All animal procedures were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals (Publication 85–23, Revised 1996) and were performed according to the University of Maryland-approved Institutional Animal Care and Use Committee protocols and the ARRIVE guidelines. C57BL/6 mice, Tac1 ^−/− (Jax, 004103), and Calca ^Cre/Cre mice (Jax, 033168) were obtained from the Jackson Laboratory. Calca^Cre is a knock-in/knock-out line in which Cre is targeted to the first coding exon of the mouse Calca gene, and therefore homozygote knock-in mice do not express Calca. Both Tac1 ^−/− and Calca^Cre/Cre mice procreate normally. Eight-week-old male and female mice were used in each experimental group. In all assays, the animals were randomly allocated to the experimental groups. Animals were group-housed under standard conditions with ad libitum access to water and food. The experimenters were blinded for the treatment groups during the analysis of the data in each assay.

Mice were anesthetized using ketamine/xylazine (100-150 mg/kg of ketamine and 10-20 mg/kg of xylazine). A 5-0 silk ligature was placed around the second maxillary molar (M2), which remained in place in all mice throughout the experimental period. The suture was tied gently to prevent damage to the periodontal tissue. The ligatures remained in place in all mice throughout the indicated experimental period.

The animals were anesthetized with ketamine/xylazine and euthanized via transcardial perfusion using 4% PFA (4). Maxillae were hemisected and micro-focus computed tomography (µCT) images were obtained using a Siemens Inveon Micro-PET/SPECT/CT (Siemens, Ann Arbor, MI) at a 9 µm spatial resolution. Siemens Inveon Research Workplace 4.2 software was used for image acquisition and processing, 2-D and 3-D image viewing, and quantitative analysis. Unless otherwise indicated, bone loss was evaluated from the buccal side. To assess the levels of periodontal bone loss, four linear measurements were taken following 3-D reconstruction of the µCT scans. Bone loss was evaluated on the distal side of the distobuccal root of the first molar (M1), the mesial side of the mesiobuccal root of the second molar (M2), the distal side of the distobuccal root of M2, and the distal side of the distobuccal root of the third molar (M3). Unless otherwise indicated, distances from the cement-enamel junction (CEJ) to the most apical site of bone destruction were measured. Four measurements were averaged to obtain the distance from the CEJ.

Immunohistochemical assays of the trigeminal ganglion (TG) and maxillae were performed as previously described (4, 21–23). Maxillae were decalcified in 0.5M Ethylenediaminetetraacetic acid (EDTA) in PBS (pH 7.4) over the course of 14 days. Tissue processing was performed at the Histology Core at the University of Maryland School of Medicine for paraffin embedding and the tissues were sectioned at 5 µm. Some tissues were cryoprotected and cryosectioned at a thickness of 12 µm for TG and 30 µm for the decalcified maxillae. Immunohistochemical assays were performed using rat anti-substance P (Sigma-Aldrich, #MAB356), rabbit anti-TRPV1 (kindly provided by Dr. Michael Caterina at Johns Hopkins University), anti-CGRP (Millipore, #C7113, rabbit or Penninsula lab, guinea pig). pan-cytokeratin (BLD, #914204; mouse), and rabbit anti-CD45 (abcam, #ab10558; rabbit). The specificity of the anti-TRPV1 and anti-CGRP was previously verified (24, 25). SP primary antibody was validated by using Tac1 ^-/- mice. The sections were further incubated with appropriate secondary antibodies for two hours at room temperature. Tooth sections were stained with 4’,6-diamidino-2-phenylindole (DAPI) to visualize the cellular nuclei.

