Exercise-Activated mPFC Tri-Synaptic Pathway Ameliorates Depression-Like Behaviors in Mouse

Abstract

Exercise is considered as playing a pivotal role in the modulation of emotional responses. However, a precise circuit that mediates the effects of exercise on depression have yet to be elucidated. Here, a molecularly defined tri-synaptic pathway circuit is identified that correlates motor inputs with antidepressant effects. With this pathway, initial inputs from neurons within the dorsal root ganglia (DRG) project to excitatory neurons in the gracile nucleus (GR), which in turn connect with 5-HTergic neurons in the dorsal raphe nucleus (DRN), eventually coursing to excitatory pyramidal neurons within the medial prefrontal cortex (mPFC). Exercise activates this pathway, with the result that depressive- and anxiety-like behaviors in mice are significantly reduced. In addition, it is found that exercise may exert antidepressant effects through regulating synaptic plasticity within this tri-synaptic pathway. These findings reveal a hindbrain-to-forebrain neuronal circuit that specifically modulates depression and provides a potential mechanism for the antidepressant effects of exercise.

Keywords: depression; dorsal raphe nucleus; exercise; medial prefrontal cortex; synaptic plasticity.

Figure 1

Exercise decreases depressive‐like behaviors induced by chronic stress. A) Schematic of the experimental design for the CRS model. B–E) Depressive‐ and anxiety‐like behaviors in the different experimental groups of the CRS model (n = 8 animals per group). SPT (B), FST (C), EPM (D), and OFT (E). F) Schematic of the experimental design for the CSDS model. G) Social interactions in the different experimental groups of the CSDS model. H–L) Depressive‐, stress‐induced‐, and anxiety‐like behaviors in the different experimental groups of the CSDS model (n = 8 animals per group). EPM (H), SPT (I), FST (J), and OFT (K). For all statistical tests: One‐way analysis of variance (ANOVA) with Bonferroni post‐hoc test. *, P < 0.05; **, P < 0.01; NS, no significant difference. Data are presented as Means ± SEMs. Dots represent individual mice.

Figure 2

Exercise‐induced activation of GR CaMKII‐α⁺ neurons directly triggers DRN 5‐HT⁺∩CaMKII‐α⁺ neurons. A) Typical image of Fos expression in DRG, GR, DRN, and mPFC. Scale bar: 50 µm. B) Quantification of c‐Fos+ cell numbers in control and runing mice (n = 8 mice per group). C) Schematic of virus injection and typical images of mCherry expression in GR. Scale bar, 100 µm. (n = 6 mice per group). D) Typical image and quantification of DRN‐projecting GR neurons co‐expressed with CaMKII/GAD2 and quantification percent. Scale bar: 20 µm. (n = 6 mice per group). E) Schematic of virus injection and typical images of mCherry expression in DRN. Scale bar: 100 µm. F) Typical image and quantification of DRN postsynaptic neurons co‐expressed with CaMKII∩TPH2 or GAD2∩TPH2 and quantification percent. Scale bar: 20 µm. G) Schematic for monitoring calcium activities of GR‐projecting DRN. Scale bar: 100 µm. H,I) Heatmap H) and quantifications (I) of calcium signal changes of GR‐projecting DRN, as aligned to the onset (time = 0 s) of running (n = 5 bouts from 5 mice). Two‐tailed unpaired Student's t‐test for (B), (D), and (F). ***p < 0.001; ****p < 0.0001; NS, not significant. Data are presented as Means ± SEMs.

