T Cell Polarization toward TH2/TFH2 and TH17/TFH17 in Patients with IgG4-Related Disease

IgG4-related disease (IgG4-RD) is a fibro-inflammatory disorder involving virtually every organ with a risk of organ dysfunction. Despite recent studies regarding B cell and T cell compartments, the disease’s pathophysiology remains poorly understood. We examined and characterized subsets of circulating lymphocytes in untreated patients with active IgG4-RD. Twenty-eight consecutive patients with biopsy-proven IgG4-RD were included in a prospective, multicentric study. Lymphocytes’ subsets were analyzed by flow cytometry, with analysis of TH1/TH2/TH17, TFH cells, and cytokine release by peripheral blood mononuclear cells. Results were compared to healthy controls and to patients with primary Sjögren’s syndrome. Patients with IgG4-RD showed an increase of circulating T regulatory, TH2, TH17, and CD4+CXCR5+PD1+ TFH cell subsets. Accordingly, increased levels of IL-10 and IL-4 were measured in IgG-RD patients. TFH increase was characterized by the specific expansion of TFH2 (CCR6−CXCR3−), and to a lesser extent of TFH17 (CCR6+CXCR3−) cells. Interestingly, CD4+CXCR5+PD1+ TFH cells normalized under treatment. IgG4-RD is characterized by a shift of circulating T cells toward a TH2/TFH2 and TH17/TFH17 polarization. This immunological imbalance might be implicated in the disease’s pathophysiology. Treatment regimens targeting such T cells warrant further evaluation.

associated with fibrosis (5). Tertiary germinal center formation is also frequently observed in diseased tissues. Mikulicz's disease, otherwise considered as a disease manifestation of primary Sjögren's syndrome (pSS) (6) is also part of the spectrum of IgG4-RD. It has now been clearly demonstrated that despite common features shared by both diseases, clinical, biological, pathological, and immunohistological findings differ between pSS and IgG4-RD (7). Yet, in daily practice, pSS remains a frequent differential diagnosis of IgG4-RD.
Despite ongoing genetic and mechanistic studies, the pathogenesis of IgG4-RD remains poorly understood. Rather than a vector of tissue damage, IgG4 is considered to be a marker of the disease. Yet, elevated levels of serum IgG4 have also been reported in other conditions, and approximately 30% of patients with biopsy-proven IgG4-RD have normal serum levels of IgG4 (8). Hence, the identification of novel biomarkers is a timely topic in the field of IgG4-RD.
Besides studies regarding B cells and immunoglobulin biology, careful attention is currently being paid on the role of T cells in the disease. Conflicting results have been reported regarding the polarization of T helper (TH) cells in IgG4-RD, especially in patients with pancreatic involvement and Mikulicz's disease (9,10). Because regulatory T cells (Treg) are a source of IL-10 and TGF-β (i.e., key cytokines involved, respectively, in the differentiation of IgG4-producing B cells and in the genesis of fibrosis), the roles of these cells have also been investigated in a few preliminary studies regarding IgG4-related hepatic and pancreatic involvements (11)(12)(13) More recently, T follicular helper cells (TFH), an important cell subset involved in the development of germinal centers as well as in antibody production, have also been studied in a small series of IgG4-RD patients (14) suggesting changes of this compartment in the disease. Yet, these results need to be confirmed in a new and larger cohort, including patients with other than salivary and lachrymal-restricted involvements, and with a more precise phenotypic definition of TFH subsets, especially concerning PD1 molecule expression.
In this study, we investigated circulating lymphocytes in untreated patients with IgG4-RD, with a focus on TH1/TH2/ TH17 balance and PD1 + TFH cells, and compared them to pSS patients.

PaTienTs anD MeThODs study subjects
Patients with IgG4-RD were identified according to the Comprehensive Diagnosis Criteria (CDC) for IgG4-RD or to the International Consensus Diagnostic Criteria (ICDC) for autoimmune pancreatitis (15). Patients with pSS were identified according to the 2002 American-European Consensus Group criteria. Patients with either IgG4-RD or pSS who received steroids or disease-modifying antirheumatic drugs within 3 months prior to study entry or who received rituximab within 6 months prior to enrollment were excluded from the study. Written informed consent was obtained from all patients with IgG4-RD or pSS as well as healthy controls (HC) prospectively included in the study. The study was approved by the local ethics committee [CPP Marseille I (Comité de Protection des Personnes Marseille I)].

