A stress-sensing circuit signals to the central pacemaker to reprogram circadian rhythms

Abstract

The circadian system provides a temporal framework for animals to anticipate environmental events, including threats. However, the effects of stressors on the circadian system remain poorly understood. Here, we demonstrate that, in mice, stressors shift the phase of the central pacemaker, housed in the suprachiasmatic nucleus (SCN), through glutamatergic inputs from the anterior paraventricular nucleus of the thalamus (aPVT). Unlike light, which can phase delay or advance the central pacemaker, stressors consistently induce delays, effects attenuated by inhibiting aPVT neurons. Stressors robustly activate AVP-expressing neurons within the SCN and are associated with inhibition of VIP-expressing neurons, whereas light strongly activates VIP-expressing neurons with minimal effects on AVP-expressing neurons. Pairing stressors with light reveals distinct time-dependent interactions, enhancing phase delays at early night but abolishing phase advances at late night. Our findings uncover distinct SCN microcircuits that differentially encode light and stressors, providing insights into how environmental cues modulate circadian timing.

Fig. 1.. Stressors phase shift circadian rhythms differently than light.

(A) Schematic of experimental paradigm. (B to D) Wheel running activity profiles of control mice (white rectangle), and mice subjected to footshock session (red rectangle) at CT14 (B), CT22 (C), and CT6 (D). Rectangles represent day and time of test. Blue solid line depicts activity onset before test; blue dotted line depicts expected activity onsets, and pink regression line depicts shift in activity onset posttest. Black arrow indicated the first day posttest. Scatter plots represent quantification of the phase shifts (hours) in controls and mice subjected to footshocks at CT14 (B), CT22 (C), and CT6 (D). Data are means ± SEM. CT14, *P = 0.01 (n = 9 to 14 mice). CT22, *P = 0.03 (n = 13 to 14 mice). CT6, nonsignificant (n.s., n = 15 to 17 mice). Student’s t test. (E) Comparison between phase shift (hours) in response to light (n = 6 mice) and footshock (n = 13 to 17) at CT6, CT14, and CT22. Data are means ± SEM. ***P = 0.002; ****P < 0.0001, Student’s t test. (F) Running-wheel activity profiles of control mice (white), and mice subjected to a forced swim session (dark red) and quantification of the phase shifts (hours) in controls and mice subjected to forced swim at CT14. Data are means ± SEM (n = 5 to 6 mice). *P = 0.04, Student’s t test. (G) Schematic of fiber implantation in the SCN of NMS-Cre;GCaMP8s mice and representative image of GCaMP8s expression and fiber placement. Scale bar, 100 μm. (H) Schematic of experimental paradigm. (I) Representative calcium activity profiles in response to a sham and a footshock session. Acrophase (pink circles) was used to calculate phase shifts. Quantification of the phase shifts (hours) in controls, and mice subjected to footshocks. Data are means ± SEM (n = 7 mice). *P = 0.04, two-tailed paired-sample t test. Cntrl, control; ITI, intertrial interval.

Fig. 2.. Pairing light and stressors results in additive phase delays at CT14 and reduces light-induced phase advances at CT22.

(A) Schematic of experimental paradigm; mice were placed in constant dark (DD) for 6 to 9 days and then subjected to a light pulse (middle) or light pulse and footshock session (right) at CT14 and CT22. Wheel running activity was recorded for an additional 7 days. (B and D) Running-wheel activity profiles of mice exposed to a light pulse alone (yellow) or light pulse paired with footshocks (orange) at CT14 (B) and CT22 (D). Blue solid line depicts activity onset before test, blue dotted line depicts expected activity onsets, and pink line depicts shift in activity onset posttest. Black arrow indicated the first day posttest. (C and E) Quantification of the phase shifts (hours) in light only (yellow) or light paired with footshock (orange) groups at CT14 (C) and CT22 (E). Data are means ± SEM. CT14 (n = 13 mice), *P = 0.028, Student’s t test. CT22 (n = 6 mice), *P = 0.035, two-tailed paired t test. ITI, intertrial interval.

Fig. 3.. SCN^NMS+ neurons are differentially responsive to aversive stimuli across CT times.

