The role of aromatase inhibitors in ameliorating deleterious effects of ovarian stimulation on outcome of infertility treatment.

We hypothesize that aromatase inhibitors can be used to improve the treatment outcome after ovarian stimulation either alone, or in combination with IUI and assisted reproductive technology. The use of aromatase inhibitors during ovarian stimulation may have several benefits including: (1) The enhancement of implantation by lowering the supraphysiological levels of estrogen attained during ovarian hyperstimulation that is believed to adversely affect the development of the endometrium, oocytes and embryo, as well as other possible targets. (2) Reducing gonadotropins dose required for achievement of optimum ovarian stimulation. This would reduce possible deleterious direct effects of exogenous gonadotropins injection in addition to reducing the cost of treatment. (3) Other possible benefits such as improvement of ovarian response to FSH stimulation in poor responders, and prevention of premature endogenous gonadotropin surge, as well as lower risk of severe ovarian hyperstimulation syndrome. The improvement in implantation, as well as the reduced cost of treatment by decreasing the gonadotropins dose required for ovarian hyperstimulation would encourage the policy of transferring one embryo to minimize the risk of multiple gestation. This would have a tremendous economic impact on the practice of assisted reproduction, as well as cost of health care for multiple gestation worldwide. In the following section, we present brief outline on the aromatase enzyme, estrogen biosynthesis and the development of aromatase inhibitors followed by a summary of the available data concerning the use of aromatase inhibitors for ovarian stimulation. Aromatase is a microsomal member of the cytochrome P450 hemoprotein-containing enzyme complex superfamily (P450 arom, the product of the CYP19 gene) that catalyzes the rate-limiting step in the production of estrogens, that is, the conversion of androstenedione and testosterone via three hydroxylation steps to estrone and E 2 respectively [ 393 ]. Aromatase activity is present in many tissues, such as ovaries, brain, adipose tissue, muscle, liver, breast tissue, and in malignant breast tumors. The main sources of circulating estrogens are the ovaries in premenopausal women and adipose tissue in postmenopausal women [ 394 , 395 ]. Aromatase is a good target for selective inhibition because estrogen production is a terminal step in the biosynthetic sequence. Several aromatase inhibitors have been utilized in clinical studies over the last 20 years. The most successful, third generation aromatase inhibitors are licensed for breast cancer treatment [ 396 ]. Aromatase inhibitors have been classified in a number of different ways, including first-, second-, and third-generation; steroidal and non-steroidal; reversible (ionic binding), and irreversible (suicide inhibitor, covalent binding) [ 397 ]. Table 4 lists the different classes of aromatase inhibitors. Different classes of aromatase inhibitors: The different degrees of whole body aromatase inhibition by the various aromatase inhibitors Steroidal aromatase inhibitors are androstenedione analogues that act as a false substrate and bind irreversibly to the androgen-binding site of the enzyme [ 398 ]. Non-steroidal aromatase inhibitors exert their function through binding to the heme moiety of the cytochrome P450 enzyme [ 399 ]. The first of these inhibitors to be used clinically was aminoglutethimide, which induces a medical adrenalectomy by inhibiting many other enzymes involved in steroid biosynthesis [ 400 ]. Although aminoglutethimide is an effective hormonal agent in postmenopausal breast cancer, its use is complicated by the need for concurrent corticosteroid replacement, in addition to side effects like lethargy, rashes, nausea and fever that result in 8–15% of patients stopping treatment [ 401 , 402 ]. The lack of specificity and unfavorable toxicity profile of aminoglutethimide led to the search for more specific aromatase inhibitors. In addition, the above mentioned aromatase inhibitors were not able to completely inhibit aromatase activity in premenopausal patients. The third-generation anti-aromatase agents commercially available include two non-steroidal preparations, anastrozole and letrozole, and a steroidal agent, exemestane. Anastrozole and letrozole are often referred to as aromatase inhibitors, whereas exemestane is called an aromatase inactivator [ 403 , 404 ]. Anastrozole, ZN 1033, (Arimidex ® ) and letrozole, CGS 20267, (Femara ® ) are selective aromatase inhibitors, available for clinical use in North America, Europe and other parts of the world for treatment of postmenopausal breast cancer. These triazole (antifungal) derivatives are reversible, competitive aromatase inhibitors, which are highly potent and selective. At doses of 1–5 mg/day, they inhibit estrogen levels by 97% to >99% resulting in estrogen concentrations below detection by most sensitive immunoassays. Table 5 shows the relative potencies of different aromatase inhibitors. Aromatase inhibitors are completely absorbed after oral administration, with mean terminal half-life (t 1/2 ) of approximately 45 hours (range, 30–60 hours). They are cleared from the systemic circulation mainly by the liver. Gastrointestinal disturbances account for most of the adverse events, although these have seldom limited therapy. Other adverse effects are asthenia, hot flashes, headache, and back pain [ 406 ]. The average wholesale cost for a month's supply (thirty tablets) of aromatase inhibitors is about $150 to $250 [ 407 ]. Table 6 lists the clinical advantages of the third generation aromatase inhibitors as regards potential use for ovarian