| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Endocrinology and Reproductive Medicine (A.B., N.M., F.K., P.T.), Hôpital Necker, 75743 Paris, France; Unité de Recherche (G.M.), Institut National de la Santé et de la Recherche Médicale, Unité 135, Hormones, Genes and Reproduction; and Unité de Recherche (M.M.), Institut National de la Santé et de la Recherche Médicale, Unité 120, Hôpital Bicêtre, 94275 Le Kremlin Bicêtre, France
Address all correspondence and requests for reprints to: Dr. Philippe Touraine, Department of Endocrinology and Reproductive Medicine, Hôpital Necker, 149 rue de Sèvres, Paris Cedex 15, 75743, France. E-mail: philippe.touraine{at}nck.ap-hop-paris.fr.
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
We have undertaken a retrospective analysis of androgen levels in women presenting with POF and a normal karyotype. Interestingly, despite ovarian failure, in some of our patients we found androgen levels above those considered normal for postmenopausal women. To identify the source of androgens in these women, we decided to study the available ovarian biopsy samples of POF patients with elevated androgen levels using immunohistochemistry with a panel of antibodies directed against the main steroidogenic enzymes. As a control group we used ovarian samples obtained from POF patients without elevation of androgen levels.
| Patients and Methods |
|---|
|
|
|---|
The androgen profiles of 143 consecutive POF patients without history of chemotherapy and/or pelvic radiotherapy and with normal karyotype, referred to our center between 1995 and 2004, were reviewed. In these patients, spontaneous POF was diagnosed before age 40 yr and was defined by at least 4 months of amenorrhea and two FSH levels above 40 mIU/ml, at least 1 month apart. Among them were three patients who had an FSH receptor gene mutation. They have been previously described (patients 2, 3, and 13) (10, 11, 12). Thirteen patients had thyroid autoimmune disease. None had ovarian autoimmune disease. Among these 143 consecutive patients, 23 patients (16.1%) had androgen levels above those in postmenopausal women, defined by plasma testosterone (T) levels greater than 1.4 nM/liter,
4-androstenedione (A) levels greater than 7 nM/liter, and/or DHEA-S levels greater than 6800 nM/liter (13). Ovarian biopsies were available from 15 of those patients. POF duration was established starting from the first determination of elevated FSH levels to the time of ovarian biopsy. Blood samples were taken the day before the ovarian biopsy, except for two patients, who did not have elevated androgen levels. Informed consent was obtained from all patients for serum samples, cytogenetic studies, and surgical procedures. Fifteen patients matched for age and mean POF duration, with normal androgen levels and available ovarian biopsies, served as controls.
Hormone measurements
Plasma levels of FSH and LH were measured by conventional RIA (Immunotech Beckman Coulter, Marseille, France). Estradiol (E2), T, and A were determined after previous plasma extraction [DiaSorin (Milan, Italy), Schering CisBio International (Gif-sur-Yvette, France), and Immunotech Beckman Coulter, respectively]. Plasma levels of DHEA-S were measured by conventional RIA (Immunotech Beckman Coulter). The normal range was based on hormone measurements obtained from healthy women during a normal menstrual cycle and provided by the laboratory of hormonal investigations at Necker Hospital. The intraassay coefficient varied from 2.68.1%, whereas the interassay coefficient varied from 3.711.9%. The inhibin B concentration was measured in duplicate in serum samples using a solid phase sandwich ELISA. The inhibin B assay used a capture monoclonal antibody raised against a sequence near the C terminal of the human ß-subunit, immobilized on a hydrazide plate, as described by Groome et al. (14) and a second monoclonal antibody specific for the ß-subunit of inhibin coupled to alkaline phosphatase. The assay detection limit was 5 pg/ml.
