SPF-grade male Sprague-Dawley (SD) rats weighing 200 $\pm$ 20 g were used. The rats were purchased from Zhejiang Viton Lihua Experimental Animal Technology Co., Ltd., and maintained at the Animal Experimental Center of Shanghai University of Traditional Chinese Medicine. We performed the animal experiments in accordance with the procedures approved by the Animal Experimental Center (Animal Ethics No.: PZSHUTCM200911014) and implemented the relevant provisions from the Guidance Suggestions for the Care and Use of Laboratory Animals formulated by the Ministry of Science and Technology of the People’s Republic of China.

We used the rat model of AA established as previously described [16]. Briefly, 50 $\mu$L complete Freund’s adjuvant (CFA) (Sigma-Aldrich, St. Louis, MO, USA) was injected into the left ankle joint cavity of rats anesthetized with 1.5% isoflurane. Local swelling and behavioral disability appeared within 24 h.

A single course of AP treatment for 20 min was performed at Zusanli (ST 36) acupoint of the affected side of conscious rats. According to our researches and related studies [4], the ST 36 has a desirable anti-nociceptive effect on AA rats. ST 36 is located at the posterolateral side of knee joint, about 5 mm below the fibula capitulum. A stainless-steel needle (0.18 mm $\times$ 13 mm, Cloud Dragon Medical Equipment Company, Wujiang, China) was gently inserted to an approximate depth of 7 mm. AP procedure was dominated by twirling-rotating motion (~100 times/min), supplemented with lifting-thrusting motion (~80 times/min).

Pain thresholds of the injured hind plantar were determined to reflect the pain levels. Mechanical allodynia (paw withdrawal threshold, PWT) and thermal hyperalgesia (paw withdrawal latency, PWL) were evaluated to determine the pain level. The timeline and methods for behavioral tests was referred to our published work [16, 18]. Briefly, CFA was administrated on day 0 and AP was performed 2 days later. Pain thresholds were determined for three times, just before CFA injection, just before AP intervention, and after 30 min-AP treatment. Thermal hyperalgesia was measured 2 h after the mechanical allodynia assessment.

All stock solutions were stored at –20 °C and diluted into working solutions to final concentrations when used. Sodium cromolyn (CRO, final concentration: 0.02 g/mL, 20 $\mu$L/paw, #C0399, Sigma-Aldrich, St. Louis, MO, USA), a MC stabilizer was pre-injected into the left ST 36 in layers 10 min prior to AP treatment, with half receiving 5 mm subcutaneous injections into the muscle and the other half receiving 2 mm subcutaneous injections under the dermis. Granisetron hydrochloride injection (Gran., 1 mg/mL, 20 $\mu$L/paw, National medicine permission number: H20043610) was procured from Guorui Pharmaceutical Co. Ltd. (China National Pharmaceutical Group, Beijing, CN). Methiothepin mesylate (Meth., #T2190, TopScience Co., Ltd., Shanghai, CN) were dissolved in dimethyl sulfoxide (DMSO), and the final concentration was 5 mg/mL, 20 $\mu$L/paw. WAY-100635 (WAY., 2 mg/mL, 20 $\mu$L/paw, #W108, Sigma-Aldrich, St. Louis, MO, USA), ARL67156 (ARL, 100 $\mu$M, 50 $\mu$L/paw, #A265, Sigma-Aldrich, St. Louis, MO, USA), 2-Chloro-N6-cyclopentyladenosine (CCPA, 0.04 mg/mL, 20 $\mu$L/paw, #C7938, Sigma-Aldrich, St. Louis, MO, USA), Suramin (100 $\mu$M, 50 $\mu$L/paw, #574625, Sigma-Aldrich, St. Louis, MO, USA) and Pyridoxal phosphate-6-azo tetrasodium salt hydrate (PPADS, 100 $\mu$M, 50 $\mu$L/paw, #P178, Sigma-Aldrich, St. Louis, MO, USA) were prepared with distilled water. All the agonists and antagonists were pre-injected into the left ST 36 in layers prior to AP treatment. CRO (25 $\mu$M), Compand 48/80 (16 $\mu$M, #C2313, Sigma-Aldrich, St. Louis, MO, USA), Meth. (100 nM), Gran. (100 nM) and WAY. (100 nM) were introduced to cells 20 minutes prior to the application of mechanical stimulation or exo-5-HT and were present during the entire process. Preparation of bath solution and 50% hypotonic solution was described in our previous work [5].

In order to intervene certain corresponding signals at the acupoints, 20–50 $\mu$L related reagents were pre-injected into the left ST 36 in layers 20 min prior to AP treatment.

