Distinct patterns of alcohol and water consumption depending on sex and alcohol concentration

Here we used a modified DID binge drinking paradigm to assess the effects of sex on water and alcohol consumption and preference across multiple concentrations of alcohol (Fig. 1a). We first measured raw cumulative intake of EtOH solution or H2O in the 20% EtOH DID, 10% EtOH DID, and H2O DID mice (Fig. 1c–e). A 2xRM-ANOVA on raw cumulative intake of 20% EtOH showed main effects of sex (F (1, 18) = 5.97, *P = 0.025) and cycle (F (7, 126) = 1141.0, ****P < 0.0001) and an interaction between the two (F (7, 126) = 4.45, ***P = 0.0002). Post hoc t-tests with H–S corrections showed that raw intake volume was higher in F than M on cycles 5–8 (cycle 5: t144 = 2.67, *P = 0.050; cycle 6: t144 = 3.08, *P = 0.019; cycle 7: t144 = 3.60, **P = 0.005; cycle 8: t144 = 3.5, **P = 0.006; all other cycles: Ps > 0.30; Fig. 1c). In contrast, cumulative raw intake of 10% EtOH was not different in M and F, as there was an effect of DID cycle (F (7, 126) = 702.30, ****P < 0.0001) but no effect of sex or interaction (Ps > 0.5) in a 2xRM-ANOVA (Fig. 1d). And, cumulative raw intake of water was not different between M and F in H2O DID (2xRM-ANOVA: main effect of DID cycle (F (7, 126) = 311.90, ****P  < 0.0001) but no effect of sex or interaction (Ps > 0.40; Fig. 1e). These results show that raw solution intake was greater in F than M only for high concentration (20%) EtOH. When EtOH or water consumption was normalized to bodyweight, females cumulatively consumed more 20% EtOH, 10% EtOH, and H2O than M (Fig. 1f–h). This emerged by cycle 3 for 20% EtOH DID, cycle 4 for 10% EtOH DID, and cycle 6 for H2O DID. A 2xRM-ANOVA for 20% EtOH DID showed main effects of sex (F (1, 18) = 29.37, ****P < 0.0001) and cycle (F (7, 126) = 862.3, ****P < 0.0001) and an interaction between the two (F (7, 126) = 28.42, ****P < 0.0001). Post hoc t-tests with H–S corrections showed that normalized consumption was higher in F than M on cycles 3–8 (*Ps < 0.05, as indicated; cycles 1 and 2: Ps > 0.15; Fig. 1f). A 2xRM-ANOVA for 10% EtOH DID showed main effects of sex (F (1, 18) = 14.11, **P = 0.0014) and cycle (F (7, 126) = 810.6, ****P < 0.0001) and an interaction between the two (F (7, 126) = 16.26, ****P < 0.0001). Post hoc t-tests with H–S corrections showed that normalized consumption was higher in F than M by cycles 4–8 (*Ps < 0.05, as indicated; cycles 1–3: Ps > 0.10; Fig. 1g). A 2xRM-ANOVA for H2O DID show main effects of sex (F (1, 18) = 9.73, **P = 0.006) and cycle (F (7, 126) = 368.3, P < 0.0001) and an interaction between the two (F (7, 126) = 10.42, ****P < 0.0001). Post hoc t-tests with H–S corrections showed that normalized H2O consumption was higher in F than M by cycles 6–8 (*Ps < 0.05, as indicated) and trended toward higher on cycle 5 ($P = 0.062; Fig. 1h), suggesting that females drank more (cumulative) water than males across the entire mDID paradigm, but increased bodyweight-normalized consumption was more robust for increasing EtOH concentrations. Examination of the average EtOH/water consumption for the first 2 h/day on each DID cycle (average of days 1–3 and the first 2 h on day 4) showed that females consumed more EtOH than males but not more water than males during this early access period (Fig. 1i–k). A 2xRM-ANOVA for 20% EtOH DID showed main effects of sex (F (1, 18) = 24.22, ***P = 0.0001) and cycle (F (7, 126) = 6.604, ****P < 0.0001) and an interaction between the two (F (7, 126) = 2.12, *P = 0.046). Post hoc t-tests with H–S corrections showed that average 2 h consumption was higher in F than M on cycles 2–8 (*Ps < 0.05, as indicated; cycle 1: P = 0.141; Fig. 1i). For 10% EtOH DID, a 2xRM-ANOVA showed main effects of sex (F (1, 18) = 18.19, ***P = 0.0005) and cycle (F (7, 126) = 2.09, *P = 0.049) but no interaction between the two (P = 0.1707). Post hoc t-tests with H–S corrections showed that average 2 h consumption was higher in F than M on cycles 3, 4, and 6 (*Ps < 0.05, as indicated) but not other cycles (Ps > 0.14; Fig. 1j). For H2O DID, a 2xRM-ANOVA showed a main effect of cycle (F (7, 126) = 4.66, ****P = 0.0001) but no effect of sex or interaction (Ps > 0.25; Fig. 1k). These results confirm that females’ increased consumption of EtOH compared to males was more robust for 20% than 10% EtOH. Further, they suggest while females display higher cumulative fluid consumption across the mDID paradigm for all EtOH/H2O conditions (Fig. 1f–h), their higher consumption of EtOH during the first two hr of each drinking day (Fig. 1i,j) was not attributable to higher fluid consumption, including water, during that time (Fig. 1k).

