TRPA1 agonist-responsive afferents contribute to central sensitization by suppressing spinal GABAergic interneurons through Somatostatin 2A receptors

Abstract

Altered nociception, a key feature of nociplastic pain, often involves central sensitization. We previously found that central sensitization underlying a nociplastic pain state in female mice depends on the ongoing activity of TRPA1 agonist-responsive afferents. Here we investigated how the activity of these afferents induces and maintains central sensitization at the spinal level. We hypothesized that in the superficial dorsal horn where somatostatin (SST) is expressed in excitatory interneurons and the SST2A receptor (SST2A-R) in GABAergic inhibitory interneurons (GABAn), TRPA1 agonist-responsive afferents stimulate SST-expressing excitatory interneurons (SSTn), leading to GABAn suppression through SST2A-R and resulting in altered nociception. We tested this hypothesis using ex vivo Ca²⁺ imaging of dorsal root-attached spinal cord slices expressing GCaMP6f in either SSTn or GABAn and in vivo assessment of mechanical hypersensitivity. The dorsal root was chemically (with allyl isothiocyanate, AITC) and electrically stimulated to activate TRPA1-expressing nociceptors and all afferents, respectively. The stimulation of dorsal root with AITC excited SSTn. During activation of AITC-responsive afferents, a subset of SSTn showed potentiated responses to both low- and high-threshold afferent inputs, whereas a subset of GABAn showed suppressed responses to those afferents in an SST2A-R-dependent manner. Intrathecally administered SST2A-R antagonist inhibited the development of mechanical hypersensitivity by intraplantar AITC injection and alleviated persistent mechanical hypersensitivity in the murine model of nociplastic pain. These results suggest that the activity of TRPA1 agonist-responsive afferents induces and maintains central sensitization by activating dorsal horn SSTn and suppressing GABAn via SST2A-R, resulting in altered nociception that manifests as mechanical hypersensitivity.

Perspectives:

This article presents experimental evidence that TRPA1 agonist-responsive afferents induce and maintain central sensitization at the spinal level by activating somatostatin (SST)-expressing excitatory interneurons and suppressing GABAergic inhibitory interneurons via SST2A receptors. Spinal SST2A-R may represent a promising target for treating mechanical pain hypersensitivity due to central sensitization by TRPA1 agonist-responsive afferents.

1. Introduction

Chronic pain is a significant global health problem affecting millions of people with a higher prevalence in women^1,2. Many chronic pain patients present with nociplastic pain arising from ‘altered nociception’ in the absence of apparent tissue damage activating nociceptors or somatosensory system damage causing the pain³. While it is well known that intense nociceptor activity, such as that from an injury, sensitizes the central nociceptive system to alter nociception (i.e., central sensitization), it remains unclear how central sensitization is maintained in the absence of injury, leading to nociplastic pain. To facilitate the preclinical investigation of the maintenance mechanisms, we have developed a murine nociplastic pain model in which an acute injury-induced central sensitization is prolonged beyond the normal resolution time by postinjury stimulation⁴. Our prior research using this model revealed that the ongoing nerve activity at the previous injury site, particularly those responsive to the TRPA1 agonist allyl isothiocyanate (AITC), maintains the nociplastic pain state in females^4,5.

Building upon the finding, the present study sought to understand how the activity of TRPA1 agonist-responsive afferents alters central nociceptive processing in the superficial dorsal horn (SDH) where the initial central processing of primary afferent inputs occurs. Recent advances in Ca²⁺ imaging offer an efficient means to monitor the activities of multiple neurons simultaneously during sensory processing in the peripheral or central nervous systems⁶. In this study, we used an ex vivo Ca²⁺ imaging technique for neuron type-specific measurement of SDH neuronal responses to afferent inputs generated by electrical and chemical stimulation of dorsal roots. Because the abovementioned nociplastic pain model manifests the prolongation of mechanical, but not thermal, hypersensitivity⁴, we focused on SDH somatostatin-expressing excitatory interneurons (STTn) known for their critical role in mechanical pain processing⁷. Additionally, we examined SDH GABAergic inhibitory interneurons (GABAn), the malfunction of which produces mechanical allodynia^7–9. In the SDH, virtually all glycinergic neurons are GABAn, and about 50% of SDH GABAn express SST2A receptors (SST2A-R) while nearly all SST2A-R-expressing neurons are GABAn¹⁰. The activation of the SST2A-R hyperpolarizes GABAn, suppressing the neuron’s excitability¹¹. Considering that TRPA1 agonist-responsive afferents significantly contribute to mechanical pain^12,13, we hypothesized that the activity of TRPA1 agonist-responsive afferents stimulates SDH SSTn, which in turn suppresses SST2A-R-expressing GABAn through the receptors (feedback disinhibition, Fig. 1A) based on the gate control theory of pain¹⁴. We tested this hypothesis using Ca²⁺ imaging of SDH SSTn and GABAn ex vivo. Portions of the data have been presented in abstract form¹⁵.

Fig. 1. Schematic illustration of our hypothesis and ex vivo spinal cord slice Ca²⁺ imaging.

(A) We hypothesized that TRPA1-expressing afferents excite somatostatin-expressing neurons (SSTn) in the dorsal horn, which in turn suppresses GABAergic inhibitory neurons (GABAn) through SST2A receptors (SST2A-R). This feedback disinhibition results in increased mechanical pain (altered mechanical nociception). (B) Schematic diagram of ex vivo Ca²⁺ imaging configuration. The sagittal spinal slice with a dorsal root was placed in an imaging chamber superfused with oxygenated artificial cerebrospinal fluid (ASCF). The dorsal root was electrically stimulated with a suction electrode and chemically stimulated by a drug locally applied via a PE tube. (C) Dye application using this local application approach stained only the dorsal root (broken line box), not the spinal cord. (D) A representative series of images showing a Ca²⁺ transient evoked by dorsal root stimulation (delivered immediately before ‘b’, marked as an electric bolt). The graph shows the corresponding fluorescent calcium signal.

