Decoding gonadotropin-releasing hormone (GnRH) pulse frequency

Within the body, cells communicate with one another using chemical signals such as hormones and neurotransmitters. These are often secreted in pulses and their effects are dependent upon pulse frequency so understanding how cells decode pulse frequency is fundamental to understanding how information is conveyed between (and within) cells. The brain's control of reproduction provides an excellent example and model for scientific exploration. Here, a neurohormone called GnRH (gonadotropin-releasing hormone) acts on cells in the pituitary gland to stimulate the synthesis and release of two other hormones (LH and FSH) that, in turn control the production of eggs and sex steroids in the gonads. A fundamental feature of this system is that GnRH secretion is pulsatile. Pulses of GnRH can be used to stimulate LH and FSH secretion and this is exploited in assisted reproduction. In contrast, sustained stimulation with GnRH ultimately reduces LH and FSH secretion. This, in turn reduces synthesis of sex steroids enabling treatment of hormone-dependent cancers (i.e. breast, ovary and prostate cancers). Thus, there is a "bell-shaped" frequency-response relationship (where sub-maximal GnRH pulse frequency elicits maximal responses) that underlies exploitation of the system, but remarkably little is known about the cellular, molecular or mathematical basis of this relationship. To explore this we have recently developed novel methods for monitoring effects of GnRH pulses on two intracellular biochemical pathways that mediate GnRH effects on gene expression (ERK and NFAT pathways). Using automated fluorescence microscopy to monitor these pathways in live cells we found that they are not GnRH frequency decoders (because they do not exhibit the negative feedback previously thought to underlie the bell-shaped frequency response relationship). However, we used this experimental data to develop and validate a sophisticated mathematical model for the mechanisms of GnRH action at the cellular level, and this model predicts that frequency decoding actually reflects the convergence of these pathways on the DNA elements that mediate GnRH effects on gene expression. Our unique wet-lab data and mathematical modelling has generated a novel theoretical frame-work that we believe represents a major breakthrough in understanding pulsatile GnRH signalling. In essence we are proposing that GnRH pulse frequency decoding is an emergent feature of the GnRH cell signalling network (rather than a characteristic of a single protein or pathway within the network) but we are still at a very early stage, as the mathematical model has not yet been tested experimentally. One of the most intriguing aspects of the modelling is the prediction that GnRH frequency-response relationships will be regulable rather than fixed (i.e. that the optimal pulse frequency for GnRH effects could differ before and after puberty, or could vary through the menstrual cycle) and this application aims to explore this possibility. Using the mathematical model for hypothesis generation, we now plan to define how some of the key model variables (such as GnRH receptor number and exposure to sex steroids) influence GnRH frequency-response relationships. We also plan to use the wet-lab data to refine the model, and to use a more formal mathematical approach for development and extension of the model. The direct importance of the planned work lies in the potential for greater understanding of GnRH signalling with physiologically relevant stimulation and for identifying novel targets for manipulation in the context in human and veterinary medicine as well as agriculture and aquaculture. The work is also likely to have widespread application because the structures and mechanisms considered are widespread in biological systems.

