HMGB1 in the mPFC governs comorbid anxiety in neuropathic pain

Abstract

Background: Whether neuroinflammation causes comorbid mood disorders in neuropathic pain remains elusive. Here we investigated the role of high mobility group box 1 protein (HMGB1), a proinflammatory cytokine, in the medial prefrontal cortex (mPFC) in anxiety comorbidity of neuropathic pain.

Methods: Neuropathic pain was induced by partial transection of the infraorbital nerve (p-IONX) or partial sciatic nerve ligation (PSL) in mice and evaluated by measuring nociceptive thresholds to mechanical and heat stimulation. Anxiety-like behaviors were assessed by elevated plus maze, light dark box and open field tests. Aversive or anti-aversive effect was detected by conditioned place preference test. Neuronal activity was evaluated by single-unit and patch clamp recordings. The contribution of mPFC pyramidal neurons to anxiety was further examined by selectively inhibiting them by optogenetics. HMGB1 expression was measured by immunohistochemistry and western blotting. Antagonism of HMGB1 was achieved by injecting anti-HMGB1 monoclonal antibody (mAb) intracerebrally or intraperitoneally.

Results: Anxiety-like behaviors were presented earlier after p-IONX than after PSL. HMGB1 expression was upregulated in the mPFC temporally in parallel to anxiety onset, rather than in other regions associated with anxiety. The upregulation of HMGB1 expression and its translocation from the nucleus to cytoplasm in the mPFC occurred predominantly in neurons and were accompanied with activation of microglia and astrocytes. Infusion of anti-HMGB1 mAb into the mPFC during the early and late phases after either p-IONX or PSL alleviated anxiety-like behaviors and aversion without changing pain sensitization, while local infusion of exogenous ds-HMGB1, the proinflammatory form of HMGB1, into the mPFC induced anxiety and aversion but not pain sensitization in naïve mice. In addition to reversing established pain sensitization and anxiety simultaneously, intraperitoneal injection of anti-HMGB1 mAb reduced HMGB1 upregulation and suppressed the hyperexcitability of layer 2/3 pyramidal neurons in the mPFC after p-IONX. Moreover, optogenetic inhibition of mPFC pyramidal neurons alleviated anxiety in p-IONX mice.

Conclusion: These results demonstrate that HMGB1 in the mPFC drives and maintains anxiety comorbidity in neuropathic pain by increasing the excitability of layer 2/3 pyramidal neurons, and justify antagonism of HMGB1, e.g., neutralization by mAb, as a promising therapeutic strategy for neuropathic pain with anxiety comorbidity.

Keywords: Anxiety; Comorbid mood disorders; HMGB1; Neuroinflammation; Neuropathic pain; mPFC.

Fig. 1

Early onset anxiety was induced by p-IONX. a Schedule of experiment procedures. b Schematic of p-IONX and PSL surgery on the left side. c Paw withdrawal thresholds to mechanical stimulation and noxious heat stimulation in the ipsilateral hind paw and head withdrawal threshold to noxious heat stimulation in the ipsilateral vibrissal pad before and after surgery. *, ** and *** indicate P < 0.05, 0.01 and 0.001, respectively, compared with the baseline (BL), Kruskal–Wallis test with repeated measures and Dunn’s post hoc. ## and ### indicate P < 0.01 and 0.001, respectively, compared with the respective sham group at the same time points, two-way ANOVA with Bonferroni post hoc. d-f Anxiety-like behaviors measured by EPM, LDB and OFT tests on D8/9 (d), D15/16 (e) and D29/30 (f) after surgery. *, ** and *** indicate P < 0.05, 0.01 and 0.001, respectively, compared with the sham group, unpaired t test. g, h Representative heat maps from the sham, PSL and p-IONX groups in EPM (g) and OFT (h) tests. i Depression-like behaviors measured by TST, FST and OFT on D30 after surgery. n = 8/group

Fig. 2

HMGB1 expression in the mPFC was increased temporally in parallel to comorbid anxiety in neuropathic pain. a Schedule of experimental procedures. b Representative photomicrographs of HMGB1 immunostaining in the mPFC, BLA, vHPC and PBN. Scale bar, 200 μm. c Quantification of HMGB1 fluorescence intensity in the mPFC, BLA, vHPC and PBN on D9 PO. d, e Representative images of protein bands in western blotting (left) and quantification of HMGB1 expression (right) in the mPFC (d) and BLA (e) on D9 and D30 PO. *, ** and *** indicate P < 0.05, 0.01 and 0.001, respectively, compared with the indicated groups, ordinary one-way ANOVA with Tukey post hoc or Kruskal–Wallis test with Dunn’s post hoc in (c) and ordinary one-way ANOVA with Dunnett post hoc in (d) and (e). n = 4/group

Fig. 3

Cellular and subcellular distribution of HMGB1 in the mPFC after p-IONX in MRL/MPJ mice. a-c Representative photomicrographs indicating the cellular distribution of HMGB1 in neurons (NeuN⁺, a), microglia (Iba1⁺, b) and astrocytes (S100b⁺, c) in the mPFC on D9 PO. DAPI was used to label nuclei. Scale bar, 50 μm. d Percentage of HMGB1-positive cells in neurons (NeuN⁺), microglia (Iba1⁺) and astrocytes (S100b⁺) on D9 PO in the mPFC. e Percentage of cells with HMGB1 cytoplasmic translocation in neurons (NeuN⁺), astrocytes (S100b⁺) and microglia (Iba1⁺) on D9 PO in the mPFC. f Quantification of Iba1 and S100b fluorescence intensity in the mPFC. g-h Representative images of western blot bands (g) and quantification of HMGB1 (h) in nuclear and cytoplasmic fractions from the mPFC on D9 PO. * and ** indicate P < 0.05 and 0.01, respectively, compared with the indicated groups, unpaired t test. n = 4/group

