γδ T cells exacerbate intestinal fibrosis in mice with experimental colitis via activation of the CD73-adenosine receptor-cyclic AMP signaling pathway

Abstract

BACKGROUND

Intestinal fibrosis is a prevalent complication of chronic inflammatory bowel disease, affecting approximately one-third of patients with Crohn’s diseases. It represents a leading cause of stricture formation, luminal obstruction, and surgical intervention. Despite its clinical significance, no approved anti-fibrotic therapies are currently available, and the underlying pathogenic mechanisms remain incompletely understood. Although γδ T cells, a major population of innate-like T lymphocytes in the intestine, have been implicated in pulmonary and liver fibrosis, their functional role in intestinal fibrosis has not been elucidated.

AIM

To investigate how CD73 modulates the innate immune function of γδ T cells in a mouse model of experimental intestinal fibrosis.

METHODS

Intestinal fibrosis was induced in mice using chronic 2,4,6-trinitrobenzenesulfonic acid administration. The pathogenic role of γδ T cells was assessed via antibody-mediated depletion (GL3) and adoptive transfer. The involvement of the CD73-adenosine receptor-cyclic AMP (cAMP) axis was validated through pharmacological intervention using specific inhibitors (PSB-12379, SQ22536) and activators (forskolin).

RESULTS

In the 2,4,6-trinitrobenzenesulfonic acid-induced intestinal fibrosis model, antibody-mediated depletion of γδ T cells ameliorated fibrosis, evidenced by reduced clinical scores, decreased extracellular matrix deposition, attenuated pro-fibrotic cytokine levels, and elevated anti-fibrotic cytokines. Conversely, reinfusion of γδ T cells significantly exacerbated intestinal fibrosis. Mechanistically, γδ T cell depletion downregulated CD73 expression in fibrotic mice, and pharmacological CD73 inhibition similarly attenuated fibrosis. Interventions targeting the CD73-cAMP signaling pathway using cAMP activators and antagonists correspondingly attenuated or exacerbated intestinal fibrosis.

CONCLUSION

This study identifies γδ T cells as pivotal drivers of intestinal fibrosis via the upregulation of the CD73-adenosine receptor-cAMP signaling pathway. Targeting this axis offers a promising therapeutic strategy for fibrostenotic Crohn’s disease.

Key Words: Intestinal fibrosis; γδ T cells; CD73; Adenosine receptor; Cyclic AMP

Core Tip: This study elucidated a critical pathogenic role for γδ T cells in driving intestinal fibrosis. Mechanistically, these cells promote fibrosis most likely by upregulating the CD73-adenosine receptor-cyclic AMP signaling pathway. Depletion of γδ T cells or pharmacological inhibition of CD73 attenuated fibrogenesis, as evidenced by reduced collagen deposition and profibrotic cytokine expression. Conversely, adoptive transfer of γδ T cells exacerbated disease severity. These findings nominate the γδ T cell-CD73-adenosine receptor-cyclic AMP pathway as a novel and promising therapeutic target for intestinal fibrosis.

INTRODUCTION

Intestinal fibrosis is a frequent and debilitating complication of chronic inflammatory bowel disease (IBD), affecting approximately one-third of patients with Crohn’s disease[1] and representing the primary driver of stricture formation and surgical resection. Up to 70% of patients require surgery within 10 years of diagnosis, yet no approved anti-fibrotic therapies currently exist[2], underscoring the necessity to study the underlying pathogenic mechanisms.

Pathologically, intestine fibrosis is characterized by the excessive accumulation of extracellular matrix (ECM) components, disrupting tissue architecture and increasing malignant transformation risk[3-5]. While chronic inflammation, immune dysregulation, and dysregulated tissue repair are recognized as key drivers of fibrosis[6], including the activation of mesenchymal cells responsible for collagen I/III deposition[7], the precise contributions of specific immune subsets, particularly innate-like lymphocytes abundant in the intestinal mucosa, to profibrotic signaling remain incompletely defined. Despite evidence that targeted pathway inhibition can mitigate fibrosis[1,8], yet effective interventions for intestinal fibrosis remain elusive.

γδ T cells, an innate-like T cell subset, are abundant in the intestinal mucosa, comprising up to about 40% of intestinal T cells[9,10]. Functionally bridging innate and adaptive immunity, γδ T cells play critical roles in mucosal homeostasis, epithelial repair, and barrier integrity following injury[11-15]. However, their role in fibrosis is context-dependent. Some subsets of γδ T cells promote inflammation and fibrogenesis via interleukin (IL)-17A production[12,13,16,17], while others may resolve fibrosis through secretion of IL-22[18] and interferon γ (IFN-γ)[16]. Although extensively studied in pulmonary and hepatic fibrosis[18], the function of γδ T cells in intestinal fibrosis remains elusive. Some studies suggested that γδ T cells was the causative factor in rats with colitis, and human IBD associated studies suggest that the abundance, subsets and secreted cytokines of γδ T cells undergo changes[19-23]. Emerging evidence suggest CD73 was correlated with pathogenic Th17 cells in IBD[23-25]. However, controversary has been found as the role of γδ T cells varies under different context in IBD[20]. Given their abundance and immunomodulatory capacity, defining the direct role of γδ T cells in intestinal fibrogenesis is essential for developing novel immunotherapies.

Inflammation has been well-studied for its role in tissue injury and immune pathology, but also contributes to the protection by removing detrimental stimuli and initiating tissue healing[26]. Among the regulators, adenosine, a purine nucleoside generated from extracellular ATP during tissue stress or injury, serves as a key immunomodulatory signal. Its production is tightly regulated by ectonucleotidases CD39 (which hydrolyzes ATP/ADP to AMP) and CD73 (5’-ectonucleotidase, NT5E), which converts AMP to adenosine[27]. Adenosine then signals through four G protein-coupled adenosine receptors (ARs, including A1R, A2AR, A2BR, A3R), modulating inflammation, immunosuppression, and fibrotic responses[28-30]. Notably, A2AR plays a crucial role in the regulation of inflammatory patterns via increasing the intracellular cyclic AMP (cAMP) levels, which activate the downstream PKA/CREB signaling pathway[31]. Importantly, cAMP levels can feedback to regulate CD39-CD73-adenosine signaling[31]. Immune cells, including γδ T cells[13], are highly expressed ARs and adenosine[32]. Elevated levels of CD73 have been observed in intestinal tissues from patients with fibrostenotic Crohn’s disease[33]. and adenosine signaling via A1R/A2AR has been shown to ameliorate colitis in murine models[34]. Nevertheless, whether CD73-driven adenosine production contributes to fibrogenesis, and whether this process is orchestrated by specific immune populations such as γδ T cells, remains unexplored.

