Alterations in Gamma-Delta T Cells in Patients With Primary Biliary Cholangitis

Gamma-delta (γδ) T cells are involved in the development of diverse liver and autoimmune diseases, whereas the role of γδ T cells in primary biliary cholangitis (PBC) remains unclear. Methods We analyzed the number, phenotypes, and functional molecules of γδ T cells in PBC patients (n = 74) and sex- and age-matched healthy controls (HCs) (n = 74) by ow cytometric analysis. Results We identied two distinct functional subsets of circulating γδ T cells according to the CD3/TCRγδ complex: the TCRγδ high and TCRγδ low subsets. Approximately three-quarters of cells in the TCRγδ high subset were Vδ1 T cells, while Vδ2 T cells were enriched in the TCRγδ low subset in HCs. The frequency and absolute number of circulating TCRγδ low cells was signicantly decreased in PBC patients compared with HCs (p < 0.001). Furthermore, the frequency of TCRγδ low cells was negatively correlated with disease severity and positively correlated with the ursodeoxycholic acid response. TCRγδ low cells exhibited a similar apoptotic and proliferative phenotype but enhanced liver-homing chemokine receptor (CXCR6) expression in PBC patients compared with HCs. In addition, both TCRγδ high and TCRγδ low subsets were more activated in PBC compared with HCs, characterized by elevated expression levels of CD69 and HLA-DR. Finally, we found an increased granzyme B (GZMB) production and similar IFN-γ and TNF-α production of TCRγδ low cells in PBC patients compared with HCs. The TCRγδ low subset might be a potential marker for disease and treatment response in PBC, which may play a crucial role in liver injury through increased CXCR6 expression and GZMB production. AST, PLT,


Introduction
Primary biliary cholangitis (PBC) is an autoimmune liver disease with an increasing tendency of prevalence worldwide [1]. Although the pathogenesis of PBC has not been fully clari ed, innate and adaptive immune responses play a key role in liver injury [2].
Gamma-delta (γδ) T cells are a unique population of unconventional T cells with γ and δ glycoprotein chains. In humans, γδ T cells represent 2-10% of T cells in the peripheral blood and 5-15% of T cells in the liver, which regulates in ammation, pathogen clearance, and tumor immunity through the production of various cytokines [3][4][5][6][7]. Previous studies have demonstrated that γδ T cells are closely involved in the development of viral hepatitis and autoimmune liver diseases [5,8,9]. However, the frequency of γδ T cells in PBC remains controversial in previous studies with small sample sizes [9][10][11][12]. The function of γδ T cells in PBC also needs to be further explored.
Additionally, most works on human γδ T cells have focused on Vδ1 and Vδ2 T cells, which are discriminated by different types of γ and δ chains [4]. Yokobori N et al. reported a novel way to identify two distinct subsets of γδ T cells, called the TCRγδ high and TCRγδ low subsets, in patients with tuberculous pleurisy according to CD3/TCRγδ complex expression [13]. In line with this, a recent study found different expression levels of transcription factors, including PLZF and RORγt, between human circulating TCRγδ high and TCRγδ low cells [14], indicating that the two distinct subsets might have different phenotypes and functions. However, the role of the TCRγδ high and TCRγδ low subsets in PBC remains unclear.
Therefore, our study aimed to explore the number, phenotype, and functional molecules of γδ T cells in PBC, with a special focus on the TCRγδ high and TCRγδ low subsets.

