| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Endocrinology and Oncology Research Center, Laval University Medical Center (Centre Hospitalier de l’Université Laval) and Laval University, Québec City, Québec G1V 4G2, Canada
Correspondence: Address all correspondence and requests for reprints to: Prof. Fernand Labrie, Molecular Endocrinology and Oncology Research Center, Laval University Medical Center (Centre Hospitalier de l’Université Laval), 2705 Laurier Boulevard, Québec City, Québec G1V 4G2, Canada. E-mail: fernand.labrie{at}crchul.ulaval.ca
 |
Abstract |
|---|
 Top Abstract I. Androgens and Their...II. DHEA Is Predominantly...III. Androgens Inhibit Breast...IV. DHEA Inhibits Breast...V. Rationale for the...References |
|---|
|
I. Androgens and Their Role in Women |
|---|
|
TopAbstractI. Androgens and Their...II. DHEA Is Predominantly...III. Androgens Inhibit Breast...IV. DHEA Inhibits Breast...V. Rationale for the...References |
|---|
Despite the well known beneficial effects of estrogen therapy on menopausal symptoms (9 10 11 ) and their role in reducing bone loss and possibly coronary heart disease (12 13 14 15 16 17 ), compliance is low. The majority of women decide not to take estrogens or stop treatment early because of the fear of breast and uterine cancer (11 ) and of symptoms associated with this therapy, namely uterine bleeding, breast tenderness, and fluid retention.
The almost exclusive focus on the role of ovarian estrogens at menopause has removed the attention from the progressive and dramatic fall in circulating dehydroepiandrosterone (DHEA), which starts early at the age of 30–40 yr (18 19 20 21 22 23 ). Because DHEA is transformed into both androgens and estrogens in peripheral tissues, such a fall in the serum concentration of the steroid precursors DHEA and DHEA sulfate (DHEA-S) explains why postmenopausal women, as discussed later, are not only lacking estrogens but are also deprived from androgens. Moreover, women taking contraceptives or estrogen replacement therapy (ERT) have reduced ovarian androgen secretion attributable to inhibition of gonadotropin secretion, as well as reduced androgen bioavailability attibutable to increased SHBG levels (24 ).
B. Decrease of serum DHEA, androgens, and their metabolites with age
Until recently, because of assay difficulties, only a limited number of circulating adrenal and gonadal steroids have been measured during advancing age, thus limiting the evaluation of the relative role of different sources of sex steroids. This analysis is of special importance in postmenopausal women in whom the sex steroids of adrenal origin gain particular importance after the arrest of estrogen secretion by the ovaries at menopause (25 ). It is important to recall that in the 50- to 60-yr-old age group, serum DHEA has already decreased by 70%, compared with the 20- to 30-yr-old peak values (Ref. 23 ; Fig. 1
). It is thus quite remarkable that most of the important decline in circulating DHEA, DHEA-S, androst-5-ene-3ß,17ß-diol (5-diol), 5-diol-G, androstenedione (4-dione), and the conjugated metabolites of androgens, namely androsterone glucuronide (ADT-G), androstane-3
,17ß-diol glucuronide (3-diol-G), and androstane-3ß,17ß-diol glucuronide (3ß-diol-G), occurs between the age ranges of 20–30 and 50–60 yr, whereas relatively smaller changes occur after the age of 60 yr (23 ). It is important to realize, as illustrated in Fig. 2, not only that serum DHEA and DHEA-S decrease by 50% between the ages of 21 and 40 yr but also that a similar decrease is observed for serum testosterone (26 ). Such data could well suggest that HRT with androgens should start early at menopause to compensate for this early fall in the secretion of androgen precursors by the adrenals and the parallel decrease in serum testosterone.
|
|
-diol-G, and 3ß-diol-G as estimates of total androgens, the average sum of the serum concentrations of these conjugated metabolites of dihydrotestosterone (DHT) are 37.5, 8.47, and 30.2 nM in men compared with 32.5, 4.28, and 17.3 nM in women (23 ). The average serum concentrations of ADT-G, 3-diol-G, and 3ß-diol-G, measured in women between the ages 20 and 80 yr are thus 86.6% (ADT-G), 50.5% (3-diol-G), and 57.2% (3ß-diol-G), compared with those found in men of the same age (20–80 yr; Table 1; Ref. 23 ). Although the metabolic clearance rates of the three main androgen metabolites are likely to show differences between men and women, an estimate of the relative amount of total androgens in women and men calculated on the basis of the sum of the serum concentrations of these three metabolites suggests that total androgen production in women is more than two thirds, or 71%, of that observed in men (23 27 ). Such an approach is based on the knowledge that active androgens are inactivated to glucuronide derivatives before their diffusion from the intracellular compartment into the circulation where they can be measured as ADT-G, 3-diol-G, and 3ß-diol-G.
|
After the recognition that such a large proportion of androgens and estrogens in men and women originate from DHEA and DHEA-S of adrenal origin (25 ), we have studied the serum concentration of a large series of androgens and estrogens as well as their metabolites after percutaneous administration of DHEA in 60- to 70-yr-old men and women (27 ). We then observed that changes in serum DHEA within the physiological range of young adult men and women led only to small or nonsignificant changes in serum testosterone, DHT, or E2, whereas, on the other hand, the concentration of the conjugated metabolites of DHT were markedly increased (27 ). Such data clearly indicate the poor value of measurements of serum androgens and estrogens as parameters of total androgenic and estrogenic activities in men and women.
