outcome measures can be seen with selective ER modulators (SERMs), such as tamoxifen or raloxifene, which function as either ER agonists or antagonists depending on the cell type and coregulator complement (29, 35). SERMs, acting through ER isoforms, have been shown to regulate different gene subsets compared with the classic ovarian estrogen, 17ß-estradiol (E₂) (36). Receptor isoform-specific regulation of gene expression, which has been demonstrated for the PR isoforms as well as for ER isoforms, is being exploited for structure-based design for ligands with individualized tissue effects (34, 37, 38, 39, 40, 41). This work and other research, for example investigating specific coregulator recruitment (42, 43), use mechanism-based approaches that show great promise for producing therapeutic tools that are function-specific. These approaches also are being explored for extranuclear steroid receptors, for example in the design of ER ligands that are cytosolic signaling pathway-selective, with the goal of target-specific therapy (44). The area of extranuclear steroid receptors and rapid signaling, including the possibility of nonreceptor-mediated effects, is too complex and unique a topic to be succinctly summarized here, but we strongly recommend recent excellent and comprehensive reviews of the subject (45, 46).

Although SERMs and selective PR modulators represent well-known examples of compounds that induce unique receptor conformations and exhibit distinct properties in specific cell types (35), what about the estrogens and progestins commonly used in HT? By far the most frequently prescribed treatment in the United States for menopausal HT is oral CEE, with or without MPA (Table 1) (47). CEE, which is derived from the urine of pregnant horses, is comprised of at least 10 different estrogens, as well as some androgens and progestins (48, 49). In addition to sulfated forms of estrone (E₁) and E₂, the estrogen component of CEE includes several sulfated estrogens that are unique to the horse, the most prominent being equilin (49, 50). The critical question is what are the functional, tissue-specific effects of the CEE component estrogens on the function of ER isoforms in women? The few studies that directly addressed this question suggest a wide range of activities reflecting individual metabolic clearance rates as well as estrogen type-specific activity at the ER that can translate into cell type-specific responses (48, 50, 51). For example, a CEE component, ^8,9-dehydroestrone sulfate, has a tissue selectivity and activity profile distinct from that for E₂ with full agonist effects on several central nervous system endpoints and little or no efficacy in certain hepatic or vascular endpoints (48, 52). In general, the CEE components demonstrate the principle that binding affinity for the ER does not necessarily predict biological activity (48). Thus, the clinical response to CEE is a composite of the pharmacokinetic profiles and individual actions of the steroid components as well as the interactions of the components at ER isoforms that could either mitigate or magnify a response.

Drospirenone (DRSP) is a progestin with antialdosterone and antiandrogenic activities that is in active use in HT and contraceptive formulations in many countries outside the United States, and especially Europe (67). In clinical trials, DRSP is effective against menopausal symptoms and is able to provide endometrial protection and maintain amenorrhea in a majority of women. In the doses used in HT formulations, DRSP also has a modest blood pressure-lowering effect, which may prove to be beneficial in some patients. The role of the mineralocorticoid receptor in cardiovascular physiology currently is receiving increasing attention, and the presence of functional mineralocorticoid receptors regulating gene expression in vascular tissues has been reported recently (68), raising the possibility that some of the protective effects of antialdosterone compounds like spironolactone or eplerenone in CVDs (69, 70) may be exerted in nonrenal tissues, including the blood vessel (71, 72).

2. HT delivery route.
In postmenopausal literature, the term "HT" frequently is used as if it were a single entity. However, in addition to the profound effect that the specific estrogen or progestin formulation may have on outcome, a perhaps equally critical variable is how HT is delivered. In 2003, an estimated 80% of prescriptions dispensed in the United States for HT were for estrogen +/– progestin formulations delivered orally, and the majority of these contained CEE as the estrogen component (Table 1) (47). In Europe, micronized E₂ is one of the most common estrogens used orally, but transdermally administered E₂ is widely used as well (19, 73, 74, 75). For basic research in animal models, steroid hormones are usually given as E₂ and progesterone in sc or im depots and are rarely administered orally. This difference in drug delivery methods is a critically important factor when extrapolating knowledge gained from animal models to humans, for which oral HT is by far the most common delivery route.

Oral CEE and oral E₂ are similar in that a peak in plasma concentration is reached within the first 3 h, followed by a decline, and both oral formulations result in higher plasma E₁ levels than E₂, which is unlike the normal menstrual cycle ratio. The pharmacokinetics of oral CEE is complex due to the multiple estrogen components, with varying binding affinities for transport proteins and different metabolic clearance rates (49). For transdermal HT, the estrogen component is E₂. Transdermal application of E₂ avoids the exaggerated peaks and nadirs in plasma estrogen concentration that are a consequence of oral HT and results in a rate of conversion to E₁ that produces an E₂:E₁ ratio more like that found in a menstrual cycle.

It has long been recognized that estrogens taken orally have the potential to modulate liver function due to first-pass effects on hepatic ERs. Clinical studies reporting side-by-side or crossover comparisons provide specific and striking examples of effects induced by oral but not transdermal estrogen on liver production of proteins involved in inflammation, lipid profile, thrombosis, and metabolism (Table 2). A key to these differences is the high concentration of oral estrogen that is required to escape first-pass hepatic metabolism and achieve a therapeutic concentration in peripheral circulation. For E₂, oral administration requires doses that are 20- to 40-fold higher than doses used for transdermal therapy to achieve comparable systemic E₂ levels. For the examples shown in Table 2 comparing oral and transdermal E₂, the resulting circulating E₂ levels averaged 40–150 pg/ml for either treatment modality; the difference was in the dose delivered to the liver.