2. Materials and Methods

2.1. Animals

All animal use protocols were approved by the University of Texas Medical Branch’s Institutional Animal Care and Use Committee and conducted following the National Institute of Health guidelines. The protocols included humane endpoints for animals showing orbital tightening with failure to move upon stimulation, weight loss ≥20%, or hind limb motor deficit.

SST-Cre (JAX# 013044)¹⁶, GAD2-Cre (JAX# 010802)¹⁷, and Ai95D (GCaMP6f with a floxed STOP cassette, JAX #028865)¹⁸ mouse lines were purchased from the Jackson Laboratory and subsequently bred in-house. Each Cre mouse line was crossbred with the Ai95D line to create heterozygous offspring. These offspring were crossbred again with the homozygous Ai95D line to create Cre-expressing homozygous Ai95D offspring which were used in Ca²⁺ imaging experiments (female N=27, male N=19).

C57BL/6N mice (8–10 weeks of age; N=24 for each sex) were bred in-house or purchased from Charles River (Houston, TX). The mice were housed in groups of up to 5 mice per cage under a 12 hr light–12 hr dark cycle and fed ad libitum. There were no animal inclusion or exclusion criteria except for the strain (and genotype) and age range described above. There was no patient or public involvement in this animal study.

2.2. Materials

Allyl isothiocyanate (AITC), a TRPA1 agonist, was purchased from Fluka (Germany). CYN 154806, an SST2A-R antagonist, and (1R,1’S,3’R/1R,1’R,3’S)-L-054264 (L-054264), an SST2A-R agonist, were obtained from Tocris Bioscience (Bristol, UK). CYN 154806 has been used in the range of 1–20 μM^11,19, and L-054264, 1–10 μM^20,21 for ex vivo or in vitro electrophysiological or Ca²⁺ imaging experiments; in this study, we used 2 μM of CYN 154806 and 5 μM of L-054264 for Ca²⁺ imaging.

2.3. Spinal Cord Slice Preparation

Mice were deeply anesthetized with isoflurane, and the spinal cord was quickly removed to ice-cold, oxygenated (95% O₂ and 5% CO₂) modified artificial cerebrospinal fluid (ACSF) containing (in mM): KCl 2.5, NaH₂PO_4.H₂O 1.2, NaHCO₃ 30, D-glucose 25, HEPES 20, N-Methyl-D-glucamine 93, Na Ascorbate 5, Thiourea 2, Sodium pyruvate 3.0, N-acetyl-L-cystein 5.0, CaCl_2.2H₂O 0.5, and MgSO_4.7H₂O 10 (pH 7.4 and 310–320 mOsm). Sagittal lumbar spinal cord slices, which were 400 μm thick and had dorsal roots (L3 or L4, 8–12 mm) attached, were prepared using a vibratome VT1200S (Leica, Germany). Then, slices were incubated for about 1 hr at 35°C in oxygenated ACSF that contains (in mM): NaCl 124, KCl 2.5, NaH₂PO_4.H₂O 1.2, NaHCO₃ 15, D-glucose 12.5, HEPES 5, CaCl_2.2H₂O 2, and MgSO_4.7H₂O 2 (pH 7.4 and 310–320 mOsm). The slices were transferred into an imaging chamber and perfused with oxygenated ACSF at a rate of 10 ml/min before Ca²⁺ imaging at room temperature (22–25 °C).

2.4. Ca²⁺ imaging

SST-Cre; Ai95D and GAD2-Cre; Ai95D mice were used for Ca²⁺ imaging of SDH SSTn and GABAn, respectively. The sagittal spinal cord slice with the dorsal root was placed in an imaging chamber, and green fluorescent Ca²⁺ signals in these neurons were detected with an Olympus (BX51W1) microscope equipped with an ORCA Flash4.0LT camera (Hamamatsu, Japan) and PRIOR Lumen 300 LED light source (Cambridge, UK). The imaging chamber was continually perfused with ACSF at room temperature (22–25°C). A series of 512 × 512-pixel fluorescence images were acquired with HC Image software (Hamamatsu) at a 2 Hz sampling rate using a 40X objective immersed in ACSF. A suction electrode was used for electrical stimulation of the dorsal root (Fig. 1B). The distance from the tip of the suction electrode to the entrance of the attached dorsal root was ~7 mm. The dorsal root was stimulated at various intensities (i.e. 10, 20, 50, 100, 300, and 500 μA; 0.5 ms pulse) to incrementally activate from low- to high-threshold afferents. We recorded the dorsal root stimulation-evoked Ca²⁺ transients before, during, and after drug treatment. AITC was locally applied to the dorsal root, away from the dorsal horn, via a PE tube attached to the suction electrode. We found that a dye applied in this manner only stained the dorsal root (Fig. 1C). CYN 154806 and L-054264 were applied to the spinal cord through bath superfusion.

Image files were imported to ImageJ (version 1.53v, NIH) for analysis. Individual neurons showing fluorescence changes upon dorsal root stimulation were identified, and the degree of change (ΔF) from the baseline fluorescence (F₀) was normalized to F₀ (ΔF/F₀) (Fig. 1D). When comparing the dorsal root stimulation-evoked responses (i.e., hereafter, termed ‘Ca²⁺ transients’) between the before, during, and after drug treatment phases, we converted the minimum and maximum ΔF/F₀ values obtained in the ‘before treatment’ phase to 0% and 100%, respectively, and expressed Ca²⁺ transients in percentage.

