Spinal somatostatin-positive interneurons transmit chemical itch

Abstract

Recent studies have made significant progress in identifying distinct populations of peripheral neurons involved in itch transmission, whereas the cellular identity of spinal interneurons that contribute to itch processing is still a debate. Combining genetic and pharmacological ablation of spinal excitatory neuronal subtypes and behavioral assays, we demonstrate that spinal somatostatin-positive (SOM⁺) excitatory interneurons transmit pruritic sensation. We found that the ablation of spinal SOM⁺/Lbx1⁺ (SOM^Lbx1) neurons caused significant attenuation of scratching responses evoked by various chemical pruritogens (chemical itch). In an attempt to identify substrates of spinal itch neural circuit, we observed that spinal SOM⁺ neurons partially overlapped with neurons expressing natriuretic peptide receptor A (Npra), the receptor of peripheral itch transmitter B-type natriuretic peptide. Spinal SOM⁺ neurons, however, did not show any overlap with itch transmission neurons expressing gastrin-releasing peptide receptor in the dorsal spinal cord, and the gastrin-releasing peptide–triggered scratching responses were intact after ablating spinal SOM⁺ neurons. Dual ablation of SOM^Lbx1 and Npra⁺ neurons in the spinal cord reduced chemical itch responses to a greater extent than ablation of SOM^Lbx1 or Npra⁺ neurons alone, suggesting the existence of parallel spinal pathways transmitting chemical itch. Furthermore, we showed that SOM peptide modulated itch processing through disinhibition of somatostatin receptor 2A–positive inhibitory interneuron. Together, our findings reveal a novel spinal mechanism for sensory encoding of itch perception.

In Brief

We report the functional synergy of parallel spinal circuits in the transmission of chemical itch and disinhibitory role of pruritic peptide SOM in mediating itch sensation.

1. Introduction

Itch is defined as an unpleasant sensation that causes a desire to scratch. There are 2 forms of itch: chemical itch and mechanical itch.^11,13 Chemical itch can be evoked by a wide range of chemical compounds and enzymes, including small molecules, peptides, cytokines, lipid components, miRNA, and proteases.^11,17 These pruritogens activate sensory nerve endings in the skin, which trigger electrical impulses that are conducted to the spinal cord and the brain, generating itch sensation.^4,26

Functional assessments have been used to characterize subtypes of neurons that can transmit chemical itch information. In the periphery, 3 types of nonpeptidergic nociceptors NP1, 2, and 3 are proposed to signal itch,^9,27,48 including NP1 neurons expressing Mgr-related G-protein–coupled receptor D, which responds to β-alanine;²⁸ NP2 neurons expressing MrgprA3 that mediates itch-evoked responses to histamine, chloroquine, and bovine adrenal medulla 8–22 (BAM8-22)^16,29; and NP3 neurons expressing B-type natriuretic peptide ([BNP] or natriuretic polypeptide b [Nppb]) or somatostatin (SOM).^20,31 However, the spinal mechanisms of sensory encoding of pruritic information received from the periphery are still under debate.

Gastrin-releasing peptide (GRP) has emerged as a key modulator in chemical itch transmission.^45,46 Ablation of GRP receptor (GRPR)⁺ neurons in the superficial dorsal horn abolishes scratching response to a wide variety of pruritogens.⁴⁶ But whether the source of GRP is dorsal root ganglia (DRG) neurons or spinal interneurons remains controversial. Various studies have shown GRP is expressed in a subset of primary sensory neurons and activates GRPR⁺ interneurons in the spinal cord.^3,37,45,46 However, other researches have reported undetectable GRP expression in DRG.^9,27,31,48 Instead, neurotransmitter Nppb secreted by peripheral NP3 itch neurons is supposed to trigger the release of GRP from spinal natriuretic peptide receptor A (Npra)⁺ neurons and then excites GRPR⁺ neurons.³¹ These results demonstrate the essential role of the GRPR⁺ neurons for chemical itch transmission but do not conclusively show other spinal circuits that are involved in itch transmission. Interestingly, recent mouse genetic studies have shown that deletion of the transcription factor T cell leukemia homeobox 3 (Tlx3) or the testicular orphan nuclear receptor (TR4) attenuates the scratching responses to pruritogens due to the loss of excitatory interneurons expressing GRP, GRPR, SOM, and preprotachykinin 1,^50,52 suggesting that multiple spinal pathways contribute to neural encoding of chemical itch.

