The Psychiatric Genomics Consortium (PGC; https://pgc.www.med.unc.edu/pgc) is delivering an increasing flow of discoveries about the fundamental basis of psychiatric disorders. Moving from discovery of GWAS loci towards delivering new therapeutic approaches is a challenge which requires a new wave of multidisciplinary work and close liaison between academia and industry. The PGC has recently initiated a new research program to deliver “actionable” findings that (a) reveal the fundamental biology of psychiatric disorders, (b) inform clinical practice, and (c) deliver new therapeutic targets. Working with industry will be essential, but there is a lack of a forum to exchange ideas with academia. With the aim of filling this gap, the first PGC Pathways to Drugs industry-academia workshop was held at the Institute of Psychiatry, Psychology and Neuroscience, King’s College London on March 2–3, 2017. Here, we review the main advances happening in the field (see abstracts(1)).
The PGC works on schizophrenia, bipolar disorder, major depressive disorder (MDD), autism spectrum disorder (ASD), eating disorders, post-traumatic stress disorder (PTSD), attention deficit hyperactivity disorder (ADHD), Alzheimer’s disease, obsessive compulsive disorder (OCD), Tourette’s syndrome, and substance use disorders. The PGC has identified over 300 genetic associations in 7 disorders and the aim of the next 5 years is to increase sample sizes to >100,000 cases for each psychiatric disorder and find new treatments(1). There is considerable potential for the knowledge derived from GWAS to revitalize industry drug discovery efforts(2). In addition, the complex nature of psychiatric disorders means that the study of biological pathways is also of paramount importance(3).
Strong examples of translation are emerging, exemplified by the new anorexia GWAS results, which are inspiring a reconceptualization of anorexia nervosa as both a metabolic and psychiatric disorder(4). Major contributions to the PGC from the iPSYCH initiative, and others, are advancing the genetics of ADHD and ASD(1). Likewise, UK Biobank(5) (www.ukbiobank.ac.uk) is an extensive resource of ~500,000 genotyped individuals. UK Biobank mental health phenotyping on 157,000 is available and will reach >300,000, meaning that UK Biobank will become the largest cohort with mental health questionnaire data worldwide(1). Multiple psychiatric disorders are achieving successful GWAS - leading to the hope that new therapeutic leads can be identified. Finally, the PGC is developing a pipeline for the identification of drug candidates using GWAS results(6). However, these hypotheses are only a first step in the drug discovery process and need to be followed up by functional and pharmacological studies, in collaboration with industry, and validated by clinical trials.
Model systems and systems biology: generating mechanistic hypotheses
Determining mechanistic hypotheses arising from a GWAS association is a general problem across medicine. Some regions occasionally implicate a single gene with clarity; e.g., genetic evidence for MDD has highlighted LRFN5, TCF4, PTPN1, and NEGR1. Other loci are intergenic and far from any known genes; still other loci can contain many genes or lie within regulatory regions. Connecting GWAS findings to salient genes and mapping regulatory regions is crucial to identifying targets.
Discussions at the conference focused on the integration of three sets of methods. Firstly, statistical methods to evaluate gene sets, which have been used to demonstrate the salience of MDD and schizophrenia GWAS findings to the targets of antidepressants and antipsychotics(6, 7). Secondly, bioinformatics approaches to leverage functional genomic data (gene expression(2), DNA-DNA looping(8), and epigenomics(9)). Data from the NIMH psychENCODE consortium on brain samples from people with severe psychiatric disorders will further enable this intention(9). Finally, large scale genetic editing may be needed(10). Together, the above steps can be used to generate hypotheses, narrow an interval to include regulatory regions, or to identify key elements.
