Computational prediction of vulnerable points and interventions for dysfunctional synaptic plasticity in neuropsychiatric disorders
Lead Research Organisation:
University of Ulster
Department Name: Sch of Computing & Intelligent Systems
Abstract
Neuropsychiatric disorders such as Autism Spectrum Disorder (ASD) and Schizophrenia are widespread, with around 1.7% of children in the United States diagnosed with ASD, and around 0.7% of people being diagnosed with Schizophrenia at some point in life. Mental health problems in general account for more than 20% of disabilities in the UK and are estimated to cost the economy between £70-100 billion per year. Current behavioural and pharmaceutical treatments for neuropsychiatric disorders are effective for only a subset of patients, often carry unwanted side-effects, and treatment success is difficult to predict from patient to patient. These shortcomings reflect the fact that almost all existing drug treatments were discovered by chance, rather than being designed based on an understanding of disorder mechanisms.
However, a recent wave of discoveries of 70-100 genetic mutations linked to each of ASD and Schizophrenia has given promising clues to the origins of these disorders. Many of the genes code implicated are important for synapses - the connections between neurons that mediate learning and memory in the brain. This implies that many neuropsychiatric disorders may in fact be disorders of synaptic plasticity.
Academic and corporate laboratory researchers worldwide are now trying to figure out what brain changes the discovered genetic mutations cause, typically by studying genetically altered mice that should ideally mimic the human patients. However, most neuroscience research methods are painstakingly difficult and low-throughput, so progress is slow. In this NIRG we will instead use data-driven computational simulations of neurons and synapses, because they work much faster, and let us perform detailed virtual experiments that researchers might like to do in the lab, but can't. The computer simulations will be based on data from the lab of our experimental collaborators within the University of Bristol. This project will shortlist a number of highly vulnerable components of synapses, that can be used to direct future wet lab experiments. Finally, I will use the results to develop a new theory of dysfunctional information transmission at synapses in neuropsychiatric disorders, that could guide broader research in the field.
However, a recent wave of discoveries of 70-100 genetic mutations linked to each of ASD and Schizophrenia has given promising clues to the origins of these disorders. Many of the genes code implicated are important for synapses - the connections between neurons that mediate learning and memory in the brain. This implies that many neuropsychiatric disorders may in fact be disorders of synaptic plasticity.
Academic and corporate laboratory researchers worldwide are now trying to figure out what brain changes the discovered genetic mutations cause, typically by studying genetically altered mice that should ideally mimic the human patients. However, most neuroscience research methods are painstakingly difficult and low-throughput, so progress is slow. In this NIRG we will instead use data-driven computational simulations of neurons and synapses, because they work much faster, and let us perform detailed virtual experiments that researchers might like to do in the lab, but can't. The computer simulations will be based on data from the lab of our experimental collaborators within the University of Bristol. This project will shortlist a number of highly vulnerable components of synapses, that can be used to direct future wet lab experiments. Finally, I will use the results to develop a new theory of dysfunctional information transmission at synapses in neuropsychiatric disorders, that could guide broader research in the field.
Technical Summary
Neuroscience research groups worldwide are currently studying a new generation of animal models of neuropsychiatric disorders, based on genetic mutations discovered in human patients. The hope is that these animal models will recapitulate the primary biochemical, cellular, and neural circuit changes, and if we discover the key phenotypes in the animal models, we can design better interventions to rescue cognitive symptoms in patients. However, most standard neuroscience techniques are notoriously low-throughput, and it is not obvious where to look in a given animal models for disease phenotypes. These challenges have led to a neuropsychiatric drug development crisis. This project will aim to dramatically speed up this progress by developing a detailed computational model of biochemical signalling during synaptic plasticity induction, that can be used to identify vulnerable points in the cascade. The computational model will be based on abundant data in the literature, and fit to reproduce recent experimental data from rodent in vitro studies from the laboratory of our collaborators at Bristol.
The project will deliver three important outputs: first, we will use the computational model to identify a shortlist of synaptic components for which the rules of synaptic plasticity are most sensitive. This list will give predictions to guide experiments on disorder animal models. Second, the computational model itself will be made publicly available for researchers around the world to use and generate predictions for their animal model of interest. Third, based on the simulation results I will develop a new theory for why some synaptic components are more important than others in synaptic plasticity induction, based on the principle of faulty information transmission along biochemical signalling cascades at synapses. This general theory has the potential to guide research on dysfunctional synaptic plasticity in psychiatric disorders for decades to come.
