Mice with chronic constriction injury (CCI) of the left sciatic nerve showed the expected hypersensitivity reflected in the decrease of mechanical pain thresholds measured in the ipsilateral hind paw after 16 days (Figure 1a and b). We examined mGlu5 receptor signaling in regions of the contralateral pain matrix using an in vivo method for the assessment of mGlu5-receptor-mediated PI hydrolysis (Zuena et al., 2018). Mice were pretreated with lithium ions and then challenged with a selective mGlu5 receptor-positive allosteric modulator (PAM, VU0360172, 30 mg/kg, i.p.). mGlu5-mediated PI hydrolysis (i.e. the extent of VU0360172-stimulated InsP formation) was amplified in the contralateral infralimbic cortex, prelimbic cortex, anterior cingulate cortex, and basolateral amygdala (BLA), but not in the contralateral ventrobasal thalamus (VPL), of CCI mice compared to the corresponding regions of sham-operated mice (Figure 1c–g). mGlu5 receptor protein levels were up-regulated in the contralateral infralimbic cortex, but not in other brain regions, of CCI mice (Figure 1—figure supplement 1). The findings suggest that mGlu5 receptors become hyperactive at least in four regions of the pain matrix after induction of neuropathic pain.

We applied photopharmacology to study brain region-specific functions and pharmacological manipulations of mGlu5 receptors by implanting optic fibers in the contralateral infralimbic cortex, prelimbic cortex, anterior cingulate cortex, amygdala (BLA), and thalamus (VPL) of CCI and sham mice 12 days after surgery (Figure 2a and b). Four days later, mice received a systemic injection of JF-NP-26 (10 mg/kg, i.p.), alloswitch-1 (10 mg/kg, i.p.), or their respective vehicle (Figure 2b, i and o). JF-NP-26 is an inactive pro-drug of the mGlu5 receptor NAM raseglurant and can be activated by blue-violet light (405 nm) (Font et al., 2017). The trans-isomer of alloswitch-1 is a systemically active mGlu5 receptor NAM, which can be isomerized into inactive cis-alloswitch-1 by blue-violet light (405 nm) and converted back into the active trans-isomer by green light (520 nm) (Pittolo et al., 2014; Figure 2c).

Light-induced activation of JF-NP-26 in the infralimbic cortex, prelimbic cortex, anterior cingulate cortex, or thalamus (VPL) contralateral to CCI caused rapid and robust antinociception in neuropathic mice measured as increased mechanical thresholds, whereas, unexpectedly, JF-NP-26 activation in the amygdala (BLA) further reduced mechanical pain thresholds, indicating pronociceptive effects (Figure 2d–h). JF-NP-26 (10 m/kg, i.p.) was given systemically 25 min before light-induced activation (Figure 2i). Interestingly, light-induced activation of JF-NP-26 in the prelimbic cortex contralateral to CCI also enhanced the mechanical pain threshold in the uninjured hind limb (ipsilateral to light delivery), whereas hyperalgesia of the uninjured limb was induced by JF-NP-26 activation in the BLA contralateral to CCI (Figure 2—figure supplement 1). No changes in mechanical thresholds were induced by the activation of JF-NP-26 in any brain region of sham-operated mice (Figure 2—figure supplement 2).

The data suggest that mGlu5 receptor blockade in medial prefrontal cortical regions (infralimbic, prelimbic, or anterior cingulate cortices) or thalamus, but not amygdala, is sufficient for antinociceptive effects in neuropathic pain. Knowing that mGlu5 receptors in the BLA shape susceptibility to stress and fear in rodents (Shallcross et al., 2021; Kim et al., 2023), we also measured depression-like and risk-taking behavior after light-induced activation of JF-NP-26 in the BLA of neuropathic mice. Light-induced activation of JF-NP-26 decreased risk-taking hence increased anxiety-like behavior in CCI mice as shown by the decreased number of entries into, and reduced time spent in, the light compartment of the light-dark box (Figure 2—figure supplement 3a–c). Depression-like behavior assessed with the tail-suspension test was unchanged in CCI mice after light-induced irradiation of JF-NP-26 in the BLA (Figure 2—figure supplement 3d).

