Experiment 1 – Model of binge-type eating Intake during a 2-h period

In the Binge group, there was significantly greater consumption of Ensure than chow during the Ensure access period (F1, 32 = 531.96, p < 0.001, ηp2 = 0.94) (data not shown). Males demonstrated an 87.8% preference and females demonstrated a 90.2% preference for Ensure. Due to this high preference, further analyses of intake in the Binge group during the 2-h Ensure access session examined only Ensure consumption.

Examining 2-h intake on days with Ensure access, comparison of Ensure consumption in the Binge group to chow intake at the same time in the Control group revealed a main effect of access day (F11.16, 412.94 = 11.06, p < 0.001, ηp2 = 0.23), group (F1, 37 = 427.97, p < 0.001, ηp2 = 0.92), and sex (F1, 37 = 31.20, p < 0.001, ηp2 = 0.46), as well as an interaction effect between access day and group (F11.16, 412.94 = 4.03, p < 0.001, ηp2 = 0.10) (Fig. 2a). Pairwise comparisons between days showed that both male and female mice in the Binge group increased their intake over the course of the experiment, having significantly higher intake of Ensure calories per body weight on the final day than on the first day (p < 0.001). Regression analysis of Ensure intake in males revealed a significant linear model (F1, 23 = 46.01, p < 0.001), with an R2 of 0.68, indicating that access day explained approximately 68% of the variance in Ensure consumption (data not shown). The regression equation was: Ensure consumption = 13.37 + 0.22(access day), confirming that there was an escalation of Ensure intake over time for males. Regression analysis of Ensure intake in females also revealed a significant linear model (F1, 23 = 31.48, p < 0.001), with an R2 of 0.59, indicating that access day explained 59% of the variance in Ensure consumption (data not shown). The regression equation for females was: Ensure consumption = 15.94 + 0.26(access day), confirming that there was also an escalation of Ensure intake over time for females. In contrast, in the Control group, there was no difference in intake on the first day compared to the final day (p = 0.650). Mice in the Binge group consumed significantly more Ensure during the 2-h session compared to mice in the Control group during the same 2-h session (p < 0.001), and pairwise comparisons between sexes revealed that females compared to males in the Binge group consumed significantly more Ensure per body weight (p < 0.001), and females compared to males in the Control group consumed significantly more chow per body weight (p = 0.003) (Fig. 2 (a, inset)). Examining binge scores also revealed a significant main effect of access day (F7.66, 168.51 = 15.41, p < 0.001, ηp2 = 0.41) and sex (F1, 22 = 8.20, p = 0.009, ηp2 = 0.27), as well an interaction effect between access day and sex (F7.66, 168.51 = 3.64, p < 0.001, ηp2 = 0.14). Pairwise comparisons between days showed that binge score increased over the course of the experiment (p < 0.001) (Fig. 2b). Females demonstrated significantly greater binge scores than males (p = 0.009) (Fig. 2 (b, inset)). These results suggest that on days of access to Ensure, both male and female mice engage in binge-type eating, and that females binge to a greater extent than males.

Fig. 2

Model of binge-type eating (Experiment 1). a For 2-h intake on days with Ensure access, the Binge group consumed more calories per body weight than the Control group, with Binge intake increasing over the course of the experiment, and females in both the Binge and Control groups consuming more than males in the respective groups. Inset is average 2-h intake for each subject across all access days. b Binge scores increased over the course of the experiment, and females had higher binge scores than males. Inset is average binge score for each subject across all access days. c For 2-h chow intake on days without Ensure access, the Binge group consumed more calories of chow per body weight than the Control group, and females in the Binge group consumed significantly more calories per body weight than females in the Control group. Inset is average 2-h chow intake for each subject across all days Ensure access. d For 2-h intake across both days with Ensure access and days without Ensure access, intake varied in the Binge group depending on access. e For 24-h chow intake on days with Ensure access, the Binge group consumed fewer calories of chow per body weight than the Control group, and females consumed more than males. Inset is average 24-h chow intake for each subject across all access days. f For 24-h chow intake on days without Ensure access, females consumed significantly more calories per body weight than males. Inset is average 24-h chow intake for each subject across all days without Ensure access. g For 24-h calories consumed on days of Ensure access, females consumed significantly more total calories per body weight than males. Inset is average 24-h calories consumed for each subject across all days of Ensure access. h Body weight increased over the course of the experiment, with females gaining more weight relative to their initial body weight than males. Values are Mean ± S.E.M. $$$p < 0.001 vs Day 1, ***p < 0.001, **p < 0.01, *p < 0.05