From mechanism to drug: genetics and chemoinformatics
After identifying potentially relevant genes, a key challenge is to determine which of the corresponding proteins are “druggable”, to find drugs that could bind to these proteins, regulate their production, prevent protein-protein interactions, or in any way have an influence on their biological mechanism. These steps are enabled by chemoinformatics and bioinformatics approaches encompassing machine learning, molecular modelling, and data mining. There is an increasing quality of data available in databases collecting drug/target affinities(3) (ChEMBL, PHAROS, Ki DB, etc.). Open Targets (https://www.targetvalidation.org/) is a relatively new free platform that collects target/phenotype associations from various sources (including UniProt, GWAS Catalog, European Variation Archive, Gene2Phenotype). These public databases, many partly funded by industry, are complemented by new open-source drug laboratory resources that can be used to allow screening of the entire druggable GPCRome (see abstracts(1)).
Polygenic risk scores in clinical trials
Polygenic risk scores (PRS) summarize an individual’s genetic loading for a disorder in a single measure. Their power is increasing, and it now possible to calculate PRS for common side effect profiles, such as hypertension and type 2 diabetes. In designing clinical trials, use of PRS could allow the recruitment of individuals at high genetic risk, or for stratifying response. Post hoc analysis of polygenic risk scores in completed clinical trials may improve understanding of drug response profiles and encourage the pharmaceutical industry to move genetics earlier within the trial pipeline, to ultimately develop companion diagnostics with PRS. Using the power of clinical trial samples will require interaction between the research community and industry to test the utility of PRS for improved prescribing and clinical trials.
Conclusion
Our discussions pointed out the need to integrate multiple data sources to discover new treatments for psychiatric disorders: gene expression in different tissues, drug/target affinities, GWAS results, bioinformatics, chemoinformatics and pharmacogenomics. New drug development could be further aided by the adoption of precision medicine for psychiatric clinical trials, as exemplified by polygenic scores and “precision psychiatry”. However, ultimately, new drug discovery in psychiatry will be best enabled if academia and industry join forces and if open data and sample sharing initiatives are encouraged. Joint industry-academia conferences will be essential to achieve these goals, identify main research objectives and find a shorter route for drug discovery pipelines.
Acknowledgments and Financial disclosures
This work is funded in part by the U.K. National Institute for Health Research (NIHR) Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health. The authors acknowledge funding from Lundbeck and Illumina.
References
- 1.Abstracts: PGC meeting: Pathways to Drugs. Retrieved January 4, 2018, from https://sites.google.com/view/pgc-pharma-meeting/abstracts. [Google Scholar]
- 2.Fromer M, Roussos P, Sieberts SK, Johnson JS, Kavanagh DH, Perumal T, et al. (2016): Gene expression elucidates functional impact of polygenic risk for schizophrenia. Nat Neurosci. 19: 1442–1453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Breen G, Li Q, Roth BL, O’Donnell P, Didriksen M, Dolmetsch R, et al. (2016): Translating genome-wide association findings into new therapeutics for psychiatry. Nat Neurosci. 19: 1392–1396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Duncan L, Yilmaz Z, Gaspar H, Walters R, Goldstein J, Anttila V, et al. (2017): Significant Locus and Metabolic Genetic Correlations Revealed in Genome-Wide Association Study of Anorexia Nervosa. Am J Psychiatry. appiajp201716121402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. (2015): UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 12: e1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gaspar HA, Breen G (2016, December 2): Pathways analyses of schizophrenia GWAS focusing on known and novel drug targets. bioRxiv. [Google Scholar]
- 7.Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium (2017): Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depressive disorder (Submitted). . [Google Scholar]
- 8.Won H, de la Torre-Ubieta L, Stein JL, Parikshak NN, Huang J, Opland CK, et al. (2016): Chromosome conformation elucidates regulatory relationships in developing human brain. Nature. . doi: 10.1038/nature19847. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.PsychENCODE Consortium, Akbarian S, Liu C, Knowles JA, Vaccarino FM, Farnham PJ, et al. (2015): The PsychENCODE project. Nat Neurosci. 18: 1707–1712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cai L, Fisher AL, Huang H, Xie Z (2016): CRISPR-mediated genome editing and human diseases. Genes & Diseases. 3: 244–251. [DOI] [PMC free article] [PubMed] [Google Scholar]