Intrathecal administration of SOM peptide can induce biting and scratching in rats,⁴² implicating a crucial function of this peptide in the itch neural circuit. Recently, spinal SOM neurons and SOM peptide released from the DRG and the spinal cord have been reported to regulate pain and itch transmission.^10,20 Here, we used an intersectional genetic strategy to selectively ablate spinal SOM⁺/Lbx1⁺ (SOM^Lbx1) excitatory interneurons in the adult mice, which represent about 80% SOM⁺ neurons in the dorsal spinal cord.¹² Subsequent behavioral and histochemical studies showed that parallel neural pathways in the spinal dorsal horn, including SOM^Lbx1 excitatory interneurons, are required to transmit chemical itch. Pharmacological studies revealed that the neuropeptide SOM triggered scratching responses by perturbing neuronal activity of SOM receptor Sst2A⁺ inhibitory interneurons in the spinal cord.

2. Methods

2.1. Genetic ablation of SOM^Lbx1 neurons

All animal experiment protocols were approved by the Unit for Laboratory Animal Medicine, University of Michigan, and all the animal handling was performed in compliance with the guidelines of the Institutional Animal Care & Use Committee, University of Michigan. SOM-IRES-Cre (SOM^Cre), ROSA26^{CAG-loxP-STOP-loxP-tdTomato}(Rosa^tdTomato, Ai14), Lbx1-Flpo (Lbx1^Flpo), and Tau^{LoxP-STOP-LoxP-FRT-STOP-FRT-DTR}(Tau^DTR) were cross-bred to obtain a quadruple mouse line.¹² In the Lbx1^Flpo strain, Flpo is driven from the promoter of Lbx1, which encodes transcription factor Lbx1 that determines neuronal cell fate and identity in the developing dorsal spinal cord and dorsal hindbrain.^15,35 Hence, the Lbx1^Flpo strain can specifically mark the neuronal population of dorsal spinal cord neurons and dorsal hindbrain. The intersectional design enabled the expression of diphtheria toxin receptor in SOM^Lbx1 neurons in the dorsal spinal cord and dorsal hindbrain. At the age of 6 to 10 weeks, animals were injected with diphtheria toxin (DTX) intraperitoneally to ablate the SOM^Lbx1 neurons. A dose of 1.5 μg of DTX per 30 g of animal weight was administered, and the toxin dose was repeated 3 days after the first injection. The efficiency of ablation was assessed by analyzing the tdTomato signal. Behavioral and histochemical experiments were performed 4 weeks after the DTX injection.

2.2. Pharmacological ablation of Npra neurons

Npra receptor–expressing spinal cord interneurons were ablated through intrathecal injection of Nppb-tagged saporin toxin (Advanced Targeting Systems, San Diego, CA. 5 μg in 10 μL), whereas control littermates were injected with blank saporin (Advanced Targeting Systems, 10 μL). The toxin was injected in the lumbar segment 3 to 4. Behavioral analyses were performed 1 week after the toxin injection.

2.3. c-Fos induction

c-Fos experiments were conducted keeping the mouse head restrained with an Elizabethan collar (Harvard Apparatus, Holliston, MA) to avoid the biting and/or licking response. To assess the c-Fos activation in SOM^Lbx1-ablated and control mice, compound 48/80 (MilliporeSigma, St. Louis, MO. 20 μg in 10 μL) or chloroquine (MilliporeSigma, 30 μg in 10 uL) was injected in the hairy skin of hind paw of the animal. After 2 hours, the lumbar spinal cord was dissected. Sagittal sections of the spinal cord were further processed for c-Fos immunodetection using rabbit anti-c-Fos (1:1,000, MilliporeSigma).

2.4. In situ hybridization

In situ hybridization (ISH) was performed as described previously.¹² All the images were acquired with Leica DFC9000 GT (Germany). The tdTomato fluorescent signal was photographed first, followed by ISH for respective mRNA. The pseudofluorescent ISH signals (for GRP, GRPR, and Npra) were converted from bright field images and later merged onto the fluorescent tdTomato images using Photoshop software.

2.5. Behavioral testing

Itch behavioral tests were performed as described previously (Liu et al., 2010). All animals were habituated to the Hargreaves chamber twice before the pruritus assessment. On the day of the experiment, animals were placed in Hargreaves chamber and habituated for 15 minutes. The baseline behavior of the mice was video recorded for 15 minutes. Compound 48/80 (MilliporeSigma, 20 μg), or chloroquine (MilliporeSigma, 50 μg), or SLIGRL (Bachem, 100 μg), or β-alanine (MilliporeSigma, 50 mM) or Me-5-HT (MilliporeSigma, 50 μg) in 50 μL of sterile saline was injected intradermally into the nape area, and the behavior was recorded for 30 minutes. We have interpreted the scratching in response to these chemicals as a behavioral read out for itch.^30,44 Nppb (Bachem, Bubendorf, Switzerland. 1.5 nmol in 10 μL), SOM (Abcam, Cambridge, United Kingdom. 3 nmol in 10 μL), or GRP (Bachem, 1 nmol in 10 μL) was injected intrathecally, and the scratching responses were recorded for 30 minutes after the injection. Sst2A-selective antagonist CYN154806 (Tocris, Bristol, United Kingdom. 3 nmol in 10 μL) was used to block Sst2A receptor in the spinal cord. CYN154806 was preinjected intrathecally 30 minutes before pruritus assessment.