(A) Schematic of the stereotaxic injection to selectively target expression of GCaMP7s to NMS-expressing SCN neurons (left). Representative image of GCaMP7s expression in the SCN and fiber placement (right). Scale bar, 100 μm. (B) Schematic of experimental paradigm. (C to E) Heatmaps of average calcium responses to footshocks (five trials) per mouse and average responses for mice displaying increases in fluorescence levels (shock activated; green) and decreases in fluorescence levels (shock inhibited; blue) at CT14 (C), CT22 (D), or CT6 (E). (F to H) Quantification of GCaMP7s responses to footshocks. Area under the curve (AUC) comparison between baseline (0 to 10 s; circle) and footshock (10 to 20 s; square) in shock activated and shock inhibited groups. Data are means ± SEM. Shock activated (n = 4 to 5 mice), CT14, **P = 0.005 (F). CT22, nonsignificant (G). CT6, nonsignificant (H). Shock inhibited (n = 3 mice), CT14, nonsignificant (F). CT22, nonsignificant (G). CT6, *P = 0.004 (H). Two-tailed paired-sample t test. (I to K) Latency to reach peak activation (green) or inhibition (blue) from the onset of footshock. Data are means ± SEM. CT14, shock activated, (n = 5 mice), shock inhibited (n = 3 mice), ***P = 0.0009 (I). CT22, shock activated (n = 5 mice), shock inhibited (n = 3 mice), **P = 0.0036 (J). CT6, shock activated, (n = 4 mice), shock inhibited, (n = 3 mice), *P = 0.023 (K). Student’s t test. Light pulse (15 s) and footshock duration (2 s) are depicted by dotted line and shaded area. (L to N) Heatmaps of average calcium responses to light (five trials per mouse) and average GCaMP7s responses at CT14 (L), CT22 (M), and CT6 (N). Continuous yellow in heatmaps represents time points where response exceeded maximum z-score. ITI, intertrial interval.

Fig. 4.. SCN-projecting PVT neurons are activated by footshocks.

(A) Schematic of the retrograde tracing strategy used for labeling SCN-projecting PVT neurons (left) and representative images showing injection sites for eGFP in the SCN (right). Scale bar, 100 μm. (B) Representative images showing eGFP expression in PVT neurons across anteroposterior axis. Scale bar, 100 μm. (C) Quantification of PVT neurons projecting to SCN. Data are means ± SEM (n = 3 mice). **P = 0.0012; ***P = 0.0003, one-way ANOVA, Tukey’s test. (D) Schematic of the viral vector strategy for anterograde tracing of PVT projections and representative image showing injection site for mCherry in the PVT. Scale bar, 500 μm. (E) Representative images showing CAMKII-mCherry expression in axon terminals within the SCN along the anteroposterior axis. Scale bar, 100 μm. (F) Quantification of the density of PVT axonal inputs in the SCN. Data are means ± SEM (n = 4 mice). Not significant, one-way ANOVA, Tukey’s test. (G) Schematic of tracing strategy to specifically express GCaMP7s in PVT neurons that project to the SCN (left). Representative image of injection site in the SCN and GCaMP7s expression in SCN-projecting PVT neurons and fiber placement. Scale bar, 100 μm (middle). Schematic of the footshock session used in fiber photometry experiments (right) (H) Heatmap of average z-score of ∆F/F of five footshock trials per mouse and average GCaMP7s responses to footshocks (n = 5 mice). Footshock duration (2 s) is depicted by shaded area. (I) Area under the curve (AUC) comparison between baseline (0 to 10 s; black) and footshock (10 to 20s; gray) groups. Data are means ± SEM (n = 5 mice). *P = 0.0169, two-tailed paired-sample t test. ITI, intertrial interval.

Fig. 5.. PVT is necessary for the effects of aversive stimuli on the central pacemaker.

(A) Schematic of the stereotaxic injection to target expression of inhibitory DREADDS to aPVT neurons (left). Representative image of hM4D(Gi)-mCherry expression in the aPVT (right). Scale bar, 100 μm. (B) Schematic of experimental paradigm; mice were placed in constant darkness (DD) for 6 to 9 days, and, then, at CT13.5, saline or CNO was injected intraperitoneally. Thirty minutes later, mice were subjected to a footshock session and allowed to free run for an additional 7 days. (C) Running-wheel activity profiles of control mice injected with saline (white rectangle) or CNO (dark gray rectangle) and subsequently subjected to a footshock session (red rectangle) at CT14 (left). Quantification of the phase shifts (hours) in response to footshocks in saline (Sal)– and CNO-injected mice at CT14 (right). Data are means ± SEM (n = 10 to 11 mice). *P = 0.04, Student’s t test. ITI, intertrial interval.

Fig. 6.. PVT neurons send glutamatergic inputs to SCN to recruit local inhibitory microcircuits.

(A) Schematic of anterograde tracing strategy used for expressing ChrimR-TdT in aPVT axons innervating the SCN and image depicting TdT expression in the PVT. Scale bar, 200 μm. (B) Sample trace showing optically evoked responses in the same SCN neurons following aPVT optogenetic stimulation: IPSCs (blue) and EPSC (green). Comparison between EPSC and IPSC amplitude from cells that display both EPSC/IPSC (right). Data are means ± SEM (n = 13 neurons and n = 7 mice). ***P = 0.0001, two-tailed paired t test. (C) Proportion of SCN neurons that receive excitatory/or inhibitory inputs from aPVT (n = 22 neurons). (D) Sample trace of optically evoked EPSC (green) and delayed IPSC (blue) elicited in SCN neurons (left). IPSC failures are represented by black and gray traces (right). Synaptic latency of EPSC (n = 17) and IPSC (n = 14) after optogenetic stimulation of aPVT axons (left). Data are means ± SEM. ****P < 0.0001, two-tailed paired t test. (E) Sample trace showing optically evoked responses in SCN neurons in the presence of TTX and 4-AP (right). Comparison between EPSC and IPSC amplitude (right). Data are means ± SEM (n = 21 neurons and n = 5 mice). ****P < 0.0001, two-tailed paired t test. (F) Sample trace showing optically evoked EPSC in SCN neurons pre and post bath application of NBQX and D-AP5 (left). Amplitude in SCN neurons receiving excitatory input from PVT pre– and post–D-AP5 and NBQX application (right). Data are means ± SEM (n = 5 neurons, n = 5 slices, and n = 5 mice). ****P < 0.0001, two-tailed paired t test. (G) Sample trace of SCN neurons showing optically evoked IPSC (left). IPSC amplitude pre and post application of NBQX and D-APV (right). Data are means ± SEM (n = 7 neurons and n = 6 mice). **P = 0.0072, two-tailed paired t test.

