All animal experiments were approved by the Animal Research Committees of Tokushima University (Permit No: T2020-09) and Nantong University (Permit No: S20210308-004). All animal experiments were conformed to the ethical guidelines of the International Association for the Study of Pain (Zimmermann, 1983) and ARRIVE. Male ICR mice (Charles River, Yokohama, Japan) aged 7–11 weeks were used in all the experiments. All the mice were housed in plastic cages under standard laboratory conditions (12 h dark/light cycle, temperature controlled between 23 and 24°C) and provided with water and food ad libitum. An overview of the experimental design and workflow is presented in Supplementary Figure S1.

To induce NP, PSL was performed according to a well-characterized method (Seltzer et al., 1990). Briefly, mice were deeply anesthetized with a mixture of 5.0 mg/ml Vetorphale (Meiji Seika, Tokyo, Japan), 1.0 mg/ml Domitor (Zenoag, Fukushima, Japan), 5.0 mg/ml midazolam (Sandoz, Yamagata, Japan) and diluted with distilled water for injection (79% of the total volume). All the mice were anesthetized with an intraperitoneal injection of the three mixtures (10 ml/kg). The right sciatic nerve (SCN) was exposed at the high thigh level. The dorsum of the SCN was carefully freed from the surrounding tissues, and its dorsal aspect was carefully supported with a glass rod without pressing the nerve against the underlying structures. A 3/8 curved mini-needle with 5–0 silk suture was inserted into the nerve, and 1/3–1/2 of the nerve thickness was tightly ligated. The incision was then closed using three to four skin sutures (4–0). In the sham controls, the SCN was exposed with a small incision but was not ligated before the incision was closed.

Recombinant MCP-1 (Peprotech, London, United Kingdom) and sSiglec-9 (R&D Systems, Minneapolis, MN, United States) were diluted with phosphate-buffered saline (PBS) to a concentration of 500 ng/ml for each protein. CM or MCP-1/sSiglec-9 was intravenously administered from tail vein, during day 0 to day 6 in the early phase model, day 7 to day 13 in the middle phase model, and day 14 to 20 in the late phase model (Figures 1A,D,F).

SHEDs were isolated and cultured as described previously (Sakai et al., 2012). In brief, deciduous teeth from 6- to 12-year-old individuals were collected at the Nagoya University and Tokushima University Hospital. We not only confirmed donor’s intentions, but also obtained the consent of their parents or guardians. This study was approved by the Institutional Ethical Committee of Nagoya University and Tokushima University Hospital and performed in accordance with the principles of the Declaration of Helsinki (Permit No H-73 and No: 3268 for Nagoya and Tokushima University, respectively). All participants provided written informed consent. After separating the crown and root, the dental pulp was isolated and digested in a solution of 3 mg/ml collagenase type I and 4 mg/ml dispase for 1 h at 37°C. Single-cell suspensions (3 × 10⁴ cells/ml) were plated on culture dishes in Dulbecco’s modified Eagle’s medium (DMEM, Sigma, St. Louis, MO, United States) supplemented with 10% fetal calf serum (Sigma), and then incubated at 37°C in an atmosphere of 5% CO₂ at 100% relative humidity. The human skin fibroblast line, derived from a 36-year-old individual, was obtained at passage 12 from the Health Science Research Resources Bank (Osaka, Japan). The seeding density of SHED (passage 9) and fibroblasts (passage 13) for CM preparation was 1 × 10⁵.

After several passages, SHEDs or fibroblasts at 70–80% confluency were washed twice with PBS, and the culture medium was replaced with serum-free DMEM. After a 48 h incubation, the medium was collected and centrifuged for 3 min at 440 × g. The supernatant was then collected and centrifuged for 3 min at 17,400 × g. The resulting supernatant was used as SHED-CM or fibroblast-CM (Fibro-CM) in different experiments. The protein concentration in the SHED-CM and Fibro-CM was measured using the bicinchoninic acid protein assay kit (Pierce, Rockford, IL, United States) and adjusted to 3 μg/ml with DMEM.

