The genetical genomic path to understanding why rats and humans consume too much alcohol.

Background: At the invitation of the Journal, we are providing a summary of our published work that has followed the publication in 2009 of our manuscript entitled "Genetical Genomic Determinants of Alcohol Consumption in Rats and Humans". Our initial premise, which has been maintained throughout, is that knowledge regarding gene transcription would greatly enhance GWAS of alcohol-related phenotypes. We chose to concentrate our studies on the quantitative phenotype of alcohol consumption since high levels of alcohol consumption are a prerequisite for the development of alcohol use disorder (AUD). We also structured our studies to focus on "predisposition" to higher levels of alcohol consumption. We defined predisposition as a genetic structure and transcriptional pattern that is inherent in an organism and present prior to exposure to an environmental stimulus that engenders a physiological/behavioral response. In studies using humans, this interest in predisposition usually requires prolonged periods of cohort follow-up. On the other hand, studies with animals can use resources such as panels of recombinant inbred (RI) animals (in our case, the HXB/BXH rat panel) to capture the transcriptional landscape of animals not exposed to alcohol and compare this transcriptional landscape to levels of alcohol consumption collected from a different cohort of animals that are the same age, have an identical genetic composition, and are raised in an identical environment. The other benefit is that the stable genetic structure of inbred strains allows for a chronological expansion of information on these animals. This characteristic of the HXB/BXH RI rats allowed us to add important information as technology and analytical methods developed over time.

Methods findings and conclusions: Our initial studies relied on hybridization arrays for RNA quantification in brain, an initial set of polymorphic markers for the rat genome, and a standard behavioral (b)QTL analysis for alcohol consumption. What we added to the conceptual basis for analysis and interpretation was the calculation of transcript expression (e)QTLs and the requirements that: 1. the eQTL overlapped the location of the bQTL; and 2. the transcript levels were significantly correlated with the quantitative levels of alcohol consumption across rat strains. These criteria were used to identify genes (transcripts) as "candidate" contributors to the alcohol consumption phenotype. We soon realized that the search for candidate genes as unique determinants of a complex trait is irrational, since these phenotypes are best characterized by differences in genetic networks. Therefore, we incorporated Weighted Gene Coexpression Network Analysis (WGCNA) in our further work. We also realized the limitations of hybridization arrays for breadth of transcriptome coverage and quantification, and in the more current work used total RNA-Seq-derived data for characterizing nearly all of the brain transcriptome. Finally, we participated in the efforts for whole genome sequencing of the strains of the HXB/BXH panel, generating an extensive new panel of markers for remapping of the QTLs. We also realized that the biological determinants of a behavioral phenotype do not have to reside in brain and, by examining the liver transcriptome, we found that the gut-liver-brain axis was, in part, involved in predisposition to higher levels of free-choice alcohol consumption. In all, from the first exploration of the genetical genomics of the alcohol consumption phenotype, to the current status of our work, the function of the brain immune system, with emphasis on microglia and astrocytes, even prior to the animal being offered alcohol, has emerged as a most significant genetic contributor to the amount of alcohol an animal will consume on a daily basis. Particularly prominent was a cluster of inflammasome (NLRP3)-modulating transcripts (P2rx4, Ift81, Oas1b, Txnip) and a long noncoding transcript, "Lrap" that repeatedly appeared within a gene coexpression module associated with alcohol consumption levels. Interestingly, data from post-mortem tissue from brain of humans suffering from AUD also indicates a hyperactive neuroimmune function. The data from studies with animals may indicate that neuroimmune hyperactivity may be a trait rather than a state marker for AUD.