Patients And Methods
Patients PBC patients were diagnosed and enrolled at Beijing Friendship Hospital, Capital Medical University. This study was approved by the Institutional Committee for Human Research of Beijing Friendship Hospital, and written informed consent was obtained from all participants.
Patients will be included if they (1) were diagnosed with PBC based on international diagnostic criteria [2,15] and (2) were treatment-naïve or received monotherapy with UDCA at a dose of 13-15 mg/kg/day. Patients were excluded if they (1) had coexistent other liver diseases, such as autoimmune hepatitis, chronic hepatitis B or C infection, drug-induced liver injury, alcoholic liver disease or metabolic fatty liver disease; (2) had coexistent malignant tumors, diabetes, or other autoimmune diseases, such as Sjogren's syndrome or autoimmune thyroiditis; or (3) were treated with corticosteroid or immunosuppressive drugs in the last 4 weeks.
Healthy controls (HCs) were included if they (1) were age-and sex-matched with PBC patients and (2) had normal routine blood, liver and kidney function, glucose, and lipid test results, routine urine test results, and abdominal ultrasound results in the last six months. HCs were excluded if they (1) had coexistent liver diseases, autoimmune diseases, diabetes, tumors or other diseases that can shorten life expectancy or (2) suffered from alcohol abuse.
Blood sample collection and PBMC preparation Blood samples were obtained from 74 PBC patients and 74 healthy volunteers. Peripheral blood mononuclear cells (PBMCs) were freshly isolated by Ficoll-Hypaque density centrifugation.
Flow cytometry was performed using a FACSAria II ow cytometer (BD Biosciences, CA, USA), and the data were analyzed with FlowJo software (Treestar, Ashland, OR, USA).

Statistical analysis
Categorical variables are expressed as counts and percentages. Continuous variables are all described as medians with interquartile ranges since they did not t the normal distribution. The Mann-Whitney U-test was performed to compare the differences in continuous variables between PBC patients and HCs. The Wilcoxon matched-pairs signed rank test was performed to compare the differences between TCRγδ high and TCRγδ low cells. The Kruskal-Wallis tests was performed to compare the differences among treatment-naïve patients, nonresponders and responders. The chi-squared test was applied to compare the categorical variables between PBC patients and HCs. A two-sided P value < 0.05 was considered signi cant. The analyses were conducted by using SPSS statistics version 26.

Demographic and clinical characteristics of the subjects
The demographic and clinical characteristics of PBC patients and HCs are shown in Table 1. A total of 74 PBC patients and 74 age-and sex-matched HCs were enrolled in the study. Signi cant differences were observed in serum levels of alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT), transaminase, albumin (ALB), total bilirubin (TBIL) and platelet count (PLT) between PBC patients and HCs. Among PBC patients, 12 were treatment-naïve, and 33 (44.6%) responded to UDCA therapy according to Paris I (for cirrhotic PBC patients) and Paris II criteria (for noncirrhotic PBC patients).  (Fig. 1a). The frequency of circulating TCRγδ low cells among CD3 + cells was signi cantly higher than that of TCRγδ high cells among CD3 + cells both in PBC patients (median 2.75% vs. 0.94%) and HCs (median 5.84% vs. 0.88%) (both p < 0.001).
The frequency and absolute number of circulating TCRγδ low cells were signi cantly decreased in PBC patients (n = 74) compared with HCs (n = 74) (p < 0.001), but the frequency and number of circulating TCRγδ high cells were similar between the two groups (p > 0.05) (Fig. 1b).
The frequency of circulating TCRγδ low cells was correlated with disease severity and UDCA response We found a signi cant positive correlation between the frequency of circulating TCRγδ low cells and the serum level of PLT (p = 0.001, r = 0.396), but no correlations were found between the frequency of circulating TCRγδ low cells and the levels of TBIL, ALP or aspartate aminotransferase (AST) (Fig. 1c).
Furthermore, the frequency and absolute number of circulating TCRγδ low cells were signi cantly decreased in cirrhotic PBC patients (n = 28) compared with noncirrhotic PBC patients (n = 46) (Fig. 1d).
Additionally, subgroup analysis showed that the frequency and absolute number of circulating TCRγδ low cells were the lowest in UDCA treatment-naïve PBC patients, followed by nonresponders and responders.

The proportion of Vδ2 T cells in the circulating TCRγδ low subset was decreased in PBC patients
We further detected the expression of Vδ1 and Vδ2 in the TCRγδ high and TCRγδ low subsets (Fig. 2a).
Approximately 75.1% of TCRγδ high cells were Vδ1 T cells and 19.3% were Vδ1 − Vδ2 − T cells in HCs. The proportion of Vδ1 T cells in the TCRγδ high subset was not different between HCs and PBC patients (p > 0.05, Fig. 2b).
In contrast, the majority of TCRγδ low cells were Vδ2 T cells in HCs. The proportion of Vδ2 T cells in the TCRγδ low subset was decreased (89.4% vs. 97.4%), and the proportion of Vδ1 − Vδ2 − T cells was increased (7.3% vs. 2.1%) in PBC patients compared with HCs (p < 0.001, Fig. 2c).

