Chemogenetic activation of CRF neurons as a model of chronic stress produces sex-specific physiological and behavioral effects

Abstract

Trauma and chronic stress exposure are the strongest predictors of lifetime neuropsychiatric disease presentation. These disorders often have significant sex biases, with females having higher incidences of affective disorders such as major depression, anxiety, and PTSD. Understanding the mechanisms by which stress exposure heightens disease vulnerability is essential for developing novel interventions. Current rodent stress models consist of a battery of sensory, homeostatic, and psychological stressors that are ultimately integrated by corticotropin-releasing factor (CRF) neurons to trigger corticosteroid release. These stress paradigms, however, often differ between research groups in the type, timing, and duration of stressors utilized. These inconsistencies, along with the variability of individual animals' perception and response to each stressor, present challenges for reproducibility and translational relevance. Here, we hypothesized that a more direct approach using chemogenetic activation of CRF neurons would recapitulate the effects of traditional stress paradigms and provide a high-throughput method for examining stress-relevant phenotypes. Using a transgenic approach to express the Gq-coupled Designer Receptor Exclusively Activated by Designer Drugs (DREADD) receptor hM3Dq in CRF-neurons, we found that the DREADD ligand clozapine-N-oxide (CNO) produced an acute and robust activation of the hypothalamic-pituitary-adrenal (HPA) axis, as predicted. Interestingly, chronic treatment with this method of direct CRF activation uncovered a novel sex-specific dissociation of glucocorticoid levels with stress-related outcomes. Despite hM3Dq-expressing females producing greater corticosterone levels in response to CNO than males, hM3Dq-expressing males showed significant typical physiological stress sensitivity with reductions in body and thymus weights. hM3Dq-expressing females while resistant to the physiological effects of chronic CRF activation, showed significant increases in baseline and fear-conditioned freezing behaviors. These data establish a novel mouse model for interrogating stress-relevant phenotypes and highlight sex-specific stress circuitry distinct for physiological and limbic control that may underlie disease risk.

Fig. 1. CNO induces dose-responsive corticosterone release in CRF-Cre + /DREADD+ mice.

A Corticosterone levels were measured in response to 4 CNO doses in DREADD+ males (2-way RM ANOVA; F_time(1.698,15.28) = 15.91, p  < 0.001; F_dose(3,9) = 2.933, p  = 0.092; F_time*dose(9,27) = 1.278, p  = 0.293; n = 3–4). Corticosterone was elevated in response to 5 mg/kg CNO compared to 0.25 mg/kg CNO at 60 mins (p  = 0.020) and 180 mins (p  = 0.041) post-injection. B Area under the curve analysis of total corticosterone release did not show a significant effect of CNO dose (1-way ANOVA; F_dose(3,9) = 2.998, p  = 0.088; n = 3–4). C CNO dose did not significantly affect corticosterone levels at the HPA axis recovery timepoint in DREADD+ males (1-way ANOVA; F_dose(3,9) = 2.403; p  = 0.135; n = 3–4). D Corticosterone responses in DREADD+ females were significantly affected by CNO dose (2-way RM ANOVA; F_dose(3,13) = 18.28, p  < 0.0001; F_time(2.066,26.86) = 55.10, p  < 0.0001; F_time*dose(9,39) = 6.684, p  < 0.0001; 120 mins post-injection: 5 mg/kg vs 0.5 mg/kg p  = 0.009; 5 mg/kg vs 0.25 mg/kg p  = 0.005; 180 mins post-injection: 5 mg/kg vs 1 mg/kg, p  = 0.007; 5 mg/kg vs. 0.5 mg/kg p  = 0.013; 5 mg/kg vs 0.25 mg/kg p  = 0.0008; n = 3–6). E CNO dosing significantly affected the total amount of corticosterone released in DREADD+ females (1-way ANOVA; F_dose(3,13) = 17.04, p  < 0.0001; 0.25 mg/kg vs. 1.0 mg/kg p  = 0.021; 0.25 mg/kg vs. 5.0 mg/kg p  = 0.0001; 0.5 mg/kg vs. 5.0 mg/kg p  = 0.0006, 1.0 mg/kg vs. 5.0 mg/kg p  = 0.013; n = 3–6). F 5 mg/kg CNO significantly elevated corticosterone at the HPA axis recovery timepoint in DREADD+ females (1-way ANOVA; F_dose(3,13) = 13.51, p  = 0.0003; 5 mg/kg vs 1.0 mg/kg p  = 0.001; 5 mg/kg vs. 0.5 mg/kg p  = 0.005; 5 mg/kg vs 0.25 mg/kg p  = 0.0005; n = 3–6). (**** p  < 0.0001, ***p  < 0.001, **p  < 0.01, ^##p  < 0.01, *p  < 0.05, ^p  < 0.05, ^#p  < 0.05, @ main effect of genotype). *An example week of the variable-dose CNO treat paradigm is shown in Table S1.

Fig. 2. DREADD+ males display physiological stress features and have increased c-Fos expression in the PVN and central amygdala following repeated high-dose CNO.