Recent studies have generated insight into mechanisms responsible for some of the hepatic actions of estrogen. For example, the cytokine signaling pathway used by GH to stimulate production of IGF-I by liver cells has a built-in brake system via a parallel induction of suppressor proteins from the SOCS (suppressor of cytokine signaling) family. In vitro studies now show that E₂, acting through nuclear ER, can suppress GH signaling by induction of SOCS-2, thus providing one possible target mechanism for elevated hepatic estrogen (78). It will be important to establish that a concentration-dependent estrogen suppression of GH signaling occurs in vivo in the liver and to determine whether this estrogen-stimulated mechanism is operative at other GH targets, e.g., affecting GH-stimulated lipolysis in adipose tissue.

Determining the impact of oral progestins on hepatic proteins is complex due to the multiple formulations and delivery routes used therapeutically and also to the experimental challenge of sorting out effects attributable to the androgenicity of the compounds. It has been shown, however, that the most commonly prescribed progestin, oral MPA, blunts the increase in high-density lipoprotein (HDL) induced by oral CEE in a dose-dependent manner, yet oral micronized progesterone has no effect on this endpoint (79, 80). For the more androgenic progestins, such as norethindrone or levonorgestrel, both of which can be taken either orally or transdermally, the few side-by-side trials in postmenopausal women suggest that the differences in outcomes such as thromboembolic events or effects on lipid profile are more related to chemical structure than to delivery route (81, 82, 83).

3. HT timing.
Two design aspects of the WHI and HERS that have been particularly controversial are the age of the participants and the duration of estrogen deficiency before HT was initiated. With a mean age of 63 yr (with nearly 70% between the ages of 60 and 79) and estrogen deficiency for more than a decade in the majority of the women, the WHI in effect becomes a study more of aging than of menopause (84, 85).

The physiological consequence of estrogen reintroduction depends on an interaction between the target system, the health of that system, and the duration of estrogen deprivation. For example in the cardiovascular system, premenopausal ovariectomy or premature ovarian failure has been associated with an increase in peripheral vascular resistance and blood pressure, impaired endothelial function, and increased risk of coronary heart disease that can be reversed by early replacement therapy with estrogen (86, 87, 88). That the timing of estrogen therapy (ET) initiation is critical was demonstrated in a monkey model of diet-induced atherosclerosis. After surgically induced menopause in these monkeys, if ET was initiated immediately there was a reduction in coronary artery atherosclerosis compared with placebo-treated ovariectomized monkeys, but delaying ET by an equivalent of about 6 human years completely eliminated the beneficial reduction in atherosclerosis (89). The effects of estrogen reintroduction after a disease process has begun will vary depending on the end organ target, its plasticity, and any consequences aging may have had in the interim. Evaluation of these system-dependent variables must be made before conclusions can be drawn regarding beneficial, neutral, or harmful effects of HT.

In summary, the initial aim of HT was to alleviate the obvious signs of estrogen withdrawal such as vasomotor flushes and vaginal dryness. Over the years, these relatively simple aims have grown into a long and complex list of targets outside of the reproductive system for which a "one formulation/one size fits all" approach is not tenable. CEE/MPA or other HT formulations were not meant to be a panacea for aging, nor were they designed to be optimal for the physiological/endocrinological status of all women over the age of 50. When some risk to health for HT users is suggested, as in the WHI, all estrogens and progestins should not be denounced because each of the formulations has a unique action profile and pharmacokinetic properties. Additionally, oral HT has a wide range of effects on liver proteins, and the consequences of these changes and their impact on other organ systems must be a central consideration in the interpretation of clinical studies as well as in comparing human studies with animal model studies, for which the steroids rarely are administered orally. Other influential factors include a woman’s age and health status if HT is initiated after a lapse of several years. To paraphrase Stevenson and Whitehead (90), survival of the human species over 2 million years implies that ovarian hormones by themselves are not a health hazard; if harm is suspected, the judicious research inquiry should focus on the types and doses of hormone substitutes being used as well as their delivery routes and treatment protocols. And we would add that an additional focus should be to understand the underlying physiology for the many targets of E₂ and progesterone outside of the reproductive system and the adaptive roles they have in the aging population.

B. Estrogen/progestin therapy before menopause
The most common use of exogenous estrogens and progestins in premenopausal women is for contraception. The pharmacology of the wide-ranging choices for types of contraceptive drugs and delivery routes is outside the scope of this review, but a limited examination of common oral formulations and the recently available transdermal preparation is presented for comparison with the regimens and outcomes of postmenopausal HT.

1. Oral contraceptive (OC) hormones.
Preparations containing 17-ethinyl estradiol (EE) and a progestin (e.g., norethindrone, levonorgestrel, or norgestimate) are the most frequently prescribed oral hormonal contraceptives. Since U.S. approval 46 yr ago, the primary changes in OCs have been a decrease in EE content and the development of new generations of progestins in an effort to reduce side effects while maintaining LH/FSH inhibition to prevent ovulation. Current low-dose oral EE is no greater than 35 µg/d.

How does EE compare to E₂? At the nuclear ER, the two ligands have similar binding affinity (91, 92) and ability to induce rat uterine target genes (93), although this similarity in efficacy has not been established for all systems or for nonnuclear ERs. In fact, there are surprisingly few studies evaluating specific actions of EE outside of the reproductive system. Earlier studies suggested that EE had enhanced hepatic effects and systemic potency compared with that following oral E₂, but this is likely related to the increased bioavailability of EE due to its marked resistance to metabolism by hepatic 16-hydroxylation as well as reduced conversion to E₁, thus resulting, in effect, in a higher concentration of EE presented to hepatic cells (see Refs. 49, 93 , and 94 and references therein). Another crucial contribution to the bioavailability of EE is its minimal binding to SHBG, instead being transported by albumin (49). A 35-µg dose of oral EE in premenopausal women results in an average total serum concentration of about 50 pg/ml, a large fraction of which is free due to the low-affinity binding of EE to albumin (95). In comparison, at least 1000 µg of oral micronized E₂ would be required to achieve comparable average total serum levels. However, in contrast to EE, less than 5% of total E₂ in circulation is free due to SHBG binding (96).