The performance of microdialysis technology referred to our published work [19]. Briefly, after rat was anesthetized with 1.5% isoflurane, a linear microdialysis probe (CMA 31, CMA Microdialysis AB, Kista, Sweden) was implanted at a distance of 4–6 mm from ST 36. The dialysis solution, Hank’s balanced salt solution (HBSS, Corning Cellgro, Manassas, VA, USA), was controlled by syringe pumps (NE-1000, New Era Pump Systems, Farmingdale, NY, USA) and continuously perfused at a rate of 1 $\mu$L/min. Subsequently, rat was restrained in the rat fixator and was gradually awakened from anesthesia. After a 2 h recovery period, microdialysis outflow was collected for 30 minutes per sample.

5-HT and ATP in the microdialysate were determined by ELISA kit (HZ-5-HT-RA, Zhen Shanghai and Shanghai Industrial Co., Ltd., CN) and luciferin-luciferase assay (L-L, Sigma-Aldrich, St. Louis, MO, USA), respectively. The procedures were performed according to the manufacturer’s instructions. The light emission was measured by a microplate reader (Synergy Mx, BioTek, Winooski, VT, USA). Calibrations were executed before and after each dertermination.

Adenosine levels was performed by high-performance liquid chromatography (HPLC) as previously described [4]. The adenosine standard was checked both before and after the measurement, and the concentrations were 3000 nM, 1000 nM, 300 nM, and 100 nM. The detection results of the animal samples were compared to the standard according the peak areas, and the concentration was calculated.

Tissue sectioning and MC Staining referenced to our previous article [4]. Briefly, after the final behavioral test, a 5 $\times$ 5 $\times$ 5 mm specimen of ST 36 was taken and 5 $\mu$m paraffin sections were prepared. After staining with toluidine blue, the MCs were dark purple in color. MCs with a blurred boundary, or dispersed granules were counted as degranulated MCs.

The performance of western blotting referred to related research [20]. We used rabbit monoclonal antibodies: GAPDH (#ab8245, Abcam, Waltham, MA, USA), 5-HT${}_{\text{1A}}$R (#ab85615, Abcam, MA, USA), 5-HT${}_{\text{1D}}$R (#ab140486, Abcam, MA, USA), 5-HT${}_{\text{1B}}$R (#PA1-41069, Thermo Fisher Scientific, Waltham, MA, USA), 5-HT${}_{\text{1F}}$R (#PA5-106542, Thermo Fisher Scientific, MA, USA), Connexin 43 (#71-0700, Thermo Fisher Scientific, MA, USA), Pannexin 1 (#488000, Thermo Fisher Scientific, MA, USA) and P2Y${}_1$ (#P6487, Sigma-Aldrich, St. Louis, MO, USA). The proteins were then incubated with peroxidase-conjugated goat anti-rabbit IgG (#7074, Cell Signaling Technology, MA, USA). A visualizer was then used to observe the stripe, and the grayscale value was calculated using Image J v1.8.0 (National Institutes of Health, Bethesda, MD, USA).

The human MC (HMC-1) cell line was kindly provided by Dr J. H. Butterfield (Mayo Clinic) and cultured in Roswell Park Memorial Institute (RPMI) 1640 medium. Rat basophilic leukemia (RBL-2H3) cell line was purchased from the Cell Bank of Chinese Academy of Sciences (Shanghai, CN) and cultured in Minimum Essential Medium (MEM) medium. The culture medium was supplemented with 10% Fetal Bovine Serum (FBS), 100 units/mL penicillin and 100 units/mL streptomycin in a 5% CO${}_2$ humidified atmosphere at 37 °C.

Hypotonic shock was introduced to HMC-1 cells to induce degranulation, and tryptase or ATP release. The entire procedure referenced to our article [15]. Briefly, 1–2 mL of HMC-1 cell suspension (0.5–1.5 $\times$ 10${}^6$ cells/mL) was gently introduced into the polycarbonate filter chamber, and superfused with bath solution at 1.3 mL/min. After an equilibration period, hypotonic solution (200 mOsm/Kg H${}_2$O) was introduced to the cells and the perfusate was collected at 1 min. Hypotonic environment is a common method to stimulate cells with force [21]. The cell becomes swollen and its cellular membrane is extended, then the corresponding downstream signals are initiated. In our previous work, we used hypotonic shock to activate mast cells in vitro to mimic manual AP [5, 15, 22]. Tryptase or ATP were assessed by ELISA (#IMM001, Sigma-Aldrich, St. Louis, MO, USA) and luciferase-luciferin assay. For 5-HT release measurements, we stimulated the cells by medium displacement. HMC-1 cells grew in the 96-well plate containing 100 $\mu$L culture medium. Half of the medium was gently pipetted up and down ten times with a pipette gun. Subsequently, 5-HT in the cell suspension was qualified by ELISA assay.