immunophenotyping by Flow cytometry
Lymphocyte populations (total lymphocytes, T cells, CD4 + T cells, CD8 + T cells, B cells, and natural killer (NK) cells) were quantified with 6-Color BD Multitest and BD Trucount Technologies (Becton Dickinson, Le Pont de Claix, France) according to the manufacturer's instructions.
For intracellular markers, fixed and permeabilized cells were stained using Alexa Fluor 488-anti-FoxP3 (BD Biosciences) to analyze T regulatory cells defined as CD4 + FoxP3 + ; PE-anti-IL-17, FITC-anti-IFNγ, and APC-anti-IL-4 (BD Bioscience) were used to characterize the functionally polarized CD4 + T cell subsets after stimulation with PMA-ionomycin for 5 h according to the manufacturers' instructions. The cells were washed with phosphate-buffered saline and then analyzed on a BD FACS Canto II.

analysis of cytokine Production
Levels of IL-4, IL-10, IL-17, and IFNγ in the supernatant of PBMCs following a 24-h stimulation with PMA-ionomycin were assessed using a multiplexed bead-based immunoassay (CBA ® kit, BD Biosciences) following the manufacturers' protocol.

statistical analysis
Continuous variables are shown as median ± SD. Multiple group comparisons were analyzed using the Kruskal-Wallis test and the Mann-Whitney U-test was used for comparison between two groups. Correlations were analyzed using Spearman's correlation coefficient. The Wilcoxon test was used to analyze the changes in values over time. Statistical analyses were performed with Prism 6 (GraphPad Software, San Diego, CA, USA).
For multiple testing, adjusted p-values were calculated using the false discovery rate procedure with the PROC MULTTEST statement (16). Statistical tests for multiple testing were made with the SAS 9.4 software.

lymphocyte subsets in Patients with igg4-rD
Analysis of patients with IgG4-RD showed no difference in the number and proportion of CD4 + , CD8 + T cell subsets, and NK cells compared to HC ( Figure 1A). Interestingly, both the number (1,685 ± 418 cells/μl versus 351 ± 66 cells/μl; p = 0.001) and proportion of circulating plasmablast cells were increased in patients with IgG4-RD compared to HC, although B cell counts were similar within the three groups ( Figure 1B).

T h 1, T h 2, and T h 17 cell Distribution in Patients with igg4-rD
To analyze the distribution of functionally polarized CD4 + T cell subsets in patients with IgG4-RD, PBMCs were stimulated for 5 h using PMA-ionomycin (Figures 3A,B). The percentage of IL-4producing CD4 + T cells (TH2) was significantly higher in patients with IgG4-RD compared to HC and pSS patients (3.4 ± 0.36% versus 1.5 ± 0.14 and 2.2 ± 0.5%; p < 0.0001 and p = 0.03, respectively). In addition, IL-17 + -producing CD4 + T cells (TH17) were significantly higher in patients with IgG4-RD compared to HC (1.0 ± 0.17% versus 0.7 ± 0.17%; p = 0.02) but not to patients with  pSS. Conversely, the proportion of TH1 IFNγ + producing cells was similar in all three groups of patients.

cytokine secretion Profile in Patients with igg4-rD
In order to further confirm the trend toward a TH2/TH17 balance in IgG4-RD, PBMCs' cytokines production following a 24-hour stimulation with PMA-ionomycin was measured using a multiplexed CBA assay. As shown in Figure 3C,  cD4 + cXcr5 + PD1 + T Fh are specifically increased in Patients with igg4-rD The ectopic lymphoid structures found in tissues of patients with IgG4-RD led us to investigate for the presence of TFH  Figure 4A). This subset was significantly increased in patients with IgG4-RD when compared to HC and pSS patients in both percentages (3.7 ± 0.3% versus 1.7 ± 0.1 and 2 ± 0.2%; p < 0.0001 and p = 0.001, respectively) and numbers (35.4 ± 4.8 cells/μl versus 13.4 ± 0.9 and 13.9 ± 1.6 cells/μl; both p < 0.0001, respectively) ( Figure 4B). We also evaluated PD1 + TFH cell number in 11 IgG4-RD patients who were treated with steroids alone (n = 4), steroids plus azathioprine (n = 1), or steroids plus rituximab (n = 6). Interestingly, disease remission in these treated IgG4-RD patients was associated with reduced circulating TFH cells ( Figure 4C).

