Insula→Amygdala and Insula→Thalamus Pathways Are Involved in Comorbid Chronic Pain and Depression-Like Behavior in Mice

Abstract

The comorbidity of chronic pain and depression poses tremendous challenges for the treatment of either one because they exacerbate each other with unknown mechanisms. As the posterior insular cortex (PIC) integrates multiple somatosensory and emotional information and is implicated in either chronic pain or depression, we hypothesize that the PIC and its projections may contribute to the pathophysiology of comorbid chronic pain and depression. We show that PIC neurons were readily activated by mechanical, thermal, aversive, and stressful and appetitive stimulation in naive and neuropathic pain male mice subjected to spared nerve injury (SNI). Optogenetic activation of PIC neurons induced hyperalgesia and conditioned place aversion in naive mice, whereas inhibition of these neurons led to analgesia, conditioned place preference (CPP), and antidepressant effect in both naive and SNI mice. Combining neuronal tracing, optogenetics, and electrophysiological techniques, we found that the monosynaptic glutamatergic projections from the PIC to the basolateral amygdala (BLA) and the ventromedial nucleus (VM) of the thalamus mimicked PIC neurons in pain modulation in naive mice; in SNI mice, both projections were enhanced accompanied by hyperactivity of PIC, BLA, and VM neurons and inhibition of these projections led to analgesia, CPP, and antidepressant-like effect. The present study suggests that potentiation of the PIC→BLA and PIC→VM projections may be important pathophysiological bases for hyperalgesia and depression-like behavior in neuropathic pain and reversing the potentiation may be a promising therapeutic strategy for comorbid chronic pain and depression.

Significance Statement

Treatment of chronic pain is quite challenging because of the involvement of central sensitization and emotional disorders. It is keenly demanding to search for brain circuits commonly or separately contributing to hyperalgesia and emotional dysfunction in chronic pain. The posterior insular cortex (PIC) is involved in both pain and emotional processing. But the relationship between the malfunction in the PIC and the comorbidity of chronic pain and depression has not been elucidated. Here, we demonstrate that in chronic pain, PIC neurons became hyperactive, and their projections to the basolateral amygdala and the ventromedial nucleus of the thalamus were augmented, resulting in hyperalgesia and depression-like behaviors. Therefore, these pathways may be potential therapeutic targets for comorbid chronic pain and depression.

Introduction

Chronic pain, affecting ∼15% of the global population, is a persistent distress that is often accompanied by psychiatric disorders such as depression (Bair et al., 2003; Rayner et al., 2016; GBD 2017 Disease Injury Incidence Prevalence Collaborators, 2018). The comorbidity of these two conditions contributes to heightened disability and an unfavorable prognosis compared with that of either condition alone (Bair et al., 2003; Roughan et al., 2021). It is recognized that neuroplasticity mechanisms underlying chronic pain are also associated with an increased risk of depression (Vogt, 2005; Meerwijk et al., 2013; Zhou et al., 2019; Llorca-Torralba et al., 2022; Yin et al., 2022; Becker et al., 2023; Shen et al., 2023; Ji et al., 2023a). Therefore, investigating the extent to which these convergent neuroplasticity changes contribute to the development of pain hypersensitivity and pain-related negative emotions holds the potential for developing targeted interventions for individuals with comorbid chronic pain and psychiatric disorders.

The posterior insular cortex (PIC) is a major cortical area to integrate multimodal somatosensory information, including emotional and pain processing. It receives nociceptive information from the spinal cord dorsal horn directly and indirectly. The insula controls spinal networks and encodes pain intensity (Peyron et al., 2000; Segerdahl et al., 2015; Tan et al., 2017). The strong interconnection between the insular cortex and the limbic system indicates an essential role of the insula in emotional processing. Consistent with this role, a large body of evidence supports the association of alterations in PIC neuronal activity with either chronic pain or major depressive disorder (MDD) in patients and in rodent models (Tan et al., 2017; Bergeron et al., 2021; He et al., 2022; Schnellbacher et al., 2022). For instance, lesion or inhibition of the PIC elevated pain thresholds and alleviates MDD (Garcia-Larrea et al., 2010; Bergeron et al., 2021), whereas activation of this structure is sufficient to trigger emotional dysfunction in naive animals (Gehrlach et al., 2019). These results suggest that the PIC is involved in both chronic pain and depression. However, further investigations are highly demanded to address the neural circuit bases for the role of the PIC in comorbid chronic pain and depression.

The reciprocal connections between the PIC and subcortical regions form complex neural circuits that are essential for various brain functions (Segerdahl et al., 2015; Tan et al., 2017; Schiff et al., 2018; Gehrlach et al., 2019, 2020; Labrakakis, 2023). For instance, the PIC is involved in anxiety-like behavior and learning of anticipatory avoidance via mobilizing the central nucleus of the amygdala, and it regulates feeding through the nucleus accumbens (Schiff et al., 2018; Gehrlach et al., 2019). Stimulation of the PIC projection to the bed nucleus of the stria terminalis enhances a reinforcing behavior in a dopamine-dependent manner, but inhibition of this projection induces aversion and anxiety without affecting food consumption (Girven et al., 2021). Knowledge about the neural circuit basis of the physiological function of the PIC has been expanding. However, the role of projections from the PIC in comorbid chronic pain and depression has been understudied.

Here, we report that PIC neurons responded to pain, aversive, and stressful stimulation and modulated pain-, reinforcement-, and depression-like behaviors in both naive and neuropathic pain mice. We also found that excitatory projections from PIC neurons to the basolateral amygdala (BLA) and the ventromedial nucleus (VM) of the thalamus modulated pain and depression-like behaviors; these projections were potentiated in neuropathic pain, and inhibition of these projections conferred analgesia, conditioned place preference (CPP), and antidepressant-like effect. The present study implicates the PIC→BLA and PIC→VM projections to pain modulation and emotional processing and suggests that inhibition of these projections may be an intervention strategy for the treatment of comorbid chronic pain and depression.

Materials and Methods

Animals

The care and use of animals and the experimental protocols in this study were approved by the Institutional Animal Care and Use Committee and the Office of Laboratory Animal Resources of Xuzhou Medical University. These approvals were granted in accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals (1988) in China.

Male C57BL/6J wild-type (WT) and CaMKII-Cre mice [B6.Cg-Tg^{(Camk2a-cre)T29-1Stl/J}], aged 8–10 weeks, were bred in the animal facility of Xuzhou Medical University. The mice were group-housed (<=5 per cage) in a controlled environment with a stable temperature (21–23°C) and humidity (45–70%). They were maintained on a 12 h light/dark cycle and provided with ad libitum access to water and food. All necessary measures were taken to minimize animal suffering and reduce the total number of animals used.

Viral vectors

AAV-CaMKII-GCaMP6s, AAV-CaMKII-eYFP, AAV-CaMKII-NpHR-eYFP, AAV-CaMKII-ChR2-eYFP, AAV-hSyn-DIO-GFP-synaptophysin-mRuby, AVV retro-hSyn-eGFP, and AAV retro-hSyn-mCherry were purchased from BrainVTA and Brain Case. The titers of the adeno-associated virus (AAV) vectors were 2–9 × 10¹² viral genome per milliliter.

Stereotaxic surgeries and injection

Mice were deeply anesthetized with isoflurane (3% for induction and 1.5% for maintenance), placed on a heating pad, and stabilized on a stereotaxic apparatus (RWD Life Science). Small holes were drilled in the skull above the brain regions of interest. The viral vectors were injected (120 nl of virus per site at 50 nl/min) with a 10 μl Hamilton syringe driven by a microinjection pump (KD Scientific).

The coordinates (relative to the bregma) for viral injection were the following (in mm): −0.5 AP, 4.05 ML, and 4.0 DV for the PIC; −1.40 AP, 3.10 ML, and 4.75 DV for the BLA; and −1.60 AP, 0.7 ML, and 4.2 DV for the VM. Optical fiber implants (200 μm in diameter, NA 0.37; Inper) were placed 200 µm above (for optogenetic manipulation) or at (for fiber photometry recording) the injection site and were fixed to the skull with a dental cement. Mice with virus injections and optical fiber implants were allowed to recover for at least 3 weeks before electrophysiological recording, behavioral test, and morphological assay. Viral expression and the position of optical fiber implants in each mouse were confirmed histologically after the termination of the experiments. We only included mice with viral expression confined to the PIC and optical fibers in right places for optogenetic modulation and fiber photometry recording.