The project will deliver three important outputs: first, we will use the computational model to identify a shortlist of synaptic components for which the rules of synaptic plasticity are most sensitive. This list will give predictions to guide experiments on disorder animal models. Second, the computational model itself will be made publicly available for researchers around the world to use and generate predictions for their animal model of interest. Third, based on the simulation results I will develop a new theory for why some synaptic components are more important than others in synaptic plasticity induction, based on the principle of faulty information transmission along biochemical signalling cascades at synapses. This general theory has the potential to guide research on dysfunctional synaptic plasticity in psychiatric disorders for decades to come.
Planned Impact
Pharmaceutical companies.
Pharmaceutical companies developing treatments for neuropsychiatric disorders will benefit from this research in similar ways to the academic research community: for example, they can use the vulnerable synaptic component shortlist to guide animal model research, and can use the model itself to test effects of genetic perturbation of interest. The added benefit is that this group of beneficiaries could use the model itself as a large-scale screening tool to test the likely effects of candidate compounds on synaptic plasticity induction. A more sophisticated approach might even be to take the computer simulation of a disease system, and work backwards from the model's phenotype to design an ideal intervention, giving a target for research chemists that develop new compounds.
Research charities.
Neuropsychiatric disorder research charities could use the outputs of this research to help allocate their often limited funding towards projects that may have a higher translational likelihood, by using the shortlist of vulnerable synaptic components as a guide.
Society at large.
This project is directly designed to speed up drug development for neuropsychiatric disorders, so is of inherent value to the public. However given the ubiquity of these disorders across society, the general public has a thirst for knowledge on the latest scientific research in this area - especially on the likelihood of new treatments. The new theory we will develop in the course of the project offers an opportunity for science communication. Although technical in the details, the basic idea of "good signalling gone bad" at neural connections is a plausible story to promote to the lay person, and somewhat more digestible than a list of risk gene names.
Reduction in animal use.
This project will help reduce the number of animals being used in neuropsychiatric disorder research and drug development, in three ways: 1) we will directly list which components of synapses are the most likely places to look for phenotypes in animal models, so targeting the trial-and-error search currently employed by most lab researchers; 2) the computational model can be used to screen the efficacy of hypothetical candidate interventions, potentially removing the need for animal testing in ineffective cases, 3) the general theory proposed in the project, if proved useful, could in the long run alter the course of neuropsychiatric disorder research and potentially save years of animal experimental work.
PDRA training.
I will work closely with the hired PDRA throughout the 3-year project. Whether or not the PDRA wishes to pursue a career in academic research following the completion of the project, the statistical, computational modelling, and neurobiological knowledge imparted by the PI will help boost the PDRA's skills, and so contribute to the UK's knowledge and/or health economies.
Pharmaceutical companies developing treatments for neuropsychiatric disorders will benefit from this research in similar ways to the academic research community: for example, they can use the vulnerable synaptic component shortlist to guide animal model research, and can use the model itself to test effects of genetic perturbation of interest. The added benefit is that this group of beneficiaries could use the model itself as a large-scale screening tool to test the likely effects of candidate compounds on synaptic plasticity induction. A more sophisticated approach might even be to take the computer simulation of a disease system, and work backwards from the model's phenotype to design an ideal intervention, giving a target for research chemists that develop new compounds.
Research charities.
Neuropsychiatric disorder research charities could use the outputs of this research to help allocate their often limited funding towards projects that may have a higher translational likelihood, by using the shortlist of vulnerable synaptic components as a guide.
Society at large.
This project is directly designed to speed up drug development for neuropsychiatric disorders, so is of inherent value to the public. However given the ubiquity of these disorders across society, the general public has a thirst for knowledge on the latest scientific research in this area - especially on the likelihood of new treatments. The new theory we will develop in the course of the project offers an opportunity for science communication. Although technical in the details, the basic idea of "good signalling gone bad" at neural connections is a plausible story to promote to the lay person, and somewhat more digestible than a list of risk gene names.
Reduction in animal use.
This project will help reduce the number of animals being used in neuropsychiatric disorder research and drug development, in three ways: 1) we will directly list which components of synapses are the most likely places to look for phenotypes in animal models, so targeting the trial-and-error search currently employed by most lab researchers; 2) the computational model can be used to screen the efficacy of hypothetical candidate interventions, potentially removing the need for animal testing in ineffective cases, 3) the general theory proposed in the project, if proved useful, could in the long run alter the course of neuropsychiatric disorder research and potentially save years of animal experimental work.
PDRA training.
I will work closely with the hired PDRA throughout the 3-year project. Whether or not the PDRA wishes to pursue a career in academic research following the completion of the project, the statistical, computational modelling, and neurobiological knowledge imparted by the PI will help boost the PDRA's skills, and so contribute to the UK's knowledge and/or health economies.
