Chronic psychosocial stress triggers microglial-/macrophage-induced inflammatory responses leading to neuronal dysfunction and depressive-related behavior

Abstract

Chronic environmental stress and traumatic social experiences induce maladaptive behavioral changes and is a risk factor for major depressive disorder (MDD) and various anxiety-related psychiatric disorders. Clinical studies and animal models of chronic stress have reported that symptom severity is correlated with innate immune responses and upregulation of neuroinflammatory cytokine signaling in brain areas implicated in mood regulation (mPFC; medial Prefrontal Cortex). Despite increasing evidence implicating impairments of neuroplasticity and synaptic signaling deficits into the pathophysiology of stress-related mental disorders, how microglia may modulate neuronal homeostasis in response to chronic stress has not been defined. Here, using the repeated social defeat stress (RSDS) mouse model we demonstrate that microglial-induced inflammatory responses are regulating neuronal plasticity associated with psychosocial stress. Specifically, we show that chronic stress induces a rapid activation and proliferation of microglia as well as macrophage infiltration in the mPFC, and these processes are spatially related to neuronal activation. Moreover, we report a significant association of microglial inflammatory responses with susceptibility or resilience to chronic stress. In addition, we find that exposure to chronic stress exacerbates phagocytosis of synaptic elements and deficits in neuronal plasticity. Importantly, by utilizing two different CSF1R inhibitors (the brain penetrant PLX5622 and the non-penetrant PLX73086) we highlight a crucial role for microglia (and secondarily macrophages) in catalyzing the pathological manifestations linked to psychosocial stress in the mPFC and the resulting behavioral deficits usually associated with depression.

Keywords: chronic stress; depressive-related behavior; innate immunity; macrophages; microglia; neuroinflammation; neuronal dysfunction; susceptibility/resilience to stress; synaptic pruning.

Figure 1 |. Chronic stress induces recruitment, proliferation and morphological activation of microglia in mPFC post-RSDS.

(a) Experimental outline of the repeated social defeat stress (RSDS) paradigm in mice. BrdU was administered ad libitum during the 10 days of RSDS to label proliferating cell populations. Behavioral testing (BH) was performed between D11-D15. Schematic of the sampled area within the mPFC; red-dashed rectangle demarcates the regions examined (Cg1, PrL, IL, MO). (b). Summary table of the behavioral effects of chronic stress in Susceptible (SD-Sus) and Resilient (SD-Res) groups. Relative comparisons to the Con groups: ↑ depicts increase, ↓ depicts decrease, <─> depicts no change. (c) Representative images of microglial/macrophage (Iba1; purple) density in the mPFC post-RSDS (D15) in low (10x magnification on the left; overlayed mouse brain areas examined at Bregma: 2.10 μm; Scale bar: 100 μm) and high (40x magnification on the right; Scale bar: 20 μm) magnifications. (d) Microglial/macrophage density (Iba1⁺; F_2,9 = 60.89, ****P ≤ 0.0001). (e) Representative images of microglial (CX3CR1-GFP; green) proliferation (BrdU; red) in the mPFC post-RSDS; Scale bar: 20 μm. (f) Microglial/macrophage density (CX3CR1⁺; F_2,8 = 109.9, ****P ≤ 0.0001). (g) Percentage of microglial proliferation (CX3CR1⁺BrdU⁺) during the RSDS (% BrdU⁺ of CX3CR1⁺; F_2,8 = 54.25, ****P ≤ 0.0001; indicated by white arrows). (h) Representative images of Iba1⁺ 100x stacks depicting microglial morphology with Sholl analysis micrographs (visualized by colored branch segments) in mPFC at D15. (i) Microglial branch intersections in relation to the distance from cell soma (F_16,238 = 143.8, ****P ≤ 0.0001). (j) Microglial cell body surface area (μm²) (F_2,14 = 62.66, ****P ≤ 0.0001). (k) Microglial branch complexity (F_2,14 = 12.46, ***P = 0.0008). For panels (d-g,j-k): one-way ANOVA with Tukey’s multiple comparisons. For panel (i): two-way ANOVA with Tukey’s multiple comparisons. For panel (d) data are reported as mean ± S.D. Con: n=4, SD-Sus: n=4, SD-Res: n=4. For panels (f-g) data are reported as mean ± S.D. Con: n=4, SD-Sus: n=4, SD-Res: n=3.For panels (i-k) data are reported as mean ± S.D. Con: n=5, SD-Sus: n=7, SD-Res: n=5 (12–16 traced cells /mouse). For panels (d-g,j-k): *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. For panel (i): *P ≤ 0.05 (Control vs SD-Sus), ^{^}P ≤ 0.05 (Control vs SD-Res) ^#P ≤ 0.05 (SD-Sus vs SD-Res). Statistical analysis details are reported in Supplementary Table 1; mPFC: medial prefrontal cortex; Cg1: cingulate cortex; PrL: prelimbic cortex; ΙL: infralimbic cortex; ΜΟ: medial orbital cortex; M1, M2: motor cortices.