For postoperative pain relief, meloxicam (4 mg/kg; Aladdin Biochemical Technology) was administered subcutaneously once a day for 3 d.

Fiber photometry

A fiber photometry instrument (Thinkertech; Wang et al., 2019; Wu et al., 2020) was used to monitor GCaMP6 signal in PIC neurons. The excitation light was set to 50 μW. The responses in PIC neurons were measured when mice encountered stimulation with a von Frey filament, a heating block, an air puff, a body restraint, etc. To summarize the responses, we calculated the mean and standard deviation (SD) of 3 s GCaMP6 signal prior to the response and used these numbers to calculate the Z-score [(F − mean) / SD] for each point in the GCaMP6 trace. We then measured the peak of the Z-score plot to quantify the response of PIC neurons.

Optogenetic manipulation

For ChR2-mediated optogenetic stimulation, 473 nm light pulses (5 ms, 20 Hz, 4 mW) from a laser generator (Inper) were delivered. For NpHR-mediated optogenetic inhibition, a 3 mW 589 nm light from a laser generator (Inper) was kept on continuously for 1–2 min. All optogenetic manipulations were performed unilaterally in the right hemisphere. Therefore, the contralateral side refers to the left side of the body and the ipsilateral side refers to the right side.

Evaluation of nociceptive responses

The mice were acclimatized in the test compartment with a mesh floor for at least 1 h before the experiments. The mechanical paw withdrawal threshold (PWT) on both hindpaws was measured using a series of the von Frey filaments (Anesthesio) with varying fiber forces (ranging from 0.02 to 2.0 g). The 50% threshold was determined using the up–down method (Zhou and Luo, 2015; Jia et al., 2022).

The thermal paw withdrawal latency (PWL) on both hindpaws was recorded to evaluate the thermal nociceptive response. The mice were habituated in a test compartment placed on a glass surface. A focused beam was directed to the hindpaws of the mice from beneath the glass. The intensity of the beam was adjusted using a plantar anesthesia tester (Boerni). If no response to the beam was observed within 20 s, the beam was turned off to avoid potential tissue damage. The measurement was repeated three times and an average PWL was calculated for each mouse.

Spared nerve injury

The chronic neuropathic pain model was established with SNI of the sciatic nerve (SN) according to a previously reported protocol (Decosterd and Woolf, 2000; Yin et al., 2022). Mice were deeply anesthetized using isoflurane (3% for induction, 1.5% for maintenance; RWD Life Science). The fur in the surgical area, extending from the knee to the hip, was shaved, and the skin was sterilized with 75% alcohol. A longitudinal incision was made in the shaved region, allowing for blunt dissection of the biceps femoris muscle to expose the SN and its branches (sural, common peroneal, and tibial nerves). Two nylon sutures 3 mm apart were tightly ligated around the common peroneal and tibial nerves, and the nerves between the sutures were subsequently cut and removed. The mice were then allowed to recover on a heating pad.

Mice that did not receive nerve ligation and severing were used as sham controls. Pain thresholds were assessed using von Frey filaments and a heating beam targeting the skin area innervated by the sural nerve.

CPP test

The CPP was performed in a custom-made two-chamber box (length × width × height, 40 × 20 × 30 cm³): the right chamber had vertical black-and-white stripes on the walls and a smooth floor, whereas the left chamber had horizontal black-and-white stripes on the walls and a mesh floor. The CPP test was performed according to a previous study with some modifications (Jia et al., 2022).

On Day 1 (preconditioning day; precondition), mice were given free access to the two chambers, and the time that the mice spent in each chamber was recorded. On Days 2–4, the mice were placed in one chamber for 20 min and received optogenetic modulation. At least 4 h later, the mice were placed in the other chamber for 20 min without optogenetic modulation. On Day 5 (test day), the mice were allowed to freely explore the two chambers for 20 min, and the time spent in each chamber was recorded. On the precondition and test days, the animal's movement was video-tracked and analyzed online or offline with an EthoVision XT video tracking software (Noldus Information Technology; Fan et al., 2023; Ji et al., 2023b). We calculated the time spent in the light-paired side on the precondition and test days. The mice were excluded from the experiments if they spent >75% of the total time in one chamber on the precondition day.

Tail suspension test

A mouse was suspended by taping its tail onto a horizontal bar 50 cm above the floor. The mouse was allowed to hang undisturbed for 6 min, and its behavior was video-recorded. The total duration that the mouse remained immobile in the last 5 min was used to evaluate a depression-like behavior.

Forced swim test

Mice were placed in a 1 L glass beaker filled with 26°C water. The movement of the mice was recorded for 5 min using a camera placed beside the beaker, and the immobility time was measured.

Brain slice electrophysiology

Brain slice electrophysiological recording was conducted with minor modifications according to previously reported methods (Jia et al., 2022, 2023a,b). Coronal slices (300 µm thick) containing the ACC, BLA, or VM were prepared using a vibratome (Leica VT-1200S) in an ice-cold modified sucrose-based artificial cerebral spinal fluid (sACSF) saturated with 95% O₂/5% CO₂ (carbogen), containing the following (in mM): 85 NaCl, 75 sucrose, 2.5 KCl, 1.25 NaH₂PO₄, 4.0 MgCl₂, 0.5 CaCl₂, 24 NaHCO₃, and 25 glucose. The brain slices were transferred into a carbogenated sACSF at 32°C and allowed to recover for 60 min and then placed in a normal carbogenated ACSF containing the following (in mM): 125 NaCl, 2.5 KCl, 1.2 NaH₂PO₄, 1.2 MgCl₂, 2.4 CaCl₂, 26 NaHCO₃, and 11 glucose at 26°C for at least 30 min prior to use.

Neurons in brain slices were visualized under an upright microscope (FN-1, Nikon), equipped with a CCD camera (Flash 4.0 LTE, Hamamatsu). Whole-cell patch-clamp recordings were obtained using an electrophysiological setup composed of a dual-channel MultiClamp 700B amplifier, a Digidata 1550B analog-to-digital converter, and a pClamp 10.7 software (Molecular Devices). The patch electrodes had a resistance of 4–6 MΩ when filled with a low-chloride intrapipette solution containing the following (in mM): 135 K-gluconate, 0.2 EGTA, 0.5 CaCl₂, 10 HEPES, 2 Mg-ATP, and 0.1 GTP, pH 7.2 and osmolarity 290–300 mOsm. All recordings were performed at 32 ± 1°C maintained by a dual-channel temperature controller (TC-344C, Warner Instruments).

Light-evoked excitatory postsynaptic currents (eEPSCs) in the presence of tetrodotoxin (TTX, 1 μM) and 4-aminopyridine (4-AP, 0.3 mM) were recorded at −50 mV. To confirm whether glutamatergic connections were involved, we bath applied 6-cyano-7-nitroquinoxaline-2,3-dione disodium salt hydrate (CNQX, 20 µM). Firing in response to current injections (1 s, 20–200 pA steps with a 20 pA increment with a 30 s intersweep interval) was recorded in the current-clamp mode.

For light-evoked responses, blue light (460 nm, 2 mW) or yellow light (560 nm, 2 mW) was delivered through an optical fiber (200 μm, NA 0.37) connected to a PlexBright LED light source (Plexon).

Histology

Mice were killed in a CO₂ chamber and then subjected to cardiac perfusion with phosphate-buffered saline (PBS), followed by 4% paraformaldehyde (PFA) in PBS. Mouse brains were removed and postfixed in 4% PFA overnight at 4°C. Brain samples were placed in 30% sucrose in PBS until they sank for cryoprotection, then imbedded in OCT, frozen at −17°C, and cut into 30 μm sections with a Leica CM1950 cryostat. The brain sections were mounted onto glass slides. For immunostaining, the brain sections were incubated in a blocking buffer containing 5% donkey serum and 0.1% Triton X-100 for 90 min at room temperature. Then the sections were incubated with a primary antibody diluted with a blocking buffer for 24 h at 4°C [rabbit anti-c-Fos IgG, 1:2,000, Cell Signaling Technology; rabbit anti-vesicular glutamate transporter 1 (VgluT1) IgG, 1:200, Proteintech]. After washing three times (10 min each) in PBS, the sections were incubated with secondary antibodies (Alexa 488- or Alexa 555-conjugated donkey anti-rabbit IgG, Alexa 488-conjugated donkey anti-mouse IgG, Jackson ImmunoResearch) for 90 min at room temperature. The sections were washed three times (10 min each) in PBS, dried in the dark, and then coverslipped in a mounting medium (Meilunbio).