Here, we hypothesized that γδ T cells promote intestinal fibrosis by upregulating the CD73-adenosine signaling pathway, and that targeting this axis could represent a novel therapeutic strategy. To test this hypothesis, we employed a 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced mouse model of colitis-associated intestinal fibrosis pursued the following specific aims: (1) To determine whether γδ T cells are causally involved in fibrosis development using antibody-mediated depletion and adoptive transfer approaches; (2) To investigate whether CD73 expression and activity are modulated by γδ T cells during fibrogenesis; (3) To assess the therapeutic potential of pharmacological CD73 inhibition; and (4) To dissect the downstream signaling mechanisms by targeting the cAMP pathway. Our findings demonstrated that γδ T cells promoted intestinal fibrosis primarily through the CD73-AR-cAMP signaling pathway. Critically, both pharmacological inhibition of CD73 and antibody-mediated depletion of γδ T cells therapeutically suppressed fibrosis, identifying this axis as a novel therapeutic target.

MATERIALS AND METHODS

Animals

Female C57BL/6J mice, aged 7-8 weeks, were sourced from the Hubei Provincial Laboratory Animal Research Center. Housing and care occurred at the Hainan Provincial Experimental Animal Center under specific pathogen-free conditions. Animals were maintained at ambient room temperature with a standardized 12-hour light/dark cycle, humidity of 50% ± 10%, and free access to standard chow and water. After one week of acclimatization, mice were randomly assigned to experimental groups.

At the start of the experiments (6-8 weeks of age), the mice in all groups had a comparable average body weight, which was within the normal range (18-20 g). All mice were weight-matched at the beginning of the study to minimize baseline variation. After the treatment, weight changes, watery stool and rectal bleeding were recorded every week.

TNBS-induced intestinal fibrosis

Mice were fasted (with free access to water) for 24 hours. Subsequently, the mice were anesthetized with ether, and a 3.5F catheter was inserted approximately 5.5 cm into the colon via the anus. Each mouse received an enema of 100 μL TNBS (Sigma, P2291-10 mL, MA, United States)/50% ethanol solution. Following administration, the mice were held in an inverted position for 60 seconds to prevent reflux of the solution. This procedure was repeated once weekly for 6 consecutive weeks to induce chronic inflammation and fibrosis, with TNBS doses progressively increased as follows: 1.5 mg TNBS per dose at week 1-2, 2.0 mg TNBS per dose at week 3-4, 2.5 mg TNBS per dose at week 5-6.

Following the completion of the 6-week enema protocol, mice were received intraperitoneal injections of ZM241385 solution (MCE, 22530, NJ, United States), PSB-12379 solution (MCE, 34427, NJ, United States), forskolin (FSK) solution (MCE, 118191, NJ, United States), and SQ22536 solution (MCE, 22969, NJ, United States), respectively (200 μL per mouse). Injections were administered twice weekly for 4 consecutive weeks.

Disease activity index analysis

Mice were carefully monitored daily during the experimental period, including weight loss, stool consistency, and presence of blood in the stool. Scores ranged from 0 (healthy) to 4 (severe colitis). The following parameters are the bases of the measurement: (1) Weight loss (0, less than 1%; 1, 1%-5%; 2, 5%-10%; 3, 10%-15%; 4, > 15%); (2) Stool consistency (0, normal; 2, mushy; 4, diarrhea); and (3) Rectal bleeding (0, negative; 2, positive; 4, visible rectal bleeding). The disease activity index (DAI) was calculated according to a standard scoring system[35].

Depletion and reinfusion of γδ T cell

The animal was administrated with GL3 antibodies [200 μg in 200 μL in sterile phosphate buffered saline (PBS)] by intraperitoneal injection. Antibodies against murine TCR-GL3 were purchased from BD Biosciences (La Jolla, CA, United States). Then γδ T cells were detected in the cells from spleen and lymphoid by fluorescence-activated cell sorting (FACS). The γδ T cells in the intestinal tissue were detected by immunofluorescence staining (anti-TCR, Invitrogen, 14-5711-82, CA, United States). Depletion was initiated one week before the first TNBS administration and repeated every two weeks throughout the experiment.

Depletion efficiency was also assessed at the experimental endpoint by immunohistochemical staining for TCRδ on frozen colon sections. Briefly, 5-μm colon sections were fixed in cold acetone, blocked with 10% normal goat serum, and incubated overnight at 4 °C with anti-TCRδ antibody (clone GL3, 1:100, BioLegend, Cat 118102, CA, United States). Images were captured using a Nikon Eclipse 80i fluorescence microscope at 200 × magnification. TCRδ-positive cells were quantified in five randomly selected high-power fields per section using ImageJ software. Depletion efficiency was calculated as the percentage reduction [mean optical density (OD)] in TCRδ-positive cells compared to TNBS-treated controls.