Fig. 1

Alcohol and water consumption are sex- and alcohol concentration-dependent. A Schematic depicting the modified Drinking in the Dark (DID) binge drinking paradigm in C57BL/6J mice. B Bodyweight gain was similar across DID cycles in H2O, 10% EtOH, and 20% EtOH DID mice of both sexes. C Cumulative intake of 20% EtOH solution (ml) was higher in females compared to males on DID cycles 5–8. D There was no sex difference in cumulative 10% EtOH intake (ml) across DID cycles. E There was no sex difference in cumulative H2O intake (ml) across DID cycles. F Cumulative EtOH consumption normalized to bodyweight (g/kg) was higher in females compared to males on DID cycles 3–8. G Females consumed more 10% EtOH (g/kg) compared to males on DID cycles 4–8. H Females consumed more H2O (ml/kg) compared to males during DID cycles 6–8, with a trend on cycle 5. I Females displayed higher average 20% EtOH consumption (g/kg) during the first 2 h /day of access on DID cycles 2–8. J Females displayed higher average 10% EtOH consumption (g/kg) on the first 2 h /day of access on DID cycles 3, 4, and 6. K There was no difference between males and females on average H2O consumption during the first 2 h/day of water access. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in 2xRM-ANOVA main effects and interactions of sex and DID cycle, as well as post hoc t-tests with H–S corrections between M and F. $P < 0.10 for post hoc t-tests with H–S corrections between M and F

To further characterize drinking patterns, we next dissected Day 4 two bottle choice (2-BC) drinking behavior, examining both consumption and preference across the 2 h, 4 h, and 24 h time points (Figs. 2 and 3). During the first 2 h of Day 4 of 20% EtOH DID, females drank more EtOH but not water than males (Fig. 2a). A 2xRM-ANOVA for EtOH consumption showed main effects of sex (F (1, 18) = 16.29, ***P = 0.0008) and cycle (F (7, 126) = 5.72, ****P < 0.0001) but no interaction between the two (P = 0.268). Post hoc t-tests with H–S corrections showed that Day 4 2-h consumption was higher in F than M on cycles 2, 5, and 7 (*Ps < 0.05, as indicated) and trended toward an increase on cycles 4, 6, and 8 ($Ps < 0.10) but did not differ on cycles 1 and 3 (Ps > 0.30; Fig. 2a). In contrast, water consumption did not differ between females and males, as there was a main effect of cycle (F (7, 126) = 4.25, ***P = 0.0003) but no effect of sex or interaction (Ps > 0.25). Four-hour consumption showed a similar pattern (Fig. 2b), with a 2xRM-ANOVA for 4-h 20% EtOH showing main effects of sex (F (1, 18) = 24.75, ****P < 0.0001) and cycle (F (7, 126) = 4.13, ***P = 0.0004) but no interaction between the two (P = 0.714). Post hoc t-tests with H–S corrections showed that 4-h consumption was higher in F than M on cycles 2, 3, 4, 5, and 8 (*Ps < 0.05, as indicated) and trended toward an increase on the remaining cycles ($Ps < 0.10). In contrast, for water consumption there was a main effect of cycle (F (7, 126) = 4.03, ***P = 0.0005) and a trend for a main effect of sex (F (1, 18) = 3.74, $P = 0.069) but no interaction (P > 0.55). For 24-h 20% EtOH consumption (Fig. 2c), a 2xRM-mixed effects model showed main effects of sex (F (1, 18) = 28.12, ****P < 0.0001) and cycle (F (7, 116) = 2.40, *P = 0.025) but no interaction between the two (P = 0.495). Post hoc t-tests with H–S corrections showed that 24-h consumption was higher in F than M on cycles 2–5 (*Ps < 0.05, as indicated) and trended toward an increase on cycle 7 ($P = 0.087). There was a trend towards a main effect of sex on water consumption (F (1, 18) = 4.11, $P = 0.058) as well as a main effect of cycle (F (7, 116) = 5.06, ***P < 0.0001) but no interaction between the two (P > 0.50 at this timepoint. During the first 2 h of Day 4 of 10% EtOH DID, females drank more EtOH but not water than males (Fig. 2d). A 2xRM-ANOVA for EtOH consumption showed a main effect of sex (F (1, 18) = 21.77, ***P = 0.0002) and a trend toward a main effect of cycle (F (7, 124) = 2.0, P = 0.061) but no interaction between the two (P = 0.619). Post hoc t-tests with H–S corrections showed that Day 4 2-h consumption was higher in F than M on cycles 2, 3, 4, and 7 (*Ps < 0.05, as indicated) but did not differ on cycles 1, 5, 6, and 8 (Ps > 0.15). In contrast, water consumption did not differ between females and males, as there was a trend toward an effect of cycle (F (7, 124) = 1.83, $P = 0.0871) but no effect of sex or interaction (Ps > 0.40). Four-hour 10% EtOH consumption showed a similar pattern (Fig. 2e), with a 2xRM-ANOVA for 4-h 10% EtOH showing a main effect of sex (F (1, 18) = 28.23, ****P < 0.0001) and a trend toward the effect of cycle (F (7, 124) = 2.06, $P = 0.053) and an interaction between the two (F (7, 124) = 2.29, *P = 0.0312). Post hoc t-tests with H–S corrections showed that 4-h 10% EtOH consumption was higher in F than M on cycles 2, 3, 4, and 7 (*Ps < 0.05, as indicated) but not the other cycles (Ps > 0.10). In contrast, water consumption did not differ between females and males, as there were no effects or interaction (Ps > 0.10). For 24-h 10% EtOH consumption (Fig. 2f), a 2xRM-ANOVA showed main effects of sex (F (1, 18) = 15.81, ***P = 0.0009) and cycle (F (7, 123) = 5.00, ****P < 0.0001) but no interaction between the two (P = 0.188). Post hoc t-tests with H–S corrections showed that 24-h consumption was higher in F than M on cycles 2, 3, and 7 (*Ps < 0.05, as indicated) and trended toward an increase on cycles 6 ($P = 0.074) and 8 ($P = 0.072) but was not different on cycles 1, 4, and 5 (Ps > 0.20). For water consumption, there was a trend towards a main effect of sex (F (1, 18) = 3.63, $P = 0.073) and a main effect of cycle (F (7, 123) = 3.00, *P = 0.006) but no interaction (P = 0.705).