All the candidate mice were habituated to the testing environment 2 h daily for at least 2 days before baseline testing. Mice with stable data were chosen for the subsequent von Frey test. They were individually placed on a 5 × 5 mm metal mesh floor covered with small plastic containers and allowed approximately 2–3 h for acclimatizing to the environment before testing.

PSL-induced tactile allodynia was evaluated by the von Frey test using the up-down method according to a method described in a previous study (Chaplan et al., 1994). The investigator was blinded to the treatment groups. Mechanical hypersensitivity was measured using an electronic von Frey device (Ugo Basile S.R.L, Varese, Italy). The von Frey filament was applied to the middle of the plantar surface of the hind paw, and withdrawal responses were measured five times for each hind paw. A positive reaction was recorded if the mice exhibited a brisk paw withdrawal reaction upon stimulation. Motor function was tested using a rotarod (LE8500, Bio Research Center, Nagoya, Japan). Before the rotarod test, the mice were habituated with the apparatus. The rotating cylinder was accelerated at 4 to 40 rotations per minute and observed for 1 min. The time until the mice fell from the rod was recorded.

Fresh sciatic nerve, 1 cm length covering the ligation site on early phase day 3, or human Schwann cells were harvested and placed in a 1.5 ml RNase-free tube and homogenized with Isogen (Nippon Gene, Tokyo, Japan). Chloroform was added, followed by centrifugation at 4°C for 30 min. The aqueous phase containing RNA was transferred to a fresh tube, and RNA was precipitated with isopropyl alcohol and 75% ethanol by centrifugation. Purified total RNA was quantified using a spectrophotometer (Denovix, Wilmington, DE, United States), and RNA integrity was checked on 1.5% agarose gels. The purified total RNA was converted to cDNA by reverse transcription using M-MLV reverse transcriptase (Thermo Fisher, Carlsbad, CA, United States) and random primers (Thermo Fisher). cDNA was used as a template for qPCR with the THUNDERBIRD SYBR qPCR Mix (Toyobo, Osaka, Japan), and the amplification was performed on the StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, United States). The primers used are listed in Table 1. The fluorescence intensity of the intercalated SYBR Green was measured and normalized to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

Mice from each group were deeply anesthetized before intracardiac perfusion with PBS, followed by 4% paraformaldehyde. After that, fixed SCN, L4/L5 DRG, and L3-L4 spinal cord were collected from mice, post-fixed in 4% paraformaldehyde, and dehydrated overnight in 25% sucrose at 4°C. Frozen tissue embedded in the OCT compound (Sakura Finetek, Torrance, CA, United States) and SCN and DRG were cut longitudinally into 10 μm thick sections, and the spinal cord was cut into 30 μm thick sections and mounted on glass slides. The sections were permeabilized with 0.3% Triton X-100 in PBS for 15 min, blocked with 5% normal donkey serum at 25–27°C for 1 h and incubated overnight with the following primary antibodies: rat anti-F4/80 (1:800, ab6640, Abcam, Cambridge, United Kingdom), rabbit anti-CD206 (1:1000, ab64693, Abcam), rabbit anti-Iba1 (1:1000, Wako, Osaka, Japan), rabbit anti-GFAP (1:1000, ab7260, Abcam), mouse anti-S100β (1:500, AMAB91038, Sigma), and rabbit anti-TNF-α (1:1000, ab9739, Abcam). The sections were washed with 0.3% Triton X-100 in PBS on the following day and incubated with fluorescence-conjugated secondary antibodies (anti-rabbit Alexa Fluor 488, Thermo Fisher, Eugene, OR, United States; anti-rabbit Alexa Fluor Plus 555, Thermo Fisher; anti-rat Alexa Fluor 555, ab150166, Abcam; anti-mouse Alexa Fluor Plus 488, Thermo Fisher) at ∼25°C for 1 h. The sections were washed with 0.3% Triton X-100 in PBS and incubated with DAPI (1:500, Sigma) at room temperature for 5 min. Finally, the sections were mounted on a slide using a mounting medium and covered with a cover slip. The images of the tissues were captured using a universal fluorescence microscope (BZX700, Keyence, Osaka, Japan) and a confocal microscope (FV1200, Olympus, Tokyo, Japan). CD206/F4/80- and TNF-α/S100β-positive cells 2 mm distal to the injury site in proximal SCN and CD206/F4/80- positive cells in DRG were counted at ×400 magnification. Iba1-positive cells in spinal cord dorsal horn and ventral horn were counted at ×200 magnification. All the positive cells were counted from at least three non-overlapping sections obtained from 8 animals per group.