γδ T cells enrichment

γδ T cells were isolated from 8-10-week-old female mice (same strain and sex as experimental animals)[13,36]. Following aseptic splenectomy, tissues were maintained in ice-cold PBS prior to mechanical dissociation for splenocyte preparation. γδ T cell enrichment was achieved through negative selection using the EasySep™ Biotin Positive Selection Kit (StemCell, Canada) targeting CD11b+, B220+, CD4+, and CD8+ lineages. The magnetic separation protocol initiated with a 15-minute room temperature incubation of splenocytes with biotinylated antibodies against CD11b, B220, CD4, and CD8 (each at 1:100 dilution, 10 μL/mL concentration). Cells were subsequently washed with 10 mL of PBS containing 0.526 mmol/L EDTA and 2% foetal bovine serum, followed by centrifugation at 1400 × g for 5 minutes at 4 °C. After supernatant removal, the pellet was resuspended in the same buffer at 1 × 10⁸ cells/mL. The EasySep™ Biotin Selection Cocktail (100 μL/mL) was introduced, mixed thoroughly, and incubated for 15 minutes at room temperature. Following brief vortexing (15-30 seconds), EasySep™ Magnetic Nanoparticles (50 μL/mL) were added to the suspension and incubated for 10 minutes at room temperature. Tubes were then placed in EasySep™ magnets for 5 minutes at ambient temperature, with the unbound cellular fraction collected for downstream processing. Enriched cells were quantified before fluorescence-activated cell sorting. For sorting, cells underwent staining for 30 minutes at 4 °C with the following antibody panel: Anti-γδ TCR-Alexa Fluor 488, anti-CD8-PE, anti-CD4-PerCP-Cy5.5, anti-CD11b-APC, and anti-B220-PE-Cy7 (all at 1:100 dilution). Post-staining washing with PBS was performed, with final resuspension in sort buffer at 5 × 10⁶ cells/mL prior to FACS isolation.

Hematoxylin and eosin staining

Following CO₂ asphyxiation and confirmatory cervical dislocation, the entire colon was excised. Luminal contents were cleared by irrigation with PBS. Tissues were immersion-fixed in 4% paraformaldehyde (Sinopharm Group, Cat 80096618, Beijing, China) for 24 hours at 4 °C. Fixed colons underwent standard dehydration, were paraffin-embedded, and sectioned at 5 μm. Sections were mounted and stained with hematoxylin and eosin for comprehensive histopathological evaluation.

Immunohistochemistry

Paraffin-embedded colon sections were subjected to deparaffinization in xylene followed by rehydration through a graded ethanol series. Antigen retrieval was performed to unmask epitopes. Sections were incubated with primary antibodies overnight at 4 °C. After washing, species-matched secondary antibodies were applied. Immunoreactivity was visualized using a chromogenic detection system according to the manufacturer’s protocol. Finally, sections were counterstained with Mayer’s hematoxylin, dehydrated, cleared, mounted with resinous medium, and imaged under bright-field microscopy. Antibody information was the following: Anti-rabbit IL17A (18 KD) (Affinity, DF6127, 1:1000, RRID AB_2838094, OH, United States), anti-rabbit IFN-γ (19 KD) (Affinity, DF6045, 1:1000, RRID AB_2838015, OH, United States), anti-rabbit IL10 (19 KD) (Affinity, DF6894, 1:1000, RRID AB_2838853, OH, United States), anti-rabbit transforming growth factor-beta1 (TGF-β1, 44/15 KD) (Affinity, AF1027, 1:1000, RRID AB_2835389, OH, United States).

Immunohistochemistry (IHC) staining was semi-quantified by measuring the mean OD, which is defined as the integral OD per unit area within the region of interest. This parameter reflects the average concentration of the chromogenic product (brown 3,3’-diaminobenzidine), thereby providing an indirect measure of target antigen expression. For each sample, the mean OD was calculated from 5-8 randomly selected high-power fieldsusing ImageJ software.

Masson’s trichrome staining

Collect colorectal tissue samples from intestine fibrosis mice. According to the guideline (kit, Baso, BA4079, China), perform the staining. Fix the tissues in Bouin’s solution for 24 hours at room temperature. Stain the sections with Biebrich scarlet-acid fuchsin solution for 10 minutes. Rinse in distilled water. Differentiate in phosphotungstic/phosphomolybdic acid solution for 10 minutes. Stain the sections with aniline blue solution for 5 minutes.

Enzyme-linked immunosorbent assay

Blood samples were collected via orbital draw blood at the time of euthanasia. Serum was sitting for 30 minutes and separated by centrifugation at 3000 rpm for 10 minutes and stored at -80 °C until analysis. For tissue sample, rinse the tissue with pre-cooled PBS to remove residual blood, then weigh and mince the tissue. Transfer the minced tissue into a tissue grinder along with the corresponding volume of PBS (typically at a 1:9 weight-to-volume ratio, e.g., 1 g of tissue sample to 9 mL of PBS). Grind thoroughly on ice. Finally, centrifuge the homogenate at 3000 rpm for 10 minutes and collect the supernatant for analysis. Levels of cytokines or tissue supernatant [Mouse ATP Kit (Meimian; MM-43789M2, China), Mouse PC III Kit (Jonlnbio; JL20169, China), Mouse IV-C Kit (Meimian; MM-45054M2, China), Mouse LN Kit (Meimian; MM-0161M2, China), Mouse HA Kit (Cusabio; CSB-E08121m, China), Mouse cAMP Kit (Meimian; MM-0544M2, China)] were measured using enzyme-linked immunosorbent assay kits according to the manufacturer’s instructions.

5’-nucleotidase activity assay kit

For tissue sample, rinse the tissue with pre-cooled PBS (0.01 M, pH = 7.4) to remove residual blood, then weigh and mince the tissue. Transfer the minced tissue into a tissue grinder along with the corresponding volume of PBS (typically at a 1:9 weight-to-volume ratio, e.g., 1 g of tissue sample to 9 mL of PBS). Grind thoroughly on ice. Finally, centrifuge the homogenate at 3000 rpm for 10 minutes and collect the supernatant for analysis (Solarbio, BC4595, Beijing, China) according to the manufacturer’s instructions.

Hydroxyproline content assay kit

For tissue sample, rinse the tissue with pre-cooled PBS (0.01 M, pH = 7.4) to remove residual blood, then weigh and mince the tissue. Transfer the minced tissue into a tissue grinder along with the corresponding volume of PBS (typically at a 1:9 weight-to-volume ratio, e.g., 1 g of tissue sample to 9 mL of PBS). Grind thoroughly on ice. Finally, centrifuge the homogenate at 3000 rpm for 10 minutes and collect the supernatant for analysis (Solarbio, BC0250, Beijing, China) according to the manufacturer’s instructions.

