Functional Genomic Analysis of Amphetamine Sensitivity in Drosophila

Abstract

Abuse of psychostimulants, including amphetamines (AMPHs), is a major public health problem with profound psychiatric, medical, and psychosocial complications. The actions of these drugs at the dopamine transporter (DAT) play a critical role in their therapeutic efficacy as well as their liability for abuse and dependence. To date, however, the mechanisms that mediate these actions are not well-understood, and therapeutic interventions for AMPH abuse have been limited. Drug exposure can induce broad changes in gene expression that can contribute to neuroplasticity and effect long-lasting changes in neuronal function. Identifying genes and gene pathways perturbed by drug exposure is essential to our understanding of the molecular basis of drug addiction. In this study, we used Drosophila as a model to examine AMPH-induced transcriptional changes that are DAT-dependent, as those would be the most relevant to the stimulatory effects of the drug. Using this approach, we found genes involved in the control of mRNA translation to be significantly upregulated in response to AMPH in a DAT-dependent manner. To further prioritize genes for validation, we explored functional convergence between these genes and genes we identified in a genome-wide association study of AMPH sensitivity using the Drosophila Genetic Reference Panel. We validated a number of these genes by showing that they act specifically in dopamine neurons to mediate the behavioral effects of AMPH. Taken together, our data establish Drosophila as a powerful model that enables the integration of behavioral, genomic and transcriptomic data, followed by rapid gene validation, to investigate the molecular underpinnings of psychostimulant action.

Keywords: DGRP; Drosophila; S6K (70-kDa ribosomal protein S6 kinase); amphetamine; dopamine transporter; mammalian target of rapamycin (mTOR); psychostimulants; transcriptomic (RNA-Seq).

Figure 1

Differential gene expression in response to AMPH in DAT mutant and isogenic control. (A–D) are gene expression scatterplots for the two strains of flies: WT (w¹¹¹⁸ isogenic strain) and DAT mutant. (A,B) are MA plots and (C,D) are volcano plots. In (A–C), differentially expressed transcripts (DE transcripts, those that change their expression in treatment vs. vehicle) are shown in teal; transcripts with changes below the statistical cut-off are in red, while transcripts for which the differential expression status could not be resolved are shown in gray. (E) Overlap between the DE transcript sets between WT (gray circle) and DAT mutant (black circle). At FDR < 0.1, there were 717 DE transcripts in the WT strain and 629 DE transcripts in the DAT mutant flies. The two lists of DE transcripts had 308 genes that were DE only in WT (blue), 220 transcripts that were DE only in DAT (green) and 409 transcripts in common (red). (F) Comparison of the expression level changes between the WT and DAT mutant strains. The 409 DE transcripts shared between the two strains are shown in red; the 220 DAT mutant-specific DE transcripts are shown in green, and the 308 WT-specific DE transcripts are shown in blue.

Figure 2

Comparison of the top GO terms that are enriched in DE genes in the WT strain. GO terms are compared by their p-values; the centers of the circles correspond to –log10(p) = 0, which increases outwards. (A) Downregulated genes; (B) upregulated genes. Gray: all DE genes in WT (these correspond to the 717 DE transcripts in Figure 1E, gray circle); red: DE genes that are shared between WT and DAT mutant (these correspond to the 409 transcripts depicted in red in Figures 1E,F); blue: DE genes that are unique to WT and not shared with DAT mutant (these correspond to the 308 transcripts depicted in blue in Figures 1E,F). The diagrams were created using the R package ggradar (36).

Figure 3

Unique GO and KEGG terms enriched with the genes that are upregulated in response to AMPH in a DAT-dependent manner. Functional category/term names are on the left. Circle sizes are proportional to the number of DE genes (gene counts), which were annotated as belonging to the corresponding functional category.

Figure 4

Functional validation of candidate genes. (A) Gene network depicting physical and genetic interactions between candidate genes identified in RNA-seq and GWA analyses. Orange lines indicate reported physical interaction, blue arrows indicate enhancing genetic interaction, and red arrows indicate suppressing genetic interaction. Network generated using the esyN webtool. S6K and Thor encode direct targets of the MTOR signaling pathway, S6 Kinase and 4E-BP, which are known to interact with ribosomal proteins to modulate mRNA translation. hppy encodes MAP4K3 which regulates the phosphorylation of S6K and 4E-BP. Eip75B, srl, and trbl all encode modulators of insulin-mTOR S6K signaling. (B) RNAi-mediated knockdown of candidate genes was targeted to dopamine neurons using TH-GAL4. Bar graphs depict change in activity in response to 5 mM AMPH (blue) or 10 mM AMPH (yellow). Error bars indicate SEM. Statistical significance was determined by Kruskal-Wallis ANOVA (p < 2e-16). Asterisks indicate pairwise significance compared to genotype control after AMPH treatment, as determined by post-hoc Dunn's Test with a Benjamini-Hochberg correction for multiple testing, ****p.adj < 0.0001, ***p.adj < 0.001, **p.adj < 0.01, *p.adj < 0.05.