Early NGF–TrkA–CGRP signaling-related changes in dorsal root ganglia are associated with mechanical hypersensitivity in a zymosan A-induced SKG model prior to overt arthritis Authors: Yumi Kawata, Hiroshi Funakoshi Published: 2026 ABSTRACT AbstractPain in axial spondyloarthritis (axSpA) may precede clinically detectable inflammation or persist despite adequate control of inflammation. However, the mechanisms underlying this early pain state remain incompletely understood. We investigated nerve growth factor (NGF)–TrkA–calcitonin gene-related peptide (CGRP) signaling in dorsal root ganglia (DRG) before the onset of overt clinical arthritis, using a zymosan A (ZyA)-induced SKG mouse model of inflammatory arthritis with SpA-like features. Female SKG mice received ZyA, with or without a prophylactic anti-NGF monoclonal antibody treatment. Mechanical sensitivity, NGF protein levels in skin tissue, and expression of CGRP, TrkA and phosphorylated TrkA in DRG neurons were evaluated on day 10. ZyA-treated mice developed mechanical hypersensitivity by day 10, while no increase in clinical arthritis scores was observed during this period.In DRG neurons, an altered size distribution of CGRP-positive neurons toward larger-diameter populations was observed in association with early pain-related behavioral changes. These changes may involve downstream TrkA signaling. Prophylactic anti-NGF treatment attenuated the decrease in mechanical thresholds and reduced CGRP-neuronal size distribution changes and TrkA signaling changes. These findings suggest that NGF-associated DRG sensitization contributes to early pain-related behavioral changes during a prodromal phase of ZyA-induced disease, before overt clinical arthritis is established. Our study supports a mechanistic link between early pain-related behavioral changes and NGF-associated DRG sensitization, providing a rationale for further investigation of NGF-targeted strategies for pain management in axSpA.Keywords: Axial spondyloarthritis, Residual pain, Mechanical hypersensitivity, Nerve growth factor, TrkA signaling, Calcitonin gene-related peptide, Dorsal root ganglion Introduction Spondyloarthritis (SpA) is a group of chronic inflammatory diseases affecting the axial skeleton. It includes axial SpA (axSpA) such as ankylosing spondylitis and peripheral subtypes (Sieper, 2009). Pain is often the earliest symptom at joints and entheses and remains inadequately controlled by analgesics (Sieper and Poddubnyy, 2017; van der Heijde et al., 2017) or biologic therapies (Al Mohamad et al., 2024). Even in patients achieving remission or low disease activity, persistent residual pain is frequently reported (D'Angelo et al., 2022; Webers et al., 2023). This refractory pain diminishes quality of life and functional capacity, representing a major unmet clinical challenge in axSpA. Despite growing clinical attention to pain in axSpA, the mechanisms underlying this pain remain unclear. Although few animal studies have examined pain during the early phase preceding overt clinical arthritis, focusing on this period may provide insight into pain mechanisms in SpA. Nerve growth factor (NGF) is a neurotrophin involved in neuronal survival, sensory plasticity, and pain modulation. Previous studies reported dynamic regulation of neurotrophin-related pathways following peripheral nerve injury (Funakoshi et al., 1995; Funakoshi et al., 1993). Insights from nerve-injury models demonstrate that NGF acting through its high-affinity receptor TrkA can induce ectopic calcitonin gene-related peptide (CGRP) expression in large-diameter neurons of the dorsal root ganglia (DRG), promoting a pro-nociceptive phenotype and mechanical allodynia (Nitzan-Luques et al., 2013; Yu et al., 2012). Although these findings are largely derived from nerve-injury models, whether NGF–TrkA–CGRP signaling is associated with early mechanical hypersensitivity in primary inflammatory arthritis remains unclear. In this study, we used the zymosan A (ZyA)-induced SKG-Jcl mouse model of inflammatory arthritis with SpA-like features and intervened before the onset of overt joint swelling. We assessed pain-related behaviors and molecular changes in DRGs to determine whether NGF–TrkA–CGRP