2.5. Classification of neurons based on response profiles

Because of the heterogeneity within SSTn and GABAn populations^22–24, we expected heterogeneous responses to the same drug treatment in each neuronal population. Therefore, we categorized them into three classes based on their maximum Ca²⁺ transients in the ‘during treatment’ phase. Specifically, we first defined the ‘normal variability range’ of the maximum Ca²⁺ transients by applying only the vehicle (ACSF). Ninety-five percent of sampled maximum values during ACSF treatment were within the range of 66.2%-123.1% of the maximum Ca²⁺ transients observed in the ‘before treatment’ phase. Accordingly, we regarded neurons showing their maximum Ca²⁺ transients within this range during drug treatment as being “within the normal variability range,” those of <66.2%, “decreased,” and those of >123.1%, “increased.”

2.6. Immunohistochemistry

After euthanized by isoflurane followed by cervical dislocation, GAD2-Cre;Ai95D mice (N=3) were transcardially perfused with 0.01M phosphate buffered saline and then ice-cold Lana’s fixative (4% paraformaldehyde, 14% picric acid in 0.4 M phosphate buffer). The spinal cord was cryoprotected in 30% sucrose and sectioned at 20 μm thickness. Tissue sections were incubated with chicken anti-GFP antibody (ab13970, 1:1000, Abcam, Cambridge, UK; to detect GCaMP6f) plus rabbit anti-PAX2 antibody (71–6000, 1:500, Invitrogen, Waltham, MA; to detect inhibitory interneurons) for 1–3 days at 4°C. They were then incubated with AlexaFluor488-conjugated goat anti-chicken and AlexaFluor568-conjugated goat anti-rabbit antibodies (both 1:500, Invitrogen) overnight at 4°C. Immunostained tissue sections were photographed using a confocal microscope (A1R, Nikon Instruments Inc., Melville, NY).

2.7. Animal models of pain

To induce central sensitization by activating TRPA1 agonist-responsive afferents, we injected AITC (3 μL, 0.1% in 10% ethanol, 10% Tween-20, and 80% saline) into the plantar side of a hind paw under 2–2.5% isoflurane anesthesia. In experiments using this pain model, mice were randomly allocated to one of the following 4 groups: (1) mice (N=5 for both sexes) treated with the vehicle of CYN 154806 (i.th.) followed by the vehicle of AITC (intraplantar), (2) mice (female N=4, male N=5) treated with CYN 154806 (i.th.) followed by the vehicle of AITC (intraplantar), (3) mice (N=4 for both sexes) treated with the vehicle of CYN 154806 (i.th.) followed by AITC (intraplantar), and (4) mice (female N=5, male N=4) treated with CYN 154806 (i.th.) followed by AITC (intraplantar). In another set of experiments examining the susceptibility of female and male mice to AITC-induced central sensitization, 0.01% AITC was used (N=3 for both sexes).

In addition, we used a mouse model of nociplastic pain produced by intraplantar capsaicin injection followed by stimulation of the injected paw with 40°C warm water⁴. In experiments using this pain model, mice were randomly allocated to one of two groups: (1) mice (N=3 for both sexes) treated with the vehicle of CYN 154806 (i.th.) and (2) mice (N=3 for both sexes) treated with CYN 154806 (i.th.).

2.7. Von Frey Filament Assays

Nociceptive behaviors reflecting mechanical hypersensitivity were assessed at predetermined times after drug treatment (e.g., 0.5, 1, 2, 3, and 4 hours post-injection) using von Frey filaments (VFF) (North Coast Medical, Inc., CA, USA) in a dedicated behavioral study room by researchers blinded to the nature of experimental groups (except for the animal sex and AITC injection which resulted in noticeable swelling of the injection site). Before testing, mice were habituated to the plastic enclosure and the mesh platform^5,25. A calibrated VFF was perpendicularly pressed up against the mid-plantar and the base of the 3^rd and 4^th toe for approximately 3 s. The numbers of withdrawals were determined from 10 applications of VFF forces, with a minimum interval of 30 seconds between applications. Rapid paw withdrawals, biting, shaking, and licking of the paw during or immediately after the stimulus were regarded as a positive nociceptive response. The degree of nociceptive response was expressed as a percent withdrawal.

2.8. Statistical analysis

The intensities of Ca²⁺ signals between experimental conditions were statistically compared using a linear mixed model (LMM) for nested (multiple neurons in single subjects) and repeated design (multiple electrical stimulations of the dorsal root before, during, and after drug treatment) of experiments with the AR1 covariance structure for repeated measures and random slopes for subject variance. The magnitudes of Ca²⁺ transients at each dorsal root stimulation intensity were compared pairwise between the before and during drug treatment phases by Sequential Sidak multiple comparison tests.

The proportions of neurons showing their maximum Ca²⁺ transients decreased, increased, or within the normal variability range during drug treatment were compared between sex or experimental conditions using the Chi-square test; proportions in each category were further compared using z-test with Bonferroni-adjusted p-values.

The numbers of animals for behavioral studies were determined by power analysis using lme4 and simr R packages or G*Power (v. 3.1.9.6) based on initial trial data. Nociceptive responses (% withdrawal) were analyzed using a generalized linear mixed model (GLMM) with a logit link function for binomial distribution, the AR1 covariance structure for repeated measures, and random slopes for subject variance. The GLMM included Group (vehicle vs. drug treatments) and Time (repeated measures) factors. Sequential Sidak correction was used for multiple comparisons between groups at each time point. Statistical analyses were conducted using SPSS version 28 (IBM, NY, USA). No data were excluded from these analyses. Data are expressed as mean ± standard deviation (for Ca²⁺ imaging data) or standard error of the mean (for behavioral data) with n representing the number of neurons and N, that of mice. A p-value ≤ 0.05 was considered statistically significant.

3. Results

3.1. TRPA1 agonist-responsive afferents activate the dorsal horn SSTn and modulate their responses to other afferent inputs.