Histological and immunocytochemical studies of the ovaries
For each patient at least two bilateral ovarian biopsies of 35 mm were available (mean, 2.50; range, 15).After formol fixation and paraffin embedding, the biopsies were serially sectioned at 5 µm, then stained with hematoxylin/eosin and Massons Trichrome stains. Histological examination was performed with a conventional optical microscope (Provis, Olympus, Japan).
Immunohistochemistry was performed as follows. Briefly, serial tissue sections were deparaffinized by successive baths in xylene and alcohol, and antigen retrieval was performed before immunohistochemistry in a commercial microwave oven at full power in pH 6 citrate buffer for 15 min. The sections were then incubated overnight with the following primary antibodies: polyclonal antibodies anti-P450 side-chain cleavage (P450scc; dilution, 1:3000) (15), anti-3ß-HSD (dilution, 1:4000) (16), anti-P450c17
(dilution, 1:5000) (17), and antiaromatase (Hauptman-Woodward Medical Research Institute, Inc., Buffalo, NY; dilution, 1:3000) at 4 C in a humid chamber. After primary antibody incubation, endogenous peroxidases were quenched with 3% H2O2 in PBS (pH 7.4) for 5 min, and the bound antibodies were revealed with a secondary biotinylated antibody and peroxidase-labeled streptavidin (LSAB2 immunostaining kit, DakoCytomation, Carpinteria, CA) according to the manufacturers instructions. Aminoethylcarbazol (Sigma-Aldrich Corp., Lyon, France) was used as a chromogen. The sections were counterstained with Mayers hematoxylin. Sections of a biopsy of a normal ovary from a 26-yr-old woman were used as controls. Each experiment was performed at least twice; immunostaining for P450c17
was performed two additional times on nonconsecutive sections.
Statistical analysis
Analyses were processed with StatView version 5 (Abacus Concepts, Inc., Berkeley, CA). Descriptive statistics were performed for each variable; quantitative results are presented as the median (range), and qualitative results are presented as a distribution of a number of patients. Hormonal and ovarian parameters were compared using the Mann-Whitney test. Proportions for the two groups were compared using the
2 test. P < 0.05 was accepted as significant.
| Results |
|---|
|
|
|---|
Among the 30 patients, most presented with secondary amenorrhea (n = 24; 80%). Three of the patients with androgen levels above the norm for postmenopausal women also had clinical hyperandrogenism. Of these three patients, one had acnea, and two had mild hirsutism with Ferriman-Gallwey scores of 12 and 13, respectively. The median age (range) at diagnosis in the POF patients with elevated androgen levels was 22 yr (1639 yr) and was not significantly different from that in control POF patients. There was no significant difference between the groups in the median POF duration [16 months (range, 3132 months) in patients with elevated androgen levels and 21 months (range, 3138 months) in control POF patients]. The median body mass index was 24.3 (range, 21.437.5) in control patients and 21.2 (range, 17.330.1) in POF patients with elevated androgen levels.
In POF patients with elevated androgen levels, median T, A, and DHEA-S levels were significantly higher than in control POF patients [T: 1.60 nM/liter (range, 0.422.54 nM/liter) vs. 0.69 nM/liter (range, 0.170.94 nM/liter); P < 0.001; A: 7.68 nM/liter (range, 3.8415.00 nM/liter) vs. 3.84 nM/liter (range, 1.476.63 nM/liter); P < 0.001; DHEA-S: 6316 nM/liter (range, 13169746 nM/liter) vs. 3751 nM/liter (range 13146408 nM/liter); P < 0.05, respectively; Tables 1
and 2
]. No significant difference in median FSH and LH levels was observed between the two groups [67 U/liter (range, 20129 U/liter) and 31 U/liter (range, 651 U/liter), respectively, in the POF patients with elevated androgen levels vs. 64 U/liter (range, 34138 U/liter) and 25 U/liter (range, 1051 U/liter), respectively, in the control POF patients]. However, median E2 levels were significantly higher in patients with elevated androgen levels compared with the control POF patients [75 pM/liter (range, 18551 pM/liter) vs. 18 pM/liter (range, 1895 nM/liter); P < 0.05; Tables 1
and 2
]. Inhibin B levels were also more elevated in POF patients with elevated androgen levels, ranging from 5150 ng/ml, and were undetectable in all control POF patients (Tables 1
and 2
).
|
|
The ovarian biopsy of all patients was studied by immunohistochemistry for expression of the steroidogenic enzymes P450scc, 3ß-HSD, P450c17
, and P450 aromatase. Three different patterns of immunostaining could be identified in the POF patients with elevated androgen levels (Tables 1
and 2
).