Western blotting

Protein expression was analyzed by western blotting. Total protein was extracted from homogenized ileum tissues using RIPA lysis buffer (Beyotime Biotechnology, P0013B, China). Protein concentration was quantified using a BCA assay kit (Beyotime Biotechnology, P0010, China). Samples (40 μg protein/Lane) were resolved by sodium-dodecyl sulfate gel electrophoresis and electrotransferred onto polyvinylidene fluoride membranes (Millipore, IPVH00010; ISEQ15150, MA, United States). Membranes were blocked with 5% BSA or non-fat dry milk in TBST for 1 hour at room temperature, followed by overnight incubation at 4 °C with primary antibodies. After three TBST washes, membranes were incubated for 1 hour at room temperature with species-matched HRP-conjugated secondary antibodies. Following three additional TBST washes, protein bands were visualized using ECL chemiluminescent substrate. Relative protein expression levels were determined by quantifying band intensities with ImageJ software, normalized to β-actin.

Antibody information was the following: Anti-mouse monoclonal β-actin [Proteintech (Wuhan Sanying), 66009-1-Ig, 1:2000, China], CD73 [anti-rabbit, Proteintech (Wuhan Sanying), 12231-1-AP, 1:2000, China], HRP-conjugated secondary antibodies [anti-mouse protein antibodies (Bosterbio; BA1051, China), goat anti-rabbit antibodies as the secondary antibodies (Bosterbio; BA1054, China)].

RNA extraction and quantitative polymerase chain reaction

Total RNA was isolated from ileum tissues using TRIzol reagent (Ambion, Cat 15596-026, TX, United States). Quantitative polymerase chain reaction (qPCR) amplification was performed on a QuantStudio 6 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific, MA, United States) with SYBR Green master mix (Vazyme, Cat Q111-02, China). Transcript levels were quantified via the comparative ΔΔCt method, normalized to endogenous reference genes and expressed relative to untreated controls. Primer sequences are detailed in Table 1.

Table 1 Primer sequences for quantitative polymerase chain reaction.

Gene	Primer	Sequence (5’-3’)	PCR products
Mus β-actin	Forward	CACGATGGAGGGGCCGGACTCATC	240 bp
	Reverse	TAAAGACCTCTATGCCAACACAGT
Mus CD73	Forward	CAGCGATGACTCCACCAAGT	144 bp
	Reverse	CAGATGGTGCCCTGGTACTG

PCR: Polymerase chain reaction.

Statistical analysis

Statistical analysis was performed with Prism software (version 8, GraphPad). Data are presented as means ± SD. Normality was assessed using the Shapiro-Wilk test. For comparisons between two groups, an unpaired two-tailed Student’s t-test was used for normally distributed data, while the Mann-Whitney U test was applied for data that did not meet normality assumptions. For multiple group comparisons, one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was used for normally distributed data with equal variances (assessed by Brown-Forsythe test); for non-normally distributed data, the Kruskal-Wallis test followed by Dunn’s post hoc test was employed. Longitudinal data (body weight changes and DAI scores over time) were analyzed using two-way repeated measures ANOVA with Bonferroni’s post hoc test. A P-value < 0.05 was considered statistically significant (^aP < 0.05, ^bP < 0.01, ^cP < 0.001).

RESULTS

Depletion of γδ T cells prevented intestinal fibrosis

To investigate the role of γδ T in intestinal fibrosis, we intraperitoneally injected the GL3 antibody to deplete γδ T cells in a well-established murine model of fibrosis[37]. Intestinal fibrosis was induced using TNBS dissolved in 50% ethanol[37-40], which disrupts the mucosal barrier and enhances TNBS penetration. Mice were divided into three groups: TNBS-treated mice with or without GL3 antibody depletion and a control group. Disease severity was assessed by monitoring weight loss, rectal bleeding, colon length, and histological activity index.

The experimental procedure was shown in Figure 1A. First, we confirmed successful depletion of the γδ T cells via immunostaining (Figure 1B). Weight loss began after five TNBS enema administrations, regardless of GL3 treatment. However, while mice treated with TNBS alone exhibited progressive weight loss (reaching -5% by week 14), GL3-pretreated mice began regaining weight at week 10 and ultimately showed a 7% increase by week 14 (Figure 1C). Additionally, TNBS-alone mice displayed more severe clinical symptoms, including watery stool and rectal bleeding. The DAI score peaked at 8 in TNBS-treated mice but remained significantly lower (maximum of 5) in GL3-pretreated mice (Figure 1D), indicating that γδ T cell depletion ameliorated colon lesions.

Figure 1 Depletion of γδ T cells prevented intestinal fibrosis. A: Experimental scheme of 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced intestinal fibrosis model. γδ T cells were depleted using GL3 antibody before TNBS administration; B: TCR staining of the colon tissues after GL3 antibody administration; C: Body weight changes in the mice administrated with TNBS or TNBS + GL3 during the whole experimental period (n = 5); D: Disease activity index score evaluated in the mice administrated with TNBS or TNBS + GL3 (n = 5); E: The representative image for hematoxylin and eosin and Masson staining of colon in each group. Orange arrow indicated areas of immune cell infiltration. Scale bars: 100 μm; F: The concentrations of hydroxyproline levels in the serum of each group at week 14 (n = 3); G: The levels of hyaluronic acid, type IV collagen, laminin, and procollagen III in each group at week 14 (n = 3). Data were presented as mean ± SD. ^bP < 0.01, ^cP < 0.001. TNBS: 2,4,6-trinitrobenzenesulfonic acid; DAI: Disease activity index; H&E: Hematoxylin and eosin; HA: Hyaluronic acid; IV-C: Type IV collagen; LN: Laminin; PCIII: Procollagen III.