The sections were imaged with a confocal microscope (LSM 880, Zeiss), and the images were processed with ImageJ (NIH; Schneider et al., 2012). The output to each brain region from the PIC was quantified by dividing the number of fluorescence pixels in that region by the total number of axon-occupied pixels in the entire brain (Fig. 5).

Figure 5.

Axonal projections of PIC projection neurons. A, Schematic tracing strategy and example images showing that AAV-EF1α-DIO-eGFP-synaptophysin-mRuby was injected into the PIC. B, Percentages of labeled axonal terminals in multiple cortical and subcortical brain regions relative to the sum of labeled axonal terminals in these regions (n = 5 mice). The axonal terminals were quantified by pixels of mRuby. C, Example images showing eGFP-labeled axonal fibers in green and mRuby-labeled terminals in red. Scale bar, 100 µm. AIC, anterior insular cortex; BLA, basolateral amygdala; BLP, basolateral posterior amygdala; BST, bed nucleus of the stria terminalis; CeL, central amygdala lateral part; CPu, striatum; Ect, entorhinal cortex; LEnt, lateral entorhinal cortex; M1, primary motor cortex; MIC, medial insular cortex; PIC, posterior insular cortex; Pn, pontine nuclei; PRh, perirhinal cortex; RPO, rostral periolivary region; S1, primary somatosensory cortex; S2, secondary somatosensory cortex; VM, ventromedial nucleus of the thalamus; VO, ventral orbital cortex; VPM, ventral posterior medial nucleus of the thalamus.

Chemicals

4-AP, CNQX, and TTX were purchased from MedChem Express.

Experimental design and statistical analysis

GraphPad Prism 7.0 was used for statistical analyses. Clampfit 10.7 (Molecular Devices) was used for the analysis of electrophysiological and fiber photometry data. Figures were prepared with Adobe Illustrator CS6.

If the data passed both normality and equal-variance tests, a parametric analysis was applied. Or else, a nonparametric analysis was used. Two-tailed paired t tests were used to analyze changes in parameters measured from mice or neurons before and after modulation. Two-tailed unpaired t tests were used for comparison of a parameter between two groups of mice or neurons. A one-way or two-way ANOVA followed by Tukey's post hoc tests was used for comparison of parameters among multiple groups of mice. For data collected multiple times from the same group of mice or neurons, a one-way or two-way repeated measures ANOVA was used to analyze the data. All summarized data are expressed as mean ± SEM. The mean, SEM, n (the number of animals), statistical test, and t, F, and p values are presented in the figure legends. A value of p < 0.05 was considered statistically significant. The minimal number of mice used in each experiment was calculated in a priori power analysis (StatMate 2.0), and the power of each experiment was set to 0.8. The sample sizes in each experiment are larger than the minimal numbers.

Results

PIC excitatory neurons are activated by pain-like and emotional stimulation

Previous clinical and preclinical studies demonstrate that PIC neurons are activated by nociceptive stimulation (Coghill et al., 1994; Garcia-Larrea and Peyron, 2013; Frot et al., 2014; Gehrlach et al., 2019). However, studies to systemically explore the responses of PIC neurons to mechanical, thermal, aversive, and stressful stimulation in mice are lacking. To monitor the responses of PIC neurons to these stimuli, we performed fiber photometry recording from PIC neurons in mice. We transfected a genetically encoded calcium sensor (GCaMP6s) or eYFP into PIC excitatory neurons by injecting into the PIC with a recombinant AAV vector under the control of the CaMKII promoter (AAV-CaMKII-GCaMP6s or AAV-CaMKII-eYFP) and implanted an optical fiber above the injection site (Fig. 1A,D). Immunohistochemistry data showed that a great majority (97%; 309 out of 317 GCaMP6 neurons in four slices from four mice) of GCaMP6-positive neurons in the PIC were labeled with an antibody of VgluT1, a marker protein of glutamatergic neurons (Fig. 1B,C). This result confirms that the CaMKII promoter drives GCaMP6 expression in glutamatergic neurons in the PIC with a high specificity. Three weeks after viral injection, we recorded GCaMP6 signal from PIC neurons in freely moving mice upon mechanical and thermal stimulation (Fig. 1E). The fluorescent signal in the PIC was not altered when GCaMP6 and eYFP mice were moving in the testing chamber (Fig. 1F–H), suggesting that motion artifacts may not confound changes in GCaMP6 signal upon other stimulation. In contrast, GCaMP6 signal of PIC neurons elevated robustly when the mice showed pain-like responses to a von Frey filament stimulation (a force of 2 g) on either hindpaw; 0.16 g von Frey stimulation that did not evoke pain-like response evoked an increase in GCaMP6 signal of PIC neurons much smaller than the 2 g von Frey filament stimulation-evoked ones (Fig. 1I–K). Similarly, a significant increase in GCaMP6 signal was also detected upon a thermal stimulation with a 50°C heating block on either hindpaw (Fig. 1L–N). Note that such fluorescent signal changes were not observed in eYFP control mice (Fig. 1J,K,M,N). The data indicate that PIC excitatory projection neurons participate in pain-like behaviors. In mice subjected to SNI of the SN, we observed elevation in GCaMP6 signal when they were subjected to mechanical stimulation by a von Frey filament with a fiber force corresponding to mechanical threshold (0.16 g) and a 48°C heating stimulus on the hindpaws (Fig. 1O–Q). These data indicate that PIC neurons are activated by both mechanical and thermal stimulation in physiological and neuropathic pain conditions.

Figure 1.

PIC neurons are activated by pain and emotional stimuli. A, Schematic diagram showing virus injection into the PIC for fiber photometry recording. B, C, Example images of GCaMP6 expression and VGluT1 immunostaining in the PIC. D, E, A representative image of GCaMP6 expression (D) and schematic diagram (E) of fiber photometry recording from awake-behaving mice in which PIC neurons were virally transfected with GCaMP6 or eYFP. F–H, Heat maps (F), averaged normalized traces (G), and quantification (H, peak responses in G) of changes in PIC fluorescent signal in GCaMP6 and eYFP mice aligned to paw lifting during voluntary movement. t = 0.15, p = 0.88, n = 5 mice in each group. I–N, Heat maps (I,L), averaged normalized traces (J,M), and quantification of changes (K,N) in fluorescent signal in the PIC in GCaMP6 and eYFP mice aligned to von Frey (0.16 g or 2 g) and heating (50°C) stimulation on hindpaws. K, Peak responses in J: F_(5,131) = 29.55, p < 0.0001. N, Peak responses in M: F_(3,51) = 21.44, p < 0.0001). n = 5 mice in each group. O–Q, Heat maps (O), averaged normalized traces (P), and quantification (Q) of changes in fluorescent signal in the PIC in GCaMP6 (n = 5) and eYFP mice (n = 5) aligned to von Frey (0.16 g) and heating (48°C) stimulation on hindpaws 2 weeks after SNI surgery. F_(5,65) = 11.82, p < 0.0001. Two-tailed unpaired t test for H. One-way ANOVA with Tukey's post hoc analysis for (K,N,Q). **p < 0.01. ^##p < 0.01. ns, not significant. Dashed line indicates stimulus onset. Contra, contralateral; Ipsi, ipsilateral; Cl, claustrum; Opto, optical fiber. Scale bar, 100 µm.

Next, we explored the alterations of the activity of PIC neurons in response to stimulation associated with negative emotional states. When we applied 1 s air puff (aversive stimulation) onto the face of individual mice (Fig. 2A), we recorded a significant increase in GCaMP6 signal (Fig. 2B–D). Stressful stimuli, including body restraint (Fig. 2E) and being handled by an unfamiliar experimenter (Fig. 2I), and appetitive stimulation [sweet water; 20% sucrose (w/v); Fig. 2M] also reliably induced a dramatic increase in GCaMP6 signal in PIC neurons (Fig. 2F–H,J–L,N–P). These data indicate that PIC excitatory neurons respond to aversion, stress, and reward stimulation.

Figure 2.

PIC neurons are activated by aversion, stress, and reward. A–P, Schematic diagram (A,E,I,M), heat maps (B,E,J,N), averaged traces (C,G,K,O), and quantification (D,H,L,P) of fluorescent signal in the PIC of GCaMP6 and eYFP mice in response to 1 s air puff, body restraint, being handled by unfamiliar experimenter, and sweet water (20% sucrose) licking. D, t = 3.08, p = 0.0067. H, t = 3.80, p = 0.0014. L, t = 5.51, p < 0.0001. P, t = 2.25, p = 0.028. n = 5 mice in each group.