We used optical inactivation and reactivation of alloswitch-1 to determine the brain region(s) necessary for analgesia caused by systemic blockade of mGlu5 receptors with alloswitch-1. Systemic injection of alloswitch-1 (10 mg/kg, i.p.) consistently increased mechanical pain thresholds of CCI mice, indicating antinociceptive effects. Blue-violet-light-induced inactivation of alloswitch-1 in the contralateral infralimbic cortex, prelimbic cortex, and thalamus (VPL) abolished alloswitch-1-induced antinociception, which was restored by green-light induced reactivation of alloswitch-1 (Figure 2j, k and n). In contrast, optical inactivation of alloswitch-1 in the amygdala (BLA) further increased the antinociceptive effects of systemic alloswitch-1, and this effect was reversed by reactivation of alloswitch-1 (Figure 2m). Optical modulation of alloswitch-1 in the anterior cingulate cortex had no effect on pain thresholds (Figure 2l). Alloswitch-1 was given systemically 25 min before light-induced inactivation (Figure 2o). The data suggest that mGlu5 receptor blockade in infralimbic and prelimbic cortices and thalamus, but not anterior cingulate cortex, is necessary for antinociceptive effects in neuropathic pain conditions, whereas mGlu5 blockade in the BLA has paradoxical pain-facilitating effects.

To understand the neural basis for the antinociceptive behavioral effects of optical modulation of alloswitch-1 in the medial prefrontal cortex, we conducted whole-cell patch-clamp electrophysiological recordings from pyramidal neurons in layer 5 of the prelimbic cortex in brain slices from CCI mice 16 days post-injury. Excitatory and inhibitory postsynaptic currents (EPSCs and IPSCs) were evoked in visually identified pyramidal neurons by stimulation of presumed afferents from BLA that traverse the infralimbic cortex in layer 4 (Figure 3a–c), as described previously (Ji et al., 2010). EPSCs and IPSCs were blocked by bath application of a glutamate receptor antagonist (CNQX, 20 µM) confirming glutamatergic EPSCs and glutamate-driven IPSCs that are part of a feedforward inhibitory circuitry from the BLA to regulate prefrontal cortical output (Neugebauer et al., 2023; Ji et al., 2010) (IPSCs were additionally blocked by bicuculline, not shown). Synaptic responses were measured before (in artificial cerebrospinal fluid, ASCF) and during the application of alloswitch-1 to the brain slice, and during blue-violet and green light illumination to inactivate and reactivate, respectively, alloswitch-1 (Figure 3d and e).

Administration of mGlu5 NAM alloswitch-1 (100 nM) increased EPSCs compared to predrug ACSF values. Alloswitch-1 inactivation with blue-violet light (5 min, 0.5 Hz, 500 ms) slightly decreased EPSCs so that they were not significantly different from baseline values. Reactivation of alloswitch-1 with green light (5 min, 0.5 Hz, 500 ms) led to a significant increase in EPSCs compared to predrug (Figure 3d and f). No significant effects were observed on IPSCs (Figure 3e and g). As a result, the E/I ratio showed a pattern that reflected the changes in EPSCs with increases by alloswitch-1 and by reactivation with green light, and no change from baseline during blue-violet light application to inactivate alloswitch-1 (Figure 3h). These results suggest that mGlu5 receptor controls the excitatory drive of prelimbic output neurons, possibly by activation of synaptic inhibition of excitatory inputs; and mGlu5 receptor inhibition restores the excitatory drive-in neuropathic pain.

To explore if the antinociceptive behavioral effects of mGlu5 receptor inhibition in the prelimbic cortex could be explained by engaging descending pain control systems as a consequence of increased excitatory drive and output, we performed in vivo single-unit recordings of so-called ON-cells in the rostroventromedial medulla (RVM). RVM ON-cells are a critical element of top-down control of pain modulation and serve pronociceptive functions (Chen and Heinricher, 2022). Systemic administration of alloswitch-1 (10 mg/kg, i.p.) caused a trend towards decreased excitation of ON cells. Optical inactivation of alloswitch-1 in the prelimbic cortex increased ON-cell excitation whereas reactivation of alloswitch-1 with green light decreased excitation (Figure 3—figure supplement 1). The effect size of optical modulation in the prelimbic cortex on RVM activity, according to Cohen’s d calculation from t-tests, is shown in the Table 1 (see Supplementary file 1).