The mannosylated-Clodrosome Macrophage Depletion Kit (m-Clodrosome: 5 mg/ml Clodronate Disodium Salt, 18.8 mg/ml L-α-Phosphatidylcholine, 4.2 mg/ml Cholesterol, and 1 mg/ml 4-Aminophenyl-alpha-D-Mannopyranoside; Encapsula NanoSciences, Brentwood, TN, United States) was used to deplete M2 macrophages from the PSL mice. After PSL, SHED-CM was injected daily for 7 consecutive days, and from day 4 to day 6, 2 mg/kg mannosylated-Clodrosome (m-Clo) was orally administered. Control mice received mannosylated liposomes without Clodronate Disodium Salt (m-Encapsome: 18.8 mg/ml L-α-Phosphatidylcholine, 4.2 mg/ml Cholesterol, and 1 mg/ml 4-Aminophenyl-alpha-D-Mannopyranoside) in the same volume. m-Clodrosome depletes M2 macrophages, which express a mannose receptor (MR, MMR, or CD206). Ten animals per group were sacrificed for tissue collection 7 days afte PSL.

Bone marrow cells from the femurs of 8-weeks-old male mice were plated on 6 cm dishes (2.0 × 10⁶ cells per dish) and differentiated into macrophages in DMEM supplemented with 20 ng/ml macrophage colony-stimulating factor (MCSF, R&D) at 37°C in an atmosphere of 5% CO₂ for 7 days. The macrophages were then incubated with serum-free DMEM and SHED-CM for 24 h. Phase contrast images of the induced macrophages were captured using a digital microscope camera (Leica DFC290 HD, Leica, Wetzlar, Germany). The mRNA expression of M2-type cell markers or trophic factors was examined by qPCR analysis (Supplementary Figure S2). SHED-CM-induced macrophages were designated as M2 macrophages. The induced macrophages were washed twice with PBS, and the culture medium was replaced with serum-free DMEM. After 24 h of incubation, the medium was collected and centrifuged for 5 min at 440 × g. The supernatant was then collected and centrifuged for 5 min at 17,400 × g. The resulting supernatant was used as M2-CM in subsequent in vivo and in vitro experiments.

Human Schwann cells (Passage 5) were obtained from ScienCell Research Laboratories (Cat. #1700; Carlsbad, CA, United States). The cells were seeded in 6 cm dishes at a density of 3 × 10⁵ cells per dish. After incubation for 24 h in 3 ml serum-free DMEM or M2-CM with 10 ng/ml TNF-α, the cells were harvested for gene analysis.

SHED-CM-treated mouse bone marrow-derived macrophages were scraped with a cell scraper (IWAKI, Tokyo, Japan) and centrifuged at 400 × g for 5 min. The cells were washed with the FACS incubation buffer (25 μg/ml DNase I (Sigma) and 1% Bovine serum albumin (BSA; Wako) containing PBS and incubated with the CD206-PE and F4/80-FITC antibodies (Biolegend, San Diego, CA, United States) in the FACS incubation buffer. Subsequently, they were analyzed using a flow cytometer (FACS Canto, BD Biosciences, San Jose, CA, United States).

The Student’s t-test was used for statistical analysis of the mean values of the two groups, and Tukey’s multiple comparison test was used to test the difference between the means of three or more groups. All statistical analyses were performed using the statistical analysis software R (version 4.0.2; available as a free download from https://www.r-project.org/). A p-value of <0.05 was used to declare the statistical significance.