Figure 2 |. Susceptibility/resilience to chronic stress is associated with the inflammatory state and responses of microglia/macrophages in the mPFC.

(a) Flow cytometry gating for CD11b (y axis) and CD45 (x axis) of mPFC cells isolated from CX3CR1⁺ mice (left), and graphical outline of experiment (top right). (b) Flow cytometry quantifications of the macrophage population (% CD45^hi /CD45⁺CD11b⁺) in the mPFC post-RSDS (F_2,6 = 10.53, *P = 0.0109). (c) Immunoblot analyses of mPFC protein levels for VCAM-1 and ICAM-1 at D15 post RSDS. (d) VCAM-1 (F_2,10 = 26.95, ****P ≤ 0.0001), ICAM-1 (F_2,10 = 17.79, ***P = 0.0005) (e) Flow cytometry histogram of pro-inflammatory marker CD86 in microglia and macrophages (normalized to mode). (f) Microglial CD86⁺ expression (MFI; F_2,6 = 17.81, **P = 0.003) (g) Macrophage CD86⁺ expression (MFI; F_2,6 = 19.28, **P = 0.0024) (h) Flow cytometry histogram of the anti-inflammatory marker CD206 in microglia and macrophages (normalized to mode). (i) Microglial CD206⁺ expression (MFI; F_2,6 = 20.06, **P = 0.0022) (j) Macrophage CD206⁺ expression (MFI; F_2,6 = 28.02, ***P = 0.0009) (k) Immunoblot analyses for CD86 and CD206 in extracts from the mPFC at D15 post RSDS. (l) CD86 (F_2,10 = 39.54, ****P ≤ 0.0001), CD206 (F_2,8 = 342.3, ***P = 0.0005) (m) Immunoblot analyses for iNOS and Arg1 in extracts from the mPFC at D15 post RSDS. (n) iNOS (F_2,10 = 20.96, ****P = 0.0003), Arg1 (F_2,8 = 17.05, ***P = 0.0006) (o) Representative images of microglial/macrophage reactivity state in the mPFC on D15 post-RSDS [(Iba1; purple) CD68 (phagolysosome; yellow) and iNOS (pro-inflammatory marker; cyan)]; Scale bar: 20 μm. (p) Microglial CD68⁺ integrated density (F_2,9 = 253.3, ****P ≤ 0.0001) (q) Microglial iNOS⁺ integrated density (F_2,9 = 2393, ****P ≤ 0.0001) (r) Microglial CD68⁺iNOS⁺ integrated density (F_2,9 = 1998, ****P ≤ 0.0001; indicated by white arrows and displayed in the insets of (ο). One-way ANOVA was performed with Tukey’s multiple comparisons. For panels (b, f-j) data are reported as mean ± S.D. Con: n=3, SD-Sus: n=3, SD-Res: n=3. For panels (d, l top, n) data are reported as mean ± S.D. Con: n=5, SD-Sus: n=4, SD-Res: n=4.For panel (l bottom) data are reported as mean ± S.D. Con: n=3, SD-Sus: n=5, SD-Res: n=3. For panels (p-r) data are reported as mean ± S.D. Con: n=4, SD-Sus: n=4, SD-Res: n=4. For all graph panels: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Statistical analysis details are reported in Supplementary Table 1; MFI: mean fluorescent intensity. The representative gels displayed in panels (c, k, m) are cropped and the original images are shown in Supplementary Fig. 8.