As well demonstrated in a long series of preclinical studies, supplementation with physiological amounts of exogenous DHEA permits the biosynthesis of androgens (essentially testosterone and DHT) and estrogens only in the target tissues that contain the specific steroidogenic enzymes (25 33 ). The widespread tissue distribution of steroidogenic enzymes is illustrated in Table 2 (34 ). In fact, in 22 peripheral tissues of the monkey, steroid sulfatase, 3ß-hydroxysteroid dehydrogenase (HSD), androgenic 17ß-HSD, estrogenic 17ß-HSD, aromatase, and 5-reductase are all present in 114 of 132 (86%) possible sites. Genomic studies are in progress to determine the identity of all families of steroidogenic enzymes in the various peripheral target tissues.
|
-diol-G, and 3ß-diol-G, these metabolites being produced locally in the peripheral intracrine tissues that possess the appropriate steroidogenic enzymes to synthesize DHT from the adrenal precursors DHEA and DHEA-S. These peripheral target tissues also contain the steroid-inactivating enzymes required to metabolize DHT into inactive and more water-soluble conjugates, especially glucuronide derivatives (25 35 ). Such local biosynthesis and action of androgens in target tissues eliminates the exposure of other tissues to androgens and thus minimizes the risks of undesirable masculinizing or other androgen-related side effects. The same applies to estrogens, although we feel that a reliable parameter of total estrogen secretion (comparable to the glucuronides for androgens) has not yet been identified.
C. Androgens and bone physiology
1. Role of androgens and estrogens in bone physiology.
A predominant role of androgens in bone physiology has already been suggested (36 ). In fact, both testosterone and DHT increased the transcription of (I) procollagen mRNA in osteoblast-like osteosarcoma cells (37 ). Treatment with DHT has also been shown to stimulate endochondral bone development in the orchiectomized rat (38 ). Androgens stimulate osteoblast differentiation, these cells being known to contain androgen receptors (AR; Refs. 39 40 41 ). Moreover, bone mineral density measured in the lumbar spine, femoral trochanter, and total body was increased more by estrogen plus testosterone implants than by E2 alone over a 24-month treatment period in postmenopausal women (42 ). In agreement with these data, biomarkers of bone formation were increased compared with estrogen alone when methyltestosterone was added to estrogen (43 ).
The essential role of androgens in bone mineralization is illustrated by the reduced bone mineral density in patients with the androgen insensitivity syndrome (44 45 46 ). In such patients having an inactive AR, estrogens are unable to increase bone mineral density (44 45 ). Thus, at doses of estrogen able to restore bone mineral density in hypogonadal women, estrogens could not exert a similar effect in patients with androgen insensitivity. Such data suggest that both estrogens and androgens are required to acquire normal bone mineral density. In fact, a correlation has been found between androgens and bone mineral density in premenopausal women (31 47 ).
In established osteoporosis, anabolic steroids have been reported to help prevent bone loss (48 ). Moreover, androgen therapy, as observed with nandrolone decanoate, increases vertebral bone mineral density in postmenopausal women (49 ). Similarly, sc E2 and testosterone implants have been found to be more efficient than oral estrogen in preventing osteoporosis in postmenopausal women (50 ). Although the difference has been attributed to the different routes of administration of the estrogen, the cause of the difference could well be the action of testosterone. Studies have convincingly shown that androgen plus estrogen was more efficient than estrogen in improving bone mineral density in postmenopausal women (42 43 50 52 53 54 55 ).
Although androgens are gaining increasing support because of their unique actions in postmenopausal women, virilizing effects are observed with the use of supraphysiological doses of testosterone (56 57 ). The availability of a compound such as DHEA, an inactive precursor that is transformed into active androgens only in specific target tissues, would be an important advantage over androgens exerting systemic effects in all tissues possessing AR.
D. Other roles of androgens in women
1. General.
It is likely that the androgens produced from DHEA have other beneficial effects in postmenopausal women. The detailed benefits of androgens added to ERT or HRT have been described on general well-being, energy, mood, and general quality of life (58 59 ). Improvements in the major psychological and psychosomatic symptoms, namely irritability, nervousness, memory, and insomnia, have been reported after addition of androgens to ERT (60 ). In addition, androgenic compounds have been found to be beneficial for the treatment of the mastalgia frequently caused by HRT (61 ). In fact, ERT may result in severe breast pain that may lead to discontinuation of therapy.
2. Libido and sexual satisfaction.
Loss of libido and/or sexual satisfaction are common in early postmenopause. The addition of androgens to HRT is known to have beneficial effects on these problems (42 53 57 58 62 63 64 ). Moreover, a series of studies have shown the beneficial effects of androgens on libido in postmenopausal women (42 65 66 67 ). In women who have undergone oophorectomy and hysterectomy, transdermal testosterone improves sexual function and psychological well-being (68 ). Similar findings have been observed with DHEA administered to women with adrenal insufficiency, this steroid being the most important precursor of androgens in postmenopausal women (69 ). On the other hand, mood and fatigue were significantly improved after DHEA replacement therapy in Addison’s disease (70 ).
3. Hot flashes.
The addition of androgens has been found to be effective in relieving hot flashes in women who had unsatisfactory results with estrogen alone (71 ). Androgen therapy is also successful in reducing hot flashes in hypogonadal men (72 ). In agreement with its transformation into androgens (27 ), DHEA has been found useful in reducing hot flashes (73 74 ). In fact, marked improvements in the vasomotor symptoms were observed in early postmenopausal women who received 50 mg DHEA orally daily from an average score of 18.4 before treatment to a score of 4.5 at 6 months (74 ).