TNBS administration also significantly reduced colon length compared with controls, whereas GL3 pretreatment partially preserved colon length (Supplementary Figure 1), suggesting protection against intestinal damage. Histopathological analysis revealed greater immune cell infiltration and mucosal thickening in TNBS-alone mice compared to GL3-pretreated mice (Figure 1E). Importantly, Masson’s trichrome staining demonstrated that GL3 pretreatment reduced ECM deposition in TNBS-treated mice (Figure 1E). Consistent with Masson staining, hydroxyproline levels, a marker of collagen deposition, were also increased in fibrotic mice but reduced with GL3 treatment (Figure 1F). Furthermore, enzyme-linked immunosorbent assay revealed elevated levels of fibrosis-associated ECM components [e.g., hyaluronic acid (HA), type IV collagen (IV-C), laminin (LN), and procollagen III (PCIII)] in TNBS-treated mice, which were significantly attenuated by GL3 pretreatment (Figure 1G). In summary, these findings demonstrate that γδ T cells contribute to the progression of intestinal fibrosis.

Reinfusion of γδ T cells aggravated intestinal fibrosis

To further elucidate the role of γδ T cells in intestinal fibrosis, we performed adoptive transfer of γδ T cells into TNBS-treated mice. Mice were divided into four groups, all receiving TNBS to induce fibrosis. At week 6, γδ T cells were isolated, purified, and reinfused into the mice, with or without GL3 antibody pretreatment, to assess whether γδ T cell-mediated effects could be blocked.

The experimental procedure was outlined in Figure 2A. First, γδ T cells were isolated and enriched via FACS (Supplementary Figure 2A and B). TCR staining confirmed that γδ T cell reinfusion significantly increased intestinal γδ T cell numbers compared with TNBS-alone mice, whereas GL3-mediated depletion markedly reduced their abundance. Quantification revealed that TNBS treatment significantly increased γδ T cell infiltration into the colon relative to control mice (Figure 2B). GL3 antibody administration profoundly reduced intestinal γδ T cells by 87.2% in the GL3-only group and 72.4% in the GL3 + γδ T group, respectively (Supplementary Figure 2C). In the GL3 + γδ T group, γδ T cell levels were slightly but not significantly higher than in the GL3-only group (Figure 2B, Supplementary Figure 2B).

Figure 2 Reinfusion of γδ T cells aggravated intestinal fibrosis. A: Experimental scheme of 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced intestinal fibrosis model. γδ T cells were depleted using GL3 antibody before TNBS administration. For reinfusion experiment, γδ T cells were reinfused via intraperitoneally injection after TNBS administration; B: TCR staining of the colon tissues following TNBS treatment in each group at week 14. Scale bars: 50 μm; C: Body weight changes in the mice in each group during the experimental period (n = 5); D: Disease activity index score evaluated in the intestinal fibrosis treated with GL3 antibody administration or γδ T cells reinfusion (n = 5); E: Colon morphologies in each group at the end of week 14; F: The representative image for hematoxylin and eosin and Masson staining of colon in each group. Orange arrow indicated areas of immune cell infiltration. Scale bars: 100 μm. Data were presented as mean ± SD. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. TNBS: 2,4,6-trinitrobenzenesulfonic acid; DAI: Disease activity index; H&E: Hematoxylin and eosin.

Notably, mice receiving γδ T cell reinfusion exhibited significantly exacerbated disease severity compared to TNBS-alone mice. While TNBS-alone mice showed moderate weight loss (-2.1% by week 13), γδ T cell-reinfused mice displayed sustained and more severe weight loss (-14.3% by week 13) (Figure 2C). In contrast, GL3-pretreated mice gained weight (6.3% by week 13), and GL3 pretreatment partially mitigated weight loss in γδ T-reinfused mice (-7.6%). DAI scores followed a similar trend, with γδ T cell-reinfused mice showing higher scores throughout the experimental period compared to TNBS-alone mice (Figure 2D).

Clinically, TNBS + γδ T-reinfused mice displayed more severe symptoms, including watery stool, rectal bleeding, and swollen, diarrhea-like colonic morphology (Figure 2E). This contrasted sharply with TNBS + GL3 mice, which had longer colons containing normal stool. though colon length did not differ significantly from TNBS-alone mice (Figure 2E, Supplementary Figure 2D), suggesting that colon length, a macroscopic parameter often associated with acute inflammation, may be less sensitive to the incremental fibrotic changes induced by γδ T cell transfer. Histological analysis revealed exacerbated pathology in γδ T-reinfused mice, including pronounced immune cell infiltration, ulceration, crypt damage, and submucosal thickening, findings that were more severe than those observed in TNBS-alone mice (Figure 2F). In contrast, GL3 pretreatment preserved near-normal intestinal architecture in both TNBS and TNBS + γδ T groups, with reduced inflammation and epithelial damage.

Moreover, Masson’s trichrome staining demonstrated marked ECM deposition in TNBS + γδ T-reinfused mice that was substantially greater than in TNBS-alone mice, whereas GL3 pretreatment significantly reduced ECM accumulation (Figure 2F). These findings confirm that γδ T cells drive intestinal fibrosis progression, with γδ T cell reinfusion exacerbating fibrosis beyond the level induced by TNBS alone.

CD73 was downregulated after γδ T cells depletion in mice with intestinal fibrosis

Given that CD73-adenosine signaling plays a unique role in inflammation and tissue damage[33,41,42], we hypothesized that γδ T cells could exacerbate intestinal fibrosis via upregulating CD73-AR signaling pathway. Given the pronounced recruitment of γδ T cells to fibrotic intestine (Figure 2B), we found a robust association between γδ T cells and fibrotic process. To determine whether this pathogenic γδ T cell accumulation functionally engages the CD73 pathway, we evaluated CD73 expression and activity in the TNBS-induced fibrosis model. The mRNA levels of CD73 were increased to 3-folds (Supplementary Figure 3A). Also, the protein expression levels were upregulated in the colon tissues of TNBS-treated mice compared to controls as determined by both western blot and IHC staining (Figure 3A and B, Supplementary Figure 3B and C), Strikingly, CD73 protein expressions were significantly diminished following γδ T cell depletion (Figure 3A and B), suggesting the potential role for CD73 in γδ T cell-driven intestinal fibrosis progression. Consistently, the activity of CD73 was increased to about 4-fold higher in TNBS-treated group than the controls as determined by 5’-nucleotidase activity assay (Supplementary Figure 3D). Remarkably, GL3 pretreatment significantly abrogated CD73 enzymatic acitivity (Supplementary Figure 3D), further confirming the potential involvement of CD73 in γδ T cells-induced intestinal fibrosis.

