Microglial Engulfment of Spines in the Ventral Zona Incerta Regulates Anxiety-Like Behaviors in a Mouse Model of Acute Pain

Abstract

Although activation of microglial cells is critical in developing brain disorders, their role in anxiety-like behaviors in pain is still vague. This study indicates that alteration of microglia's neuronal spine engulfment capacity in ventral zona incerta (ZI _V ) leads to significant pain and anxiety-like behaviors in mice 1-day post-injection of Complete Freud's Adjuvant (CFA1D). Performing whole-cell patch-clamp recordings in GABAergic neurons in the ZI _V (ZI _V ^GABA ) in brain slices, we observed decreased activity in ZIv ^GABA and reduced frequency of the miniature excitatory postsynaptic currents (mEPSCs) in ZI _V ^GABA of CFA1D mice compared with the saline1D mice. Besides, chemogenetic activation of ZI _V ^GABA significantly relieved pain and anxiety-like behaviors in CFA1D mice. Conversely, in naïve mice, chemogenetic inhibition of ZI _V ^GABA induced pain and anxiety-like behaviors. Interestingly, we found changes in the density and morphology of ZI _V ^Microglia and increased microglial engulfment of spines in ZI _V of CFA1D mice. Furthermore, pain sensitization and anxiety-like behaviors were reversed when the ZI _V ^Microglia of CFA1D-treated mice were chemically inhibited by intra-ZI _V minocycline injection, accompanied by the recovery of decreased ZI _V ^GABA excitability. Conclusively, our results provide novel insights that dysregulation of microglial engulfment capacity encodes maladaptation of ZI _V ^GABA , thus promoting the development of anxiety-like behaviors in acute pain.

Keywords: GABAergic neurons; anxiety-like behaviors in pain; chemogenetic manipulation; dendritic spines; inflammatory pain model; microglial engulfment; zona incerta.

FIGURE 1

CFA1D mice demonstrate anxiety-like behaviors in acute pain. (A) Schematic diagram of the experiment paradigms for the CFA injection. (B) Time course of CFA-induced mechanical sensitivity on days 1, 3, 7, 10, and 14 post injection. Mechanical pain threshold; two-way ANOVA with Bonferroni post hoc analysis [Time × group interaction, F(5,854) = 3.930 and P = 0.0016]. (C) Performance of stimulus intensity of the contralateral and ipsilateral hind paw in saline1D- and -CFA1D mice (t₁₄ = 1.741 and P = 0.1037 in contralateral hind paw, t₁₄ = 4.393 and P = 0.006 in ipsilateral hind paw). Two-tailed unpaired t-test. (D) Representative animal heat tracks from saline1D- and CFA1D mice in EPM and OFT. (E,F) Performance of EPM and OFT on saline1D- and CFA1D mice (time in open arm; t₁₄ = 6.200 and P < 0.001, open arm entries; t₁₄ = 3.005 and P = 0.095, time in center; t₁₄ = 3.76 and P = 0.0021, distance traveled; t₁₄ = 0.56 and P = 0.58). (G) Performance of stimulus intensity of the contralateral and ipsilateral hind paw in saline7D- and CFA7D- mice (t₁₄ = 1.185 and P = 0.2556 in contralateral hind paw, t₁₄ = 4.193 and P = 0.0009 in ipsilateral hind paw). Two-tailed unpaired t-test. (H) Representative animal heat tracks from saline7D- and CFA7D mice in EPM and OFT. (I,J) Performance of EPM and OFT on saline7D- and CFA7D- mice (time in open arm; t₁₆ = 1.118 and P = 0.2799, open arm entries; t₁₄ = 1.414 and P = 0.1792, time in center; t₁₄ = 0.5309 and P = 0.6038, distance traveled; t₁₄ = 0.9066 and P = 0.38). All data are expressed as mean ± SEM. n = 8 mice per group. All data are expressed as mean ± SEM. **P < 0.01, and ***P < 0.001; ns, not significant.

