All animal experiments were approved by the Institutional Animal Care and Use Committee in Jilin University and complied with relevant ethical guidelines. C57BL/6 male mice weighing 18–20 g were purchased from the Animal Centre of Jilin University. All mice were housed at 25°C in a humidity-controlled room on a 12 h light–dark cycle with ad libitum to food and water. Experiments were conducted on mice aged between 8 and 12 weeks old. Animals were randomly assigned to a treatment group and were allowed to acclimatize to experimental conditions for 3 days before the experiments.

Our study used H-151 (10 nM, MedChemExpress, HY-112693) and 7-BIA (10 or 20 mg/kg, MedChemExpress, HY-115496). Both reagents (H-151, 7-BIA) were diluted in 10% DMSO and 90% corn oil. H-151 was administered via intrathecal injection. All intrathecal injections were performed under brief anesthesia induced by isoflurane. Briefly, a small area was shaved on the back of the animal and a spinal puncture was carried out at L4–L5 using a microsyringe needle to deliver a maximum of 10 μl reagents into the subarachnoid space. A tail movement confirmed a successful intrathecal injection. 7-BIA was administrated by intraperitoneal injection.

The CCI neuropathic pain model was induced in 8- to 10-week-old male C57 mice under isoflurane anesthesia, as previously described (Bennett and Xie, 1988; Gopalsamy et al., 2019). Briefly, the right thigh was shaved and sterilized with iodophor. A transverse incision was made under the long head of the biceps femoris and the sciatic nerve was exposed via blunt dissection. Under a microscope, three ligatures (8-0 Prolene) were placed around the sciatic nerve 1 mm apart from proximal to the trifurcation. Ligatures were loosely tied to prevent an arrest of the epineural blood flow. In the sham group, the same procedure was performed without ligation of the sciatic nerve.

All sensory behavioral testing in mice was performed between 9 am and 6 pm in an isolated experimental room maintained at 21–25°C. Mice were placed in the experimental room at least 3 day prior to baseline behavioral testing and all tests were performed by the same experimenter who was blinded to treatment. To assay mechanical sensitivity, the paw withdrawal threshold (PWT) of the hind paw was recorded using a series of von Frey filaments (from 0.16 to 2.0 g, Aesthesio, UGO) by perpendicular stimulation of the central plantar surface of the paw at resting state. A positive response was recorded if the paw was sharply withdrawn. To test the 50% PWT, we used the up–down method, as previously described (Michaelidou et al., 2013). Typically, 0.16 g filaments do not elicit paw withdrawal in mice and were used to test the paw withdrawal frequency (PWF). Mice were repeatedly stimulated 10 times, in 1 min intervals (Aesthesio, UGO).

To assay temperature sensitivity, thermal paw withdrawal latency (PWL) was assessed using a hot plate (BIO-CHP-ER, Bioseb). Briefly, the mice were placed on a 52°C metallic plate surrounded by an acrylic container. A sensory response was recorded by flinching, licking one of the hind paws, or jumping. The test was stopped if the paw was not withdrawn within 20 s. The hot plate test was repeated three times for each mouse, with at least 20 min intervals between each repetition.

Motor coordination was evaluated by rotarod testing (Panlab, LE8505), as previously described (Bohlen et al., 2009). Rotarods were accelerated from 4 to 40 rpm over 300 s. Over the course of 3 days, each mouse underwent three trials per day separated by 15 min intervals. The fall latency was recorded on the third day.

Two different target shRNAs were designed using BLOCK-iT™ RNAi Designer and cloned into the lentiviral EGFP-expressing PLL3.7 plasmid. The shRNA sequences were as follows: shPTPRD1, 5′-GGTTCAGATGACTCCGGTTAC-3′; shPTPRD2, 5′-GGTTGAAAGCAAATGATAA-3′. For lentiviral packaging, plasmids containing shRNAs (including shRNAs against PTPRD and scrambled shRNA) were transfected into the packaging cell line 293T with the packaging plasmid psPAX2 (#12260, Addgene) and the envelop plasmid pMD2.G (#12259, Addgene) at a ratio of 4:3:1. Viruses were harvested 48 or 72 h later by filtering cell lysates with 0.45 μm filters (Millipore) followed by ultracentrifugation for 4 h at 25,000 rpm (Beckman SW28 rotor) and resuspension in phosphate-buffered saline (PBS, 50 μl). The efficiency of the two shRNAs in 293T cells was tested using quantitative real-time PCR (qRT-PCR) and Western blotting.