stimulation. Advantages of the third generation aromatase inhibitors: In the late 1990s, we explored the hypothesis that it might be possible to mimic the action of clomiphene citrate, without depletion of ER, by administration of an aromatase inhibitor in the early part of the menstrual cycle. We hypothesized that the result of blocking estrogen production would be release of the hypothalamic/pituitary axis from estrogenic negative feedback, thereby increasing gonadotropin secretion and resulting in stimulation of ovarian follicular development. This first hypothesis is referred to as CENTRAL hypothesis. The selective non-steroidal aromatase inhibitors have a relatively short half-life (approximately 40 hours) compared to clomiphene citrate, and would be ideal for this purpose since they are eliminated from the body rapidly [ 408 , 409 ]. In addition, no adverse effects are expected on estrogen target tissues, since no ER down-regulation occurs as observed in clomiphene citrate treatment cycles. We subsequently developed a second hypothesis, which was referred to as the PERIPHERAL hypothesis to explain another mechanism of action of the aromatase inhibitors in ovarian stimulation. We believe that aromatase inhibitors also act locally in the ovary to increase follicular sensitivity to FSH stimulation. This may result from the temporary accumulation of intraovarian androgens, since conversion of androgen substrate to estrogen is reversibly blocked by aromatase inhibition. There are data supporting a stimulatory role for androgens in early follicular growth in primates [ 410 ]. Testosterone was found to augment follicular FSH receptor expression in primates suggesting that androgens promote follicular growth and estrogen biosynthesis indirectly by amplifying FSH effects [ 411 , 412 ]. In addition, androgen accumulation in the follicle may stimulate insulin-like growth factor I (IGF-I), along with other endocrine and paracrine factors, which may synergize with FSH to promote folliculogenesis [ 413 - 416 ]. This hypothesis still waits for evidence to support it. In the last few years, we have worked on the development of aromatase inhibitors for ovarian stimulation and infertility management and reported interesting success in different applications as follows: Based on our CENTRAL hypothesis of using an aromatase inhibitor for induction of ovulation (as explained above), the success of an aromatase inhibitor in inducing ovulation in patients with PCOS was reported [ 417 - 420 ]. After demonstrating success in inducing ovulation in anovulatory women, we proceeded to test whether aromatase inhibition might enhance the release of endogenous gonadotropins enough to stimulate the development of multiple follicles in ovulatory women and result in augmentation of ovulation or even controlled ovarian hyperstimulation. The use of an aromatase inhibitor for augmenting ovulation in patients with ovulatory infertility was successful in women with unexplained infertility, endometriosis, and women undergoing therapeutic donor insemination, and in ovulating partners of infertile men [ 419 , 420 ]. For over 40 years, clomiphene citrate has been the most commonly used treatment for the induction and augmentation of ovulation, accounting for about two thirds of the fertility drugs prescribed in the United States [ 421 , 422 ]. In spite of the high ovulation rate associated with the use of clomiphene citrate, the pregnancy rate is much lower than anticipated. This is particularly true when considering pregnancy rate per cycle after three cycles of CC treatment [ 421 ], with higher than expected incidence of miscarriage in conception cycles [ 255 ]. Such discrepancy is believed, as explained earlier, to be due to the peripheral antiestrogenic effect of clomiphene citrate particularly at the level of the cervical mucus and endometrium, and which manifest themselves even in the presence of gonadotropin treatment superovulation. The accumulation in the body of the isomers of clomiphene citrate due to the long half-life (several days to weeks) adds to the persistence of the antiestrogenic effect [ 261 , 262 ]. In order to improve the outcome of clomiphene citrate treatment, various approaches have been suggested to overcome the antiestrogenic effect including concomitant estrogen administration. Some investigators reported increased endometrial thickness and improved pregnancy rates with this approach [ 423 - 425 ] while others have reported no benefit [ 426 ] or even a deleterious effect of estrogen administration [ 427 ]. Another approach was to administer clomiphene citrate earlier during the menstrual cycle [ 428 ], to allow the anti-estrogenic effect to wear off earlier before the critical period of fertilization and implantation. A third approach has been to combine another selective ER modulator like tamoxifen, which has more estrogen agonistic effect on the endometrium with clomiphene citrate [ 429 ] or to use it as an alternative to clomiphene citrate [ 430 ]. However, none of these strategies have proved to be completely successful in avoiding the peripheral antiestrogenic effects of clomiphene citrate. In addition to a discrepancy between ovulation and pregnancy rates with clomiphene citrate treatment, 20% to 25% of anovulatory women are resistant to clomiphene citrate and fail to ovulate at doses up to 200 mg daily. The success of aromatase inhibition in inducing and augmenting ovulation encouraged trying aromatase inhibitors in cases of clomiphene citrate failure that was found to be successful in achieving ovulation and pregnancy [ 419 , 420 ]. As explained above, the significantly shorter half-life of the third generation aromatase inhibitors compared to clomiphene citrate, allows rapid elimination from the body [ 408 , 409 ]. In