2.6. Spinal cord slice preparation

Adult mice (P23-30) were deeply anesthetized with isoflurane (VetOne, Boise, Idaho) and decapitated. The lumber spinal cord was then removed and transferred to preoxygenated ice-cold cutting solution that contained (in mM) NaCl 80, KCl 2.5, NaHCO₃ 25, NaH₂PO₄ 1.25, MgCl₂ 3.5, sucrose 75, CaCl₂ 0.5, sodium ascorbate 1.3, and sodium pyruvate 3.0 (pH at 7.4 and osmolality at 310-320 mOsm). Dura mater was removed, and all the dorsal and ventral roots were cut; the pia-arachnoid membrane was then removed. Spinal cord slices (300 μm) were cut transversely by a VT1200s vibratome (Leica, Germany) and then incubated at 33°C for 1 hour in preoxygenated cutting solution that was then replaced by recording solution slowly, which contained (in mM) NaCl 125, KCl 2.5, NaHCO₃ 26, NaH₂PO₄ 1.25, MgCl₂ 1, D-glucose 25, CaCl₂ 2, sodium ascorbate 1.3, and sodium pyruvate 3.0 (pH at 7.2 and osmolality at 310-320 mOsm). All chemicals were purchased from MilliporeSigma (St. Louis, MO).

2.7. Whole-cell patch-clamp recordings

Spinal cord slices were transferred to a recording chamber after incubation, perfused with preoxygenated recording solution at a rate of 5 mL/minutes at room temperature. Whole-cell patch-clamp recordings were then made on random pickup dorsal horn neurons using borosilicate glass pipettes (Sutter instrument, Novato, CA) with resistance of 3 to 6 MΩ, filled with internal solution that contained (in mM) potassium gluconate 130, KCl 5, Na₂ATP 4, NaGTP 0.5, HEPES 20, and EGTA 0.5 (pH 7.28 and osmolality at 310-320 mOsm). Biocytin (MilliporeSigma, 0.5%) was added in the internal solution to label the neurons recorded. pClamp 10.0 software (Molecular Devices, San Jose, CA) with MultiClamp 700B patch-clamp amplifier was used to acquire the signals. The signals were digitized by Digidata 1550B (Molecular Devices, San Jose, CA), low-pass filtered on-line at 2 kHz and digitized at 5 kHz. Somatostatin peptide (Abcam, 2 μM) was added in the perfusion solution to test its effect on dorsal horn neurons.

2.8. Biocytin labeling

Biocytin-filled electrodes were withdrawn from the recorded neurons after a minimum of 20 minutes of whole-cell recording. The slice was immersed in 4% paraformaldehyde for 2 hours at room temperature and washed in PBS (3 × 20 minutes). The tissue was then incubated with Alexa-488 conjugated streptavidin (1:200; Life technologies, Carlsbad, CA) for 2 hours at room temperature. After washing in PBS (3 × 20 minutes), fluorescent signals were photographed under a fluorescent microscope (Leica DMi8, Germany).

2.9. Statistics

Data shown are represented as mean ± SEM. Statistical analysis was performed using Prism (GraphPad). Statistical significance between different experimental groups was calculated using Student t test; P < 0.05 was considered as statistically significant. For multiple comparisons among different ablated groups, one-way analysis of variance with Tukey's post hoc test was used.

3. Results

3.1. Spinal SOM^Lbx1 neurons are crucial to transmit chemical itch

To study the role of spinal SOM^Lbx1 neurons in itch transmission, we used a genetically engineered mouse-line, Tau^DTR; Rosa^tdTomato; Lbx1^Flpo; and SOM^Cre.¹² The specific design was adapted to target spinal SOM^Lbx1 neurons for ablation through Flpo- and Cre-dependent strategic expression of diphtheria toxin receptor.¹² With the administration of diphtheria toxin (DTX), we were able to ablate SOM^Lbx1 neurons with high specificity (Fig. 1A), as the expression of Cre and Flpo was transcriptionally coupled to the expression of SOM and Lbx1, respectively.