Cells often communicate using pulsatile signals. Here, information is conveyed in pulse frequency as well as amplitude but mechanisms underlying frequency decoding are poorly understood. We have been studying this using GnRH, a neuropeptide that mediates control of reproduction by the CNS, It is secreted in brief pulses to exert frequency-dependent effects on hormone synthesis and secretion in gonadotropes. We have used automated live cell imaging to monitor activation of ERK (ERK2-GFP reporter) and NFAT (NFAT1c-EFP reporter) during pulsatile stimulation and have developed a mathematical model for activation of these pathways. Our key finding is that the Raf/MEK/ERK and calmodulin/calineurin/NFAT modules are not stand-alone GnRH frequency decoders. Instead, our modelling predicts that frequency decoding could reflect co-operative convergence of these modules at gene promoters, generating genuine frequency decoding (at the transcription level) without the upstream feedback assumed to underlie it. This novel theoretical framework for understanding GnRH signalling (and biological frequency decoding in general) has not been tested, so our objective is to do so. Specifically, we hypothesise that that GnRH frequency-response relationships are regulable, and plan to test this by determining how they are influenced by GnRHR number, NFAT concentration, ERK concentration, ERK activation/inactivation kinetics, GnRH pulse duration and gonadal steroids. Here, the existing mathematical model has been used for hypothesis generation, but we will also use the wet-lab data to develop and refine the model. In addition a systematic parameter analysis is planned (to simplify the model before extending it to encompass steroidal modulation), and we also plan to incorporate stochasticity and diffusion into the model, hypothesising that co-operative convergence at the transcriptome enables diffusion-limited graded responses to drive stochastic transcriptional responses in individual cells.

Who will benefit from the research?
The immediate beneficiaries are reproductive endocrinologists and academic researchers working on cell signalling, frequency decoding and/or mathematical modelling of cell signalling pathways. Outside academia, the main beneficiaries will be the Pharma and Biotech industries, the general public and, in the long term, patients receiving gonadotrope-targeted treatments for assisted reproduction and/or hormone-dependent cancers. Our staff and students will also benefit from the planned research.

How will they benefit from the research?
Surprisingly little is known about mechanisms of GnRH action with physiological pulsatile stimulation or, indeed, how gonadotropes decode GnRH pulse frequency. Our wet-lab and mathematical modelling provide a novel frame-work for understanding the latter so the reproductive endocrine research community will benefit from the work we have planned to prove or disprove our hypotheses. Moreover, pulsatile GPCR activation and the downstream activation of transcription via NFAT and ERKs occur in many biological systems including, for example, the central nervous system and cardiovascular system. Testing, refining and simplifying our mathematical model will therefore also benefit a broad range of scientists. The fundamental prediction is that complex frequency decoding behaviours can be generated as emergent features of the system/network (even when the modules and elements forming the network show no such behaviour). The architecture and mechanisms considered are abundant in biological systems so the modelling will likely benefit scientists working on frequency decoding in many systems.

Interest in mechanisms of GnRH action is driven largely by the potential for therapeutic manipulation of post-receptor targets and our modelling highlights the possibility that GnRH frequency-response relationships will be regulable (i.e. that they will shift as cells adapt to their environment). This implies that changes in pulse frequency for GnRH secretion could be co-ordinated with changes in optimal pulse frequency for GnRH action. If so, manipulations that prevent or favour such co-ordination could be used to reduce or increase fertility. Although this is a speculative (and unlikely to be exploited in <10 years), it is the most obvious means by which the research could be applied for human or veterinary medicine. A related issue is that small molecule peptide antagonists for GnRHR are currently undergoing clinical trials. They typically increase cell surface GnRHR expression (by stimulating trafficking) so the information we provide on relationships between GnRHR number and optimal pulse frequency may well inform development of these novel compounds.

We will also engage with the public, by publication of our data in accessible formats (including University research theme web pages) and by involvement in outreach activities such as the "Science Alive" fair held in the Bristol city centre, the "Head Start" day for GCSE students and workshops (mathematical enrichment workshops for GCSE and AS students and cell biology workshops for FP2 medics) as well, as in invited talks at schools.

The RAs working on the projects will also benefit from the training provided (notably training in high content imaging, as this is rarely available in academic labs, but is highly sought after in the pharmaceutical industry). They will be also trained in extremely important for successful future career multidisciplinary skills such as ability to interact with experimentalist/modellers, understand their language and respect their priorities. Similarly, the applicants embed their research into student teaching (notably GnRH signalling lectures to Developmental Biology MSc students, High Content Imaging workshops for MRC DTA students, and Systems Biology lectures to Engineering Maths MEng students and ESPRC Complexity Science PhD students), so our students will also benefit from the work.