4. Cardiovascular function and lipids.
There is also evidence that androgens may improve endothelium-dependent and -independent vasodilation in postmenopausal women (75 ). In fact, parenteral testosterone therapy improved brachial artery vasodilatation in postmenopausal women using long-term estrogen therapy. It is also of great interest that the addition of parenteral testosterone does not negate the favorable effects of estrogen on low-density lipoprotein cholesterol (76 ).
|
II. DHEA Is Predominantly Converted into Androgens in Women |
|---|
|
TopAbstractI. Androgens and Their...II. DHEA Is Predominantly...III. Androgens Inhibit Breast...IV. DHEA Inhibits Breast...V. Rationale for the...References |
|---|
). Adrenal secretion of DHEA and DHEA-S increases during adrenarche in children at the age of 6–8 yr, and maximal values of circulating DHEA-S are reached between the ages of 20 and 30 yr. Thereafter, serum DHEA and DHEA-S levels decrease markedly (Fig. 1
; Refs. 18 and 20 21 22 ). In fact, as mentioned earlier, at 70 yr of age, serum DHEA-S levels are decreased to approximately 20% of their peak values, whereas they can decrease by 95% by the age of 85–90 yr (22 ). The 70–95% reduction in the formation of DHEA and DHEA-S by the adrenals during aging results in a dramatic reduction in the formation of androgens and estrogens in peripheral target tissues. Such a marked decrease in the formation of sex steroids in peripheral tissues could well be involved in the pathogenesis of diseases associated with aging.
|
Transformation of the adrenal precursor steroids DHEA-S and DHEA into androgens and/or estrogens in peripheral target tissues depends upon the level of expression of the various steroidogenic and metabolizing enzymes in each of these tissues. This sector of endocrinology that focuses on the intracellular hormone formation and action has been called intracrinology (Refs. 25 and 78 ; Fig. 4). This situation of a high secretion rate of adrenal precursor sex steroids in men and women is thus completely different from all animal models used in the laboratory, namely rats, mice, guinea pigs, and all others (except monkeys) in which the secretion of sex steroids takes place exclusively in the gonads (77 80 ). A major problem that is at least partially responsible for the delayed progress in the recognition of the formation of a major proportion of sex steroids in peripheral tissues or intracrinology is the fact that the animal models usually used in the laboratory do not secrete significant amounts of adrenal precursor sex steroids, thus focusing all attention on the testes and ovaries as the exclusive sources of androgens and estrogens. The term intracrinology was thus coined (78 ) to describe the synthesis of active steroids in peripheral target tissues in which the action is exerted in the same cells where synthesis takes place without release of the active steroids in the extracellular space and general circulation (25 ).
|
B. Structure of the human steroidogenic enzymes
As mentioned above, transformation of DHEA and DHEA-S into active androgens and/or estrogens in peripheral target tissues depends on the level of expression of the various steroidogenic and metabolizing enzymes in each cell type. Elucidation of the structure of most of the tissue-specific genes that encode the steroidogenic enzymes responsible for the transformation of DHEA and DHEA-S into androgens and/or estrogens has permitted rapid progress in this area (Refs. 33 and 83 84 85 86 ; Fig. 5). The major importance of DHEA and DHEA-S is illustrated by the finding that approximately 50% of total androgens in the prostate of adult men derive from these adrenal precursor steroids (77 87 88 ). Our best estimate of the intracrine formation of estrogens in peripheral tissues in women is in the order of 75% before menopause and close to 100% after menopause (25 ). Although testosterone of ovarian and adrenal origin can act directly in peripheral tissues, its transformation into estrogens requires the action of the peripheral or intracrine steroidogenic enzymes, especially aromatase (89 ).
|
-reductase, and/or aromatase (Fig. 5).
1. Human 3ß-HSD isoenzymes and their genes.
Despite its essential role in the biosynthesis of all classes of hormonal steroids, the structure of the 3ß-HSD/
5-4 -isomerase gene family, hereafter called 3ß-HSD, was only elucidated relatively recently (84 93 94 95 96 ). The membrane-bound enzyme 3ß-HSD catalyzes an essential step in the transformation of all 5-pregnen-3ß-ol and 5-androsten-3ß-ol steroids into the corresponding 4-3-keto-steroids, namely progesterone as well as the precursors of all androgens, estrogens, glucocorticoids, and mineralocorticoids.
Experiments performed using microsomes and purified enzymes show that 3ß-HSD can catalyze the interconversion of 3ß-hydroxy- and 3-keto-5-androstane steroids (97 ). On the other hand, experiments performed under more physiological conditions (i.e., in intact transfected cells in culture without added cofactor) indicate that 3ß-HSD catalyzes almost exclusively the oxidation of 3ß-hydroxy- into 3-keto-5-androstane steroids (98 ). The reverse reductive reaction is catalyzed by another enzyme, namely 3(
ß)-hydroxysteroid epimerase [3(ß)-HSE; Refs. 98 and 99 ] and type 7 17ß-HSD (our unpublished data).