FIGURE 2

CFA1D mice demonstrate a decrease in the ZI_V^GABA activity. (A) Confocal images of ZI_V^GABA from GAD-tdTOM mice. The white box depicts the area shown in the box of the ZI_V. Scale bar, 50 μm. (B) Sample traces and (C) summarized statistical data for action potential firing recorded from ZI_V^GABA in mice treated with saline1D or CFA1D. Two-way ANOVA with Bonferroni post-tests [Time × group interaction, F(5,180) = 6.735 and P < 0.05]. (D–E) The alterations in action potential properties in (D), rheobase; t₃₆ = 0.53 and P = 0.6012; (E) V_rest (mV); t₃₆ = 0.527 and P = 0.6. (F) Sample traces of the mEPSCs from saline1D or CFA1D groups. (G,H) Summarized statistical data of (G) the frequency; t₆₆ = 2.23 and P = 0.02; (H) the amplitude; t₆₆ = 1.12 and P = 0.26 of CFA1D compared with saline1D mice. Action potential parameters were estimated from the initial evoked spike. n = 17–36 neurons per group. (I) Confocal images of ZI_V^GABA from GAD-tdTOM mice. The white box depicts the area shown in the box of the ZI_V. Scale bar, 50 μm. (J) Sample traces and (K) summarized statistical data for action potential firing recorded from ZI_V^GABA in mice treated with saline7D or CFA7D. Two-way ANOVA with Bonferroni post-tests [Time × group interaction, F(1,21) = 1.403 and P = 0.2495]. (L,M) The alterations in action potential properties in (L), rheobase; t₃₈ = 1.558 and P = 0.127; (M) V_rest (mV); t₃₈ = 0.526 and P = 0.6. (N) Sample traces of the mEPSCs from saline7D or CFA7D groups. (O,P) Summarized statistical data of (O) the frequency; t₄₅ = 1.26 and P = 0.21; (P) the amplitude; t₄₅ = 0.72 and P = 0.47 of CFA7D compared with saline7D mice. Action potential parameters were estimated from the initial evoked spike. n = 22–25 neurons per group. All data are presented as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant.

FIGURE 3

Manipulation of ZI^GABA activity regulates pain perception and anxiety-like behaviors. (A) Schematic showing unilateral injection of ZI_V for chemogenetic manipulation and (B) diagram of viral injection of AAV-DIO-hM3Dq-mCherry (hM3Dq-mCherry) in the ZI_V of GAD2-Cre mice. The scale bar is 200 μm. (C) A representative trace displays that bath application of CNO (10 μM) depolarized ZIv^GABA, and statistics (right) show the average magnitude of depolarization (One sample t-test, t₄ = 11.45 and P = 0.03). (D–G) Chemogenetic activation of ZIv^GABA in GAD2-Cre mice affects anxiety-like behaviors in acute pain (Time in open arms: mCherry vs. hM3Dq, P = 0.0469, time in the center; mCherry vs. hM3Dq, P = 0.0036). (D) Mechanical pain threshold; two-way ANOVA with Bonferroni post-tests [Time × group interaction, F(5,65) = 2.235 and P = 0.01612]. (E) Representative animal heat tracks in mCherry vs. hM3Dq mice in EPM and OFT. (F) Time in open arms; ordinary one-way ANOVA with Turkeys’ post-test (mCherry vs. hM3Dq, P = 0.0469). (G) Time in the center; ordinary one-way ANOVA with Turkeys’ post-test (mCherry vs. hM3Dq, P = 0.0015). n = 7 mice per group. (H) Schematic of viral injection in the ZI_V for chemogenetic manipulation and (I) diagram of bilateral injection of Cre recombinase-dependent AAV virus expressing hM4Di or mCherry AAV-DIO-hM4Di-mCherry (hM4Di-mCherry) into the ZI_V of GAD2-Cre mice. The scale bar is 200 μm. (J) A representative trace from a whole-cell current-clamp electrophysiological recording shows that bath application of CNO (10 μM) hyperpolarizes ZIv^GABA. Statistics (right) show the average magnitude of hyperpolarization (One sample t-test, t₆ = 11.71 and P < 0.0001). (K–N) Chemogenetic inhibition of ZIv^GABA in GAD2-Cre mice induces anxiety-like behaviors and pain in naïve mice. (K) Mechanical pain threshold; two-way ANOVA with Bonferroni post-tests [Time × group interaction, F(5,65) = 3.199 and P = 0.0121]. (L) Representative animal heat tracks in mCherry vs. hM4Di mice in EPM and OFT. (M) Time in open arms; ordinary one-way ANOVA with Turkeys’ post-test (mCherry vs. hM4Di, P = 0.0036). (N) Time in the center; ordinary one-way ANOVA with Turkeys’ post-test (mCherry vs. hM4Di, P = 0.0024). n = 7–8 mice per group. All data are expressed as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, and ns, not significant.