signaling is associated with early mechanical hypersensitivity and whether preventive NGF neutralization modulates these responses. Methods Animals All animal experiments were approved by the Institutional Animal Care and Use Committee of Asahikawa Medical University (protocol numbers: R5–115-04, R6–090-02, R7–076) and conducted in accordance with national regulations and the ARRIVE guidelines. Female SKG-Jcl mice were purchased from CLEA Japan (Tokyo, Japan) and maintained under SPF conditions. Zymosan A (ZyA; Sigma-Aldrich, #Z4250-1G, St. Louis, MO, USA), a yeast β-glucan extract, was administered by a single intraperitoneal (i.p.) injection to induce inflammatory arthritis with SpA-like features in SKG mice (Jeong et al., 2018; Ruutu et al., 2012). Only female SKG mice (8 weeks old) were used because females develop more severe arthritis with lower variability than males (Keith et al., 2013; Keller et al., 2013), consistent with 3R principles (Russell et al., 1959). Induction of SpA-like symptoms and anti-NGF treatment Mice were randomly assigned to three groups and injected as follows: •Control: PBS (i.p.) + saline (subcutaneous injection at nape of neck, s.c.) Control: PBS (i.p.) + saline (subcutaneous injection at nape of neck, s.c.) •ZyA: ZyA (3 mg/mouse, i.p.) + saline (s.c.) ZyA: ZyA (3 mg/mouse, i.p.) + saline (s.c.) •ZyA + anti-NGF: ZyA (3 mg/mouse, i.p.) + anti-NGF monoclonal antibody (frunevetmab, 21 μg/mouse [1 mg/kg]) (s.c.) ZyA + anti-NGF: ZyA (3 mg/mouse, i.p.) + anti-NGF monoclonal antibody (frunevetmab, 21 μg/mouse [1 mg/kg]) (s.c.) All mice first received intraperitoneal injections, immediately followed by subcutaneous injections. Clinical arthritis score was assessed at baseline and endpoint using Sakaguchi's criteria (Sakaguchi et al., 2003). Frunevetmab (Solensia; Zoetis) is a felinized monoclonal antibody targeting NGF for the control of osteoarthritis pain in cats (U.S. Food and Drug Administration, 2022) and has been reported to cross-react with and neutralize murine NGF in vitro, confirming its activity against mouse NGF (Gearing et al., 2016). The antibody was provided by Zoetis Japan (Tokyo), aliquoted into 3 μL portions, stored at −80 °C, and diluted in sterile saline immediately before administration. Von Frey filament test Mechanical sensitivity was assessed using the up–down method to calculate the 50% paw withdrawal threshold (PWT) at the plantar surface of the hindpaw under blinded conditions (0.4–4.0 g; starting filament 0.6 g) (Chaplan et al., 1994; Dixon, 1965). PWT was primarily assessed using the right hindpaw, but the contralateral hindpaw was used if measurement was prevented by excessive movement. Mice were placed individually in chambers on a wire-mesh platform and acclimated for 30 min before testing. Tissue collection Mice were euthanized with pentobarbital sodium (150 mg/kg, i.p.; Kyoritsu Seiyaku, Tokyo, Japan) and perfused transcardially with ice-cold PBS. For ELISA, hindpaws were amputated approximately 1 cm proximal to the heel, the right hindpaw was degloved, and skin tissues were snap-frozen in liquid nitrogen and stored at −80 °C. For histological analysis, in the same animals, mice were sequentially perfused with 4% paraformaldehyde (PFA) in PBS, L4–L5 DRGs were dissected, post-fixed in 4% PFA, and cryoprotected sequentially in 10% and then 20% sucrose in PBS at 4 °C, each step performed overnight. ELISA Frozen skin tissues were pulverized in liquid nitrogen and homogenized in SDS-free radioimmunoprecipitation (RIPA) buffer supplemented with a protease inhibitor cocktail (Roche; #04693124001) according to the manufacturer's protocol. Homogenates were sonicated and centrifuged (15,000 rpm at 4 °C). Total protein concentration was determined using a BCA protein assay (Thermo Fisher Scientific; #23227) and NGF levels were normalized to total protein content. NGF concentrations were measured using the NGF Rapid ELISA Kit, which preferentially detects mature β-NGF with minimal cross-reactivity to proNGF (Biosensis; #BEK-2213-1P). Because samples were not subjected to acid dissociation to separate ligand–receptor complexes, the assay primarily reflects free (unbound) NGF rather than total NGF content and was therefore