| Endocrine and metabolic changes in women with polycystic ovaries and with polycystic ovary syndrome | ||
|---|---|---|
| Prev | Chapter 6. Discussion | Next |
Obese women with PCOS displayed a distinctly different steroidogenic response pattern to a single dose of hCG compared with the control women followed up for 4 days. Peak peripheral serum A and T concentrations were achieved at 48 hours after an injection of hCG in the PCOS women, with peak levels of 17-OHP and E2 at 24 hours. In contrast, all steroids measured in the control women reached their maximum serum concentrations at 96 hours.
The rapid 17-OHP and E2 responses to hCG (with similar biological properties as LH) in PCOS women are in accordance with the results of previous studies in which short-term responses (up to 24–48 h) to hCG (Ibanez et al. 1996, Gilling-Smith et al. 1997, Levrant et al. 1997) or to GnRHa (Barnes et al. 1989b, Rosenfield et al. 1994, Ibanez et al. 1996) have been studied. These findings have been interpreted to imply that "dysregulation" of the enzyme P450c17α, leading to enhanced activities of both 17α-hydroxylase and 17,20-lyase, plays a central role in the pathogenesis of the ovarian hyperandrogenism associated with PCOS. The male-type steroidogenic response pattern to hCG seems to be related to PCOS and not to obesity, since a similar response to hCG was found in both the lean and the obese control women.
The steroid responses to hCG in this study were followed longer than in previous studies, and it was observed that the levels of 17-OHP and E2 began to decline at 48 hours, in contrast to the control women, who demonstrated peak 17-OHP and E2 values at 96 hours after hCG. The steroid response patterns observed in obese PCOS women were identical to those found in normal men (Smals et al. 1979, Martikainen et al. 1980). It has been suggested that E2 inhibits 17-lyase activity, leading to an accumulation of 17-OHP, possibly preventing its further conversion to T in the human testis (Forest et al. 1979, Martikainen et al. 1980, Tapanainen et al. 1983). It is tempting to speculate that a similar regulatory mechanism may be operative in the polycystic ovary, with increased theca cell activity and mass. Since a physiological decline of follicular cohort has been shown to occur while ageing (Faddy et al. 1992), it would be interesting to see if the responses to hCG are different in younger women than in older women. Furthermore, the time course of 17-OHP and E2 responses indicates that E2, in contrast to a previous suggestion (Ibanez et al. 1996), may be an important regulator of 17,20-lyase activity in the human ovary. These statements are still speculative, however, and further studies are required to support them.
The observed rapid and significant increase in E2 after hCG in obese PCOS women is not in good agreement with the two-cell model of ovarian steroidogenesis in which LH stimulates theca cell androgen production and FSH mainly stimulates granulosa cell estrogen production. There is a considerable amount of data, however, indicating that human theca cells are capable of forming E2 throughout the life span of the antral follicle (McNatty et al. 1979, Gilling-Smith et al. 1994). Thus, the rapid release of E2 from polycystic ovaries after hCG-stimulation observed in this study may reflect the existence of a releasable pool of steroids in the theca cells. Alternatively, hCG may directly stimulate ovarian aromatase activity (Tapanainen et al. 1991), or androgens might exert a rapid paracrine effect on granulosa cells by up-regulating their aromatase activity (Haning, Hackett et al. 1993). Furthermore, the response of E2 to hCG may also be due to a direct activation of granulosa cells, since granulosa cells from small individual follicles obtained from anovulatory polycystic ovaries have been shown to be prematurely responsive to LH (Willis et al. 1998).
In conclusion, the present study indicates that obese PCOS women have a male-type steroidogenic response pattern to a single dose of hCG, which may be explained by higher theca cell activity or mass in polycystic ovaries.
It is obvious that higher amounts of gonadotrophins in the hMG preparation (FSH 75IU and LH 75IU) than in the FSH preparation (FSH 75IU and LH < 0.1 IU) have an influence on the stimulation results. It is noteable that FSH administration alone also led to a distinctly increased production of 17-OHP, A and T in both endocrinologically normal and PCOS women. According to the two-cell theory of ovarian steroidogenesis, LH stimulates the theca cell production of A and T and FSH stimulates granulosa cell function. As there was no significant supply of LH in the FSH preparation used for stimulation in this study, it was likely that FSH alone is also capable of stimulating ovarian theca cell steroid synthesis under these experimental conditions. This was confirmed by stimulations with recombinant FSH preparation. Our results are in accordance with the study by Tanbo et al. (Tanbo et al. 1990), in which the experimental conditions were similar to ours.
