Chronic pain-induced neuronal plasticity in the bed nucleus of the stria terminalis causes maladaptive anxiety

Abstract

The comorbidity of chronic pain and mental dysfunctions such as depression and anxiety disorders has long been recognized, but the underlying mechanisms remain poorly understood. Here, using a mouse model of neuropathic pain, we demonstrated neuronal plasticity in the bed nucleus of the stria terminalis (BNST), which plays a critical role in chronic pain-induced maladaptive anxiety. Electrophysiology demonstrated that chronic pain increased inhibitory inputs to lateral hypothalamus (LH)-projecting BNST neurons. Chemogenetic manipulation revealed that sustained suppression of LH-projecting BNST neurons played a crucial role in chronic pain-induced anxiety. Furthermore, using a molecular genetic approach, we demonstrated that chronic pain elevated the excitability of a specific subpopulation of BNST neurons, which express cocaine- and amphetamine-regulated transcript (CART). The elevated excitability of CART-positive neurons caused the increased inhibitory inputs to LH-projecting BNST neurons, thereby inducing anxiety-like behavior. These findings shed light on how chronic pain induces psychiatric disorders, characterized by maladaptive anxiety.

Fig. 1.. Chronic pain induces anxiety-like behavior.

(A) Spared nerve injury model. Nerve injuries were induced by ligating and cutting of the tibial and common peroneal nerves, leaving the sural nerve intact. (B) Schedule of experiments. (C) Mechanical threshold of sham-operated control and nerve injury mice (sham: n = 23; nerve injury: n = 22). Data are expressed as means ± SEM. Statistical significance was evaluated using two-way repeated-measures analysis of variance (ANOVA) followed by Bonferroni’s post hoc test. ****P < 0.0001. (D to F) Behavioral analyses using OFT (D; sham: n = 20; nerve injury: n = 19), EPM (E; sham: n = 19; nerve injury: n = 18), and LDB (F; sham: n = 12; nerve injury: n = 12). (G) Emotionality z scores were calculated using the data from animals that were subjected to all three tests (sham: n = 9; nerve injury: n = 9). Box-whisker plots show the median, interquartile range, and 10th to 90th percentiles. Statistical significance was evaluated using two-tailed unpaired Student’s t test. *P < 0.05; **P < 0.01; ***P < 0.001. Details of statistical data are provided in table S1.

Fig. 2.. Chronic pain increases inhibitory synaptic inputs to LH-projecting adBNST neurons.

(A) Scheme of experiments. Inserted confocal images show retrograde-labeled LH-projecting neurons (red) in the BNST (left) and a biocytin-labeled (cyan) recorded adBNST neuron (right) filled with retrobeads (red). Scale bars, 200 and 10 μm. (B) Representative traces of membrane potential in response to a hyperpolarizing current injection (−40 pA, 400-ms duration) recorded from I_h-negative (top) and I_h-positive (bottom) neurons. (C) Proportion of I_h-negative and I_h-positive neurons in LH-projecting adBNST neurons of sham-operated (top) and nerve injury (bottom) mice (P = 0.58, sham versus nerve injury, chi-square test). (D) Representative traces of sIPSCs from I_h-negative LH-projecting adBNST neurons of sham-operated (top) and nerve injury (bottom) mice. (E and F) Cumulative probability plots of the inter-event interval and means ± SEM of the sIPSC frequency (E) and amplitude (F). (G) Representative traces of spontaneous excitatory postsynaptic currents (sEPSCs) from I_h-negative LH-projecting adBNST neurons of sham-operated (top) and nerve injury (bottom) mice. (H and I) Cumulative probability plots of the inter-event interval and means ± SEM of the sEPSC frequency (H) and amplitude (I). Kolmogorov-Smirnov test was used to analyze cumulative probability plots. Two-tailed unpaired Student’s t test or two-tailed Mann-Whitney test was used to analyze the frequency/amplitude of sIPSC/sEPSC. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; n.s., not significant. Details of statistical data are provided in table S1.

Fig. 3.. Inhibition of LH-projecting adBNST neurons induces anxiety-like behavior.

(A) Scheme of experiments. (B) Representative images of hM4Di-mCherry expression in the adBNST (left) and AAV2 injection site visualized by coinjection of fluorescent beads in the LH (right). Scale bars, 300 μm (left) and 100 μm (right). AC, anterior commissure; EP, entopeduncular nucleus; lv, lateral ventricle, MEA, medial amygdala; opt, optic tract; Str, striatum. (C) Voltage-clamp recordings of resting membrane potential and membrane resistance in hM4Di-expressing adBNST neurons. Data are expressed as means ± SEM (n = 6 cells from 4 mice). Statistical significance was evaluated using two-tailed paired Student’s t test. *P < 0.05. (D to G) Behavioral analyses using OFT (D), EPM (E), and LDB (F), and emotionality z score across these tests (G). (H) Mechanical thresholds in the von Frey test. Box-whisker plots show the median, interquartile range, and 10th to 90th percentiles. Statistical significance was evaluated using two-tailed unpaired Student’s t test or two-tailed Mann-Whitney test (mCherry: n = 11; hM4Di: n = 10). *P < 0.05; **P < 0.01; ****P < 0.0001. Details of statistical data are provided in table S1.