Similar to previous reports, these results implicate PIC excitatory neurons in pain-like behavior and emotional processing.

Activation of PIC excitatory neurons induces hyperalgesia and negative emotion

Optogenetic stimulation of glutamatergic synaptic inputs from the cingulate nuclei to the PIC induces mechanical and thermal hypersensitivity (Tan et al., 2017; Bouchatta et al., 2022), whereas lesion or pharmacological and optogenetic inhibition of the PIC alleviates hyperalgesia in both neuropathic pain and capsaicin-induced pain models (Benison et al., 2011; Tan et al., 2017; Nagasaka et al., 2022). These studies suggest that PIC neurons may be hyperactive in the chronic pain state. In the present study, we provided direct evident for this.

First, we established SNI mice. These mice developed mechanical and thermal hypersensitivities on Day 7 post-SNI and remained stable until Day 28 post-SNI (Fig. 3A–C). Four weeks after the SNI surgery, SNI mice developed depression-like behaviors showing prolonged immobility in the tail suspension test (TST) and forced swim test (FST), relative to sham mice (Fig. 3D,E). We then counted PIC neurons expressing c-Fos (protein of an immediate early gene, a widely used marker for activation of neurons) in SNI and sham mice. We observed that SNI mice exhibited a significant increase in the total number of c-Fos(+) PIC neurons in both hemispheres 4 weeks after the SNI surgery was performed on the left side of the mice (Fig. 3F–G). Second, we labeled PIC excitatory neurons with a viral vector (AAV-CaMKII-eYFP; eYFP mice) and performed a sham or SNI surgery in these mice (Fig. 3H). Four weeks later, we performed the whole-cell patch-clamp technique to record firing from eYFP-labeled PIC neurons in acute slice preparations from these mice. As showed in Figure 3H,I, PIC excitatory neurons in SNI mice displayed higher firing rates in response to depolarizing current injections than those in sham mice. These data suggest that PIC neurons in SNI mice are hyperactive relative to those in sham mice.

Figure 3.

Hyperactivity of PIC neurons is sufficient to induce hyperalgesia and conditioned place aversion in mice. A, Mice were subjected to SNI or sham surgery on the left SN. B, C, Time courses of mechanical PWT and thermal PWL after sham and SNI surgery. PWT: F_(4,82) = 24.94, p < 0.0001. PWL: F_(4,82) = 475.9, p < 0.0001. Sham, n = 9. SNI, n = 10. D, E, Quantification of immobility time in the TST (D, t = 4.63, p = 0.0003) and FST (E, t = 4.22, p = 0.0006) in sham (n = 9) and SNI (n = 10) mice. F, G, The mouse brain was collected 2 weeks after SNI or sham surgery for c-Fos-immunostaining. Representative images (F) and quantification (G) of c-Fos(+) neurons in the ipsilateral and contralateral PIC 2 weeks after sham or SNI surgery on the left side. G, Ipsilateral, t = 7.52, p < 0.0001; contralateral, t = 4.05, p = 0.0037; n = 5 mice in each group. H, Top, Schematic diagram for AAV-CaMKII-eYFP injection into the PIC bilaterally of SNI or sham mice. Bottom, Representative traces of depolarizing current injection-evoked firing in PIC neurons 4 weeks after sham (black) or SNI (red) surgery. I, Summary of depolarizing current injection-evoked firing of PIC neurons in SNI and sham mice. F_(1,26) = 8.478, p = 0.0073; n = 5 SNI mice, n = 12 sham mice. J, Left, Top panel, Schematic diagram of virus injection and optical fiber implantation for optogenetic activation of PIC neurons; Bottom panel, Timeline of experiments in L–U. Right, A representative image showing ChR2-eYFP expression in the PIC. Cl, claustrum; Opto, optical fiber. K, Example trace and quantification of ChR2 currents recorded in PIC neurons (n = 5 mice). L–O, Effect of optogenetic activation of PIC neurons on mechanical PWT (L, F_(2,36) = 39.88, p < 0.0001; M, F_(2,24) = 0.01, p = 0.99) and thermal PWL (N, F_(2,36) = 21.61, p < 0.0001; O, F_(2,24) = 0.31, p = 0.73) on both hindpaws in ChR2 mice (n = 10) and eYFP mice (n = 7). P, Schematic diagram of CPP test. Q–S, Representative heat maps (Q) and quantification of time spent (R) and velocity (S) in the blue light-paired chamber during precondition (Day 1) and test sessions in ChR2 mice (n = 7) and eYFP mice (n = 7). R, Time, F_(1,12) = 7.94, p = 0.016. S, Velocity, F_(1,12) = 0.64, p = 0.44. T, U, Quantification of immobility time in the TST (T, t = 0.78, p = 0.45) and FST (U, t = 0.44, p = 0.67) in ChR2 (n = 8) and eYFP (n = 6) mice during the blue light illumination (473 nm, 20 Hz, 5 ms pulse width, 4 mW) of the PIC. *p < 0.05; **p < 0.01. Two-tailed unpaired t test for D, E, G, T, and U. Two-way repeated measures ANOVA with Tukey's post hoc analysis for B, C, I, L–O, R, and S. Scale bar, 100 µm.

Considering that the PIC is only one of the hyperactive brain regions in pain states (Meerwijk et al., 2013; Simons et al., 2014; Zhou et al., 2019; Luan et al., 2020; Jia et al., 2022, 2023a; Yahiro et al., 2023), we performed optogenetic stimulation of PIC neurons to test whether elevation of the activity in PIC neurons is sufficient to reduce pain threshold. We unilaterally injected AAV-CaMKII-ChR2-eYFP or AAV-CaMKII-eYFP into the PIC of mice (ChR2 or eYFP mice) and implanted an optical fiber above the injection site (Fig. 3J). Using the whole-cell patch-clamp technique in the voltage-clamp mode, we confirmed that blue light pulses evoked time-locked inward currents in all recorded ChR2-expressing PIC neurons in live brain slices (Fig. 3K). In pain-like behavioral tests, we observed that blue light illumination of the PIC reduced mechanical and thermal thresholds on both hindpaws in ChR2 mice, but not in eYFP mice (Fig. 3L–O). Consistent with previous studies showing stimulation of excitatory synaptic inputs to the PIC facilitates pain perception (Tan et al., 2017; Bouchatta et al., 2022), these data support that stimulation of PIC neurons reduces pain thresholds in naive mice.

We next examined whether activation of PIC neurons has an impact on the affective aspect of pain. We employed two paradigms to address this issue. First, we utilized a CPP/conditioned place aversion (CPA) paradigm (Fig. 3P). On Day 1, the mice were first allowed to roam in the two chambers in the CPP/CPA box for 20 min in the absence of light stimulation. In Days 2–4, blue light pulses (473 nm, 20 Hz, 5 ms, 4 mW; 15 min/day) were delivered to the PIC when the mice were in their favored chamber. On Day 5, ChR2 mice displayed significant CPA, while control eYFP mice did not show either CPP or CPA (Fig. 3Q,R). We also observed that blue light illumination of the PIC did not alter the traveling velocity in the blue light-paired chamber in either ChR2 or eYFP mice (Fig. 3S). Second, we determined whether hyperactivity in PIC projection neurons is sufficient to induce depression-like behaviors in naive mice. We observed that optogenetic activation (473 nm, 5 ms pulses, 20 Hz) of PIC neurons in mice did not alter the immobility time in the TST and FST (Fig. 3T,U).

These data demonstrate that enhancement of the activity of PIC neurons is sufficient to induce hyperalgesia and aversion, but not depression-like behavior.

Inhibition of PIC neurons suppresses hyperalgesia and depression-like behavior in the chronic pain state

To explore whether PIC neuron activation is necessary for the maintenance of mechanical and thermal thresholds in naive mice, AAV-CaMKII-NpHR-eYFP or AAV-CaMKII-eYFP was injected into the PIC (NpHR or eYFP mice; Fig. 4A,B). We performed patch-clamp recording to confirm that yellow light stimulation (589 nm, 1 s, 2 mW) evoked outward currents in NpHR-eYFP–expressing PIC neurons (Fig. 4B). In NpHR mice, instead of eYFP-mice, yellow light illumination (continuous light, 589 nm laser) of the PIC dramatically elevated mechanical and thermal thresholds on both hindpaws (Fig. 4C–F). In the CPP test, we found that photoinhibition of PIC neurons significantly increased the time spent in the light-paired chamber without altering the locomotor activity (Fig. 4G–I). Furthermore, yellow light illumination (589 nm, 3 min on/30 s off for 6.5 min, 3 mW) of the PIC throughout the test session shortened the immobility time in the TST and FST in NpHR mice relative to eYFP mice (Fig. 4J,K). Our data indicate that inhibition of PIC neurons alters pain thresholds, reward processing, and depression-like behavior in naive mice.