We tightly ligated 1/3 to 1/2 of the SCN on the right side of mice to induce NP. A decrease in the threshold for tactile stimuli was observed after nerve ligation. The threshold was reached at least 3 days after PSL and was maintained at the level for at least 2 weeks. Daily intravenous administration of SHED-CM, but not Fibro-CM, immediately after PSL (early phase), inhibited PSL-induced mechanical allodynia (Figures 1A,B, Supplementary Figure S3A). These antinociceptive effects of SHED-CM were detected even 3 days after the administration of SHED-CM. In the von Frey test, the threshold for the right hindpaw on day 3 was 6.73 ± 1.50 g in the SHED-CM group and 4.29 ± 1.49 g in the DMEM group. On day 7, the threshold of the SHED-CM group was 8.39 ± 0.99 g, which was significantly higher than that in the DMEM group (4.30 ± 1.78 g). Furthermore, the antinociceptive effects of SHED-CM disappeared after 7 days from the last injection. Locomotor function analyzed by Rotarod testing revealed reduced motor function in DMEM-treated PSL mice, whereas that of SHED-CM mice was equivalent to uninjured wild type mice during the treatment period (days 1–7), which was immediately decreased after the last SHED-CM injection (Figure 1C).

In previous study, it has been reported that the development of NP occurred within a week after PSL surgery when the number of Iba1⁺ activated microglia increased in spinal cord. In the subsequent maintenance phase, the number of microglia was decreased; however, that of GFAP⁺ activated astrocytes greatly increased (Tanga et al., 2004). In our experimental setting, the early phase and middle/late phase were development and maintenance phase, respectively (Supplementary Figure S4). Furthermore, to evaluate the antinociceptive activity of SHED-CM against the maintenance of PSL-induced NP, we carried out the von Frey test in middle and late phases. SHED-CM treatment exhibited significant antinociceptive effects in both the middle and the late phase model. In the middle phase, the thresholds of the SHED-CM and DMEM groups on 14 days after PSL were 8.07 ± 0.84 g and 4.82 ± 0.92 g, respectively, (Figures 1D,E). In the late phase, the thresholds on 21 days after PSL were 9.47 ± 0.74 and 5.07 ± 0.67 g, respectively (Figures 1F,G). However, during the SHED-CM treatment, the von Frey test data of the contralateral side did not show any change (Supplementary Figure S3B). These results demonstrated the potential of SHED-CM to inhibit both the development and maintenance of NP.

Immunohistochemical analysis of the ipsilateral L4/L5 DRG, 7 days after PSL, revealed that SHED-CM treatment significantly increased the number of CD206⁺ F4/80⁺ M2 macrophages compared with the DMEM control treatment (Figure 3E). The cell count analysis showed that the number of M2 cells in the SHED-CM group was significantly higher than that in the DMEM group (Figure 3F), indicating that M2 not only accumulated in the ipsilateral SCN but also in the ipsilateral DRG.

To investigate the analgesic mechanism of SHED-CM, we examined the mRNA expression profiles of genes involved in pro- and anti-inflammatory responses in the early phase PSL. Three days after PSL, expressions of inflammatory genes, TNF-α, IL-1β, and inducible nitric oxide synthase (iNOS), was increased, which was suppressed by SHED-CM treatment. The SHED-CM treatment had no effect on the expression of pan macrophage markers F4/80, but increased significantly M2-specific molecules, CD206 and arginase-1 (Arg-1). The SHED-CM treatment also elevated the expression of an array of neurotrophic and immunosuppressive factors, nerve growth factor (NGF), glial cell-derived neurotrophic factor (GDNF), and transforming growth factor-β1 (TGF-β1) (Figure 2). These results showed that the SHED-CM treatment converted the proinflammatory microenvironment of PSL to an anti-inflammatory and tissue-protective microenvironment.

Furthermore, immunohistochemical analysis revealed that the SHED-CM treatment significantly increased the number of CD206⁺ F4/80⁺ macrophages in the proximal side of the ligated SCN. However, TNF-α⁺S100β⁺ proinflammatory Schwann cells and TNF-α⁺F4/80⁺ M1 macrophages were greatly decreased by the SHED-CM treatment (Figures 3A–D; Supplementary Figures S5, S6).

