Stress experience and hormone feedback tune distinct components of hypothalamic CRH neuron activity

Abstract

Stress leaves a lasting impression on an organism and reshapes future responses. However, the influence of past experience and stress hormones on the activity of neural stress circuits remains unclear. Hypothalamic corticotropin-releasing hormone (CRH) neurons orchestrate behavioral and endocrine responses to stress and are themselves highly sensitive to corticosteroid (CORT) stress hormones. Here, using in vivo optical recordings, we find that CRH neurons are rapidly activated in response to stress. CRH neuron activity robustly habituates to repeated presentations of the same, but not novel stressors. CORT feedback has little effect on CRH neuron responses to acute stress, or on habituation to repeated stressors. Rather, CORT preferentially inhibits tonic CRH neuron activity in the absence of stress stimuli. These findings reveal how stress experience and stress hormones modulate distinct components of CRH neuronal activity to mediate stress-induced adaptations.

Fig. 1. Optical recordings of CRH neuron activity in freely behaving mice.

a Image showing PVN expression of CRH-tdTomato reporter (left), AAV-driven GCaMP6s (center), and merged image (right). Scale bar: 100 μm. b Schematic illustration of the fiber photometry setup. c Blood CORT concentrations obtained from tail tip samples, while mice remained within experimental testing box in the absence of white noise (WN) stress; n = 8, p > 0.05, paired t test. d Blood CORT concentrations following WN stress; n = 8, ***p < 0.001, paired t test. e Photometry recordings of CRH neurons from three individual mice displaying continued activity after termination of WN. f Photometry recordings of CRH neurons from three individual mice displaying rapid cessation of activity after termination of WN. g Peak ΔF/F at WN stress onset, n = 64. h Time to peak from the onset of WN onset, n = 64. i Mean ΔF/F of CRH neuron activity during 5 min WN stress from all individual mice tested, n = 64. j Mean CRH neuron ΔF/F in 10 sec bins from all individual mice; n = 64, repeated-measures (RM) one-way ANOVA ***p < 0.001 vs. baseline, Dunn’s post hoc test. All data are presented as mean ± SEM.

Fig. 2. CRH neuron shut-off and adaptation occur regardless of fast CORT feedback.

a Mean photometry signals of CRH neuron activity induced by two sequential WN stressors 30 min apart from vehicle and metyrapone-treated mice. b Heatmap of mean CRH neuron activity from all individual mice in 20 s bins. c Peak ΔF/F at WN onset; RM two-way ANOVA, *p < 0.05 vs. WN1, Holm–Sidak; ANOVA interaction p = 0.14, ANOVA main effect of group p = 0.86. d Average ΔF/F across 5 min of CRH neuron activity before, during, and after each WN; n = 13 per group, RM two-way ANOVA, *p < 0.05 vs. baseline, ^†p < 0.05 vs. respective WN1 timepoint, Holm–Sidak; ANOVA interaction p = 0.90, ANOVA main effect of group p = 0.38. e CORT levels before and after WN stress following metyrapone treatment; n = 8, paired t test. f Percentages of CRH neuron activity during WN2 relative to WN1; Veh vs. MET, Mann–Whitney test. g Averaged photometry recordings of CRH neuron activity from all vehicle and metyrapone-treated mice showing the response to WN2. h Cumulative integrated ΔF/F from the time of WN2 onset; RM two-way ANOVA, *p < 0.05 Veh vs. MET, Holm–Sidak; ANOVA interaction p < 0.001, ANOVA main effect of group p = 0.09. Lines indicate points at which statistical significance was reached and its duration. Gray shaded area indicates duration of WN. All data are presented as mean ± SEM, */^†p < 0.05, **/^††p < 0.01, ***/^†††p < 0.001.

Fig. 3. CORT feedback slowly inhibits tonic CRH neuron excitability without affecting habituation.