Fig. 4

Local neutralization of HMGB1 in the mPFC reduced anxiety-like behaviors and aversion but not pain sensitization after p-IONX. a Schedule of procedures of mPFC drug delivery experiments in the early phase after p-IONX. b Schematic (upper) and representative photomicrograph (lower) of the cannula implantation in bilateral mPFC for local infusion of agents. Dashed lines indicate the trace of implanted canula. Scale bar, 100 μm. c Paw withdrawal thresholds to mechanical stimulation and noxious heat stimulation at the left hind paw and head withdrawal threshold to noxious heat stimulation at the left vibrissal pad before and after p-IONX. *, ** and *** indicate P < 0.05, 0.01 and 0.001, respectively, compared with the baseline (BL), Kruskal–Wallis test with repeated measures and Dunn’s post hoc. #, ## and ### indicate P < 0.05, 0.01 and 0.001, respectively, compared with the sham + mAb group at the same time points, two-way ANOVA with Bonferroni post hoc. d, e Anxiety-like behaviors measured by EPM, LDB and OFT on D8/9 (d) and D15/16 (e) after p-IONX, respectively. * indicates P < 0.05, compared with the indicated groups, ordinary one-way ANOVA with Tukey post hoc. f Schedule (left) and schematic (right) of the experiment procedures for the CPP test. g Representative heat maps in the CPP test. h-j Time spend by the mice in the control IgG- and anti-HMGB1 mAb-paired boxes (h), preference (i) and shift preference (j) for the anti-HMGB1 mAb-paired box in pre- (Pre) and post-conditioning (Post) phases. *, ** and *** indicate P < 0.05, 0.01 and 0.001, respectively, compared with the indicated groups, unpaired t test. ## indicates P < 0.01, compared with the sham group in (i), two-way ANOVA with Bonferroni post hoc. n = 8/group

Fig. 5

HMGB1 upregulation in the mPFC maintained comorbid anxiety in neuropathic pain. a Schedule of procedures of mPFC drug delivery experiments in the late phase after p-IONX. b, c Anxiety-like behaviors measured by EPM, LDB and OFT on D29/30 and D36/37 after p-IONX and PSL, respectively. * indicates P < 0.05, compared with the IgG-treated groups, unpaired t test. n = 8/group

Fig. 6

HMGB1 in the mPFC was sufficient to induce anxiety-like behaviors and aversion but not pain sensitization in naïve mice. a Schedule of experimental procedures. b Paw withdrawal thresholds to mechanical stimulation and noxious heat stimulation at the left hind paw and head withdrawal threshold to noxious heat stimulation at the left vibrissal pad. c, d Anxiety-like behaviors measured by EPM, LDB and OFT on D8/9 (c) and D15/16 (d) after administration. e Representative heat maps in the CPP test. f Time spend by mice in the saline- and ds-HMGB1-paired boxes. *, ** and *** indicate P < 0.05, 0.01 and 0.001, respectively, compared with the saline-treated group in (c) and pre-conditioning in (f), unpaired t test. n = 8/group

Fig. 7

Systemic anti-HMGB1 mAb reversed the hyperexcitability of pyramidal neurons in mPFC layer 2/3 after p-IONX. a Schedule of experimental procedures. Anti-HMGB1 mAb was administered intraperitoneally. b Example image of whole cell patch-clamp recordings on mPFC layer 2/3 pyramidal neurons. Scale bar, 150 μm. c Example traces of sEPSCs and sIPSCs in sham + saline, p-IONX + saline and p-IONX + anti-HMGB1 mAb groups. d, e Amplitude, frequency and charge transfer of sEPSCs (d) and sIPSCs (e). f Quantification of E/I ratio (∣CsEPSC∣/CsIPSC). g, h Resting potential (g) and rheobase (the lowest current that evoked action potential firing, h). i, j Representative traces of firing (i) and number of spikes (j) of action potentials evoked by injecting current at two-fold of rheobase. k Firing rate of action potentials evoked by step-current injection. l Autocorrelation and waveform (inset) of a representative glutamatergic neuron in in vivo single-unit recordings. m Mean frequency of spontaneous firing. *, ** and *** indicate P < 0.05, 0.01 and 0.001, respectively, compared with the indicated groups, unpaired t test or Mann–Whitney test. n = 9–10 neurons from 4 mice/group in (d-k) and n = 20 neurons from 8 mice/group in (m)

Fig. 8

Systemic anti-HMGB1 mAb simultaneously alleviated pain sensitization and anxiety after p-IONX. a Schedule of experimental procedures. b Paw withdrawal thresholds to mechanical stimulation and noxious heat stimulation at the left hind paw and head withdrawal threshold to noxious heat stimulation at the left vibrissal pad before and after p-IONX. ## and ### indicate P < 0.01 and 0.001, respectively, compared with the saline-treated group at the same time points, two-way ANOVA with Bonferroni post hoc. c Anxiety-like behaviors measured by EPM, LDB and OFT on D15/16 after p-IONX. * and *** indicate P < 0.05 and 0.001, respectively, compared with the saline-treated group, ordinary one-way ANOVA with Dunnett post hoc. n = 8/group