For each experiment, RPMI 1640 and MEM medium was used to further dilute 10 mM 5-HT stock solutions to acquire various required concentrations of 5-HT. 5-HT-sensitive ATP released from HMC-1 cells in response to exogenous 5-HT (exo-5-HT, #H9523, Sigma-Aldrich, St. Louis, MO, USA) in the range of 1 nM–10 $\mu$M for 15 min. Released ATP levels in the fractions were quantified by L-L assay, as described in 2.8. Based on our preliminary results, when exo-5-HT concentration $\ge$50 $\mu$M, cell mortality tended to be above 80%, and the large increase of ATP at this time might be caused by cell death.

The measurement data are expressed as the mean $\pm$ standard error (SE) values. The data were analyzed using SPSS 25.0 (IBM Co., Armonk, NY, USA). The figures were prepared by GraphPad Prism 8.0 software (GraphPad Software, San Diego, CA, USA). Differences among multiple groups were tested with one-way or two-way analysis of variance (ANOVA), followed by Fisher’s least significant difference (LSD.) test. Where only two groups were compared, a two-sample t-test was used. When data were non-normally distributed, a Mann-Whitney U test was performed. We considered p 0.05 to be statistically significant.

We revealed that 30 min-needling caused 5-HT accumulation in the interstitial space of treated acupoint. Inters. 5-HT concentration remarkably increased from 5.5 ng/mL $\pm$ 0.2 ng/mL (time = 0) to 11.4 ng/mL $\pm$ 0.4 ng/mL (t = 30 min) (n = 4, p 10${}^{-15}$) gradually restored to the baseline level (Fig. 1A). Inters. 5-HT amount was significantly potentiated from baseline of 156.8 pg $\pm$ 5.8 pg (-30–0 min) to 252.5 pg $\pm$ 7.2 pg (0–30 min) (p = 1.2 $\times$ 10${}^{-12}$, n = 4) (Fig. 1B). Such potentiation effect lasted for at least 120 min. AP-induced 5-HT mobilization was canceled by the presence of CRO (0.02 g/mL, 20 $\mu$L), a stabilizer of MCs, which was pre-injected in ST 36 before AP intervention (Fig. 1A,B). These results suggest that needling activate MCs in acupoint to release 5-HT.

Fig. 1.

AP mobilized 5-HT release from MCs in acupoint, which contributed to AP-analgesic effect (n = 4–10, mean $\mathrm{\pm }$ SE). (A) Time-course of inters. 5-HT concentration in response to needling on AA model rats in the absence and presence of CRO (0.02 g/mL, 20 $\mu$L) pre-injected in ST 36 (n = 4). Inters. 5-HT was collected by microdialysis technique and determined by ELISA. ### p 0.001 vs. baseline (time = 0 min); ** and ***, p 0.01 and 0.001 vs. +CRO, respectively. Two-way ANOVA and LSD. test were used. (B) Changes of inters. 5-HT amount in response to AP in the absence and presence of CRO (n = 4). They were calculated based on the area under the curve in (A). ## and ###, p 0.01 and 0.001 vs. baseline (time = -30–0 min); *, ** and ***, p 0.05, 0.01 and 0.001 vs. +CRO, respectively. (C,D) Effects of AP treatment on CFA-induced tactile allodynia (PWT) and thermal hyperalgesia (PWL) of injured-side hindpaws, respectively. ### p 0.001 vs. baseline (Day 0). $\bullet$$\bullet$$\bullet$p 0.001 vs. Day 2, before AP. Two-way ANOVA and LSD. test were used to perform within-group comparisons. (E,F) Effects of pre-co-injection of Methiothepin (Meth.) and Granisetron (Gran.) in ST 36 on AP-relieved PWT and PWL, respectively. Data illustrate the final behavioral tests. ** and ***, p 0.01 and 0.001, respectively. One-way ANOVA and LSD. test were used to perform between-group comparisons. (G) AP/Mechanical stimulation induced MCs to degranulate. Allows point to degranulated MCs. (i) Paraffin sections prepared from ST36 of AA model group (left) and AP group (right). MCs were stained with toluidine blue and were marked in purple. (ii) Connective tissue slices isolated from ST 36 of normal rat before (left) and after (right) hypotonic shock. (iii) HMC-1 cells before (left) and after (right) hypotonic shock. In (ii) and (iii), the tissue slice and HMC-1 cells were incubated in 200 mOSm/kg H2O-hypotonic solution for 25 min and 15 min, respectively. The plasma membrane of degranulated MCs became rough and the granules scattered around the cells. Scale bar: 50 $\mu$m. Mech. stim.: Mechanical stimulation. (H) Release of tryptase, ATP and 5-HT from HMC-1 cells in response to 50% hypotonic shock (n = 7–10). Unpaired t-test was performed for two groups’ comparison. *** p 0.001. In (E–G), numbers above each column denote sample size (n).

