GLP-1 Receptor Agonists in PCOS: Insulin-Androgen Axis, Ovarian GLP-1R, and Cycle Restoration

[Image: CYP17A1 enzyme pathway in theca cells showing insulin-driven androgen synthesis]

The dominant androgenic mechanism in PCOS involves hyperinsulinemia-driven upregulation of CYP17A1 in theca cells. CYP17A1 encodes the bifunctional enzyme 17α-hydroxylase/17,20-lyase, which catalyzes the rate-limiting step in androgen synthesis: conversion of progesterone to 17α-hydroxyprogesterone and subsequently to androstenedione (the precursor to testosterone). Insulin and IGF-1 synergistically upregulate CYP17A1 transcription via PI3K/Akt and MAPK signaling, amplifying androgen output proportional to circulating insulin levels.

GLP-1RAs reduce fasting insulin and HOMA-IR through enhanced glucose-stimulated insulin secretion, reduced glucagon release, and — critically — direct peripheral insulin sensitization via GLP-1Rs in skeletal muscle and adipose tissue. As hyperinsulinemia resolves, CYP17A1 transcription in theca cells is normalized. The downstream effect is a reduction in circulating total testosterone (mean reduction 0.41 nmol/L in 16-week trials) and free androgen index.

Sex hormone-binding globulin (SHBG) increases by a mean of 38% with semaglutide treatment in PCOS cohorts, driven by reduced hepatic suppression of SHBG synthesis as insulin decreases. Because free testosterone — not total — determines androgenic activity at the androgen receptor, the SHBG increase amplifies the clinical benefit beyond what total testosterone reductions alone would predict. Free testosterone decreases by a proportionally greater magnitude than total testosterone in most published series.

[Image: Ovarian follicle cross-section showing GLP-1R expression in granulosa and theca cell layers]

Immunohistochemical and mRNA profiling studies have confirmed GLP-1R expression in both granulosa cells and theca cells of human ovarian follicles. This is mechanistically significant: it establishes a direct endocrine-to-gonadal signaling axis that operates independently of systemic insulin sensitization. GLP-1R is a Gs-coupled receptor; agonist binding activates adenylyl cyclase, increases intracellular cAMP, and activates PKA. In theca cells, cAMP/PKA signaling modulates StAR (steroidogenic acute regulatory protein) and CYP17A1 expression in a context-dependent manner — the elevated insulin milieu drives these in a pathological direction, while restored GLP-1R signaling under normalized insulin appears to counteract this.

In granulosa cells, GLP-1R activation enhances FSH receptor expression and sensitivity, supporting more complete folliculogenesis. FSH receptor sensitivity is impaired in PCOS granulosa cells partly due to the androgen-rich, hyperinsulinemic microenvironment. GLP-1 signaling appears to partially restore FSH responsiveness, improving dominant follicle selection and reducing the arrested small follicle accumulation that gives PCOS its characteristic polycystic appearance on ultrasound.

The clinical significance of direct ovarian GLP-1R signaling is supported by observations that ovulation rates improve in some PCOS women before substantial weight loss occurs. If the mechanism were purely weight-mediated, ovulation restoration would lag weight loss by the time required to normalize insulin — but the data shows earlier onset hormonal changes consistent with direct ovarian receptor engagement.

[Image: Arcuate nucleus KNDy neuron network showing GLP-1R modulation of GnRH pulse frequency]

The neuroendocrine disruption in PCOS centers on the KNDy (Kisspeptin/Neurokinin B/Dynorphin) pulse generator in the arcuate nucleus of the hypothalamus. KNDy neurons control GnRH pulse frequency through a reciprocal excitatory (kisspeptin, NK3R-mediated neurokinin B) and inhibitory (dynorphin via KOR) system. In PCOS, hyperinsulinemia and hyperandrogenism alter this balance toward excess GnRH pulse frequency, resulting in LH hypersecretion relative to FSH and the characteristic LH:FSH ratio elevation (typically >2:1) that disrupts normal folliculogenesis.

GLP-1Rs are expressed on arcuate nucleus kisspeptin neurons. GLP-1R activation in these neurons modulates the excitatory drive to the KNDy system, reducing aberrant GnRH pulse frequency when insulin is normalized. The mechanism is complementary to direct androgen withdrawal: as testosterone decreases, the AR-mediated negative modulation of dynorphin is partially restored, increasing inhibitory tone on the pulse generator. GLP-1R signaling provides an additional modulatory input that works in parallel.

Phase-specific implications are important clinically. During the follicular phase, normalized GnRH pulse frequency allows FSH to rise appropriately and drive dominant follicle selection — the step that fails in anovulatory PCOS. In the luteal phase, women with PCOS commonly have progesterone insufficiency because the corpus luteum forms inadequately from an anovulatory or sub-optimally ovulated cycle. As GLP-1RAs restore ovulatory cycles, luteal phase progesterone output normalizes, reducing the downstream effects of progesterone deficiency including endometrial instability and cycle irregularity.

[Image: Hormonal axis diagram showing follicular and luteal phase GLP-1 effects across the menstrual cycle]

Insulin resistance in PCOS creates phase-specific disruptions that GLP-1RAs address through temporally consistent insulin normalization. During the follicular phase, elevated insulin impairs FSH receptor signaling in granulosa cells via competitive PI3K pathway occupation — insulin and FSH share downstream Akt/FOXO1 signaling, and hyperinsulinemia effectively saturates this pathway, reducing the proportional FSH-specific transcriptional response. The clinical result is impaired dominant follicle selection and arrested small follicle development.

GLP-1RA treatment normalizes follicular phase insulin, restoring FSH receptor sensitivity. FOXO3a — a transcription factor critical for primordial follicle dormancy and activation — whose localization is disrupted by hyperinsulinemia, shows more appropriate nuclear compartmentalization with insulin sensitization. This may partly explain why GLP-1RAs appear to improve not just ovulation rates but reported egg quality metrics in IVF populations, though this data is more preliminary.

Luteal phase progesterone deficiency in anovulatory or oligo-ovulatory PCOS creates a hormonal environment characterized by relative estrogen dominance and progesterone insufficiency. As GLP-1RAs restore ovulatory cycles, corpus luteum formation improves and mid-luteal progesterone levels normalize — typically to >10 ng/mL, the threshold for adequate luteal phase support. This restoration has downstream effects on endometrial receptivity, cycle symptom profile, and premenstrual symptom burden, since many PCOS-related premenstrual symptoms are driven by the relative progesterone-to-estrogen imbalance.