Fig. 2

Females consume more alcohol than males across EtOH concentrations at 2-, 4-, and 24-h timepoints on Day 4 access but have similar alcohol preference. A–C EtOH (top row) and water (bottom row) consumption at 2-h, 4-h, and 24-h time points on Day 4, 2-bottle choice (2-BC) access in 20% EtOH DID across cycles. A During the first 2 h, females rapidly escalated EtOH drinking and consumed more 20% EtOH (g/kg) than males during DID cycles 2, 5, and 7, with a trend towards increased consumption on cycles 4 and 6. There was no difference in water consumption between the sexes. B The sex difference in 20% EtOH consumption was even more robust for the first 4 h of Day 4 access, with females consuming more than males on DID cycles 2–5, and 8 and a trend towards increased consumption cycles 1, 6, and 7, with a trend towards a difference in water consumption between the sexes. C Females consumed more 20% EtOH (g/kg) than males during the 24-h period of Day 4 access on DID cycles 2–5, with a trend towards a difference in water between males and females. D–F EtOH (top row) and water (bottom row) consumption at 2-h, 4-h, and 24-h time points on Day 4, 2-bottle choice (2-BC) access in 10% EtOH DID across cycles. D During the first 2 h, females consumed more 10% EtOH (g/kg) than males on DID cycles 3, 4, and 6, with a trend towards increased consumption in cycles 2 and 7. There was no difference in water consumption. E During the first 4 h, females consumed more 10% EtOH (g/kg) than males during DID cycles 2, 3, 4, 6, and 7. There was no difference in water consumption. F During the 24-h access period, females consumed more 10% EtOH than males during DID cycles 2, 3, 6, and 7, with a trend toward increased consumption in cycle 8. Water consumption was overall higher in females compared to males. G–I Average EtOH preference in 20% and 10% DID at 2-h (G), 4-h (H), and 24-h (I) time points, showing that males and females similarly displayed higher EtOH preference for 10% compared to 20% EtOH. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in 2xRM-ANOVA main effects and interactions of sex and DID cycle (A–F) and 2xANOVA main effects of EtOH concentration (G–I) and post hoc t-tests with H–S corrections. $P < 0.10 for post hoc t-tests with H–S corrections between M and F (A–F) and EtOH concentration within sex (G–I)

Fig. 3

Higher female water consumption emerges across the day. A–C Water consumption at 2-h, 4-h, and 24-h time points on Day 4, 2-bottle choice (2-BC) access in H2O DID across cycles. A There were no sex differences in H2O consumption (ml/kg) during the first 2 h of Day 4 access. B Females consumed more H2O (ml/kg) than males during the first 4 h of Day 4 on cycles 3 and 8, with a trend towards increased consumption during cycle 6. C Females consumed more H2O (ml/kg) than males during the 24-h period on cycles 2–8. D–F Average total volume consumption for Day 4 EtOH and water in 20% EtOH, 10% EtOH, and H2O DID. D During the first 2 h, total volume consumption was highest for 10% EtOH, and females consumed more fluid than males in 10% EtOH DID. E Within 4 h, females consume greater volume than males for all conditions, and consumption was higher for H2O, particularly in females. F At the 24-h time point, females consumed more fluid than males in all DID conditions; there were no differences in fluid consumption between DID conditions in either sex. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in 2xRM-ANOVA main effects and interactions of sex and DID cycle (A–C) and 2xANOVA main effects of sex and EtOH concentration (D–F), as well as post hoc t-tests with H–S corrections as indicated. $P < 0.10 for post hoc t-tests with H–S corrections between M and F