addition, since no ER down-regulation occurs, any adverse effects on estrogen target tissues, as observed in clomiphene citrate treated cycles, is expected. We investigated the idea of combining an aromatase inhibitor with FSH injections to reduce the dose of FSH required to achieve optimum controlled ovarian hyperstimulation, without adverse antiestrogenic effects [ 431 , 432 ]. A significant reduction in the FSH dose required (from 45% to 55% in women with PCOS and unexplained infertility) has been reported [ 433 ] without evidence of peripheral antiestrogenic effects [ 432 , 433 ]. Reducing FSH dose required for optimum controlled ovarian hyperstimulation encouraged us to explore the use of an aromatase inhibitor in conjunction with FSH to improve response to ovarian stimulation in poor responders. In a selected group of women who had a poor response to FSH stimulation in at least two prior treatment cycles, adding an aromatase inhibitor resulted in improvement in the response to FSH stimulation. All patients developed a significantly greater number of mature ovarian follicles and almost a third of the treatment cycles resulted in pregnancy. In addition, the dose of FSH required to achieve such optimum response was significantly less than the dose used in the prior cycles in which FSH was used alone [ 434 , 435 ]. During all the above-mentioned studies, the aromatase inhibitor, letrozole, was administered orally as a daily dose of 2.5 mg from day 3 to day 7 of the menstrual cycle. Based on the pharmacokinetics of the new aromatase inhibitors (almost 100% bioavailability after oral administration, ~2 days half-life and no accumulation or significant metabolite accumulation), we thought of a potentially more convenient method of administering an aromatase inhibitor for ovulation induction. We hypothesized that significantly higher concentration of the aromatase inhibitor can be achieved early in the menstrual cycle with faster clearance later in the menstrual cycle with the administration of a high single dose of the aromatase inhibitor early in the menstrual cycle such as on day 3. A single dose regimen would satisfy two goals: first, achieving maximum estrogen suppression early in the menstrual cycle when it is desired, and second, to allow clearance of the aromatase inhibitor before the critical final stage of fertilization and embryogenesis, to maximize safety and avoid any possible undesirable effects of the aromatase inhibitors. Single dose administration of an aromatase inhibitor was found successful in inducing ovulation with ovulation and pregnancy rates comparable to the 5-day regimen [ 436 ]. In all our clinical trials described above, results were not obtained from randomized, prospective, placebo controlled studies, the optimum research design. However, due to the experimental nature of the use of the aromatase inhibitors for ovarian stimulation, which to our knowledge, has never been used before, we believed that the present observational trials were mandatory before proceeding into any definitive randomized studies. The encouraging results of our data have led other investigators from different centers world-wide to study the use of aromatase inhibitors for ovarian stimulation and in general have reported findings similar to ours in terms of the success of aromatase inhibitors in infertility treatment [ 437 - 450 ]. In a randomized, prospective study, a superior uterine environment has been found in patients treated with an aromatase inhibitor compared with clomiphene citrate, reflecting the lack of the antiestrogenic effects with aromatase inhibitor treatment. Although a non-significant increase in pregnancy was observed in patients who received aromatase inhibitor treatment (16.7 versus 5.6% per patient, P = 0.55), this almost three-fold increase in pregnancy rate would have required about ten more patients per group to reach statistical significance. The superiority of a single dose administration of an aromatase inhibitor was also reported when compared to clomiphene citrate [ 439 ]. In other studies, the use of an aromatase inhibitor in conjunction with FSH was found to reduce the FSH dose [ 440 ] and improve response to ovarian stimulation in poor responders [ 441 , 442 ], confirming our previous reports. In another prospective randomized trial, Biljan [ 438 ] studied two doses (2.5 and 5 mg per day) of the aromatase inhibitor, letrozole for ovarian stimulation in patients with unexplained infertility. They found patients treated with a higher dose of letrozole developed more follicles without a detrimental effect on endometrial development. The potential for different doses and regimens of administration of aromatase inhibitors for ovarian stimulation still requires a lot of study, but the use of aromatase inhibitors at high doses should be cautiously considered. In all reports and most of studies of other investigators, the aromatase inhibitor, letrozole was the one used. However, anastrozole, another third generation aromatase inhibitor similar to letrozole, was used in other studies [ 443 , 444 ]. Due to the similar pharmacokinetics and pharmacodynamics, including similar potencies and specificity in inhibiting the aromatase enzyme, we believe there is likely to be no difference between the third generation aromatase inhibitors in their efficacy for ovarian stimulation. It does, however, need to be determined what dose of anastrozole is required to be equivalent to letrozole. It seems that there are many interesting areas of research that need exploration as regards the development of aromatase inhibitors for infertility management. These directions for research would include: confirming the available preliminary data on the success of aromatase inhibition in induction and augmentation of ovulation, as