Maxillae were fixed overnight with 4% paraformaldehyde at 4°C and decalcified in 0.5M ethylenediaminetetraacetic acid (EDTA) for two weeks at room temperature. The decalcified tissues were dehydrated, embedded in paraffin, and sectioned into 5 µm slices. TRAP and alkaline phosphatase (ALP) staining was conducted using a commercial kit (#29467001; Fujifilm Wako Pure Chemical Corporation, Richmond, VA, USA), following the manufacturer’s protocol, with some modifications. Briefly, deparaffinized and rehydrated sections were incubated in acetate-tartaric acid buffer containing naphthol ASMX phosphate and Fast Red TR (TRAP substrate). The staining process was monitored until the color reaction became distinct. After washing, tissues were counterstained with the nuclear staining solution. In an experiment, ALP staining was performed together with TRAP staining, but it was not quantified. Digital images were scanned using an Aperio scanning system (Leica, Wetzlar, Germany). To quantify the numbers of osteoclasts in the interradicular alveolar bone surface, we selected images showing the tooth roots of both the first and second molars and the second and third molars. Dense, purple-colored, TRAP-stained multinucleated osteoclasts on the surface of bone between the roots from two teeth were counted. The number of osteoclasts from each region was normalized to the perimeter of the bone surface assessed, to calculate the number of osteoclasts per millimeter (TRAP+ cells/mm) of each sample.

In C57BL/6 mice anesthetized with ketamine/xylazine, fluoro-gold (FG) (Fluorochrome LLC, Denver, CO, USA) was injected into the gingiva around the maxillary first molar to retrogradely label the periodontal afferents in the TG as described previously (23). FG was dissolved in 0.9% saline at a concentration of 4%. A 50 μl Hamilton syringe was used to slowly inject 5 μl of tracer into five sites (1 μl per site) on the gingiva around the distobuccal groove, buccal groove, mesial groove, palatal groove, and the distopalatal groove of the first maxillary molar. On seven days after injection, the ligature was placed around the maxillary first molar. The unligatured side was used as control. After two weeks, the mice were euthanized by transcardial perfusion for further histological study. After performing IHC, the FG signal was identified under a gold filter (11006v2, Chroma, Bellows Falls, VT, USA). We used Image J to calculate the surface area of the cells. We followed the criteria for neuronal size classification (small, < 300 μm²; medium, 300 to 600 μm²; large, > 600 μm²) (22, 26).

Mice were anesthetized using ketamine/xylazine and transcardial perfusion was performed using >20 ml PBS to flush out the immune cells from the vasculature. To dissect out the gingival blocks, vertical incisions were made immediately anterior to the maxillary M1 and posterior to the M3, and horizontal incisions were made at the border of the gingiva on both the buccal and palatal sides. Gingival tissues from the unligatured side were used as controls. The gingival tissues were processed as previously described (4, 27). The gingival tissues were digested in a mixture of 3.2 mg/ml type IV collagenase (Invitrogen, 17104019) and 0.15 μg/ml DNase (Sigma, DN25) in media for 25 min at 37°C with gentle shaking. Then EDTA was added to a final concentration of 2 mM, and the solution was incubated for 5 min. The enzymes were prepared in a complete media containing RPMI with 2 mM L-glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin, and 10% fetal bovine serum. The enzyme-digested gingival tissues and media were mashed through a cell strainer with a pore size of 70 μm, and additional cold DNase media was added. The cell suspension was centrifuged at 4°C, at 314×g for 6 min, and the pellet was resuspended in 100 µl cold PBS containing 0.5% FBS. Gingival tissues and single cell suspensions obtained from two separate mice in the same experimental group were pooled for one flow cytometry experiment. Single cell suspensions obtained from two separate mice in the same experimental group were pooled for one flow cytometry experiment. CD45-PE, CD11b-BV421, and Ly6G-PE-Cy7 antibodies were used to identify neutrophils. Flow cytometry analysis was performed within 2 hours on Cytek Aurora flow cytometer (3 lasers: 405nm, 488nm, 640nm; Cytek Biosciences, San Jose, CA, USA). The spectral data were unmixed based on single color compensation beads controls and analyzed using FCS Express 7 software (De Novo Software, Pasadena, CA, USA). Neutrophils were defined by 7-AAD^-/CD45⁺/Ly6G⁺/CD11b⁺. Percentages of each cell population in live, single cells were calculated for comparisons between groups.

Gingival injections of drugs were performed using an insulin syringe with a 31G needle under isoflurane anesthesia. SP (Sigma-Aldrich, # S6883; 1 µg/site), QWF (Tocris, #6645; 2 µg/site), neurokinin A (NKA; MilliporeSigma, # 86933746; 1 µg/site) or vehicles (PBS or dimethyl sulfoxide) were injected into the proximal and distal areas of the gingiva around M2.

