Alexander S. Hatoum¹, Alex Miller², Aaron J. Gorelik³, David A.A. Baranger¹, Emma C. Johnson¹, Arpana Agrawal¹, Ryan Bogdan³

Background: Alcohol use disorder (AUD) presents with substantial heterogeneity.

Rational: Attempts to distinguish SUDs such as AUD by clinical features alone have had limited translational success.

Hypothesis: Objective biomarkers, like those derived from neuroimaging, will help parse AUD heterogeneity based on biology.

Methods. We developed neuroimaging derived cortical thickness biotypes of AUD in the European subset of the UK biobank (N=22,321) and conducted GWAS of biotype membership vs. controls (i.e., AUD biotype 1 vs. controls from biotype 2 and non-affected individuals). We also conducted GWAS contrasting biotypes in all individuals with neuroimaging data (N = 32,287). Results: Differences in regional cortical thickness within the salience network (e.g., regions) differentiated two AUD biotypes from controls. AUD biotype 1 was characterized by greater cortical thickness and biotype 2 had less cortical thickness, within the salience network. GWAS of biotype 1 (vs controls) identified one suggestive hit (rs60657889, p = 5.93e-08). GWAS contrasting the two biotypes revealed two significant hits including lead variant rs12355217 on FAM107B, (P=1.77e-09), a gene known to influence neurological pathways.

Discussion. Using a data-forward approach, we can parse heterogeneity in AUD and discover potential genetic pathways contributing to heterogeneity. Future analyses will examine loci from a joint analysis of biotypes with the largest GWAS of problematic alcohol use using a multivariate GWAS (N=1,079,947) and conduct multi-ancestry fine-mapping to determine likely causal variants. Finally, genetic correlations and drug repurposing analyses will determine the clinical and pharmacotherapeutic separability of brain-imaging derived biotypes.

1. Washington University School of Medicine, Department of Psychiatry, 2. Indiana University School of Medicine, Department of Psychiatry, 3. Washington University in St. Louis, Department of Psychological & Brain Sciences. Funding: NIAAA K01030083

William B. Lynch1,2,3, Jacob A. Beierle1,3,4, Bryan McKiver5, Arthur J. Vanvalkenburg6, Jared Mann5, Emily J. Yao1, Milad Mortazavi7, Yu-Yu Ren7, Binh-Minh Nguyen1, Kayla T. Richardson1,8, Abraham A. Palmer7, W. Evan Johnson6, M. Imad Damaj5, Camron D. Bryant1,3,4

Chemotherapy-induced peripheral neuropathy (CIPN) is a debilitating side effect of cancer treatment with an unclear genetic etiology. Paclitaxel (PAC) is an anti-mitotic agent that causes CIPN in 60-70% of patients. Quantitative genetics in near-isogenic mouse substrains [C57BL/6J (J) and C57BL/6NCrl (N)] can facilitate gene, biological, and therapeutic discovery. J, but not N, showed robust PAC-induced mechanical hypersensitivity (von Frey) and cold hypersensitivity (acetone) that emerged, peaked, and recovered at 7, 14, and 28 days post-first PAC treatment (4X 2 mg/kg, i.p.). We identified a genome-wide significant quantitative trait locus (QTL) for mechanical hypersensitivity on chromosome 1 (LOD=7.7; p=0.001; 40-78 Mb; 11% variance explained) and a suggestive QTL for cold hypersensitivity on chromosome 14 (LOD=4.4; p=0.095; 100-124 Mb; 8% variance explained). Polymorphic, positional, differentially expressed genes (effect of Substrain, Substrain x Treatment) across DRG, PAG, and spinal cord (SC) at 7, 14, and 28 days post-first PAC revealed Cryba2, Aox3, Utp14b, Il1r2, and Wdfy1 as candidate genes for mechanical hypersensitivity and Slitrk1, Slitrk6, Lmo7, Cldn10, Hs6st3, Mbnl2, and Ndfip2 for cold hypersensitivity. Threshold-free analysis of gene expression signatures across substrains, tissues, treatments and time using rank-rank hypergeometric overlap indicated earlier, post-PAC changes in DRG and SC and later changes in PAG. Dynamic, biological pathways associated with peak substrain behavioral differences and recovery included mitochondrial function (NADH:ubiquinone oxidoreductase and Cytochrome c Oxidase genes) and ribosomal function. To summarize, we identified genetic and biological sources of variation following the emergence and recovery of CIPN that have implications for understanding risk and time-dependent therapeutics.