DRG injections were performed as previously described (Fischer et al., 2011). L4 and L5 vertebrae were exposed by a 2 cm longitudinal incision and blunt dissection of multifidus and longissimus lumborum muscles. The processus accessorius and parts of the left processus transversus were removed to expose L4/L5 DRGs. The spinal column was fixed in a stereotaxic frame and lentiviruses (300 nL for each DRG) were injected at 30 nL/min using a 36 G NanoFil needle on a NanoFil syringe, controlled by a micropump (World Precision Instruments, Sarasota, FL, United States).

Mice were deeply anesthetized with isoflurane and perfused via injection of 20 mL PBS followed by 4% paraformaldehyde (PFA) into the ascending aorta. After perfusion, L–L5 DRGs were removed and fixed in the pre-cooled 4% PFA. DRG samples were dehydrated in a 30% sucrose solution, and 12 μm transverse sections were cut using a cryostat microtome (CM1950, Leica, Germany). After washing sections three times in PBS, tissues were permeabilized in 1% Triton X-100 at room temperature and blocked with 5% BSA in PBS containing 0.5% Tween 20 (PBS-T). Samples were incubated with the following primary antibodies overnight at 4°C: rabbit anti-PTPRD (1:200; Novus, NBP2-94767), rabbit anti-NeuN (1:100, CST, D4G40), mouse anti-Tuj1 (1:1000, Abcam, 2G10), mouse anti-S100β (1:1000, Sigma, S2532), or mouse anti-GFAP (1:200, Abmart, MB0345S). The following day, sections were washed three times in PBS-T, and primary antibodies were visualized using secondary antibodies labeled with Alexa-488 and Alexa-546 (1:800; Invitrogen). Fluorescence (DM4B, Leica, Germany) and laser scanning confocal microscopes (A1HD25, Nikon, Japan) were used to capture images.

Mice were deeply anesthetized with isoflurane and L4–L5 DRGs were dissected and placed in RIPA assay lysis buffer (C50008, Sangon Biotech). Protein concentration was measured using a bicinchoninic acid (BCA) assay (P0010, Beyotime Biotechnology, China). Proteins were separated by 10% SDS-PAGE and transferred onto PVDF membranes. PVDF membranes were blocked for 90 min at room temperature using 5% BSA in TBS-T and subsequently probed with the following primary antibodies at 4°C overnight: rabbit anti-PTPRD (1:1000, Novus, NBP2-94767), rabbit anti-STING (1:1000, CST, D2P2F), or mouse anti-GAPDH (1:1000, TransGen Biotech). The membranes were subsequently incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:1000, Beyotime) for 120 min at room temperature. Finally, images were acquired using a GS800 Densitometer Scanner. The optical density of protein bands was quantified using ImageJ, with GAPDH used as the loading control to normalize protein expression levels. Each experiment was repeated three times.

Total RNA from L4 to L5 DRGs or cells was extracted using the Eastep™ Super Total RNA Extraction Kit (LS1040, Promega), and cDNA was synthesized using the Tranc-Script One-Step cDNA Synthesis SuperMix (AT311, TransGen Biotech, China). Primers (shown in Supplementary Table 1) were synthesized by Genewiz Biotech. For PCR reactions, the TB Green™ Premix Ex Taq™ (RR420A, TaKaRa) mix was used. PCR reactions were carried out on a Real-Time PCR System (CFX96, Bio-Rad, United States), and the 2^{–Δ Δ Ct} method was used to calculate relative mRNA expression. mRNA levels were normalized to Gapdh.