Figure 4.

Inhibition of PIC neurons improves hyperalgesia and emotions in SNI mice. A, Schematic diagram and a representative image showing virus injection and fiber implantation for optogenetic inhibition of PIC neurons. Cl, claustrum. B, Example traces and quantification of NpHR currents recorded in PIC neurons (n = 5 mice). C–F, Effect of yellow light illumination (589 nm, continuous, 3 mW) of PIC neurons on mechanical PWT (C,D) and thermal PWL (E,F) in naive NpHR (n = 15) and eYFP mice (n = 6). C, F_(2,60) = 65.65, p < 0.0001. D, F_(2,20) = 0.24, p = 0.79. E, F_(2,56) = 40.95, p < 0.0001. F, F_(2,20) = 0.06, p = 0.94. G–I, Representative heat maps (G) and quantification of time spent (H) and velocity (I) in the yellow light-paired chamber during the precondition (Day 1) and test sessions for NpHR mice (n = 10) and eYFP mice (n = 6). H, Time, F_(1,14) = 13.24, p = 0.0027. I, Velocity, F_(1,14) = 2.19, p = 0.26. J, K, Quantification of immobility time in the TST (J, t = 2.55, p = 0.02) and FST (K, t = 4.34, p = 0.001) in NpHR (n = 8) and eYFP (n = 6) mice during yellow light illumination of PIC neurons. L, Timeline of experiments for M–U. M–P, Effect of yellow light illumination of the PIC on mechanical PWT and thermal PWL on the injured side in NpHR (n = 8) and eYFP (n = 6) mice subjected to SNI surgery on the side contralateral or ipsilateral to light stimulation. M, F_(2,28) = 21.15, p < 0.0001; N, F_(2,20) = 0.11, p = 0.89; O, F_(2,28) = 53.34, p < 0.0001; P, F_(2,20) = 0.56, p = 0.58. Q–S, Representative heat maps (Q) and quantification of time spent (R) and velocity (S) in the yellow light-paired chamber during the precondition (Day 1) and test sessions in NpHR (n = 10) and eYFP (n = 6) mice 4 weeks after SNI surgery. R, Time, F_(1,14) = 19.15, p = 0.0091. S, Velocity, F_(1,14) = 4.21, p = 0.38. NpHR: mice. T, U, Quantification of immobility time in the TST (T, t = 6.4, p < 0.0001) and FST (U, t = 4.13, p = 0.0012) in NpHR (n = 10) and eYFP (n = 6) mice 4 weeks after SNI surgery during the yellow light illumination of PIC neurons. V, W, Example images (V) and summary (W) showing c-Fos expression in the PIC of SNI mice 4 weeks after SNI surgery with (n = 7 mice) or without (n = 6 mice) 30 min yellow light delivery into the PIC. t = 4.7, p = 0.0006. **p < 0.01. ns, not significant. Two-way repeated measures ANOVA with Tukey's post hoc analysis for C–F, H, I, M–P, R, and S. Two-tailed unpaired t test for J, K, T, U, and W). Scale bar, 100 µm.

We next assessed whether inhibition of PIC neurons in SNI mice results in an analgesic effect (Fig. 4L). In NpHR mice, optogenetic inhibition of PIC neurons in the right hemisphere significantly elevated mechanical and thermal thresholds on the injured side in mice subjected to the SNI surgery on either the left (contralateral) or the right (ipsilateral) side (Fig. 4M,O). In eYFP mice, delivery of yellow light into the PIC in the right hemisphere did not change mechanical and thermal thresholds on the injured side (left or right) in SNI mice (Fig. 4N,P). Four weeks after the SNI surgery, NpHR and eYFP mice were subjected to the CPP test. After 3 d of conditioning paired with yellow light delivery into the PIC (589 nm, 3 min on/30 s off, 3 mW, 15 min/day), NpHR mice, but not eYFP mice, displayed significant CPP (Fig. 4Q–S). As chronic pain induces depression-like behavior (Fig. 3D,E), we explored the effect of the inhibition of PIC neurons on depression-like behavior in SNI mice. We observed that yellow light delivery (589 nm, 3 min on/30 s off for 6.5 min, 3 mW) into the PIC throughout the test session shortened the immobility time in the TST and FST in NpHR mice compared with that in eYFP mice (Fig. 4T,U). These behavioral effects may be associated with the inhibition of PIC projection neurons because yellow light illumination (589 nm, 3 min on/30 s off for 30 min, 3 mW) of the PIC resulted in a significant decrease in the number of c-Fos(+) PIC neurons in NpHR mice subjected to the SNI surgery (Fig. 4V,W).

Taken together, we provide evidence for the significant and bidirectional impact of the PIC on pain-related behaviors. Additionally, the analgesic and antidepressant-like effects of PIC neuron inhibition on neuropathic pain suggest that persistent hyperactivity in PIC neurons may contribute to comorbid chronic pain and depression.

The PIC→BLA and PIC→VM projections are enhanced in chronic pain

The PIC projections are among the major components of the descending pathway en route to the somatosensory cortex, raphe magnus nucleus, and spinal cord dorsal horn to facilitate pain perception (Benison et al., 2011; Tan et al., 2017). The PIC also projects to the thalamus and amygdala (Shi and Cassell, 1998a; Gehrlach et al., 2020). The latter is relevant for its role in affective pain processing (Li et al., 2013; Simons et al., 2014; Gandhi et al., 2020; Becker et al., 2023). We next used AAV viral vector to trace the major downstream nuclei and performed optogenetic experiments to probe those mediating the outcomes of optogenetic modulation of PIC neurons in naive and SNI mice.

We first traced the axonal projections of PIC neurons by injecting AAV-hSyn-DIO-mGFP-synaptophysin-mRuby into the PIC of CaMKII-Cre mice (Fig. 5A). This viral vector labels PIC neurons and their processes with mGFP and axonal terminals with mRuby (Beier et al., 2015; Ji et al., 2023a). Consistent with previous studies (Gehrlach et al., 2020), mRuby-labeled structures were distributed in a large collection of cortical and subcortical areas (Fig. 5B,C). Notably, there was a dense signal of both mGFP and mRuby in the BLA and VM of the thalamus. Considering the involvement of the BLA and VM in pain transmission and modulation, and emotional processing (Craig et al., 1994; Simons et al., 2014; Kuramoto et al., 2015; Gandhi et al., 2020; Becker et al., 2023; Habig et al., 2023), we focused on the BLA and VM in the present study.

To understand the organization of the circuit from the PIC to the BLA and VM, we examined whether the same PIC neurons sent collateral axonal projections to both regions or distinct subpopulations of PIC neurons individually innervated these two nuclei. We injected AAV retro-hSyn-eGFP and AAV retro-hsyn-mCherry, respectively, into the BLA and VM in the same mice to simultaneously label PIC neurons projecting to either the BLA, VM, or both (Fig. 6A–D). The data showed that BLA-projecting PIC (PIC→BLA) neurons and VM-projecting PIC (PIC→VM) neurons were largely separated: ∼57.5% (1,250/2,173) and 60.8% (1,321/2,173) of neurons were labeled by eGFP and mCherry, respectively, and ∼18.5% (402/2,173) of neurons were colabeled by eGFP and mCherry (Fig. 6E,F). Given the activation of PIC neurons in SNI mice and the well-documented occurrence of long-term synaptic modifications in brain regions implicated in chronic pain (Jia et al., 2022, 2023a), we next examined whether PIC→BLA neurons and PIC→VM neurons were hyperactive in SNI mice. We performed whole-cell patch-clamp recording from retrogradely labeled PIC neurons projecting to either BLA (labeled by eGFP only), VM (labeled by mCherry only), or both (labeled by eGFP and mCherry) 4 weeks after SNI or sham operation (Fig. 6G,I,K). Our data showed that these three groups of PIC neurons in SNI mice responded to depolarizing current injections with higher firing rates than those in sham mice (Fig. 6G–L).