Behavioral test revealed that CFA-treated rats suffered from pain hypersensitivity, exhibiting tactile allodynia (PWT) (reduced from 54.9 g $\pm$ 1.1 g to 11.1 g $\pm$ 1.4 g, n = 6, p = 2.4 $\times$ 10${}^{-13}$ vs. baseline) and thermal hyperalgesia (PWL) (reduced from 19.6 s $\pm$ 0.8 s to 6.6 s $\pm$ 0.1 s, n = 6, p = 3.6 $\times$ 10${}^{-8}$ vs. baseline), but clearly benefited from 30 min-AP at ST 36 (PWT: 43.2 g $\pm$ 2.8 g, n = 6, p = 2.4 $\times$ 10${}^{-13}$ vs. before AP; PWL: 15.7 s $\pm$ 1.4 s, n = 6, p = 1.7 $\times$ 10${}^{-4}$ vs. before AP) (Fig. 1C,D).

Based on our previous research, the microdialysis recovery percentage was in the range of 25–40% [19]. Considering 20 $\times$ dilution by diffusion and/or convection into the bulk extracellular volume [19, 23], the real concentration of 5-HT at the cell surface near the AP site was more than 500 ng/mL (2.8 $\mu$M), indicating some 5-HT receptors might be activated [17, 24, 25]. Considering the expressions and effects in pain processing in acupoints [11, 12], we tried to verify the role of 5-HT${}_1$Rs and ${}_3$R. Co-injection of Methiothepin (5 mg/mL, 20 $\mu$L), a nonselective antagonist of 5-HT${}_1$Rs, and Granisetron (1 mg/mL, 20 $\mu$L), a selective antagonist of 5-HT${}_3$R, in ST 36 prior to AP suppressed the anti-nociceptive effect of AP on PWT (13.2 g $\pm$ 1.0 g, n = 6, p = 1.3 $\times$ 10${}^{-7}$ vs. AP) and PWL (6.3 s $\pm$ 0.7 s, n = 6, p = 0.0014 vs. AP) (Fig. 1E,F). Additionally, compared with normal rats, CFA induced more 5-HT mobilization in rats (Supplementary Fig. 1), suggesting MCs in acupoint changed dynamically with the ankle arthritis.

AP-induced subcutaneous MCs degranulation is the basis for triggering the analgesic mechanism [3, 4], which was confirmed in the current work (Fig. 1G,i). In addition, MCs in the connective tissue slice ex vivo prepared from rat ST 36 (Fig. 1G,ii) and cultured HMC-1 cells in vitro (Fig. 1G,iii) also degranulated in response to mechanical stimulation of hypotonic shock (200 mOsm/Kg H${}_2$O) that was used to mimic needling stimulation. Determination of the mediator in cellular suspension showed that when HMC-1 cells were perturbed with 50% hypotonic shock, tryptase release, a classic index for MCs degranulation [26], was potentiated from 23.5 ng/mL $\pm$ 2.4 ng/mL of control to 41.5 ng/mL $\pm$ 2.9 ng/mL (n = 10, p = 0.0004, vs. Control) (Fig. 1H). In addition, secretion of stress-sensitive molecular ATP and 5-HT were also strengthened (ATP: from 16.1 $\pm$ 1.2 ng/mL of control to 32.1 $\pm$ 3.2 ng/mL, n = 8–10, p = 0.0001 vs. Control; 5-HT: from 4.1 ng/mL $\pm$ 0.2 ng/mL of control to 7.4 ng/mL $\pm$ 0.3 ng/mL, n = 7–8, p = 1.1 $\times$ 10${}^{-6}$ vs. Control) (Fig. 1H). 5-HT is well known to be released in the manner of degranulation [27]. While, unexpectedly, mechanosensitive ATP release was neither suppressed by CRO nor enhanced by Compand 48/80 (Supplementary Fig. 2A), but tryptase release was remarkably affected (Supplementary Fig. 2B), indicating that ATP secreted from MCs in a degranulation-independent manner.