Levels of inflammatory cytokines IL-6, IL-1β, TNF-α, IL-10, and IFN-α in DRG tissues were determined by enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s instructions (Anoric Bio-technology Co., Ltd., Tianjin, China). Absorbance at 450 nm was measured using a microplate reader (Thermo Fisher K3, United States). Each experiment was performed three times.

Statistical analysis was performed using SPSS 21.0 for windows (SPSS, Inc., Chicago, IL, United States). All data are presented as mean ± SEM. Behavioral data were analyzed using two-way ANOVAs followed by post hoc Tukey’s tests to compare multiple groups. Unpaired Student’s t-tests were used for comparisons between two groups. All experiments were performed at least three times independently. P-values less than 0.05 were considered statistically significant. Asterisks correspond to the following significance levels: *p < 0.05, ^**p < 0.01, ^***p < 0.001.

RNA sequencing (RNA-seq) experiments previously indicated that PTPRD is upregulated in rat DRGs 14 days after CCI (Zhou et al., 2017; Cao et al., 2019; Sun et al., 2020). To confirm these findings in mice and further investigate the role of PTPRD in neuropathic pain, we established a mouse CCI model (Figure 1A). Behavioral testing confirmed that the hot plate test PWL was decreased after CCI surgery, indicating that the induction of the neuropathic pain model was successful. Likewise, von Frey filament tests demonstrated that the PWT decreased, while the PWF increased, from day 3 to day 21 after CCI surgery (Figures 1B–D). The above behavioral results indicated a successful induction of neuropathic pain in mice following CCI.

We next evaluated which cells in the DRG expressed PTPRD by co-staining sections from 7 days post CCI with PTPRD and astrocyte marker GFAP, satellite glial cell marker S100β, or neuronal marker Tuj1. This revealed that PTPRD was exclusively expressed in DRG neurons (Figure 1E).

To investigate the temporal dynamics of PTPRD expression in more detail, we conducted immunohistochemical staining on sections collected before surgery (day 0) and on days 1, 3, 7, and 14 after CCI surgery. PTPRD exhibited only low expression in the uninjured DRG (before surgery), but expression significantly increased by day 7 and 14 after CCI (Figure 2A). We next validated our immunohistochemistry results by Western blotting. Western blotting indicated that PTPRD expression was slightly elevated by day 3 and significantly increased by days 7 and 14 after CCI (Figures 2B,C). Importantly, the expression of PTPRD in contralateral DRGs did not change and was maintained at basal levels for each timepoint we evaluated (Figures 2D,E). These results suggested a potential correlation between elevated PTPRD protein levels and the onset of neuropathic pain.

To confirm whether PTPRD participated in the development neuropathic pain following CCI, we generated lentiviruses containing one of two PTPRD-targeting shRNAs (shPTPRD-1 and shPTPRD-2) or a control shRNA (shCtrl). The silencing efficiency of the two shPTPRDs was assessed by qRT-PCR (Supplementary Figure 1A) and Western blotting (Supplementary Figures 1B,C), and shPTPRD-2 was chosen for subsequent experiments as it performed better. Viruses containing shCtrl or shPTPRD-2 (shPTPRD in short hereafter) were injected into L4-L5 DRGs. DRGs were isolated and stained 14 days after transfection, which revealed that the majority of NeuN⁺ DRG neurons were successfully transfected (Figure 3A). qRT-PCR confirmed an in vivo PTPRD mRNA silencing efficiency of 75.11 ± 2.55% (Figure 3B). The silencing ability of shPTPRD was further tested by Western blotting, which showed a 46.83 ± 4.12% decrease in PTPRD protein levels (Figures 3C,D). These results showed that the lentiviral shRNA delivery successfully knocked down RNA and protein levels of PTPRD in DRG tissues in vivo.