We evaluated whether sex or EtOH concentration affected EtOH preference at the 2, 4, and 24-h time points on Day 4 2-BC (Fig. 2g–i). While EtOH preference was significantly above chance for both sexes at both 10% and 20% concentrations (one-sample t-tests compared to the null hypothesis of 0.5 preference score: all Ps < 0.01, not indicated), there were differences between concentrations that depended on sex. We found that 2-h EtOH preference (averaged across all cycles) was higher for 10% than 20% EtOH (Fig. 2g). A 2xANOVA showed a main effect of EtOH concentration (F (1, 36) = 4.56, *P = 0.040) but no effect of sex or interaction (Ps > 0.65). However, post-hoc t-tests with H–S corrections showed no differences in preference for 10% versus 20% EtOH preference in either females (P = 0.156) or males (P = 0.2282). This phenotype was stronger at 4 h (Fig. 2h), as a 2xANOVA showed a main effect of EtOH concentration (F (1, 36) = 15.61, ***P = 0.0003) but no effect of sex or interaction (Ps > 0.40). Post-hoc t-tests with H–S corrections showed a higher 10% than 20% EtOH preference in females (t36 = 3.36, **P = 0.004) and males (t36 = 2.23, *P = 0.032). At 24 h, both sexes robustly displayed a stronger 10% than 20% EtOH preference (Fig. 2i). A 2xANOVA showed a main effect of EtOH concentration (F (1, 36) = 35.05, ****P < 0.0001) but no effect of sex or interaction (Ps > 0.55). Post-hoc t-tests with H–S corrections showed higher 10% than 20% EtOH preference in both sexes (F: t36 = 3.77, ***P = 0.0006; M: t36 = 4.61, ****P < 0.0001).

We also examined water consumption at these 2, 4, and 24 h time points in mice that underwent H2O DID for comparison (Fig. 3a–c). For 2-h consumption (Fig. 3a), a 2xANOVA revealed a main effect of cycle (F (7, 125) = 4.81, ****P < 0.0001) but no effect of sex or interaction (Ps > 0.15). For 4-h H2O consumption (Fig. 3b), a 2xANOVA showed a main effect of sex (F (1, 18) = 11.86, **P = 0.003) and cycle (F (7, 125) = 4.92, ****P < 0.0001) and a trend for interaction between the two (F (7, 126) = 2.20, *P = 0.039). Post hoc t-tests with H–S corrections showed that Day 4 4-h consumption was higher in F than M on cycles 3, 7, and 8 (*Ps < 0.05, as indicated) and trended toward higher on cycle 6 ($P = 0.059) but did not differ on cycles 1, 2, 4, 5, or 7 (Ps > 0.25). For 24-h H2O DID water consumption (Fig. 3c), a 2xRM-mixed effects model showed main effects of sex (F (1, 18) = 30.65, ****P < 0.0001) and cycle (F (7, 115) = 3.27, **P = 0.003) but no interaction between the two (P = 0.158). Post hoc t-tests with H–S corrections showed that 24-h consumption was higher in females than males on all cycles but the first (cycle 1: P = 0.266; all other cycles: *Ps < 0.05, as indicated). These results show that water consumption in females is higher than in males, but this emerges across the 24-h period after the first 2 h of access. Given these results, we further assessed total volume consumed across 2, 4, and 24 h time points of Day 4 averaged across cycles for all mice (Fig. 3d–f). For the 2-h time point (Fig. 3d), a 2xANOVA showed main effects of sex (F (1, 54) = 24.46, ****P < 0.0001) and EtOH concentration (F (2, 54) = 25.23, ****P < 0.0001) but no interaction (P = 0.154). Post hoc t-tests with H–S corrections showed that total consumption for the first 2 h on Day 4 was higher in females than males for 10% (t54 = 4.42, ***P = 0.0001) and 20% EtOH (t54 = 2.39, *P = 0.04, with a trend towards a sex difference for H2O DID mice (P = $0.085). Further, total volume consumed in 10% EtOH DID mice was higher than in 20% EtOH and H2O DID for females (***Ps ≤ 0.0001, as indicated) and than 20% DID mice for males (**P = 0.004; all other Ps > 0.10). For the 4-h time point (Fig. 3e), a 2xANOVA showed main effects of sex (F (1, 54) = 39.94, ****P < 0.0001) and EtOH concentration (F (2, 54) = 12.71, ****P < 0.0001) but no interaction (P = 0.675). Post hoc t-tests with H–S corrections showed that total consumption for the first 4 h on Day 4 was higher in females than males for all three groups (**Ps < 0.01, as indicated). Further, total volume consumed in 20% EtOH DID mice was lower than in 10% EtOH and H2O DID for females (**Ps < 0.01, as indicated) and than in H2O DID mice for males (*P = 0.035), with a trend towards lower consumption than 10% EtOH ($P = 0.071). For the 24-h time point (Fig. 3f), a 2xANOVA showed a main effect of sex (F (1, 54) = 87.29, ****P < 0.0001) but not EtOH concentration or interaction (Ps > 0.50). Post hoc t-tests with H–S corrections showed that total consumption for the first 2 h on Day 4 was higher in F than M for all three groups (****Ps < 0.0001). Thus, total volume intake converged across the last 20 h of access for all groups, leading to similar overall volume intake across DID conditions that was higher in females than males.