FIGURE 4

Increased microglial cell activity and engulfment of ZI_V^GABA spines in CFA1D mice. (A) Fluorescence images displaying Iba-1 immunostaining (red) and 3D reconstruction of microglia in the ZI_V of saline1D-, CFA1D-, saline7D- and CFA7D mice. The scale bars are 40 μm (overview) and 10 μm (inset and rendering). (B,C) Statistical data for the number of Iba-1⁺ cells per cubic millimeter (t₁₀ = 2.46 and P = 0.029) and Iba-1 intensity (t₁₀ = 5.42 and P = 0.003) in the ZI_V from saline1D- and CFA1D mice. (D,E) Summary data of microglia branches endpoints (t₁₈ = 4.07 and P = 0.0007) and total process lengths per cell of Iba-1⁺ microglia (t₁₈ = 4.43 and P = 0.0003) in ZIv from saline1D and CFA1D mice. (F,G) Statistical data for the number of Iba-1⁺ cells per cubic millimeter (t₁₄ = 0.9354 and P = 0.365) and Iba-1 intensity (t₁₄ = 1.765 and P = 0.0994) in the ZI_V from saline7D- and CFA7D mice. (H,I) Summary data of total microglia endpoints (t₁₄ = 0.06 and P = 0.953), and total branches lengths per cell of Iba-1⁺ microglia (t₁₄ = 0.965 and P = 0.3505) in ZIv from saline7D and CFA7D mice. (J,K) Representative images (left) and quantitative analyses (right) of immunostaining for Iba-1 (green) and MHCII (red) (t₁₀ = 6.178 and P = 0.0001) in the ZI_V of saline1D- or CFA1D-treated mice. Scale bar, 10 μm. (L,M) Representative images (left) and quantitative analyses (right) of immunostaining for Iba-1 (green) and MHCII (red) (t₁₀ = 1.799 and P = 0.1022) in the ZI_V of saline7D- or CFA7D-treated mice. Scale bar, 10 μm. (N) qPCR analysis of IL-1β (t₁₀ = 2.719 and P = 0.0216), IL6 (t₁₀ = 2.901 and P = 0.0158), and TNFα (t₁₀ = 2.740 and P = 0.0208) mRNA in ZI_V tissues from saline1D and CFA1D mice. (O) Illustrative images and 3D reconstruction surface rendering of Iba-1⁺ microglia (green) including PSD95⁺ puncta (red) and DAPI (blue) in the ZI_V from saline1D- and CFA1D- treated mice. Scale bars, 10 μm (overview) and 2 μm (inset and rendering). (P) PSD95⁺ puncta’s quantification of microglia in slices (t₂₆ = 2.71 and P = 0.011, n = 14–16 slices per group from four mice) as showed in (O). (Q,R) Representative images (left) and quantitative analyses (right) of immunostaining for Iba-1 (green), CD68 (purple), and PSD95 (red) (t₁₀ = 3.212 and P = 0.0093) in the ZI_V of saline1D- or CFA1D-treated mice. Scale bar, 10 μm. All data are presented as mean ± SEM. n = 6–10 slices per group from three mice. Unpaired-t test. *P < 0.05; **P < 0.01, and ***P < 0.001; ns, not significant.