The mechanism(s) by which FSH stimulates theca cell function is unclear. It has been postulated that when granulosa cells are exposed to FSH they can generate estrogens only when supplied with androgens as the aromatase substrate (Erickson & Ryan 1976). In gonadotrophin-deficient woman it was shown that FSH alone, without any LH, is not able to stimulate follicular steroid synthesis (Schoot et al. 1992). In the presence of profound gonadotrophin deficiency, pharmacological doses of highly purified FSH with minute LH contamination have been reported to be capable of stimulating ovarian follicular maturation (Couzinet et al. 1988). FSH is known to potentiate LH action by inducing LH receptors (Knecht et al. 1986). Thus, minute amounts of endogenously secreted or exogenously administered LH, because of its presence in the highly purified FSH preparation, could result in a significant but subnormal production of ovarian androgens (Couzinet et al. 1988, Teissier et al. 1999). LH suppression by GnRHa is known to be incomplete (Chang et al. 1983, Matikainen et al. 1992, Bützow et al. 1999) and endogenously secreted LH therefore probably had an influence on ovarian androgen production in our study.
Estradiol has been shown to augment A production and the stimulatory effect of FSH on steroid production in theca cells may thus be mediated by E2 produced in granulosa cells (Gilling-Smith et al. 1997). FSH also seems to be capable of influencing ovarian androgen synthesis via a paracrine mechanism involving an enhanced expression of thecal/interstitial P450c17α. It is unclear which locally produced factor(s) mediates this action of FSH. Likely candidates are regulatory proteins produced by granulosa cells responding to a direct stimulation by FSH (i.e. inhibin, IGF-1) (Adashi et al. 1985, Hsueh et al. 1987, Smyth et al. 1993). It has recently been suggested that theca cells secrete factor(s) inhibiting the differentation of immature, while promoting that of matured granulosa cells; those results also suggested that granulosa cells secrete factor(s) promoting both the differentation and growth of theca cells throughout the follicular maturation process (Yada et al. 1999).
In this study with daily injections of gonadotrophins after pituitary suppression, women with PCOS did not show a distinctly exaggerated steroidogenic response to gonadotrophin stimulation compared to endocrinologically normal women. We were unable to detect any difference in the serum A and 17-OHP concentrations or in the ?17-OHP/?A ratio between the two patient groups (control vs. PCOS) studied, indicating unchanged 17-20-lyase activity in PCOS. This result is in contrast with those of short-term GnRHa or hCG stimulation studies where PCOS women showed an exaggerated 17-OHP response compared with control women (Barnes et al. 1989b, Rosenfield et al. 1994, Ibanez et al. 1996, Gilling-Smith et al. 1997).
The discrepancies between this and previous studies is most likely due to experimental conditions. GnRHa used for suppression prior to gonadotropin stimulation may modify ovarian responsivness to gonadotrophins, for example, by inhibiting LH receptor formation (Hsueh & Erickson 1979, Rabin & McNeil 1980, Amsterdam et al. 1981). However, PCOS women have been shown to display 17-OHP hyperresponsiveness to a single dose of hCG (10 000 IU) also after 1 month of GnRH agonist treatment (Gilling-Smith et al. 1997). The lesser increase in 17-OHP and A in the present study may be due to the lower dosage of gonadotrophin (150-300 IU/daily) given. On the other hand, considering the heterogeneity of PCOS, the sample size in this study may be too small to give a significant result. One explanation for our results may be ovarian desensitization to gonadotrophins or the down-regulation of LH receptors after repeatedly injected exogenous gonadotrophins. It has been shown that stimulation of male rat testicular Leydig cells with exogenous gonadotrophins (LH or hCG) is followed by a loss of LH receptors and a decreased maximum T response (Hsueh et al. 1976, Hsueh et al. 1977, Tsuruhara et al. 1977). It is tempting to suggest, however, that the rapid 17-OHP response to hCG stimulation in previous studies (Barnes et al. 1989b, Rosenfield et al. 1994, Ibanez et al. 1996, Gilling-Smith et al. 1997) may possibly only reflect the releasable pool of steroids from the thecal cells of polycystic ovaries and not an increased P450c17α-activity in the single thecal cell.
A low estrogen/androgen ratio has a negative influence on follicle developement in PCOS women (Cataldo 1997). The ΔE2/ΔT ratio was slightly - but not - significantly higher in both controls and in PCOS women in hMG stimulated cycles than in FSH stimulated cycles, suggesting that hMG does not have unfavourable effects on aromatase activity as compared to FSH. The higher increase of serum A concentrations in hMG stimulated cycles may, however, have some unfavourable effects on endometrial cell growth, as has recently been reported (Tuckerman et al. 2000). Nevertheless, the clinical results (including pregnancy rate) were similar in all treatment groups, which is in accordance with previous studies (Homburg et al. 1990, McFaul et al. 1990, Tanbo et al. 1990).