Figure 3 CD73 was downregulated after γδ T cells depletion in mice with intestinal fibrosis. A and B: Immunohistochemistry (IHC) staining of CD73 and the quantification of average optical density (OD) of CD73 in the colon tissues. Quantification of IHC staining expressed as mean OD. Scale bars: 50 μm; C and D: Western blot and the quantification result of the expression levels of interleukin (IL)-17A, transforming growth factor-beta, interferon γ and IL10 using colon tissues in each group (n = 3); E and F: IHC staining of IL-17A, transforming growth factor-beta, interferon γ and IL10 and the quantification of average OD of these proteins in the colon tissues. Quantification of IHC staining expressed as mean OD. Scale bars: 50 μm. Data were presented as mean ± SD. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. TNBS: 2,4,6-trinitrobenzenesulfonic acid; IL: Interleukin; TGF: Transforming growth factor; IFN: Interferon.

It was reported that intestinal γδ T cells are associated with cytotoxicity, innate-like rapid IFN-γ[43] and IL-17A[16,17] production, playing important roles in autoimmune, inflammatory responses and fibrosis[14]. Therefore, we specifically detected the levels of pro-fibrotic cytokines, including IL17A[44] and TGF-β[45]. Moreover, IFN-γ[46] and IL-10[47], anti-fibrosis cytokines, was also analyzed. Mice treated with γδ T cells reinfusion showed increased levels of IL17A and TGF-β while in mice-pretreated with GL3 antibody, the levels of IL17A, TGF-β were even significantly downregulated in the colon tissue as analyzed by western blot and IHC staining (Figure 3C-F). Consistently, the expression levels of anti-fibrosis factor IFN-γ and IL10 were upregulated in GL3-pretreated mice (Figure 3C-F), further confirming the important role of γδ T cells in regulating intestinal fibrosis.

Inhibition of CD73 attenuated intestinal fibrosis

To further validate the essential role of CD73 in intestinal fibrosis, we administered PSB-12379, a selective CD73 inhibitor, to TNBS-treated mice (Figure 4A). As expected, CD73 expression was elevated in fibrotic mice but significantly reduced by PSB-12379 treatment, as confirmed by qPCR and IHC staining (Figure 4B, Supplementary Figure 4A and B), which was further validated by the lower enzyme activity of CD73 (Figure 4C). Additionally, Furthermore, PSB-12379 administration restored the downregulated ATP levels observed in TNBS-induced fibrotic mice (Figure 4D), suggesting that the lower CD73 expression impaired its catalytic function in the ATP-AMP-adenosine metabolic pathway. Consistent with the phenotypic effects observed following γδ T cell depletion, PSB-12379 treatment promoted weight recovery (Figure 4E) and reduced DAI scores (Figure 4F). Notably, the shorter length of colon induced by TNBS treatment was significantly improved by PSB-12379 (Supplementary Figure 4C and D). Histological analysis revealed that PSB-12379 mitigated crypt damage and inflammation compared to TNBS-treated controls (Figure 4G). TCR staining corroborated these findings, showing a significant reduction in intestinal γδ T cell numbers after PSB-12379 treatment (Figure 4G and H), implying that the antifibrotic effect of PSB-12379 is likely mediated, at least in part, through the modulation of γδ T cell function or abundance.

Figure 4 Inhibition of CD73 attenuated intestinal fibrosis. A: Experimental scheme of 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced intestinal fibrosis model. PSB-12379 were intraperitoneally injected after TNBS administration; B: Quantitative polymerase chain reaction analysis of the expression of CD73 gene in mice treated with TNBS and TNBS + PSB-12379 injection. Normal colon tissues were used as the control (n = 3 in each group). Data are presented as relative mRNA levels normalized to β-actin; C: The 5’-nucleotidase activity of CD73 in each group, normal colon tissues were used as the control (n = 3); D: The concentrations of ATP in the colon tissues; E: Body weight changes in the mice administrated with TNBS or TNBS + PSB-12379 during the experimental period (n = 5); F: Disease activity index score evaluated in the mice treated with TNBS or TNBS + PSB-12379 (n = 5); G: The representative image for hematoxylin and eosin, Masson staining and immunohistochemistry staining of TCR in the colon tissues in the mice treated with TNBS or TNBS + PSB-12379. Orange arrow indicated areas of immune cell infiltration. Scale bars: 100 μm or 50 μm; H: Quantification of immunohistochemistry staining (Figure 4G), described as mean optical density; I: The concentrations of hydroxyproline levels in the serum at week 14 (n = 3); J: The levels of hyaluronic acid, type IV collagen, laminin, and procollagen III in the serum at week 14 (n = 3). Data were presented as mean ± SD. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. TNBS: 2,4,6-trinitrobenzenesulfonic acid; NT5E: 5’-nucleotidase; DAI: Disease activity index; H&E: Hematoxylin and eosin; HA: Hyaluronic acid; IV-C: Type IV collagen; LN: Laminin; PCIII: Procollagen III.

Moreover, Masson’s trichrome staining demonstrated that PSB-12379 suppressed ECM accumulation in fibrotic mice (Figure 4G). This was consistent with reduced levels of fibrosis-associated ECM components (HA, IV-C, LN, and PCIII) and decreased hydroxyproline content following PSB-12379 treatment (Figure 4I and J). Moreover, the levels of pro-fibrotic cytokines IL17A and TGF-β were reduced in mice treated with PSB-12379 administration compared with mice treated with TNBS while the concentrations of IFN-γ and IL-10, these anti-fibrosis cytokines, were slightly increased in mice treated with PSB-12379 compared with TNBS-treated mice (Supplementary Figure 4E and F). Taken together, inhibition of CD73 attenuated intestinal fibrosis.