We hypothesized that TRPA1 agonist-responsive afferents excite SDH SSTn and consequently augment their responses to other afferent inputs (through feedback disinhibition, Fig. 1A). To test this hypothesis using a Ca²⁺ imaging approach, we crossbred SST-Cre mice with Ai95D mice. The Cre-dependent reporter gene expression in this SST-Cre line was previously shown to co-localize with naïve Sst mRNA expression extensively (82±2%)²³. In the SDH of the crossbreed offspring, we recorded Ca²⁺ signals from SSTn (Fig. 2A) while locally applying AITC (50 μM) to the dorsal root in female mice. This concentration of AITC was previously shown to evoke action potentials effectively in mouse DRG neurons while not significantly desensitizing TRPA1-mediated currents even in the presence of extracellular Ca^{2+ 26,27}. As shown in Fig. 2B, SSTn were rapidly and robustly activated by AITC-responsive afferents. After the cessation of the stimulation and washing out, the activity of SSTn readily returned to the baseline level.

Fig. 2. TRPA1 agonist-responsive afferents excite SSTn in the dorsal horn and alter their responses to dorsal root electrical stimulation in female mice.

(A) A schematic showing Ca²⁺ imaging of SSTn (under an objective) and afferent manipulations for this experiment (AITC application to the dorsal root and dorsal root electrical stimulation [DR e-stim], LT, low-threshold, and HT, high-threshold). (B) The average (top) and individual (bottom) traces of fluorescent Ca²⁺ signals (ΔF/F₀) from SSTn in the SDH before, during, and after local application of the TRPA1 agonist AITC (50 μM for 3 minutes) to the dorsal root. The black bar indicates the AITC treatment. (C) Ca²⁺ transients were evoked by DR e-stim (0.5 ms pulse, 10–500 μA); three classes of SSTn were observed: their maximum Ca²⁺ transients increased (32.5%), decreased (10.0%), or within the normal variability range (57.5%) during AITC application. The top panels are representative images of SDH SST neurons activated by DR e-stim (500 μA) before, during, and after the AITC application. The magenta, white, and black arrows indicate neurons showing their Ca²⁺ transients increased, decreased, and within normal variability range activity, respectively (scale bar=20 μm). See Methods 2.5 for classification details. Data are expressed as mean (line) ± SD (shaded area). *p<0.05 and **p<0.01 vs. corresponding Ca²⁺ transients in the ‘before’ phase by sequential Sidak multiple comparison tests following linear mixed model analysis.

We then determined if the ongoing stimulation of AITC-responsive afferents influences the degree of SSTn activation by low- to high-threshold afferents, which is expected in the presence of feedback disinhibition (i.e., nociceptors activate excitatory interneurons which in turn suppress inhibitory interneurons that normally reduces the nociceptor inputs to the excitatory interneurons). In 32.5% of SSTn (n=26 of 80, N=6) responding to dorsal root electrical stimulation, AITC application to the dorsal root increased the maximum Ca²⁺ transients evoked by dorsal root electrical stimulation with elevated resting Ca²⁺ levels. In other words, during the AITC application, the maximum Ca²⁺ transients were >123.1% of those obtained before the AITC application in these SSTn (see Method section 2.5 for classification methods). In 57.5% of SSTn (n=46 of 80, N=6), AITC application did not significantly affect the maximum Ca²⁺ transients (i.e., their maximum responses were within the normal variability range). However, even in this population, their responses to low-threshold (20–50 μA) afferent inputs were found to be significantly increased during AITC application (t(810)=2.855, p=0.004 at 20 μA; t(810)=3.312, p=0.002 at 50 μA by sequential Sidak tests). Therefore, while AITC-responsive afferents were being stimulated, a total of 90% of SSTn showed augmented responses to other afferent inputs. The remaining 10% of SSTn (n=8 of 80, N=6) showed decreased responses to high-threshold (100–500 μA) afferent inputs during AITC application (Fig. 2C).

We then asked if SDH SSTn in male mice would similarly respond to AITC-responsive afferents (note that a nociplastic pain state in male mice was found to be dependent on reactive spinal microglia, not AITC-responsive afferent inputs arising from the previous injury area^5,28,29). AITC-responsive afferents effectively activated SSTn in males (Fig. 3A–B) as in females, and the relative proportion of SSTn showing increased maximum Ca²⁺ transients (19.0%, n=8 of 42, N=6) during AITC application was not significantly different from its female counterpart. However, we found a greater proportion (81.0%, n=34 of 42, N=6) of SSTn falling into the class showing their maximum Ca²⁺ transients within the normal variability range during AITC application in males than in females, with no SSTn showing decreased maximum Ca²⁺ transients during AITC application (Table 1 and Fig. 3C). Another difference was that male SSTn in the “within the normal variability range” class did not show increased responses to low-threshold afferent inputs, unlike their female counterparts.

Fig. 3. TRPA1 agonist-responsive afferents excite SSTn in the dorsal horn and alter their responses to dorsal root electrical stimulation in male mice.

(A) A schematic showing Ca²⁺ imaging of SSTn (under an objective) and afferent manipulations for this experiment (AITC application and dorsal root electrical stimulation [DR e-stim], LT, low-threshold, and HT, high-threshold). (B) The average (top) and individual (bottom) traces of fluorescent Ca²⁺ signals (ΔF/F₀) from SSTn in the SDH before, during, and after local AITC application to the dorsal root (50 μM, 3 min). The black bar indicates the AITC treatment. (C) Ca²⁺ transients were evoked by DR e-stim (0.5 ms pulse, 10–500 μA); two classes of SSTn were observed: maximum Ca²⁺ transients increased (19.0%) or within the normal variability range (81.0%) during AITC application. The top panels are representative images of SDH SST neurons activated by DR e-stim (500 μA) before, during, and after the AITC application. The magenta and black arrows indicate neurons showing their Ca²⁺ transients increased and within normal variability range activity, respectively (scale bar=20 μm). Data are expressed as mean (line) ± SD (shaded area). **p<0.01 vs. corresponding Ca²⁺ transients in the ‘before’ phase by sequential Sidak multiple comparison tests following linear mixed model analysis.