Figure 1.: Induced acute chemical itch sensation in control and SOM^Lbx1-ablated animals. (A) Fluorescent images showing efficacy of ablation of SOM^Lbx1 neurons (marked by tdTomato expression) in the dorsal spinal cord, scale 50 μm. Bar graphs showing itch generation after administration of (B) compound 48/80 (n = 11 for control; n = 8 for ablated), (C) chloroquine (n = 11 for control; n = 6 for ablated), (D) SLIGRL (n = 7 for control; n = 7 for ablated), or (E) Me-5-HT (n = 6 for control; n = 6 for ablated). (F) β-alanine (n = 10 for control; n = 12 for ablated). Data shown are a representation of mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001 with respect to control, Student unpaired t test.

To investigate whether spinal SOM^Lbx1 neurons are a component of any specific pruritogen-activated itch neural circuit, we then assessed itch responses after subcutaneous injection of various pruritogens in control and SOM^Lbx1-ablated animals. A significant reduction in the itch responses was observed after the administration of compound 48/80, chloroquine, protease-activated receptor 2–activating peptide Ser-Leu-Ile-Gly-Arg-Leu-NH2 (SLIGRL), 2-Methyl-5-HT (Me-5-HT), or β-alanine (Fig. 1B–F), indicating that spinal SOM^Lbx1 neurons are required to transmit itch elicited by chemically heterogeneous pruritogens. Also, the selected pruritogens can activate particular or multiple populations of dorsal root ganglion neurons. β-alanine can activate NP1, chloroquine, and SLIGRL can activate NP2, serotonin (5-HT) can activate NP3, whereas compound 48/80 can activate the NP2 and NP3.^11,48

To test whether pruritogen administration can activate spinal SOM⁺ neurons, we injected compound 48/80 or chloroquine in the hairy skin of hind paw and monitored the activation of spinal cord neurons by c-Fos induction. We found that compound 48/80 treatment and chloroquine can induce c-Fos expression in the lumbar dorsal horn. In the compound 48/80-administered group, ∼14% of these c-Fos⁺ neurons were SOM^tdTomato+, whereas the percentage was reduced to ∼1% in the ablated group (Fig. 2A and B). In the chloroquine-administered group, ∼24% of the c-Fos⁺ neurons were SOM^tdTomato+, which was decreased to ∼7% in the ablated group (Fig. 2A and C). In SOM^Lbx1-ablated mice, the total number of c-Fos⁺ neurons was largely reduced to a mere ∼27% in compound 48/80 group and ∼39% in the chloroquine group (Fig. 2B and C). These results indicate that SOM⁺ spinal neurons are activated during pruritogen-triggered itch transmission and are required to transmit chemical itch.

Figure 2.: Activation of SOM^Lbx1 neurons during chemical itch transmission. (A) Compound 48/80 or chloroquine was injected in the hind paw of the animal, followed by assessment of c-Fos activity in the dorsal horn. Left panel (sagittal sections) shows histochemistry images of c-Fos immunoreactivity (green) in control and ablated animals after Compound 48/80 and chloroquine administration. Right panel shows magnified insets. Arrows indicate c-Fos–positive SOM^tdTomato+ neurons, whereas arrowheads indicate c-Fos–negative SOM^tdTomato+ neurons, scale 50 μm. (B) Left panel bar graph represents the quantified number of c-Fos–positive cells detected per section in control and SOM^Lbx1-ablated animals upon 48/80 administration. Right panel bar graph represents quantified the c-Fos–positive SOM^tdTomato+ neurons over total activated neurons in control and ablated animals upon compound 48/80 injection. (C) Left panel shows bar graph representing quantified c-Fos activation in response to chloroquine in control and ablated animals. Right panel shows bar graph representing the c-Fos–positive SOM^tdTomato+ neurons in response to chloroquine over total chloroquine-induced activation of neurons. Control (compound 48/80: n = 3 mice, chloroquine: n = 3 mice. 3-5 sections from each mouse). Ablated (compound 48/80: n = 3 mice, chloroquine: n = 3 mice. 3-5 sections from each mouse). Data shown are a representation of mean ± SEM, ***P < 0.0005 and *P < 0.05 with respect to control, Student unpaired t test.

3.2. Identification of spinal SOM⁺ neurons as a component of itch circuit

The complex mode and hierarchy of action of different peptidergic molecules released by itch-encoding neurons is not completely understood. Gastrin-releasing peptide has been highly implicated as an itch-transmitting peptide.⁴⁵ Natriuretic polypeptide b (Nppb) is a periphery-derived itch neurotransmitter. Exogenous administration of Nppb can elicit increased scratching bouts.³¹ Through ISH, we investigated whether the population of SOM⁺ neurons was distinct or had any overlap with the reported itch transmission neural elements. SOM^tdTomato+ neurons had a small overlap with GRP-positive neurons (16%, Fig. 3A) and no overlap with GRPR⁺ neurons (Fig. 3A). Interestingly, 37% of Npra⁺ neurons also expressed tdTomato (Fig. 3A). Concomitant expression of SOM with other itch-implicated substrates strongly suggests SOM⁺ neurons are functional units of the itch-encoding neural pathway.