These results suggest that increased prelimbic cortical output with mGlu receptor inhibition (brain slice physiology data) can decrease excitation of pronociceptive RVM ON-cells resulting in antinociceptive behavioral effects (see Figure 3—figure supplement 1). These findings highlight the importance of prefrontal cortical-RVM connectivity in mediating the pain-relieving effects associated with mGlu5 receptor inhibition in the prelimbic cortex.

We initially hypothesized that the neural basis for the pronociceptive behavioral effects of optical modulation of alloswitch-1 in the BLA could be the decrease of excitatory drive from BLA to the prefrontal cortex. Our previous work showed increased feedforward inhibition of prefrontal cortex due to increased activity in BLA in pain, and therefore, restoring prefrontal cortical output was able to inhibit pain behaviors (Neugebauer et al., 2023). We conducted in vivo electrophysiological recordings of individual pyramidal neurons in the prelimbic cortex in anesthetized mice 16 days after CCI induction. Systemically applied alloswitch-1 (10 mg/kg, i.p.) increased neuronal activity, but inactivation of alloswitch-1 with blue-violet light illumination in the BLA had no significant effect, arguing against a contribution of mGlu5 in the BLA. Furthermore, reactivation of alloswitch-1 with green light illumination in the BLA did not mimic the effect of systemic alloswitch-1 (Figure 3—figure supplement 1). Therefore, we explored another BLA output and tested the hypothesis that inhibition of mGlu5 in the BLA would decrease feedforward inhibition of neurons in the central nucleus (CeA) and hence increase their activity (Figure 3i), which has been shown to correlate positively with pain behaviors through projections to several brain areas involved in behavioral modulation (Neugebauer, 2020).

To do so we performed whole-cell patch-clamp recordings of neurons in the laterocapsular division of the central nucleus of the amygdala (CeLC) in brain slices obtained from neuropathic mice 16 days after CCI induction (Figure 3j and k). CeA neurons receive direct excitatory and feed-forward inhibitory inputs from the BLA (Neugebauer, 2020). Glutamatergic-driven IPSCs and monosynaptic EPSCs were evoked in CeLC neurons by focal electrical stimulation in the BLA as described previously (Ren et al., 2013). Synaptic responses were evaluated before (ACSF) and during alloswitch-1 bath application, and under blue-violet or green light illumination to block and reinstate, respectively, the pharmacological activity of alloswitch-1 (Figure 3l and m). Both EPSCs and IPSCs were blocked by a glutamate receptor antagonist (CNQX, 20 µM), confirming glutamatergic EPSCs and glutamate-driven IPSCs (IPSCs were additionally blocked by bicuculline, not shown). Alloswitch-1 (100 nM) had significant inhibitory effects on IPSCs at the BLA-CeLC synapse (Figure 3l and n), which was reversed by inactivation with blue-violet light (5 min, 0.5 Hz, 500 ms) so that IPSCs were not significantly different from baseline. Reactivation with green light (5 min, 0.5 Hz, 500 ms) decreased IPSCs. Alloswitch-1 by itself and its inactivation and reactivation had no significant effects on EPSCs (Figure 3m and o). Accordingly, alloswitch-1 decreased the inhibition/excitation (I/E) ratio significantly; this effect was reversed with blue-violet light for optical inactivation and was reinstated with green light for reactivation (Figure 3p). The data suggest that mGlu5 blockade reduces BLA-driven feed-forward inhibition onto CeLC neurons that have been linked mechanistically to pain behaviors, which could explain the hypersensitivity observed with photopharmacological blockade of mGlu5 receptors.

The aim of this study was to determine the region-specific behavioral and electrophysiological effects of mGlu5 receptor blockade with caged and photoswitchable NAMs in neuropathic pain using photopharmacology. While the effect of optical manipulations of photosensitive mGlu5 NAMs in individual brain regions in pain models has been studied before (Font et al., 2017; Notartomaso et al., 2022; Gómez-Santacana et al., 2017), the comparative analysis across different brain regions and their contribution to systemically applied NAMs represents a novel concept.