The first pattern was characterized by an intense expression of steroidogenic enzymes in a relevant number of cells. This pattern was found in five biopsies (patients 15); however, there were some differences in both the arrangement of steroidogenic cells and the types of expressed enzymes. Large bands of luteinized thecal cells arranged around the irregular antrum of atretic follicles were evident in four biopsies. The cells expressed the enzymes P450scc, P450c17
, and 3ß-HSD with similar intensities (Fig. 1A
, P450scc; Fig. 1B
, 3ß-HSD; Fig. 1C
, P450c17
). In only one of the four cases were a few residual granulosa cells expressing aromatase observed in the antrum of an atretic follicle (Fig. 1D
, P450 aromatase). A recent corpus luteum was found in the fifth case. It was characterized by the expression of P450c17
in peripheral theca-derived lutein cells, P450 aromatase immunopositivity of the more central granulosa-derived lutein cells, and an ubiquitous expression of P450scc and, very intensely, of 3ß-HSD on both theca- and granulosa-derived lutein cells (Fig. 1E
, P450scc; Fig. 1F
, 3ß-HSD; Fig. 1G
, P450c17
; Fig. 1H
, P450 aromatase).
|
, whereas 3ß-HSD expression was observed with less frequency and in a smaller number of cells, and aromatase was not expressed. This pattern characterized five biopsies (patients 610), which also exhibited different histologies. Two were totally devoid of normal or atretic follicular structures and of corpora albicantia (Fig. 2A
). One showed an isolated primary follicle of normal appearance, one was characterized by the presence of an antral atretic follicle (Fig. 2B
), and one exhibited a corpus albicans (Fig. 2C
).
|
); two exhibited a variable number of small follicles up to the primary stage (Fig. 2E
). One biopsy showed the presence of remnants of an antral atretic follicle (Fig. 2F
). Analogous distribution patterns were also observed in the control POF patients, but with different proportions. The biopsies of two controls POF patients (no. 16 and 17), showed significant steroidogenic activity of the hypertrophic theca interna cells of antral atretic follicles. Aromatase was expressed in a few residual follicular granulosa cells in only one of the two. Six other patients of this group (no. 1823) exhibited a small number of steroidogenic cells in their biopsies, associated with remnants of antral atretic follicles or corpora albicantia in only half the cases. Finally, seven patients were devoid of steroidogenic cells (patients 2430). Four of those patients lacked follicular structures or remnants in their ovary specimens, one of them exhibited a primary follicle, and corpora albicantia were found in the two others.
| Discussion |
|---|
|
|
|---|
We found a heterogeneous pattern of expression ranging from the presence of antral atretic follicles with a hypertrophic, functional, androgen-producing theca interna to the total absence of steroidogenic cells. Between these extremes there were several biopsies exhibiting small numbers of interstitial androgen-producing cells. This heterogeneous expression was detected in both groups of patients and is not correlated with the age at diagnosis (median duration of POF), because the two groups of patients with or without elevated androgen levels were comparable in this respect. In particular, the patients presenting with intense expression of steroidogenic enzymes in a relevant number of cells in their ovarian biopsies were not the patients with the minimal duration of POF. However, detectable E2 and inhibin B levels were found in some POF patients with elevated androgen levels, suggesting residual follicular activity in these patients.