Not unexpectedly, because of the first-pass hepatic phenomenon and increased bioavailability, oral EE affects liver proteins similarly to oral E₂ (Table 2). For example, OCs are associated with increased incidence of thromboembolic disease (81, 82, 97). One of the most common risk factors for venous thromboembolism is hereditary resistance to hepatic protein C, which when activated essentially down-regulates thrombin formation. OCs have been shown to induce resistance to activated protein C (APC) in the absence of a hereditary mutation, so-called acquired APC resistance, and this extends to oral postmenopausal HT as well (Table 2) (98, 99, 100, 101). Although oral estrogen by itself can lead to increased APC resistance, the progestin component, particularly third generation progestins such as norgestimate, can contribute to the thrombotic effects of oral estrogen (81, 82, 98, 99, 102). In addition to its role in antithrombin formation, APC has been shown to have antiinflammatory activity (103, 104), but whether this function is affected by OC- or oral HT-induced APC resistance has not been established.

Venous thrombosis is a complex process involving procoagulant, anticoagulant, and fibrinolytic elements, and oral estrogen has some beneficial effect on other components of the pathway despite its overall adverse effect on thrombosis. This becomes more apparent when first-pass liver effects are bypassed; studies in postmenopausal women suggest that transdermally administered E₂ has beneficial effects on several hemostatic markers for anticoagulant activity (105). It is significant that transdermal E₂ used for HT, unlike oral E₂, is not associated with risk for venous thrombosis and does not induce APC resistance (Table 2) (75, 100, 106). It remains to be determined whether the same applies to transdermal EE (see Section II.B.2) and to premenopausal women.

2. Transdermal contraceptive hormones.
A recent addition to the contraceptive armamentarium is a combined estrogen/progestin formulation given as a transdermal patch. Instead of containing E₂ as in the patch used for HT, the currently available contraceptive patch contains EE. The formulation results in average serum levels of 50–70 pg/ml with the same increased bioavailability as oral EE but without the peaks and troughs and, consequently, an overall higher 24-h level of serum EE compared with that found with oral EE as described above (95, 107). The progestin in the contraceptive patch is norelgestromin, the primary active metabolite of a common component of OCs (norgestimate).

Given the importance of overall health status and the indeterminate effects of aging in the response of organ systems to the modulating actions of estrogen, the premenopausal population represents an excellent opportunity for a direct comparison of oral vs. transdermal administration of the same estrogen/progestin formulation in younger women. This would be particularly useful for the examination of cardiovascular and central nervous system endpoints affected by the confounding variable of oral estrogen’s hepatic action. For example, a study of oral vs. transdermal E₂ rather than EE in premenopausal women to evaluate indicators of venous and arterial thromboses could be highly informative for comparison with existing data for oral vs. transdermal E₂ in postmenopausal women.

In summary, hormone treatment can be considered as a continuum from contraception through postmenopause. Although similarities are many, the differences may be informative. For example, unlike most regimens for HT, the standard protocol for contraceptive hormones includes a periodic interruption for a week without exogenous hormones, which may have consequences for ER isoform expression level and ratios. Additionally, some differences in formulation and dose are found between contraceptive hormones and HT, but the significance of these variances to targets outside the reproductive system has not been firmly established. Confounding such analysis are endogenous hormones. Unlike that in postmenopausal women, the ovarian contribution to circulating estrogen levels still can be significant in premenopausal women on contraceptive HT. However, the current trend toward an increase in body mass index (BMI) for both pre- and postmenopausal women may mute this difference in endogenous hormone levels because of conversion of adrenal androgens to estrogens by adipocyte aromatase. And perhaps the most critical difference between premenopausal and postmenopausal populations is that of overall health status and background into which HT is introduced, a consideration particularly relevant for subclinical disease progression when HT initiation is delayed.

	III. The Inflammatory Process

Top Abstract I. Introduction II. Pharmacology of Estrogens... III. The Inflammatory Process IV. Adipose/Metabolic System and... V. Cardiovascular System and... VI. Central Nervous System... VII. Summary/Conclusions References

 
A. Inflammation in brief
Inflammation research has experienced remarkable progress over the last decade and has generated increased appreciation for the complexity of the inflammatory process. A detailed presentation of the area is well beyond the scope of this review, but an outline of some of what is known of the ground rules and the players in inflammation is necessary for considering how sex steroid hormones might affect this process.

As a simple working definition, inflammation is a well-coordinated response of the immune, endocrine, and metabolic systems to tissue damage. Although an acute inflammatory response, for instance to a skin injury or infection, is the type that most readily comes to mind, at the other end of the spectrum is chronic inflammation, which is thought to be at the base of disorders such as arteriosclerosis, osteoarthritis, inflammatory bowel diseases, and several neurodegenerative diseases. The highly regulated acute inflammatory response that results in damage control and tissue restoration normally includes antiinflammatory mechanisms that contribute to the termination of the response. However, if the initiating injury or perturbation is recurrent or nonresolving or if a genetic alteration interferes with the sequence, the result is chronic inflammation.

The primary signaling molecules in the inflammatory response are cytokines produced on site (by, e.g., macrophages, monocytes, lymphocytes, activated microglial cells) and peripherally (e.g., visceral adipose tissue). The proinflammatory cytokines (e.g., TNF, IL-1, IL-6, and interferon-) along with growth factors (e.g., macrophage colony-stimulating factor) and chemokines (e.g., IL-8) initiate a coordinated sequence including, for example, recruitment of appropriate leukocytes and immune cells, up-regulation of pattern-recognition receptors for innate immunity including scavenger receptors and Toll-like receptors, positive feedback for continued production of proinflammatory cytokines, and increased synthesis of prostaglandins and leukotrienes. In contrast to this group of mediators, as part of the turn-off mechanism, antiinflammatory cytokines (e.g., IL-4, IL-10) inhibit production or block actions of the proinflammatory cytokines.