Figure 3

The GR‐DRN pathway directly innervates mPFC CaMKII‐α⁺ neurons. A) Experimental design used for labeling axons and presynaptic terminals of DRN postsynaptic neurons. Representative image of injection site in DRN. Scale bar: 100 µm. B) Synaptic puncta in the mPFC, CA1, DG, LHb, and amygdala from DRN postsynaptic neurons. Scale bar: 50 µm (n =  6 brain slices from 3 mice per group). C) Schematic of virus injection and typical images of mCherry expression in the mPFC. Scale bar, 100 µm. D) Typical image and quantification of DRN‐projecting mPFC neurons co‐expressed with CaMKII/GAD2 and quantification percent. Scale bar: 50 µm (n = 6 mice per group). E) Schematic depicting recordings of 5‐HT levels in the mPFC and representative image of virus expression. Scale bar: 100 µm. F,G) Heatmap (F) and quantifications (G) of 5‐HT level changes in DRN‐projecting mPFC neurons, as aligned to the onset (time = 0 s) of running (n = 5 bouts from 5 mice). H) Schematic for monitoring calcium activities of DRN‐projecting mPFC. Scale bar: 100 µm. I‐J) Heatmap (I) and quantifications (J) of calcium signal changes of DRN‐projecting mPFC, as aligned to the onset (time = 0 s) of running (n = 5 bouts from 5 mice). Two‐tailed unpaired Student's t‐test for (D). ****p < 0.0001; Data are presented as Means ± SEMs.

Figure 4

Activation of the GR‐DRN‐mPFC pathway decreases depressive‐like behaviors induced by chronic restraint stress. A) Schematic of the experimental design for the CRS model. B) Schematic of virus injection and typical images of mCherry expression in DRN. Scale bar: 100 µm. C) Representative images and quantification of DRN showing c‐Fos expression in neurons expressing hM3Dq in response to an i.p. injection of saline or CNO. Scale bar: 20 µm (n = 6 mice per group). D) Representative images and quantification of mPFC showing c‐Fos expression in response to an i.p. injection of saline or CNO. Scale bar: 50 µm (n = 6 mice per group). E) Representative traces of voltage responses in DRN mCherry⁺ neurons to injection currents from 0 to 100 pA with 20 pA steps before and after bath application of CNO (5 µm). Red traces indicate minimal currents needed to induce APs. F) CNO increases spiking of hM3Dq‐expressing neurons (n = 6 cells from 3 mice per group). G,H) Depressive‐like and stress‐induced behaviors in the different experimental groups of the CRS model (n = 8 animals per group). SPT (G) and FST (H). I) Schematic of virus injection and typical images of mCherry expression in the mPFC. Scale bar: 100 µm. J) Representative images and quantification of the mPFC showing c‐Fos expression in neurons expressing hM3Dq in response to an i.p. injection of saline or CNO. Scale bar: 20 µm. N = 6 mice per group. K) Representative traces and analyses of voltage responses in mPFC mCherry⁺ neurons to injection currents from 0 to 100 pA with 20 pA steps before and after bath application of CNO (5 µm). Red traces indicate minimal current needed to induce APs. L) CNO decreases spiking of hM3Dq‐expressing neurons (n = 6 cells from 3 mice per group). M,N) Depressive‐like behaviors in the different experimental groups of the CRS model (n = 8 animals per group). SPT (M) and FST (N). Two‐tailed unpaired Student's t‐test for (C), (D), and (J) and paired Student's t‐test for (F) and (L). For (G), (H), (M), and (N): One‐way analysis of variance (ANOVA) with Bonferroni post‐hoc test. *, P < 0.05; **, P < 0.01; ***p < 0.001; ****p < 0.0001; NS = no significant difference. Data are presented as Means ± SEMs.