Figure 3.: Spinal SOM^Lbx1 neurons are a component of spinal itch circuit. (A) ISH images showing mRNA localization (green) of GRP, GRPR, and Npra, exhibiting overlap with tdTomato signal (red); scale 50 μm. Right panel shows magnified images of the respective insets. The percentage is calculated as double positive neurons (mRNA and SOM^tdTomato+) over total number of indicated mRNA⁺ neurons or SOM^tdTomato+ neurons. Arrows indicate SOM^tdTomato+ neurons positive for GRP or Npra, whereas arrowheads indicate GRP⁺, GRPR⁺, or Npra⁺ neurons negative for tdTomato signal. GRP (n = 3 mice, 3 sections from each mouse, total number of tdTomato⁺ analyzed = 937), GRPR⁺ (n = 3 mice, 3 sections from each mouse, total number of tdTomato⁺ analyzed = 961), and NPRA (n = 4 mice, 3-4 sections from mouse, total number of tdTomato⁺ analyzed = 1621). (B and C) Bar graphs showing itch responses to intrathecally injected itch neuropeptides GRP (n = 6 for control; n = 6 for ablated) and Nppb (n = 5 for control; n = 5 for ablated). Data shown are a representation of mean ± SEM, *P < 0.05 with respect to control, Student unpaired t test. GRPR⁺, gastrin-releasing peptide receptor; ISH, in situ hybridization; Nppb, natriuretic polypeptide; Npra, natriuretic peptide receptor A.

To further address our hypothesis, we analyzed itch behavior with intrathecal injection of GRP and Nppb in control and SOM-ablated animals. With GRP administration, there was no effect on itching behavior, suggesting GRPR is downstream to SOM^Lbx1-mediated itch signaling (Fig. 3B). Albeit, with intrathecal administration of Nppb in SOM^Lbx1-ablated animals, the itch was drastically reduced (Fig. 3C), hinting that spinal SOM⁺ neurons are required for Nppb-mediated itch transmission.

As shown previously, there was a notable population of neurons concordantly expressing SOM and Npra, coexisting with distinct groups of SOM- and Npra-expressing neurons. To answer a probable query whether these neuronal subpopulations act synergistically or in parallel to mediate itch, we analyzed how the individual or concomitant ablation of these neurons affected the itch behavior. To specifically ablate the Npra neurons, we used cell toxin saporin tagged with Npra ligand (Nppb-Sap). We then confirmed the efficiency of the ablation through ISH for Npra mRNA (Fig. 4A), the ablation had an efficacy of more than 90% (control = 40.88 ± 3.14 in hemisection of the dorsal spinal cord; Npra-ablated = 3.33 ± 0.52 in hemisection of the dorsal spinal cord). Generation of itch was assessed with compound 48/80 or chloroquine in the dual-ablated, single-ablated, and blank saporin–treated control mice. Dual-ablated mice showed significantly reduced itch response in comparison with spinal Npra- or SOM^Lbx1-ablated mice (Fig. 4B). Interestingly, the dual ablation of spinal SOM^Lbx1 and Npra⁺ neurons did not abolish the itch completely, implying other spinal GRP-secreting neurons are a part of the chemical itch circuit and/or an appreciable quantity of GRP is being released by the peripheral sensory neurons that can diffuse to act on the spinal cord.

Figure 4.: Spinal neurons transmitting chemical-induced pruritus. (A) Images showing the efficacy of ablation of SOM^Lbx1 and Npra neurons in the dorsal horn of the spinal cord. Upper panel shows ISH signals for Npra mRNA, and lower panel shows tdTomato signal in control and dual-ablated; scale 100 μm. (B) Bar graph represents quantified itch bouts in the indicated group of animals in response to compound 48/80 (n = 6 for blank-sap; n = 6 for blank-sap+SOM-abl; n = 8 for Npra-abl; n = 9 for dual-abl) and chloroquine CQ (n = 10 for blank-sap; n = 6 for blank-sap+SOM-abl; n = 7 for Npra-abl; n = 8 for dual-abl). Given data are a representation of mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001 with respect to control; #P < 0.05, ##P <0.005 with respect to single-ablated groups. One-way analysis of variance, Tukey's post hoc analysis. ISH, in situ hybridization; Npra, natriuretic peptide receptor A.