The initiating signals that stimulate production of proinflammatory molecules can range from peroxidation products from excessive lipid oxidation to mechanical stress or a combination of cell-specific insults. Regardless of the particular perturbation, nuclear factor-B (NF-B)/Rel transcription factors play pivotal roles in transducing a signal that leads to production of cytokines. The NF-B system also acts as mediator of inflammatory cytokine action, as do the MAPK signaling pathways and c-Jun N-terminal kinase (JNK1) (reviewed in Refs. 108, 109, 110, 111). In addition to cytokines such as TNF-, JNK1 also is activated by free fatty acids, and the activity of this kinase is abnormally elevated in mouse models of obesity (112).

Another key player in the regulation of inflammatory events is nitric oxide (NO), a diffusible uncharged gas with a half-life in seconds and having multiple molecular targets, for example cytokine production, inhibition of immune cell proliferation, antimicrobial activity, cyclooxygenase-2 induction, and vascular permeability and vasodilation (see Refs. 113, 114, 115 and references therein). Depending on its concentration and the cell- and event-specific activity, NO can have either pro- or antiinflammatory effects. NO production is regulated by isoforms of NO synthase (NOS): inducible NOS (iNOS), endothelial NOS (eNOS), and neuronal NOS (nNOS). All three isoforms operate in the immune system and are expressed in multiple cell types. Regulation of iNOS is via transcription activated by cytokines and costimulatory molecules signaling through several pathways, including the NF-B system. Activation of eNOS and nNOS predominantly is via a change in activity of existing protein triggered by a wide array of stimuli, including vasoactive substances, neurotransmitters, shear stress, and cytokines.

B. Estrogen and the inflammatory process
Experimental evidence for the role of estrogen in inflammation within the adipose/metabolic, cardiovascular, and neural systems will be reviewed in subsequent sections, but it is useful here to consider recent studies that establish the mechanistic potential for estrogen to affect the inflammatory process. ER and, in some cases, ERß are present in front line immune and cytokine-producing cells, such as macrophages and microglia, and E₂-activated ER has been shown in vitro to affect release of proinflammatory cytokines from these cells and to interfere with the action of cytokines (see Refs. 116, 117, 118, 119, 120 and references therein). These critical E₂ actions can be explained, at least in part, by the ability of ER to function as a transcriptional repressor by inhibiting the activity of NF-B through a protein-protein interaction of agonist-bound ER with a subunit of activated NF-B (121, 122, 123). The extent of the inhibitory action of E₂ on NF-B function may be context- and target gene-selective, thus opening intriguing possibilities for mechanism-based design of SERMs (30, 123, 124).

E₂ also is implicated in the activation of eNOS and nNOS as well as the regulation of expression of all three NOS isoforms in the immune, cardiovascular, and central nervous systems (see Refs. 37, 115 , and 125, 126, 127 and references therein). Both nuclear and extranuclear pathways are used by E₂ to affect the NOS isoforms. For example, the rapid release of NO in endothelial cells that results in vasodilation can be accomplished by E₂ activation of membrane ER and subsequent signaling through a phosphatidylinositol 3-kinase/Akt signaling pathway leading to eNOS activation (reviewed in Refs. 45 and 46). Progesterone has been shown to potentiate E₂ effects on the generation of NO in human endothelial cells in vitro, whereas MPA impairs E₂ signaling in this model (61).

In summary, the widespread presence and participation of ERs in multiple cell types of the immune system and the inflammatory response is remarkable and represents just one facet of the pleiotropic action of estrogens. Although much work remains to be done in defining the complex interaction of estrogens with the inflammatory process, the studies to date establish E₂ as a significant player, as will be explored in the following.

	IV. Adipose/Metabolic System and Estrogens

Top Abstract I. Introduction II. Pharmacology of Estrogens... III. The Inflammatory Process IV. Adipose/Metabolic System and... V. Cardiovascular System and... VI. Central Nervous System... VII. Summary/Conclusions References

 
A. Visceral adipose tissue
Visceral fat produces an array of hormones involved in energy homeostasis, the fibrinolytic system, the immune system, vascular homeostasis, and the inflammatory response. This multifunctional endocrine/paracrine organ has been the focus of many excellent reviews (for example, Refs. 128, 129, 130, 131), and only a few highlights will be considered here.

Recent interest has been sparked, in part, by the observation that increased visceral adiposity is closely associated with insulin resistance, hypertension, dyslipidemia, and CVD. Adipose-derived hormones (adipokines) involved in these disorders have been examined in clinical, animal knockout, and in vitro studies, and prominent on any list have been the inflammatory cytokines and related proteins (Fig. 1) (131, 132, 133). For many of the adipokines shown in Fig. 1, there is a direct relationship between visceral adipose mass and the amount of product. An important exception is adiponectin, which has an inverse relationship to fat mass; in addition, an inverse association between plasma adiponectin levels and both insulin resistance and chronic inflammation has been reported (131, 134).

View larger version (36K): [in this window] [in a new window]   FIG. 1. Overview of some of the endocrine and paracrine products of visceral adipose tissue. The products are grouped into broad functional categories with examples shown for each category. Many of the secretory products (adipokines) have targets within visceral fat as well as in the periphery. The enzyme 11ß-hydroxysteroid dehydrogenase-1 expressed in adipocytes is responsible for the conversion of circulating cortisone (inactive) to cortisol (active); adipose aromatase converts circulating androgens to estrogens. Receptors for cortisol and for E2 are present in adipocytes. PAI-1, Plasminogen activator inhibitor-1; FFA, free fatty acids; VEGF, vascular endothelial growth factor.