Because females’ greater alcohol consumption was apparent at 2 h while sex differences in water consumption did not emerge until later timepoints, we investigated the patterns of EtOH and H2O consumption (Fig. 4). We found that mice of both sexes drank more alcohol in the first two hours than the second two hours, but this was a more prominent effect in 10% compared to 20% EtOH. In 20% EtOH DID females (Fig. 4a), a 2xRM-ANOVA showed main effects of the first vs. second 2 h (“time”; F (1, 9) = 27.25, ***P = 0.0005) and DID cycle (F (7, 63) = 2.97, **P = 0.009), as well as an interaction (F (7, 63) = 2.41, *P = 0.030). Post hoc t-tests with H–S corrections showed that consumption was higher on the first two compared to the second 2 h on cycles 2, 4, 5, 6, and 7 (**Ps < 0.01, as indicated) but not cycles 1, 3, or 8 (Ps > 0.15). We observed a similar, but more robust, pattern for 10% EtOH DID females (Fig. 4b). A 2xRM-ANOVA showed a main effect of time (F (1, 9) = 226.8, ****P < 0.0001) and DID cycle (F (7, 63) = 2.39, *P = 0.031) but no interaction (P = 0.149). Post hoc t-tests with H–S corrections showed that consumption was higher in the first two compared to the second 2 h on all cycles (***Ps < 0.001, as indicated). In contrast, water consumption for 20% and 10% EtOH females did not differ between the first vs. second 2 h of access on Day 4 or DID cycle, as 2xRM-ANOVAs showed no effects or interactions (all Ps > 0.10; Fig. 4c, d). In H2O DID females (Fig. 4e), a 2xRM-ANOVAs showed an effect of cycle (F (7, 63) = 5.33, ****P < 0.0001) and a trend for time (F (7, 63) = 5.33, $P = 0.093) but no interaction (P > 0.20). These results suggest that females specifically frontload their consumption for EtOH, particularly at lower concentrations. While there was a trend for this front-loading behavior in water for females, it was less robust than alcohol consumption, likely because of the highly variable levels of water consumed. In concert with our results showing that total volume consumption is highest in 10% EtOH DID females in the first 2 h and then equilibrates by 24 h (Fig. 3d–f), alcohol frontloading may be performed to achieve similar levels to 20% EtOH consumption. Males also displayed frontloading behavior for EtOH, especially 10%, however, they also displayed a mild frontloading of H2O. In 20% EtOH DID males (Fig. 4f), a 2xRM-ANOVA showed main effects of time (F (1, 9) = 27.68, ***P = 0.0005) but not DID cycle or interaction (Ps > 0.15) on EtOH consumption. Post hoc t-tests with H–S corrections showed that consumption was higher on the first two compared to the second 2 h on cycles 2, 3, 4, and 6 (*Ps < 0.05, as indicated) but not cycles 1, 5, 7, or 8 (Ps > 0.15). As in females, we observed a more robust pattern for 10% EtOH DID than 20% DID males (Fig. 4g). A 2xRM-ANOVA for EtOH consumption in 10% EtOH DID males showed a main effect of time (F (1, 9) = 79.30, ****P < 0.0001) but no effect of DID cycle or interaction (Ps > 0.20). Post hoc t-tests with H–S corrections showed that consumption was higher in the first two compared to the second 2 h on all cycles (**Ps < 0.01, as indicated). In contrast to females that only significantly frontloaded EtOH, males’ water consumption was modestly higher in the first 2 h compared to the second 2 h in 10% EtOH DID and H2O DID (but not 20% EtOH DID). A 2xRM-ANOVA for water consumption on 20% DID in males showed a main effect of cycle (F (7, 63) = 4.36, ***P = 0.0005) but no effect of time or interaction (Ps > 0.65; Fig. 4h). However, for 10% EtOH DID (Fig. 4i), there was a main of time (F (1, 9) = 5.13, *P = 0.0498) but not cycle or interaction (Ps > 0.05). Post hoc t-tests with H–S corrections showed that consumption trended toward being higher on the first two compared to the second 2 h on cycles 1 and 3 (Ps < 0.10) but no other cycles (Ps > 0.35). Similarly, for H2O DID (Fig. 4j), there was a main effect of time (F (1, 9) = 6.08, *P = 0.036) but not cycle or interaction (Ps > 0.30). Post hoc t-tests with H–S corrections showed no differences in the first two versus second 2 h at any DID cycle (Ps > 0.10).

Fig. 4

Both sexes display frontloading behavior in alcohol but not water consumption. A-E) EtOH and water consumption in females during the first vs. second 2 h of Day 4 for 20% EtOH, 10% EtOH, and H2O DID. A Females consumed more 20% EtOH (g/kg) during the first vs. second 2 h of Day 4 DID on cycles 2 and 4–7. B Females consumed more 10% EtOH (g/kg) during the first vs. second 2 h of Day 4 DID on all cycles. C–E There was no difference in female H2O consumption (ml/kg) during the first vs. second 2 h in 20% EtOH DID (C), 10% EtOH DID (D), or H2O DID (E). F–J EtOH and water consumption in males during the first vs. second 2 h of Day 4 for 20% EtOH, 10% EtOH, and H2O DID. F Males consumed more 20% EtOH (g/kg) during the first vs. second 2 h of Day 4 DID on cycles 2–4 and 6. G) Males consumed more 10% EtOH (g/kg) during the first vs. second 2 h of Day 4 DID on all cycles. H–J There was no difference in male H2O consumption (ml/kg) during the first vs. second 2 h in 20% EtOH DID (H), but consumption was higher on the first vs. second 2 h for 10% EtOH DID (I) and H2O DID (J). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in 2xRM-ANOVA main effects and interactions of time point and DID cycle, as well as post hoc t-tests with H–S corrections between the first and second 2 h