3.3. Neuropeptide SOM-mediated disinhibition enhances itch processing in the spinal cord

As observed that the spinal SOM^Lbx1 neurons are critical players for itch transmission, we further queried the mechanism of action of spinal SOM^Lbx1 neurons. Inhibitory GPCR, SOM receptor 2A (Sst2A), is the dominantly expressed form of SOM receptor subtype in Bhlhb5 inhibitory neurons of dorsal spinal cord and has been implicated in itch signal processing.²² To address the role of SOM-Sst2A pathway in itch transmission, we performed electrophysiological analysis to study the effect of SOM peptide on Sst2A⁺ inhibitory interneurons. Patch-clamped Sst2A⁺ neurons exhibiting tonic firing pattern were filled with biocytin and then later processed for Sst2A immunostaining. Application of SOM peptide induced hyperpolarization of the outward current and decreased threshold of current injection–induced action potential in Sst2A⁺ neurons (Fig. 5A–D). Consistent with previous studies,^42,51 we found that intrathecal SOM administration elicited scratching behavior in mice. Interestingly, antagonist CYN154806-mediated blockade of SOM receptor Sst2A abolished SOM-triggered scratching behaviors (Fig. 5E), suggesting Sst2A receptor is the major form of SOM receptors to exert the inhibitory function. Moreover, ablating spinal SOM^Lbx1 neurons abolished SOM-triggered scratching behaviors, indicating spinal SOM^Lbx1 neurons are the itch transmission neurons for SOM-induced itch. Our results suggest the SOM^Lbx1 neuron–mediated disinhibition may play essential role in itch transmission. To further examine the function of SOM in itch transmission, we tested the effect of Sst2A antagonist CYN154806 in chemical itch. Intrathecal administration of CYN154806 significantly downregulated the histamine-dependent and histamine-independent itch transmission (Fig. 5F and G). Hence, the presented results indicate that SOM facilitates itch processing in the spinal cord through Sst2A-mediated disinhibition.

Figure 5.: SOM peptide inhibits the itch-gating neurons. (A) Micropictographs showing the biocytin-filled Sst2A⁺ tonic inhibitory neurons analyzed electrophysiologically following SOM peptide administration. (B) Evoked outward current trace recorded from Sst2A⁺ inhibitory neurons. SOM peptide application (2 μM) leads to hyperpolarization. Horizontal bar indicates peptide application period of 1 minute. (C) Analysis of current injection induced firing patterns of Sst2A⁺ inhibitory neurons prebath and postbath application of SOM peptide (2 μM). (D) Bar graph represents quantified AP threshold (pA) before and after SOM peptide application. (E) Analysis of pruritic response upon intrathecal administration of SOM peptide with or without SOM receptor selective antagonist (CYN154806) in control or ablated animals (n = 6 for control; n = 6 for SOM-abl; n = 5 for SOM + Cyn). *P < 0.05 with respect to control, NS, not significant, Student unpaired t test. (F and G) Analyzed scratching bouts in response to (F) compound 48/80 or (G) chloroquine with or without administration of CYN154806 (for compound 48/80: n = 11 for control; n = 11 for Cyn; for CQ: n = 8 for control; n = 8 for Cyn). Data sets are cumulative of independent experiments (mean ± SEM), *P < 0.05, **P < 0.01 with respect to control, Student unpaired t test.

4. Discussion

4.1. Heterogeneous subpopulations of spinal SOM⁺ excitatory interneurons transmitting pain and itch

The heterogeneous populations of excitatory and inhibitory interneurons in the spinal dorsal horn have important roles in processing distinct sensory modalities, such as pain and itch.^13,23 Different subpopulations of SOM⁺ neurons coexist in the superficial dorsal horn, as defined by differences in firing properties, molecular marker profiles, and morphology.¹² Our previous studies reveal that spinal SOM⁺ neurons are functionally heterogeneous.^8,12 The SOM^Lbx1 neurons in lamina IIo, which receive monosynaptic Aδ and C fiber inputs, are involved in the processing of acute intense mechanical pain.¹² The SOM^Lbx1 neurons in lamina II, which coexpress calretinin (calbindin 2), might contribute to the transmission of acute punctate mechanical pain.¹² The SOM^Lbx1 neurons present in laminae IIi-III, which receive Aβ-fiber inputs and transiently express VGLUT3, are essential to transmit dynamic mechanical allodynia after nerve injury or inflammation.^8,12 In this article, we report that the ablation of SOM^Lbx1 neurons results in loss of chemical itch, strengthening the notion of functional diversity of these neurons. Consistently, optogenetic activation of spinal SOM⁺ neurons increased histamine-induced itch and mechanical pain.¹⁰ To dissect the precise role of different SOM population in regulating different somatosensations, the selective manipulation of functionally dedicated spinal SOM^Lbx1 neurons is further warranted.