Rodent models have revealed that the selective surgical resection of visceral fat pads led to marked improvements in insulin resistance, which was not seen with the removal of equivalent amounts of sc adipose tissue (142), supporting the concept that sc adipose tissue is a less important determinant of insulin action and lipid metabolism than visceral adipose tissue. Consistent with the functional distinctiveness of adipose depots is the recent identification of visfatin, a novel adipokine preferentially produced by human abdominal visceral adipose that facilitates adipogenesis but intriguingly also acts as an insulin-mimetic (143, 144, 145). Although much work remains in establishing its physiological roles, visfatin offers new prospects in obesity research.

The regulation of visceral adipose activity is complex, which is not surprising given its multifunctional and integrative characteristics. In addition to receptors for multiple cytokines and metabolic hormones, visceral adipocytes express most members of the nuclear hormone receptor family, notably peroxisome proliferator-activated receptor- (PPAR), liver X receptor (LXR), thyroid hormone receptor, glucocorticoid receptor, AR, PR, and ER. An example of visceral adipose as a complex, functionally integrated unit comes from studies of glucocorticoid receptor and its ligand cortisol, which is produced locally from inactive cortisone by 11ß-hydroxysteroid dehydrogenase-1 (11ßHSD-1) (Fig. 1). This enzyme, which is highly expressed in human adipocytes, results in increased cortisol concentration within adipose without significantly affecting systemic cortisol levels (reviewed in Refs. 128, 146 , and 147). In addition to its effect on adipocyte lipolysis and sensitivity to insulin, cortisol promotes differentiation of human preadipocytes into mature adipocytes (148). Although questions remain about the regulation and dysregulation of 11ßHSD-1 and its role as cause or effect in visceral obesity (146, 149, 150), this cortisol-generating system within adipose is significant for its local integration of energy homeostasis under normal conditions and for its potential as a target for pharmaceutical manipulation.

B. Influence of estrogen on adipose tissue
1. Estrogen and body fat distribution.
It is now clear that sex steroid hormones are major determinants of body fat distribution and that reproductive hormones in general may be considered adiposity signals. There is a gender dimorphism in body fat distribution, because women generally accumulate sc fat in the gluteal and femoral regions, whereas men develop an android pattern of body fat distribution with weight gain. Women generally have a higher percentage of total body adiposity, whereas BMI-matched men have about twice as much visceral adipose tissue as premenopausal women (151). The lower amount of visceral adipose tissue in women is thought to contribute to the lower prevalence of dyslipidemia, hypertension, diabetes, and CVD in premenopausal women compared with men and postmenopausal women.

The influence of hormones on body fat distribution appears to be related both to adipose tissue-specific expression of steroid receptors and to local tissue steroid hormone metabolism. ER and ERß are expressed in both sc and visceral fat tissues; however, ERß appears to be preferentially expressed in sc adipose (152). Steroidogenic enzymes that metabolize sex steroid hormones, e.g., 17-hydroxylase and aromatase, are differentially expressed in sc and visceral adipose depots and regulate local steroid conversion (153). Adipose aromatase itself can be regulated, for example, by locally produced cytokines and prostaglandins (154). In human adipose stromal cells, cortisol produced locally from cortisone is a potent stimulator of the conversion of androgens to estrogens through an action on aromatase (155).

Lipoprotein lipase (LPL), a triglyceride hydrolase that directs the deposition of triglyceride into adipocytes cells, is thought to play an important role in regional body fat distribution. Tchernof et al. (156) recently showed regional differences in adipocyte metabolism with menopausal status, independent of age, body fat mass, and visceral adipose tissue accumulation. The authors found that omental adipocytes from postmenopausal women were larger and had higher LPL activity compared with premenopausal women, but they found no menopause-related differences in sc adipocytes, reflecting a shift toward visceral fat storage. A role for estrogen in LPL regulation is supported by studies in aromatase knockout mice, which have a marked increase in visceral adiposity associated with an increase in LPL expression and a metabolic syndrome phenotype. E₂ treatment resulted in a marked decrease in visceral fat mass and a profound inhibition of LPL expression in the aromatase knockout (reviewed in Ref. 157).

2. Estrogen deficiency.
Menopausal estrogen deficiency is associated with significant changes in metabolic parameters that appear to be partially related to the shifts in body fat distribution with menopause (158). Premenopausal women have a less "atherogenic" lipid profile than men due to higher HDL, higher levels of large antiatherogenic HDL₂ subspecies, and lower triglyceride levels, all of which are closely associated with lower central fat accumulation. Menopause, either natural or surgical, is coupled with rapid adverse changes in lipid metabolism, because reduced HDL and HDL₂ and increased triglyceride and low-density lipoprotein (LDL) are seen within 3 months of amenorrhea (159, 160). The rapidity of the lipid changes implies that there may be both direct effects of sex steroid hormones and indirect effects of visceral fat accumulation on lipid metabolism.

Cross-sectional (156, 161) and longitudinal studies (162, 163) have shown that the menopausal transition is associated with an increase in visceral adiposity, and this effect of menopause is independent of the effect of age and total body adiposity or BMI (164, 165, 166). Establishing a mechanistic relationship in humans between the absence of E₂ and an increase in visceral adipose has been elusive, due in part to the complexities contributed by a paradoxical increase in serum E₂ associated with increases in adiposity in older menopausal women (see Ref. 167 and references therein). An additional factor in attempting to sort out the cause/effect role of estrogen and adiposity is the negative correlation of serum SHBG levels with BMI in peri- and postmenopausal women, resulting in a greater fraction of bioactive E₂ in circulation in obese individuals (168, 169). Another consideration is the source of postmenopausal E₂, which primarily is from extragonadal conversion of testosterone by aromatase. Consistent with the positive correlation of serum E₂ and BMI in postmenopausal women is the observation that adipose aromatase activity increases with age (170). So influential is the regulation by aromatase that the argument has been made that, in postmenopausal women, E₂ functions not as a circulating hormone but as a paracrine factor at its sites of production (e.g., adipose, brain, breast, and bone) (157). Clearly, the availability of testosterone and its derivation from adrenal and ovarian androstenedione and dehydroepiandrosterone must be major considerations in attempting to decipher the complex relationship of E₂, body weight, and metabolism in menopausal women.

