Abstract
Arginine-phenylalanine related-peptide (RFRP) neurons are a small population of neurons found in and around the dorsomedial nucleus of the hypothalamus (DMH) in mammals. RFRP neurons have been implicated in the neuroendocrine control of reproduction through their anatomical connections with key neurons involved in the hypothalamo-pituitary gonadal (HPG) axis, gonadotropin-releasing hormone (GnRH) neurons and both anteroventral periventricular nucleus (AVPV) and arcuate nucleus kisspeptin neurons. In support of this, GnRH neurons are known to be inhibited by the peptide arginine-phenylalanine related-peptide 3 (RFRP-3) which is the primary peptide secreted by RFRP neurons. A plethora of functional data has investigated the role of RFRP peptide signalling in fertility, with complex findings both between and within species, sexes and model types. More recently, the involvement of RFRP neurons in neuroendocrine stress responses has also been investigated. RFRP neurons are known to be both stress receptive and stress reactive. A variety of physiological and psychological stressors have been found to activate RFRP neurons and the production of Rfrp mRNA, and RFRP neurons are known to express both glucocorticoid receptors and receptors for corticotrophin releasing hormone (CRH). Additionally, central delivery of RFRP peptides has been shown to activate hypothalamo-pituitary adrenal (HPA) axis hormonal cascades. To date, the functions of RFRP neurons have primarily been inferred through administration of RFRP-3.
In this thesis, I have explored the role of RFRP neurons with the overall hypothesis that RFRP neurons act as a relay or feed-forward mechanism for stressors – propagating stressor information through communication with several brain regions to maintain allostasis following stressor exposure. I have investigated specifically this role in the HPA axis, the behavioural stress system and the HPG axis, to determine whether there is anatomical, physiological and behavioural evidence to support this role in the propagation of stressor information. Rather than using central injection of exogenous RFRP peptides as the experimental approach as has typically been done to date, I have focused primarily on the function of RFRP neurons a population, by utilising chemogenetic techniques to activate them. Using designer receptors exclusively activated by designer drugs (DREADDs) to non-invasively activate RFRP neurons through the peripheral injection of the designer drug clozapine-N-oxide (CNO) allowed me to do this, and two different DREADD expression models were compared and utilised to target RFRP neurons.
I found that chemogenetic activation of the RFRP neuronal population stimulated corticosterone secretion and that 15% of PVN CRH neurons, the master output neurons of the HPA axis, express the receptor for RFRP peptides, GPR147. Additionally, I found that GPR147 is not expressed by CRH neurons of the extended amygdala, which are known to be involved in anxiety-like behaviours in rodents. Following up this finding, I used patch-clamp brain slice electrophysiology to record responses of PVN CRH neurons to bath application of RFRP-3. I did not find any effects of RFRP-3 on this population, but as only 15% of the population respond to the peptide it is conceivable that any potential effects were not uncovered, or that another signalling molecule produced by RFRP neurons may be responsible for activating the HPA axis.
I then determined the effect of activating RFRP neurons on anxiety-like and depression-like behaviours using a battery of behavioural tests. I found that in male (but not female) mice, acute activation of RFRP neurons resulted in significantly increased anxiety-like and depression-like behaviours in two different DREADD models, respectively. This indicates that RFRP neurons in male mice can function to produce protective behaviours like freezing, escape, and avoidance when acutely activated, and they therefore have the potential to cause maladaptive changes in these circuits if chronically activated, which could result in anxiety and depression in humans. I investigated the role of RFRP peptide signalling in behavioural stress responses by using mice with GPR147 receptors specifically knocked out of CRH neurons. Following exposure to a five minute swιm stressor, the performance of these knock-out mice was assessed in anxiety-like behavioural tests. I did not see any alterations in behavioural stress responses in these mice, indicating that PVN CRH neurons are not likely to facilitate any behavioural changes seen by directly modulating GPR147 signalling.
Finally, I examined role of RFRP neurons in the activity of the HPG axis. I found that activation of RFRP neurons reduced the luteinising hormone (LH) pulse frequency of female mice by 50%, but that male LH pulse dynamics were not affected. This indicates that RFRP neurons in female mice can act to reduce reproductive function when acutely active, but have the potential to negatively impact fertility if chronically active. I did not find an effect on the pre-ovulatory LH surge following activation of RFRP neurons, perhaps due to inappropriate timing of CNO delivery relative to the surge. I also aimed to characterise the activity of RFRP neurons at the onset of the oestradiol-induced LH surge following acute restraint stress administered at discrete time points throughout the day to profoundly suppress LH surge secretion. I determined that acute restraint stress administered two to zero hours prior to the expected onset of the LH surge blocked the surge, elevated circulating corticosterone concentrations four-fold and reduced periventricular kisspeptin neuronal activity by 33%. Additionally, acute restraint stress administered from six to four hours before the expected LH surge was sufficient to abolish the surge, irrespective of low circulating corticosterone levels or relatively normal levels of AVPV kisspeptin neuron activity at the time of collection. I did not find an effect of stressor exposure on Rfrp mRNA production. Unfortunately I was not able to definitively determine the level of Rfrp neuronal activity.
Taken together, these data support the hypothesis that RFRP neurons function allostatically to propagate acute stressor information to the HPA axis, behavioural stress system, and HPG axis. Under normal conditions these changes occur to conserve energy and maintain safety when faced with psychological or potentially physical threats. However, pathophysiological alterations to the RFRP system, caused by allostatic load, have the real potential to negatively impact several aspects of human health. This thesis has further elucidated the role of RFRP neurons in mice; in stimulating HPA axis responses and producing acute anxiety-like and depression-like behaviours, as well as supporting a role for RFRP neurons in stress induced reproductive inhibition, paving the way for future research to unlock the therapeutic potential of the RFRP system.