well as reducing the dose of FSH needed for ovarian stimulation, improving response in poor responders, and finding the optimum regimen for administering aromatase inhibitors for infertility treatment Moreover, the use of aromatase inhibitors for new applications including in-vitro maturation and prevention of severe ovarian hyperstimulation syndrome and endometriosis-related infertility are interesting future avenues for aromatase inhibitors potential use in infertility management. In addition the use of aromatase inhibitors to improve the outcome of treatment after assisted reproduction as discussed earlier in this review is an exciting area of application. In addition to using aromatase inhibitors, ALONE, for ovarian stimulation, their use during assisted reproduction carries several potentials to improve the treatment outcome. We hypothesize that aromatase inhibition during assisted reproduction may improve the implantation rate mainly by reducing the estrogen levels attained during COH. In addition, two other mechanisms may apply including reducing the dose of FSH required for optimum COH as well as applying much simpler stimulation protocols that do not require the use of GnRH analogues, hence avoiding any possible direct deleterious effects of FSH and GnRH analogues on the endometrium. As discussed above, the undesirable effects of ovarian stimulation on the outcome of infertility treatment are believed to be due to the supraphysiological levels of estrogen irrespective to whatever mechanisms explain for that (various postulated mechanisms were discussed above). So, lowering estrogen levels may be associated with improved outcome of treatment in terms of improving the implantation and pregnancy rates in addition to lowering risk of sever ovarian hyperstimulation syndrome. Reducing estrogen synthesis by aromatase inhibition seems to be an exciting idea to ameliorate the deleterious effects of the supraphysiological levels of estrogen on the endometrium, the developing oocyte and embryo as well as the luteal. Until recently there was no suitable aromatase inhibitor that could be used clinically to reduce estrogen levels during ovarian stimulation. This is because the available aromatase inhibitors were not safe for clinical application during ovarian stimulation due to lack of specificity in inhibiting the aromatase enzyme without inhibiting other steroidogenesis enzymes (e.g. aminoglutethimide). The other aromatase inhibitors (steroidal androstenedione analogues) were irreversible in their effect on the aromatase enzyme besides being parentrally administered. Most important, these old aromatase inhibitors were not potent enough to inhibit the aromatase enzyme and lower estrogen levels in women of the reproductive age group. However, the third generation non-steroidal aromatase inhibitors group is very potent and specific in inhibiting the aromatase enzyme reversibly. These new aromatase inhibitors are orally administered with very high safety profile and well tolerability. Moreover, they are cheap with a relatively short half-life [~45 hours], and already approved for clinical use to reduce estrogen production in postmenopausal women with breast cancer. They have not been used in women of the reproductive age group. However, we have found these aromatase inhibitors to be effective in inhibiting the aromatase enzyme and effectively decrease estrogen levels in women of the reproductive age group during their successful use for ovarian stimulation [ 417 , 418 , 431 , 432 ]. In our experience with the use of an aromatase inhibitor for ovarian stimulation, estrogen levels were significantly lower (especially E 2 level/mature follicle) when compared with conventional stimulation protocols (clomiphene citrate, FSH and clomiphene citrate plus FSH). Such low E 2 levels may be beneficial and explains at least partially the improved outcome of treatment in terms of achieving promising high pregnancy rates during aromatase inhibitor treatment. It may not be only the supraphysiological estrogen levels, which explain totally for the reduced implantation rate during ovarian stimulation cycles. Local paracrine factors, or possibly other undetermined factors may be responsible for the reduced implantation rate during COH. In this case the aromatase inhibitors may not offer a complete solution to overcome such drawbacks of ovarian stimulation. However, the idea is exciting and seems to be promising and warrant trials because high estrogen levels may explain, at least in part, for the deleterious effects of ovarian stimulation on treatment outcome. In ddition, we think that, reducing estrogen levels achieved during induction of ovulation by using aromatase inhibitors may prevent the significant increase in leptin concentrations avoiding its possible deleterious effects on the outcome of treatment as explained above. Because elucidation of leptin's specific role in reproductive function has been challenging, with conflicting results reported by various investigators, it is still quite early and highly speculative to hypothesize a link between aromatase inhibitors use and the role of leptin during induction of ovulation. However, the available strong data about a possible role of leptin in mediating reproductive disorders especially in obese women and the firm findings of a regulatory effect of estrogen on leptin production make this hypothesis exciting and interesting enough to warrant future investigation. As mentioned earlier, the use of an aromatase inhibitor significantly reduces the dose of FSH required for optimum COH. In addition to the significant economic benefit, we believe that reducing the dose of FSH may improve the treatment outcome of ovarian stimulation by reducing any possible deleterious effect of the exogenously