Targeting Elevated GnRH Pulsatility to Treat Polycystic Ovary Syndrome.

Polycystic ovary syndrome (PCOS) is the foremost cause of anovulatory subfertility affecting 3% to 13% of women. Increased gonadotropin-releasing hormone (GnRH) pulsatility is a characteristic feature of PCOS; women with PCOS have an approximately 40% increase in GnRH pulse frequency (from every 90 minutes to hourly), leading to increased luteinizing hormone (LH) secretion from the pituitary gland, and in turn increased ovarian androgen production (1). Variants in genes encoding both for LH (β subunit) and its receptor have been implicated in PCOS. Approximately half of women with PCOS have increased LH levels; more so in the presence of menstrual disturbance (1). Reducing the elevated LH level using sustained administration of GnRH agonists attenuates the associated hyperandrogenism.

Increased GnRH pulsatility in PCOS is mediated by reduced sensitivity to negative feedback from sex steroids, in particular to, progesterone and estradiol, which, at least in part, is due to elevated androgen levels. Thus, a vicious circle is established whereby GnRH pulsatility is less restrained by sex steroids, leading to increased LH levels, and in elevated androgen levels, thus perpetuating the lack of restraint on GnRH pulsatility by sex steroids. Notably, androgen exposure induces a PCOS-like phenotype in animal models via an action at the brain rather than the ovary (2). However, GnRH neurons do not have receptors for androgens nor progesterone, and thus sex steroids are believed to act via upstream neurons.

Kisspeptin neurons in the arcuate nucleus of the hypothalamus (analogous to infundibular nucleus in humans) are recognized to be key for pulsatile GnRH secretion. These kisspeptin neurons, which coexpress neurokinin B and dynorphin (termed KNDy neurons), do express the androgen receptor. KNDy neurons are believed to be the GnRH pulse generator and act in an auto/paracrine manner to induce pulsatile secretion of GnRH. The predominant action of neurokinin B (NKB) on these neurons is stimulatory via its receptor (NK3R). Indeed, genetic variants reducing NKB signaling manifest as hypogonadotropic hypogonadism with decreased GnRH pulsatility. Applying a “network interactome” approach to PCOS, genes related to kisspeptin signaling were posited as being noteworthy in that they were one of the few to act upstream of the central network hub of insulin (3).

Overall, antagonizing the stimulatory action of NKB is hypothesized to temper the increased GnRH pulsatility observed in women with PCOS, leading to a decrease in LH and subsequently androgen levels. This approach was first investigated using the NK3R antagonist, MLE4901 (previously known as AZD4901), at doses between 20 and 80 mg for 28 days in 65 women with PCOS (4). Women were required to have all 3 major features of PCOS (ie, hyperandrogenism, oligomenorrhea/amenorrhea, and polycystic ovarian morphology on ultrasound) to be eligible (4). Of note, only 67 of 403 women screened met these inclusion criteria (4). After 7 days of treatment with the highest dose of MLE4901 tested, LH pulse frequency was reduced by 0.44 pulses per hour, basal LH secretion by 79%, and total testosterone by 29% (4). Of women receiving MLE4901, 8 of 42 (19%) ovulated, as compared with 1 of 11 (9%) women receiving placebo; however, the study was of a short duration and did not aim to assess clinical outcomes such as ovulation (4).

NK3R antagonists are also in development for the treatment of postmenopausal hot flashes, although MLE4901 has been discontinued because of concerns over liver dysfunction. Fortunately, this appears to be an agent-specific side effect rather than a class effect, and other NK3R antagonists such as fezolinetant (previously known as ESN364) appear to be less susceptible to causing this. Fezolinetant has a shorter half-life than MLE4901 (t1/2 4-7 vs ~ 8.5 h), but effectively attenuated hot flashes at doses between 15 and 180 mg daily in 287 postmenopausal women. It was well tolerated with potential side effects including headache, nausea, diarrhea, and fatigue. In healthy volunteers receiving fezolinetant, maximal LH and testosterone suppression was observed at 6 to 12 hours following administration, and these reverted to near baseline levels by 24 hours (5).