Circulating TCRγδ low cells exhibited similar apoptotic and proliferative phenotypes and enhanced CXCR6 expression in PBC patients
To explain the reduction in TCRγδ low cells in PBC patients, we assessed the apoptosis, proliferation, and expression of liver-homing chemokine receptors of γδ T cells (Fig. 3a). No difference was observed in the expression of annexin-V (AV) and Ki67 in either circulating TCRγδ high cells or TCRγδ low cells between HCs and PBC patients (p > 0.05, Fig. 3b and 3c).
Compared with TCRγδ high cells, the expression of liver-homing chemokine receptors, including CXCR6 and CXCR3, was signi cantly higher in TCRγδ low cells in both HCs and PBC patients (all p < 0.001). The expression of CXCR6 in TCRγδ high cells and TCRγδ low cells was signi cantly increased in PBC patients compared with HCs (p < 0.05, Fig. 3d), but the expression of CXCR3 in TCRγδ high cells and TCRγδ low cells was comparable between the two groups (p > 0.05, Fig. 3e).

Circulating TCRγδ low cells demonstrated an activated phenotype and enhanced GZMB production in PBC patients
The expression of activation markers, including CD69, HLA-DR, and CD25, in circulating γδ T cells was analyzed in PBC patients and HCs. Circulating TCRγδ high cells were more activated than TCRγδ low cells in both HCs and PBC patients. Both the circulating TCRγδ high cells and TCRγδ low cells in PBC were more activated than those in HCs, characterized by higher expression of CD69 and HLA-DR (p < 0.05, Fig. 4a).
In addition, circulating TCRγδ high cells produced higher GZMB but lower IFN-γ and TNF-α than TCRγδ low cells in both HCs and PBC patients. We found an increased GZMB production of TCRγδ low cells in PBC patients (n = 28) compared with HCs (n = 30) (p < 0.05, Fig. 4c). The production of IFN-γ and TNF-α by TCRγδ low cells in PBC patients was lower than that in HCs, but the results remained statistically insigni cant (p > 0.05, Fig. 4c).