Figure 6.

The excitability of PIC neurons and their projections to the BLA and VM are enhanced in SNI mice. A–D, Schematic diagram of representative coronal sections showing virus injection into the VM (AAV-retro-hSyn-mCherry) and BLA (AAV-retro-hSyn-eGFP) to retrogradely label PIC neurons. E, F, Numbers of PIC→VM (red) and PIC→BLA (green) neurons at a series of anterior to posterior levels (n = 5 mice). G–L, Representative images of GFP(+) (G, top panel), mCherry(+) (I, top panel), or GFP(+)–mCherry(+) (K, top panel) PIC neurons for whole-cell current-clamp recording. Representative traces (G, I, and K, bottom panels) and summary (H,J,L) of depolarizing current injection-evoked firing in PIC→BLA neurons, PIC→VM neurons, and colabeled PIC→BLA→VM neurons 4 weeks after sham (black) or SNI (red) surgery. H, F_(1,91) = 134.2, p < 0.0001; J, F_(1,247) = 128.2, p < 0.0001; L, F_(1,104) = 72.96, p < 0.0001. M, Representative images showing the injection of AAV-CaMKII-ChR2-eYFP into the PIC and eYFP-labeled fibers in the VM and BLA. N, O, Example traces (N) and peak amplitude (O) of blue light-evoked (5 ms, 2 mW) photo-EPSCs in BLA neurons from sham and SNI mice 2 weeks after surgery. Amplitude (O), t = 4.82, p < 0.0001. n = 14 cells from four mice in each group. P–R, Representative traces (P) and summary of evoked firing rate (Q) and summary of resting membrane potentials (R) in BLA neurons receiving PIC projections in sham or SNI mice 2 weeks after sham or SNI surgery. Q, F_(1,15) = 20.08, p = 0.0004, n = 5 mice in each group; R, t = 3.59, p = 0.0022. Sham, n = 12. SNI, n = 7. S,T, Example traces (S) and peak amplitude (T) of blue light-evoked (5 ms, 2 mW) photo-EPSCs in VM neurons from sham and SNI mice 2 weeks after surgery. Amplitude (T), t = 2.69, p = 0.016. n = 9 cells in each group. U–W, Representative traces (U) and summary of evoked firing rate (V) and summary of resting membrane potentials (W) in VM neurons receiving PIC projections in sham or SNI mice 2 weeks after surgery. V, F_(1,19) = 13.99, p = 0.0014, n = 5 mice in each group. W, t = 0.28, p = 0.78. Sham, n = 12. SNI, n = 9. *p < 0.05. **p < 0.01. Two-way repeated measures ANOVA for H, J, L,Q, and V. Two-tailed unpaired t test for O, T, R, and W. Scale bar, 100 µm.

To explore potential alterations in the PIC→BLA and PIC→VM projection in chronic pain, we injected AAV-CaMKII-ChR2-eYFP into the PIC and performed unilateral SNI surgery in mice. Four weeks later, SNI mice were killed, and dense eYFP-labeled fibers and terminals were visualized in the VM and BLA (Fig. 6M). We then performed whole-cell patch-clamp recording of blue light-evoked (5 ms, 2 mW) postsynaptic currents (photo-EPSCs) in BLA and VM neurons. We observed photo-EPSCs in 71% (17 of 25) of BLA neurons and 75% (9 out of 12) of VM neurons at a holding potential of −50 mV (Fig. 6N,S). The peak amplitude of photo-EPSCs in both BLA and VM neurons was significantly increased in SNI mice compared with that in sham mice (Fig. 6N,O,S,T). These data suggest that the PIC→BLA and PIC→VM projections are enhanced in SNI mice. We next performed current-clamp recording from BLA and VM neurons innervated by the PIC. These neurons were identified by the appearance of EPSCs upon optogenetic stimulation of PIC axonal fibers (Fig. 6N,S). We observed that injections of depolarizing currents led to higher firing rates in BLA and VM neurons from SNI mice than sham control mice (Fig. 6P,Q,U,V). Additionally, more depolarized resting membrane potentials were observed in BLA neurons, but not in VM neurons, in SNI mice, compared with those in sham mice (Fig. 6R,W). These data suggest that the excitability of both BLA and VM neurons are enhanced in SNI mice.

The BLA and VM mediate control of pain thresholds and depression-like behavior by the PIC

We proposed that the PIC→BLA and PIC→VM pathways might be involved in the regulation of pain behaviors. To investigate this, we performed a bidirectional projection-specific optogenetic modulation of these pathways.

To investigate the effect of activation of the PIC→BLA projection on pain-related behavior, we injected AAV-CaMKII-ChR2-eYFP or AAV-CaMKII-eYFP into the PIC and implanted an optical fiber above the BLA (Fig. 7A,B). We observed that optogenetic activation (473 nm, 5 ms, 20 Hz, 4 mW) of the PIC→BLA projection in one hemisphere significantly reduced mechanical and thermal thresholds on both hindpaws (Fig. 7C1–D2). These results are similar to the outcomes observed upon direct manipulation of PIC neurons. Additionally, activation of the PIC→BLA projection did not result in the development of CPA or CPP (Fig. 7E–G) but had a remarkable effect on depression-like behavior, reflected by a significant increase in the immobility time in the TST and FST in ChR2 mice (Fig. 7H,I).

Figure 7.

The PIC→BLA projection modulates pain thresholds and depression-like behavior. A, Schematic diagram of unilateral virus injection and fiber implantation for optogenetic modulation of the PIC→BLA projection. B, Example images of ChR2-eYFP labeling in the PIC and BLA. Opt, optical fiber. C1–D2, Effect of unilateral blue light illumination (473 nm, 20 Hz, 5 ms pulse width, ∼5 mW) of the PIC→BLA projection on mechanical PWT and thermal PWL in naive ChR2 mice (n = 9) and eYFP mice (n = 8). C1, F_(2,32) = 11.77, p = 0.0001; C2, F_(2,28) = 11.33, p = 0.28; D1, ChR2: F_(2,32) = 18.27, p < 0.0001; D2, eYFP: F_(2,28) = 1.55, p = 0.23. E–G, Representative heat maps (E) and quantification of time spent (F) and velocity (G) in the blue light-paired (in the BLA) chamber during precondition (Day 1) and test sessions in ChR2 (n = 8) and eYFP (n = 8) mice. F, F_(1,14) = 0.12, p = 0.74; G, F_(1,14) = 0.73, p = 0.41. H, I, Summary of immobility time in the TST (H, t = 4.08, p = 0.001) and FST (I, t = 2.36, p = 0.035) in ChR2 (n = 8) and eYFP mice (n = 8) during unilateral blue light illumination (473 nm, 20 Hz, 5 ms pulse width, 4 mW) of the PIC→BLA projection. J, Representative images showing NpHR-eYFP in the PIC and BLA after injection of AAV-CaMKII-NpHR-eYFP into the PIC and optical fiber (Opt) implantation into the BLA. K1–L2, Effect of yellow light illumination (589 nm, continuous, 3 mW) in the PIC→BLA projection on mechanical PWT and thermal PWL in NpHR mice (n = 10) and eYFP mice (n = 6). K1, NpHR: F_(2,36) = 17.41, p < 0.0001; K2, eYFP: F_(2,20) = 0.28, p = 0.76; L1, NpHR: F_(2,36) = 19.7, p < 0.0001; L2, eYFP: F_(2,20) = 0.2, p = 0.82. M–O, Representative heat maps (M) and quantification of time spent (N) and velocity (O) in the yellow light-paired (in the BLA) chamber during precondition (Day 1) and test sessions in NpHR (n = 10) and eYFP (n = 8) mice. M, F_(1,16) = 0.062, p = 0.81; O, F_(1,16) = 0.37, p = 0.55. P, Q, Quantification of immobility time in the TST (P, t = 3.07, p = 0.0073) and FST (Q, t = 4.52, p = 0.0007) in NpHR mice (n = 10) and eYFP mice (n = 8) during unilateral yellow light illumination of the PIC→BLA projection. R1–S2, Effect of yellow light illumination of the PIC→BLA projection in the right hemisphere on mechanical PWT and thermal PWL on the injured side in NpHR mice (n = 10) or eYFP mice (n = 8) subjected to SNI surgery on the left (contralateral) or right (ipsilateral) side. R1, NpHR: F_(1.17,19.83) = 28.65, p < 0.0001; R2, F_(1.93,26.95) = 0.046, p = 0.95; S1, F_(4.84,31.19) = 18.04, p < 0.0001; S2, eYFP: F_(1.69,23.62) = 0.38, p = 0.64. T–V, Representative heat maps (T) and quantification of time spent (U) and velocity (V) in the yellow light-paired chamber during the precondition (Day 1) and test sessions in NpHR mice (n = 10) and eYFP mice (n = 8) 4 weeks after SNI surgery contralateral to light stimulation. U, Time, F_(1,16) = 7.32, p = 0.016; (V) Velocity, F_(1,16) = 0.0077, p = 0.93. W, X, Quantification of immobility time in the TST (W, t = 3.5, p = 0.0032) and FST (X, t = 5.76, p < 0.0001) during the yellow light illumination of the ACC→STN projection in NpHR mice (n = 9) and eYFP mice (n = 8) 4 weeks after SNI surgery on the side contralateral to light stimulation. *p < 0.05. **p < 0.01. ns, not significant. Two-way repeated measures ANOVA with Tukey's post hoc analysis for C1–D2, F, G, K1–L2, N, O, R1–S2, U, and V). Two-tailed unpaired t test for H, I, P, Q, W, and X. Scale bar, 100 µm.