Abnormal, but functional, follicular structures, characterized by a highly steroidogenic hypertrophied theca interna and apoptotic or absent granulosa cells, were detected in five POF patients with elevated androgen levels. Our observations suggest that the ovaries of some patients with POF may be capable of advanced follicular development. The abnormal follicular structures are functionally active and capable of steroidogenesis, but orientated mainly toward androgen production. The hypertrophic and hyperplastic androgen-producing thecal cells observed are secondary to the action of persistently elevated LH levels on developing thecae and have been previously described in the ovaries of POF subjects (5). The absence of aromatase-positive granulosa cells in those anomalous follicles could be secondary to the elevated thecal androgen synthesis, because androgen excess has been implicated in the induction of follicular atresia (18). The hypertrophied thecae could well be the source of androgens in POF patients with high androgen levels, especially in the presence of an elevated number of those structures in the ovarian cortex. Taken together, our results suggest that patients with POF may develop follicles at an advanced stage of development and that those structures are capable of steroidogenesis. However, the finding of very similar (both qualitatively and quantitatively) androgen-producing, abnormal follicles in the ovaries of POF patients either with or without hyperandrogenism suggests that the androgen production of those abnormal follicles is not always sufficient to elevate serum androgen levels in POF patients.
Scarce ovarian steroidogenic cells were found in the biopsies of five POF patients with elevated androgen levels and six control POF patients. These ovaries had an aspect similar to that of some normal (nonpathological) postmenopausal ovaries (9). The few androgen-synthesizing cells observed in these ovaries are probably the thecal remnants of past episodes of advanced follicular maturation. The finding of groups of interstitial steroidogenic cells in POF patients, therefore, suggests the existence of reserve follicles capable of being recruited. This limited follicular reserve might already be totally spent, as in truly postmenopausal ovaries (19), or conversely, some follicles might still persist and be susceptible to additional development. This small group of interstitial androgen-producing cells, however, is insufficient to explain the elevated androgen levels found in some POF patients with this pattern of expression of steroidogenic enzymes. Moreover, an identical immunohistochemical aspect is found in a similar number of control POF patients (and in normal postmenopausal women).
Finally, the ovarian biopsies of five POF patients with elevated androgen levels and seven control POF patients had no evidence of steroidogenic activity by immunohistochemistry. Some of these biopsies from patients 14 and 15 and from patients 2730 did not contain reserve follicles and showed no traces (fibrous tissue scars, glassy membranes, or corpora albicantia) of past antral follicular development, suggesting complete exhaustion of the follicular reserve and, possibly, absence of follicular development beyond the secondary stage.
The source of the elevated androgen levels found in some POF patients might also be extrafollicular. Interestingly, the ovarian failure in one of the patients with elevated androgen levels devoid of steroidogenic cells was secondary to a totally inactivating mutation of the FSH receptor gene (10). This woman, who exhibited only primordial and primary follicles at the ovarian biopsy, is incapable of follicular development beyond the primary stage. In this particular case, the existence in the ovaries of thecal or interstitial (theca derived) steroid-producing cells capable of androgen production is very unlikely. Conversely, two patients with elevated androgen levels with only partial inactivation of the FSH receptor, and therefore capable of more advanced follicular development (11, 12), exhibited hypertrophied thecae.
A possible intraovarian source of steroids independent of thecal cells could be the ovarian hilus cells, which are similar to testicular Leydig cells. Hilus cells are capable of androgen production (20) and might be sensible to elevated LH levels, as are their testicular homologs (21). There is one report of a POF patient with hyperandrogenemia associated with ovarian hilus cells hyperplasia (22). However, an unavoidable limit of the ovarian biopsy is the lack of sampling of structures in the deep portions of the cortex and the medullary; therefore, it was not possible to detect an abnormal proliferation of hilus cells in our patients. In contrast, the examination of postmenopausal ovaries, removed whole, allows assessment of the respective contributions to steroidogenesis of hilus and cortical cells (9).