DiscUssiOn
IgG4-related disease is characterized by tissue infiltration by numerous IgG4 + plasmocytes, increased numbers of circulating plasmablasts and elevated titers of both serum IgG4, IgE, as well as other IgG subclasses (2). B cell activation appears to be T celldependent because activated B cells harbor enhanced somatic mutation, and relapses after rituximab (a B-cell-depleting agent) are characterized by the emergence of new plasmablast clones (17). T CD4 + cells are numerous in disease infiltrates and are thought to trigger B cell polyclonal expansion and fibrosis. Prior studies regarding blood and/or tissue analysis of IgG4-RD patients suggest changes of T helper cells, Treg, and more recently oligoclonal CD4 + effector/memory cytotoxic T lymphocytes (9-11, 18, 19). Hence, we conducted a systematic analysis of lymphocyte subsets in untreated patients with IgG4-RD. Major T cell subsets, including memory and naïve T cells, NK, and B cells were normal in patients with IgG4-RD compared to HC. Concordant with previous reports, plasmablast cell counts were higher in patients with IgG4-RD as compared to HC (17). Albeit not reaching significance, plasmablast cells were also higher in patients with IgG4-RD as compared to pSS patients. Next, in line with a previous study, there was a clear trend toward an increase of circulating CD4 + Foxp3 + Tregs in IgG4-RD patients, suggesting that in such patients Treg expansion is not restricted to diseased tissues (18). Treg were also shown to be increased in blood, without reaching significance, in an independent study (19). In IgG4-RD, Treg expansion might contribute to the genesis of fibrosis through TGF-β release and to the increase of IL-10 release by PBMCs. From that perspective, IgG4-RD is different than most autoimmune diseases where Treg are commonly decreased (20).
Over-production of TH2-related cytokines (IL-4, IL-13, and IL-5) in IgG4-RD tissues has previously been reported (11) and correlates with IgE and IgG4 class switch, blood eosinophilia, and eosinophilic infiltrates that are common features of IgG4-RD (4,11). Hence, and despite previous studies that provided conflicting results regarding the TH1/TH2 balance in PBMCs from IgG4-RD patients, it has been suggested that TH2 cells might be key players in the disease's pathophysiology (9,10). Yet, the central role of such cells is debated because IL-4 production upon re-stimulation has been shown to be restricted to CD4 + GATA3 + T cells in patients with known atopy, suggesting that the role of TH2 responses in IgG4-RD might be confounded by concomitant allergic disease (21). The present analysis of TH1, TH2, and TH17 cells in IgG4-RD showed an increase of both TH2 and TH17 cells. Interestingly, the increase of TH2 cells in the present study was not restricted to a unique organ involvement. Moreover, only 4 out of 28 patients (16%) with IgG4-RD also complied with the definition of the European Academy of Allergy and Clinical Immunology for atopy (22). TH2 and TFH2 cell number, IgE levels, and eosinophil numbers were not significantly different between atopic and non-atopic patients (data not shown). Hence, according to our data, TH2 cell expansion in IgG4-RD patients cannot be restricted only to patients with concomitant atopy. Next, the analysis of cytokine release by PBMCs upon stimulation were in line with these findings and showed that patients with IgG4-RD produced higher amounts of IL-4 but also of IL-10 and IL-17 as compared with HC and patients with pSS. Both TH2 and Treg are thought to contribute to IL-4 and IL-10 release, but we did not expect to report on both high IL-17 levels and elevated TH17 cells. Analysis of the co-expression of intracellular IL-17 and IL-4 after stimulation in CD4 + T cells showed that double positive cells were extremely rare, and therefore that the source of both cytokines were differentiated TH2 or TH17 cells (data not shown). Extranodal germinal centers are often found in IgG4-RD tissues, and IgG4 + plasmablasts harbor somatic hypermutation (17), a process occurring during B cell differentiation in germinal centers. Because TFH are essential for germinal center formation and Ig class switch in humans and mice, they could be implicated in the pathogenesis of IgG4-RD (23). The circulating TFH compartment in human is heterogeneous based on PD1, CCR6, and CXCR3 expression (23,24). Here, we show that the CD4 + CXCR5 + PD1 + TFH cells are specifically expanded in IgG4-RD. Interestingly, since only one patient presented with an increase of TFH after treatment with rituximab, expansion is reversible in patients with IgG4-RD that are efficiently treated. Expanded PD1 + TFH have been previously reported in human autoimmune diseases such as systemic lupus erythematosus, pSS, and juvenile dermatomyositis (25,26). In the latter, both TFH2 and TFH17 are increased and correlate with plasmablast numbers and disease activity (26). In our analysis, CXCR5 + PD1 + TFH cells were not expanded in untreated pSS patients but were expanded in IgG4-RD. TFH expansion was mostly related to the TFH2 CCR6 − CXCR3 − subset and to a lesser extent to the TFH17 CCR6 + CXCR3 − subset. TFH2 and TFH17 cells have been shown to efficiently induce in vitro naïve B cells to proliferate and differentiate into plasmablasts and produce all IgG subclasses, in contrast to TFH 1 (25). TFH2 specifically produces IL-4, IL-5, and IL-13, which are important cytokines for the class switching to IgE and IgG4. The expansion of TFH 2 is consistent with pathological and biological abnormalities reported in IgG4-RD patients. Our study showed that TFH2 cell numbers correlated positively with serum IgG4 (r = 0.64; p = 0.0004), IL-4 (r = 0.55; p = 0.01), and IL-10 (r = 0.49; p = 0.03) ( Table 3). Moreover, an increase of the CD4 + CXCR5 + CD45RA − TFH and TFH 2 cells in IgG4-RD has been reported in another series of 15 patients (14). However, PD1 expression was not analyzed. The specific expansion of CXCR5 + PD1 + TFH observed in our study could be related to some unique functional properties inherent to IgG4-RD's pathogenesis. Indeed, PD1 + TFH require less activation than PD1 − TFH to differentiate into functional helpers and, by opposition to PD1 − TFH, PD1 + TFH express low levels of CCR7 (24). The PD1 + CCR7 low TFH population is required for T cells to migrate into B cell follicles (27). Thus the specific expansion of PD1 + TFH in IgG4-RD could be an important trigger to B cell activation, class switch, and plasmablast generation. Interestingly, it has been shown in rheumatoid arthritis that PD1 + TFH is maintained by plasmablasts by an IL-6-dependent positive feedback loop that should be investigated in IgG4-RD (28). The findings reported in our study consist of correlations and causation of these T cells changes in the pathophysiology of IgG4-RD have to be confirmed by further functional studies. It has been recently shown in Japanese patients with predominant salivary and lachrymal glands involvement that CD4 + CD45RA − CXCR5 + CCR6 − CXCR3 − TFH2 cells were more efficient in inducing differentiation into plasmablasts and led to higher IgG4 production by autologous naïve B cells in active, untreated IgG4-RD patients than in HC (29), suggesting a functional role of these cells in the disease. The same authors found in a previous study a correlation between the increased number of circulating TFH2 cells and the number of plasmablasts (14), which was not found in our study. Conversely it is also plausible that these T cells changes are secondary to yet other unknown factor(s) (e.g., a source of TGF-β) that drives T cell differentiation and IgG4 production. In this line, mast cells have recently been shown to express IL-4, IL-10, and TGF-β (30), as well as IL-13 (31) in IgG4-RD tissues, and these innate cells could contribute to the TH2/T regulatory cytokines orientation reported in the disease.
Major cytokines involved in the early TFH differentiation process from CD4 + T cells in human, including IL-12, IL-23, and TGF-β, are also supported by other STAT3-activating cytokines including IL-6, IL-21, and IL1-β (23). In human autoimmune diseases, both TH17 and TFH co-emerge and share a developmental mechanism induced by TGF-β. It has been proposed that abundant expression of TGF-β in inflammatory sites in human autoimmune diseases (28), where tertiary lymphoid organs are often formed, contribute to the generation of TFH and TH17 cells (24). Hence, the expansion of these cells could be the consequence of an initial inflammatory process. In tissues, TH17-related molecules have been reported in salivary glands of patients with IgG4-RD, albeit at low levels (32). The site where the differentiation and expansion of TFH occurs in IgG4-RD is unknown, and no link has been established with IL-1β and TGF-β producing clonal expanded CD4 + SLAMF7 + CTLs (19).
Interestingly, our results showed that PD1 + TFH cells were significantly decreased in patients with IgG4-RD who were treated with either steroids alone, steroids plus azathioprine, or steroids plus rituximab, and that the decrease in PD1 + TFH cells was always associated with clinical improvement of the disease. Significant variation of TFH cells after treatment with steroids has also been recently reported in Japanese patients with modification of TFH1 cells, which we did not observe in our larger study (29). However, organ involvement was significantly different from our study. Interestingly, the PD1 + TFH subset also comprises TFH regulatory cells co-expressing Foxp3 and an imbalance of TFH/TFH Foxp3 regulatory could also be implicated in the disease's pathophysiology and should be further explored (33).