Examining 2-h intake on days without Ensure access also revealed a main effect of day (F17, 493 = 2.97, p < 0.001, ηp2 = 0.09), but there was no discernable pattern of increase or decrease across time (Fig. 2c, d). There was a main effect of group (F1, 29 = 9.72, p = 0.004, ηp2 = 0.25), with the Binge group consuming significantly more calories of chow per body weight than the Control group (p = 0.004). There was no main effect of sex (F1, 29 = 0.93, p = 0.342, ηp2 = 0.03) and no interaction effect between sex and group (F1, 29 = 1.53, p = 0.227, ηp2 = 0.05); however, comparisons of simple main effects revealed that, in females, the Binge group consumed significantly more per body weight during the 2-h session on non-access days than the Control group (p = 0.004). This effect was not found in males (p = 0.197) (Fig. 2 (d, inset)). These results suggest that, on days without Ensure access, Binge female but not Binge male mice continue to engage in greater 2-h intake than Controls.

Intake during a 24-h period

Examining 24-h chow intake on days with Ensure access revealed a main effect of access day (F9.72, 389.15 = 2.16, p = 0.021, ηp2 = 0.05); however, there was again no discernable pattern of intake. There was also a main effect of group (F1, 40 = 32.62, p < 0.001, ηp2 = 0.45) and a main effect of sex (F1, 40 = 7.88, p = 0.008, ηp2 = 0.17), but no significant interaction effect between group and sex (F1, 40 = 53.00, p = 0.091, ηp2 = 0.07) (Fig. 2e). Females consumed more calories from chow per body weight than males (p = 0.008), and mice in the Binge group consumed significantly fewer calories from chow than mice in the Control group (p < 0.001) (Fig. 2 (e, inset)).

Examining 24-h chow intake on days without Ensure access, there was a main effect of day (F17, 306 = 1.98, p = 0.012, ηp2 = 0.10) and an interaction effect between day and group (F17, 306 = 1.78, p = 0.030, ηp2 = 0.09); however, there was no discernable pattern of increase or decrease over time. There was again a main effect of sex (F1, 18 = 5.43, p < 0.001, ηp2 = 0.64), with females consuming significantly more calories per body weight from chow than males (p < 0.001). There was no main effect of group (F1, 18 = 0.09, p = 0.767, ηp2 = 0.01), although there was a significant interaction effect between group and sex (F1, 18 = 5.43, p = 0.032, ηp2 = 0.23) (Fig. 2f). Pairwise comparisons indicated that females in the Binge group consumed more chow than males in the Binge group (p < 0.001), and females in the Control group consumed more chow than males in the Control group (p = 0.028) (Fig. 2 (f, inset)). These chow results together indicate that mice in the Binge group may have reduced chow intake on days of Ensure access, but they consume the same amount as those in the Control group on days without Ensure access.

Examining 24-h calories consumed (inclusive of both chow and Ensure, when relevant) on days of Ensure access, there was a main effect of sex (F1, 40 = 12.13, p < 0.001, ηp2 = 0.23), with females consuming significantly more total calories per body weight than males (Fig. 2 (g, inset)). There was no main effect of access day (F9.85, 393.82 = 1.11, p = 0.324, ηp2 = 0.03) and no main effect of group (F1, 40 = 1.57, p = 0.218, ηp2 = 0.04) (Fig. 2g). These results suggest that, although mice in the Binge group consume significantly more calories during their binge sessions, this does not result in general overeating.

Body weight

Examining change in bodyweight over the course of the experiment, there was a main effect of week (F2.99, 119.92 = 42.26, p < 0.001, ηp2 = 0.51), with weight increasing over the course of the experiment. There was a trend for a main effect of group (F1, 40 = 3.63, p = 0.064, ηp2 = 0.08) and a significant main effect of sex (F1, 40 = 8.47, p = 0.006, ηp2 = 0.18), but there was no interaction effect between group and sex (F1, 40 = 0.01, p = 0.918, ηp2 = 0.00). The animals in the Control group showed a trend for more weight gain than animals in the Binge group (p = 0.064), and females gained more weight relative to their initial body weight than males (p = 0.006) (Fig. 2h). These results suggest that, despite their greater consumption of calories in the 2-h sessions, the Binge group did not gain significantly more weight than the Control group.