administered FSH on the endometrium, developing oocyte or other targets. It is believed that the endogenous LH surge arises once serum estrogen levels surpass a set threshold for a certain period of time [ 451 - 453 ]. The supraphysiological estrogen levels attained during ovarian stimulation are believed to cause premature release of the endogenous LH surge resulting in cycle cancellation during assisted reproduction. For that reason, GnRH analogues have been the standard of practice during most of the stimulation protocols for assisted reproduction to prevent the occurrence of endogenous LH surge either by direct inhibition of LH surge (GnRH antagonists) or by down-regulation of the GnRH receptors at the pituitary levels resulting in pituitary desensitization and prevention of the release of endogenous LH [ 454 , 455 ]. Unfortunately, the use of GnRH analogues (which is a crucial part of the stimulation protocol to prevent cycle cancellation as explained above), is associated with several problems including increasing the dose of FSH required to achieve optimum COH (due to suppressing the endogenous gonadotropin secretion, and possible peripheral effect at the level of the ovaries), as well as the luteal phase defect due to a dysfunctional corpus luteum function secondary to persistent endogenous LH suppression as explained earlier. In addition, there is rising evidence of a possible direct deleterious effect of the GnRH analogues, especially the antagonist at the level of the endometrium [ 456 ]. The use of an aromatase inhibitor to reduce estrogen levels attained during COH may be effective in preventing the occurrence of premature ovarian surge. This would avoid the use of GnRH analogues during stimulation protocols for assisted reproduction which has several advantages including prevention of the possible deleterious effects of these agents as mentioned above in addition to reducing the cost of treatment as well allowing implementing much simpler stimulation protocols during assisted reproduction. The expression of aromatase enzyme in endometriotic tissues with the significant role of locally produced estrogen in endometriosis progression [ 457 ] suggests a benefit of aromatase inhibitors in endometriosis-related infertility. The inhibition of local estrogen production in endometrial implants, and the lower peripheral estrogen levels associated with the use of aromatase inhibition for ovulation induction, is expected to protect, to some degree, against progression of endometriosis during infertility treatment. There are three main concerns that may arise regarding the use of the aromatase inhibitors to improve the outcome of treatment after ovarian stimulation and assisted reproduction. They include the possible deleterious effect of the low follicular estrogen milieu on the development of the oocytes, the possible direct effect of the aromatase inhibitors on the oocyte development, fertilization or embryogenesis, and the accumulation of the androgens that may result from inhibition of their conversion into estrogens. Palter et al. have reviewed the question whether an estrogen-free (or at least very low estrogen) intrafollicular environment is compatible with follicular development, ovulation, and corpus luteum formation [ 458 ]. The authors concluded that markedly reduced to nonexistent intrafollicular and circulating concentrations of estrogen are compatible with follicular "expansion", retrievable and fertilizable oocytes, as well as with cleavable and apparently transferable embryos. The authors drew their conclusion after discussing lessons from data in the literature including cases of deficiency of the 17 -hydroxylase/17–20 lyase [ 459 - 466 ] and 3β-hydroxysteroid dehydrogenase [ 467 , 468 ] and the aromatase enzyme [ 469 - 473 ] as well as cases of severe hypogonadotropism [ 474 - 478 ]. In addition, data on the use of aromatase inhibitors in animals were reviewed [ 479 - 482 ]. 17alpha-hydroxylase/17–20 lyase deficiency is one form of congenital adrenal hyperplasia, which is associated with marked impairment of glucocorticoid, androgen, and estrogen biosynthesis [ 459 ]. Women suffering from this enzyme deficiency suffer from hypergonadotropic hypogonadism and sexual infantilism. However, early reports noted the presence of many primary and secondary follicles in ovarian material [ 460 ] and many of them had bilateral multicystic ovaries at the time of laparotomy [ 461 - 463 ]. It is interesting that Rabinovici reported on a patient afflicted with virtually complete 17-hydroxylase/17–20 lyase deficiency that, despite castrate levels of estrogens, underwent an apparently successful induction of ovulation associated with progressive follicular expansion and oocyte retrieval, IVF-ET, and early embryonic cleavage followed suit [ 464 ]. The deficiency of 3β-HSD is associated with markedly reduced levels of estrogen. However, the existence of normal ovulatory function in a woman with late-onset of a mild form of 3β-HSD has been reported [ 483 ]. In a group of cycling female Rhesus monkeys exposed to ovulation induction with hFSH and hFSH + hLH in the absence or presence of the 3β-HSD inhibitor trilostane given on days 1–8 of the menstrual cycle [ 467 ], apparently healthy oocytes were obtained by follicular aspiration 34 h after hCG administration. Importantly, treatment with the 3β-HSD inhibitor, trilostane, led to a reduction in serum E 2 levels to 7% of that of control animals throughout the follicular phase. Despite this dramatic reduction in E 2 levels, neither the total number of large antral follicles per animal (17 ± 1 vs. 18 ± 2) nor their size distribution differed significantly from 3β-HSD inhibitor-untreated controls. Furthermore, treatment with the 3β-HSD inhibitor did not alter the overall