5-HT is reported to induced ATP release from type II cells of the rat carotid body [17]. Mechanosensitive ATP release is discovered in various mammalian cells [21], including MCs [5]. Needling-induced ATP transient accumulation in treated acupoint is supposed as one of the primary steps toward AP-analgesic mechanism by providing the precursor for adenosine [16, 19]. Besides primary mechanosensitive ATP release, secondary ATP secretion had been found in MCs in our previous work [15]. Subsequently, we tried to verified whether 5-HT could mediate MCs to release ATP. Fig. 2A shows the dependency of ATP generation on the concentration of exogenous 5-HT (exo-5-HT) (1 nM–10 $\mu$M). When the concentration of exo-5-HT reached 1 $\mu$M and 10 $\mu$M, extracellular ATP level was significantly potentiated from baseline of 15.7 nM $\pm$ 2.5 nM to 43.6 nM $\pm$ 5.3 nM and 107.3 nM $\pm$ 3.4 nM, respectively (Fig. 2A). 5-HT concentration greater than 10 $\mu$M was not determined because cells viability was affected (Supplementary Fig. 3).

Fig. 2.

5-HT receptors mediated ATP release and contributed to AP analgesia (n = 3–8, mean $\mathrm{\pm }$ SE). (A) Concentration-dependent mediation of ATP release from HMC-1 cells by 1 nM–10 $\mu$M exo-5-HT). ### p 0.001 vs. baseline (absent of exo-5-HT). (B,C) Effects of 5-HT receptors inhibition on exo-5-HT induced ATP release from HMC-1 cells. Meth. (100 nM) or Gran. (100 nM) was introduced to cells 10 min before 5-HT intervention and was present during the entire process. ** p 0.01. (D) Response of 5-HT release from HMC-1 cells to 10 nM–100 $\mu$M exogenous ATP intervention. (E) WB determination of 5-HT${}_{\text{1A, B, D, F}}$ subtypes identification on HMC-1 cells. These are representative blots from n = 4–6 recordings. (F) Effects of 5-HT${}_{\text{1A}}$R inhibition on exo-5-HT induced ATP release from HMC-1 cells. WAY-100635 (WAY., 100 nM) was introduced to cells 10 min before exo-5-HT intervention and was present during the entire process. * p 0.05 and *** p 0.001. (G,H) Effect of 5-HT${}_{\text{1A}}$ receptor-inhibition on AP-relieving tactile allodynia and thermal hyperalgesia of the injured plantar, respectively. WAY. (2 mg/mL, 20 $\mu$L) was pre-injected at ST 36. Data illustrate the final behavioral tests. ** and ***, p 0.01 and 0.001, respectively. Numbers above each column/circle denote sample size (n). In (F), Unpaired t-test was performed. In the rest panels, one-way ANOVA and LSD. test were used.

Pretreatment HMC-1 with Methiothepin (100 nM) significantly impaired ATP release induced by 1 $\mu$M- and 10 $\mu$M-exo-5-HT as well. In contract, Granisetron (100 nM) had null effect (Fig. 2B,C). Similar results were also obtained from rat MCs, RBL2H3 cell line (Supplementary Fig. 4). However, exogenous ATP had null effect on 5-HT release (Fig. 2D). These results suggest that 5-HT mediates ATP release from MCs via activating 5-HTRs, especially 5-HT${}_1$Rs.

5-HT${}_1$Rs family comprises A, B, D, E and F subtypes. Currently, all subtypes except for E have been reported to be involved in transducing of peripheral pain [28]. Among them, subtype A has been repeatedly confirmed to be expressed in human and rat MCs. Our WB determination revealed that 5-HT${}_{\text{1A}}$R markedly expressed in HMC-1 cells (Fig. 2E). Subsequently, we examined the effect of WAY-100635 (WAY.), a specific antagonist of 5-HT${}_{\text{1A}}$R. The result uncovered that pretreatment HMC-1 with WAY. (100 nM) significantly impaired ATP release induced by 1 $\mu$M (n = 5–6, p = 2.8 $\times$ 10${}^{-6}$ vs. -WAY.) and 10 $\mu$M (n = 5, p = 0.0020 vs. -WAY.) exo-5-HT (Fig. 2F). Taken together, these results indicate that 5-HT mediates ATP release from MCs via activating 5-HT${}_{\text{1A}}$R subtype.

Based on these findings, we returned to the test in vivo. Behavioral tests found that pre-injection of WAY. (2 mg/mL, 20 $\mu$L) in ST 36 canceled AP analgesic effect on PWT (n = 5–6, p = 8.1 $\times$ 10${}^{-5}$ vs. AP) and PWL (n = 4–6, p = 0.0003 vs. AP) (Fig. 2G,H). Taken these results together, we infer that contribution of MC-associated 5-HT release to AP analgesia mechanism might be due to mediate ATP release from MCs or other adjacent cells by activating 5-HT${}_{\text{1A}}$R.