FIGURE 5

Inactivating microglial cells in ZI_V of CFA1D mice reverses anxiety-like behavior in pain and the dynamic modifications of microglial status. (A) Schematic diagram of the experiment and cannula implantation in ZI_V. (B) Stimulus intensity of the ipsilateral (P = 0.002) and contralateral (P > 0.9) hind paws CFA1D mice during intra- ZI_V injection of minocycline (10 μg/μL) or ACSF, Ordinary one-way ANOVA with Bonferroni’s multiple comparisons tests. (C–E) Heat map visualization and statistics data from EPM [Time in open arm (P = 0.0024), open-arm entries] and OFT [Time in the center (P = 0.0114) and distance traveled] apparatuses in CFA1D mice during intra- ZI_V injection of minocycline (10 μg/μL) or ACSF, n = 8 mice per group. (F) Fluorescence images of Iba-1⁺ cells from CFA1D mice with saline or minocycline, scale bar, 20 μm. Sections show a detailed view of the white box; scale bar, 10 μm. (G–J) Statistical data for the Iba-1⁺ cell’s number, Iba-1 intensity, endpoints, and process lengths per cell. n = 6 slices per group from three mice. (G,H) Statistical data for the number of Iba-1⁺ cells per cubic millimeter (Iba-1⁺ cells; t₁₀ = 2.448 and P = 0.0344) and Iba-1 intensity (Iba-1 intensity; t₁₀ = 3.29 and P = 0.0081) in the ZI_V from saline- and minocycline treated CFA1D mice. (I,J) Summary data of microglia total branches endpoints (t₁₆ = 3.054 and P = 0.0076) and all process lengths per cell of Iba-1⁺ microglia (t₁₆ = 3.65 and P = 0.0022) in ZI_V from saline- and minocycline treated CFA1D. All data are expressed as mean ± SEM. n = 6–11 slices per group from three mice. *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant.

FIGURE 6

Blockade of microglial activation affects ZI_V^GABA activity and microglial engulfment of ZI_V neuronal spines in CFA1D mice. (A) Sample traces and (B) summarized statistical data for action potential firing recorded from ZI_V^GABA in CFA1D mice treated with saline or minocycline. Two-way ANOVA with Bonferroni post-tests [Time × group interaction, F(5,175) = 3.78 and P = 0.0028]. (C,D) The alterations in action potential properties in (C) rheobase (t₃₅ = 0.98 and P = 0.33); (D) V_rest (mV) (t₂₆ = 0.98 and P = 0.33); (E) sample traces of the mEPSCs from CFA1D mice treated with saline or minocycline. (F,G) Summarized statistical data of (F) the frequency (t₃₁ = 3.05 and P = 0.0046); (G) the amplitude (t₃₀ = 1.92 and P = 0.06) of CFA1D mice treated with saline or minocycline. Action potential parameters were estimated from the initial evoked spike. n = 16–21 neurons per group. (H) Representative images and quantification of Iba-1⁺ microglia (green) containing PSD95⁺ puncta (red) in the ZI_V from CFA-treated mice treated with minocycline or saline. Scale bars, 10 μm (overview) and 2 μm (inset and rendering). (I) PSD95⁺ puncta’s quantification of microglia in slices (t₂₆ = 2.55 and P = 0.016). n = 14–16 slices per group from four mice. (J,K) Representative images (left) and quantitative analyses (right) of immunostaining for Iba-1 (green), CD68 (purple), and PSD95 (red) in the ZI_V of saline1D- or CFA1- mice treated with minocycline slices (t₁₀ = 3.117 and P = 0.0109). Scale bar, 10 μm. Unpaired t-test. All data are presented as mean ± SEM. *P < 0.05 and **P < 0.01; ns, not significant.

FIGURE 7

The interactions of the microglial processes and dendritic spines of ZI_V^GABA. (A) Representative immunofluorescence images and 3D rendering of CD68 (purple), Iba-1 (red), and EGFP (green) of AAV-sparse-CSSP-YFP-8E3 virus into ZI_V in the indicated groups. (B) Quantification of EGFP⁺ dendritic spines (green) containing Iba-1⁺ microglia (red) in the ZI_V from CFA1D and saline mice (t₁₀ = 3.21 and P = 0.0093). (C) Quantification of EGFP⁺ dendritic spines (green) Iba-1⁺ microglia (red) in the ZI_V from saline and minocycline treatment of CFA1D mice (t₁₀ = 3.117 and P = 0.0109). Scale bars, 10 μm. Unpaired t-test. *P < 0.05 and **P < 0.01; ns, not significant.