JF-NP-26 (10 mg/kg, i.p. in 6% DMSO, 6% Tween 80 in saline) and alloswitch-1 (10 mg/kg, i.p. in 20% DMSO, 20% Tween 80 in saline) were kindly provided by Professor Llebaria (IQAC-CSIC, Barcelona, Spain). Lithium chloride was purchased from Sigma-Aldrich (Milan, Italy). VU0360172 [N-cyclobutyl-6-(2-(3-fluorophenyl)ethynyl) pyridine-3-carboxamide] was purchased from Tocris Bioscience (Bristol, United Kingdom).

All experiments were carried out according to the European (2010/63/EU) and Italian (D. Lgs. 26/2014) guidelines for animal care. The Italian Ministry of Health (n° 804/2018-PR) approved the experimental protocol. Experiments carried out at Texas Tech University Health Sciences Center (TTUHSC) were approved by the Institutional Animal Care and Use Committee (IACUC) at TTUHSC (14006) and conform to the guidelines of the International Association for the Study of Pain (IASP) and National Institutes of Health (NIH). Every effort was made to minimize animal suffering and the number of animals used. Adult male C57BL/6 J mice (20–25 g, b.w.) were purchased from Charles River (Calco, Italy) for the behavioral assays and in vivo electrophysiological recordings, while wild-type C57BL/6 J mice (original breeding pairs from Jackson lab) were used for brain slice electrophysiology experiments. All animals were housed in groups of four in standard cages with ad-libitum access to food and water and maintained under 12 hr dark/light cycle (starting at 7:30 AM), 22 °C temperature, and 66% humidity (standard conditions). Sample size was based on our previous studies and experience with similar behavioral, electrophysiological, and molecular assays. Animals were randomly assigned to different experimental groups and a researcher blinded to drug treatments carried out all animal experimentation. All the experiments were conducted by at least two investigators. Some of the electrophysiological experiments represent replicates.

Induction of the sciatic nerve CCI model (Bennett and Xie, 1988) and assessment of mechanical hypersensitivity are described in more detail in Bio-protocol (Notartomaso et al., 2018). In brief, CCI induction was carried out under isoflurane anesthesia (5% for induction and 2% for maintenance), using a modified version of a previously described method (Bennett and Xie, 1988). The biceps femoris and the gluteus superficialis were separated by blunt dissection, and the left sciatic nerve was exposed. CCI was produced by tying two ligatures (6–0 silk, Ethicon, LLC, San Lorenzo, PR, USA) around the sciatic nerve. The ligature was tied loosely around the nerve, until it elicited a brief twitch in the respective hind limb. Over-tightening of the ligation was avoided to preserve epineural circulation. The incision was cleaned and the skin was closed with 2–3 ligatures of 5–0 dermalon. In sham-operated mice, the left sciatic nerve was exposed without ligature. Mechanical allodynia was assessed 16 days after surgery by measuring the hind paw withdrawal response to von Frey filament stimulation. Mice were placed in a dark box (20 × 20 × 40 cm) with a wire grid bottom through which the von Frey filaments (North Coast Medical, Inc, San Jose, CA, USA; bending force ranging from 0.008 to 3.5 g) were applied by using a modified version of the up-down paradigm previously described (Chaplan et al., 1994). Each filament was applied five times (3 min intervals) perpendicularly to the plantar surface of the hind paw until it bent. The filament that evoked at least three paw withdrawals was assigned as the pain threshold in grams.

Mice were anesthetized with isoflurane (5% for induction and 2% for maintenance) and implanted with monolateral LED optic fibers (MFC_400/430–0.57_mm_SMR_C45 Doric Lenses Inc, Quebec, Canada) using dental cement and surgical screws (Agnthos, Lidingö, Sweden) in a stereotaxic frame (Stoelting Co., Wood Dale, IL, USA). The following brain regions in the right hemisphere were targeted (Paxinos and Franklin, 2008): prelimbic cortex (1.50 mm anterior to the bregma, 0.3 mm lateral to midline, 2.7 mm ventral from surface of the skull), infralimbic cortex (1.70 mm anterior, 0.3 mm, 3 mm), anterior cingulate cortex (0.5 mm anterior, 0.2 mm, 2 mm), basolateral amygdala (1.34 mm posterior, 2.9 mm, 4.7), ventral posterolateral thalamus (1.8 mm posterior, 1.5 mm, 3.5 mm). The length (mm) of the optic fibers was selected according to the targeted brain area. Previous studies from our group showed that stereotaxic insertion of optical fibers into the infralimbic cortex leaves cortical connectivity intact (Ji and Neugebauer, 2012).