3ß-HSD is found not only in the classical steroidogenic tissues (placenta, adrenal cortex, ovary, and testis) but also in several peripheral tissues, including the skin, adipose tissue, breast, lung, endometrium, prostate, liver, kidney, epididymis, and brain (34 84 91 100 ), thus catalyzing the first step in the intracrine transformation of DHEA into 4-dione, the precursor of both androgens and estrogens. The existence of multiple members of the 3ß-HSD gene family offers the unique possibility of tissue- and/or cell-specific expression of this enzymatic activity.
After purification of 3ß-HSD from human placenta and development of antibodies against the enzyme in rabbits (101 ), we have isolated and characterized a first 3ß-HSD cDNA type (93 ) and its corresponding gene (94 ). The second 3ß-HSD cDNA type, which corresponds to the almost exclusive mRNA species expressed in the adrenals and gonads, was chronologically designated human type 2 3ß-HSD (95 ). The structure of the corresponding human type 2 3ß-HSD gene has also been elucidated (96 ). The human 3ß-HSD genes corresponding to human cDNAs type 1 and 2 contain four exons and three introns within a total length of 7.7–7.8 kb. These genes were assigned by in situ hybridization to the p13.1 region of chromosome 1 and are closely linked to D1S514 located at 1–2 cM of the centromeric marker D1Z5 (102 ).
We have observed that mutations in the type 2 3ß-HSD gene are responsible for classic 3ß-HSD deficiency, a form of congenital adrenal hyperplasia that impairs steroidogenesis in both the adrenals and gonads (103 104 105 ). However, the absence of mutations in the type 1 gene provided the long-awaited molecular explanation for the persistence of peripheral steroidogenesis in these 3ß-HSD type 2-deficient patients, thus demonstrating the importance of peripheral sex steroid formation or intracrinology.
2. Human 17ß-HSDs.
The 17ß-HSDs are responsible for the formation and inactivation of all active androgens and estrogens. As discussed above for 3ß-HSD, until recently, 17ß-HSDs as well as almost all other dehydrogenases were considered to be reversible enzymes that catalyze the interconversion of substrates and products, mainly because the enzymatic activity was first characterized using tissue homogenates, subfractions, or purified proteins with added oxidized (NAD+, NADP+) or reduced (NADH, NADPH) cofactors. These exogenous cofactors drive the reaction in the oxidative or reductive direction depending on their oxidized or reduced state, respectively. However, using a more physiologically relevant method of enzymatic activity analysis, namely intact transfected cells in culture without the addition of exogenous cofactors, the transfected enzyme catalyzes the reaction in a unidirectional manner (85 98 99 106 107 ). These findings agree with the isolation of multiple types of 17ß-HSDs in which approximately half catalyze the reductive reaction (types 1, 3, 5, and 7) and half catalyze the oxidative reaction (types 2, 4, 6, and 8).
a. Type 1 17ß-HSD.
The molecular structure of the human type 1 17ß-HSD gene and mRNA, which encode a predicted protein of 327 amino acids, was the first of the 17ß-HSDs to be elucidated (Refs. 108 109 110 111 ; Fig. 6). This enzyme is a member of the short-chain alcohol dehydrogenase superfamily. The type 1 17ß-HSD enzyme is a cytosolic protein that exists in a homodimeric form that catalyzes predominantly the interconversion of estrone (E1) to E2 using NADP(H) as cofactor (112 113 ).
|
b. Type 2 17ß-HSD.
The structure of a cDNA encoding a second type of 17ß-HSD cDNA was then reported (117 118 ). This cDNA encodes a predicted protein of 387 amino acids with a molecular weight of 42,782 (Fig. 6). This protein is most likely associated with the membranes of the endoplasmic reticulum. The enzyme catalyzes the conversion of E2 to E1, testosterone to 4-dione, and 5-diol to DHEA. This enzyme, chronologically designated type 2 17ß-HSD, is also a member of the short-chain alcohol dehydrogenase superfamily, but it shares only about 20% sequence identity with the type 1 17ß-HSD cytoplasmic enzyme (109 ). This enzyme uses NAD(H) as a cofactor (117 ) and is less specific than type 1 17ß-HSD, both estrogens and androgens acting as substrates. This enzyme inactivates the estrogens and androgens made after the reductive action of type 2, 3, and 5 17ß-HSDs.
c. Type 3 17ß-HSD.
A third type of human 17ß-HSD cDNA encoding a predicted protein of 310 amino acids with a molecular weight of 34,513 was then characterized (119 ). Type 3 17ß-HSD, a microsomal isozyme, using NADP(H) as a cofactor, is expressed predominantly in the testes, where it synthesizes testosterone from 4-dione. This enzyme, which shares 23% sequence identity with the two other 17ß-HSD enzymes, is the site of the mutations responsible for male pseudohermaphroditism resulting from 17ß-HSD deficiency (119 ).
d. Type 4 17ß-HSD.
Human type 4 17ß-HSD is a 736-amino- acid protein of molecular mass 80 kDa that can transform E2 to E1 and 5-diol to DHEA (120 121 ). The human type 4 17ß-HSD mRNA is expressed in virtually all human tissues examined by Northern blot, including the liver, heart, prostate, testis, lung, skeletal muscle, kidney, pancreas, thymus, ovary, intestine, placenta, and several human breast cancer cell lines. This enzyme possibly plays a role in the inactivation of estrogens in a large series of peripheral tissues, although its activity is low and its importance in steroid formation in the human remains to be established. Indeed, mutations in type 4 17ß-HSD gene lead to a fatal form of Zellweger syndrome (122 ).
e. Type 5 17ß-HSD.