In this issue of JCEM, Fraser and colleagues conducted a phase 2a multicenter, double-blind, placebo-controlled trial randomly assigning 73 women (64 completed the study), aged 18 to 45 years, who met the Rotterdam criteria for PCOS (all women had hyperandrogenemia, ie, total testosterone > 1.7 nmol/L) to once-daily oral administration of either fezolinetant 60 mg, or 180 mg, or placebo (1:1:1) for 12 weeks (6). LH levels at 12 weeks were dramatically reduced from baseline levels by 8.2 IU/L after 60 mg and 10.2 IU/L after 180 mg of fezolinetant (vs by 3.2 IU/L after placebo). The interpretation of grouped LH levels is challenging because not all women were anovulatory and may have been at different points in their cycle at the fixed every-3 weekly assessment time points. Testosterone levels are more stable over the menstrual cycle and fell at 3 weeks from baseline levels by 19% with 60 mg, and by 32% with 180 mg of fezolinetant, and these reductions were maintained for the duration of the 12-week intervention period. Alterations in GnRH pulsatility are known to affect LH more than follicle-stimulating hormone (FSH), and fezolinetant (180 mg) induced only a small decrease in FSH levels by 1.5 IU/L. There were no changes in leptin or sex hormone–binding globulin at 12 weeks, but there was a trend toward a reduction in antimüllerian hormone (which has been posited to directly stimulate GnRH pulsatility).

As in previous studies in postmenopausal women, fezolinetant was well tolerated with potential treatment-related side effects including headache, paresthesia, and rash. Only one woman treated had an increase in alanine transaminase to more than 3 times the upper limit of normal, but this was transient and resolved spontaneously. This is reassuring, as although normal liver function was requisite for entry to this study, there is a high prevalence of metabolic-associated fatty liver disease among women with PCOS.

The study authors felt that the 12-week study duration was too short to assess an impact on clinical outcomes such as ovulation. That said, there was no change in cycle-length, ovulation, or progesterone levels. Likewise, the frequency of menses per 4-week period was not altered, and if anything trended downward. MLE4901 treatment was recently reported to not reverse PCOS features in a 5-dihydrotestosterone (DHT)-induced mouse model, although it did appear to have a beneficial effect on metabolic status by preventing the weight-gain associated with DHT treatment (7). Based on the data thus far, it seems probable that these agents can reduce symptoms associated with hyperandrogenism if studied for longer durations, given the robust decreases in testosterone levels observed, but larger, longer-term, dedicated studies are needed to assess their potential to treat anovulation.

Factors to be considered regarding the success of these agents for rectifying anovulation include that PCOS is a heterogeneous condition, and although GnRH pulsatility is regarded as a central feature of PCOS, only half of women having increased LH levels among women with PCOS diagnosed by current criteria, and similarly raised LH levels, do not form part of the diagnostic criteria for PCOS. Furthermore, obesity reduces LH pulse amplitude in women with PCOS, and thus it is notable that at least some women in this study had body mass indices up to 50. Consequently, it would seem logical to hypothesize that NK3R antagonists may be of greater benefit in select women with raised LH and androgen levels at baseline, and thus evaluating NK3R in specific PCOS subgroups may be relevant during further evaluation. Indeed, lowering LH levels excessively could also be detrimental, and therefore more detailed study of the degree of LH reduction, with consideration of baseline LH levels, and its impact on subsequent cycle characteristics would be of interest.

Furthermore, it is not clear that a reduction in GnRH pulsatility by NK3R antagonism would be desirable throughout the cycle. Whereas arcuate KNDy neurons are important for GnRH pulsatility, kisspeptin neurons in the preoptic area are responsible for the midcycle LH surge. Indeed, injection of senktide (an NK3R agonist) into that area can induce an LH surge. Thus, it seems feasible that NK3R antagonists may initially be of benefit by reducing LH levels, and thus ameliorating the relative FSH deficiency, but once follicular development is established, they may not need to be continued such that they do not also inhibit the midcycle LH surge. Accordingly, it is notable that fezolinetant was dosed on the day of the LH surge in 2 women in this study and caused LH levels to rapidly decline over the first few hours postdosing (5). Likewise, studies of fezolinetant in healthy women found that NK3R antagonism delayed or prevented ovulation, prolonged the cycle, and led to reduced luteal progesterone levels (5). Thus, consideration of individual endocrine profiles and protocols using shorter courses of NK3R antagonists until follicular development is established could be relevant for future studies.

Although NK3R antagonism predominantly exerts its effects via the hypothalamus, NKB agonism can directly stimulate granulosa cell aromatase expression and activity. However, NKB, NK3R, and kisspeptin expression were reduced in mural granulosa and cumulus cells of women with PCOS (8). Interestingly, although serum estradiol fell by 136 pmol/L at 180 mg of fezolinetant, this more likely reflects changes in gonadotropin levels than the direct inhibition of aromatase.

In summary, further larger, longer-term studies are highly anticipated to evaluate the potential of this novel approach to rectify the central neuroendocrine dysfunction underpinning the major hormonal abnormalities that characterize PCOS.