Mice were treated with JF-NP-26, (10 mg/kg, i.p. in 6% DMSO, 6% Tween 80 in saline) or alloswitch-1 (10 mg/kg, i.p. in 20% DMSO, 20% Tween 80 in saline). Individual brain areas (infralimbic cortex, prelimbic cortex, anterior cingulate cortex, ventrobasal thalamus, and basolateral amygdala contralateral to the side of injury) were illuminated through the implanted optic fiber with conical tip (Doric Lenses Inc) with blue-violet light (405±6.5 nm) to activate JF-NP-26 or inactivate alloswitch-1 and with green light (520±19 nm) to reactivate alloswitch-1 (Pittolo et al., 2014) (see Figure 2c). The selection of wavelengths was based on our previous work (Pittolo et al., 2014; Ricart-Ortega et al., 2020). Pulsed light illumination (or dark control) was for 5 min (0.5 Hz, 500ms), 20 min after drug administration of the caged or photoswitchable compound. The light spread with a conical tip cannula is less than 200 µm according to the manufacturer (Doric Lenses Inc).

The light-dark box consisted of two compartments: an illuminated compartment and a dark compartment (50 × 20 cm) connected by a shuttle door (12 × 5 cm) in the center of the partition at floor level. The light box was open at the top, painted white, and illuminated by a 60 W light bulb located 30 cm above the apparatus, providing illumination in the range of 380–470 lux measured on the floor of the white compartment. The dark box was painted matte black and had a removable black lid at the top. Each animal was placed at the center of the dark compartment and was allowed to explore freely both compartments for 10 min. Total number of crossings between the two compartments (defined as all four paws being within a given compartment), latency to enter the illuminated compartment, and time spent in the illuminated compartment were measured. Movement of each animal was recorded using a video camera and analyzed by a blinded operator. Neuropathic mice were treated with JF-NP-26 (10 mg/kg, i.p. in 6% DMSO, 6% Tween 80 in saline) with or without violet light delivered into the BLA. In the TST, each mouse was suspended by its tail using adhesive tape and attached to a hook. The duration of immobility during the 5 min test period was measured.

Tissues were dissected out and homogenized at 4 °C in RIPA buffer containing protease inhibitors cocktail (Merck Millipore, Milano, Italy) for 30  min, and an aliquot was used for protein determination. Equal amounts of proteins (20  μg) from supernatants were separated by 8% SDS polyacrilamide gel at 100  V for 1  hr for the detection of mGlu5 receptors, using a mini-gel apparatus (Bio-Rad Mini Protean II cell, Milano, Italy). Proteins were then electroblotted on Immuno PVDF membranes (Bio-Rad) for 7  min using a Trans Blot Turbo System (Bio-Rad). Filters were washed three times and blocked for 1  hr in Tris-Tween buffered saline (TTBS) containing 5% non-fat dry milk. The following primary antibodies were used: rabbit monoclonal anti-mGlu5 receptor antibody (1:200, Abcam, Cambridge, UK, Cat #AB76316, lot. GR45647-18). Filters were washed three times with TTBS buffer and then incubated for 1  hr with secondary peroxidase-coupled anti-rabbit antibody (1:7000, Millipore, Cat #401393–2  ML). Immunostaining was revealed by enhanced chemiluminescence luminosity (Amersham Pharmacia Biotech, Arlington Height, IL). The blots were re-probed with monoclonal anti-β-actin antibody (1:50000, Sigma-Aldrich, Cat # A5441, lot. 116M4801V). Statements of validation and references are on the manufacturers' websites; antibodies were validated in our previous studies, including with knock-out mice (PMID: 30190524). ImageLab 6.1 software was used for the acquisition and analysis of the Western blot image data.