Although type 3 17ß-HSD synthesizes testosterone from 4-dione in the Leydig cells of the testes, thus providing approximately 50% of the total amount of androgens in men, the same enzymatic reaction is catalyzed in the peripheral target tissues in both men and women as well as in the ovary by a different enzyme, namely type 5 17ß-HSD (106 ). This enzyme is highly homologous with types 1 and 3 3-HSD as well as 20-HSD (106 ) and thus belongs to the aldo-keto reductase family.
In the postmenopausal ovary, hypertrophied stromal cells are localized mainly at the periphery and hilus (123 ). These stromal cells contain both 3ß-HSD and type 5 17ß-HSD, thus permitting the transformation of DHEA into 4-dione and then into testosterone. The amount of stromal hyperplasia in postmenopausal ovaries is correlated with the ovarian vein levels of 4-dione and testosterone (124 ). These hyperplastic stromal cells are thus responsible for the synthesis of 4-dione and testosterone in the postmenopausal ovary.
Type 5 17ß-HSD is not only expressed in the ovary but is also present in a large series of peripheral tissues including the mammary gland. The epithelium lining the acini and ducts of the mammary gland is composed of two layers, an inner epithelial layer and an outer discontinuous layer of myoepithelial cells. By immunocytochemistry, 3ß-HSD is seen in the epithelial cells of acini and ducts as well as in stromal fibroblasts (Fig. 7A). Immunostaining is also observed in the walls of blood vessels, including the endothelial cells. In the positive cells, the labeling is mainly cytoplasmic. No significant labeling could be detected in the myoepithelial cells. As shown in Fig. 7B, immunostaining for type 5 17ß-HSD gives results almost superimposable to those obtained for 3ß-HSD, the cytoplasmic labeling being observed in both epithelial and stromal cells and blood vessel walls (125 ). Studies performed at the electron microscopic level revealed that in sections stained for 3ß-HSD or type 5 17ß-HSD, labeling was not associated with any specific membrane-bound organelles in the different reactive cell types (126 ).
|
-diol. Among the many clones obtained, one type, named type 6 17ß-HSD, catalyzes selectively the oxidation of 3-diol to androsterone. The transformation of other C19-steroids, namely DHT to androstanedione and testosterone to 4-dione, also occurs but at an approximately 50- to 100-fold lower rate. Type 6 17ß-HSD shares 65% homology with rat type 1 retinol dehydrogenase and thus belongs to the retinol dehydrogenase family. The human counterpart has not yet been described, and its role remains to be established.
g. Type 7 17ß-HSD.
Type 7 17ß-HSD was first cloned from a rat corpus luteum cDNA library and was identified as prolactin receptor-associated protein (PRAP; Ref. 128 ). With the use of expression cloning of a mouse mammary epithelial (HC11) cell cDNA library, a clone that shares 89% identity with rat PRAP and catalyzes selectively the transformation of E1 to E2 has been isolated (129 ). After transfection into HEK-293 cells, Nokelainen et al. (129 ) also found that rat PRAP catalyzes efficiently and selectively the transformation of E1 to E2, whereas the transformation of C19 steroids was much weaker.
Human type 7 17ß-HSD cDNA is 1.5-kb long and encodes a protein of 37 kDa or 341 amino acids (130 ). With the use of RT-PCR, this enzyme is detected in the ovary, breast, placenta, testis, prostate, and liver. Comparison with other 17ß-HSDs indicates that it shares less than 20% identity, a typical percentage for the other members of the 17ß-HSD family. The human type 7 17ß-HSD gene spans 21.8 kb and consists of nine exons and eight introns. The gene is assigned to human chromosome bands 10p11.2 (130 ). It is noteworthy that type 5 17ß-HSD is also mapped to human chromosome 10 (bands 10p1514). The importance of this enzyme remains to be established.
h. Type 8 17ß-HSD.
Type 8 17ß-HSD is also known as the product of the Ke6 gene, which is found in the HLA region (131 ). This area is well known to contain genes encoding the human major histocompatibility complex. This complex is thought to be involved in polycystic kidney disease because aberrant gene expression has been found in two different models of polycystic kidney disease mice (132 ). Recently, Fomitcheva et al. (133 ) have found that the overproduced protein fused with GST catalyzes efficiently the transformation of E2 to E1. The transformation of testosterone to 4-dione is about 25% of that of E2 into E1. Using HEK-293 cells stably transfected with human type 8 17ß-HSD, we have shown recently that this enzyme selectively converts E2 to E1, the transformation of E1 as well as of androgen substrates being negligible (134 ).
3. Human 5-reductase isoenzymes.
The enzyme 5-reductase catalyzes the 5-reduction of 4-dione, testosterone, and other 4-ene-3-keto-steroids to the corresponding 5-dihydro-3-keto-steroids. The best known role of this enzyme is the transformation of testosterone into DHT, the most potent androgen, which is responsible for the differentiation of the male external genitalia and prostate as well as virilization at puberty. The major impact of 5-reductase in men, however, is its role in prostate cancer and benign prostatic hyperplasia. Two types of human steroid 5-reductases, chronologically identified as type 1 and type 2, were isolated from human prostatic cDNA libraries (135 136 ). The structure of the human type 1 5-reductase gene was first elucidated (137 ). This gene is not responsible for 5-reductase deficiency and is relatively insensitive to the inhibitor finasteride (136 ). Type 2 5-reductase, on the other hand, is the isozyme responsible for male pseudohermaphroditism from 5-reductase deficiency and is sensitive to finasteride (136 138 ).