maturation pattern of collected oocytes (atretic, prophase I, metaphase I, or metaphase II). However, the authors found a reduction in the percentage of metaphase II oocytes that were successfully fertilized (15 vs. 65%). Moreover, metaphase oocytes that required more than 8 h to complete meiosis in vitro failed to fertilize in three of four animals receiving 3β-HSD inhibitor relative to controls (31%). These observations suggest that follicular development and the completion of meiosis may be unaffected by the low estrogen levels but that cytoplasmic oocyte maturation and/or function could be unfavorably affected [ 467 ]. Obviously, aromatase (estrogen synthase) enzyme deficiency is associated with marked decrease or almost absence of estrogen production. Extreme examples of complete aromatase deficiency due to mutations in the aromatase gene, CYP19 gene, in adult human females, however, were reported [ 469 , 470 ]. The affected patients suffered from ambiguous external genitalia, primary amenorrhea, sexual infantilism, and multicystic ovaries. Morishima [ 471 ] reported on the aromatase deficiency in a 28-yr-old 46 XX proband followed since infancy. Null mutant mice for aromatase gene [ArKO] were generated [ 484 ], thereby affording the opportunity to examine the role of estrogen in the follicular development in the mouse ovary. Evaluation of the ovaries revealed the presence of many large follicles filled with granulosa cells and evidence of antrum formation, but no corpora lutea. The ovarian phenotype degenerated with age upon the appearance of hemorrhagic cystic follicles and the loss of secondary and antral follicles coincident with the infiltration of macrophages and with stromal hyperplasia [ 485 - 488 ]. Therefore, the ArKO females are infertile, due primarily to a complete lack of ovulation. However, recently, Jones et al., reported that oocytes were harvested from the ovaries of 4- and 7-week old ArKO, wild type and heterozygote mice stimulated with 5 IU PMSG. The number of immature oocytes harvested from ArKO females did not differ from the number collected from wild type or heterozygote littermates of either age group. Oocyte in vitro maturation rates also did not differ between the three genotypes or two age groups, with almost 75% of the immature oocytes progressing to metaphase II. Chromatin staining confirmed the arrest of these oocytes at the second meiotic division with chromatin staining clearly present in the oocyte and polar body. Mature oocytes were inseminated and IVF-ET rates did not differ between the three genotypes or two age groups, with fertilization occurring in approximately 67% of oocytes. Fertilized oocytes were cultured to blastocysts. Again, blastocyst development rates did not differ between the groups, with approximately 65% of the zygotes developing into blastocysts. Blastocyst morphology was similar across all of the groups [ 489 ]. These results indicate that ArKO oocytes are competent to develop to at least the implantation stage. The authors concluded that estrogen might not be required for the production and maturation of developmentally competent oocytes. Rather, its role in folliculogenesis is probably via regulation of the hypophysial pituitary gonadal axis and thus gonadotropin secretion [ 489 ]. Regarding the effect of an estrogen-free/poor intrafollicular environment on gametogenic maturation, Palter et al. [ 458 ] concluded that the effect is a negative one. Their conclusion was based on a number of primate studies, which indicated that an estrogen-free/poor intrafollicular environment is associated with marked decrements in the rates of meiotic maturation and fertilization [ 467 , 490 ]. In addition data derived from rodent models suggested that further suggested a compromise in early embryonic development [ 491 ]. In cycling female rhesus monkeys, the aromatase inhibitor, 1,4,6-androstatrien-3, 17-dione (ATD) was used to inhibit estrogen production during gonadotropin ovarian stimulation [ 479 ]. Animals treated with ATD displayed a drastic reduction in serum 17β-E 2 levels to 37% of that of controls within 8 h of ATD treatment and to 16% of control by the day of hCG injection. In turn, the circulating levels of androstenedione rose. Despite the drastic reduction in the circulating levels of E 2 and the increase in the circulating levels of androgens, the overall number of large antral follicles [16 ± 3 for controls and 20 ± 3 for ATD-treated] and their size distribution (as assessed by ultrasonography) proved comparable for control and ATD-treated animals. Similarly, no difference was noted in the number of oocytes collected or in the proportion of oocytes reinitiating meiosis (MI at the time of collection). In contrast, ATD-treated animals displayed a marked increase (31 vs. 11%) in the proportion of prophase I oocytes. Moreover, ATD-treated oocytes displayed retarded in vivo completion of maturation to MII (4% vs. 26%). Interestingly, the latter retardation was not observed in vitro. Furthermore, two of the four ATD-treated animals yielded oocytes that were morphologically abnormal. Finally, oocytes from ATD-treated animals displayed significantly reduced rates of fertilization (9% vs. 25%) as compared with controls. However, the cleavage rate after successful fertilization was similar for ATD-treated vs. ATD-untreated controls. In vitro cultures of granulosa cells collected at the time of oocyte aspiration revealed equivalent 24-h progesterone production in treated and control animals. The authors concluded that these findings suggest that the acute reduction in E 2 levels during the terminal stage of gonadotropin stimulation had little effect upon follicular recruitment and expansion but an apparent detrimental effect upon gametogenic function may, in