Our present study showed that distinct subsets of spinal SOM^Lbx1 neurons act in synergy to transmit chemical itch. We found that ∼37% of Npra⁺ neurons are SOM⁺ (Fig. 3A), and ∼22% SOM⁺ neurons express Npra (Fig. 3A), suggesting a partial overlap between the 2 populations. By ablating spinal SOM⁺ neurons, Nppb-induced scratching was drastically reduced, whereas GRP-induced scratching was unaltered (Fig. 3B and C), demonstrating that SOM⁺/Npra⁺ neurons function upstream of GRPR⁺ neurons in the itch circuit. Interestingly, dual ablation of both SOM^Lbx1 and Npra⁺ neurons in the spinal cord results in dramatic loss of acute itch behavior, compared with the ablation of SOM^Lbx1 neurons or Npra⁺ neurons alone. Thus, our results suggest that 3 parallel excitatory neuronal populations converge to GRPR⁺ neurons in the spinal cord for itch transmission: SOM⁺/Npra⁻, SOM⁺/Npra⁺, and SOM⁻/Npra⁺ (Fig. 6). However, our study does not exclude the possibility of neuropeptide GRP being released from DRG neurons and/or other spinal circuits.⁴⁵

Figure 6.: Proposed schematic representing the cellular and molecular cascade of SOM^Lbx1 neurons facilitated itch information processing. The 3 different populations of primary sensory itch afferents NP1, NP2, and NP3 communicate the pruritus information to 3 parallel pathways in the spinal cord, including SOM⁺/Npra⁺/GRP^+/−, SOM⁺/Npra⁻/GRP^+/−, and SOM⁻/Npra⁺/GRP^+/− neurons, to be relayed to the downstream itch circuit. The SOM^Lbx1 neurons and downstream GRPR⁺ neurons are gated by Sst2A⁺ inhibitory interneurons. Both spinal SOM^Lbx1 neurons and NP3 sensory neurons that express Nppb/SOM release SOM peptide to supress Sst2A⁺ inhibitory interneurons. TRPV1⁺ pain fibers can impede the itch transmission by activating Sst2A⁺-mediated inhibition of spinal SOM^Lbx1 neurons. Black arrows indicate undefined monosynaptic or polysynaptic pathways. PN: projections neurons that transmit itch information to the brain. GRPR⁺, gastrin-releasing peptide receptor; Npra, natriuretic peptide receptor A.

4.2. Neuropeptide SOM as a neuromodulator for itch transmission

Recent studies have implicated multiple neurotransmitters/neuromodulators in acute and chronic itch transmission, including glutamate, GRP, substance P, Nppb, and neuromedin B.^{1,2,24,31,34,45,49} Antagonism of spinal AMPA receptors largely attenuates itch transmission,^1,24 and the deletion of glutamate transporter vglut2 in the parabrachial nucleus has been shown to impair acute and chronic itch,³⁴ indicating that spinal SOM⁺ excitatory interneurons probably transmit itch information mainly through the fast-acting neurotransmitter glutamate.

The function of SOM in pain has been extensively studied, with reports of pronociceptive and antinociceptive effects.^{7,21,32,33,39–41,51} Strikingly, previous studies found that SOM peptide can elicit scratching responses in rats,^42,51 but the exact function of this peptide in the itch transmission is still unexplored. Our present study indicates that SOM acts as a neuromodulator to amplify itch processing. SOM is an inhibitory peptide, which binds to specific cell-surface SOM receptors.³⁸ Among the 5 types of SOM receptors, Sst2 transcripts are dominantly detected in the dorsal spinal cord (Allen Brain Atlas, http://mousespinal.brain-map.org/). Sst2A⁺ neurons are GABAergic, and more than 80% of these neurons express glycine,⁴⁷ indicating Sst2A is exclusively expressed in inhibitory interneurons. A recent study has shown that optogenetic activation of spinal SOM⁺ neurons results in an increase in histamine-evoked itch behavior that may be mediated through SOM-Sst2A signaling.¹⁰ Consistently, our whole-cell patch-clamp recordings in Sst2A⁺ neurons showed that SOM peptide can evoke hyperpolarization and could decrease the actional potential threshold generated by depolarizing current pulses (Fig. 5A–D). Thus, our present results revealed that the propruritic effect of SOM in the spinal cord is mediated primarily through Sst2A receptors, as blockade of Sst2A receptors abolishes SOM-induced scratching responses (Fig. 5E). Moreover, blockade of Sst2A downregulates pruritogen-evoked itch responses (Fig. 5F and G), suggesting a tonic inhibition that is provided to the itch circuit by Sst2A⁺ inhibitory interneurons. When the tonic inhibition is subdued by inhibitory peptide SOM (disinhibition), there is a heightened relay of itch sensation. But when the tonic inhibition is strengthened through elimination of restrain on the tonic inhibition, there is reduced transmission of itch signal (Fig. 6).