In conclusion, the present findings indicate that both hMG and FSH stimulate ovarian androgen production after pituitary suppression in normal and PCOS women. The PCOS women did not show a distinctly exaggerated steroid response to gonadotrophins and no sign of steroidogenic defects was observed under these experimental conditions.
The improvement of hyperandrogenism observed during metformin treatment has been considered to be the result of a decrease in serum insulin concentrations (Velazquez et al. 1994, Nestler & Jakubowicz 1996, Morin-Papunen et al. 1998, Pasquali et al. 2000), although not all studies support this concept (Acbay & Gundogdu 1996, Ehrmann et al. 1997a). In the present study, we observed a slight, but non-significant decrease in basal serum androgen concentrations during metformin treatment. In addition, we observed significant decreases in AUCT and AUCA after hCG during metformin treatment. During that treatment fasting serum insulin concentration decreased and insulin sensitivity improved in obese and insulin-resistant PCOS women, suggesting that the slight alleviation of hyperandrogenism brought about by metformin may be mediated by a decreased insulin action. No change in BMI or in WHR was observed during the study.
Given that significant decreases in GnRHa and hCG-stimulated 17-OHP responses have been observed during metformin treatment, it has been suggested that an improvement of hyperinsulinemia by metformin treatment could reduce ovarian cytochrome P450c17α activity (Nestler & Jakubowicz 1996, Nestler & Jakubowicz 1997, la Marca et al. 2000). In contrast to these studies, however, the response of 17-OHP to hCG was not significantly decreased by metformin in the present study, although insulin levels decreased. The results indicate that a 2-month treatment period with metformin does not modify the response pattern of 17-OHP. These data suggest that - if the increased response of 17-OHP to hCG in women with PCOS is partly due to hyperactivity of P450c17α (Rosenfield et al. 1994, Levrant et al. 1997) - short-term metformin treatment may not improve hyperandrogenism by affecting this step of androgen biosynthesis. Insulin is known to augment the expression of P450c17α (Ehrmann et al. 1995), possessing both 17α-hydroxylase and 17,20-lyase activities in the ovary, and thus, decreased serum insulin concentrations after metformin treatment in our study may be the result of an improvement of hyperandrogenism.
The present results, as well as those of our previous studies (Morin-Papunen et al. 1998, Morin-Papunen et al. 2000), showed that serum SHBG concentrations and the FAI did not change significantly during the 2 months of metformin treatment. This is in agreement with a recent study of long-term metformin treatment (Pasquali et al. 2000) and in contrast to some previous studies where increased serum SHBG concentrations during metformin treatment have been observed (la Marca et al. 1999, Unluhizarci et al. 1999a,b, la Marca et al. 2000). One explanation for this discrepancy may be that abdominal obesity is associated with reduced SHBG concentrations (Pasquali et al. 1990), and since no weight loss was observed after the treatment in our study, SHBG did not change. Moreover, the women with PCOS in our study were probably still relatively hyperinsulinemic after their treatment and no change in SHBG was therefore noticed.
The present results suggest that the slight alleviation of hyperandrogenism brought about by metformin therapy may be explained by decreased ovarian steroidogenesis, probably due to a reduced serum insulin concentration.
During metformin treatment a significant decrease of serum insulin and leptin concentrations at two months of treatment was observed in parallel with T changes. This is in accordance with previous studies (Morin-Papunen et al. 1998, Morin-Papunen et al. 2000, Pasquali et al. 2000). Messenger RNA for leptin receptors has been found in both ovarian granulosa and theca cells (Agarwal et al. 1999), suggesting a possible direct role of leptin in ovarian function. This may explain the parallelism observed between changes in leptin and T concentrations.
Insulin stimulates leptin production in vitro (Kennedy et al. 1997) and in vivo (Kolacynski et al. 1996). The improvement of hyperinsulinemia in our study may also decrease serum leptin concentrations. The relationship between hyperinsulinemia, insulin resistance and serum leptin concentrations is not clear, however, and as the role of leptin in PCOS is not understood, more studies are needed to clarify whether the improvement of hyperinsulinemia can decrease serum leptin concentrations.
As metformin may inhibit lipolysis in adipose tissue and thus play some role in the metabolism of fat cells (Riccio et al. 1991), a direct effect of metformin on the secretion of leptin in fat tissue cannot be excluded.