Increased cAMP levels prevented intestinal fibrosis

Given that CD73 catalyzes the hydrolysis of AMP to adenosine in the adenosine metabolic pathway[42], we hypothesized that CD73 exerts its regulatory functions through ARs-mediated signaling. As all ARs are G protein-coupled receptors and mainly exert their function via the cAMP pathway[31]. Therefore, the downstream of adenosine signaling were intervened by targeting cAMP signaling. Following TNBS-induced colitis, mice received either the adenylyl cyclase activator FSK or the cAMP synthesis inhibitor SQ22536 (Figure 5A). First, we analyzed the inhibition and activation effects. The concentrations of cAMP were upregulated in FSK-treated mice while downregulated after SQ22536 treatment (Figure 5B). Meanwhile, the concentrations of ATP were upregulated in FSK-treated mice while downregulated after SQ22536 treatment (Figure 5B). Of note, depletion of γδ T cells with GL3 antibody similarly increased cAMP and ATP concentrations to levels comparable with FSK treatment (Figure 5B), suggesting that γδ T cell ablation alleviates fibrosis, at least partially, by disinhibiting the CD73-Ars-cAMP signaling axis. Intriguingly, cAMP modulation regulated CD73 expression and activity. FSK treatment downregulated CD73 mRNA and protein levels as confirmed by qPCR and IHC staining (Figure 5C, Supplementary Figure 5A and B). Concordantly, CD73 enzymatic activity decreased post-FSK but increased significantly with SQ22536 (Figure 5C).

Figure 5 Increased cyclic AMP levels prevented intestinal fibrosis. A: Experimental scheme of 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced intestinal fibrosis model. Forskolin and SQ22536 were intraperitoneally injected twice weekly for 4 weeks after TNBS administration; B: The concentrations of cyclic AMP and ATP in the colon tissues. Normal colon tissues were used as the control (n = 3 in each group); C: Quantitative polymerase chain reaction analysis of the expression of CD73 gene and the 5’-nucleotidase activity of CD73 in each group. Normal colon tissues were used as the control (n = 3 in each group); D: Body weight changes in the mice treated with TNBS or TNBS + forskolin or TNBS + SQ22536 during the experimental period (n = 5); E: Disease activity index score evaluated in the mice treated with TNBS or TNBS + forskolin or TNBS + SQ22536 (n = 5); F: The representative image for hematoxylin and eosin, Masson staining and immunohistochemistry staining of TCR within colon tissues in the mice treated with TNBS or TNBS + forskolin or TNBS + SQ22536. Orange arrow indicated areas of immune cell infiltration. Scale bars: 100 μm or 50 μm; G: The concentrations of hydroxyproline levels in the serum in the mice treated with TNBS or TNBS + forskolin or TNBS + SQ22536 at week 14 (n = 3); H: The levels of hyaluronic acid, type IV collagen, laminin, and procollagen III in the serum at week 14 (n = 3). Data were presented as mean ± SD. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. TNBS: 2,4,6-trinitrobenzenesulfonic acid; FSK: Forskolin; cAMP: Cyclic AMP; NT5E: 5’-nucleotidase; DAI: Disease activity index; H&E: Hematoxylin and eosin; LN: Laminin; IV-C: Type IV collagen; PCIII: Procollagen III; HA: Hyaluronic acid.

Following the intervention, SQ22536-treated mice exhibited exacerbated weight loss at week 14 compared with TNBS controls, while FSK improved weight recovery, phenocopying the protective effect of the CD73 inhibitor PSB-12379 (Figure 5D). DAI scores followed similar trends (Figure 5E). Histopathological analysis further supported these findings. SQ22536 administration intensified crypt damage and inflammation, whereas FSK preserved near-normal intestinal architecture (Figure 5F). Consistently, intestinal γδ T cell infiltration (by TCR staining) was markedly reduced by FSK but amplified by SQ22536 relative to TNBS controls (Figure 5F). Critically, ECM accumulated markedly in SQ22536-treated mice but was prevented by FSK co-administration with TNBS (Figure 5F). In addition, SQ22536 also dramatically elevated hydroxyproline concentrations and ECM components (HA, IV-C, LN, PCIII); these effects were abrogated by FSK (Figure 5G and H). IHC staining further validated these findings, showing upregulated pro-fibrotic cytokines (TGF-β, IL-17A) and downregulated anti-fibrotic cytokines (IFN-γ, IL-10) in SQ22536-treated colons (Supplementary Figure 5C-E). Taken together, γδ T cell-driven intestinal fibrosis is most likely dependent on the CD73-AR-cAMP axis. FSK-mediated cAMP elevation attenuated intestinal fibrosis, while SQ22536-induced cAMP reduction exacerbated it, identifying cAMP modulation as a novel therapeutic strategy.

DISCUSSION

Intestinal fibrosis is a major contributor to intestinal dysfunction and carcinogenesis. Here, our study elucidates a previously unrecognized role for γδ T cells as pivotal drivers of intestinal fibrosis, mechanistically linked to the CD73-AR-cAMP signaling axis. Depletion of γδ T cells or targeting CD73-adenosine signaling markedly attenuated TNBS-induced fibrosis, as evidenced by reduced collagen deposition, hydroxyproline levels, and key ECM components, revealing novel therapeutic targets to treat intestinal fibrosis. Beyond the immediate findings, this work has several broader implications for the field. First, it highlights the pathogenic potential of innate-like lymphocytes in tissue fibrosis, expanding our understanding of immune contributions to fibrogenesis beyond classical Th17 and macrophage pathways. Second, it reveals a previously unrecognized role for CD73-adenosine signaling in promoting fibrosis, challenging the prevailing view that this pathway is exclusively immunosuppressive and tissue-protective. This context-dependent duality of adenosine signaling, protective in acute inflammation but potentially detrimental in chronic fibrosis, warrants further investigation.

