Tata B, Mimouni NEH, Barbotin AL, et al. Elevated prenatal anti-Müllerian hormone reprograms the fetus and induces polycystic ovary syndrome in adulthood. Nat Med. 2018;24(6):834-846. doi: 10.1038/s41591-018-0035-5
Human retrospective correlational study and mouse treatment study
To determine whether anti-Müllerian hormone (AMH) remains elevated during pregnancy in women with polycystic ovary syndrome (PCOS) and to further investigate the association of prenatal exposure to excess AMH with PCOS incidence in offspring.
Previously collected data was obtained from the Uppsala Biobank of Pregnant Women in Sweden. Included for analysis were records of 66 pregnant women, aged 27 to 39 years, with PCOS diagnosis based on Rotterdam Criteria. Women were categorized according to BMI, androgen status, and age. Records from pregnant, healthy controls (n=63) were matched for age and BMI.
Timed-pregnant adult mice were included for initial experimentation. Those, along with their adult female offspring, were utilized for data analysis.
Pregnant mice received injections of either phosphate-buffered saline (PBS) as the control, PBS + human recombinant prenatal AMH (PAMH), or PBS + pro-AMH (PAMH) + gonadotropin-releasing hormone (GnRH) antagonist (cetrorelix acetate) during the late gestational period (embryonic days 16.5 to 18.5). The female offspring of the PAMH-treated group received injections of PBS + GnRH antagonist during adulthood.
Pregnant women: serum AMH (pmol/L) measured during gestational weeks 16 through 19
Pregnant mice: plasma testosterone (ng/mL), luteinizing hormone (LH; ng/mL), estradiol (pg/mL), and progesterone (pg/mL)
Adult female offspring of treated mice: BMI; anogenital distance (mm); onset of puberty (vaginal opening and onset of first estrous cycle); fertility status (time of first litter following pairing, total number of litters, number of pups per litter); estrous cyclicity (time spent in each phase of the cycle); plasma testosterone and LH (ng/mL); LH pulsatility (number of pulses/2h); and ovarian histologic findings (number of corpora lutea and late antral follicles). The latter 4 measures were obtained both at puberty and after injection of GnRH.
Pregnant women with PCOS exhibited elevated AMH compared to pregnant controls (P≤0.0005). When stratified according to BMI (lean BMI≤25; obese BMI≥30), androgen status (normoandrogenic, hyperandrogenic), and age (27-34 years; >34 years), hyperandrogenic lean PCOS was the most likely phenotype to present with elevated AMH levels. No significant correlation was found with age.
Compared to controls, the AMH-treated adult pregnant mice had significantly higher testosterone (P=0.0023) and LH (P=0.004), and lower estradiol (P<0.0001) and progesterone (P<0.0001).
The PAMH-treated adult female offspring exhibited longer anogenital distance, delayed puberty (delayed vaginal opening and later onset of first estrous cycle), severely disrupted estrous cyclicity (decreased time in the preovulatory stage and prolonged time in metestrus and diestrus), elevated testosterone and LH, heightened LH pulsatility, abnormal ovarian histologic findings (fewer corpora lutea and late antral follicles), and impaired fertility (delay in first litter, fewer litters, fewer pups per litter) compared to controls. Following injection with the GnRH antagonist, hormone levels normalized, estrous cyclicity regulated, and ovarian histologic morphology improved.
Anti-Müllerian hormone, also referred to as Müllerian-inhibiting substance, is a glycoprotein belonging to the transforming growth factor (TGF)-beta family. When produced by fetal Sertoli cells, regression of the Müllerian duct occurs and sexual differentiation of the male reproductive tract results.1 In females, the hormone is synthesized by ovarian granulosa cells in the preantral and small antral follicles.2 It inhibits recruitment of primordial follicles and follicle-stimulating hormone (FSH)–dependent growth of the small developing follicles, thus acting as a regulator of folliculogenesis.3
The present study aimed to confirm the hypothesis that AMH will remain elevated in PCOS women during pregnancy, despite the usual decline during normal pregnancy.