interpreted as a pharmacodynamic measure of target engagement. Absorbance was read at 450 nm using a SpectraMax iD3 microplate reader (Molecular Devices), and NGF concentrations were calculated using the GainData software (Arigo Biolaboratories). Immunofluorescence staining Cryoprotected DRGs were embedded in O.C.T. compound (Sakura Finetek; #4583) using a dry ice–hexane bath (FUJIFILM Wako; #082–00426) and stored at −80 °C. Cryosections (6 μm) were prepared using Kawamoto's film method with Cryofilm type 2C(9) (SECTION-LAB, Yokohama, Japan) (Kawamoto, 2003). Sections were fixed sequentially in Drysol (Kanto Chemical; #73492) and 4% PFA in PBS according to Kawamoto's film method, which includes this additional fixation step to improve tissue integrity and prevent section detachment during processing. Sections were then blocked with 5% normal donkey serum (Sigma-Aldrich; #S30-100ML) containing 0.5% Triton X-100 in PBS for 1 h at room temperature. Sections were incubated with primary antibodies against TrkA, phosphorylated TrkA (pTrkA), or CGRP, followed by Alexa Fluor–conjugated secondary antibodies. Sections were counterstained with DAPI and mounted for confocal imaging. Primary antibodies were goat anti-TrkA (1:1000; R&D Systems, AF1056), rabbit anti-pTrkA (Antibodies.com, #A50553–50; 1:200) and mouse anti-CGRP (Santa Cruz, #sc-57053; 1:50). Signal specificity, including that of nuclear pTrkA staining, was verified using negative control sections processed without primary antibodies. Image acquisition Images were acquired using a confocal laser-scanning microscope (LSM900, ZEISS) at 20× magnification using identical acquisition settings across groups. For each section, six optical planes were captured at 1 μm intervals and reconstructed into a 6 μm projection. Laser power, gain, and exposure were kept constant across groups to ensure comparability. All image acquisition was performed under blinded conditions. Quantification and statistics for imaging Fluorescence intensity and cell size were quantified using ImageJ (v1.54m, NIH). Images were acquired at 512 × 512 pixels (330.3 × 330.3 μm). Image processing included median filtering (radius = 0.5 pixel) and background subtraction (rolling-ball radius = 50 pixels), except for membrane analysis. Single-channel images were binarized using Otsu's method. For TrkA membrane analysis, eroded binarized images were subtracted from dilated images, and the mean intensity was measured. For pTrkA nuclear signal analysis, nuclear localization was identified based on co-localization with DAPI. Pseudocoloring (pTrkA, green; DAPI, red) was applied in RGB color space, and co-localized signals (yellow) were quantified as mean fluorescence intensity within overlapping nuclear regions. For cell size analysis, Feret diameters of neuronal somata in L4–L5 DRGs were measured. Size distributions were initially binned at 2 μm intervals, and percentages of cell numbers were calculated using 5 μm bins. For CGRP+ soma size–distribution analyses, images were obtained from two independent staining batches performed under identical experimental conditions. Because binarization thresholds were optimized separately for each batch, size distributions were analyzed after normalization within each staining set, and animals were treated as the statistical unit. Statistical analyses Data are presented as mean ± standard error of the mean (SEM) for bar graphs and as median with interquartile range (IQR) for box plots; whiskers indicate values within 1.5 × IQR. Comparisons among three or more groups were performed using the Kruskal–Wallis test followed by Steel–Dwass post hoc tests. Statistical analyses were conducted using EZR (version 1.61; based on R and R Commander, Jichi Medical University, Japan) (Kanda, 2013). Principal analyses were performed with the following sample sizes per group: ELISA (n = 6, 6, and 4 per group); 50% PWT (n = 11 per group); immunofluorescence for TrkA and pTrkA (n = 3 animals per group; two sections per animal); and CGRP-positive cell size-distribution analysis (n = 5 animals per group; two sections per animal). Results Mechanical hypersensitivity emerges prior to overt arthritis and is attenuated by prophylactic anti-NGF