Genetic basis of variation in cocaine and methamphetamine consumption in outbred populations of Drosophila melanogaster

Abstract

We used Drosophila melanogaster to map the genetic basis of naturally occurring variation in voluntary consumption of cocaine and methamphetamine. We derived an outbred advanced intercross population (AIP) from 37 sequenced inbred wild-derived lines of the Drosophila melanogaster Genetic Reference Panel (DGRP), which are maximally genetically divergent, have minimal residual heterozygosity, are not segregating for common inversions, and are not infected with Wolbachia pipientis We assessed consumption of sucrose, methamphetamine-supplemented sucrose, and cocaine-supplemented sucrose and found considerable phenotypic variation for consumption of both drugs, in both sexes. We performed whole-genome sequencing and extreme quantitative trait locus (QTL) mapping on the top 10% of consumers for each replicate, sex, and condition and an equal number of randomly selected flies. We evaluated changes in allele frequencies among high consumers and control flies and identified 3,033 variants significantly (P < 1.9 × 10^-8) associated with increased consumption, located in or near 1,962 genes. Many of these genes are associated with nervous system development and function, and 77 belong to a known gene-gene interaction subnetwork. We assessed the effects of RNA interference (RNAi) on drug consumption for 22 candidate genes; 17 had a significant effect in at least one sex. We constructed allele-specific AIPs that were homozygous for alternative candidate alleles for 10 single-nucleotide polymorphisms (SNPs) and measured average consumption for each population; 9 SNPs had significant effects in at least one sex. The genetic basis of voluntary drug consumption in Drosophila is polygenic and implicates genes with human orthologs and associated variants with sex- and drug-specific effects.

Keywords: Drosophila Genetic Reference Panel; RNA interference; advanced intercross population; extreme QTL genome-wide association mapping.

Fig. 1.

Capillary feeder (CAFE) assay. Individual flies are collected and placed in vials containing standard fly medium 24 h prior to beginning the experiment. Flies are then transferred to vials containing 1.5% agar medium, plugged with a foam plug. A capillary filled with the appropriate treatment and a mineral oil cap is then inserted through the foam plug. The initial liquid level is marked on each capillary. Vials are placed in wire racks, and the racks are placed inside a plastic container. The plastic container is placed inside a large transparent bag to create a humidified chamber. After 18 h of feeding, the final liquid levels are marked, and the differences between the initial and final liquid levels are quantified.

Fig. 2.

Genes associated with methamphetamine and cocaine consumption. (A) Numbers of genes with allele frequencies that are significantly different between high-consuming flies and random control flies. The dots on the x axis indicate the treatment(s) in which the candidate genes were detected. The vertical lines indicate genes that were identified in more than one treatment. The horizontal bars indicate the number of significant genes for each treatment. (B) Genetic network without missing genes constructed from a subset of candidate genes identified in the xQTL mapping analysis at a Bonferroni-corrected significance level. The network consists of 77 interconnected genes (P < 0.001), of which greater than 83% possess a human ortholog (oval nodes; DIOPT ≥ 3). The blue nodes indicate genes associated with nervous system development or function. (C) A genetic interaction network constructed from the human orthologs of genes in B. The rectangles represent the input human orthologs. The triangles represent computationally recruited genes. The colored nodes indicate the gene is involved in apoptosis (purple), epidermal growth factor receptor (EGFR) signaling (blue), nerve growth factor (NGF) signaling (gray), vascular endothelial growth factor (VEGF) signaling (green), integrin cell surface interactions (red), or interactions of immunoglobin superfamily (IGSF) member proteins (yellow).

Fig. 3.

Differences in methamphetamine (A), cocaine (B), and sucrose (C) consumption between RNAi and control genotypes for 22 candidate genes. Data are presented as the difference between the amount consumed in the RNAi line and the amount consumed in the control line. Deviations below zero indicate less consumption in the RNAi line versus the control, and deviations above zero indicate greater consumption in the RNAi line. The blue bars denote males, and the red bars denote females. The bars with an asterisk indicate a significant difference from control (ANOVA, P < 0.05).

Fig. 4.

Consumption differences in H versus C allele-specific AIPs. Mean consumption of methamphetamine (A) or cocaine (B) is shown for H and C allele-specific AIP populations. The blue dots represent average consumption for C or H males; the red dots represent average consumption for C or H females. (C) Differences in consumption of cocaine or methamphetamine (METH) for males and females in H and C populations for each pair of allele-specific AIPs. Black indicates the drug was not tested for that SNP/gene. Blue indicates significantly greater consumption in the H population versus the C population. Red indicates significantly less consumption in the H population.