We next went on to assess the impact of PTPRD knockdown on development of neuropathic pain. Fourteen days before CCI, baseline performance (BL) in the von Frey and heat plate tests was recorded, after which lentiviruses containing shRNAs were administered. After injection, mice were allowed to recover for 14 days. Behavioral tests were then again conducted prior to CCI surgery (day 0) or 1, 3, 5, 7, 14, and 21 days after surgery (Figure 3E). shPTPRD did not influence mechanical or thermal sensitivity prior to CCI, with animals showing similar PWT, PWL, and PWF compared to shCtrl-injected animals 14 days after lentiviral injection, which suggested that PTPRD did not participate in modulating or inducing neuropathic pain under physiological condition. In the shCtrl group, mechanical and thermal hypersensitivity were observed as early as 1 day after CCI and lasted for 21 days, when animals were sacrificed. Conversely, PTPRD silencing significantly attenuated mechanical and thermal hypersensitivity (Figures 3F–H). Knockdown of PTPRD did not alter motor behaviors (Supplementary Figure 2A). The above results suggested that upregulation of PTPRD participated in the development of neuropathic pain following CCI in mice.

We subsequently aimed to investigate the molecular mechanism through which PTPRD contributed to the development of neuropathic pain. Previous studies suggested that inflammation and the balance of pro- and anti-inflammatory cytokines play an important role in regulating neuropathic pain in injured nerves (Xu et al., 2018; Zheng et al., 2019). We hypothesized that the amelioration of neuropathic pain following PTPRD knockdown in CCI mice was associated with changes in neuroinflammation. To test this hypothesis, we evaluated levels of multiple inflammatory cytokines 7 days after CCI using ELISA (Figure 4A). DRG levels of IL-6, IL-1β, TNF-α, and IL-10 were not significantly altered following PTPRD knockdown (Figures 4B–E). Conversely, levels of IFN-α were significantly increased (Figure 4F) compared to shCtrl DRGs. These results were confirmed by qRT-PCR (Figure 4G). Based on this finding, we aimed to determine how PTPRD was involved in the regulation of IFN-α expression in the DRG after CCI. Previous reports have described that mitochondrial antiviral signaling protein (MAVs), acid-inducible gene I (RIG-I), stimulator of interferon genes (STING), Toll-like receptor-7 (TLR-7), and Toll/IL-1 receptor domain-containing adaptor (TRIF) can promote the release of IFN-1. We therefore assessed mRNA levels of these potential upstream regulators by qRT-PCR and found that the expression of STING, but not MAVs, RIG-1, TLR-7, or TRIF, was increased after PTPRD knockdown (Figure 4H). Consistently, Western blotting also demonstrated an enhanced protein expression of STING in shPTPRD DRGs (Figures 4I,J).

To confirm whether the development of neuropathic pain caused by PTPRD was mediated via the STING-IFN-I pathway, we intrathecally administered animals with the STING-specific palmitoylation inhibitor H-151 at 7 days after CCI, when hypersensitivity had typically already improved in PTPRD knockdown mice (Figure 5A). As early as 4 h after the injection, the attenuation of hypersensitivity in shPTPRD animals was reversed and no differences in PWT, PWL, and PWF were observed between the control and shPTPRD + H-151 groups. This effect lasted for 48 h (Figures 5B–D). Therefore, H-151 administration disrupted the increased expression and secretion of IFN-α (Figures 5E,F), suggesting that PTPRD mediated neuropathic pain via the STING-IFN-I pathway in mice following CCI.

7-Butoxy illudalic acid analog is a small molecule drug recently synthesized to specifically target PTPRD as part of addiction therapy. Here, we investigated whether 7-BIA could also serve as a treatment for neuropathic pain. Seven days after CCI surgery, we intraperitoneally (i.p.) administered 7-BIA at 10 and 20 mg/kg and performed behavioral testing at 1.5, 6, 24, and 48 h after injection (Figure 6A). 7-BIA elicited a dose-dependent analgesic effect in CCI mice as early as 1.5 h after i.p. injection and lasted for 24 h, reducing both mechanical and thermal sensitivity (Figures 6B–D). Additionally, 7-BIA treatment resulted in increased DRG IFN-α levels after CCI (Figure 6E) as well as significantly increased protein levels of STING in the DRG (Figures 6F,G). Animals exhibited no changes in motor coordination (Supplementary Figure 2B). Together, these results suggested that PTPRD might be involved in the development of neuropathic pain, and its specific small molecule inhibitor 7-BIA could represent a novel and safe treatment for the clinical management of neuropathic pain without addiction risk.