AP-induced adenosine accumulation and activation of local adenosine A${}_1$Rs contribute to initiate analgesia [14]. Such adenosine mobilization might partially rely on the hydrolyzation of ATP. Our microdialysate determination revealed that both adenosine and ATP in the interstitial space of treated acupoint were significantly potentiated by needling (adenosine, from baseline of 58.4 nM $\pm$ 6.9 nM to 76.8 nM $\pm$ 9.1 nM, n = 9, p = 0.0250; ATP, from baseline of 3.4 nM $\pm$ 1.4 nM to 56.0 nM $\pm$ 12.5 nM, n = 4, p = 0.0269). More importantly, the presence of ARL (100 $\mu$M, 50 $\mu$L), a non-specific inhibitor of ecto-ATPase, in acupoint further potentiated AP-induced ATP accumulation (136.3 nM $\pm$ 22.7 nM, n = 4, p = 0.0258 vs. AP group) (Fig. 3A), indicating the activities of ecto-ATPase and hydrolyzation of inters. ATP existed in the treated acupoint. Further behavioral tests demonstrated that AP analgesic effect was impaired by ARL (100 $\mu$M, 50 $\mu$L) (n = 4, p = 0.0233), and was duplicated by CCPA (0.04 mg/mL, 20 $\mu$L), an agonist of adenosine A1Rs (n = 5, p = 0.9773) (Fig. 3B). Taken together, these findings suggest that mobilization of inters. ATP via primary or secondary release contributes to anti-nociceptive mechanism of AP by promoting adenosine production.

Fig. 3.

AP-induced the mobilization of adenosine (Ado) and inters. ATP in the treated acupoint contributed to the anti-nociceptive effect of AP (n = 4–9, mean $\pm$ SE). (A) Changes of Adenosine and ATP levels in the interstitial space in ST 36 of AA model rats before (-AP) and after (+AP) needling. Adenosine was determined with HPLC. ATP was assayed with L-L method. ARL (100 $\mu$M, 50 $\mu$L) was pre-injected at ST 36 20 min prior to AP. * p 0.05 by paired t-test. # p 0.05 by unpaired t-test. (B) Involvement of inters. ATP hydrolyzation and Adenosine A1 receptors in AP-analgesic effect. ARL (100 $\mu$M, 50 $\mu$L) or CCPA (0.04 mg/mL, 20 $\mu$L) was pre-injected at ST 36. ** and ***, p 0.01 and 0.001, respectively. (C,D) Effects of P2 receptor-inhibition on AP-analgesia in tactile allodynia and thermal hyperalgesia, respectively. *, ** and ***, p 0.05, 0.01 and 0.001, respectively. In (B–D), data illustrate the final behavioral tests and one-way ANOVA and LSD. test were used. Numbers above each column denote sample size (n).

Calcium wave propagation (CWP) is supposed as the basis of ATP regeneration, which facilitates the localized propagation of biological signals [29]. CWP has been uncovered in mouse treated acupoint [30]. Our previous work had found a similar CWP and cascaded ATP release in HMC-1 cells when perturbed by hypotonic shock, in which primary mechanosensitive ATP released induced secondary ATP release via activating P2Y${}_{13}$ or P2X${}_7$ receptors [15].

The present work found an additional 5-HT-induced secondary ATP release. We hypothesized that besides as the precursor of adenosine, such cascaded ATP release could amply and propagate the localized needling signals. Our WB determination revealed that some components for forming CWP/ATP regeneration, including Panx-1, connexin 43 and P2Y${}_1$ were expressed in HMC-1 cells (Supplementary Fig. 5). Subsequently, further behavioral tests observed that inhibition of local purinergic receptors (P2 receptors) by Suramin (100 $\mu$M, 50 $\mu$L) (Relative PWT: reduced from 0.9 $\pm$ 0.0 to 0.7 $\pm$ 0.1, n = 4, p = 0.0057 vs. AP and Relative PWL: reduced from 0.9 $\pm$ 0.0 to 0.6 $\pm$ 0.0, n = 4, p = 0.0269 vs. AP) or PPADS (100 $\mu$M, 50 $\mu$L) (Relative PWT: reduced from 0.9 $\pm$ 0.0 to 0.7 $\pm$ 0.1, n = 4, p = 0.0009 vs. AP and Relative PWL: reduced from 0.9 $\pm$ 0.0 to 0.8 $\pm$ 0.1, n = 4, p = 0.0749 vs. AP) partially suppressed the AP-relieved tactile allodynia and thermal hyperalgesia (Fig. 3C,D). This finding implies that besides adenosine and A1Rs, ATP/ADP and related P2 receptors in acupoint also play a role to mediate AP analgesia and its underlying mechanism might be the ATP-related CWP.