Next, we examined microglial activation 7 days after PSL. As shown in Figure 4, the numbers of Iba1⁺ microglia in the L3/4 ipsilateral dorsal and ventral horn were increased compared with that in the contralateral side. They showed activated microglial morphology with a hypertrophied soma and thicker and retracted processes in the ipsilateral side, whereas that of the contralateral side seemed to be at the quiescent stage with smaller soma and ramified processes. Notably, in the SHED-CM-treated group, the number of Iba1⁺ microglia in both ipsilateral dorsal and ventral horn was reduced, and their morphologies were very similar to those on the contralateral side (Figures 4A–D). These results demonstrated that SHED-CM treatment suppressed the PSL-induced microglial activation in the spinal cord.

Next, we investigated the roles of M2 macrophages on the antinociceptive activity of SHED-CM by depleting them with m-Clo. The time course of M2 depletion in the early phase model is shown in Figure 5A. The results demonstrated that, on day 5, the antinociceptive effects of m-Clo group, but not m-Enc, was significantly decreased. On day 7, the von Frey test thresholds of the m-Clo, m-Enc, and DMEM groups were 5.49 ± 0.92, 7.90 ± 1.33, and 4.57 ± 1.00 g, respectively (Figure 5B). The thresholds of contralateral side of PSL treated with m-Clo showed no hypersensitivity (Supplementary Figure S7).

Furthermore, immunohistochemical analysis showed that m-Clo, but not m-Enc, reduced the number of CD206⁺ F4/80⁺ macrophages in SCN and DRG while increasing TNF-α⁺ S100⁺ proinflammatory Schwann cells in SCN and Iba1⁺ activated microglia in the spinal cord (Figures 6, 7). These results suggest that SHED-CM-induced M2 macrophages are crucial for the suppression of proinflammatory response and antinociceptive activity.

We next examined the biological activity of the secretion from M2 induced by SHED-CM. MCSF-treated bone marrow cells differentiated into macrophages, which were subsequently stimulated with SHED-CM for 24 h. In addition, more than 68.48% of the cells differentiated into CD206⁺ F4/80⁺ M2 macrophages following the same procedure (Figure 8A). We harvested the secretion from M2 induced by SHED-CM as M2-CM.

Schwann cells first detect nerve injury and play a critical role in the development and maintenance of NP. Under proinflammatory conditions, they express cytokines TNF-α and IL-1β, chemokine MCP-1, and transient receptor potential ankyrin 1 (TRPA1) channels, which accelerate neuroinflammation and mechanical allodynia. We treated human Schwann cells with TNF-α and M2-CM or DMEM for 24 h and analyzed the gene expression profile. The TNF-α treatment increased the expression of TRPA1, TNF-α, IL-1β, and MCP-1, whereas the M2-CM treatment strongly suppressed this upregulation (Figure 8B).

To examine the antinociceptive activity of M2-CM, we administered it to the PSL mice. We found that M2-CM, but not DMEM, prevented PSL-induced allodynia and proinflammatory responses in the SCN (Figures 9A–C). In addition. M2-CM also suppressed TRPA1 expression in Schwann cells in vivo (Figures 9D,E). The results also showed that Iba1⁺ positive cells on the ipsilateral side of L3/4 were extensively decreased in the M2-CM group compared with that in the DMEM group (Figures 9F–H). Collectively, these results suggest that SHED-CM suppressed the neuroinflammation and mechanical allodynia in part through the effect of M2.

To confirm the therapeutic effects of a set of M2 inducers, MCP-1 and sSiglec-9, in SHED-CM, were administered to the PSL mice intravenously. We found that, in the middle phase setting, the thresholds (von Frey test) of the right hindpaw of the MCP-1/sSiglec-9 group on day 14 was 7.32 ± 1.75 g, which was significantly higher than that of the DMEM and PBS groups but was lower than that of the SHED-CM group (Figures 10A,B). These results suggest that the promising analgesic ability of SHED-CM might partly rely on MCP-1/sSiglec-9.