To investigate the effect of inhibition of the PIC→BLA projection on pain-related behavior, we injected AAV-CaMKII-NpHR3.0-eYFP or AAV-CaMKII-eYFP into the PIC and placed an optical fiber above the BLA (Fig. 7J). Yellow light in the BLA led to increased mechanical and thermal thresholds on both hindpaws (Fig. 7K1–L2), as well as an antidepressant-like effect in NpHR mice (Fig. 7P,Q), but without any observed effect in the CPP paradigm (Fig. 7M–O). These data indicate that the PIC→BLA projection regulates pain threshold and depression-like behavior in naive mice.

The analgesic effects elicited by inhibition of the PIC→BLA projection prompted us to investigate whether this pathway could modulate pain-related behaviors in the context of neuropathic pain. Consistent with the observations in naive mice, optogenetic inhibition of the PIC→BLA projection in the right hemisphere led to a significant increase in mechanical and thermal thresholds on the hindpaw of the injured side in SNI mice subjected to nerve injury on the left (contralateral) or right (ipsilateral) side (Fig. 7R1–S2). Furthermore, inhibition of this pathway induced CPP (Fig. 7T–V) and exhibited antidepressant effects in the TST and FST (Fig. 7W,X). These data suggest that inhibition of the PIC→BLA projection may be a promising strategy to improve hyperalgesia and depression-like behavior in chronic pain.

We also performed bidirectional optogenetic modulation of the PIC→VM projection (Fig. 8A,B,J). We observed that unilateral activation or inhibition of the PIC→VM pathway correspondingly reduced or elevated mechanical and thermal thresholds on the ipsilateral hindpaw, but not on the contralateral hindpaw in naive mice (Fig. 8C1–D2,K1–L2). Additionally, stimulation and inhibition of the PIC→VM projection exerted depression-like and antidepressant-like effect in the TST and FST, respectively (Fig. 8H,I,P,Q). Similar to the optogenetic manipulation of the PIC→BLA projection, neither activation nor inhibition of the PIC→VM projection resulted in the development of CPP or CPA (Fig. 8E–G,M–O). In SNI mice, optogenetic inhibition of the PIC→VM projection in the right hemisphere increased mechanical and thermal thresholds on the right (ipsilateral) hindpaw (injured side) but not on the left (contralateral, injured side) hindpaw (Fig. 8R1–S2), established CPP (Fig. 8T–V), and reduced the immobility time in the TST and FST (Fig. 8W,X).

Figure 8.

Optogenetic manipulation of the PIC→VM projection modulates pain threshold and depression-like behaviors. A, Schematic diagram of unilateral virus injection and fiber implantation for optogenetic modulation of the PIC→VM projection. B, Example images of ChR2-eYFP-labeled neurons and axonal fibers in the PIC and VM, respectively. Cl, claustrum. C1–D2, Effect of unilateral blue light illumination (473 nm, 20 Hz, 5 ms pulse width, 4 mW) of the PIC→BLA projection on mechanical PWT and thermal PWL on either side in ChR2 mice (n = 9) and eYFP mice (n = 9). C1, ChR2: F_(2,32) = 5.22, p = 0.01; C2, eYFP: F_(2,32) = 0.33, p = 0.73; n = 9 eYFP mice. D1, ChR2: F_(2,32) = 15.67, p < 0.0001; D2, eYFP: F_(2,32) = 0.94. E–G, Representative heat maps (E) and quantification of time spent (F) and velocity (G) in the blue light-paired (in the VM) chamber during precondition (Day 1) and test session in ChR2 (n = 9) and eYFP (n = 8) mice. F, F_(1,15) = 0.36, p = 0.056; G, F_(1,15) = 0.15, p = 0.70. H, I, Quantification of immobility time in the TST (H, t = 3.39, p = 0.0013) and FST (I, t = 2.32, p = 0.037) in ChR2 (n = 9) and eYFP (n = 8) mice during unilateral blue light illumination (473 nm, 20 Hz, 5 ms pulse width, 4 mW) of the PIC→VM projection. J, Representative images of NpHR-eYFP in the PIC and VM 3 weeks after AAV-CaMKII-NpHR-eYFP injection into the PIC. Cl, claustrum. K1–L2, Effect of unilateral yellow light illumination (589 nm, continuous, 3 mW) of the PIC→VM projection on mechanical PWT and thermal PWL on either side in NpHR (n = 10) and eYFP (n = 6) mice. K1, NpHR: F_(2,36) = 4.58, p = 0.017; K2, eYFP: F_(2,20) = 0.048, p = 0.95; L1, NpHR: F_(2,36) = 16.14, p < 0.0001; L2, eYFP: F_(2,20) = 0.44, p = 0.65. M–O, Representative heat maps (M) and quantification of time spent (N) and velocity (O) in the yellow light-paired (in the VM) chamber during precondition (Day 1) and test sessions in NpHR (n = 9) and eYFP (n = 8) mice. N, F_(1,15) = 0.015, p = 0.90; O, F_(1,15) = 0.94, p = 0.35. P, Q, Quantification of immobility time in the TST (P, t = 3.31, p = 0.0044) and FST (Q, t = 2.14, p = 0.04) in ChR2 (n = 10) and eYFP (n = 8) mice during unilateral yellow light illumination of the PIC→VM projection. R1–S2, Effect of yellow light illumination of the PIC→BLA projection on mechanical PWT and thermal PWL on the injured side in NpHR (n = 10) and eYFP (n = 7) mice subjected to SNI surgery on the side contralateral or ipsilateral to light illumination. R1, NpHR: F_(2,34) = 20.55, p < 0.0001; R2, eYFP: F_(2,24) = 0.0088, p = 0.99; S1, NpHR: F_(2,34) = 8.84, p = 0.0008; S2, eYFP: F_(2,24) = 0.91, p = 0.42. T–V, Representative heat maps (T) and quantification of time spent (U) and velocity (V) in the yellow light-paired chamber during precondition (Day 1) and test sessions in NpHR (n = 9) and eYFP (n = 6) mice 4 weeks after SNI surgery ipsilateral to virus injection. U, Time, F_(1,13) = 5.55, p = 0.035. (V) Velocity, F_(1,13) = 1.63, p = 0.22. W, X, Quantification of immobility time in the TST (W, t = 4.9, p = 0.0002) and FST (X, t = 2.47, p = 0.027) during yellow light illumination of PIC→VM projection in NpHR (n = 9) and eYFP (n = 7) mice 4 weeks after SNI surgery. *p < 0.05. **p < 0.01. Two-way repeated measures ANOVA with Tukey's post hoc analysis for C1–D2, F, G, K1–L2, N, O, R1–S2, U, and V. Two-tailed unpaired t test for H, I, P, Q, W, and X. Scale bar, 100 µm.

The preceding paragraphs reveal that the PIC→BLA and PIC→VM projections modulate pain thresholds and depression-like behaviors in naive mice and mitigate hyperalgesia and depression-like behaviors in SNI mice; they differ from PIC projection neurons in the processing of aversion, reward, and depression-like behavior in naive mice.

Note that the physiological outcomes following optogenetic stimulation of the PIC→BLA and PIC→VM projection may be confounded by the back-propagation of spikes on the terminals, and more sophisticated techniques overcoming this issue may be applied to confirm these findings.