cOnclUsiOn
We show that patients with active, untreated IgG4-RD presenting with variable organ involvements present specific changes of T cells in peripheral blood. Circulating PD1 + TFH are expanded together with TH2 and TH17, associated with elevated IL-4, IL-10, and IL-17 release. Both PD1 + Tfh2 and Tfh17 subsets have been shown to be prone to migrate into B cell follicles and/or inflammatory sites and could therefore contribute to B cell activation and IgG4 class switch in IgG4-RD tissues. Yet, the precise role of these T cell changes together with the role of clonal expanded cytotoxic T cells, their interactions with B cells and other immune cells, and the induction of fibrosis remains to be further assessed in IgG4-RD. Indeed, in a near future, it is likely that specific treatments targeting T cells will be assessed in IgG4-RD. aUThOr cOnTribUTiOns AG, ME, FV, and NS: conception and design of the work; acquisition, analysis, and interpretation of data for the work; drafting the work and revising it critically for important intellectual content. CP, CF, J-RH, and NC-C: acquisition, analysis, and interpretation of data for the work; revising the work critically for important intellectual content. MG, AR, MS, BT, NM, SA, FM, JG, MH, AF, SP, EB, and BB: acquisition of data for the work; revising the work critically for important intellectual content. All authors meet following criteria: final approval of the version to