Alcohol-induced aggression in Drosophila

Abstract

Alcohol-induced aggression is a destructive and widespread phenomenon associated with violence and sexual assault. However, little is understood concerning its mechanistic origin. We have developed a Drosophila melanogaster model to genetically dissect and understand the phenomenon of sexually dimorphic alcohol-induced aggression. Males with blood alcohol levels of 0.04-mg/ml BAC were less aggressive than alcohol-naive males, but when the BAC had dropped to ~0.015 mg/ml, the alcohol-treated males showed an increase in aggression toward other males. This aggression-promoting treatment is referred to as the post-ethanol aggression (PEA) treatment. Females do not show increased aggression after the same treatment. PEA-treated males also spend less time courting and attempt to copulate earlier than alcohol-naive flies. PEA treatment induces expression of the FruM transcription factor (encoded by a male-specific transcript from the fruitless gene), whereas sedating doses of alcohol reduce FruM expression and reduce male aggression. Transgenic suppression of FruM induction also prevents alcohol-induced aggression. In male flies, alcohol-induced aggression is dependent on the male isoform of the fruitless transcription factor (FruM). Low-dose alcohol induces FruM expression and promotes aggression, whereas higher doses of alcohol suppress FruM and suppress aggression.

Keywords: Drosophila; aggression; alcohol-induced aggression; alcohol-use disorder; fruitless; sexually dimorphic behaviour.

Figure 1.. Low dose of ethanol (PEA paradigm) enhances aspects of male aggression.

(A) Alcohol-treatment protocol. (B) Blood alcohol levels measured immediately after alcohol exposure (Low Dose Mean +/− SEM; 0.047 +/− 0.002 mg/mL, n = 24) and one hour after alcohol exposure (PEA; Mean +/− SEM; 0.0144 +/− 0.002 mg/mL, n = 20). (C) PEA treatment increases the number of male-male encounters (p = 0.0029). (D) PEA treatment increases the time spent in male:male encounters (p = 0.0276), (E) the time spent high fencing (p = 0.0196), (F) the time spent shoving (p = 0.0292, ), (G) the number of lunges (p = 0.0086), and (H) the fraction of flies that lunged (p = 0.0068). However, the period of time spent (I) low fencing (p = 0.0572), (J) tussling (p = 0.1625), or (K) boxing (p = 0.3557) were not increased by PEA treatment. For (C) through (K), N = 17 or 19. In (B)-(G) and (I)-(K) the middle line is the mean and outer brackets are SEM and the Mann-Whitney test was used to evaluate significance. For H, the Mantel-Cox log-rank test was used to evaluate significance.

Figure 2.. PEA treatment increases dominance but reduces courtship behavior in males.

(A) Positive Retreat Index indicates that alcohol-naive flies are more likely to retreat, and negative values indicate that PEA flies are more likely to retreat (one-sample Wilcoxon Signed Rank test, μo = 0, p = 0.0229, n = 36). (B) Positive Lunge Index indicates that PEA flies are more likely to lunge, and vice versa for negative values (one-sample Wilcoxon Signed Rank test, μ_o = 0, n = 18 p = 0.2890). (C) Increased lunging and dominance are associated with PEA treatment. Shown is Lunge Index vs. Retreat Index. More positive ordinate values represent greater incidence of lunging by PEA flies and more positive abscissa values represent greater incidence of retreat by control flies. Conversely, more negative ordinate values represent greater incidence of lunging by control flies and more negative abscissa values represent greater incidence of retreat by PEA flies. Heat map represents frequency. Incidence of PEA lunging is correlated with dominance by the PEA individual (Spearman r test, p = 0.0090, N = 27 XY pairs). (D) PEA treatment reduces the time spent performing UWEs (Mann Whitney test p = 0.0107, n = 39, 38) and the (E) latency to attempt mating (Mann Whitney test p = 0.0777, N = 29). (F) Three categories of fru splice variants. Transformer suppresses FruM production and induces DsxF expression. (G) Time spent in male:male interaction plotted across genotypes (includes all forms of male:male aggression; + indicates no UAS or Gal4 transgene) and treatment (C = control, PEA = Post-Ethanol Aggression treatment; Mann Whitney test, * p < 0.05, ** p < 0.01, n = 10–24). (H) Fold change fru^M transcript abundance calculated using ΔΔCT method (Mann Whitney test, * p=0.0395, n = 40, 16). (I) Merged optical stacks across whole-brain volume for Control and PEA flies stained for FruM protein. Scale bar = 100 um. (J) Whole-brain relative intensity of FruM antibody stained fly brains. Relative intensity (normalized to control) of FruM signal (Mann Whitney test, * p = 0.0434, n = 8, 10).