To dissect the relationship between alcohol and water consumption in males and females, we performed correlational analyses at the 2, 4, and 24-h timepoints (Fig. 5). We found that total volume intake was positively correlated with alcohol consumption in the 20% EtOH female and 10% EtOH male groups at the 2-h (Fig. 5a; 20% EtOH F: R = 0.89, ***P = 0.0006; 10% EtOH M: R = 0.90, ***P = 0.0004; all other P > 0.14) and 24-h (Fig. 5c; 20% EtOH F: R = 0.69, *P = 0.0271; 10% EtOH M: R = 0.81, **P = 0.0048; all other P > 0.15) timepoints. At the 4-h (Fig. 5b) timepoints there was a positive correlation between alcohol consumption and total volume intake for all groups (20% EtOH F: R = 0.79, **P = 0.0061; 20% EtOH M: R = 0.65, *P = 0.0440; 10% EtOH F: R = 0.66, *P = 0.0366; 10% EtOH M: R = 0.76, *P = 0.0109). There was a trend towards a negative correlation between alcohol and water consumption in the 10% EtOH male cohort at 4 h (Fig. 5b; R = − 0.63, $P = 0.051) while there was a negative correlation at the 24-h timepoint (Fig. 5c) in this group with no significant correlations in the other groups (R = − 0.79, **P = 0.007; all other P > 0.15). Alcohol consumption and EtOH preference were positively correlated in the 4 (Fig. 5b; R = 0.75, *P = 0.0132) and 24-h (Fig. 5c; R = 0.83, **P = 0.0027) 10% EtOH males only, with a trend at 2-h (Fig. 5a; R = 0.62, $P = 0.056) for this group and in the 20% EtOH males at 24-h (R = 0.62, $P = 0.058; all other P > 0.18). These results suggest that males’ alcohol consumption was overall related to higher preference, while females’ alcohol consumption was unrelated to their water consumption at all time points.

Fig. 5

Alcohol consumption is correlated with water consumption and EtOH preference in males but not females. A–C R values reported from linear regressions for correlations between alcohol consumption versus total volume intake, water consumption, and EtOH preference. A At the 2-h timepoint, there was a positive correlation between alcohol consumption and total volume in the 10% EtOH males and 20% EtOH females, and a trend towards a positive correlation between alcohol consumption and preference in the 10% EtOH males. B At the 4-h timepoint, alcohol consumption was positively correlated with total volume intake for both sexes in both the 20% and 10% EtOH cohorts, while there was a trend towards a negative correlation between alcohol and water consumption in the 10% EtOH male group and positive correlation between alcohol consumption and EtOH preference in this group only. C At the 24-h timepoint, alcohol consumption and total volume intake was positively correlated in the 20% EtOH female and 10% EtOH male groups. Alcohol and water consumption were negatively correlated while alcohol consumption and EtOH preference were positively correlated in the 10% EtOH male cohort only, with a trend towards a positive correlation between alcohol and EtOH preference in the 20% EtOH male group. *P < 0.05, **P < 0.01, ***P < 0.001 for correlations between alcohol consumption and total volume intake, EtOH preference, or water consumed at 2, 4 and 24-h timepoints

No effect of chronic alcohol drinking on reward sensitivity

Following 20% EtOH or H2O DID, mice underwent a battery of behavioral assays to assess the effects of chronic binge alcohol drinking, beginning with sucrose preference starting 72 h after the last EtOH/water exposure (Fig. 6). Mice underwent 3 days of 1% sucrose preference, followed by 3 days of 2% sucrose preference, and we observed that females consumed more sucrose but not water than males at both doses, but EtOH drinking did not affect this behavior. A 3xRM-ANOVA on sucrose and water consumption for the 1% sucrose preference test revealed main effects of substance (F (1, 29) = 609.8, ****P < 0.0001) and sex (F (1, 29) = 48.38, ****P < 0.0001), as well as an interaction between the two (F (1, 29) = 20.03, ****P < 0.0001); however, there was no effect of or interactions involving EtOH as a variable (Ps > 0.20; Fig. 6b). To follow up on the substance x sex interaction, we performed 2xANOVAs within each substance. For sucrose, we found an effect of sex (F (1, 29) = 41.50, P < 0.0001) but no effect of EtOH or interaction between the two (Ps > 0.15). Post hoc t-tests with H–S corrections showed that CON and EtOH females consumed more sucrose than their male counterparts (t29 = 5.80, ****P < 0.0001; t29 = 3.43, **P = 0.002, respectively). We found similar results for 2% sucrose (Fig. 6c). A 3xRM-ANOVA showed main effects of substance (F (1, 35) = 931.1, P < 0.0001) and sex (F (1, 35) = 65.54, ****P < 0.0001), as well as an interaction between the two (F (1, 35) = 50.56, ****P < 0.0001); however, there was no effect of or interactions involving EtOH as a variable (Ps > 0.40). To follow up on the substance x sex interaction, 2xANOVAs within each substance were performed. For sucrose, we found an effect of sex (F (1, 35) = 60.76, P < 0.0001) but no effect of EtOH or interaction between the two (Ps > 0.55). Post hoc t-tests with H–S corrections showed that CON and EtOH F consumed more sucrose than their male counterparts (t35 = 5.86, ****P < 0.0001; t35 = 5.18, ****P < 0.0001, respectively). For water, we found no effects or interaction (Ps > 0.10). Analysis of preference for 1% sucrose over water showed no effects of sex or EtOH or an interaction between these variables (Ps > 0.25; Fig. 6d). Analysis of preference for 2% sucrose over water showed no effects of sex, EtOH, or an interaction between these variables (Ps > 0.10; Fig. 6e). Altogether, these results suggest that while females had higher consumption of sucrose compared to males, both sexes showed a high preference for sucrose at both concentrations, and these measures were not affected by a history of EtOH drinking.