For the assessment of mGlu5 receptor signaling, mice were injected systemically with lithium chloride (10 mM, 100 μl, i.p.) followed, after 1 hr, by i.p. injection of a positive allosteric modulator for mGlu5 (VU0360172, 30 mg/kg i.p.) to activate mGlu5 receptors. Control mice were treated with lithium chloride followed by the vehicle. Mice were sacrificed 60 min after the last injection and the brain regions were microdissected with a Vibratome (Leica Biosystems, Buccinasco, MI, Italy). The selected brain regions were weighed and homogenized by sonication in 10 μl/mg of tissue of Tris-HCl buffer (100 mM; pH 7.5) containing 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 1% SDS. Homogenates were diluted 1:50 and InsP levels were assessed with the IP-One ELISA kit (Cisbio, Codolet, France) according to the manufacturer’s instructions. Microplate manager 6 software (MPM6 Biorad) was used for plate reading.

Whole-cell patch-clamp recordings were performed in brain slices containing the right medial prefrontal cortex or amygdala obtained from CCI mice (16 days after surgery) as described previously (Ji et al., 2010; Ren et al., 2013; Yakhnitsa et al., 2022). Brains were quickly removed and immersed in an oxygenated ice-cold sucrose-based physiological solution containing (in mM): 87 NaCl, 75 sucrose, 25 glucose, 5 KCl, 21 MgCl₂, 0.5 CaCl_2, and 1.25 NaH₂PO₄. Coronal brain slices of 300 μm thickness were obtained using a Vibratome (VT1200S, Leica Biosystems VT Series, Nussloch, Germany) and incubated in oxygenated artificial cerebrospinal fluid (ACSF, in mM: 117 NaCl, 4.7 KCl, 1.2 NaH₂PO₄, 2.5 CaCl₂, 1.2 MgCl₂, 25 NaHCO₃ and 11 glucose) at room temperature (21 °C) for at least 1 hr before patch recordings. A single brain slice was transferred to the recording chamber and submerged in ACSF (31 ± 1°C) superfusing the slice at ~2 ml/min. Only one or two brain slices per animal were used and only one neuron was recorded in each brain slice. Whole-cell patch-clamp recordings were made from visually identified prelimbic or central amygdala (laterocapsular division) neurons using DIC-IR videomicroscopy as described previously (Ji et al., 2010; Ren et al., 2013). Recording electrodes (tip resistance 5–8 MΩ) were made from borosilicate glass and filled with intracellular solution containing (in mM): 122 K-gluconate, 5 NaCl, 0.3 CaCl₂, 2 MgCl₂, 1 EGTA, 10 HEPES, 5 Na₂-ATP, and 0.4 Na₃-GTP; pH was adjusted to 7.2–7.3 with KOH and osmolarity to 280 mOsm/kg with sucrose. In some experiments, 0.2% biocytin was included in the intracellular solution on the day of the experiment. Data acquisition and analysis were done using a dual 4-pole Bessel filter (Warner Instr., Hamden, CT), low-noise Digidata 1322 or 1550B interface (Axon Instr., Molecular Devices, Sunnyvale, CA), Axoclamp-2B or MulitClamp700B amplifier (Axon Instr., Molecular Devices, Sunnyvale, CA), and pClamp9 or 11 software (Axon Instr.). If series resistance (monitored with pClamp9 or 11 software) changed by more than 20%, the neuron was discarded. To characterize the properties of recorded neurons, depolarizing current pulses (500 ms, 25 pA step) were applied in current clamp mode. Neurons were voltage-clamp at –70 mV or 0 mV to study excitatory and inhibitory postsynaptic currents (EPSCs and IPSCs) evoked by the electrical stimulation of presumed afferents from basolateral amygdala (BLA) using a concentric bipolar stimulating electrode (David Kopf Instruments). Monosynaptic EPSCs and glutamate-driven IPSCs were completely blocked by the application of a non-NMDA receptor antagonist (6-cyano-7-nitroquinoxaline-2,3-dione, CNQX, 20 μM). IPSCs were blocked by bicuculline (10 μM). Alloswitch-1 was applied to the brain slice by gravity-driven superfusion in ACSF (final concentration 100 nM Pittolo et al., 2014, 2 ml/min), and electrophysiological measurements were made 10–15 min during drug application. Illumination of the target region in the slice was done through the objective of the microscope (BX51, Olympus, Waltham, MA) with an integrated LED stimulation system (X-Cite Xylis). Photoisomerization of alloswitch-1 was achieved by illumination with blue-violet light (375±14 nm) generated by a broad-spectrum LED illumination system by passing through an AT375/28 x filter (Chroma Technology Corp., Bellows Falls, VT) for 5 min (0.5 Hz, 500 ms), and was reversed by illumination with green light (545±15 nm) an AT545/30 x filter (Chroma Technology Corp., Bellows Falls, VT) for 5 min (0.5 Hz, 500 ms). The electrophysiological outcome measures were assessed after 5 min and during (blue-violet or green) light application.