The mice were euthanized following anesthesia with ketamine/xylazine, and maxillae, including the gingiva, alveolar bone, and three molars, were dissected out. Whole tissues were ground in a buffer (210 µl; 50 mM Tris, 2 mM EDTA pH 7.5) containing a protease/phosphatase inhibitor cocktail (5872S; Cell Signaling Technology, Danvers, MA, USA) using a tissue grinder (Pyrex PTFE Pestle Tissue Grinder with Steel Shaft). After spinning down for 5 min at 1000 g at 4°C, the supernatant was collected and frozen on dry ice. Luminex multiplex cytokine assays were performed using Mouse Luminex Discovery Assay Kits (R&D Systems, Minneapolis, MN, USA) and a Luminex 100 Multi-analyte System (Luminex, Chicago, IL, USA).

All data are presented as mean ± SEM. The data were analyzed using Student’s t-tests, a one-way ANOVA, and a two-way ANOVA followed by post hoc assays, as indicated in the figure legends, using GraphPad Prism 7 (GraphPad Software, San Diego, California USA). Comparisons of distributions of neuronal sizes between two groups were performed using Mann-Whitney test. A value of P< 0.05 was considered statistically significant.

A subset of TG neurons retrogradely labeled from gingiva expressed SP, CGRP, and TRPV1 ( Figures 1A, B ). The FG-labeled gingival TG neurons were largely small- (52.4%) or medium- (42.2%) sized neurons with only a few large-sized neurons (5.4%) (Figure 1C). We previously showed that 23% of TG neurons retrogradely labeled from gingiva expressed CGRP, which is highly co-expressed with TRPV1 (23). Approximately 15% of the FG-labeled neurons express SP or TRPV1, and approximately 10% of the labeled neurons co-express SP and TRPV1. The majority of SP+, CGRP+, or TRPV1+ neurons were small-sized neurons ( Figure 1D ). FG+/SP+ neurons were significantly larger than FG-negative SP+ neurons (red). FG+/TRPV1+ neurons were also significantly larger than FG-negative TRPV1+ neurons (blue). In contrast, FG+/CGRP+ neurons were not significantly different from FG-negative CGRP+ neurons (green). However, the size distribution of FG+/SP+ and FG+/CGRP+ neurons were not significantly different. SP+ nociceptive afferents abundantly project to the gingiva and periodontal ligaments ( Figures 1E, F ). In contrast, SP was undetectable in Tac1 KO mice, confirming SP deficiency in this line. CGRP+ nociceptive terminals were also detected from periodontia ( Figure 1G ). These data suggest that SP+ or CGRP+ afferents represent a subpopulation of sensory neurons innervating periodontal tissues.

To determine functional roles of SP and CGRP, encoded by Tac1 and Calca gene respectively, we used Tac1 and Calca knockout mice for ligature-induced periodontitis experiments. Tac1 knockout (KO) mice (Tac1^-/- ) were viable and exhibited a hypoalgesia phenotype (28). We first examined ligature-induced bone loss in the Tac1 KO mice via µCT. Two weeks after placing the ligature, wild type (WT) mice showed remarkable alveolar bone loss in both the buccal and palatal sides as indicated by a decrease in alveolar bone height, as well as buccal bone plate disruption ( Figures 2A–C ). In contrast, Tac1 KO mice exhibited an obviously reduced loss of the alveolar bones compared to WT controls. Specifically, alveolar bone height reduction and buccal bone disruption were less severe, and the buccal bone plate showed fenestration rather than complete resorption on the crest ( Figures 2B, C ). Consequently, the distance from CEJ to buccal alveolar bone crest (B-Crest) was significantly less in Tac1 KO than WT, although the distance from CEJ to the most apical site of bone destruction (B-max) was not significantly different. Bone loss in palatal side was also significantly less in Tac1 KO than WT. In histological sections, bone loss was more prominent in WT than in Tac1 KO after ligature, whereas alveolar bone without ligature was comparable ( Figure 2D ). Different extent of bone loss between genotypes are unlikely due to the different extent of mucosal damage since cytokeratin-expressing gingival epithelia were comparable ( Figure 2E ).