monoclonal antibody treatment To investigate molecular mechanisms relevant to SpA-associated pain, we used ZyA-treated SKG mice, an inflammatory arthritis model with some SpA-like features (Jeong et al., 2018). Joint swelling typically develops 2–3 weeks after injection. Analyses were therefore performed on day 10 after induction (Fig. 1A), a time point preceding overt clinical arthritis. No increase in clinical arthritis scores was observed during this period. Early mechanical hypersensitivity in ZyA-treated mice is attenuated by prophylactic anti-NGF treatment. Statistical analyses were performed using the Kruskal-Wallis test with Steel–Dwass post hoc comparisons (*p < 0.05, **p < 0.01). n = 11 animals per group. (A) Experimental timeline. Eight-week-old female SKG mice underwent baseline von Frey filament testing, group allocation, administration, follow-up testing, and tissue collection. ZyA or PBS was administered intraperitoneally (i.p.), followed immediately by subcutaneous (s.c.) injection of anti-NGF antibody or saline. (B) 50% paw withdrawal threshold (PWT) at day 0 and day 10. (C) Percentage change in 50% PWT from baseline (∆PWT%) in individual animals. Mechanical sensitivity was assessed using von Frey testing at baseline (day 0) and day 10. The 50% PWT was reduced in ZyA-treated mice at day 10 compared with baseline. Analysis of 50% PWT values showed no statistically significant differences between anti-NGF–treated mice and PBS-treated controls (Fig. 1B). Further, the percentage change in 50% PWT for each individual animal showed that the threshold reduction was attenuated by anti-NGF treatment (Fig. 1C). These results indicate that ZyA induced early mechanical hypersensitivity in this model and that prophylactic anti-NGF treatment attenuated the reduction in 50% PWT, although individual variability in pain sensitivity and treatment responsiveness was observed. Preventive anti-NGF treatment modulates CGRP-positive neuronal size distribution and TrkA-related signaling changes in DRGs The hindpaw skin represents a peripheral target of DRG neurons. While ZyA treatment did not alter free NGF protein levels in hindpaw skin, co-administration of anti-NGF reduced free NGF levels, confirming pharmacodynamic target engagement (Fig. 2A). Anti-NGF treatment attenuates ZyA-induced changes in CGRP-positive neuronal size distribution and TrkA–related signaling in dorsal root ganglia (DRGs). Statistical analyses were performed as in Fig. 1 (*p < 0.05, **p < 0.01, ***p < 0.001). Two sections per animal were analyzed for fluorescence analyses. The bar graph shows mean ± SEM. Box plots display medians, 25th–75th percentiles, and whiskers within 1.5 × IQR. (A) Free (unbound) NGF levels in hindpaw skin were quantified using ELISA (PBS: n = 6; ZyA: n = 6; ZyA + anti-NGF: n = 4). (B) Representative images of CGRP-positive DRG neurons (red) in L4–L5 DRGs. Scale bar, 30 μm. (C) Size distribution of CGRP-positive DRG neurons (2 μm bins). n = 5 animals per group. (D) Percentage of CGRP-positive DRG neurons with soma diameters ≥30 μm, quantified using 5 μm bins. n = 5 animals per group. (E) Representative images showing TrkA (red), phosphorylated TrkA (pTrkA; green), and DAPI (blue) in L4–L5 DRGs. Arrowheads indicate the plasma membrane regions used for TrkA analysis. Boxed regions indicate nuclei with pTrkA signal, with enlarged views. Scale bar, 30 μm. (F) Quantification of TrkA intensity on the plasma membrane. n = 3 animals per group. (G) Quantification of nuclear pTrkA/DAPI co-localization. n = 3 animals per group. Immunofluorescence analysis showed an increased proportion of CGRP-positive neurons among larger-diameter (≥30 μm) DRG neurons following ZyA treatment (Fig. 2B–D). The size-distribution histogram exhibited a bimodal pattern with a rightward shift toward larger-diameter neurons (Fig. 2C). This shift was attenuated by anti-NGF treatment. Using the animal as the statistical unit, the proportion of CGRP-positive neurons with soma diameters ≥30 μm was increased in ZyA-treated mice and attenuated by anti-NGF treatment (Fig. 2D). Immunofluorescence analysis of L4–L5 DRGs revealed increased plasma membrane localization of TrkA and increased nuclear pTrkA signals in