Fig. 6

Females display higher sucrose consumption than males regardless of alcohol history. A Experimental timeline. B, C Sucrose and water consumption during the sucrose preference test, beginning 72 h after the last EtOH/H2O drinking session, using 1% sucrose (B) and 2% sucrose (C), showing that females consumed more sucrose, but not water, than males, with no effect of EtOH drinking history. D, E Sucrose preference was similar between H2O DID CON mice and 20% EtOH DID groups and between males and females for 1% (D) and 2% (E) sucrose. **P < 0.01, ***P < 0.001, ****P < 0.0001 in 3xRM-ANOVA main effects and interactions of substance, sex, and EtOH, as well as post hoc t-tests with H–S corrections as indicated

Behavioral disinhibition following a history of binge alcohol drinking was especially robust in females

We next assessed avoidance behavior across assays performed 2–4 weeks post-EtOH/H2O DID. We found that across behavioral tests, females displayed a reduction in avoidance behavior (Fig. 7). In the open field test (OF), 20% EtOH females, but not males, spent more time in the center of the OF than their H2O DID control (CON) counterparts (Fig. 7a, b). A 2xANOVA on the % time spent in the center showed a main effect of EtOH (F (1, 36) = 4.95, *P = 0.033) but no effect of sex or interaction (Ps > 0.15). Post hoc t-tests with H–S corrections showed that EtOH mice entered the center more quickly than CON mice in females but not males (F: t36 = 2.59, *P = 0.027; M: t36 = 0.56, $P = 0.582). Examination of the % time in the center of the OF across the 30 min assay confirmed this effect (Fig. 7b). A 3xRM-ANOVA on the % time spent in the center of the open field revealed main effects of EtOH (F (1, 216) = 13.44, ***P = 0.0003) and time (F (5, 216) = 6.71, ****P < 0.0001), but no effect of sex (P > 0.40). However, there was a sex x EtOH interaction (F (1, 216) = 5.64, *P = 0.018) but no other interaction (Ps > 0.25). We followed up using 2xRM-ANOVAs within sex. In females, there were main effects of EtOH (F (1, 18) = 8.10, *P = 0.011) and time (F (5, 90) = 5.48, ***P = 0.0002), as well as an interaction between the two (F (5, 90) = 3.04, *P = 0.014). Post hoc t-tests with H–S corrections showed that EtOH females spent more time in the center in the 20–25 min time bin (t108 = 4.33, ***P = 0.002) and a trend in the 15–20 min time bin (t108 = 2.59, $P = 0.053). In males, there was a main effect of time (F (5, 90) = 5.60, ***P = 0.0002) but no effect of EtOH or interaction (Ps > 0.60). When the % time spent in the center of the OF was averaged across the 30 min test, a 2xANOVA showed a main effect of EtOH (F (1, 36) = 4.95, *P = 0.033) but no effect of sex or interaction (Ps > 0.15). Post hoc t-tests with H–S corrections showed that EtOH F spent more time in the center than CON F (t36 = 2.59, *P = 0.027) but there was no difference between CON and EtOH males (t36 = 0.56, P = 0.582). A 2xANOVA on the latency to enter the center of the OF showed a main effect of EtOH (F (1, 36) = 16.50, ***P = 0.0003) but no effect of sex or interaction (Ps > 0.20). Post hoc t-tests with H–S corrections showed that EtOH mice entered the center more quickly than CON mice in both sexes (F: t36 = 3.62, **P = 0.002; M: t36 = 2.12, *P = 0.041; Fig. 7c). A 2xANOVA on the total distance traveled during the assay showed no effects of EtOH or sex or interaction (Ps > 0.75; Fig. 7d), suggesting that differences in avoidance behavior were not due to locomotor effects.

Fig. 7

Females display reduced avoidance and increased compulsive behavior during protracted abstinence (2–4 weeks) from chronic binge alcohol drinking. A–D Open field test (OF). A EtOH females, but not males, spent a greater % time in the center of the OF compared to CONs. B This effect emerged across the 30 min assay. C EtOH mice of both sexes had a shorter latency to enter the center of the OF. D There was no effect of sex or alcohol history on locomotion in the OF. E, F Elevated plus maze (EPM). E EtOH females, but not males, spent a greater % time in the open arms of the EPM compared to CONs. F There was no effect of sex or alcohol history on total distance traveled (m) in the EPM. G–H Light–dark box (LDB). G There was an interaction between sex and alcohol history for % time spent in the light side of the LDB but no differences in direct comparisons. H There was no effect of alcohol history or sex on locomotion as measured by the number of dark side entries. I, J Novelty-suppressed feeding (NSF). I EtOH females, but not males, had a reduced latency to eat the fruit loop compared to CONs. In addition, males had shorter latencies than females. J There was no effect of sex or alcohol history on post-test home cage food consumption. K Marble burying (MB). EtOH females, but not males, buried more marbles than CONs. In addition, males buried more marbles than females. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in 2xANOVA main effects of and interactions between sex and EtOH, 3xRM-ANOVA main effects of and interactions between sex, EtOH, and time, and post hoc t-tests with H–S corrections as indicated. $P < 0.10 for post hoc t-tests with H–S corrections between M and F