We assessed if ligature-induced periodontitis produces changes in SP or TRPV1 expression in TG. The retrogradely labeled FG+ TG neurons from mice with ligature showed a slight but significant increase in neuronal sizes compared to controls ( Figure 2F ). The proportions of FG+/SP+ neurons, FG+/TRPV1+ neurons, or FG+/SP+/TRPV1+ neurons were not significantly different between the control and the ligature group ( Figure 2G ), suggesting that altered expressions of SP or TRPV1 in gingival afferents are not major contributors to the nociceptor regulation of periodontitis. Interestingly, the size distributions of FG+/SP+, FG+/TRPV1+, or FG+/SP+/TRPV1+ neurons were significantly larger in the ligature group than in controls, although the extent of such changes was modest ( Figure 2H ).

Since CGRP is another type of neuropeptide abundantly expressed in nociceptors, we therefore also examined the involvement of CGRP in ligature-induced bone loss by using Calca KO mice that are deficient of CGRP ( Figure 3 ). We found that ligature-induced bone loss in Calca KO mice was comparable to WT, suggesting that net effects of global knockout of CGRP may not significantly impact bone resorption in periodontitis.

In agreement with reduced bone loss, TRAP staining revealed that osteoclast numbers outlining the bone surfaces were also decreased in Tac1 KO mice in comparison to controls. These histological studies were conducted five days after ligature placement, when osteoclasts and bone resorption were more actively ongoing than at the two-week time point ( Figures 4A–D ). Osteoclasts activities were comparable between genotypes without ligature ( Figure 4E ). In ligature groups, ALP activities were observed along the bone surfaces in both Tac1 KO and WT mice in comparable extents, suggesting that Tac1 KO did not substantially affect the bone formation process ( Figure 4F ).

Given that the regulation of host responses by neuropeptides underlies the neural regulation of barrier tissue functions, we further investigated the role of Tac1 in the recruitment of immune cells to sites of periodontitis. We performed flow cytometry assays using cells collected from gingiva with and without ligature in WT and Tac1 KO mice. The proportion of total CD45+ leukocytes ( Figure 5A ) and neutrophils ( Figure 5B ) were significantly greater in periodontia from the ligature side than from the control side in WT mice. In comparison, the proportion of ligature-induced total CD45+ leukocytes and neutrophils were significantly reduced in Tac1 KO mice compared to WT ( Figures 5C, D ). Consistently, immunohistochemical labeling of CD45 in periodontia with ligatures also demonstrated limited recruitment of CD45+ cells in Tac1 KO mice relative to controls ( Figure 5 ). Furthermore, cytokines associated with inflammatory responses, such as tumor necrosis factor, interleukin 1β, and receptor activator of nuclear factor kappa- Β ligand (RANKL), were significantly lower in the periodontia of Tac1 KO mice than WT ( Figure 6 ). Altogether, these results indicate that Tac1 KO mice exhibit reduced host response in ligature-induced periodontitis and suggest that SP is a major neurogenic regulator of host immune responses and alveolar bone loss in periodontitis.

Since Tac1 encodes both SP and NKA and both are detected in human gingiva crevicular fluid (17, 29), we determined if local injection of exogenous SP or NKA into gingiva is sufficient to induce osteoclastic activities and gingival inflammation. Results showed that injections of SP alone without periodontal ligature increased the number of osteoclasts along the bone surfaces of the injected areas ( Figures 7A–C ). Moreover, infiltration of CD45+ cells was also increased in SP-injected areas compared to vehicle-injected areas ( Figures 7D–F ). In contrast, the injection of exogenous NKA did not induce the osteoclast activity ( Figures 8A–C ) or CD45+ cell infiltration ( Figures 8D, E ). These results support the idea that Tac1 KO phenotypes inducing the inflammatory responses and osteoclasts activation are dominantly mediated by SP, but not by NKA.

To further test the roles of exogenous SP in periodontitis, we examined ligature-induced bone loss in mice with local injections of SP or vehicle ( Figure 9 ). Consistently, the SP-injected group demonstrated more vigorous bone resorption compared to the Veh-injected group.

We also determined the effects of inhibiting SP receptors in the periodontium by performing gingival injections of QWF, a tripeptide SP antagonist, or vehicle with periodontal ligature ( Figures 10A–D ). Results showed that QWF injection significantly decreased bone loss compared to the vehicle injection, which is consistent with the effects of Tac1 KO.