ZyA-treated mice. These changes were modulated by anti-NGF treatment (Fig. 2E–G). Discussion In neuropathic pain models, the NGF–TrkA–CGRP axis has been implicated in mechanical allodynia and subsequent neuroimmune changes. Increased ectopic CGRP expression has been reported in large-diameter DRG neurons (Nitzan-Luques et al., 2013; Zou et al., 2016), and enhanced TrkA/pTrkA signaling that is sensitive to NGF neutralization has been described (Cheng et al., 2009). These processes are thought to involve ligand-induced TrkA phosphorylation, receptor internalization, and downstream membrane-to-nucleus signaling. Our findings integrate previously reported features of CGRP expression downstream of NGF–TrkA signaling within a single experimental framework and extend this literature by suggesting involvement of this signaling pathway at an early stage of disease in a primary inflammatory model lacking direct nerve injury. Although subtle inflammatory changes may already have been present, overt clinical arthritis was not yet evident. In psoriasis, a skin disease commonly associated with SpA (van der Heijde et al., 2017), CGRP contributes to skin inflammation (Kotlyar and Granstein, 2025). NGF-dependent CGRP release from DRG afferents could influence spinal processing and peripheral vasodilation, while immune cells in peripheral tissues may produce cytokines, further increasing NGF availability (Iyengar et al., 2017; Kim and Granstein, 2021). Retrograde NGF transport to DRG somata could amplify CGRP induction in neurons (Yu et al., 2012). Beyond the scope of the present study, skin–central nervous system crosstalk mediated by NGF signaling may represent an important area for future investigation in SpA. Here, we found that preventive NGF neutralization attenuated ZyA-induced molecular changes and mechanical hypersensitivity, supporting a role for altered CGRP-positive neuronal size distribution associated with NGF–TrkA signaling in early SpA-associated mechanical pain. ZyA exposure was associated with increased proportion of CGRP-positive neurons with soma diameters ≥30 μm and a bimodal shift toward larger-diameter populations, without changes in total cell number or fluorescence intensity. Because large-diameter DRG neurons are often associated with low-threshold mechanoreceptive pathways (Devor, 2009; Nitzan-Luques et al., 2011), these findings raise the possibility that normally innocuous touch circuits may become pro-nociceptive in this model. However, soma diameter alone is an imperfect classifier in this context. The observed histogram shift could reflect hypertrophy or transcriptional activation in pre-existing nociceptors rather than bona fide phenotypic conversion of mechanoreceptors. Further, exploratory analysis showed increased plasma membrane localization of TrkA and increased nuclear pTrkA signals, supporting possible involvement of NGF–TrkA–CGRP signaling in this model. Evidence for NGF/TrkA-dependent CGRP induction specifically in large-diameter neurons remains limited. Therefore, altered CGRP-positive neuronal size distribution and potentially related TrkA/pTrkA changes observed in our model should be interpreted as preliminary findings requiring further validation. NGF levels in skin tissue did not increase following ZyA administration. Because the ELISA was performed without acid dissociation to separate ligand–receptor complexes, the assay primarily reflects free (unbound) NGF. Therefore, we cannot exclude the possibility that total NGF levels in the tissue may have increased. These findings suggest that activation of NGF–TrkA signaling in this model may not depend on a detectable increase in free NGF concentration in the skin. Rather, subtle changes in local NGF availability, receptor accessibility, or the composition of NGF-producing cell populations may contribute to enhanced TrkA signaling in DRG neurons. Frunevetmab is a felinized monoclonal antibody approved for veterinary use that cross-reacts with and neutralizes mouse NGF (Gearing et al., 2016). Because pharmacokinetic data in mice are limited, we adapted a single-dose strategy equivalent to the clinically used per-kilogram dose in cats. Although penetration of large immunoglobulin G into DRGs