Abstract
The gonadotrophin-releasing hormone (GnRH) neurons represent the key output cells of the neuronal network controlling the hypothalamic pituitary-gonadal axis and thus fertility in all mammalian species. The patterns of neuronal activity by which GnRH neurons generate pulses of GnRH and thus luteinizing hormone (LH) are unknown. To date, the scattered distribution of the GnRH cell bodies remain the main limitation to investigating the cellular events that lead to pulsatile secretion of LH. As such, the numbers of GnRH neurons involved in a pulse, their location, and their patterns of electrical firing have never been determined. The ability to remotely manipulate the neuronal activity of GnRH neurons in a selective manner in vivo would facilitate our understanding of the patterns of GnRH neuron activity required to generate pulses of LH.
Channelrhodopsins (ChR2) are now well established as genetically-encodable cation channels that can be activated by blue light to excite neurons. I have used a Cre-dependent adeno- associated virus (AAV) in transgenic GnRH-Cre mice to target ChR2 to the GnRH neuronal phenotype. Taking advantage of the retrograde transport properties of AAVs, I injected the viral vector into the median eminence to target expression of ChR2 to only hypophysiotropic GnRH neurons. Immunofluorescence studies showed that 93% of all ChR2-expressing cells in the brain expressed GnRH.
I characterised the ability of ChR2 to modulate GnRH neuron excitability in vitro in response to blue light. I prepared acute brain slices from AAV-injected GnRH-Cre mice and undertook cell-attached recordings of ChR2-expressing GnRH neurons. The analysis of spontaneous firing patterns, as well as the increase in firing activity after kisspeptin application, indicated that ChR2-transfected GnRH neurons were not compromised by the presence of ChR2 channels in the membrane. ChR2-expressing GnRH neurons can be driven to fire with high spike fidelity with blue light stimulation frequencies up to 40 Hz for short second intervals and up to 10 Hz for minute periods.
Next, I determined the profile of pulsatile secretion of LH in the ovariectomized conscious mouse using a fast blood sampling collection. I aimed to replicate these LH pulses in vivo, in AAV injected ovariectomized GnRH-Cre female mice, by placing a fibre optic in the rostral preoptic area and stimulating ChR2-expressing hypophysiotropic GnRH neurons with different patterns of light. Optogenetic activation of GnRH neurons for 30 s to 5 min time periods using a range of different frequencies revealed that 10 Hz stimulation for 2 min was the minimum required to generate a pulse-like increment of LH release. The same result was found for optical activation of GnRH projections in the median eminence. Under these conditions, the dynamics of optogenetically-evoked LH secretion are very similar to that of endogenous LH pulses suggesting that the minimal parameters of GnRH neuron activation reported here are likely to be close to that occurring in vivo for pulsatile LH secretion.
Synchronised burst firing has been suggested to underlie GnRH pulsatility. As such, rPOA GnRH neurons were synchronously activated for 5 min to fire in bursts using the burst patterning previously detected for GnRH neurons in vivo. Surprisingly, this was found to have no effect on LH secretion.
In summary, I have generated a mouse model in which hypophysiotropic GnRH neurons can be selectively activated with high temporal and spatial precision in vivo. Using of an optogenetics approach combined with a fast-blood sampling technique, I have been able to define minimal parameters of activation required to generate pulsatile gonadotropin secretion in the blood. This first insight into how GnRH neurons generate a pulse of LH in vivo provides critical information for understanding and manipulating the genesis of gonadotropin pulsatility in reproductive biology.