5-HT-immunopositive MCs are present in acupoints and become more when the acupoints are sensitized [8, 9]. Acupoint sensitization means acupoints transformed dynamically from a “silent” state to an “activated” one when the body is suffering from some corresponding diseases [31]. Neurogenic inflammation is one of the biological bases for acupoint sensitization due to dorsal root reflex or axon reflex [32, 33]. In the present work, we noticed that compared to normal rats, ST 36 in model group had higher inters. 5-HT level at baseline (Supplementary Fig. 1). Here, the onset site and treated acupoint are innovated by tibial nerve and peroneal nerve, respectively, branches of sciatic nerve, suggesting that they are innovated by the same or adjacent peripheral ganglia and spinal segments. Similarly, in knee osteoarthritis rats, 5-HT together with tryptase and histamine are upregulated in the nearby acupoints, Yanglingquan acupoint and Heding acupoint rather than Weizhong acupoint [34]. The greatest benefit of acupoint sensitization is to facilitate AP to exert better therapeutic effect. Our previous work had revealed that AP modulated higher ATP accumulated in AA rats than in normal ones [19]. Systematic review demonstrates the comparative superiority of needle stimulation of sensitized points over non-sensitized points, especially for some pain syndromes [31].

Degranulation of local subcutaneous MCs induced by needling has been often reported [1]. While the certain associated mediators are not well quantitatively studied. Beside adenosine assessed in our previous work [4], the current study clearly recorded MC-associated 5-HT mobilization (Fig. 1A,B). Based on the microdialysis recovery percentage [19] and considering dilution by diffusion and/or convection into the bulk extracellular volume [23], the real concentration of inters. 5-HT at the cell surface was more than 500 ng/mL (2.8 $\mu$M). This concentration is within the range we used in our test in vitro (1 and 10 $\mu$M), which is sufficient to activate 5-HT${}_{\text{1A}}$R (Fig. 2F). Additionally, this concentration range can activate 5-HT${}_{\text{1A}}$R [24], 5-HT${}_{\text{2A}}$R [17] and 5-HT${}_3$R [25], implying various functions are involved, for example, influence the differentiation, proliferation, migration, adherence and life-span of skin cells, regulate the tonus of blood vessels and influence nerve transmission [35]. The potentiated inters. 5-HT could last for more than 2.5 h (Fig. 1A,B). It might contribute to the maintenance of AP effect.

The activation of MCs can easily establish a positive feedback loop. Matrix metalloproteinase-9 [36] or ATP [15] secreted from MCs, in turn, activate MCs. In the present work, unidirectional regulation of 5-HT on ATP release via activation of 5-HT${}_{\text{1A}}$R was another example of this self-activation mechanism (Fig. 2A). Other cells can also be activated, as long as they express 5-HT${}_{\text{1A}}$R, for example, melanocytes [37]. Similar finding has been reported in type II cells of rat’s carotid body, in which exo-5-HT increases [Ca${}^{2+}$]${}_i$ partially via activating Panx-1 channels, implying 5-HT might facilitate ATP release [17]. Although we demonstrated that MCs express Panx-1 channels (Supplementary Fig. 5), whether they assisted ATP release still need further determination.

Additionally, interaction between 5-HT mediated signaling and ATP-related P2X receptors have been characterized. While it is mainly inhibitory rather than facilitative. In submucosal neurons of guinea pigs, whole-cell recordings reveal that 5-HT${}_3$Rs and P2X channels negatively modulate each other when they are simultaneously activated [38]. 5-HT${}_{\text{3A}}$R and P2X2 channels were found to co-immunoprecipitate constitutively and be associated in clusters in cultured myenteric neurons of guinea pig as well as expression cells [39]. 5-HT${}_3$Rs in pelvic afferent neurons is up-regulated in mice lacking P2X2 or P2X3 receptor genes [25].

We demonstrated an upregulation of inters. ATP in the treated acupoint by needling (Fig. 3A), and prevention its hydrolysis by ARL impaired AP-analgesic effect (Fig. 3B). These findings imply that ATP mobilization and subsequent hydrolysis are vital steps towards antinociception during needling. The aggregated inters. ATP originates from cell lysis and non-lytic release as well. Regarding the former, needle piercing and subsequent manipulation can cause tissue damage. As reported, damage-associated molecules, high mobility group box 1 protein (HMGB1) and toll-like receptor 4 (TLR4), increases in the treated acupoint of normal rats by 2-min needling [40]. The non-lytic release of ATP results from the primary ATP secretion via activating mechanosensitive channels [1] and secondary ATP release induced by ATP or 5-HT, as we discussed above. But unlike 5-HT, mechanosensitive ATP release via non-degranulation approach (Supplementary Fig. 2A).