Experiment 2 – PACAP gene expression in the PVT relative to binge-type eating

Examining levels of PACAP mRNA in the PVT, results revealed no main effect of group (F2, 30 = 2.16, p = 0.133, ηp2 = 0.13) or sex (F1, 30 = 0.09, p = 0.762, ηp2 = 0.00), but there was a significant interaction effect between group and sex (F2, 30 = 7.08, p = 0.003, ηp2 = 0.32) (Fig. 3a). Pairwise comparisons revealed that Control females had significantly higher levels of PACAP than Control males (p = 0.014), but that non-binge day males had significantly higher levels than non-binge day females (p = 0.012), while there was no significant difference between the binge day groups (p = 0.641). Levels of PACAP in the non-binge day males were not significantly different than Control females (t10 = 0.44, p = 0.666). For males, pairwise comparisons indicated that the non-binge day group had significantly higher levels of PACAP than the Control group (p = 0.007) and that the binge day group had significantly lower levels of PACAP than the non-binge day group (p = 0.045) (Fig. 3b). In line with this, while the correlation for all male groups between average 2-h intake and level of PACAP mRNA did not reach significance (r = 0.40, p = 0.098), the intake-PACAP correlation when including only the Control and non-binge day groups was significant and positive (r = 0.64, p = 0.025) (data not shown). For females, pairwise comparisons showed a trend for lower levels of PACAP in the binge day group compared to the Control group (p = 0.078), but no significant differences between the non-binge day and Control group (p = 0.166) or the binge day and non-binge day group (p = 0.977) (Fig. 3c). In females, the correlation for all groups between average 2-h intake and level of PACAP mRNA was significant and negative (r = -0.60, p = 0.008), with a significant negative intake-PACAP correlation being found when including only the Control and binge day groups (r = -0.67, p = 0.018) (data not shown). These results indicate that, in males but not females, a history of exposure to Ensure under a binge-type eating paradigm elevates levels of PACAP, and that these levels are decreased immediately prior to an Ensure access session.

Fig. 3

Gene expression of PACAP in the PVT relative to binge-type eating (Experiment 2). a Between the Control groups, females had significantly higher levels of PACAP mRNA than males. Non-binge day males had significantly higher levels of PACAP mRNA than non-binge day females. Males in the non-binge day group had levels of PACAP mRNA that were statistically no different than Control females. b Among males, the non-binge day group had significantly higher levels of PACAP mRNA than the Control group, but the binge day group had significantly lower levels of PACAP mRNA than the non-binge day group. c Among females, there were no significant differences between groups. Values are Mean ± S.E.M. **p < 0.01, *p < 0.05

Experiment 3 – Effects of PACAP+ cell activity in the PVT on binge-type eating Validation of PACAP-Cre mice

In PACAP-Cre transgenic mice, we observed a high degree of co-localization between PACAP and Cre labeling in the PVT (Fig. 4a). Our analysis showed good penetrance of PACAP+ cells (82%), as well as good fidelity (80%). In mice that received the excitatory (Gq) DREADD in the PVT, systemic injection of CNO compared to saline resulted in a significant increase in c-Fos in the PVT (t4 = 2.88, p = 0.045, d = 2.35) (Fig. 4b and f), and 64% compared to 33% of Cre+ neurons co-labeled with c-Fos (Fig. 4g). Visual examination of tdTomato gene expression in a PACAP-Cre mouse with an Ai14(RCL-tdT) reporter in the Allen Mouse Brain Connectivity Atlas similarly demonstrated detectable levels of expression in the mouse PVT (Allen Institute for Brain Science 2004) (Fig. 4c). Histological analysis found that virus injections in this experiment were made into the middle and posterior subregions of the PVT, between bregma -0.94 and -1.70 mm (Fig. 4d and e). These results together confirm that PACAP and Cre are co-expressed in the PVT of PACAP-Cre transgenic mice, and that injection of CNO in PACAP-Cre mice that received the excitatory DREADD in the PVT resulted in excitation of PACAP+ cells in the PVT.