How do spinal Sst2A⁺ neurons inhibit itching transmission? A previous report suggested that the inhibition of itch by dorsal horn Bhlhb5⁺/Sst2A⁺ interneurons depends on the action of the opioid peptide dynorphin (Dyn).²² However, ablation of spinal Dyn⁺ neurons¹² and embryonic deletion of Dyn (Pdyn^−/−)³⁶ do not affect itching behaviors. Contrary to these findings, Huang et al.²⁰ have shown that chemogenetic activation of Pdyn^Cre neurons downmodulates itch transmission. Pdyn^Cre-derived spinal neurons can be subdivided according to transient vs persistent expression of dynorphin, which includes excitatory neurons (∼12%) and inhibitory neurons (∼88%). In the ablation study by Duan et al.,¹² an intersectional genetic strategy was used to capture Dyn⁺ inhibitory interneurons which included neurons that expressed Dyn peptide either transiently or persistently, but not Dyn⁺ excitatory neurons and galanin-positive inhibitory neurons. However, when Pdyn^Cre neurons were captured in adult animals by virus-based labeling techniques, persistent Dyn⁺ neurons, including both excitatory and inhibitory neurons, were targeted,^20,53 excluding the transiently expressing Dyn⁺ neurons. Because of these technical differences, the neurons that were ablated using the intersectional genetic strategy only partially overlapped with the neurons captured in these virus-based studies. But consistent with the ablation study, activation or silence of persistent Dyn⁺ neurons developed mechanical hypersensitivity in mice.^20,53 Thus, these studies suggest that different subsets of Pdyn^Cre-derived neurons may be required to gate pain and itch.

Previous studies have shown that the selective activation of glycinergic neurons in the dorsal horn, spinal transplantation of GABAergic precursors, or spinal administration of GABA receptor agonists is antipruritic,^5,6,14 suggesting GABA and/or glycine that is released from Sst2A⁺ interneurons act as the major inhibitory neurotransmitter to suppress itch. In addition, Sst2A is also expressed in CGRP⁺/TRPV1⁺ peptidergic nociceptors and has an antiallodynic effect in a mouse model for neuropathic pain.⁴³ Our results do not exclude the possibility of peripheral Sst2A⁺ neurons functioning as a component of itch neural circuit. Neuropeptide SOM may target peripheral Sst2A⁺ nociceptor to trigger itch (Fig. 6) through suppressing pain.^25,30

Apart from spinal excitatory interneurons, SOM is also expressed in a small population of primary sensory neurons.¹⁸ Hence our study does not rule out the potential role of peripheral SOM in itch processing. SOM released from primary sensory neurons may bind to Sst2A⁺ neurons in both DRG and spinal cord to disinhibit itch transmission as well (Fig. 6). Recently, Huang et al.¹⁹ reported that SOM is expressed in NP3 itch neurons. Deletion of SOM peptide encoding Sst gene in the spinal cord and DRG attenuated pruritogen-induced itch responses, whereas deletion of Sst either in the spinal cord or the DRG alone did not influence itch transmission.¹⁹ After the deletion of Sst gene either in the spinal cord or the DRG, SOM peptide released from primary afferents of peripheral Nppb⁺ sensory neurons or spinal SOM⁺ excitatory interneurons might be sufficient to regulate the spinal itch circuitry. Thus, SOM released by the DRG neurons and the spinal neurons synergically facilitates itch sensation. Albeit, as shown by Huang et al., spinal deletion of SOM gene did not have any effect on itch transmission, but our data show that ablation of spinal SOM^Lbx1 neurons impedes itch transmission. The plausible explanation of the observed disparity is because of the adopted technical approaches. In this study, the SOM^Lbx1 neurons were ablated, whereas Huang et al. deleted the Sst gene in the spinal cord and/or periphery. In the case of ablation of spinal SOM^Lbx1 neurons, the observed impaired itch transmission is due to nonexecutable disinhibition of itch circuit by SOM peptide given that the cellular substrate (SOM^Lbx1 neurons), which primarily received the tonic inhibition, is not present in the circuit to be acted upon and to further transmit the itch signal.

In summary, we identify multiple partially overlapped pathways in the spinal cord that convey pruritic information. Importantly, we demonstrate that neuropeptide SOM acts as an itch modulator in the spinal cord through a disinhibition mechanism. Altogether, we propose that specific and effective disruption of this disinhibition circuit might offer an alternative therapeutic approach for itch treatment.

Conflict of interest statement

The authors have no conflict of interest to declare.