a Mean photometry signals of CRH neuron activity from two sequential WN stressors 120 min apart in vehicle and metyrapone-treated mice. b Heatmap of mean CRH neuron activity from all individual mice in 30 s bins. c Average ΔF/F across 5 min of CRH neuron activity before, during, and after each WN; n = 11 per group, RM two-way ANOVA, *p < 0.05 vs. baseline, ^†p < 0.05 vs. respective WN1 timepoint, Holm–Sidak; ANOVA interaction p = 0.71, ANOVA main effect of group p = 0.39. d Cumulative integrated ΔF/F from the point of WN1 onset; RM two-way ANOVA, *p < 0.05 Veh vs. MET, Holm–Sidak; ANOVA interaction p < 0.001, ANOVA main effect of group p = 0.19. Lines indicate points at which statistical significance was reached and its duration. Gray shaded area indicates duration of WN. e CRH neuron activity during WN2 is shown as a percentage relative to WN1; Veh vs. MET, Mann–Whitney test. f Averaged photometry recordings of CRH neuron activity from vehicle- and metyrapone-treated mice showing the response to white noise 2 (WN2). g Cumulative integrated ΔF/F from the time of WN2 onset; RM two-way ANOVA, Veh vs. MET, Holm–Sidak; ANOVA interaction p < 0.001, ANOVA main effect of group p = 0.38. Gray shaded area indicates duration of WN. h Peak CRH ΔF/F at WN onset; RM two-way ANOVA, *p < 0.05 vs. WN1, Holm–Sidak; ANOVA interaction p = 0.08, ANOVA main effect of group p = 0.87. i Representative photometry traces of tonic activity (during a period between WN1 and WN2) in individual vehicle- (left) or metyrapone- (right) treated mice. j Total number of GCaMP transients and proportion of transients larger than 10% of individual peak WN ∆F/F; *p < 0.05, one-way ANOVA, Tukey. k Average ΔF/F (% of individual WN peak) of detected GCaMP transients during post-stress activity; *p < 0.05, Mann–Whitney test. l Average slope (measured by amplitude/time) of detected GCaMP transients (ΔF/F measured as % of individual WN peak); *p < 0.05, Mann–Whitney test. All data presented as mean ± SEM, */^†p < 0.05, **/^††p < 0.01, ***/^†††p < 0.001.

Fig. 4. CORT-negative feedback slowly suppresses basal but not stress-evoked CRH neuron activity.

a Mean photometry signals of CRH neuron activity from vehicle- and CORT-treated mice. Gray shaded area indicates time of injection where handling/injection stress response is evident. b Heatmap of mean CRH neuron activity from all individual mice in 30 s bins. c Mean ΔF/F changes in 40 min bins; n = 12 per group, RM two-way ANOVA, ***p < 0.001 vs. baseline, Holm–Sidak; ANOVA interaction p < 0.001, ANOVA main effect of group p < 0.001. CRH neuron activity during injection stress (5 min bin) was not included in the statistical analysis. d Cumulative integrated ΔF/F of tonic CRH neuron activity; RM two-way ANOVA, *p < 0.05 Veh vs. CORT, Holm–Sidak; ANOVA interaction p < 0.001, ANOVA main effect of group p = 0.025. Lines indicate points at which statistical significance was reached and its duration. Gray shaded area indicates stress response due to injection. e Averaged photometry recordings of the WN stress response 150 min after injection without baseline normalization. f Average ΔF/F across 5 min of CRH neuron activity before, during, and after WN stress without baseline normalization; n = 8 per group, RM two-way ANOVA, *p < 0.05 vs. baseline (unless otherwise indicated), Holm–Sidak; ANOVA interaction p = 0.45, ANOVA main effect of group p = 0.01. g Peak ΔF/F at WN onset without baseline normalization; Mann–Whitney test. h Averaged photometry recordings of the WN stress response 150 min after injection with baseline normalized to 10 min of activity prior to white noise. i Average ΔF/F across 5 min of CRH neuron activity before, during, and after WN stress after baseline normalization; n = 8 per group, RM two-way ANOVA, *p < 0.05 vs. Baseline, Holm–Sidak; ANOVA interaction p = 0.48, ANOVA main effect of group p = 0.47. j Peak ΔF/F at WN onset after baseline normalization; Mann–Whitney test. All data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 5. Fast CORT feedback has little impact on CRH neural activity but faithfully suppresses ACTH release.

a Mean photometry signals of CRH neuron activity during WN stress, 30 min after injection of either vehicle or CORT. b Heatmap of mean CRH neuron activity from all individual mice in 10 s bins. c Average ΔF/F across 5 min of CRH neuron activity before, during, and after WN; n = 12 per group, RM two-way ANOVA, ***p < 0.001 vs. baseline, Holm–Sidak; ANOVA interaction p = 0.19, ANOVA main effect of group p = 0.14. d Peak ΔF/F at WN onset; Mann–Whitney test. e Cumulative integrated ΔF/F from the time of WN stress; RM two-way ANOVA, *p < 0.05 Veh vs. CORT, Holm–Sidak; ANOVA interaction p < 0.001, ANOVA main effect of group p = 0.21. Lines indicate points at which statistical significance was reached and its duration. Gray shaded area indicates duration of WN. f Plasma ACTH levels 5 min post WN; one-way ANOVA, Tukey’s multiple comparison test. All data presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 6. CORT feedback increases first spike latency and decreases spike output.