Past research reveals that AMH levels are higher in women with PCOS compared to those without.4,5 This is unsurprising given that a characteristic finding of PCOS is follicular excess, particularly of the preantral and small antral follicles producing AMH.6 The present study aimed to confirm the hypothesis that AMH will remain elevated in PCOS women during pregnancy, despite the usual decline during normal pregnancy.7 In addition to confirming the hypothesis, the data revealed that elevations in AMH are most significant among hyperandrogenic lean PCOS pregnant women.
The mice in this study develop hyperandrogenism and additional imbalances in LH, estradiol, and progesterone following injection with PAMH, suggesting that elevated AMH was a causative factor in the hormone dysregulation mimicking PCOS. Similar hormonal patterns and resultant abnormalities in puberty onset, estrous cyclicity, and fertility were noted in PAMH-treated offspring. This further demonstrates the contributory role that excess AMH has on the development of an inherited PCOS-like phenotype.
AMH is classically used to assess ovarian reserve, though recent research reveals its potential novel role in PCOS diagnosis.5 So far in 2018, 10 papers have been published on the diagnostic potential of AMH. Indran et al proposed a simplified PCOS criteria that replaces abnormal ovarian morphologic ultrasound findings with elevated levels of AMH. Compared with the traditional Rotterdam criteria (Rotterdam 2003), AMH used as a single diagnostic biomarker revealed poor specificity (58.9%), but the simplified criteria, which uses AMH in place of ovarian morphology, demonstrated a 94% diagnostic overlap.8 Saxena et al similarly found that AMH levels, when used adjunctively with the existing Rotterdam criteria, had promising diagnostic potential.9 In the most recent paper, Yue et al suggest the optimal diagnostic threshold is 8.16 ng/mL (20-29 years) or 5.89 ng/mL (30-39 years); however, there are discrepancies about the cutoff most appropriate for diagnosis.10,11
In the present study, prenatal cotreatment of PAMH with the GnRH antagonist cetrorelix acetate prevented offspring from acquiring the neuroendocrine and reproductive defects that mice treated with PAMH-only did. This finding suggests that “prenatal programming” of the condition is GnRH-dependent. In offspring born to PAMH mice, treatment with the GnRH antagonist led to reversal of inherited PCOS-like traits. This further supports the theory that GnRH plays a significant role in PCOS pathophysiology.
The authors of this paper suggest that GnRH antagonism may be a worthwhile therapeutic target for PCOS treatment. The restoration of ovulation and fertility that was observed in cetrorelix-treated mice is a promising finding, though it remains unclear how these outcomes will translate to humans. Treatment with GnRH antagonists like cetrorelix has been studied in the PCOS population, though only in association with assisted reproduction. A 2010 study published by the Journal of Obstetrics and Gynaecology Research reveals that GnRH antagonists, compared to GnRH agonists, are more effective, safe, and well-tolerated among those with PCOS.12 Thus, it seems reasonable to believe that GnRH antagonism may become a target for PCOS treatment.
Assuming, though prematurely, that the above mouse findings can be extrapolated to humans, lowering AMH may also be a beneficial treatment approach. Numerous studies have assessed the impact of metformin on AMH levels in women with PCOS. Findings are overall inconsistent, with reported outcomes ranging from no change at all13 to statistically significant changes.14 No correlations could be made with differing variables of each study (metformin dose, treatment duration, PCOS phenotype) and the outcomes. Several studies looked at oral contraceptives alone and compared them to metformin. Results were also mixed.15-17 Studies on the following other commonly prescribed naturopathic PCOS treatments have yielded favorable results in terms of reducing AMH: aerobic exercise,18 dietary intervention,19 weight loss,20 myo-inositol plus folic acid,18 D-chiro-inositol21 and cinnamon.22
The implications of this study are significant. The results strongly suggest a potential cause, whether sole or contributory, of PCOS, and reveal 2 novel treatment approaches. To assure that outcomes are consistent among humans, further research is necessary. Until then, continuing to recommend the tried-and-true naturopathic treatments for PCOS, many of which are highlighted above, appears worthwhile for several reasons. Doing so may not only improve symptomology in these women, but may also decrease the likelihood that their children go on to develop PCOS.