Discussion

The PIC is a hub to integrate multimodal information, including somatosensory, interoceptive, and emotional inputs, from a wide range of cortical and subcortical regions (Gehrlach et al., 2020; Bergeron et al., 2021; He et al., 2022). The involvement of the PIC in pain has been well documented (Benison et al., 2011; Segerdahl et al., 2015; Tan et al., 2017; Bergeron et al., 2021; Nagasaka et al., 2022; Labrakakis, 2023). Emerging evidence convincingly supports that the PIC regulates emotions, including aversion, anxiety, and depression (Schiff et al., 2018; Gehrlach et al., 2019; Girven et al., 2021). Although chronic pain is often comorbid with emotional disorders (Bair et al., 2003; Rayner et al., 2016; GBD 2017 Disease Injury Incidence Prevalence Collaborators, 2018; Zhou et al., 2019; Roughan et al., 2021; Llorca-Torralba et al., 2022), it has not been addressed whether the PIC regulates emotional states in chronic pain or modulates pain states in emotional disorders. Consistent with the role of the PIC in processing multimodal information, we found that PIC neurons are mobilized by pain, aversive, and stressful stimulation in both physiological and neuropathic pain conditions. We also observed that PIC neurons are hyperactive in neuropathic pain mice; optogenetic stimulation of PIC neurons is sufficient to induce hyperalgesia and aversion in naive mice, whereas inhibition of PIC neurons elevated pain thresholds, established CPP, and alleviated depression-like behavior. These results suggest that the hyperactivity of PIC neurons may be associated with hyperalgesia and depression-like behavior in neuropathic pain, and inhibition of PIC neurons may be a promising strategy to mitigate comorbid pain and depression.

Note that stimulation and inhibition of PIC neurons regulate pain states in opposite directions in naive mice. This phenomenon is consistent with the fact that the PIC is a key component in pain processing neural circuitry. The PIC sends projections to join other axonal fibers to form the descending pain pathway regulating pain perception. In these routes, the PIC innervates the somatosensory cortex, periaqueductal gray, raphe magnus nucleus, locus coeruleus, and spinal cord dorsal horn (Shi and Cassell, 1998a,b; Segerdahl et al., 2015; Labrakakis, 2023). The descending pain pathway may mediate the decrease or increase of pain thresholds following stimulation or inhibition of the PIC. This is consistent with a previous study (Bouchatta et al., 2022), but different from another showing no effect on mechanical pain thresholds by a lesion of the caudal granular insular cortex (Benison et al., 2011). The difference in neuron-type selectivity and volume of tissue being affected by optogenetic inhibition and lesion may be potential explanation for this discrepancy.

In addition to alterations of pain thresholds, stimulation and inhibition of PIC neurons correspondingly established CPA and CPP and led to no effect on and alleviation of depression-like behavior, respectively. These effects may be relevant to the connection of the PIC with the limbic circuit, including the nucleus accumbens, central amygdala, BLA, and bed nucleus of stria terminalis (BNST; Schiff et al., 2018; Gehrlach et al., 2019; Girven et al., 2021). The establishment of CPP by the inhibition of PIC neurons suggests the mobilization of the reward center. On the contrary, stimulation and inhibition of the projection from the PIC to the BNST, respectively, led to reinforcing behavior and resulted in aversion and anxiety (Girven et al., 2021). It is different from our finding. Therefore, we propose that the PIC may cause the outcomes we observed in reward processing through nuclei other than the BNST. The possibility that the PIC may affect reward processing complicates the interpretation of CPP following inhibition of the PIC in SNI mice. Our data cannot discriminate whether this effect is mediated by either relief of spontaneous pain, recruitment of reward center, or both.

Besides the abovementioned nuclei, the PIC also projects to the anterior insula, anterior cingulate cortex, thalamus, and amygdala to modulate pain and emotion (Shi and Cassell, 1998a,b; Frot et al., 2014; Gehrlach et al., 2020; Labrakakis, 2023). We performed neuronal tracing and optogenetic labeling with AAV viral vectors. In fact, retrograde viral vector may not fully occupy the whole volume of the BLA and VM and may not ideally transfect PIC neurons projecting to either the BLA, VM, or both. However, our data reveal that a considerable proportion of PIC neurons innervate both the BLA and VM. Under our experimental condition, we observed that the excitability of neurons and the strength of synapses in these projections were all enhanced in neuropathic pain. These data not only provide morphological and functional evidence to confirm the existence of these connections but also hint that these projections may be involved in pathophysiology related to hyperalgesia and depression-like behavior in neuropathic pain.

The BLA is primarily composed of glutamatergic neurons, receiving inputs from multiple brain regions, such as the PIC, anterior cingulate cortex, medial prefrontal cortex, and ventral pallidum (Polepalli et al., 2020; Hajos, 2021; Ji et al., 2023a). It is involved in the processing of pain and negative emotions. Functional brain imaging studies in rodents and humans implicate the neuronal hyperactivity and altered functional connectivity in the BLA in the pathophysiology of chronic pain and depression (Drevets, 2003; Simons et al., 2014; Gandhi et al., 2020). On the contrary, lesion or inhibition of the BLA neurons mitigates chronic neuropathic pain and abolishes comorbid depression-like behavior (Li et al., 2013; Ji et al., 2023a). Although the excitatory synaptic inputs to the BLA are proved to be enhanced in pain processing, the sources of these inputs remain elusive. Our data revealed that the glutamatergic inputs from the PIC to the BLA were enhanced in neuropathic pain. We also observed that optogenetic manipulation of the PIC→BLA projection regulates pain thresholds and depression-like behaviors in both naive and SNI mice. These results are consistent with the previously reported physiological function of the BLA (Drevets, 2003; Simons et al., 2014; Gandhi et al., 2020) and support that the hyperactivity of the PIC→BLA projection is sufficient to induce and necessary to maintain the hyperalgesia state and depression-like behavior in neuropathic pain. Collectively, the PIC→BLA pathway plays an important role in both the sensory and emotional aspects of pain and comorbid chronic pain and depression as well.

The thalamus is the main relay station for nociceptive inputs to cortical and subcortical structures. The VM receives strong inputs from the basal ganglia and sends axons to the layer 1 of the motor cortex and anterior cingulate cortex (Herkenham, 1979; Kuramoto et al., 2015). This connectivity pattern involves the VM in a variety of behaviors, including learning and memory, motor control, emotion, and reward-based behavior. Here, we revealed that similar to the PIC→BLA projection, the PIC→VM projection replicates PIC neurons in pain modulation and emotional processing in both physiological and chronic pain conditions. These data also support that the hyperactivity of the PIC→VM projection was sufficient to induce and necessary to maintain the hyperalgesia state and depression-like behavior in neuropathic pain. Therefore, the hyperactivity in the PIC→VM pathway may also contribute to pathophysiology in comorbid pain and depression in neuropathic pain.

Although either the PIC→BLA projection or the PIC→VM projection replicated the function of PIC neurons in the modulation of pain in both naive and SNI mice, they differ from PIC neurons in the regulation of place preference and depression-like behavior in naive mice. Literature supports that PIC neurons regulate aversion/reward probably through the central amygdala and BNST (Schiff et al., 2018; Gehrlach et al., 2019; Girven et al., 2021). On the other hand, PIC neurons may regulate depression-like behavior not only through the BLA and VM but also through other nuclei-conferring outcomes opposite to the BLA and VM. This may explain why stimulation of PIC neurons did not induce depression-like behavior in naive mice but stimulation of the PIC→BLA and PIC→VM projections did. But further investigations are needed to clarify this hypothesis. Moreover, the PIC→BLA and PIC→VM projections may modulate pain thresholds through different downstream circuits because they regulate pain thresholds respectively on the bilateral side and on the ipsilateral side in naive and SNI mice. To discriminate the laterality of pain modulation by the PIC→BLA or PIC→VM projection, further investigations are needed to address how these projections contribute to or modulate the descending pain pathway.

In summary, this study implicates PIC neurons and their projections to the BLA and VM in pain modulation and emotional processing and demonstrates enhanced neuronal activity and synaptic strength in these pathways as key pathophysiological bases for hyperalgesia and depression-like behavior in chronic neuropathic pain, a model for comorbid pain and depression. The inability to recruit aversion/reward circuits may endow the PIC→BLA or PIC→VM projection an advantage over PIC neurons as intervention targets to improve comorbid pain and depression.