Another possible source of androgens in POF patients with elevated androgen levels could be the adrenal gland. In fact, seven POF patients had elevated DHEA-S levels, suggesting an adrenal contribution to androgen production, and five of them were devoid of ovarian steroid-producing cells. Some researchers have reported the presence of the LH receptor (LHR) gene in the fasciculata and reticularis regions of the human adrenal gland (23), and we have detected the presence of LHR in the human adrenal cortex (Meduri, G., unpublished observations). Participation of the corticosurrenal gland in the circulating androgen pool in postmenopausal women has previously been described (9), and the presence of functional LHR has recently been demonstrated in adrenocortical cell lines (24). Additionally, adrenal expression of LHR has been shown to increase with chronically elevated serum LH levels in female mice (25), and gonadotropins are chronically elevated in POF patients. Persistently elevated LH levels, however, were common to all patients: the existence of variable individual sensitivity to LH stimulation of hilus or adrenocortical androgen-producing cells, secondary to different levels of expression of LHR or of factors synergic to LH, might explain the difference.
Our findings on the steroidogenic activity of the ovaries in POF patients are limited by the possibility of incomplete sampling. The diagnostic efficacy of the ovarian biopsy, especially in assessing ovarian reserve, has been discussed and is quite controversial due to the uneven distribution of small follicles in the ovarian cortex (26, 27, 28). However, some researchers have reported the usefulness of ovarian biopsies in the diagnosis of POF (26, 29, 30). Moreover, multiple and bilateral ovarian samples more than 3 mm in diameter were available for most of our patients (78%).
In conclusion, we investigated in situ, for the first time, the steroidogenic production by ovaries of patients with POF. We found that although the steroidogenic aspect of the ovaries of some POF patients is not dissimilar to that of the ovaries of normal postmenopausal women, a substantial portion of POF patients is still capable of follicular steroid hormone production. The analysis of the particular enzymes involved indicates that theca interna- and interstitial theca-derived cells are the main source of steroids in these patients and that they synthesize mainly androgens. Our results also suggest that the androgen levels of a portion of POF patients might not always be correlated to the presence of androgen-producing cells in the ovarian biopsy or to age and ovarian failure duration. The possibility of a contribution of the ovarian hilus cells or of the adrenal gland to the elevated androgen levels of some POF patients during LH stimulation was discussed and requires additional investigation.
| Acknowledgments |
|---|
| Footnotes |
|---|
First Published Online January 25, 2005
1 A.B. and G.M. contributed equally to this work. ![]()
Abbreviations: A,
4-Androstenedione; DHEA-S, dehydroepiandrosterone sulfate; E2, estradiol; LHR, LH receptor; POF, premature ovarian failure; P450scc, P450 side-chain cleavage; T, testosterone.
Received August 31, 2004.
Accepted January 17, 2005.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. Bachelot, A. Rouxel, N. Massin, J. Dulon, C. Courtillot, C. Matuchansky, Y. Badachi, A. Fortin, B. Paniel, F. Lecuru, et al. Phenotyping and genetic studies of 357 consecutive patients presenting with premature ovarian failure Eur. J. Endocrinol., July 1, 2009; 161(1): 179 - 187. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Alevizaki, K. Saltiki, E. Mantzou, E. Anastasiou, and I. Huhtaniemi The adrenal gland may be a target of LH action in postmenopausal women. Eur. J. Endocrinol., June 1, 2006; 154(6): 875 - 881. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Saner-Amigh, B. A. Mayhew, F. Mantero, F. Schiavi, P. C. White, C. V. Rao, and W. E. Rainey Elevated Expression of Luteinizing Hormone Receptor in Aldosterone-Producing Adenomas J. Clin. Endocrinol. Metab., March 1, 2006; 91(3): 1136 - 1142. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. Arlt Androgen therapy in women Eur. J. Endocrinol., January 1, 2006; 154(1): 1 - 11. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |