1.1: The effects of chronic CBD administration during UCMS on depressive-like symptoms:
We used a 2×2×2 design with main factors of Sex, Stress (No UCMS/UCMS) and Drug (Vehicle/CBD) (see Fig. 1a for experimental design). In cases of a significant sex effect or three-way interaction, data from male and female rats were analyzed separately. See Table S3 for detailed analyses of three-way and two-way ANOVA.
In the FST (Fig. 1b), univariate ANOVA [sex×stress×drug (2×2×2)] revealed a significant effect of sex, drug, stress and the following interactions: stress×sex, sex×drug, and sex×stress×drug on immobility. No significant effect of stress×drug interaction was found. Two-way ANOVA [stress×drug (2×2)] revealed a significant effect of drug (males: F(1,39) = 19.412, p < 0.001; females: F(1,39) = 10.602, p < 0.01), stress (males: F(1,39) = 28.714, p < 0 < 001) and stressxdrug interaction (males: F(1,39) = 13.394, p < 0.01; females: F(1,39) = 7.199, p < 0,05). This suggests that in males, CBD restored the UCMS-induced increase in immobility, while in females, CBD restored the UCMS-induced decrease in immobility.
A significant effect of stress, drug and stress×drug interaction on swimming was found in males, and of drug×stress in females. This suggests that in males, UCMS decreased swimming time and that CBD prevented this effect. No effects on climbing were found (data not shown, available in Table S3).
In the OFT (Fig. 1c), univariate ANOVA revealed a significant effect on locomotion of sex, stress and the following interactions: sex × stress and sex × stress × drug. Two-way ANOVA revealed a significant effect of stress (males: F(1,39) = 242.585, p < 0.001; females: F(1,39) = 46.286, p < 0.001) and stress×drug interaction (males: F(1,39) = 6.143, p < 0.05; females: F(1,39) = 6.376, p < 0.05), with no effect of drug (males: F(1,39) = 0.001, ns; females: F(1,39) = 0.943, ns), suggesting that UCMS led to increased locomotion in both sexes, regardless of CBD treatment. Also, we found no significant effect of stress nor CBD on the time spent in the center of the arena during the first 5 minutes of the test in males and females (data not shown, available in Table S3), suggesting that UCMS did not cause anxiety-like behavior.
1.2: The effects of chronic CBD administration during UCMS on cannabinoid receptors, inflammatory markers, and estrogen receptor gene expression:
Raw ΔCt values for RT-PCR experiments are presented in Table S4.
1.2.1: cnr1:
In the vmPFC (Fig. 2a), univariate ANOVA [sex×stress×drug (2×2×2)] revealed significant effects of stress but not of the drug or any of the interactions. Two-way ANOVA [stress×drug (2×2)] revealed a significant effect of stress (males: F(1,35) = 17.396, p < 0.001; females: F(1,34) = 17.912, p < 0.001) but not of drug (males: F(1,35) = 0.011, ns; females: F(1,34) = 0.098, ns) or stress×drug interaction (males: F(1,35) = 0.072, ns; females: F(1,34) = 0.019, ns), suggesting that UCMS led to downregulation of cnr1 regardless of CBD treatment.
In the CA1 (Fig. 2b), univariate ANOVA revealed significant effects of sex, stress and drug. Two-way ANOVA revealed a significant effect of drug (F(1,36) = 9.757, p < 0.01) and stress (F(1,36) = 13.617, p < 0.001) in males but not in females (drug: (F(1,37) = 0.493, ns; stress: F(1,37) = 0.124, ns). This suggests that CBD treatment to UCMS males resulted in the upregulation of cnr1. Stress×drug interaction was not significant in males (F(1,36) = 0.019, ns), or females (F(1,37) = 0.001, ns).
In the VS (Fig. 2c), univariate ANOVA revealed a significant effect of sex. Two-way ANOVA revealed a significant effect of drug (F(1,30) = 8.394, p < 0.01) and stress×drug interaction (F(1,30) = 6.915, p < 0.05) in males but not in females (drug: (F(1,28) = 0.001, ns; stress×drug: F(1,28) = 0.001, ns). This suggests that UCMS downregulated cnr1 in male rats and that CBD prevented this effect. Stress was not significant in males (F(1,30) = 2.708, ns) or females (F(1,28) = 0.066, ns),
1.2.2: cnr2:
In the vmPFC (Fig. 2d), univariate ANOVA (2×2×2) revealed a significant effect of sex and drug but not of stress or any of the interactions. Two-way ANOVA (2×2) revealed a significant effect of drug (males: F(1,34) = 6.927, p < 0.05; females: F(1,32) = 6.706, p < 0.05) but not of stress (males: F(1,34) = 0.614, ns; females: F(1,32) = 0.233, ns) or stress×drug interaction (males: F(1,34) = 0.081, ns; females: F(1,32) = 0.454, ns), suggesting that CBD downregulated cnr2 in both sexes irrespective of UCMS.
In the CA1 (Fig. 2e), univariate ANOVA revealed a significant effect of sex but not of stress, drug, or any of the interactions. Two-way ANOVA revealed a significant effect of drug (F(1,32) = 22.341, p < 0.001) and stress (F(1,32) = 0.5.439, p < 0.05) in males, but not in females: (drug: F(1,34) = 0.110, ns; stress: F(1,34) = 0.001, ns), suggesting that CBD treatment led to upregulation of cnr2 in males. Stress×drug interaction was not significant in males (F(1,32) = 1.302, ns) or females (F(1,34) = 0.004, ns).
In the VS (Fig. 2f), univariate ANOVA revealed a significant effect of sex but not of stress, drug or any of the interactions. Two-way ANOVA revealed no significant effect of drug (males: F(1,30) = 0.208, ns; females: F(1,29) = 0.024, ns), stress (males: F(1,30) = 2.174, ns; females: F(1,29) = 0.039, ns) or stress×drug interaction (males: F(1,30) = 0.290, ns; females: F(1,29) = 0.093, ns), suggesting that in both sexes, neither UCMS nor CBD affected cnr2 expression.
1.2.3: tnf:
In the vmPFC (Fig. 3a), univariate ANOVA (2×2×2) revealed a significant effect of sex but not of stress, drug, or any of the interactions. Two-way ANOVA (2×2) revealed no significant effect of drug (males: F(1,36) = 1.060, ns; females: F(1,33) = 0.576, ns), stress (males: F(1,36) = 1.211, ns; females: F(1,33) = 0.243, ns) or stress×drug interaction (males: F(1,36) = 0.122, ns; females: F(1,33) = 1.894, ns).
In the CA1 (Fig. 3b), univariate ANOVA revealed a significant effect of sex and stress but not of drug, or any of the interactions. Two-way ANOVA revealed a significant effect of drug (F(1,36) = 6.488, p < 0.05), stress (F(1,36) = 5.668 p < 0.05) and stress×drug interaction (F(1,36) = 4.085, p < 0.05) in males, but not in females (drug: F(1,36) = 0.086, ns; stress: F(1,36) = 1.097, ns; stress×drug interaction: F(1,36) = 0.003, ns), suggesting that CBD prevented UCMS-induced upregulation of tnf in males.
In the VS (Fig. 3c), univariate ANOVA revealed a significant effect of sex and sex×stress ×drug interaction, but not of stress, drug, or other interactions. Two-way ANOVA revealed a significant effect of drug (F(1,29) = 4.592, p < 0.05) and stress×drug interaction (F(1,29) = 6.718, p < 0.05) in males but not in females (drug: F(1,28) = 0.388, ns; sex×drug: F(1,28) = 0.462, ns), suggesting that CBD prevented UCMS-induced upregulation of tnf in males. Stress was not significant in males (F(1,29) = 3.704, ns) or females (F(1,28) = 0.290, ns).
1.2.4: nfkb1:
In the vmPFC (Fig. 3d), univariate ANOVA (2×2×2) revealed a significant effect of sex and sex×drug interaction, but not of stress, drug, or other interactions. Two-way ANOVA (2×2) revealed a significant effect of drug in females (F(1,34) = 14.913, p < 0.001) but not in males (F(1,32) = 0.138, ns), with no effect of stress (males: F(1,32) = 1.565, ns; females: F(1,34) = 0.195, ns) or stress×drug interaction (males: F(1,32) = 1.264, ns; females: F(1,34) = 0.157, ns), suggesting that in females, CBD resulted in the downregulation of nfkb1 irrespective of UCMS.
In the CA1 (Fig. 3e), univariate ANOVA revealed a significant effect of sex, stress, drug and stress×drug interaction. Two-way ANOVA revealed a significant effect of stress (males: F(1,36) = 4.865, p < 0.05; females: F(1,35) = 5.283, p < 0.05). Drug was significant in males (F(1,36) = 4.735, p < 0.05) but not in females (F(1,35) = 0.966, ns). Stress×drug interaction was not significant in males (F(1,36) = 3.419, ns) or females (F(1,35) = 3.953, ns).
In the VS (Fig. 3f), univariate ANOVA revealed a significant effect of sex, stress and sex×stress interaction, but not of drug or other interactions. Two-way ANOVA revealed a significant effect of stress (males: F(1,31) = 44.963, p < 0.001; females: F(1,27) = 15.462, p < 0.001) but not of drug (males: F(1,31) = 0.430, ns; females: F(1,27) = 0.214, ns) and stress×drug interaction (males: F(1,31) = 1.547, ns; females: F(1,27) = 0.001, ns), suggesting that in both sexes, UCMS led to the upregulation of nfkb1 regardless of CBD treatment.
1.2.5: esr1:
In the vmPFC (Fig. 4a), univariate ANOVA (2×2×2) revealed no significant effect of sex, stress, drug, or any of the interactions.
Two-way ANOVA (2×2) revealed no significant effect of drug (males: F(1,35) = 0.031, ns; females: F(1,34) = 0.561, ns) stress (males: F(1,35) = 0.503, ns; females: F(1,34) = 0.762, ns) or stress×drug interaction (male: F(1,35) = 2.5601, ns; females: F(1,34) = 0.001, ns).
In the CA1 (Fig. 4b), univariate ANOVA revealed a significant effect of sex and stress but not of drug or any of the interactions. Two-way ANOVA revealed a significant effect of stress (males: F(1,35) = 5.650, p < 0.05; females: F(1,36) = 4.462, p < 0.05) but not drug (males: F(1,35) = 0.903, ns; females: F(1,36) = 0.270, ns) or stress×drug interaction (males: F(1,35) = 2.064, ns; females: F(1,36) = 1.744, ns), suggesting that in both sexes UCMS led to upregulation of est1 regardless of CBD treatment.
In the VS (Fig. 4c), univariate ANOVA revealed a significant effect of the following interactions: sex×stress, sex×drug, stress×drug and sex×stress×drug, but not of sex, stress and drug. Two-way ANOVA revealed a significant effect of drug in males (F(1,31) = 12.044, p < 0.01) and females (F(1,31) = 10.732, p < 0.01). Stress (F(1,31) = 6.010, p < 0.05) and stress×drug interaction (F(1,31) = 13.030, p < 0.01) were significant in males but not in females (stress: F(1,31) = 0.962, ns; stress×drug: F(1,31) = 0.275, ns), suggesting that in males CBD prevented the UCMS-induced downregulation of esr1.
1.2.6: esr2:
In the vmPFC (Fig. 4d), univariate ANOVA (2×2×2) revealed a significant effect of sex but not of stress, drug, or any of the interactions. Two-way ANOVA (2×2) revealed a significant effect of stress×drug interaction (F(1,35) = 4.821, p < 0.05) in males but not in females (F(1,35) = 0.088, ns). No significant effect was found of drug (males: F(1,35) = 0.004, ns; females: F(1,35) = 0.755, ns) or stress (males: F(1,35) = 2.008, ns; females: F(1,35) = 0.280, ns).
In the CA1 (Fig. 4e), univariate ANOVA revealed a significant effect of sex, stress and drug, but not any of the interactions. Two-way ANOVA revealed a significant effect of stress (males: F(1,35) = 6.485, p < 0.05; females: F(1,36) = 17.306, p < 0.001). Drug was significant in females (F(1,36) = 6.483, p < 0.05) but not in males (F(1,35) = 1.351, ns). Stress×drug interaction was not significant in males (F(1,35) = 0.023, ns) or females (F(1,36) = 0.611, ns), suggesting that UCMS led to upregulation of esr2 in both sexes.
In the VS (Fig. 4f), univariate ANOVA revealed a significant effect of sex, drug and the following interactions: sex×stress, sex×drug and stress×drug. Two-way ANOVA revealed a significant effect of drug (F(1,31) = 17.549, p < 0.001), stress (F(1,31) = 4.250, p < 0.05) and stress×drug interaction (F(1,31) = 12.1954, p < 0.01) in males, but not in females (drug F(1,31) = 0.115, ns; stress: F(1,31) = 3.955, ns; stress×drug interaction: F(1,31) = 1.687, ns), suggesting that in males CBD prevented the UCMS-induced downregulation of esr2.
The distribution of estrus phases in each group of female rats was observed on the first day of behavioral tests. A similar distribution of rats across the diestrus, proestrus, and estrus phases was noted within each group (see Table S5 for estrus phase distribution and Table S6 for correlations between estrus phase and behavioral phenotype).
To explore the association between depressive-like behavior and gene expression, Pearson bivariate correlation tests were conducted between the behavioral measurements and mRNA expression in the vmPFC, CA1, and VS in males (supplemental material, Table S7) and females (Table S8).
1.4: The effects of chronic CBD administration during UCMS on miRNA expression in male and female rats:
1.4.1: miR-9-5p:
In the vmPFC (Fig. 5a), univariate ANOVA (2×2×2) revealed a significant effect of sex but not of stress, drug, or any of the interactions.
Two-way ANOVA (2×2) revealed no significant effect of drug (males: F(1,31) = 0.068, ns; females: F(1,33) = 0.335, ns), stress (males: F(1,31) = 0.388, ns; females: F(1,33) = 1.564, ns) or stress×drug interaction (males: F(1,31) = 0.136, ns; females: F(1,33) = 0.071, ns).
In the CA1 (Fig. 5b), univariate ANOVA revealed a significant effect of sex and the following interactions: stress×drug and sex×stress×drug, but not of stress, drug, or any of the other interactions. Two-way ANOVA revealed a significant effect of stress×drug interaction (males: F(1,34) = 5.938, p < 0.05; females: F(1,34) = 17.087, p < 0.001). A significant effect of drug was found in males (F(1,34) = 5.753, p < 0.05) but not in females (F(1,34) = 1.026, ns). Stress was not significant in males (F(1,34) = 3.535, ns) or females (F(1,34) = 0.052, ns), suggesting that CBD treatment downregulated miR-9-5p in UCMS rats.
In the VS (Fig. 5c), univariate ANOVA revealed a significant effect of sex, stress and sex×stress interaction, but not of drug, or any of the other interactions. Two-way ANOVA revealed a significant effect of stress (F(1,34) = 24.911, p < 0.001) and stress×drug interaction (F(1,34) = 5.598, p < 0.05) in females, but not in males (stress: F(1,28) = 0.016, ns; stress ×drug: F(1,28) = 0.091, ns), suggesting that in females UCMS downregulated miR-9-5p. Drug was not significant in males (F(1,28) = 1.137, ns) or females (F(1,34) = 1.045, ns).
1.4.2: miR-98-5p:
In the vmPFC (Fig. 5d), univariate ANOVA (2×2×2) revealed a significant effect of sex, stress and the following interactions: stress×drug and sex×stress×drug but not of drug and any of the other interactions. Two-way ANOVA revealed a significant effect of stress (F(1,33) = 6.154, p < 0.05) and stress×drug interaction (F(1,33) = 24.91, p < 0.001) in females but not in males (stress: F(1,31) = 0.480, ns; stress×drug interaction: F(1,31) = 0.194, ns), suggesting CBD downregulated miR-98-5p in non-stressed females. Drug was not significant in both males (F(1,31) = 0.122, ns) and females (F(1,33) = 1.305, ns).
In the CA1 (Fig. 5e), univariate ANOVA revealed a significant effect of sex×stress×drug interaction, but not of stress, drug, or any of the other interactions. Two-way ANOVA revealed a significant effect of stress×drug interaction in females (F(1,34) = 6.102, p < 0.05) but not in males (F(1,33) = 0.096, ns). Stress and drug were not significant in males (stress: F(1,33) = 0.281, ns; drug: F(1,33) = 0.064, ns) or females (stress: F(1,34) = 1.203, ns; drug: F(1,34) = 0.344, ns),
In the VS (Fig. 5f), univariate ANOVA revealed a significant effect of sex, stress and the following interactions: sex×drug and sex×stress×drug, but not of drug and the other interactions. Two-way ANOVA revealed a significant effect of drug in males (F(1,28) = 5.955, p < 0.05) but not in females (F(1,34) = 2.378, ns), and a significant effect of stress in females (F(1,34) = 17.302, p < 0.001) but not in males (F(1,28) = 0.587, ns), suggesting that CBD led to upregulation of miR-98-5p in UCMS males, while in females UCMS led to downregulation of miR-98-5p. Stress×drug interaction was not significant in males (F(1,28) = 3.296 ns) or females (F(1,34) = 1.097, ns).
1.4.3: miR-146a-5p:
In the vmPFC (Fig. 5g), univariate ANOVA (2×2×2) revealed a significant effect of sex and sex×stress×drug interaction, but not of stress, drug, or any of the other interactions.
Two-way ANOVA (2×2) revealed a significant effect of stress×drug interaction in females (F(1,33) = 7.823, p < 0.01) but not in males (F(1,31) = 0.064, ns). Stress and drug were not significant in males (stress: F(1,31) = 0.160, ns; drug: F(1,31) = 1.246, ns) or females (stress: F(1,33) = 0.005, ns; drug: F(1,33) = 2.812, ns).
In the CA1 (Fig. 5h), univariate ANOVA revealed a significant effect of sex and sex×stress interaction, but not of stress, drug, or any of the other interactions. Two-way ANOVA revealed a significant effect of stress (F(1,34) = 13.529, p < 0.001) in males but not in females (F(1,34) = 0.465, ns), suggesting that in males UCMS led to upregulation of miR-146a-5p. Drug and stress×drug interaction were not significant in males (drug: F(1,34) = 0.286 ns; stress x drug: F(1,34) = 0.017, ns) or females (drug: F(1,34) = 0.259, ns; stress×drug: F(1,34) = 0.028, ns(.
In the VS (Fig. 5i), univariate ANOVA revealed a significant effect of sex and sex×stress interaction, but not of stress, drug, or any of the other interactions. Two-way ANOVA revealed a significant effect of stress (F(1,31) = 9.293, p < 0.01) in females but not in males (F(1,28) = 0.888, ns), suggesting that in females UCMS led to downregulation of miR-146a-5p. Drug and stress×drug interaction was not significant in males (drug: F(1,28) = 0.001, ns; stress×drug: F(1,28) = 0.527, ns) or females (drug: F(1,31) = 0.048, ns; stress×drug: F(1,31) = 2.262, ns).
To explore the association between depressive-like behavior and microRNA expression, Pearson bivariate correlation tests were conducted between the behavioral measurements and microRNA expression in the vmPFC, CA1, and VS in males (supplemental material, Table S9) and females (Table S10).