fact, exist. However, we cannot extrapolate from this study that the findings are simply due to the acute reduction of estrogen production as it is important to realize that there was a concomitant significant rise in androstenedione. Moreover, the time of administering an aromatase inhibitor during the latter part of the follicular phase as well as the irreversible nature of aromatase inhibition constitute important differences from our model of using a REVERSIBLE aromatase inhibitor, TEMPORARILY, EARLY in the menstrual cycle that has a SHORT half-life. Selvaraj [ 480 , 481 ] and Shetty [ 482 ] examine the effects of blocking estrogen biosynthesis during the follicular phase on follicular maturation in the adult female bonnet monkey. In their studies, they used one of the third-generation REVERSIBLE, aromatase inhibitors (CGS 16949A) starting EARLY on day 3 of the menstrual cycle. There were 53% and 70% reductions in the basal and surge levels of E 2 , respectively without obvious effect on follicular maturation, ovulation, and luteal function as assessed by serum hormone profiles as well as by laparotomy. Moreover, the concurrent administration of FSH and an aromatase inhibitor resulted in the suppression of the FSH-induced increase in the circulating levels of E 2 (by 100%) with no effect noted on either the number of follicles developed or their size relative to control. In addition, granulosa and theca cells, removed on day 9 of the treatment cycle, were responsive to gonadotropins in vitro, disclosing no evidence of a deleterious effect on the cellular development and maturation of follicular cells. Terry studied a possible role of high E 2 levels in mediating the adverse effects of hyperstimulation with PMSG on early embryonic development in the rat [ 491 ]. They used the aromatase inhibitor, 4-hydroxyandrostenedione (4-OHA), to inhibit endogenous E 2 production. The authors conducted three experiments. In the first, varying doses of 4-OHA were administered either concurrently with hCG to pro-estrus female rats hyperstimulated at early diestrus stage with 20 IU PMSG or alone into non-hyperstimulated pro-oestrus females. At high doses of 1000, 2000, or 5000 mg/rat, 4-OHA substantially improved the survival of embryos in hyperstimulated females with optimum protection at 2000 mg, while low doses of 100 and 500 mg/rat were ineffective. When administered alone, only the highest dose of 5000 mg/rat 4-OHA increased embryo count. In the second experiment, higher doses of PMSG were studied (30 or 40 IU), with or without 5000 mg/rat 4-OHA given at the time of hCG injection. PMSG proved to be more detrimental with increasing dose, and high doses of 4-OHA (5000 mg/rat) was needed to rescue embryos from death in the 30, but not 40, PMSG group. In the third experiment, the influence of the timing of 4-OHA treatment on its ability to improve the embryo count in hyperstimulated females was examined by introducing 4-OHA 24 h earlier, rather than at the time of hCG treatment. The results showed the importance of timing of 4-OHA administration, as 5000 mg/rat 4-OHA was able to restore embryo survival in the 40 PMSG hyperstimulated group only when it was administered 24 h before hCG injection. These results highlighted that 4-OHA, when administered at the appropriate time and dose, could reverse the negative effects of hyperstimulation from PMSG on early embryonic development. The authors concluded that this might be due to the suppression of estrogen production, thereby alleviating the supraphysiological level of E 2 , which is typically present in PMSG-treated females, which supports the hypothesis that excessive E 2 is responsible for the negative effects of hyperstimulation with PMSG on early embryonic development. The high success rates of ovulation and achievement of pregnancy in our reports on the use of aromatase inhibitor in ovarian stimulation, despite significantly lower estrogen levels could be due to several reasons. First: We used one of the third-generation "REVERSIBLE" aromatase inhibitors. Second: The aromatase inhibitor was administered EARLY in the follicular phase and for a limited period of time, which would allow the rapid clearance of the aromatase inhibitor from the body due to its SHORT half-life. Third: the reversible nature of the aromatase enzyme inhibition and the rising levels of FSH, which induce the expression of the aromatase enzyme, both would not allow for the estrogen levels to drop drastically below the low physiologic range. We believe that these low estrogen levels attained during the use of aromatase inhibitors for ovarian stimulation are compatible with healthy development of the oocytes, fertilization, embryo development, implantation and achievement of pregnancy. Fourth: the absence of significant rise in androgen levels would have prevented any possible unwanted effects on the terminal part of the oocyte maturation and ovulation. As explained above, the short half-life of the aromatase inhibitors and limiting the administration to the early part of the follicular phase would allow the rapid clearance of the medications from the body before the important stage of fertilization and embryogenesis. This in addition to the absence of accumulation of the aromatase inhibitors of any of their metabolites would make them safe for the ovarian stimulation. We have reported a preliminary data on the pregnancy outcome after the use of aromatase inhibitors for ovarian stimulation supporting the safety of using these medications for such indication [ 492 ]. Data have been accumulating regarding the absence of deleterious effects in association with aromatase inhibitor treatment on follicle/oocyte maturation and embryo development in mice [ 492 , 493 ]. Treatment with the aromatase inhibitor, anastrozole was