Considering the crucial role of type 2 5-reductase, we have elucidated the structure of its corresponding gene (83 ). The type 2 5-reductase gene contains five exons and four introns and shows splicing sites identical to those of the type 1 gene. Its coding region shares 57% homology with that of the type 1 5-reductase gene. Type 1 5-reductase is the predominant form expressed in human skin (139 ).
C. Women produce about two thirds of the androgens synthesized in men
1. Decline in serum androgen precursors and metabolites occurs well before menopause.
To gain a better knowledge of the role of DHEA and DHEA-S transformation in both men and women, we have analyzed the serum levels of 18 conjugated C21- and C19-steroids (23 ). The data obtained show a dramatic decline in the circulating levels of DHEA, DHEA-S, 5-diol, and 5-diol fatty acid esters between the ages of 20 and 80 yr (Fig. 1). As mentioned earlier, in the 50- to 60-yr-old group, serum DHEA has already decreased by 70% from its 20- to 30-yr-old peak values in women (Fig. 1). It should be added that between the ages of 21 and 40 yr, mean serum testosterone in normal women decreases from approximately 1.3 to 0.61 nM (Ref. 26 ; Fig. 2). A parallel decrease is observed for serum DHEA and DHEA-S, thus suggesting the role of DHEA in the progressive decline in serum testosterone between the ages of 21 and 40 yr in normal women.
The serum concentrations of the conjugated metabolites of DHT, namely ADT-G, 3-diol-G, and 3ß-diol-G, are the most reliable parameters of the total androgen pool in women, whereas serum testosterone is mostly a measure of direct secretion of testosterone by the ovaries and/or adrenals. In fact, although the vast majority of testosterone and DHT is synthesized in the peripheral tissues in women, only a small proportion, estimated at 10–15% of the intracellular content of these androgens, diffuses out of the intracellular compartment without prior metabolism and can be measured as active androgen in the circulation. This is because testosterone and DHT, instead of being almost quantitatively released in the circulation, are rapidly glucuronidated into ADT-G, 3-diol-G, and 3ß-diol-G (Fig. 8). Because the individual glucuronosyltransferases responsible for the inactivation of androgens in the human mammary gland have not yet been identified, the human prostate is used as an example of the types of glucuronosyltransferases involved (140 141 ). These metabolites are much more water soluble than DHT and thus easily diffuse into the general circulation where they can be measured en route for their elimination mainly by the kidneys (Figs. 9 and 10). The serum concentration of the above-indicated conjugated androgen metabolites decreases by 47.5–72.7% between the 20- to 30- and 70- to 80-yr age groups in women, thus suggesting a parallel decrease in the total androgen pool with age (23 ).
|
|
|
); in women, most of these androgens originate from the transformation of DHEA and DHEA-S into testosterone and DHT in peripheral intracrine tissues. Such an estimate of the androgen pools in men and women based on the serum concentration of androgen metabolites can be influenced by possible differences in the metabolic clearance rates of these metabolites in men and women.
2. Plasma sex steroid levels are not a valid parameter of the intracellular situation in women.
Proof that changes of the intracellular concentration of sex steroids cannot be estimated by the measurement of testosterone and E2 in the circulation has been obtained in a study performed in postmenopausal women (23 ). This study analyzed in detail the serum concentrations of the active androgens and estrogens, as well as a series of free and conjugated forms of their precursors and metabolites, after daily application for 2 wk of a 10-ml 20% DHEA solution on the skin to avoid first passage of DHEA through the liver.
After daily administration of a single dose of DHEA percutaneously, serum DHEA, DHEA-S, and DHEA fatty acid esters increased approximately 175%, 130%, and 250% above control, respectively (Fig. 11), whereas serum 4-dione and testosterone increased by about 100% and 50% over control, respectively (Fig. 12). In parallel with the changes in serum DHEA, DHEA-S, and DHEA fatty acids, the most important effects (Fig. 13) were seen on the glucuronidated metabolites of ADT, 3-diol, and 3ß-diol. In fact, treatment with DHEA caused an increase in serum ADT-G, 3-diol-G, and 3ß-diol-G of approximately 125% (Fig. 13A), 140% (Fig. 13B), and 120% (Fig. 13C), respectively. No significant effect was observed on serum E1, E2, or DHT.
|
|
|
The 50% increase in serum testosterone of approximately 0.8 nM (from 1.5–2.3 nM) observed in women during DHEA treatment corresponds to a much larger increase of approximately 20 nM in serum DHEA. These data are in agreement with the information obtained in men after medical or surgical castration in which the serum levels of testosterone decreased from 15 nM to about 1.5 nM after elimination of testicular androgens. Thus, after castration, the serum levels of testosterone in 60- to 70-yr-old men became comparable to those observed in intact postmenopausal women. The 1.5 nM serum testosterone remaining after castration in men originates essentially from adrenal DHEA (77 87 ). The present data thus offer an independent measure of the amount of testosterone that diffuses into the circulation from the androgens synthesized from DHEA and DHEA-S in peripheral intracrine tissues (25 ).