A 4 days after beginning CNO treat consumption, 62.5% of DREADD+ males (n = 7) partially or fully stopped consuming CNO treats compared to 12.5% of female DREADD+ mice (n = 9). 100% of DREADD+ males partially or fully stopped consuming the treats after 5 days compared to 25% of DREADD+ females. B Weekly body weight measurements across 4 weeks of CNO in male mice (2-way RM ANOVA; F_Time(1.345,16.14) = 4.716, p  = 0.039; F_genotype(1,12) = 3.201, p  = 0. 099; F_{time*genotype}(4,48) = 2.480, p  = 0.056; n = 7 per group). C 4 weeks of CNO significantly reduced body weight (unpaired t-test, t₍₁₁₎ = 2.638, p  = 0.023; n = 6–7) and D thymus weights (unpaired t-test; t₍₁₁₎ = 2.638, p  = 0.023, n = 4–5) in DREADD+ males compared to controls. E Weekly body weight measurements across 4 weeks of CNO in female mice (2-way RM ANOVA; F_time(2.478,34.70) = 4.938, p  = 0.009; F_genotype(1,14) = 0.751, p  = 0.401; F_{time*genotype}(4,56) = 0.964, p  = 0.435; n = 7–9). F 4 weeks of CNO did not induce overall body weight change (unpaired t-test, t₍₁₄₎ = 0.952, p  = 0.357, n = 7–9) or (G) affect thymus weights (unpaired t-test; t₍₈₎₌0.721, p  = 0.492; n = 3–7) in female DREADD+ mice compared to controls. H Representative images of c-Fos immunostaining in the PVN and (I) central amygdala. Dashed lines indicate region of interest used for quantification. J Quantification of c-Fos immunoreactivity in the PVN (2-way ANOVA; F_sex(1,10) = 22.77, p  = 0.0008; F_genotype(1,10) = 20.68, p  = 0.001, F_sex*genotype(1,10) = 14.15, p  = 0.004; n = 3–4). DREADD+ males had a significantly higher proportion of c-Fos immunoreactivity compared to controls (p  = 0.0007) and DREADD+ females (p  = 0.0003) while there were no significant differences in DREADD+ females compared to controls (p  = 0.943). K Quantification of c-Fos immunoreactivity in the CeA (2-way ANOVA; F_sex(1,12) = 7.037, p  = 0.021; F_genotype(1,12) = 34.16, p  < 0.0001; F_sex*genotype(1,12) = 3.831, p  = 0.074). DREADD+ males had a significantly higher proportion of c-Fos immunoreactivity (p  = 0.0007) compared to controls and DREADD+ females (p  = 0.03) while there were no significant differences in c-Fos immunoreactivity between DREADD+ females and controls (p  = 0.073). L Representative images of HA immunostaining in the PVN and (M) CeA in male and female DREADD+ mice. By visual inspection of expression patterns, no apparent differences were noted. (***p  < 0.001, *p  < 0.05, @ main effect of genotype).

Fig. 3. Lower variable-dose chronic CNO does not induce a severe stress physiological phenotype or lead to CNO habituation.

*An example week of the variable-dose CNO treat paradigm is shown in Table S1. A 100% of male (n = 7) and female (n = 8) DREADD+ mice continued consuming daily CNO treats across 9 weeks. B Weekly body weight measurements across 9 weeks in DREADD- (n = 10) and DREADD+ (n = 7) males (2-way RM ANOVA; F_time(1.091,16.36) = 5.050, p  = 0.036, F_genotype(1,15) = 0.043, p  = .838; F_{time*genotype}(9,135) = 0.162, p  = 0.997). C Lower, variable dose CNO treats did not induce a greater overall body weight change in DREADD+ males compared to controls (unpaired t-test; t₍₁₅₎₌0.434, p  = 0.670). D Weekly body weight measurements across 9 weeks in DREADD- (n = 8) and DREADD+ (n = 8) females (2-way RM ANOVA; F_time(1.230,17.22) = 2.238, p  = 0.15; F_genotype(1,14) = 0.039, p  = 0.847; F_{time*genotype}(9,126) = 0.151, p  = 0.998). E There were no differences in overall body weight change at the end of 9 weeks in DREADD+ females compared to controls (unpaired t-test; t₍₁₄₎₌0.684, p  = 0.505; n = 8 per group). F 9 weeks of CNO did not affect HPA axis reactivity to an acute CNO injection (2-way RM ANOVA; F_{CNO chronicity}(1,9) = 0.112, p  = 0.746; F_time (1.564, 14.08) = 13.09, p  = 0.001; F_{time*CNO chronicity}(3, 27) = 0.494, p  = 0.690) or (G) total amount of corticosterone released following an acute CNO injection in DREADD+ males (n = 7) compared to CNO-naïve males (n = 3) (unpaired t-test; t₍₈₎ = 0.0467, p  = 0.964). H 9 weeks of chronic CNO did not affect HPA axis reactivity to an acute CNO injection (2-way RM ANOVA; F_{CNO chronicity}(1,10) = 0.002, p  = 0.967; F_time(1.686, 16.86) = 40.77, p  < 0.0001; F_{time*CNO chronicity}(3, 30) = 0.379, p  = 0.769) or (I) total amount of corticosterone released in female DREADD+ mice (n = 6) compared to CNO-naïve females (n = 6) (unpaired t-test; t₍₁₀₎ = 0.221, p  = 0.829). J Slope analysis of HPA axis corticosterone response from 60- to 120- mins post-injection. Female DREADD+ mice recovered faster from peak corticosterone levels than DREADD+ males when CNO-naïve (p  = 0.014) and after 9 weeks of chronic CNO administration (p  = 0.003) (2-way ANOVA; F_sex(1,18) = 18.49, p  = 0.0004; F_{CNO chronicity}(1,18) = 1.231, p  = 0.282; F_{sex*CNO chronicity}(1,18) < 0.0001, p  = 0.999). (*p  < 0.05, @ main effect of sex).