Single-unit electrophysiological recordings were performed in the prelimbic cortex or RVM in CCI mice 16 days after injury. An LED optic fiber was implanted into the BLA when recording from the prelimbic area, or into the prelimbic cortex when recording from RVM. Mice were anaesthetized with an intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg) and anesthesia was maintained with a constant continuous infusion of propofol (5–10 mg/kg/h, i.v.) throughout the experiment. A small craniotomy allowed the stereotaxic insertion of a glass-insulated tungsten filament recording electrode (3–5 MΩ) (Frederick Haer & Co., Bowdoin, ME, USA) that was lowered into the prelimbic cortex (1.50 mm anterior to the bregma, 0.3 mm lateral to midline, 1.5–2.7 mm ventral from surface of the skull) or RVM (6.48 mm posterior to bregma, 0.3–0.5 mm lateral, and 4.5–6 mm ventral) as previously described (Guida et al., 2018; Boccella et al., 2019). Spike waveforms were displayed on an oscilloscope to ensure that the unit under study was unambiguously discriminated throughout the recording period. Signals were processed using an interface (CED 1401; Cambridge Electronic Design Ltd., Cambridge, United Kingdom) connected to a Pentium III PC. Spike2 software (CED, v.4) was used to create peristimulus rate histograms on-line and to store and analyze digital records of single-unit activity off-line. In the RVM, we classified ON- and OFF-cells based on their firing patterns relative to nocifensive withdrawal responses (Fields and Heinricher, 1985). As ON-cells with a high basal firing rate can be easily misclassified as NEUTRAL-cells (Barbaro et al., 1986), potential NEUTRAL-cells with continuous spontaneous activity were verified by giving a brief bolus of anesthetic to the point that the withdrawal reflex was abolished. Indeed, the firing of spontaneously active ON-cells slows or stops with this manipulation, which unmasks reflex-related responses. Recordings were performed only from ON-cells, which have response characteristics consistent with a role in pain facilitation and accelerate firing immediately before the nocifensive reflex (Chen and Heinricher, 2022).

Specifically, cells were identified by a burst of activity in response to a brief (3 s) noxious mechanical stimulus (von Frey filament with bending force of 97.8 mN) applied to the hind paw contralateral to the RVM. RVM ON-cell burst activity (Hz) evoked by peripheral noxious stimulation was calculated as number of spikes in the 10 s interval starting from the beginning of the increase in firing frequency (which was at least double its spontaneous activity). In the prelimbic cortex, regular-spiking pyramidal cells were studied. Background activity (Hz) refers to neuronal activity in the absence of intentional stimulation. Neuronal activity (frequency of excitation of RVM neurons or background activity of prelimbic neurons) was measured every 5 min before and after systemic administration of alloswitch-1 and before and after optical manipulation, and was calculated as mean ± SEM.

Statistical analysis was performed with GraphPad Prism version 9.5.1 for Windows (GraphPad Software, San Diego, California USA). Standard two-tailed unpaired or paired student t-test and one or two-way ANOVA followed by Bonferroni or Dunnett’s posthoc tests were used as appropriate. A value of P<0.05 was considered statistically significant. All data are presented in the Figures. No data from experiments presented in this study were excluded, and all data represent biological replicates.