Perturbation of the gonadotropin-releasing hormone neuronal network in polycystic ovarian syndrome

Polycystic ovarian syndrome (PCOS) is the most common cause of infertility among women of reproductive age worldwide. While classically identified as an ovarian disorder, PCOS is associated with increased luteinising hormone (LH) pulse frequency, implicating changes in the regulation of gonadotropin-releasing hormone (GnRH) neurons, the final output cells in a large neuronal network controlling fertility. To function appropriately, GnRH neurons require feedback information from circulating steroid hormones produced by the gonads. The afferent neuronal network to GnRH neurons is thought to be responsible for relaying this critical feedback information, as GnRH neurons themselves do not express the required steroid hormone receptors. Elevated GnRH/LH secretion in women with PCOS is less responsive to gonadal steroid hormones, demonstrating that steroid hormone feedback is impaired in PCOS. This thesis utilised a mouse model of PCOS, generated by prenatal androgen (PNA) exposure, to investigate the hypothesis that impaired steroid hormone negative feedback results from alterations within the steroid hormone-sensitive afferent neuronal network. Serial blood sampling confirmed that PNA mice mirror the clinical neuroendocrine phenotype of impaired gonadal steroid feedback. PNA mice exhibited a blunted post-castration rise in luteinising hormone compared to vehicle-treated controls (LH, p<0.05), and a lack of LH suppression following estrogen or progesterone treatment. This provides strong support for impaired negative feedback regulation of gonadotropin secretion. In contrast, the ability to positively drive the preovulatory GnRH/LH surge downstream of estrogen signalling was intact in PNA animals, as detected by a significant increase in LH levels (p<0.05) and activated (cFos positive) GnRH neurons (p<0.001). To identify whether alterations in steroid hormone receptors may underlie impaired steroid hormone feedback and to locate potential brain regions involved, the receptors for estrogen and progesterone were mapped throughout the hypothalamus of control and PNA animals. Estrogen receptor alpha (ERα) expression was largely unchanged, with the exception of an unexpected increase in the number of labelled cells in the periventricular nucleus (PeN) of PNA mice (p<0.001). In contrast, PNA mice exhibited a dramatic decrease in the number of cells expressing the progesterone receptor (PR) in the anteroventral periventricular nucleus (AVPV, p<0.05), PeN (p<0.05) and the arcuate nucleus (ARN, p<0.001). Reduced progesterone sensitivity was most dramatic within the ARN, with a nearly 60% decrease in PR-positive cells. I also investigated whether elevated GnRH/LH pulse frequency correlates with an increase in excitatory synaptic input to GnRH neurons. Confocal imaging of GnRH neurons in GnRH-green fluorescent protein (GFP) transgenic mice revealed that spine density, an indicator of excitatory synaptic input in many neuronal phenotypes, was significantly greater in PNA mice compared with controls (p<0.01). Surprisingly, elevated GnRH neuron spine density did not correlate with putative glutamatergic input, which was identified by vesicular glutamate transporter 2 puncta closely apposing GnRH neurons. However, spine density in PNA mice did correlate with an increase in GABAergic input to GnRH neurons, as measured by apposition with vesicular GABA transporter puncta (vGaT, p<0.01). As GABA can excite adult GnRH neurons, these data suggest that elevated GABAergic innervation to GnRH neurons may mediate the increased GnRH/LH pulsatile release observed in the PCOS phenotype. To address whether increased GABA innervation of GnRH neurons originates in the ARN, a Cre-dependent adenoviral vector expressing farnesylated enhanced green fluorescent protein (EGFPf) was injected into the ARN of mice that express Cre specifically in GABAergic neurons (vGaT-Cre). Immunofluorescent labelling revealed that, remarkably, EGFPf-labelled ARN GABAergic projections heavily contacted and even bundled with GnRH neurons, and the density of fibres apposing the GnRH neuron dendrite was even greater in PNA mice (56%, p<0.05). Additionally, ARN GABA neurons in PNA animals exhibited significantly fewer neurons colocalised with PR compared to controls (p