In a recent study, daily oral administration of 50 mg DHEA had no significant effect on serum testosterone or DHT, whereas DHEA and ADT-G were increased to a similar extent (80–90%; Ref. 142 ). In another study, predosing serum levels of DHEA-S in postmenopausal women were increased from 0.55 µg/ml to about 1.4 µg/ml (143 ) after daily oral administration of 25 mg DHEA for 6 months. Serum DHEA and testosterone levels, however, measured 23 h after the last administration of DHEA, were not changed significantly. Similarly, the 50-mg/d oral dose of DHEA was found to lead to serum androgen levels in the premenopausal range (144 ).
Our data obtained after percutaneous administration of DHEA in normal postmenopausal women offer the first direct analysis of the correlation between the serum levels of DHEA and DHEA-S with the serum concentration of active androgens and estrogens and their corresponding glucuronidated and sulfated metabolites. It can be concluded that measurements of serum testosterone and E2 mainly reflect ovarian and/or adrenal steroid secretion, whereas the major contribution of the adrenals is not accurately represented in the circulating levels of active sex steroids. The present data clearly demonstrate that DHEA and DHEA-S are converted in a series of intracrine tissues into the active androgens and/or estrogens that exert their biological effects at their site of synthesis. These steroids are then metabolized in the same cells into inactive glucuronidated and sulfated metabolites, which finally diffuse in the extracellular compartment and can be measured in the circulation. Measurement of the conjugated metabolites of androgens is the only approach that permits an accurate estimate of the total androgen pool in women. It is likely that a similar situation exists for estrogens, although a precise evaluation of the pharmacokinetics of estrogen metabolism and identification of their metabolites remains to be completed.
3. Contribution of the postmenopausal ovary to serum 4-dione and testosterone.
It is well recognized that the postmenopausal ovary is a steroid-secreting gland (145 146 ). In fact, the postmenopausal ovary is well known to secrete testosterone, and most authors agree that it also secretes some 4-dione (147 148 ). In fact, a correlation has been observed between the degree of ovarian stromal hyperplasia and the secretion of androgens by the ovary (124 149 ). Moreover, lowering serum gonadotropins with a GnRH agonist has been shown to result in decreased serum androgen levels, thus indicating that the stromal cells of the ovary are under gonadotropin control (150 151 ). In agreement with these data, receptors for LH and FSH have been described in the ovarian stromal cells. It should be mentioned that Couzinet et al. (152 ) have reported that the postmenopausal ovary does not contribute significantly to serum androgen levels. This observation is unique and, if confirmed, will bring even more emphasis on the importance of the adrenals in sex steroid physiology after menopause.
Despite the above-described limitations of the interpretation of serum levels of sex steroids, it is of interest to provide the best available estimate of the contribution of the ovaries and adrenals to the serum levels of 4-dione and testosterone. The majority of studies show declining levels of serum testosterone and 4-dione with age (149 153 154 155 156 ). Testosterone concentration in the ovarian venous blood is 15 times higher than in peripheral blood (147 ). In fact, the production of testosterone by the ovary has been estimated to decrease from 250 to 180 µg/d after menopause (157 ).
As illustrated in Fig. 14A, although the ovaries and adrenals contribute about equally to the serum levels of 4-dione in premenopausal women (158 159 ), the contribution of the ovaries decreases to about 20% after menopause (158 159 ), despite a progressive fall in the contribution of the adrenals through transformation of declining amounts of DHEA into 4-dione, thus leading to lower total serum concentration of 4-dione after menopause. Similarly, the serum levels of testosterone in premenopausal women originate in approximately equal amounts from the ovaries and adrenals (Refs. 158 and 159 ; Fig. 15). Peripheral serum testosterone decreases by 50% after ovariectomy in postmenopausal women, thus indicating that the approximately equal contribution of the ovaries and adrenals to serum testosterone remains after menopause. In another study, human chorionic gonadotrophin stimulation and dexamethasone suppression tests in postmenopausal women have suggested that the ovary contributes about 50% of testosterone and 30% of 4-dione in the peripheral circulation (160 ).
|
|
-diol-G, and 3ß-diol-G (Figs. 8–10). Some testosterone and DHT are also glucuronidated and are found in the circulation as Testo-G and DHT-G. One can thus estimate, as illustrated schematically in Fig. 16, that after menopause the contribution of the ovaries to the intracellular concentration of testosterone is only about 10%. This estimate is based on the observation that serum levels of testosterone are reduced by 90% from 15.0 nM to about 1.5 nM after castration in men, whereas the intraprostatic concentration of DHT is reduced only by 50% to about 2.5 ng/g tissue or about 7.5 nM (77 80 ). Thus, whereas 7.5 nM intratissular DHT of testicular origin corresponds to 13.5 nM serum testosterone, 1.5 nM serum testosterone of adrenal origin corresponds to the same 7.5 nM intratissular DHT, thus requiring a multiplication factor of 9 to compensate for the poor diffusion of testosterone synthesized intracellularly from DHEA compared with the efficacy of entry of circulating testosterone in the prostatic tissue. Such calculations are in agreement with other data showing that serum ADT-G levels reflect essentially adrenal androgen secretion (161 ). In fact, Giagulli et al. (161 ) have concluded that DHEA-S accounts for 70–80% of serum ADT-G levels.
|
|
III. Androgens Inhibit Breast Cancer |
|---|
|
TopAbstractI. Androgens and Their...II. DHEA Is Predominantly...III. Androgens Inhibit Breast...IV. DHEA Inhibits Breast...V. Rationale for the...References |
|---|
A. Clinical data
Estrogens have long been known to play a predominant role in the development and growth of human breast cancer (164 165 166 ). On the other hand, well recognized observations have shown that androgens such as testosterone propionate (162 167 168 169 ), fluoxymesterone (170 171 ), and calusterone (172 ) used in the adjuvant therapy of breast cancer have an efficacy comparable to that achieved with other types of endocrine manipulations (165 169 173 174 ).