a Traces of sEPSC from individual CRH neurons from vehicle (green) or metyrapone- (red) treated mice. b Traces of paired-pulse eEPSC amplitudes from CRH neurons from vehicle (green) or metyrapone- (red) treated mice. c Mean sEPSC frequency; unpaired t test, p = 0.8. d Mean sEPSC amplitudes; unpaired t test, p = 0.3. e Mean eEPSC P1 amplitudes; unpaired t test, p = 0.9. f Paired pulse ratio (PPR); Mann–Whitney test, p = 0.3. g Traces of CRH neuron spike output from vehicle (green) or metyrapone- (red) treated mice in response to varying current steps. h Spike frequency plotted for each 5 pA step; n = 11 per group, RM two-way ANOVA, *p < 0.05 vs. vehicle, Holm–Sidak; ANOVA interaction p = 0.09, ANOVA main effect of group p < 0.001. i Input resistance from individual CRH neurons; unpaired t test, p = 0.9. j CRH neuron responses to 10 pA (left) and 20 pA (right) current steps with or without prior incubation with 4AP (Veh + 4AP in black, MET + 4AP in gray). k, l FSL during 10 pA (k) and 20 pA (l) steps from vehicle and metyrapone-treated mice with or without 4AP incubation. RM two-way ANOVA *p < 0.05 Holm–Sidak; ANOVA interaction p = 0.003, ANOVA main effect of group p < 0.001. All data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 7. Experience-gated habituation to white noise stress is dependent on stress familiarity.

a Mean photometry signals of CRH neuron activity from mice receiving footshock (FS) followed by WN with a 30 min interval. b Mean photometry signals from mice receiving WN followed by FS. c Averaged photometry recordings of CRH neuron activity during WN stress when presented either first or second. d Average ΔF/F across 5 min of CRH neuron activity before, during, and after each WN; n = 9 per group, RM two-way ANOVA, *p < 0.05 vs. baseline, Holm–Sidak; ANOVA interaction p = 0.13, ANOVA main effect of group p = 0.96. e Peak ΔF/F response to WN stress presented either first or second; Mann–Whitney test. f Averaged photometry recordings of CRH neuron activity during footshock stress when presented either first or second. g Average ΔF/F across 5 min of CRH neuron activity before, during, and after each FS; n = 9 per group, RM two-way ANOVA, *p < 0.05 vs. baseline, Holm–Sidak; ANOVA interaction p = 0.44, ANOVA main effect of group p = 0.31. h Peak ΔF/F response to FS stress presented either first or second; Mann–Whitney test. i Cumulative integrated ΔF/F from the time of WN stress; RM two-way ANOVA, p > 0.99 WN before FS vs. WN after FS at 10 min, Holm–Sidak; ANOVA interaction p > 0.99, ANOVA main effect of group p = 0.52. Gray shaded area indicates duration of WN. j Cumulative integrated ΔF/F from the time of FS stress; RM two-way ANOVA, p > 0.99 FS before WN vs. FS after WN at 10 min, Holm–Sidak; ANOVA interaction p < 0.001, ANOVA main effect of group p = 0.67. Gray shaded areas indicate timing of the two FSs (2 s duration each). k Mean ΔF/F response to the first presentations of WN or FS stress in 5 min bins; n = 9 per group, RM two-way ANOVA, *p < 0.05 WN vs. FS, Holm–Sidak; ANOVA interaction p = 0.007, ANOVA main effect of group p = 0.051. All data presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 8. Long-term adaptation of CRH neural activity to stress.

a Mean photometry signals of CRH neuron activity from mice receiving daily WN stress over 4 days (round 1: day 1; red; day 2, green; day 3, blue; day 4, orange). Black trace indicates mean CRH neuron activity from the same cohort of mice on day 0, in the absence of WN. b Mean photometry signals of CRH neuron activity from the same mice after a 3-week rest interval (with no stress) and then subsequently receiving daily WN stress over 4 more days (round 2: all recordings in black overlaid with corresponding round 1 WN response). c Mean ΔF/F changes in response to each WN or no stress (black) in 2.5 min bins during round 1. d Average ΔF/F of CRH neuron activity during WN stress across the two rounds of repeated stress; n = 6, RM two-way ANOVA, *p < 0.05 vs. round 1 WN day 1, ^†p < 0.05 vs. round 2 WN day 1, Holm–Sidak; ANOVA interaction p = 0.77, ANOVA main effect of group p = 0.09. e Peak ΔF/F at WN onset across the two rounds of repeated stress; RM two-way ANOVA, Holm-Sidak; ANOVA interaction p = 0.53, ANOVA main effect of group p = 0.60. f Correlation of mean CRH neuron activity during WN and corresponding post-stress blood CORT concentration across the two rounds (round 1 in circles and round 2 in triangles, 4 days of WN indicated by corresponding color; Pearson’s r = 0.90, r² = 0.8. All data presented as mean ± SEM, */^†p < 0.05, **/^††p < 0.01, ***/^†††p < 0.001.