1Laboratory of Addiction Genetics. Current Location: Center for Drug Discovery, Department of Pharmaceutical Sciences, Northeastern University, Boston, MA 02115, 2Graduate Program for Neuroscience, Boston University, Boston, MA USA, 3Transformative Training Program in Addiction Science, Boston University, Boston, MA USA, 4T32 Biomolecular Pharmacology PhD Training Program, Boston University Chobanian and Avedisian School of Medicine, 72 E. Concord St., L-603, Boston, MA 02118 USA, 5Department of Pharmacology and Toxicology, Molecular Medicine Research Building, Room 3044, Virginia Commonwealth University, 6Division of Infectious Disease, Department of Medicine and Center for Data Science, Rutgers University - New Jersey Medical School, Newark, NJ USA, 7Department of Psychiatry, University of California, San Diego, Biomedical Research Facility II (BRF2), 3A24, 9500 Gilman Drive, La Jolla, CA 92093-0667 USA, 8BU Post—Post-baccalaureate Research Education Program (PREP), Graduate Medical Sciences, Boston University, 72 E. Concord St., L-317, L-309, Boston, MA 02118

Joseph Liang1, Kristen Tsoi1, Anushka Sood1, Christopher Mok1, Ben Westmore1, Catharine Rankin1,2

Parkinson’s Disease (PD) is the most common neurodegenerative movement disorder that affects up to 1% of all people older than 65 years old. Our current understanding of the genetic contributions of PD has been expanded by genome-wide association studies in recent years. However, two major challenges plague researchers studying the genetic underpinnings of the disease today: There is a need to functionally characterize newly identified risk loci, and there remains a signification portion of loci for which a best gene candidate has yet to be assigned. We established a pipeline for in vivo characterization of C. elegans orthologs of newly identified PD risk loci. C. elegans have orthologs to many PD-associated and biologically relevant genes, there are strains with loss-of-function mutations available for almost every gene in the C. elegans genome. Notably, our lab developed the Multi-Worm Tracker for high-throughput characterization of behavioural and morphological phenotypes in populations of freely behaving animals in real time. Studying more than 150 mutant strains harbouring loss-of-function mutations in orthologs of PD-linked have yielded unique phenotypic profiles spanning up to 30 phenotypes ranging from morphology, baseline behaviours, non-associative learning and a dopamine-dependent behaviour for 83 PD-linked gene orthologs. From the data generated, we built a machine learning model to identify the best candidate gene for yet-to-be assigned loci. This research will establish high-throughput genotype-to-phenotype characterization of newly identified risk genes for PD to inform future disease modelling efforts and further our understanding of the biological processes underlying PD.

1Graduate Program of Neuroscience, University of British Columbia, Vancouver, B.C., Canada. 2Djavad Mowafaghian Centre for Brain Health, University of British Columbia, Vancouver, B.C., Canada. 3Department of Psychology, University of British Columbia, Vancouver, B.C., Canada.

Funding Support: Canadian Institutes of Health Research, Project Grant (CIHR MOP PJT-165947).

R-K Thériault¹, JD Manduca¹, Nolan CJ¹, J Lalonde², B Labonté³, IC Tobias¹, ML Perreault¹

Eph receptors are the largest family of tyrosine receptor kinases and are involved in neuronal growth and differentiation during development. The EphA2 receptor (EphA2R) has unique cell signalling characteristics and has been additionally implicated in processes causally linked to depression including vascular permeability, inflammation, and synaptic plasticity. In this work we explored the role of the EphA2R in regulating neuronal systems function in adults and its potential involvement in depression. Using rats, we first showed that chronic unpredictable stress (CUS) suppressed hippocampal (HIP) theta power and elevated HIP EphA2R signalling selectively in female rats that were stress-susceptible. Conversely, stress-susceptible males had lower high gamma power in the HIP and no alterations in EphA2R signalling. To further explore the functional role of prefrontal cortical EphA2Rs, we used a selective peptide to activate the receptor (5 nmoles/side). In female, but not male rats, acute prefrontal cortical EphA2R activation induced a depression-like phenotype and suppressed HIP theta power. HIP gene expression analysis using RNA-seq, analyzed 75 minutes post-injection, showed a rapid female-specific elevation in the gene expression of markers of inflammation, as well as microglial and immune activation. In human studies, postmortem analysis of frontal cortical gene expression in those that had depression showed significant elevations in the gene expression of several EphA2R downstream effectors in women, but not men. Together these findings suggest that heightened EphA2R activity may be associated with enhanced depression risk in women. A role for the EphA2R as a potential female-specific therapeutic target in depression warrants further investigation.

¹Department of Biomedical Sciences, University of Guelph, Guelph, ON, Canada

²Department of Molecular and Cellular Biology, University of Guelph, University of Guelph, Guelph, ON, Canada

³Department of Psychiatry and Neurosciences, Faculty of Medicine, Université Laval, Quebec City, QC, Canada.

Funding Support: University of Guelph Research Leadership Chairs Program, Canadian Institutes for Health Research, 450277