3. Exogenous estrogen.
Contrary to popular belief, neither postmenopausal ET nor OCs have been shown to cause weight gain (171, 172), and OCs do not appear to have an effect on body fat distribution (173). However, oral ET has been shown to have a modest effect in reducing the postmenopausal weight gain (174). Mattiasson et al. (175) showed a selective reduction in visceral fat in early postmenopausal women treated with oral E₂ plus cyclic MPA that was not seen in the placebo-treated women. The HERS and WHI studies showed that combination HT (oral CEE + MPA) led to improvements in BMI and waist circumference in women with (372) and without (176) coronary heart disease.

Rodents rapidly become obese after ovariectomy, and E₂ administration is known to prevent the increase in adiposity (177). Recent data from D’Eon et al. (178) showed that E₂ treatment in ovariectomized mice promoted a reduction in adipose tissue mass and adipocyte size compared with pair-fed ovariectomized controls. The reductions in adiposity were seen in the intraabdominal fat but not the sc fat depots. The authors found increased levels of adipocyte lipolysis and down-regulation of associated genes such as LPL and LXR- in intraabdominal fat from the E₂-treated mice. Conversely, in muscle they found that E₂ led to up-regulation of LPL and muscle-specific PPAR, as well as several of its downstream targets, suggesting that E₂ promotes the use of lipid as a fuel by promoting fat oxidation in the muscle and enhancing adipocyte lipolysis (178).

If ET is associated with decreased visceral adipose, the critical question is whether there are beneficial metabolic consequences in humans. When insulin sensitivity is used as an endpoint, results are conflicting with reports of either increases, decreases, or no change depending on the age of the women, treatment route, and methodology for assessing insulin resistance (175, 176, 179). Similar inconsistencies are found with other endpoints and underscore the complexity of interactions between the paracrine and the endocrine functions of adipose, their integration with liver metabolic activity and inflammation, and the multiple estrogen targets within these tissues. Studies to disentangle this complex area are beginning to produce some answers, but much work remains in understanding the roles for estrogen in these relationships.

C. The influence of estrogen on markers of inflammation
Acute phase proteins serve as examples of the interface between adipose tissue and the inflammatory response. Markers of subclinical inflammation, such as C-reactive protein (CRP), IL-6, and serum amyloid A (SAA), have been linked to insulin resistance, diabetes mellitus 2, and CVD. In vivo and association studies have shown a strong link between obesity and increased inflammatory proteins, and obesity is associated with increased infiltration of adipose tissue by macrophages (180, 181).

CRP is a marker of the presence of subclinical inflammation that has been proposed to be an independent risk factor for cardiac events (182) and is a hallmark of the inflammatory processes that convert "fatty streak" plaques into complex atherosclerotic lesions. Although CRP is synthesized mainly in the liver, several extrahepatic sources of CRP have been reported, for example, adipocytes, macrophages, coronary artery smooth muscle cells, and neurons (131, 183). Circulating CRP levels are positively associated with total body fat and visceral adipose, and weight loss by hypocaloric diet or surgical intervention has been shown to reduce CRP levels (reviewed in Ref. 131). Oral estrogen significantly increases CRP levels through the first-pass hepatic effect, and some have hypothesized that increased CRP may have played a role in the higher rates of cardiovascular events and stroke observed in the HERS (10) and WHI (11) trials. Transdermal E₂, however, has no effect on systemic CRP levels (Table 2). It would be important to determine whether E₂ can affect local production of CRP within adipose or other target sites.

IL-6 is a proinflammatory cytokine produced at many sites, for example monocytes/macrophages, microglia, bone, and adipose, and has multiple functions in the inflammatory cascade, including increasing CRP and SAA production. Elevated IL-6 levels have been associated with increased BMI and risk of cardiovascular death (184). IL-6 has been shown to be preferentially secreted by visceral adipocytes, and removal of sc fat by liposuction did not reduce IL-6 levels (141). In addition to the positive association with BMI, circulating levels of IL-6 correlate positively with age (116, 186). Whether menopausal ET can reverse elevated serum IL-6 levels has been the subject of many studies with conflicting results, depending on treatment formulations and the age and BMI status of the women (116, 186, 187, 188, 189). For example, data from the Postmenopausal Estrogen Progestin Intervention (PEPI) trial and other randomized clinical trials have shown no significant effect of oral ET (CEE), with or without progestin, on IL-6 levels (188), but several observational studies have shown variable effects (no change, reduction, or elevation) on IL-6 levels in postmenopausal women using oral ET (reviewed in Ref. 190). In contrast, in vitro studies establish that E₂ can decrease IL-6 production and interfere with its actions in several cell types secondary to a cell-type specific disruption of NF-B signaling (116, 123, 191, 192). That E₂ modulates IL-6 activity primarily at the local tissue level is an important consideration and requires further study.

SAA is a marker of systemic inflammation that predicts future cardiovascular events and responds simultaneously with CRP to inflammatory stimuli. The SAA family is a group of differentially expressed apolipoproteins that are synthesized primarily in the liver and associate with HDL particles (187). The association of elevated SAA with HDL has been hypothesized to promote a proatherogenic phenotype by impairing reverse cholesterol transport (193). During inflammation, SAA-associated HDL particles become depleted of ApoA-1 (the apolipoprotein of the more antiatherogenic HDL subspecies), have a lower affinity for hepatocytes than HDL, and induce activity of phospholipases that lower HDL levels (187). Like CRP and IL-6, SAA is positively associated with BMI and waist circumference and has been found to be expressed in adipose tissue (194). Both IL-6 and SAA are up-regulated in adipocytes in the insulin-resistant state (see Ref. 129 and references therein). Abbas et al. (195) recently compared the effects of 8 wk of oral estrogens (CEE) or transdermal E₂ to placebo in postmenopausal women. They found that oral estrogen significantly increased total SAA and HDL-associated SAA levels (proatherogenic phenotype), whereas transdermal E₂ resulted in a significant reduction in SAA levels compared with placebo. Thus, presumed beneficial effects of oral estrogens resulting from an increase in HDL particles may be obviated by the parallel increase in SAA and interference with HDL function. The observation that transdermal E₂ is associated with a decrease in SAA reinforces the need to establish the action of E₂ on the production of acute phase proteins at nonhepatic sites.