γδ T cells, abundant in the gut epithelium[9], exhibit functional plasticity. While they contribute to barrier repair under homeostasis[10,14,48]. Our data align with reports that they can become pathogenic in chronic inflammation, particularly through IL-17A production[6,7]. TNBS-treated mice exhibited elevated pro-fibrotic cytokines the levels of pro-fibrotic cytokines (IL17A[44] and TGF-β[45]) and suppressed anti-fibrosis cytokines (IFN-γ[46] and IL-10[47]). Notably, IL-17, a cytokine critical for epithelial integrity and overexpressed in in Crohn’s strictures cells in Crohn’s disease[49], is predominantly produced by Th17 cells and γδT cells[12,20]. Under homeostatic conditions, IL-17 production is mainly from γδ T cells[20], implicating it in fibrosis promotion. Upon γδ T cell depletion, the observed reduction in IL-17A and TGF-β, and increase in IFN-γ and IL-10, supports their net profibrotic role in this model. Our data demonstrate a clear shift in the colonic cytokine milieu upon γδ T cell manipulation; however, the cellular source of these cytokines, whether directly from γδ T cells or from other immune cells (e.g., Th17 cells, macrophages) influenced by them, remains correlative. The observed significant infiltration of γδ T cells into the fibrotic colon suggested that at least a portion of the elevated IL-17A is directly derived from these cells. However, other immune populations (e.g., Th17 cells, macrophages) likely also contribute to the observed cytokine profiles. Future studies employing intracellular cytokine staining and single-cell RNA sequencing will be necessary to definitively map the cellular sources of these cytokines in intestinal fibrosis and the heterogeneity of γδ T cells[10,19,50].

A key finding of our study is the paradoxical role of the CD73-adenosine axis. While adenosine is typically viewed as an immunosuppressive and tissue-protective molecule in acute injury[28-30], our data suggest that chronic upregulation of CD73 by γδ T cells creates a pro-fibrotic environment. Consistent with experimental autoimmune uveitis[32], we also found that CD73-AR signaling regulated local inflammation. Critically, γδ T cell depletion reduced CD73 activity by > 70%, directly linking these cells to adenosine overproduction. However, further detailed studies are required to validate the feedback mechanism via direct measurement of adenosine levels in the colon tissue. Pharmacological inhibition of CD73 (PSB-12379) or activation of AR signaling (via FSK) replicated the anti-fibrotic effects of γδ T cell depletion, validating this axis as a therapeutic target. Aligning with some reports[29,51], our data reveal a non-monotonic regulatory loop in cAMP signaling. The divergent outcomes of cAMP modulation, where supraphysiological elevation via FSK was protective while inhibition via SQ22536 was detrimental, underscore the compartmentalized nature of cAMP signaling and the potential dose-response. This context-dependency is a critical consideration for future therapeutic targeting. Furthermore, aligned with some reports[29,51], our data also reveal a paradoxical feedback loop in cAMP signaling.

While GL3 depletion phenocopies the cAMP and ATP changes seen with FSK, it should be note that GL3 and FSK likely act through a convergent signaling axis rather than identical pathways. FSK directly activates adenylyl cyclase, while GL3 depletion removes a primary source of CD73, thereby altering the adenosine metabolic landscape. While direct regulation of adenylyl cyclase by γδ T cells cannot be ruled out, the reduction in CD73-mediated adenosine signaling is the most likely driver of these tissue-level cAMP changes.

Certain limitations warrant future investigation. First, the TNBS model induces acute-on-chronic fibrosis and may not fully mimic human Crohn’s strictures. Validation in chronic dextran sulfate sodium or IL-10^-/- models is needed. Second, while our depletion/reinfusion data are compelling, conditional knockout mice (e.g., Cd73 Δγδ T, A2AR^-/-) would establish cell-intrinsic mechanisms. Third, given that distinct γδ T cell subsets (Vδ1+, Vδ2+, Vδ3+) populate the gut epithelium, their subset-specific functions, particularly cytokine profiles (e.g., IL-17 vs IFN-γ), proliferation, apoptosis, and AR expression[10,19,50], warrant detailed characterization. Fourth, the impact of gut microbiota and microbial metabolites on the γδ T cell-CD73 axis remains unexplored and is critical to understanding intestinal fibrosis pathophysiology. Finally, human validation is a crucial next step to understand the unique pathophysiology of intestinal fibrosis.

While our findings revealed that γδ T cells as central regulators of intestinal fibrosis vial upregulating CD73 expression and enzymatic activity, consistent with elevated CD73 in human fibrotic disorders[9], further studies should be performed. First, the TNBS-induced acute fibrosis model may not fully recapitulate human chronic fibrostenotic diseases (e.g., Crohn’s disease). Second, given the heterogeneity of γδ T cells, subset-specific analysis using single-cell RNA sequencing and multi-parameter flow cytometry will be crucial to identify the precise γδ T cell subpopulations (e.g., IL-17-producing vs IFN-γ-producing) that drive fibrosis and express CD73. This could enable more targeted therapeutic strategies that preserve protective subsets while eliminating pathogenic ones. Third, although γδ T cell depletion with GL3 antibody significantly reduced total CD73 protein expression and enzymatic activity compared the TNBS-group. While the reduction (approximately 50%) is striking and suggests that γδ T cells are either major expressers of CD73 or critical regulators of CD73 expression on other cells. Conditional knockout mice with γδ T cell-specific deletion of CD73 or A2AR would definitively establish cell-autonomous vs non-autonomous mechanism. Third, in vitro functional characterization of isolated γδ T cells, including proliferation, apoptosis, cytokine profiles, and AR expression, is warranted. Finally, the therapeutic efficacy of CD73 inhibitors should be evaluated in chronic, more clinically relevant fibrosis models (e.g., chronic dextran sulfate sodium or IL-10^-/- models) and the potential interplay between the γδ T cell-CD73 axis and gut microbiota, which is a critical modulator of both γδ T cell function and adenosine signaling, represents an exciting avenue for future investigation.

CONCLUSION

In conclusion, our study identifies γδ T cells as pivotal mediators of intestinal fibrosis, acting primarily through the CD73-AR-cAMP signaling pathway. Both antibody-mediated depletion of γδ T cells and pharmacological inhibition of CD73 significantly attenuated fibrosis in a pre-clinical model, establishing this axis as a novel therapeutic target for fibrostenotic Crohn’s disease. These results not only broaden our understanding of innate-like lymphocytes in chronic IBD complications but also nominate CD73 inhibition as a potential localized therapeutic intervention to prevent or reverse intestinal strictures.

ACKNOWLEDGEMENTS

We are thankful to Dr. Dong-Chun Liang for critical comments and suggestion for CD73 associated experiments. We also thank the core facilities for molecular biology and cell biology, animal care at Hainan Provincial Medical Research Institute.