Adenosine, as the downstream product of ATP, aggregates in acupoints too [14, 41]. The current work and other similar study [14] observed that pharmacologic activation of adenosine A1R in acupoints had antinociceptive effect (Fig. 3B), in which A1R are supposed to be situated on local nerve terminals [14]. Like ATP, adenosine might also release out via lytic and non-lytic approaches [42].

We inferred the presence of ecto-ATPases in the acupoint based on the reinforcing effect of ARL on inters. ATP accumulation (Fig. 3A). NTPDases family is the major nucleotide-hydrolyzing enzymes, which degrade ATP to AMP with intermediate formation of ADP [43]. Our previous work revelated that rats ST 36 acupoint expressed mRNA of NTPDase-1, -2,-3 [16]. Combined with ecto-AMPase, for example, ecto-5’-nucleotidase (Nt5e), can fulfill the production of adenosine. In the skin, Langerhans cells express NTPDase-1 [44], keratinocytes express NTPDase-3 [45] and Nt5e [46]. Behavioral evidences in mice show that nucleotidases prostatic acid phosphatase (PAP) [47] and Nt5e [48, 49] have anti-nociceptive effects at spinal cord level via promoting adenosine production. Injection of PAP into Weizhong acupoint (BL 40) has analgesic effect similar to AP treatment [50].

Needling manipulations generate mechanical stimulation that is transmitted to the wider and deeper space by subcutaneous collagen fibers twisting around the needles [51, 52]. This is the “physical propagation” of needling signals [3]. In addition, “biochemical propagation” also occurs. As reported, AP-generated acoustic shear waves can cause [Ca${}^{2+}$]${}_i$ shock and CWP in vitro and in vivo [30]. CWP permits regenerative signal amplification. It has been proposed that non-neural cells respond to mechanical stimulation by ATP release and send signals to nerve terminals via CWP [30]. CWP formation is commonly based on ATP regeneration [29]. Compelling evidence indicates that ATP-mediated ATP release involves P2YRs [53], among which P2Y1R is the most documented [54, 55]. Our previous work had observed a similar CWP phenomenon in cultured MCs in response to mechanical stimulation, which relied on the activation of P2Y${}_{13}$ and P2X${}_7$ receptors via “ATP-induced ATP release” mechanism [15]. 5-HT-mediated ATP release uncovered in this work is a good complement to ATP regeneration, then facilitate CWP formation. The necessary components for CWP, including P2Y1 receptors, Connexin 43 and Panx-1 channels are expressed in mice acupoint (ST 36) [20]. In the present work, WB determined their presence in MCs (Supplementary Fig. 5). Furthermore, Panx-1 [56] and connexin 43 [57] are been reported to express in skeletal muscle and fibroblasts that are also main tissues contained in acupoints. Hence, besides as the precursor of adenosine, inters. ATP in acupoint might exert function of transmitting the needling signals. This hypothesis was partially confirmed by the inhibitory effect of Suramin and PPADS on ATP regeneration in MCs in our previous work [15] and AP-analgesic effect in the present study (Fig. 3C,D).

We found that 5-HTRs, especially 5-HT${}_{\text{1A}}$R, in acupoint played role in the initiation mechanism of AP-analgesia (Fig. 1E,F and Fig. 2G,H). Beside facilitating needling signals as we discussed above, 5-HT might contribute to AP analgesic effect by activating 5-HT${}_{\text{1A}}$R on the adjacent peripheral nerve endings. As we investigated previously, AP-activated MCs were eventually involved in the analgesic effect via transmitting the biological signals to the nearby nerve endings, then ascending to central nerve system [58]. 5-HT${}_1$Rs class represents the most complex subtype families of 5-HTRs, which comprise of five receptor subtypes, 5-HT${}_{\text{1A}}$, 5-HT${}_{\text{1B}}$, 5-HT${}_{\text{1D}}$, 5-HT${}_{\text{1E}}$ and 5-HT${}_{\text{1F}}$. Although 5-HT${}_{\text{1A}}$R activation is reported to elicit hyperalgesia [59, 60], later researches reveal its anti-nociceptive effect. Peripheral sensory nerve endings are the main location of 5-HT${}_{\text{1A}}$ subtype [61]. Administered intrathecally with ($\pm$)-8-hydroxydipropylaminotetralin hydrobromide (8-OH-DPAT), a 5-HT${}_{\text{1A}}$R agonist, was to suppressed the second phase of formalin-induced aversive responses in the 5,7-dihydroxytryptamine treated rats [62]. Another research using 8-OH-DPAT also suggested an analgesic action is exerted by 5-HT${}_{\text{1A}}$R in the formalin model of tonic nociceptive pain in rats [63]. It still needs further work to explore whether 5-HT mediates AP analgesia by activating 5-HT${}_{\text{1A}}$R situated on nearby nerve endings.