Fig. 4.. Activation of LH-projecting adBNST neurons ameliorates chronic pain–induced anxiety-like behavior.

(A) Scheme of experiments. (B) Representative images of hM3Dq-mCherry expression in the adBNST (left) and AAV2 injection site visualized by coinjection of fluorescent beads in the LH (right). Scale bars, 300 μm. (C) Schedule of experiments. (D to G) Behavioral analyses using OFT (D), EPM (E), and LDB (F) tests, and emotionality z score across these tests (G). Box-whisker plots show the median, interquartile range, and 10th to 90th percentiles. Statistical significance was evaluated using two-way ANOVA followed by Bonferroni’s post hoc test (n = 9 for each group). *P < 0.05 and **P < 0.01. (H) Mechanical thresholds were assessed using the von Frey test in mCherry-expressing nerve injury mice (n = 9) and hM3Dq-mCherry–expressing nerve injury mice (n = 9) before and after CNO administration (1 mg/kg, intraperitoneally). Statistical significance was evaluated using two-way repeated-measures ANOVA followed by Bonferroni’s post hoc test. Details of statistical data are provided in table S1.

Fig. 5.. LH-projecting adBNST neurons receive inhibitory inputs from CART-expressing BNST neurons.

(A) Scheme of experiments. Cart-Cre mice were injected with AAV5 delivering a Cre-dependent mCherry expression construct to visualize CART-positive BNST neurons. (B) Confocal images of coronal brain sections showing mCherry-expressing neurons (red) in the anterior BNST. Approximate distances (anterior-posterior) from the bregma are shown. ac, anterior commissure; ad, adBNST; ov, ovBNST. Scale bar, 200 μm. DAPI, 4′,6-diamidino-2-phenylindole. (C) High-magnification images of the ovBNST (left) and adBNST (right). Arrows indicate mCherry expression in soma. Scale bar, 20 μm. (D) Scheme of experiments. AAV5 delivering Cre-dependent ChR2 expression construct was injected into the ovBNST of Cart-Cre mouse and then the red retrobeads were injected into the LH. Whole-cell patch-clamp recordings from retrobead-labeled adBNST neurons were performed to examine the electrophysiological responses to blue light stimulation. (E) Confocal image of a retrobead-labeled BNST neuron filled with biocytin during recording. Scale bar, 20 μm. (F) Representative trace showing light-evoked IPSCs. (G) Synaptic connectivity from ovBNST^CART neurons to I_h-negative LH-projecting adBNST neurons. Fifteen out of 18 cells from two mice showed light-evoked IPSCs. (H and I) Summary of amplitude (H) and latency to peak (I) in recorded neurons (n = 15 cells from two mice). Box-whisker plots indicate median, interquartile range, and 10th to 90th percentiles. Means are indicated by “+” in (H) and (I).

Fig. 6.. Chronic pain enhances the excitability of ovBNST^CART neurons.

(A) Representative action potentials in ovBNST^CART neurons of sham (black) and nerve-injured (blue) mice. (B to E) Intrinsic electrophysiological properties of ovBNST^CART neurons of sham (black, n = 20 cells from four mice) and nerve injury (blue, n = 22 cells from four mice) mice: resting membrane potential (B), relationship between action potential threshold and rheobase (C), membrane resistance (D), and action potential amplitude (E). Data are expressed as means ± SEM. Statistical significance was evaluated using two-tailed unpaired t test. *P < 0.05 and **P < 0.01. Details of statistical data are provided in table S1.

Fig. 7.. Involvement of enhanced excitability of ovBNST^CART neurons in increased inhibitory inputs to LH-projecting adBNST neurons and chronic pain–induced anxiety-like behavior.

(A) Scheme of experiments. (B to D) Representative traces of sIPSCs in LH-projecting adBNST neurons before and after CNO (10 μM) bath application in sham mice that expressed hM4Di (B; n = 10 from four mice), nerve injury mice that expressed EYFP alone (C; n = 10 from four mice), and nerve injury mice that expressed hM4Di (D; n = 13 from five mice). (E and F) Effect of CNO bath application on sIPSC frequency. Data are expressed as means ± SEM. Statistical significance was evaluated using two-way repeated-measures ANOVA followed by Bonferroni’s post hoc test (E) and one-way ANOVA followed by Sidak’s post hoc test (F). *P < 0.05; ***P < 0.01; ****P < 0.001. (G) Behavioral analyses using an EPM test (mCherry: n = 8; hM4Di: n = 7). Box-whisker plots indicate median, interquartile range, and 10th to 90th percentiles. Statistical significance was evaluated using two-tailed unpaired Student’s t test. *P < 0.05. Details of statistical data are provided in table S1.

Fig. 8.. Neuronal circuit involved in chronic pain–induced maladaptive anxiety.

Increased excitability (white arrow) of ovBNST^CART neurons causes a sustained suppression (black arrow) of LH-projecting adBNST neurons during chronic pain, thereby enhancing anxiety-like behavior.