Fig. 4

Effects of PACAP+ cell activity in the PVT of PACAP-Cre transgenic mice on binge-type eating (Experiment 3). a Photomicrograph demonstrating co-labeling of PACAP (blue) and Cre (magenta) in the PVT of PACAP-Cre transgenic mice. b Photomicrograph demonstrating co-labeling of PACAP (blue), Cre (magenta), the Cre-dependent excitatory DREADD (pAAV8-hSyn-DIO-hM3D(Gq)-mCherry) (red), and c-Fos (green) in the PVT of PACAP-Cre transgenic mice after systemic injection with CNO. Inset is higher-magnification image of main image, and shows a cell co-labeled for PACAP, Cre, and AAV (right) and a cell co-labeled for PACAP, Cre, AAV, and c-Fos (left, indicated by “V”). Note that cell size appears to be similar to that in (Gao et al. 2023; Kooiker et al. 2023; Paniccia et al. 2024). Scale bars = 200 μm. c Expression of tdTomato in a PACAP-Cre mouse with an Ai14(RCL-tdT) reporter in the Allen Mouse Brain Connectivity Atlas demonstrating detectable levels of expression in the mouse PVT (Allen Institute for Brain Science 2004). d Injection track of the excitatory DREADD into the PVT. The AAV here is pseudo-colored magenta. Scale bar = 200 μm. e Histological analysis found that virus injections were made between -0.94 and -1.70 mm posterior to bregma, in the middle and posterior subregions of the PVT. Dots indicate placement of on-target injections. Blue = male, red = female. f In mice that received the excitatory (Gq) DREADD, systemic injection with CNO compared to saline resulted in a significant increase in c-Fos in the PVT. g In mice that received the excitatory (Gq) DREADD, systemic injection with CNO compared to saline increased c-Fos labeling in Cre.+ neurons in the PVT. h For effects of the excitatory (Gq) DREADD on binge-type eating of Ensure, females overall consumed more Ensure than males, but injection of CNO resulted in reduced consumption of Ensure in the overall Gq group, due to a specific decrease in consumption in the males. i For effects of the excitatory (Gq) DREADD on chow intake during the 2-h binge-type eating of Ensure, the overall Gq group ate less than the overall control AAV group. j For effects of the excitatory (Gq) DREADD on 24-h chow intake on the day of Ensure access, there were no significant differences between groups. Values are Mean ± S.E.M. **p < 0.01, *p < 0.05

Effects of PACAP+ cell activity in the PVT on food intake

Examining effects of the excitatory (Gq) DREADD on binge-type eating of Ensure, results showed that there was a significant main effect of sex (F1, 15 = 10.87, p = 0.005, ηp2 = 0.42) and a trend for a significant main effect of drug injection (F1, 15 = 4.35, p = 0.054, ηp2 = 0.23), and while there was no significant main effect of virus (F1, 15 = 1.06, p = 0.319, ηp2 = 0.07), there was a significant interaction effect between drug injection and virus (F1, 15 = 7.03, p = 0.018, ηp2 = 0.32) (Fig. 4h). With females overall consuming more Ensure per body weight than males (p = 0.005), pairwise comparisons between drug injections in males and females combined revealed that, in the Gq group but not in the control AAV group, injection of CNO compared to saline resulted in significantly reduced consumption of Ensure (p = 0.004 and p = 0.701, respectively). Moreover, tests of simple main effects revealed that this decrease in Ensure intake following CNO in the Gq group was due to a significant decrease in Ensure intake in males (p = 0.004), but that the change in eating in females did not reach significance (p = 0.127) (Fig. 4 (h, right)). In contrast, when examining 2-h chow intake during the same Ensure access sessions, there was no significant main effect of sex (F1, 14 = 2.39, p = 0.144, ηp2 = 0.22) or drug injection (F1, 14 = 0.56, p = 0.468, ηp2 = 0.12), although there was a significant main effect of virus (F1, 14 = 5.01, p = 0.042, ηp2 = 0.26), with the Gq group overall eating less chow than the control AAV group overall (Fig. 4i). Examining 24-h chow intake on the injection days, there was no significant main effect of sex (F1, 13 = 4.04, p = 0.066, ηp2 = 0.22), drug injection (F1, 13 = 0.243, p = 0.630, ηp2 = 0.02), or virus (F1, 13 = 0.030, p = 0.865, ηp2 = 0.01) (Fig. 4j). These results suggest that enhancing the activity of PACAP+ cells in the PVT decreases eating, particularly binge-type eating of Ensure, and that this specifically occurs in male mice.