Figure 5

Inhibition of the GR‐DRN‐mPFC pathway prevents exercise‐induced antidepressant effects. A) Schematic of the experimental design for the CRS model. B) Schematic of virus injection and typical images of mCherry expression in DRN. Scale bar: 100 µm. C) Representative images and quantification of DRN showing c‐Fos expression in neurons expressing hM4Di in response to an i.p. injection of saline or CNO. Scale bar: 20 µm (n = 6 mice per group). D) Representative images and quantification of mPFC showing c‐Fos expression in response to an i.p. injection of saline or CNO. Scale bar: 50 µm (n = 6 mice per group). E) Representative traces of voltage responses in DRN mCherry⁺ neurons to injection currents from 0 to 100 pA with 20‐pA steps before and after bath application of CNO (5 µm). Red traces indicate minimal currents needed to induce APs. F) CNO increases spiking of hM4Di‐expressing neurons (n = 6 cells from 3 mice per group). G,H) Depressive‐like and stress‐induced behaviors in the different experimental groups of the CRS model (n = 8 animals per group). SPT (G) and FST (H). I) Schematic of virus injection and typical images of mCherry expression in the mPFC. Scale bar: 100 µm. J) Representative images and quantification of the mPFC showing c‐Fos expression in neurons expressing hM4Di in response to an i.p. injection of saline or CNO. Scale bar: 20 µm (n = 6 mice/group). K) Representative traces and analyses of voltage responses in mPFC mCherry⁺ neurons to injection currents from 0 to 100 pA with 20 pA steps before and after bath application of CNO (5 µm). Red traces indicate minimal currents needed to induce APs. L) CNO decreases spiking of hM4Di‐expressing neurons (n = 6 cells from 3 mice per group). M,N) Depressive‐like behaviors in the different experimental groups of the CRS model (n = 8 animals/group). SPT (M) and FST (N). Two‐tailed unpaired Student's t‐test for (C), (D), and (J) and paired Student's t‐test for (F) and (L). For (G), (H), (M), and (N): One‐way analysis of variance (ANOVA) with Bonferroni post‐hoc test. *, P < 0.05; **, P < 0.01; ***p < 0.001; ****p < 0.0001; NS = no significant difference. Data are presented as Means ± SEMs.

Figure 6

Exercise promotes synaptic plasticity of neurons within DRN. A) Schematic of the experimental design for the CRS model. B,C) Representative western blots showing that exercise increased expression levels of synaptic structural proteins (B) and TPH2 (C) in DRN of mice exposed to CRS. D) Quantification of protein expression levels of PSD95, Synaptophysin, BDNF and TPH2 within the DRN region. E,F) Representative western blots and quantification of p‐CREB/CREB (E) and P‐CAMKII/CAMKII (F) within the DRN region. G) Immunofluorescent staining showing VGLUT1⁺ (red) and PSD95⁺ (green) co‐localization in DRN neurons. Top: scale bar: 10 µm. Bottom: scale bar: 5 µm (n = 6 brain slices from 3 mice for each group). H) The quantification of the synaptic number in DRN with each group. I) Representative traces of sEPSCs in DRN neurons with each group. J,K) Cumulative fraction plots of sEPSCs amplitude (J) and frequencies(K) (n = 8 cells from 3 mice/group). For (B‐F): n = 6 animals/group. One‐way analysis of variance (ANOVA) with Bonferroni post‐hoc test. *, P < 0.05; **, P < 0.01; NS, no significant difference. Data are presented as Means ± SEMs.

Figure 7

Exercise promotes synaptic plasticity of neurons within the mPFC. A) Schematic of the experimental design for the CRS model. B,C) Representative western blots showing that exercise increased expression levels of synaptic structural proteins (B) and TPH2 (C) in the mPFC of mice exposed to CRS. D) Quantification of protein expression levels of PSD95, SYN, BDNF and TPH2 within the mPFC region. E,F) Representative western blots and quantification of p‐CREB/CREB (E) and P‐CAMKII/CAMKII (F) within the mPFC region. G) Immunofluorescent staining showing VMAT2⁺ (red) and PSD95⁺ (green) co‐localization in DRN neurons. Top: scale bar: 10 µm. Bottom: scale bar: 5 µm (n = 6 brain slices from 3 mice for each group). H) The quantification of the synaptic number in mPFC with each group. I) Representative traces of sEPSCs in mPFC neurons with each group. J,K) Cumulative fraction plots of sEPSCs amplitude (J) and frequencies(K) (n = 8 cells from 3 mice per group). For (B‐F): n = 6 animals per group. One‐way analysis of variance (ANOVA) with Bonferroni post‐hoc test. *, P < 0.05; **, P < 0.01; NS, no significant difference. Data are presented as Means ± SEMs.