Fig. 4. Male DREADD+ mice have heightened HPA axis reactivity while female DREADD+ mice have heightened fear responses following chronic CNO.

A Male DREADD+ mice (n = 5) had an elevated HPA axis response to 15 min of restraint stress following 9 weeks of CNO (2-way RM ANOVA; F_genotype(1,14) = 3.904, p  = 0.068, F_time(1.671,23.39) = 77.76, p  < 0.0001; F_{time*genotype}(3,42) = 5.130, p  = 0.004) with a significant elevation in corticosterone levels 15 mins (p  = 0.009) and 30 mins (p  = 0.020) after restraint onset compared to controls (n = 11). B Area under the curve analysis did not show a significant difference in total corticosterone released in DREADD+ males following 9 weeks of chronic CNO (unpaired t-test; t₍₁₄₎ = 1.921, p  = 0.075). C 9 weeks of chronic CNO did not alter the HPA axis response to acute restraint stress in DREADD+ females (n = 8) compared to controls (n = 7) (2-way RM ANOVA; F_genotype(1,13) = 0.529, p  = 0.480; F_time(2.096,27.27) = 78.92, p  < 0.0001; F_{time*genotype}(3,39) = 0.167, p  = 0.918) and (D) did not affect total corticosterone released (unpaired t-test; t_(g13)=0.729, p  = 0.479). E Using the von Frey filament test to measure tactile sensitivity, DREADD+ males (n = 7) showed a leftward shift of the paw withdrawal curve compared to controls (n = 11) (2-way RM ANOVA; F_genotype(1,16) = 4.034, p  = 0.062; F_force(14,224) = 169.8, p  < 0.0001; F_{genotype*force}(14, 224) = 2.658, p  = 0.001) with DREADD+ males having significantly more paw withdrawals at 1.0 g (p  = 0.021), 1.4 g (p  = 0.006), and 2.0 g (p  = 0.001) of force. F DREADD+ males had a lower average force required for 50% paw withdrawal (VF50) compared to controls (unpaired t-test; t₍₁₆) = 2.489, p  = 0.024). G DREADD+ females (n = 8) also showed a leftward shift of the paw withdrawal curve compared to controls (n = 8) (2-way RM ANOVA; F_genotype(1,14) = 14.93, p  = 0.002; F_force(4.396,61.55) = 371.2, p  < 0.0001; F_{genotype*force}(14,196) = 5.182, p  < 0.0001) with significantly more paw withdrawals at 1.4 g of force (p  = 0.014), and (H) had lower VF50 (unpaired t-test; t₍₁₄₎ = 4.235, p  = 0.0008) compared to controls. I Freezing behavior during auditory fear conditioning. Male DREADD+ mice (n = 7) showed no differences in freezing behavior during conditioning (2-way RM ANOVA; F_genotype(1,16) = 0.287, p  = 0.600; F_trial(2.114,33.82) = 12.26, p  < 0.0001; F_{genotype*trial}(3,48) = 0.085, p  = 0.968) or extinction (2-way RM ANOVA; F_genotype(1,16) = 0.129, p  = 0.725; F_trial(2.441,39.05) = 5.210, p  = 0.007; F_{trial*genotype}(4,64) = 1.073, p  = 0.377) compared to controls (n = 11) following 9 weeks of CNO. J Chronic CNO significantly elevated freezing behavior in female DREADD+ mice during conditioning (n = 8 per group; 2-way RM ANOVA; F_genotype(1,14) = 23.36, p  = 0.0003, F_trial(2.170, 30.38) = 25.24, p  < 0.0001; F_{genotype*trial}(3,42) = 4.20, p  = 0.011), with DREADD+ females freezing more at baseline (p  = 0.003) and conditioning trials 1 (p  = 0.002) and 3 (p  = 0.0004). Chronic CNO also increased female DREADD+ freezing behavior (n = 6 per group) during the extinction trials (2-way RM ANOVA; F_genotype(1,10) = 16.23, p  = 0.002; F_trial(2.092, 20.92) = 2.324, p  = 0.121; F_{genotype*trial}(4,40) = 0.780, p  = 0.545), with significantly higher levels of freezing compared to controls during extinction trials 1 (p  = 0.022) and 4 (p  = 0.034). (***p  < 0.001, **p  < 0.01, *p  < 0.05, @ main effect of genotype).