Table 1.

Relative proportions of SSTn and GABAn in each class based on the change in their maximum Ca²⁺ transients during drug treatment.

	SSTn		GABAn
	AITC		AITC		CYN+AITC		L-054264
	Female	Male	Female	Male	Female	Male	Female	Male
Increased	32.5% (26/80)	19.0% (8/42)	0.0% (0/63)	0.0% (0/28)	1.7% (2/115)	15.2%*# (5/33)	0.0% (0/44)	11.1%* (3/27)
Decreased	10.0% (8/80)	0.0%* (0/42)	39.7% (25/63)	10.7%* (3/28)	0.9%# (1/115)	0.0% (0/33)	15.9% (7/44)	18.5% (5/27)
Within normal variability range	57.5% (46/80)	81.0%* (34/42)	60.3% (38/63)	89.3%* (25/28)	97.4%# (112/115)	84.8%* (28/33)	84.1% (37/44)	70.4% (19/27)

^*p<0.05 between female vs. male in each treatment

^#p<0.05 between AITC vs. CYN+AITC in each sex. Numbers in parentheses indicate the number of neurons in each class per total sampled neurons.

Collectively, these findings suggest that TRPA1 agonist-responsive afferents excite SDH SSTn (thus producing nociception by the afferent inputs) and influence the neurons’ response to low- and high-threshold afferent inputs in a subset of SSTn (thus resulting in altered nociception), aligning with previous findings that optogenetic activation of spinal SSTn produces nociceptive behaviors and increases mechanical sensitivity [5].

3.2. TRPA1 agonist-responsive afferents suppress the dorsal horn GABAn via SST2A-R.

We hypothesized that ongoing activity of AITC-responsive afferents suppresses SST2A-R-expressing GABAn in the SDH (Fig. 1A). To test this hypothesis using a Ca²⁺ imaging approach (Fig. 4A), we crossbred GAD2-Cre mice with Ai95D mice and determined if the offspring express GCaMP6f in GABAn. We found robust colocalization of GCaMP6f and PAX2, a marker of GABAn in the SDH³¹; 86±3% (N=3) of GCaMP6f-expressing SDH neurons were PAX2-immunoreactive, and 77±4% (N=3) of PAX2-immunoreactive SDH neurons expressed GCaMP6f (Fig. 4B).

Fig. 4. TRPA1 agonist-responsive afferents suppress GABAn activation by other afferent inputs.

(A) A schematic showing Ca²⁺ imaging of GABAn (under an objective) and afferent manipulations for this experiment (AITC application and dorsal root electrical stimulation [DR e-stim], LT, low-threshold, and HT, high-threshold). (B) In the SDH of GAD2-Cre;Ai95D mice, GCaMP6f was expressed in PAX2-immunoreactive neurons (scale bar=50 μm). Ca²⁺ transients were evoked in GABAn by DR e-stim (0.5 ms pulse, 10–500 μA) before, during, and after AITC application in (C) females and (D) males. Two classes of GABAn were observed: maximum Ca²⁺ transients decreased (39.7% in females and 10.7% in males) or within the normal variability range (60.3% in females and 89.3% in males) during AITC application. The top panels are representative images of SDH GABAn activated by DR e-stim (500 μA) before, during, and after the AITC application. The white and black arrows indicate neurons showing their Ca²⁺ transients decreased and within normal variability range activity, respectively (scale bar=20 μm). Data are expressed as mean (line) ± SD (shaded area). *p<0.05 and **p<0.01 vs. corresponding Ca²⁺ transients in the ‘before’ phase by sequential Sidak multiple comparison tests following linear mixed model analysis.

Based on the reported proportion of SST2A-R-expressing GABAn in the mouse SDH^10,32, we expected that <50% of GABAn would be suppressed by SSTn-activating AITC-responsive afferent inputs. As presented in Table 1 and Fig. 4C, 39.7% of GABAn in female mice (n=25 of 63, N=9) showed their maximum Ca²⁺ transients decreased during AITC application. After washing out of AITC, their Ca²⁺ transients were not fully restored despite a trend toward recovery. The remaining 60.3% of GABAn in females (n=38 of 63, N=9) showed their maximum Ca²⁺ transients within the normal variability range during AITC application. Compared with females, a significantly smaller proportion of GABAn (10.7%, n=3 of 28, N=7) showed decreased Ca²⁺ transients to dorsal root electrical stimulation during AITC application in males (Table 1 and Fig. 4D).

We then examined if the suppression was meditated through SST2A-R (and thus preventable by an antagonist of the receptor) (Fig. 5A). As shown in Table 1 and Fig. 5B, the SST2A-R antagonist CYN 154806 (2 μM) significantly reduced the proportion of GABAn (0.9%, n=1 of 115, N=7) whose maximum Ca²⁺ transient was suppressed during AITC application (i.e., CYN+AITC phase in Fig. 5B) in females. Because there was only one GABAn in this class, we did not statistically test the reduction in Ca²⁺ transients. In the CYN+AITC phase, the vast majority of GABAn (97.4%, n=112 of 115, N=7) showed their Ca²⁺ transients within the normal variability range. In these GABAn, we detected a minute but significant increase by CYN 154806 itself in their Ca²⁺ transients (i.e., CYN phase in Fig. 5B; t(2664)=7.134 for the before vs. CYN phases, p<0.001 by sequential Sidak test), suggesting that SST2A-R-mediated inhibition was included in the afferent input-evoked GABAn activation at baseline in our experimental setting. When AITC was applied in this condition (i.e., CYN+AITC phase), the Ca²⁺ transients upon high-threshold (300–500 μA) afferent inputs were attenuated when compared pairwise with their counterparts in the CYN phase (t(2664)=7.766, p<0.001 at 300 μA; t(2664)=6.702, p<0.001 at 300 μA by sequential Sidak tests), suggesting that the antagonism might have been counteracted to some degree by AITC-induced release of the endogenous ligand SST. The remaining 1.7% of GABAn (n=2 of 115, N=7) showed their maximum Ca²⁺ transients increased. In these GABAn, the resting Ca²⁺ levels were noticeably elevated by CYN 154806. Because there were only two neurons in this class, we did not statistically test the increase in Ca²⁺ transients. In male mice, after blocking SST2A-R, no SSTn showed decreased Ca²⁺ transients (n=0 of 33, N=3; Table 1 and Fig. 5C); overall, an increase in Ca²⁺ transients was detected in the CYN+AITC phase.