Figure 3 |. Microglial phagocytosis of synaptic elements is exacerbated in mPFC post-chronic stress.

Representative images of microglia (CX3CR1-GFP; blue), phagolysosomes (CD68; cyan), and (a) the postsynaptic neuronal marker (PSD95; red; yz planes on the right depict microglial cells with phagocytosed PSD95), or (b) the presynaptic neuronal marker (Syn-1; yellow; yz planes on the right depict microglial cells with phagocytosed Syn-1) in the mPFC on D15 post-RSDS; Scale bar: 10 μm. (c) PSD95⁺ integrated density (F_2,8 = 10.10, **P = 0.0065) (d) % Microglial CD68⁺PSD95⁺ of total PSD95⁺ integrated density (F_2,8 = 41.67, ****P ≤ 0.0001) (e) Syn-1⁺ integrated density (F_2,8 = 14.71, **P = 0.0021). (f) % Microglial CD68⁺Syn-1⁺ of total Syn-1⁺ integrated density (F_2,8 = 37.31, ****P ≤ 0.0001). One-way ANOVA was performed with Tukey’s multiple comparisons. Data are reported as mean ± S.D. Con: n=4, SD-Sus: n=4, SD-Res: n=3. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Statistical analysis details are reported in Supplementary Table 1.

Figure 4 |. Chronic stress induces microglia phagocytosis of myelin and neuronal plasticity alterations post-RSDS.

(a) Representative images of microglia (CX3CR1-GFP; blue), phagolysosomes (CD68; cyan) and myelin (CNP; red). Zoom in micrographs on the right; Scale bar: 20 μm. (b) CNP integrated density levels (F_2,6 = 6.621, *P = 0.003). (c) % Microglial CD68⁺CNP⁺ of total CNP⁺ integrated density (F_2,6 = 101.7, ****P < 0.0001). (d) Representative images of neuronal somas (NeuN; blue) and neuronal plasticity marker (p-CREB; red). (e) % Neuronal p-CREB⁺ of total NeuN⁺ integrated density (F_2,6 = 427.4, ****P < 0.0001). For all panels: One-way ANOVA with Tukey’s multiple comparisons. Data are reported as mean ± S.D. Con: n=3, SD-Sus: n=3, SD-Res: n=3. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Statistical analysis details are reported in Supplementary Table 1.

Figure 5 |. Chronic stress induces monocytic infiltration into the mPFC, and depressive-like phenotype in mice, which is partially rescued after peripheral inhibition of CSF1R via PLX73086.

(a) Schematic of the generation of the tamoxifen (TMX)-inducible CX3CR1-Cre^ERT2-eYFP::Rosa26-tdTom mouse line. CX3CR1-Cre^ERT2-eYFP mice were crossed with Rosa26-tdTom animals to label microglia/macrophages following TMX-induced recombination. (b) Experimental outline of i.p. TMX administration in 1 month-old transgenic mice (Daily dose: 75 μg/g body weight; 5 consecutive days); RSDS paradigm was performed 1 month after the induction of Cre and behavioral tests were performed 2 weeks after the end of the paradigm (BH: D22-D25) (c) Representative images of microglia (YFP⁺tdTom⁺) and macrophages (YFP⁺tdTom^-) colocalizing with ICAM-1 (purple). (d) Microglial density (YFP⁺tdTom⁺; F_2,4 = 83.16, ***P = 0.0006). (e) % of macrophage population in the mPFC (% of tdTom^- of YFP⁺; F_2,4 = 91.03, ***P ≤ 0.0005) (f) total ICAM-1⁺ integrated density (F_2,4 = 303.0, ****P ≤ 0.0001) (g) CX3CR1-YFP⁺ICAM-1⁺ integrated density (F_2,4 = 167.0, ***P = 0.001) (h) Graphical outline of RSDS paradigm and treatment with Control or PLX73086 chow. (i) Summary table of behavioral effects for experimental groups with no treatment (left) and with PLX73086 treatment (right). (j) Social interaction ratio (F_5,34 = 60.34, ****P ≤ 0.0001).(k) Time spent in closed arms (left; F_5,34 = 48.16, ****P ≤ 0.0001) and open arms (right; F_5,34 = 46.16, ****P ≤ 0.0001). (i) Immobility time (F_5,34 = 19.49, ****P ≤ 0.0001). (m) % Sucrose preference (F_5,34 = 32.92, ****P ≤ 0.0001). For panels (j-m) only the most relevant comparisons are depicted on the graph. More details are shown in Supplementary Table 1. One-way ANOVA was performed with Tukey’s multiple comparisons. Data are reported as mean ± S.D. For panels (d-g) Con: n=3, SD-Sus: n=3, SD-Res: n=2. For panels (j-m) Control Chow: (Con: n=11, SD-Sus: n=8, SD-Res: n=2) and PLX73086 Chow: (Con: n=7, SD-Sus: n=6, SD-Res: n=6) *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Statistical analysis details are reported in Supplementary Table 1.