	V. Cardiovascular System and HT

Top Abstract I. Introduction II. Pharmacology of Estrogens... III. The Inflammatory Process IV. Adipose/Metabolic System and... V. Cardiovascular System and... VI. Central Nervous System... VII. Summary/Conclusions References

 
In postmenopausal women, the effect of HT on cardiovascular events has become an intensely controversial issue. Several major recent clinical trials, especially the WHI study, have caused basic and clinical scientists to reconsider the fundamental biology and physiological complexities that impact upon HT in postmenopausal women (3). The lack of an effect of HT on cardiovascular endpoints in WHI and several other recent prospective clinical trials has caused confusion and controversy because the outcomes of these trials seem in opposition to a large body of observational data that demonstrate a protective cardiovascular effect of hormone replacement in postmenopausal years. However, recent clinical trials like the WHI studied cardiovascular endpoints after de novo initiation of HT to a mostly older postmenopausal population without a prior history of hormone use, which contrasts sharply with most observational cohorts studied in which the initiation of HT occurred in younger women who usually were enrolled in these trials for treatment of active perimenopausal symptoms.

The strengths and weaknesses of the WHI trial have been reviewed recently (3). Consensus now has emerged that it is important to acknowledge the strengths and limitations of WHI, but also to move past the study by emphasizing the enormous need that exists to strengthen interactions between preclinical and clinical research scientists and to improve our understanding of menopause and of the biology of age and gender differences in the cardiovascular system (3, 196, 197). The biology of the perimenopause remains poorly understood at present, and enormous gaps remain in our knowledge of the gender differences of relevance to cardiovascular function that may occur during this transition (197). It remains unclear, for example, whether or not there is a critical window for hormonal treatment that will confer cardioprotection, whether the duration of any such therapeutic advantage conferred by HT is limited, and what postmenopausal time course underlies the loss of efficacy and/or emergence of adverse effects seen with administration of HT to older women. The cardiovascular implications of discontinuous vs. continuous administration of HT also are unclear, as are the hormonal formulations and selective receptor modulators that may optimize cardiovascular functions and protection (197, 198). To address these issues, it will be necessary to understand the mechanisms of sex steroid hormone effects on cardiovascular tissues in both normal physiology and disease.

A. Sex steroid hormone receptor action and cardiovascular physiology
Steroid hormones all activate members of the nuclear hormone receptor superfamily of ligand-activated transcription factors. The biology of these receptors has been reviewed in Section II.A and will not be discussed in detail here. The general presence of functional sex steroid hormone receptors (SSHR) in the cardiovascular system is well established, their expression in both heart and blood vessels having been recognized for several decades (reviewed in Refs. 199, 200, 201, 202). E₂, progesterone, testosterone, and dihydrotestosterone all bind specifically to atrial and ventricular myocardial fibers and to vascular endothelial and smooth muscle cells of human and nonhuman primates (203, 204, 205, 206, 207, 208, 209). In a classic study, specific saturable receptors for estrogens, androgens, and glucocorticoids were demonstrated in canine coronary arteries, and PRs, absent at baseline, were induced by treatment with physiological E₂ concentrations (205). PRs also are expressed early in life, in human term fetoplacental vessels (210). ER and ERß expression in vascular smooth muscle and endothelial cells is regulated by injury and by E₂ (see Refs. 199, 200 , and 202 and references therein)_. Aromatase, the enzyme that produces estrogens from androgens, is expressed in primate coronary arteries (211) and human venous and arterial tissues, including vascular smooth muscle cells and, in some but not all studies, endothelial cells (212, 213, 214, 215, 216). Vascular aromatase mediates normal vasomotion in males, and normal young males given aromatase inhibitors develop endothelial dysfunction that is reversible upon cessation of the drug (202, 217, 218, 219). Surprisingly, the effects of aromatase inhibition on vascular function in females have not yet been studied, despite the increasing use of these compounds in the therapy of breast cancer. Furthermore, the precise distribution and regulation of expression of SSHR and aromatase in heart and vascular tissues is not known, nor is it clear yet whether vascular beds differ, whether there are gender differences in the levels of expression of these proteins, and whether their level of expression is modified by the presence of atherosclerosis.

The hypothalamic-pituitary-gonadal axis undergoes perinatal maturation. Mammalian hypothalamic-pituitary-gonadal axis function begins in utero, when testosterone induces male reproductive organ development, whereas ovarian endocrine activity begins after birth (220). However, the cardiovascular developmental effects of sex steroid hormones and the time course of expression of SSHR in the development of cardiovascular tissues have not been studied. An important hypothesis in maternal-fetal medicine, the Barker or fetal programming hypothesis arose from recognition of a statistical correlation between conditions prevailing at the time of birth and morbidity and mortality due to chronic diseases late in life, such as the strong and inverse relationship observed between coronary heart disease mortality and perinatal weight (221). However, little data examining this hypothesis at the level of cardiovascular endocrinology and physiology exist at present. One recent epidemiological report suggested that low birth weight is only associated with increased carotid intimal medial thickening in young adulthood in subjects who have experienced both severe intrauterine growth retardation and exaggerated postnatal growth (222).

Another interesting perinatal issue is the recently described presence of fetal cells in myocardial tissues of the mother (223). The potential biological effects of fetal microchimerism on adult physiology have been suggested, but the cardiovascular role of these cells in gender differences for vascular or myocardial disorders also has not been explored.

