Abstract
Polycystic ovary syndrome (PCOS) is one of the most common causes of infertility in women. Approximately half of diagnosed individuals also experience metabolic syndrome. Metabolic syndrome is defined as a cluster of features that together increase the risk of cardiovascular disease and type 2 diabetes, common attributes being obesity, insulin resistance and the accompanying hyperinsulinemia. However, the relationship between PCOS and metabolic syndrome is very complex and not understood, partly due to the heterogeneity of both syndromes. Elevated circulating levels of androgens have been identified as a critical player in causing the dysregulation in PCOS, the brain being a crucial target site for this. Indeed, inducing a hyperandrogenic state in early development, either in utero and peri-pubertally, is able to induce PCOS-like phenotype in many different animal models.
As early as 1921, a hyperandrogenic state had been associated with impaired glucose regulation. Although a large amount of subsequent evidence has confirmed this relationship between androgens and metabolic function, the mechanisms by which androgens cause their dysregulation of both metabolic and reproductive function are unknown. Hyperinsulinemia has been proposed to contribute to the PCOS phenotype via insulin’s co-gonadotrophin actions on to the ovarian steroid hormone synthesis pathway. Its co-gonadotrophin actions also make it a candidate for being causative of the androgen excess observed in PCOS. Hyperinsulinemia, among other features, may be a result of hypothalamic metabolic hormone resistance, which has been reported to contribute to both metabolic and reproductive dysregulation.
In this thesis, I aim to investigate the interplay between metabolic and reproductive features of PCOS and to further the understanding of whether androgen excess in the female brain causes its PCOS-like phenotype through dysregulation in metabolically relevant pathways.
PCOS-like animal models, like the syndrome, are heterogeneous, with different androgen excess treatments during development leading to the development of different PCOS-like features, some of which are lean and some with a metabolic phenotype. Metabolic disruptions are seen in models induced by continuously administrating exogenous androgens; however, this makes it impossible to investigate the reversal of hyperandrogenemia. Therefore, in my PhD, I initially characterised multiple different PCOS-like models to determine which would be suitable for the subsequent three experiments. In the first of these experiments I aimed to investigate how hyperinsulinemia may be contributing to hyperandrogenism, specifically, whether this is through direct action on the ovary. I selected the well characterised prenatal androgenised (PNA) mouse model fed an obesogenic diet, which in the preliminary study showed both elevated circulating insulin and testosterone levels and was acyclic, making it the best model to assess the rescue of hyperandrogenemia in a hyperinsulinemic environment. Using cre-lox transgenics, I generated mice with insulin receptor deletion from androgen-producing cells (such as the theca cells) to assess whether this would reduce testosterone levels and, in turn, other PCOS-like characteristics. PNA and the obesogenic diet induced acyclicity and increased body weight, subcutaneous fat mass and fasting blood glucose levels, although PNA and the diet appeared to cause the reproductive and metabolic effects independently of each other. Insulin receptor deletion did not rescue these effects. Unfortunately, complications in reproducibility of the initial PNA + obesogenic diet phenotype (in particular elevated circulating testosterone levels) meant the role of hyperinsulinemia in inducing hyperandrogenism was not able to be tested in this thesis.
Next, I aimed to assess whether the chronic hyperandrogenic environment causes dysregulation of the central leptin and insulin signaling pathways, to determine if this contributes to the reproductive and metabolic attributes of PCOS. For this experiment, I used the peripubertal androgen excess PCOS-like model (PPA) that has well defined metabolic and reproductive dysfunction. I used the cre-lox transgenics to delete PTP1B and SOCS3, negative modulators of the leptin and insulin signaling pathways, from all forebrain neurons. Only PTP1B knockout was able to be fully validated at both the gene and protein levels. This deletion did not result in a rescue of the reproductive dysfunction but did reduce adiposity compared to control PPA mice. PPA mice did not have impaired hypothalamic leptin signaling, which is in contrast to what has been reported in PPA rats. This data suggests that peripubertal androgen excess-induced reproductive impairments are not mediated through perturbations in PTP1B signaling but may mediate its increased adiposity effects.
Finally, I aimed to determine whether androgen induced reproductive and metabolic dysregulation is mediated directly through the orexigenic AgRP-expressing neurons. These hypothalamic neurons are important for both reproductive and metabolic function, and have been shown to be upregulated in a ewe PCOS-like model. However, it is unknown how androgens regulate these neurons. Again using the PPA PCOS model and the cre-lox transgenic system, I specifically deleted androgen receptors from AgRP neurons. Only 20% of androgen receptor knockout mice showed rescue of PPA-induced acyclicity. No rescue effects of metabolic dysfunction was observed. These results suggest androgen actions directly on AgRP neurons are not crucial for the development of the PPA-induced acyclicity and anovulatory phenotype.
Overall, this work suggests peripubertal androgen excess does not induce its dysregulated reproductive phenotype through direct actions of androgens onto the AgRP neurons or through PTP1B and SOCS3 dysregulation in the forebrain neurons. An important point to note which my thesis highlights is the importance of validating models for their key PCOS-like features. It also highlights the necessity for better preclinical mouse models for understanding how metabolic dysregulation may either be causative or a contributing factor to the hyperandrogenemia observed in PCOS. This may be facilitated by improved phenotyping of clinical PCOS.