Figure 6 |. Microglial/macrophage elimination via the CSF1R inhibitor PLX5622 protects from neuronal plasticity deficits and the depressive- like behavior post RSDS.

(a) Graphical outline of RSDS paradigm and treatment with PLX5622 or Control chow. (b) Summary table of behavioral effects for experimental groups with no treatment (left) and with PLX5622 treatment (right). (c) Immunoblot analysis for TSPO in the mPFC at D15 post RSDS (top) and quantification (bottom; F_4,16 = 15.74, ****P ≤ 0.0001); PLX-Con group immunoblot analysis is displayed in Supplementary Fig. 9. (d) Representative heatmaps of social Interaction ratio (left) and corresponding quantifications (right; F_4,72 =148.6, ****P ≤ 0.0001) (e) Immunoblot analyses for PSD95 and Synapsin-1 in the mPFC at D15 post RSDS (left). Analyses (right) of PSD95 (F_4,12 = 26.68 ****P ≤ 0.0001) and Synapsin-1 (F_4,12 = 22.64, ****P ≤ 0.0001); PLX-Con group immunoblot analysis is displayed in Supplementary Fig. 9. (f) Immunoblot analyses for BDNF and 5-HTR1A in the mPFC at D15 post RSDS (left). Analyses (right) of BDNF (F_4,15 = 15,69 ****P ≤ 0.0001) and 5-HTR1A (F_4,16 = 8.751, ***P = 0.0006); PLX-Con group immunoblot analysis is displayed in Supplementary Fig. 10. (g) Graphical outline of RSDS paradigm with Control or PLX5622 chow treatment (RSDS/Treatment phase; D0-D15), and microglial repopulation (Repopulation phase; D15-D25). BH1 and BH2: behavioral tests. (h) Representation images (left) and corresponding quantifications (right; F_2,13 = 224.4 ****P ≤ 0.0001) of microglial depletion (Day 15) and repopulation (Day 25). (i) Overall social interaction ratio quantifications (left; F_5,31 = 60.77 ****P ≤ 0.0001) and pairwise quantifications (right; ****P ≤ 0.0001). Only the most relevant comparisons are depicted on the graph. More details are shown in Supplementary Table 1. (j) Summary table of behavioral effects for experimental groups with no treatment (left) and PLX5622 withdrawal (right). For all panels with graphs one-way ANOVA was performed with Tukey’s multiple comparisons, except for (i) on the right for which paired two tailed T-test was performed. Data are reported as mean ± S.D. For panel (c) Control Chow: (Con: n=3, SD-Sus: n=4, SD-Res: n=4) and PLX5622 Chow: (Con: n=3, SD-Res: n=5). For panel (d; right) Control Chow: (Con: 22, SD-Sus: n=18, SD-Res: n=5) and PLX5622 Chow: (Con: n=13, SD-Res: n=19). For panels (e-f) Control Chow: (Con: 3–4, SD-Sus: n=4–8, SD-Res: n=3–4) and PLX5622 Chow: (Con: n=3, SD-Res: n=3–5) *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Statistical analysis details are reported in Supplementary Table 1. The representative gels displayed in panels (c, e, f) are cropped and the original images are shown in Supplementary Figs. 9 and 10.