There are few effective therapies to treat drug abuse and dependence, despite the urgent need to develop more effective treatments. A major impediment to the development of such treatments is our extremely limited understanding of the biological basis of drug abuse. While human genome-wide association studies (GWAS) have begun to elucidate genes that influence various traits relevant to drug abuse, they are still unable to attribute more than a small fraction of the heritable variance to specific genes. NIDA center for genetic studies of drug abuse in outbred rats (P50DA037844) was created in 2014 to perform GWAS on numerous behavioral traits that have well-established relevance to drug abuse using outbred rats. The Center was successfully renewed in 2019. We expect to discover new genes that can influence drug abuse-related behaviors in rats, therefore improving our understanding of genetic susceptibility to drug abuse in humans. Our results will help to identify new opportunities to treat psychiatric disorders, including addiction.
Why rats? The use of rats is critical because many of the phenotypes we study were developed using rats and difficult or impossible to adapt for mice. We utilize a unique rat population: the N/NIH heterogeneous stock (HS), which is derived from 8 fully sequenced inbred rat strains and has been maintained as an outbred population for 70+ generations. This population is designed to resolve quantitative trait loci (QTL) to regions that are small enough to permit identification of specific genes. The colony is maintained by Dr. Leah Solberg Woods, Wake Forest University (Breeding Core).
What behaviors do we study? We have carefully selected a range of behavioral traits that model behavioral predispositions and inferred psychological processes that are known to influence vulnerability to drug abuse in humans. This approach is consistent with the NIMH Research Domain Criteria (RDoC) which seeks to emphasize traits rather than diagnoses. Several psychological domains are represented. Drs. Terry Robinson and Shelly Flagel from University of Michigan (Research Project 1) focus on the propensity of animals to attribute motivational value (“incentive salience”) to discrete reward cues, as indexed by variation in Pavlovian conditioned approach (PavCA) behavior. They also examine the ability of cocaine-associated contextual cues to control behavior. Dr. Hao Chen from University of Tennessee Health Sciences Center (Research Project 2) examines the role of social cues in the development and relapse of intravenous nicotine self-administration behavior in adolescent rats. He also examines reinstatement of drug seeking behavior upon re-exposure of the rat to the complex social environment previously associated with nicotine availability. Drs. Jerry Richards and Paul Meyer from University at Buffalo, Research Institute on Addictions (Research Project 3) use reward cues to probe variation in a constellation of traits including sensation seeking, impulsivity and cocaine-induced cue preferences, all of which have been associated with addiction. In each case, multiple quantitative measures from multiple behavioral tests reflecting multiple psychological constructs will be examined. We have previously demonstrated that behavioral traits we have selected are highly heritable in HS rats. Our studies will provide a unique opportunity to identify shared genetic control (a.k.a. co-heritability) among these traits and empirically test long-standing theories related to the psychological processes being modeled.
How do we genotype HS rats? We use an innovative technology called genotype by sequencing (GBS) to obtain genotypes from all rats tested in Research Projects at the lowest possible cost. GBS is a reduced representation method based on restriction enzymes that cut at more than 1M sites in the rat genome. We then anneal indexed adapters to the resulting overhangs and sequence regions proximal to the restriction sites. The variants obtained using GBS are evenly spaced throughout the rat genome and provide at least hundreds of thousands of SNPs per individual (Core C). We then use these SNPs to impute millions of SNPs, taking advantage of the fact that the eight founders of the HS have been sequenced (Core C). We currently impute approximately 7 million SNPs that segregate among the 8 founder strains.
How do we measure gene expression? Dr. Chen (Research Project 2) collected brain tissue from five brain regions that play key roles in the addiction related behaviors: nucleus accumbens core, lateral habenula, infralimbic cortex, orbitofrontal cortex and prelimbic cortex. The tissue is collected from ~ 200 behaviorally naive non-sibling rats. The Sequencing Core (Core C) will measure gene expression in RNA isolated from brain regions using next-generation sequencing. These measurements will be used to identify expression QTLs (eQTLs) for both genes and exons and will also use allele-specific expression as an independent means of identifying cis-eQTLs (Core C). We will also conduct joint analyses of multiple brain regions to increase power for eQTL studies.
How do we analyze genotype, phenotype and expression data? The identification of specific genes is a key goal of this center because genes are the unit at which we hope to translate our findings to the field of human genetics, and because genes can be most easily manipulated to test and validate the causal inferences. The data analysis will be performed by the Core C. To find QTLs associated with behavioral phenotypes, we will perform genome wide association studies (GWAS). The HS rats have complex genetic relationships to one another (siblings, cousins, etc.), therefore we use special statistical approach, using linear mixed model that accounts for relatedness among subjects. Thresholds for significance is empirically estimated using permutation. Since we phenotype equal numbers of male and female rats, we are able examine sex differences by using gender as a covariate for GWAS. In order to Identify specific genes for specific behaviors we will integrate the data collected by all Research Projects. These data will include behavioral QTLs, expression QTLs and known protein coding variants. Two complementary approaches will be used. The first will depend on data visualization tools and human integration of external data sources. This approach will be primarily driven by the PIs from the Research Projects 1-3. In addition, in the we will apply more sophisticated and comprehensive (“data-driven”) methods. We will do this by jointly modeling multiple behavioral phenotypes (to gain power and to detect pleiotropic relationships), and jointly modeling of behavioral phenotypes with expression phenotypes (to implicate specific genes).