Somatostatin-positive (SOM+) neurons have been proposed as one of the key populations of excitatory interneurons in the spinal dorsal horn involved in mechanical pain. However, the molecular mechanism for their role in pain modulation remains unknown. Here, we showed that the T-type calcium channel Cav3.2 was highly expressed in spinal SOM+ interneurons. Colocalization of Cacna1h (which codes for Cav3.2) and SOMtdTomato was observed in the in situ hybridization studies. Fluorescence-activated cell sorting of SOMtdTomato cells in spinal dorsal horn also proved a high expression of Cacna1h in SOM+ neurons. Behaviorally, virus-mediated knockdown of Cacna1h in spinal SOM+ neurons reduced the sensitivity to light touch and responsiveness to noxious mechanical stimuli in naïve mice. Furthermore, knockdown of Cacna1h in spinal SOM+ neurons attenuated thermal hyperalgesia and dynamic allodynia in the complete Freund’s adjuvant-induced inflammatory pain model, and reduced both dynamic and static allodynia in a neuropathic pain model of spared nerve injury. Mechanistically, a decrease in the percentage of neurons with Aβ-eEPSCs and Aβ-eAPs in superficial dorsal horn was observed after Cacna1h knockdown in spinal SOM+ neurons. Altogether, our results proved a crucial role of Cav3.2 in spinal SOM+ neurons in mechanosensation under basal conditions and in mechanical allodynia under pathological pain conditions. This work reveals a molecular basis for SOM+ neurons in transmitting mechanical pain and shows a functional role of Cav3.2 in tactile and pain processing at the level of spinal cord in addition to its well-established peripheral role.

Chronic pain causes both physical suffering and comorbid mental disorders such as depression. However, the neural circuits and molecular mechanisms underlying these maladaptive behaviors remain elusive. Here, we report a pathway from vesicular glutamate transporter3 neurons in the dorsal raphe nucleus to dopamine neurons in the ventral tegmental area (VGluT3DRN→DAVTA), of which population activity in response to innocuous mechanical stimuli and sucrose consumption, is respectively inhibited and attenuated by chronic neuropathic pain. Mechanistically, neuropathic pain dampens VGluT3DRN→DAVTA glutamatergic transmission and DAVTA neural excitability. VGluT3DRN→DAVTA activation alleviates neuropathic pain and comorbid depression-like behavior (CDB) by releasing glutamate, which subsequently promotes DA release in the nucleus accumbens medial shell (NAcMed) and produces analgesia and antidepressant effects via D2 and D1 receptors, respectively. In addition, VGluT3DRN→DAVTA inhibition produces pain-like hypersensitivity and depression-like behavior in healthy mice. These findings reveal a novel VGluT3DRN→DAVTA→D2/D1NAcMed pathway in establishing and modulating chronic pain and comorbid depressive-like behavior.

The high incidence of treatment-resistant pain calls for the urgent preclinical translation of new analgesics. Understanding the behavioral readout of pain in animals is crucial for efficacy evaluation when developing novel analgesics. Mas-related G protein-coupled receptor D-positive (Mrgprd+) and transient receptor potential vanilloid 1-positive (TRPV1+) sensory neurons are two major non-overlapping subpopulations of C-fiber nociceptors. Their activation has been reported to provoke diverse nocifensive behaviors. However, what kind of behavior reliably represents subjectively conscious pain perception needs to be revisited. Here, we generated transgenic mice in which Mrgprd+ or TRPV1+ sensory neurons specifically express channelrhodopsin-2 (ChR2). Under physiological conditions, optogenetic activation of hindpaw Mrgprd+ afferents evoked reflexive behaviors (lifting, etc.), but failed to produce aversion. In contrast, TRPV1+ afferents activation evoked marked reflexive behaviors and affective responses (licking, etc.), as well as robust aversion. Under neuropathic pain conditions induced by spared nerve injury (SNI), affective behaviors and avoidance can be elicited by Mrgprd+ afferents excitation. Mechanistically, spinal cord-lateral parabrachial nucleus (lPBN) projecting neurons in superficial layers (lamina I–IIo) were activated by TRPV1+ nociceptors in naïve conditions or by Mrgprd+ nociceptors after SNI, whereas only deep spinal cord neurons were activated by Mrgprd+ nociceptors in naïve conditions. Moreover, the excitatory inputs from Mrgprd+ afferents to neurons within inner lamina II (IIi) are partially gated under normal conditions. Altogether, we conclude that optogenetic activation of the adult Mrgprd+ nociceptors drives non-pain-like reflexive behaviors via the deep spinal cord pathway under physiological conditions and drives pain-like affective behaviors via superficial spinal cord pathway under pathological conditions. The distinct spinal pathway transmitting different forms of nocifensive behaviors provides different therapeutic targets. Moreover, this study appeals to the rational evaluation of preclinical analgesic efficacy by using comprehensive and suitable behavioral assays, as well as by assessing neural activity in the two distinct pathways.