Sex steroid hormones are critical determinants of cardiovascular gender differences. The majority of research has focused on the effects of estrogens and ER on cardiovascular physiology and disease, whereas progesterone and testosterone, their receptors, and the enzyme Cyp19 or aromatase all have received far less attention (202). Although SSHR interactions with their respective hormones, transcriptional coregulatory proteins, and specific DNA response elements have been studied extensively in reproductive target organs, these interactions have not yet been explored sufficiently in cardiovascular physiology. One common mechanism regulating hormone action that is likely to be of major importance for cardiovascular physiology involves coactivator and corepressor proteins that interact directly with the SSHR. The regulation of these coregulator proteins, in turn, through posttranslational modifications allows cells to alter the genes regulated by SSHR and the timing of transcriptional events (3, 21, 43). Coregulator biology also is important to the development of novel cardiovascular selective SSHR modulators (43, 198). Examples of cardiovascular coregulator specificity include the relatively specific myocardial AR coactivator, FHL2 (224) and the vascular role of the coactivator protein steroid receptor coactivator 3 (SRC-3; also known as AIB1, pCIP, ACTR, and TRAM-1) in mediating estrogen inhibition of vascular injury (225).

Yuan et al. (225) recently studied the role of SRC-3 in estrogen-mediated vascular protection after vascular injury. These authors created mice harboring a knock-in of the LacZ reporter into the SRC-3 gene. Heterozygous mice were phenotypically normal and provided a sensitive marker to characterize SRC-3 expression within the cardiovascular system, where SRC-3 expression was noted in vascular smooth muscle cells and endothelial cells but not in myocardial cells. A carotid ligation injury model was used to examine the extent of neointima formation in mice homozygous for SRC-3 disruption (SRC-3^–/– mice). Vascular injury was more exuberant in SRC-3^–/– mice than in wild-type mice, and this difference was diminished after ovariectomy. E₂ treatment after ovariectomy in wild-type mice inhibited the neointimal response, as has been reported previously (226, 227), but the ability of E₂ to inhibit the neointimal response and cellular proliferation in SRC-3^–/– mice was attenuated. Thus, the loss of SRC-3 function interfered with the protective effects of estrogen in this vascular injury model, suggesting the potential importance of this coactivator in vascular biology. SRC-3 has been shown recently to integrate genomic responses to multiple cellular signaling pathways (228), suggesting that it may be an especially critical coactivator deserving of further study in cardiovascular cells and tissues (Fig. 2).

View larger version (75K): [in this window] [in a new window]   FIG. 2. Overview of the vascular and blood cell types that are influenced by estrogens. The effects of estrogens on the blood vessel wall depend upon the extent and complexity of atherosclerotic disease present at the time therapy is initiated. As atherosclerosis evolves, the early vascular protective mechanisms of estrogen ("early HT"; e.g., increased NO and cyclooxygenase-2 and decreased TNF-, cell adhesions molecules, LDL oxidation/binding, platelet activation, and vascular smooth muscle cell proliferation) recede and are replaced by estrogen responses in the setting of underlying advanced and unstable lesions that may be deleterious ("delayed HT"). The pleiotropic effects of HT on the vascular endothelium, vascular smooth muscle cells, and inflammatory cells thus differ, depending on the stage of atherosclerosis in the underlying blood vessel (the Timing Hypothesis, Ref. 202 ). The relative roles of steroid hormone receptors, other than ERs, and of the coregulatory molecules involved in nuclear receptor action in the vasculature remain less well understood.

Another type of nuclear hormone receptor that has not yet been adequately explored for regulation by sex steroid hormones and as a contributor to the gender differences in CVDs is the LXRs that regulate cholesterol homeostasis. Like the SSHR, LXRs are members of the nuclear hormone receptor superfamily of ligand-activated transcription factors, and there is evidence that LXR expression can be affected by E₂ (232, 233, 234) (see below). The LXRs regulate genes whose protein products control cholesterol homeostasis, such as proteins in the reverse cholesterol transport pathway involved in cholesterol removal from peripheral tissues, as well as proteins involved in the breakdown and excretion of cholesterol from the body, which can directly influence the rate and progression of atherosclerosis (232). LXR- and LXR-ß, the two known LXRs, are highly expressed in liver, adipose tissue, and macrophages, and LXR-ß is expressed as well in many other tissues. Lundholm et al. (234) have recently used a gene profiling approach to examine estrogen-regulated genes in mouse adipose tissue. In these studies, E₂ decreased mRNA expression of LXR, as well as several known LXR target genes involved in cholesterol homeostasis, including sterol regulatory element-binding protein 1c, apolipoprotein E, and the ATP-binding cassette A1. That E₂ can decrease expression of LXR and several of its target genes in adipose tissue suggests that further study of LXR regulation by sex steroid hormones may be fruitful. This sort of crosstalk between SSHR and other nuclear hormone receptors important to cardiovascular physiology and disease deserves greater attention (Fig. 2). In summary, better understanding of the emerging regulatory mechanisms for SSHR is central to advancing our understanding of sex steroid hormones in cardiovascular physiology and disease (202).

B. Vascular effects of sex steroid hormones, atherosclerosis, and ischemic cardiovascular diseases
The importance of the timing of HT to the effects on vascular function has been reviewed recently (202). The largest body of scientific evidence in the field, the seminal primate studies of Clarkson and colleagues (89, 235), supports the conclusion that HT can be beneficial when initiated in the early stages of atherogenesis but that its effects are lost in the setting of more advanced atherosclerosis (202) (Fig. 2). Sex steroid hormone therapies will require better attention to timing, as well as dose and formulation of hormone, like other hormonal replacement therapies for endocrine organs that are losing or have lost their capacity to synthesize hormone (such as in thyroid disease). The effects of endogenous and exogenous hormones on cardiovascular SSHR have been d