Figure 3.. A sedating-dose of alcohol reduces fru^M expression.

(A) Position of primers used to target each fru splice variant. (B) Treatment paradigms for alcohol vapor sedation (top) and alcohol feeding (bottom). (C) There is a significant reduction in fru^M 6h and 24h after alcohol vapor sedation (One-way ANOVA p = 1.8e-6 with Dunnett post hoc test; 6h * p = 0.0213; 24h ** p < 0.0001). (D) Feeding on 20% alcohol food for 3 days also reduces fru^M levels (One-way ANOVA p = 0.0053 with Dunnett post hoc test; 0h ** p = 0.00769; 48h p = 0.80023). (E) Expression of fru^F (measured in females) and (F) fru^COM (measured in males) are not changed 24 hours post alcohol vapor sedation. (G) Expression of dsx^F in male flies remains unchanged 24 hours after alcohol vapor sedation (two-tailed unpaired Student’s t-test).

Figure 4.. Alcohol vapor sedation suppresses fru^M expression in a region-specific manner.

(A) Paradigms for alcohol vapor sedation and alcohol feeding. (B) GFP fluorescence from dissected fruGal4>eGFP adult brains stained with DAPI to visualize nuclei. Alcohol-naive brain on left and a brain from an animal sedated with alcohol vapor on the right. Scale bar = 100 μm. (C) Quantification of whole brain fruGal4>eGFP signal following alcohol vapor sedation (Kruskal-Wallis test p = 0.02 with Dunn’s Multiple Comparison Test, 24h post treatment not significant, 48h post treatment * p < 0.05, n = 12, 11, 11) and following feeding on 20% alcohol food for 3 days (Student’s t-test alcohol, p = 0.01399, n = 4). (D) Alcohol vapor sedation did not alter fruGal4>eGFP expression in optic lobes whereas mushroom body accumulation was depressed (Mann Whitney test p = 0.0352, n = 10). (E) Relative intensity of fruGal4>eGFP signal in animals fed food containing 20% alcohol for 3 days. Expression in the antennal lobe glomeruli DA1 and Va1v were depressed. (DA1: Mann Whitney p = 0.0047; Va1v Mann Whitney p = 0.0379, n = 10). (F) Neural-specific R57C10-Gal4>eGRP was not depressed by the alcohol sedation (Mann Whitney p = 0.8857, n = 4). (G) fruGal4>GFP signal is depressed in antenna 24 h after alcohol-vapor sedation (Mann Whitney p = 0.0012, n = 7). (H) Representative brains stained with FruM antibody control and 48 h after the animals were sedated alcohol vapor. Scale bar = 100 um. (I) Whole-brain relative intensity of FruM antibody stained fly brains 24 h and 48 h following alcohol vapor sedation (normalized to control; Kruskal-Wallis test p = 0.0103 with Dunn’s Multiple Comparison Test, 24h not significant; 48h * p < 0.05, n = 10, 4, 5).

Figure 5.. Venn Diagram of FruM DamID identified transcriptional targets and functionally validated alcohol genes.

There is significant overlap (Hypergeometric distribution p = 0.3.8e-11) between FruM target genes and genes functionally validated to modulate alcohol behaviors . The list of genes are those in the region of intersection.