In the elevated plus maze (EPM), a 2xANOVA on the % time spent in the open arms showed a main effect of EtOH (F (1, 16) = 12.05, **P = 0.003), a trending effect of sex (F (1, 16) = 4.17, $P = 0.058), and a trend for interaction (F (1, 16) = 3.21, $P = 0.092). Post hoc t-tests with H–S corrections showed that EtOH females spent more time in the open arms of the EPM than CON females (t16 = 3.72, **P = 0.004) but there was no difference between CON and EtOH males (t16 = 1.19, P = 0.252; Fig. 7e). A 2xANOVA on the total distance traveled showed no effects of sex or EtOH or interaction between the two (Ps > 0.25; Fig. 7f), suggesting no differences in locomotion. In the light–dark box (LDB), a 2xANOVA on the % time spent in the light side showed no effect of sex or EtOH (Ps > 0.40) but there was an interaction between the two (F (1, 36) = 5.65, *P = 0.023). Post hoc t-tests with H–S corrections showed no effects in direct comparisons (Ps > 0.05; Fig. 7g). A 2xANOVA on the number of entries to the dark side, as a measure of locomotion, showed a trending effect of sex (F (1, 36) = 3.98, P = 0.054) but no effect of EtOH or interaction between the two (Ps > 0.15; Fig. 7h). In the novelty-suppressed feeding assay (NSF), a 2xANOVA on the latency to eat the fruit loop in the novel environment revealed an effect of sex (F (1, 36) = 8.90, **P = 0.005) but not EtOH (P > 0.10), and there was an interaction between the two (F (1, 36) = 19.06, ***P = 0.0001). Post hoc t-tests with H–S corrections showed that CON females had longer latencies than EtOH females (t36 = 4.24, ***P = 0.0003) and CON males (t36 = 5.20, ****P < 0.0001) but there was no difference between CON males and EtOH males (P > 0.30; Fig. 7i). A 2xANOVA on the post-test home cage chow consumption showed no effects of sex or EtOH or interaction between the two (Ps > 0.30; Fig. 7j), suggesting there were no differences in hunger driving effects of sex and EtOH on NSF. Altogether, results from avoidance behavior assays suggest that a history of binge alcohol drinking in females, but not males, leads to a decrease in behavioral inhibition. We also examined marble burying behavior, finding that EtOH females also display increased compulsive-like behavior (Fig. 7k), as a 2xANOVA on the proportion of marbles buried showed effects of sex (F (1, 36) = 13.46, ***P = 0.001) and EtOH (F (1, 36) = 5.81, P = 0.021) and an interaction between the two (F (1, 36) = 9.24, **P = 0.004). Post hoc t-tests with H–S corrections showed that CON females buried fewer marbles than EtOH females (t36 = 3.85, ***P = 0.001) and H2O M (t36 = 5.20, ****P < 0.0001) but there was no difference between CON and EtOH males (P > 0.30).

Effects of sex, but not EtOH drinking, on coping strategy

Finally, we tested mice in the forced swim test (FST), tail suspension (TS), and paw withdrawal assay for pain 4–6 weeks following EtOH/H2O exposure (Fig. 8). We found that females displayed more active coping strategies, as a 2xANOVA on the % time immobile in the FST showed an effect of sex (F (1, 36) = 7.79, **P = 0.008) but no effect of EtOH or interaction between the two (Ps > 0.45). Post hoc t-tests with H–S corrections showed trends for greater immobility in males than females in CON and EtOH groups (CON: t36 = 1.93, $P = 0.099; EtOH: t36 = 2.02, $P = 0.099; Fig. 8a). A 2xANOVA on the % time swimming showed an effect of sex (F (1, 36) = 8.48, **P = 0.006) but no effect of EtOH or interaction between the two (Ps > 0.45). Post hoc t-tests with H–S corrections showed trends for greater swimming in female than male in CON and EtOH groups (CON: t36 = 1.95, P = 0.072; EtOH: t36 = 2.17, P = 0.072; Fig. 8b). A 2xANOVA on the % time climbing showed no effects of sex or EtOH or interaction between the two (Ps > 0.65; Fig. 8c). We also found more active coping in females than males on the TST, as a 2xANOVA on the % time passive showed an effect of sex (F (1, 36) = 9.09, **P = 0.005) but no effect of EtOH or interaction between the two (Ps > 0.15). Post hoc t-tests with H–S corrections showed that CON females displayed less time passive than CON males (t36 = 3.12, **P = 0.007) but there was no difference between EtOH F and EtOH M (P > 0.25; Fig. 8d). These results showed that females displayed more active coping while males showed more passive coping, but there were no effects of EtOH on this behavior. Finally, we measured thermal pain sensitivity, finding no effect of EtOH or sex at either low- or high-temperature stimuli (Fig. 8e). A 2xANOVA on paw withdrawal latency for 52 °C showed no effects of sex or EtOH or interaction between the two (Ps > 0.65). Similarly, at 58