Fig. 5. TRPA1 agonist-responsive afferents suppress GABAn via SST2A-R signaling.

(A) A schematic showing Ca²⁺ imaging of GABAn (under an objective) and experimental manipulations for this experiment (AITC application, dorsal root electrical stimulation [DR e-stim], and SST2A-R antagonist application, LT, low-threshold, and HT, high-threshold). Ca²⁺ transients were evoked in GABAn by DR e-stim (0.5 ms pulse, 10–500 μA) before, during, and after AITC application in the presence of the SST2A-R antagonist CYN 154806 (CYN) in (B) females and (C) males. Three classes of SSTn were observed: maximum Ca²⁺ transients increased (1.7% in females and 15.2% in males), decreased (0.9% in females only), or within the normal variability range (97.4% in females and 84.8% in males) during AITC application. The top panels are representative images of SDH GABAn activated by DR e-stim (500 μA) before, during, and after the CYN application with or without AITC. The magenta, white, and black arrows indicate neurons showing their Ca²⁺ transients increased, decreased, and within normal variability range activity, respectively (scale bar=20 μm). Data are expressed as mean (line) ± SD (shaded area). *p<0.05 and **p<0.01 vs. corresponding Ca²⁺ transients in the ‘before’ phase; #p<0.05 and ##p<0.01 vs. corresponding Ca²⁺ transients in the ‘CYN’ phase by sequential Sidak multiple comparison tests following linear mixed model analysis.

Next, we bypassed the AITC-responsive afferent inputs and directly activated the SST2A-R with its agonist L-054264 (5 μM) (Fig. 6A). As shown in Fig. 6B, the Ca²⁺ transients evoked by dorsal root electrical stimulation were decreased in 15.9% of GABAn (n=7 of 44, N=5) in the presence of the SST2A-R agonist in females. Interestingly, while the remaining 84.1% of GABAn (n=37 of 44, N=5) showed their maximum Ca²⁺ transients within the normal variability range during the agonist treatment, their responses to high-threshold (300–500 μA) afferent inputs were found to be significantly reduced by L-054264 (t(648)=2.672, p=0.008 at 300 μA; t(648)=2.856, p=0.004 at 500 μA by sequential Sidak tests). This suggests that this class includes GABAn whose Ca²⁺ transients before and during the SST2A-R agonist application are significantly different when compared pairwise, but the reduction is to a limited extent, still within the normal variability range. In males, L-054264 increased Ca²⁺ transients in 11.1% of GABAn (n=3 of 27, N=3; Table 1 and Fig. 6C) while decreasing Ca²⁺ transients in 18.5% of GABAn (n=5 of 27, N=3; Table 1 and Fig. 6C).

Fig. 6. SST2A-R agonist suppresses GABAn activation by afferent inputs.

(A) A schematic showing Ca²⁺ imaging of GABAn (under an objective) and experimental manipulations for this experiment (dorsal root electrical stimulation [DR e-stim] and SST2A-R agonist application, LT, low-threshold, and HT, high-threshold). Ca²⁺ transients were evoked in GABAn by DR e-stim (0.5 ms pulse, 10–500 μA) before, during, and after the SST2A-R agonist L-054264 in (B) females and (C) males. Three classes of GABAn were observed: maximum Ca²⁺ transients increased (11.1% in males only), decreased (15.9% in females and 18.5% in males), or within the normal variability range (84.1% in females and 70.4% in males) during the L-054246 application. The top panels are representative images of SDH GABAn activated by DR e-stim (500 μA) before, during, and after the L-054264 application. The white and black arrows indicate neurons showing their Ca²⁺ transients decreased and within normal variability range activity, respectively (scale bar=20 μm). Data are expressed as mean (line) ± SD (shaded area). *p<0.05 and **p<0.01 vs. corresponding Ca²⁺ transients in the ‘before’ phase by sequential Sidak multiple comparison tests following linear mixed model analysis.

3.3. TRPA1 agonist-responsive afferents induce mechanical hypersensitivity via SST2A-R activation at the spinal level.

As an aggregate, the above results indicate that AITC-responsive afferents activate a subset of SDH SSTn and suppress a subset of GABAn through SST2A-R, which may underlie altered mechanical nociception (i.e., mechanical hypersensitivity) in pain conditions where AITC-responsive afferents are active. We tested this idea by intrathecally injecting the SST2A-R antagonist CYN 154806 (1 μg, 5 μL) before intradermally injecting AITC (0.1%, 3 μL) in the hind paw and then mechanically stimulating the AITC-injected hind paw in the area outside the injection site with von Frey filaments (VFF). The SST2A-R antagonist acting at the spinal level significantly inhibited the development of AITC-induced mechanical hypersensitivity in both females (Fig. 7A; F(3,112)=12.412, p<0.001 between all groups; t(112)=5.444, p<0.001 for saline+vehicle vs. saline+AITC; t(112)=3.203, p=0.007 for saline+AITC vs. CYN+AITC by sequential Sidak tests) and males (Fig. 7B, F(3,112)=24.923, p<0.001 between all groups; t(112)=7.489, p<0.001 for saline+vehicle vs. saline+AITC; t(112)=5.078, p<0.001 for saline+AITC vs. CYN+AITC by sequential Sidak test), supporting the receptor’s critical role in altered mechanical nociception caused by AITC-responsive afferent activity in both sexes. As our ex vivo Ca²⁺ imaging results showed that AITC at a non-saturating concentration suppressed a smaller proportion of GABAn (i.e., lesser disinhibition) in males than in females, we then examined if a 10-times lower concentration (0.01%) of AITC would produce a lesser degree of hypersensitivity in males than in females. As shown in Fig. 7C, unlike females, males failed to develop mechanical hypersensitivity in response to this low concentration of AITC (F(1,28)=51.084, p<0.001 by GLMM analysis).