Fig. 7.. Local inhibitory microcircuit shapes SCN response to PVT inputs.

(A) Schematic of the viral vector strategies used for combined two-photon Ca²⁺ imaging of GCaMP8s fluorescence from SCN^NMS+ neurons and optogenetic manipulation of PVT terminals in the SCN. (B) Schematic (left) and representative image of SCN brain slice with expression of ChrimR-TdT in PVT axons (magenta) and GCaMP8s (green) in SCN^NMS+ neurons (right). Scale bar, 100 μm. (C) Average fluorescence responses of SCN^NMS+ neurons, categorized by response type, to aPVT optogenetic stimulation at CT14. Type I represents SCN^NMS+ neurons that showed net increases, and type II represents SCN^NMS+ neurons that showed brief increases followed by decreases in response to aPVT optogenetic stimulation. (D) Proportion of SCN neurons classified as type I, type II, or nonresponsive (n = 1077 neurons, n = 3 slices, and n = 2 mice). (E) Heatmap showing average GCaMP8s fluorescence (∆F/F, five trials) per cell in response to optogenetic stimulation (200 ms, 550 nm at 20 Hz; dotted line) in the presence of ACSF (left), TTX (500 nM), and 4-AP (100 μM, left), and D-AP5 (50 μM) and CBQX (10 μM, right). (F) Example traces showing individual trials per cell for the two distinct response types in SCN^NMS+ neurons following optogenetic stimulation. (G) Average GCaMP8s fluorescence responses to optogenetic stimulation of PVT terminals in ACSF (black) or in the presence of TTX/4-AP (blue), or APV/CNQX (orange). Optogenetic stimulation (200 ms) is depicted by shaded area. (H) Quantification of area under the curve (AUC, 30 to 90 s, bottom) for each response type in ACSF, TTX/4-AP, and APV/CNQX. Data are means ± SEM. Type I, *P = 0.033, **P = 0.002; type II, **P = 0.004, ***P = 0.0003; two-way ANOVA, Šidák’s test.

Fig. 8.. The SCN differentially encodes stress and light information in SCN^AVP+ and SCN^VIP+ neurons.

(A) Schematic of stereotaxic injection strategy to selectively express GCaMP7s in SCN^AVP+ and SCN^VIP+ neurons, with representative images showing GCaMP7s expression and fiber placement. Scale bars, 100 μm. (B) Experimental paradigm. (C and D) Average calcium responses in SCN^AVP+ neurons to footshocks (C) and light (D) at CT14. *P = 0.019. (E) Comparison of footshock responses between SCN^AVP+ and SCN^VIP+ neurons at CT14. **P = 0.008. (F and G) Average calcium responses in SCN^VIP+ to footshocks (F) and light (G) at CT14. *P = 0.023. (H) Comparison of light responses between SCN^AVP+ and SCN^VIP+ neurons at CT14. **P = 0.009. (I and J) Average calcium responses in SCN^AVP+ to footshocks (I) and light (J) at CT22. (K) Comparison of footshock responses between SCN^AVP+ and SCN^VIP+ neurons at CT22. *P = 0.027. (L and M) Average calcium responses in SCN^VIP+ to footshocks (L) and light (M) at CT22. *P = 0.028. (N) Comparison of light responses between SCN^AVP+ and SCN^VIP+ neurons at CT22. *P = 0.010. (O and P) Average calcium responses in SCN^AVP+ to footshocks (O) and light (P) at CT6. *P = 0.046; **P = 0.008. (Q) Comparison of footshock responses between SCN^AVP+ and SCN^VIP+ neurons at CT6. **P = 0.008. (R and S) Average calcium responses in SCN^VIP+ to footshocks (R) and light (S) at CT6. (T) Comparison of light responses between SCN^AVP+ and SCN^VIP+ neurons at CT6. Line graphs depict GCaMP7s response quantification using the area under the curve (AUC) during baseline (0 to 10 s), footshock (10 to 20 s), and light (10 to 40 s). Footshock duration (2 s) is indicated by a red-shaded area, and light pulse (15 s) is indicated by an orange-shaded area. Baseline-versus-stimulus responses were analyzed using a two-tailed paired-sample t test. Comparisons between AVP⁺ and VIP⁺ neurons were performed using unpaired Student’s t test. ITI, intertrial interval.