Most importantly, a higher response rate and a longer time to disease progression have been observed when androgens were combined with an antiestrogen, compared with an antiestrogen alone (171 175 ). The benefits of combined treatment with fluoxymesterone and tamoxifen vs. tamoxifen alone were observed in postmenopausal women with metastatic breast cancer (175 ), both in terms of response rate and time to progression of disease.
As summarized later, such additive inhibitory effects of an antiestrogen and androgen on breast cancer have been clearly demonstrated in a series of experimental models. The above-mentioned clinical data are also well supported by the observation of a synergistic effect of DHEA and of the pure antiestrogen EM-800 on prevention of the development of dimethylbenz(a)anthracene (DMBA)-induced mammary tumors in the rat (176 ). Moreover, the almost exclusive androgenic component in the action of DHEA on the histomorphology and structure of the rat mammary gland has recently been shown (177 ), thus supporting such an inhibitory effect of DHEA.
It should also be mentioned that androgens have been shown to induce an objective remission after failure of antiestrogen therapy and hypophysectomy. These clinical observations indicate that the benefits obtained with androgen therapy in breast cancer cannot be due solely to a suppression of pituitary gonadotropin secretion but must result, at least in part, from a direct effect on tumor growth in women. The role of androgens as direct inhibitors of breast cancer growth is well supported by the presence of AR in a large proportion of human breast cancers (178 179 180 181 ). In fact, in primary breast cancer, AR has been found in 54% of premenopausal and 48% of postmenopausal patients (180 182 ). The presence of AR has also been described in MCF-7 cells (183 184 ).
The overwhelming clinical evidence for tumor regression observed in 20–50% of pre- and postmenopausal breast cancer patients treated with various androgens (173 ) favors the view that naturally occurring androgens might constitute an as yet overlooked, direct inhibitory control of mammary cancer cell growth. It is thus reasonable to suggest that an imbalance between androgenic and estrogenic influences could modify the overall growth rate of breast tumors in much the same way as that suggested for progestins in estrogen target tissues (185 ). There is also genetic evidence in agreement with a protective role of androgens against breast cancer (186 187 ). Interestingly, the observation that an increased response rate can be obtained by combining androgens and an antiestrogen therapy in breast cancer patients (171 175 ) is in agreement with our observations summarized later that the mechanisms of the inhibition exerted by the two types of agents are different, whereas their effects, at least in part, are additive.
In this context, it has been found that Western women having a low excretion of adrenal androgenic metabolites respond more poorly to endocrine therapy and have a shorter survival time (188 189 190 ). Possibly because of the small number of cancer cases in many studies, the methodology used, the low predictive value of measurements of serum sex steroid levels, and the association in case-control studies between serum androgen levels and breast cancer risk have led to contradictory data. Thus, subnormal levels of serum androgens have been found in women with increased risk of breast cancer (191 192 193 ), whereas opposite data have also been reported (194 195 196 197 ).
It is of interest that suppression of androgens in men is associated with breast growth (198 ). Moreover, mutations in AR have been linked with breast cancer in men (199 ).
It should be added that treatment of ovariectomized monkeys with testosterone decreased by about 40% the stimulation of mammary epithelial proliferation induced by E2 (200 ). It is possible that part of the increased risk of breast cancer in BRCA-1 mutant patients is associated with the decreased efficiency of the mutated BRCA-1 gene to interact with the AR (201 ). It is also pertinent to mention that female athletes and transsexuals taking androgens show atrophy of mammary gland epithelial tissue (202 203 ).
B. Preclinical data
Lacassagne (204 ) first observed in 1936 that treatment of mice with testosterone propionate delayed the occurrence of E1-stimulated mammary tumors. In DMBA-induced tumors, high doses of DHT (0.5–4.0 mg/d) for several weeks caused the regression of 60% of established tumors (163 ). Similar effects were observed with testosterone propionate (205 ) and dromostanolone propionate (206 207 ).
In support of the early clinical data mentioned above, our previous studies have clearly demonstrated that androgens exert a direct inhibitory effect on the proliferation of human breast cancer cells (208 209 210 211 212 213 ). In fact, the first demonstration of a potent and direct inhibitory effect of androgens on human breast cancer growth was obtained in the estrogen-sensitive human breast cancer cell line ZR-75-1 (208 ). In that study, as shown in Fig. 17A, DHT not only completely blocked the stimulatory effect of E2 on cell proliferation but also reduced cell growth in the absence of estrogens. At low cell density (Fig. 17B), it can be seen that DHT completely prevented breast cancer cell growth.
|
-reductase in breast cancer tissue, an enzyme thought to be specific for androgen-dependent tissues. In ZR-75-1 cells, concentrations of DHT in the incubation medium similar to the plasma levels found in normal women (215 216 217 ) and breast cancer patients 
), or both steroids (
) were added, and cell numbers were determined at the indicated time intervals. Control cells received the ethanol vehicle only. B, Same as A, except that the initial density was 5.0 x 103 cells/2.0-cm2 well (