Fig. 7. SST2A-R antagonism at the spinal level prevents TRPA1 agonist-induced mechanical hypersensitivity.

(A) Female and (B) male mice received an intrathecal injection of SST2A-R antagonist CYN 154806 (CYN 1 μg, injection marked by a blue dotted line) or its vehicle (saline). Thirty minutes later, AITC (0.1%, 3 μL, injection marked by a red broken line) or its vehicle was injected into the plantar side of a hind paw. Mechanical sensitivity was assessed using a von Frey filament (0.1 g force) before and at predetermined times after AITC injection (N=4–6 per group). In both sexes, pretreatment with intrathecal CYN 154806 (CYN+AITC) significantly inhibited the development of AITC-induced mechanical hypersensitivity compared to vehicle pretreatment (saline+AITC). *p<0.05 and **p<0.01 for saline+vehicle vs. saline+AITC groups; ^#p< 0.05 and ^##p<0.01 for saline+AITC vs. CYN+AITC groups by sequential Sidak multiple comparison tests following generalized linear mixed model analysis. (C) AITC at a low concentration (0.01%, 3 μL, injection marked by a red broken line) induced mechanical hypersensitivity only in females. **p<0.01 vs. male groups by sequential Sidak multiple comparison tests following generalized linear mixed model analysis.

We then tested if intrathecal CYN 154806 would be also effective on the prolonged mechanical hypersensitivity in our nociplastic pain model in which only females showed increased ongoing activity of AITC-responsive afferents in the affected skin [11]. As shown in Fig. 8A, the hypersensitivity was significantly inhibited by the antagonist (F(1,24)=19.619, p<0.001 by GLMM analysis), suggesting that in this model, the ongoing activity of AITC-responsive afferents maintains the nociplastic pain state through SST2A-R-mediated disinhibition. Next, we examined the effects of intrathecal CYN 154806 in our male nociplastic pain model whose prolonged mechanical hypersensitivity depends on spinal microglia^4,5,29 Interestingly, the SST2A-R antagonist also alleviated their nociplastic mechanical hypersensitivity (Fig. 8B, F(1,24)=297.82, p<0.001 by GLMM analysis), suggesting SST2A-R-mediated disinhibition as a converging mechanism underlying this chronic pain state in both sexes.

Fig. 8. SST2A-R antagonism at the spinal level alleviates nociplastic mechanical hypersensitivity.

(A) Female and (B) male mice in a nociplastic pain state received an intrathecal injection of SST2A-R antagonist CYN 154806 (1 μg, injection marked by a blue dotted line) or saline. The SST2A-R antagonist significantly attenuated mechanical hypersensitivity in both sexes. ^##p<0.01 vs. saline group by sequential Sidak multiple comparison tests following generalized linear mixed model analysis.

4. Discussion

This study demonstrates that the TRPA1 agonist-responsive afferents excite SDH SSTn (and potentiate their response to other afferent inputs in a subset of them) while suppressing GABAn via SST2A-R. Although we did not experimentally specify the neuronal source of SST for this GABAn suppression, it likely includes SDH SSTn, considering their robust activation upon AITC-responsive afferent inputs. This feedback disinhibition provides insights into the gate control mechanism originally depicted as direct suppression of inhibitory interneurons by small-diameter fibers¹⁴, explaining how ongoing AITC-responsive afferent activity may maintain central sensitization to alter the processing of nociception via disinhibition. We acknowledge that central sensitization may be mechanistically different when it is driven by afferents other than AITC-responsive afferents. For instance, in ex vivo spinal cord slices obtained after intraplantar capsaicin injection in vivo, mainly non-SSTn in the SDH were found to sensitize, firing action potentials upon Aβ afferent inputs before the arrival of Aβ afferent-evoked inhibitory inputs, which was due to decreased A-type potassium currents in the non-SSTn rather than disinhibition³³. This is intriguing because capsaicin-induced secondary mechanical allodynia (but not mechanical hyperalgesia) disappears when the ongoing nerve activity at the injection site is inhibited (and reappears as the inhibition wears off)^34,35. It should be mentioned that the abovementioned changes in non-SSTn are likely independent of “ongoing” nociceptor activity arising from the capsaicin injection site as such inputs were severed in the ex vivo spinal cord slices. Therefore, it seems that central sensitization mechanisms dynamically maintained by ongoing nociceptor inputs differ from those independent of such inputs. Considering that TRPA1- and TRPV1-expressing afferents partially overlap^36,37, it is not unlikely that SST2A-R-mediated disinhibition also plays a role in capsaicin-induced central sensitization while the SDH circuits receive ongoing inputs from capsaicin-responsive afferents.

In rodent SDH, SSTn consists of excitatory neurons of multiple neurochemical phenotypes, and they can be subtyped by expression of neuropeptides (e.g., neurokinin B, neurotensin, gastric-releasing peptide (GRP), etc.) and G protein-coupled receptors (e.g., NPY1R, GRPR, CCKBR, etc.)^22,23. In this regard, it would be interesting to examine if there is a correlation between these neurochemical phenotypes and the three functional classes that we identified in this study by their Ca²⁺ transient responses to AITC-responsive afferent inputs. As some of the abovementioned receptors are G_q-coupled and thus able to increase intracellular Ca²⁺, their agonists may be used in future studies (as in the study by Warwick et al.³⁸) to further delineate the circuit components of central sensitization by the ongoing activity of AITC-responsive afferents.