is limited, systemic IgG reaches DRGs at low but detectable levels and can also interact with resident cells (Goebel et al., 2021; Iijima and Iwasaki, 2016; Lund et al., 2024). These results suggest that NGF-targeted therapy could help address the unmet need for inflammation-refractory pain in SpA. Candidate populations may include patients with non-radiographic axSpA (nr-axSpA), who frequently experience predominant pain without overt radiographic changes (Almansouri et al., 2024; Kiltz et al., 2012; Mogard et al., 2021), as well as those with pain refractory to anti-inflammatory agents. Anti-NGF antibodies, such as tanezumab, have shown efficacy in phase III trials for osteoarthritis and chronic low back pain (Markman et al., 2020), though concerns about rapidly progressive osteoarthritis (RPOA) have prevented their approval (Berenbaum et al., 2020; Karsdal et al., 2019). The pathophysiology of RPOA is unclear, and its relevance to SpA is uncertain; therefore, next-generation anti-NGF agents with improved safety profiles are under development (Koya et al., 2023). This study has several limitations that should be acknowledged. First, the anti-NGF antibody was administered prophylactically; therefore, therapeutic efficacy after pain establishment remains unknown. Second, sample sizes for DRG immunofluorescence analyses, particularly for TrkA/pTrkA-related endpoints, were modest; therefore, the associated findings should be interpreted cautiously and require further validation in larger studies. Finally, several important evaluations were beyond the scope of the present study and warrant investigation in future studies, including interactions with other neurotrophins (brain-derived neurotrophic factor, neurotrophin-3, or neurotrophin-4), glial activation in the spinal cord and DRG, electrophysiological assessments, and additional behavioral assessments such as thermal hypersensitivity, cold allodynia, and spontaneous pain. Conclusion Our findings suggest that early mechanical hypersensitivity in the ZyA-induced SKG mouse model is associated with NGF–TrkA–related signaling changes and alter size distribution of CGRP-positive neurons, and that preventive NGF neutralization attenuates these responses. These findings support further investigation of NGF-targeted strategies for pain management in SpA. CRediT authorship contribution statement Yumi Kawata: Writing – review & editing, Writing – original draft, Visualization, Validation, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Hiroshi Funakoshi: Writing – review & editing, Supervision, Conceptualization. Consent for publication All authors approved the manuscript for publication. Ethics approval and consent to participate All animal experiments were approved by the Institutional Animal Care and Use Committee of Asahikawa Medical University and were conducted in accordance with national regulations and the ARRIVE guidelines. This study adhered to the 3Rs principle (Russell et al., 1959). Funding This work was supported by institutional education and research funds of the Department of Advanced Medical Science, Asahikawa Medical University. The funder had no role in study design, data collection, data analysis, interpretation of data, or writing of the manuscript. Declaration of competing interest The authors declare that they have no competing interests. Acknowledgements The authors are grateful to Dr. Hirotsugu Kanda, Dr. Megumi Kanao-Kanda, and Dr. Izumi Sato for their valuable technical advice on von Frey filament testing and for generously providing the equipment necessary for the experiments. We also express our gratitude to Ms. Kaori Segawa for her excellent administrative support, and to Dr. Jun Ueda for his general encouragement. This manuscript was submitted in partial fulfillment of the requirements for the Ph.D. degree of Y.K. Contributor Information Yumi Kawata, Email: yumik@asahikawa-med.ac.jp. Hiroshi Funakoshi, Email: hfuna3@outlook.com. Data availability The datasets are available from Y.K., one of the corresponding authors, upon reasonable request. References Associated Data Data Availability Statement The datasets are available from Y.K., one of the corresponding authors, upon reasonable request.