Discussion
In the present study, we analyzed the number, phenotype, and functional molecules of γδ T cells in PBC patients and sex-and age-matched HCs by ow cytometric analysis. We found that the number and function of TCRγδ low cells were signi cantly altered in PBC patients compared with HCs, which might be involved in the pathogenesis of PBC.
We identi ed two distinct subsets of circulating γδ T cells in PBC patients according to CD3/TCRγδ complex expression: the TCRγδ high and TCRγδ low subsets. Approximately three-quarters of cells in the TCRγδ high subset were Vδ1 T cells, while the majority of the TCRγδ low subset was Vδ2 T cells in HCs, which was consistent with the results of a previous study [13]. Furthermore, these two subpopulations differed from each other in terms of the expression levels of chemokines, activation markers, and functional molecules. Compared with the TCRγδ high subset, the TCRγδ low subset expressed higher levels of liver-homing chemokine receptors, lower activation markers, lower GZMB, and higher IFN-γ and TNF-α.
The change in the quantity of circulating γδ T cells in PBC patients remains controversial in previous studies [9][10][11][12]. Studies have found a signi cantly decreased frequency of circulating γδ T cells in PBC patients compared with HCs [11,12], whereas the other two studies showed a similar frequency of γδ T cells between the two groups [9,10]. This might be due to the small sample sizes and different disease stages. The number of patients in our study was more than twice that of the previous studies. We found that the frequency and the absolute number of γδ T cells was signi cantly decreased in PBC patients compared with HCs, and the γδ T cells that were decreased in number were the those in the circulating TCRγδ low subset. The relatively large sample size allowed us to further analyze the frequency of TCRγδ low cells in subgroups. The results showed that the frequency of TCRγδ low cells was negatively correlated with disease severity and positively correlated with the UDCA response, indicating that TCRγδ low cells might be a potential marker of disease progression and treatment response.
To explore the possible reasons for the reduction in TCRγδ low cells in PBC patients, we further detected the apoptotic and proliferative phenotype and the expression of liver-homing chemokine receptors on γδ T cells. We found that circulating TCRγδ low cells exhibited a similar apoptotic and proliferative phenotype but enhanced liver-homing chemokine receptor (CXCR6) expression in PBC patients compared with HCs.
Therefore, we postulated that the reduction in TCRγδ low cells might be caused by redistribution to the liver. In line with this hypothesis, studies found an increased number and frequency of hepatic γδ T cells in autoimmune liver diseases and chronic liver diseases compared with HCs [9,16,17]. However, a recent study demonstrated a decreased frequency of hepatic γδ T cells in 13 PBC and other chronic liver diseases patients compared with that in patients with healthy livers. Of note, all 13 of these PBC patients had end-stage liver disease, which might be an explanation for the opposite results. Future studies with larger sample sizes and different disease stages are needed to explore the quantity and function of γδ T cells and their subpopulations in PBC livers.
Generally, the CXCR6-CXCL16 axis is considered to promote in ammation and disease progression in various liver diseases, including brosis[18], hepatocellular carcinoma [19], acute liver injury [20,21], and nonalcoholic fatty liver disease [22]. Our results showed that approximately one-third of circulating TCRγδ low cells expressed CXCR6, and the proportion was even higher in PBC patients than in HCs. Consistent with our results, a recent study using single-cell RNA sequencing found that CXCR6 is preferentially expressed by hepatic γδ T cells [17]. Taken together, this evidence suggests that γδ T cells might be involved in liver injury through the CXCR6-CXCL16 axis.
The function of γδ T cells in PBC has not been well elucidated. Human γδ T cells display antiviral, proin ammatory or cytotoxic abilities in different liver and autoimmune diseases through the production of various cytokines, mainly IFN-γ, TNF-α, and GZMB [4,5,8,23]. In this study, we analyzed the expression of activation markers and functional molecules of both the TCRγδ high and TCRγδ low subsets in PBC patients and HCs. We demonstrated that the circulating TCRγδ low subset exhibited an activated phenotype and enhanced cytotoxicity characterized by increased GZMB production in PBC patients, suggesting that the TCRγδ low subset might cause liver damage through GZMB production. Consistent with our results, Ferri S et al. found that γδ T cells produced higher GZMB in 47 patients with autoimmune hepatitis than in 28 HCs, and the production of GZMB was correlated with biochemistry variables of liver damage [8]. In addition, another study demonstrated that circulating Vδ2 T cell function shifted from antiviral toward cytotoxicity after HCV infection, which is characterized by lower IFN-γ but higher GZMB produced by Vδ2 T cells [24]. Furthermore, the percentage of GZMB + Vδ2 T cells positively correlated with serum ALT levels in HCV-infected patients [24]. All these results showed that γδ T cells might induce liver injury through enhanced GZMB production.
This study has several limitations. First, we failed to analyze the number and function of γδ T cells in the liver by immunohistology, and further relevant research on hepatic γδ T cells by new strategies, such as RNA scope, is needed. Second, this is a cross-sectional study that enrolled patients with different disease stages and treatment responses. Long-term follow-up of the level of γδ T cells in the same cohort before and after UDCA treatment will provide more value.
In summary, we identi ed two distinct functional subsets of circulating γδ T cells according to the CD3/TCRγδ complex in PBC patients. The TCRγδ low subset was signi cantly decreased in PBC patients and correlated with disease severity and the UDCA response, and this subset may play a crucial role in liver injury through increased CXCR6 expression and GZMB production.

Declarations
Funding This work is funded by National Natural Science Foundation of China (82000533).
Con icts of interest The authors declare that they have no con ict of interest. Consent to participate Written informed consent was obtained from all subjects enrolled in the study.
Consent for publication Written informed consent for publication was obtained from all participants.
Availability of data and material Not applicable.