While we report the role of SST2A-R in disinhibition causing “altered nociception” in this study, such a mechanism was previously recognized in the spinal processing of pruriception. SST-expressing primary afferents function as pruriceptors^39,40, and SDH SSTn is crucial for itch as well as mechanical pain^30,41,42; SST in the spinal pruriceptive circuit was shown to produce/facilitate itch through SST2A-R-mediated inhibition of a subset of GABAn³². However, it should be noted that intrathecal injection of the SST analog octreotide produces not only itch-related behaviors (such as scratching and biting) but also pain-related ones (licking)³². Likewise, optogenetic activation of SDH SSTn produces nociception and induces mechanical hypersensitivity as well as itch hypersensitivity³⁰. In line with this, activation of SDH neurons that express gastrin-releasing peptide receptor (GRPR), which are primarily regarded as a tertiary pruriceptor in the SDH, produces both itch- and pain-related behaviors; furthermore, these neurons are activated by noxious thermal and mechanical stimulation as well as pruritogens⁴³. Our study adds to a growing body of literature revealing the intersection of nociception and pruriception processing at the spinal level, focusing on the role of endogenous SST signaling through SST2A-R in increased mechanical nociception.

The foregoing suggests that endogenous SST from the spinal sensory circuitry will increase nociception. Interestingly, however, conditional knockout of SST from either primary afferents or SDH neurons increased nociception at baseline⁴⁰, indicating an “antinociceptive” role of this neuropeptide and thus being contradictory to our finding that mechanical hypersensitivity was effectively alleviated by antagonism of endogenous SST signaling via SST2A-R. While this contradiction in the role of endogenous SST in nociception remains to be further investigated, it is worth mentioning that the SST4-R is also present in primary afferents and SDH neurons, producing antinociception upon activation^44,45. Therefore, it could be that the anti- and pro-nociceptive functions of endogenous SST are determined by different types of SST receptors engaged at the spinal level in different pain states. In a normal state, sensory neuronal SST may not sufficiently activate SST2A-R signaling to increase mechanical nociception at a behavioral level. It would be interesting to examine in future studies if the conditional knockout of SST from the two sensory neuronal populations (i.e., primary afferents and SDH neurons) would influence the development of mechanical hypersensitivity in pathological pain conditions.

Compared with females, we observed a smaller proportion of GABAn whose maximum Ca²⁺ transients were suppressed during AITC application in males (the decrease in this proportion was reflected in the increased proportion of GABAn showing their maximum Ca²⁺ transients within the normal variability range during AITC application). Because ~50% of SDH GABAn were found to express SST2A-R in male mice by immunohistochemistry^10,32, our findings in male mice are not accounted for by the proportion of GABAn expressing SST2A-R. In line with this Ca²⁺ imaging result, we found that unlike in females, intraplantar injection of a low concentration (0.01%) of AITC did not induce mechanical hypersensitivity in males. In this regard, we acknowledge the limitation of our approach of not using multiple concentrations of AITC for ex vivo Ca²⁺ imaging. Considering that a high concentration (0.1%) of AITC effectively induced a comparable degree of mechanical hypersensitivity in both sexes, greater than 50 μM AITC may be needed in males to suppress GABAn through SST2A-R to the extent found in females with 50 μM AITC. While specifying the neurobiological basis for this sex difference is beyond the scope of this study, multiple reasons can be speculated, including potential sex differences in the sensitivity of TRPA1-expressing afferents to AITC (estrogen metabolites are shown to sensitize TRPA1 channel function and increase nociceptor activity⁴⁶), in the SDH microcircuits connecting the afferents to SSTn, in the expression quantity and signaling efficacy of SST2A-R in GABAn, etc.

In this study, we focused on “TRPA1 agonist-responsive afferents” because a nociplastic pain state underlying prolonged mechanical hypersensitivity in female mice depends on their ongoing activity⁵, which aligns well with clinical findings that silencing ongoing nerve activity in the previous injury site immediately turned off central sensitization in patients (4 women and 1 man) of complex regional pain syndrome^28,47. By comparison, the nociplastic pain state in male mice is maintained by spinal microglia^4,5,29. This difference, together with the abovementioned finding that a smaller proportion of SDH GABAn was suppressed by AITC-responsive afferents through SST2A-R in males, led us to expect that unlike in females, SST2A-R would not play a significant role in mechanical hypersensitivity in males. However, intrathecal CYN 154806 robustly inhibited not only intraplantar AITC-induced mechanical hypersensitivity but also nociplastic mechanical hypersensitivity in males, suggesting that spinal microglia-dependent mechanisms of central sensitization also involve SST2A-R-mediated suppression of GABAn. Understanding how reactive microglia exert such disinhibition is the topic of our next study.

In conclusion, this study demonstrates that TRPA1 agonist-responsive afferents activate SDH SSTn and inhibit GABAn via SST2A-R receptors, ultimately amplifying mechanical sensitivity. Despite sex differences in the proportion of SDH GABAn whose responses to afferent inputs are decreased by AITC-responsive afferents ex vivo and the susceptibility to develop mechanical hypersensitivity upon low degree activation of the afferents in vivo, we observed no difference in the effect of intrathecal SST2A-R antagonist on mechanical hypersensitivity between males and females at a behavioral level. Therefore, targeting SST2A-R may represent a promising avenue for developing new therapeutic agents for treating chronic pain due to central sensitization by TRPA1 agonist-responsive afferents.

Highlights.

• TRPA1 agonist-responsive afferents induce and maintain central sensitization

• These afferents activate somatostatin (SST)-expressing excitatory neurons in the dorsal horn

• These afferents suppress dorsal horn inhibitory neurons via SST2A receptors

• Antagonism of spinal SST2A receptors inhibits mechanical hypersensitivity