associated with similar number of total follicles found per animal (30.4 follicles/animal in the control group and 27 follicles/animal in the anastrozole-treated group) as well as comparable rates of development of embryos, morulae, blastocysts, and hatching blastocysts between the two groups (P = 0.20, 0.10, 0.44, and 0.38, respectively) [ 493 ]. As mentioned before, the rapid clearance of the aromatase inhibitors from the body due to their short half-life and the reversible nature of the aromatase enzyme inhibition as well as the rising levels of FSH, which induce the expression of the aromatase enzyme, all would not allow the clearance of any accumulating androgens by converting them to estrogen. The conversion of the androgens into estrogens (the step catalyzed by the aromatase enzyme) is a terminal step in the cascade of steroidogenesis. So, substrate (androgens) accumulation is not expected to be very significant, as other alternative earlier steps in the steroidogenesis pathways will work to clear out the accumulating androgens. For these reasons, pharmacokinetic studies on the new aromatase inhibitors reported the absence of significant androgen elevations or abnormal changes in other steroids in patients receiving aromatase inhibitors for breast cancer. In a subset of our patients who received an aromatase inhibitor for ovarian stimulation, we did not find significant change in the androgen levels while receiving the medication when compared to stimulation with gonadotropins alone (unpublished data). Most recently we published data on the favorable pregnancy outcome after the use of letrozole for ovarian stimulation alone or in conjunction with gonadotropins [ 494 ]. This study looked at outcome of pregnancies achieved after ovarian stimulation with the aromatase inhibitor, letrozole, alone or in conjunction with gonadotropins. The study included a cohort of 394 pregnancy cycles achieved after letrozole and other ovarian stimulation treatments in a ddition to a control group of pregnancies spontaneously conceived without ovarian stimulation. These 394 pregnancy cycles were achieved as follows: 63 pregnancies with 2.5 mg of letrozole alone or with gonadotropins, 70 pregnancies with 5.0 mg of letrozole, 113 pregnancies with clomiphene alone or with gonadotropins, 110 pregnancies with gonadotropins alone, and 38 pregnancies achieved without ovarian stimulation. The study found pregnancies conceived after letrozole treatments were associated with similar miscarriage and ectopic pregnancy rates compared with all other groups. In addition, letrozole use was associated with a significantly lower rate of multiple gestation compared with clomiphene citrate. The findings of the stuy support the favorable pregnancy outcome and low multiple gestation rate in association with the use of aromatase inhibitors for ovarian stimulation [ 494 ]. Ovarian stimulation particularly in conjunction with assisted reproductive technologies aims at stimulating the recruitment of several mature ovarian follicles that would enhance the chance of successful treatment by obtaining several embryos readily available for transfer into the uterus. It is obvious that supraphysiological levels of estrogen are inevitably attained during ovarian stimulation due to the significant contribution of estrogen production from each one of the several mature ovarian follicles. There is growing evidence that such supraphysiological estrogen levels are deleterious on the development of the endometrium, oocytes, embryos as well as other targets through different mechanisms. This is believed to explain, at least in part, the low pregnancy rates associated with assisted reproduction, mainly as a result of the persistently low implantation rates that have not improved impressively along almost three decades of assisted reproduction experience. This led several investigators to adopt the concept of minimal ovarian stimulation for assisted reproduction as an alternative approach to reduce the deleterious effects of the supraphysiological estrogen levels attained during routine controlled ovarian hyperstimulation. Unfortunately such approach, despite being logical, is associated with the major drawback of achieving fewer oocytes than desired, which, expectedly, reduces the chances of treatment success. Recently, we reported the success of the new generation aromatase inhibitors in ovarian stimulation. We found these agents to be effective in reducing the amount of estrogen production by mature ovarian follicles significantly. Furthermore, we found the use of these agents to be associated with improved ovarian response to stimulation by gonadotropins, resulting in significant reduction in the dose of gonadotropins required for controlled ovarian hyperstimulation. The novel idea of using aromatase inhibitors during ovarian stimulation for assisted reproduction combines the benefit of a significant reduction in estrogen production as well as gonadotropins dose, with the exiting advantage of achieving a good number of mature oocytes as a result of enhanced ovarian response to gonadotropins. This is expected to improve the various aspects of treatment outcomes after assisted reproduction including increased safety, reduced cost, as well as enhanced implantation rate. If confirmed in well-designed clinical trials, this would facilitate acceptance of the concept of single embryo transfer by infertile couples and practitioners, thereby reducing the epidemic of multiple births after assisted reproduction with its significant deleterious health and economical effects. This is particularly true in the light of recent data supporting the success of the approach of single embryo transfer [ 495 - 497 ].

MFMM has written the review article, RFC and MPD have reviewed the manuscript