Distinct disease-specific Tfh cell populations in two different fibrotic diseases: IgG4-related disease and Kimura’s disease

Abstract

Background:

How T follicular (Tfh) cells contribute to many different B-cell class-switching events during T cell-dependent immune responses has been unclear. Diseases with polarized isotype switching offer a unique opportunity for the exploration of Tfh subsets. Secondary and tertiary lymphoid organs (SLOs and TLOs) in patients with elevated tissue expression levels of IgE (Kimura’s disease, KD) and those of IgG4 (IgG4-related disease, IgG4-RD) can provide important insights regarding cytokine expression by Tfh cells.

Objective:

To identify disease-specific Tfh cell subsets in SLOs and TLOs expressing IL-10 or IL-13 and thus identify different cellular drivers of class switching in two distinct types of fibrotic disorders: allergic fibrosis (driven by type 2 immune cells) and inflammatory fibrosis (driven by cytotoxic T lymphocytes).

Methods:

Single-cell RNA-sequencing, in situ sequencing, and multi-color immunofluorescence analysis was used to investigate B cells, Tfh cells and infiltrating type 2 cells in lesion tissues from patients with KD or IgG4-RD.

Results:

Infiltrating Tfh cells in TLOs from IgG4-RD were divided into six main clusters. We encountered abundant infiltrating IL-10-expressing LAG3⁺ Tfh cells in patients with IgG4-RD. Furthermore, we found that infiltrating AID⁺CD19⁺B cells expressing IL-4, IL-10, and IL-21 receptors correlated with IgG4 expression. In contrast, we found that infiltrating IL-13-expressing Tfh cells were abundant in affected tissues from patients with KD. Moreover, we observed few infiltrating IL-13-expressing Tfh cells in tissues from patients with IgG4-RD, despite high serum levels of IgE (but low IgE in the disease lesions). Cytotoxic T cells were abundant in IgG4-RD, and in contrast Type 2 immune cells were abundant in KD.

Conclusions:

This single-cell dataset revealed a novel subset of IL10⁺LAG3⁺Tfh cells infiltrating the affected organs of IgG4-RD patients. In contrast, IL13⁺Tfh cells and type 2 immune cells infiltrated those of KD patients.

Capsule Summary:

Single-cell RNA sequencing analysis identified a distinct Tfh cell subset in tertiary lymphoid organs in which plasma cells predominantly express IgG4.

INTRODUCTION

T follicular (Tfh) cells are a subset of CD4⁺ T cells that assist B cells during T cell-dependent immune responses and contribute to isotype switching, somatic hypermutation, germinal center formation, and high-affinity B-cell selection in germinal centers ^{1, 2}. Tfh cells are distinguished from other CD4⁺ T cells based on their expression of ICOS and Bcl6. Tfh cells were originally observed in the light zones of germinal centers, but broadly similar cells have also been observed outside follicles at the T cell zone–B cell follicle interface (T-B interface). While some studies have referred to these broadly similar cells as Tfh cells, this is not universally accepted, as these cells reside outside the follicle; we have also referred to these cells as pre-germinal center Tfh cells ³. We and other groups have argued that class switching events also occur outside the follicle ^{4, 5}. Nevertheless, how Tfh cells or Tfh-like cells contribute to class-switching events at more than one location remains unclear.

Immunoglobulin G4 (IgG4)-related disease (IgG4-RD) is a fibrotic, systemic, inflammatory disease of unknown etiology ⁶. The expansion of circulating plasmablasts, most of which express IgG4, is a hallmark of active IgG4-RD ⁷. These blood plasmablasts are heavily somatically hypermutated, implying that they may be derived from germinal centers with assistance from Tfh cells. Histological analyses have shown that ectopic germinal centers frequently occur in affected salivary glands in patients with IgG4-RD⁸. IgG4-RD is a disease that involves polarized class switching to IgG4, but many patients also have elevated serum levels of IgG1, IgE, or both ⁶. Notably, patients with non-allergic disease who have IgG4-RD also show elevated serum IgE ⁹. Although the mechanisms underlying the switch to IgG4 in IgG4-RD are poorly understood, IL-10 is presumed to indirectly contribute to IgG4 class switching by facilitating IL-4-mediated switching to IgG4, rather than to IgE ¹⁰. Our previous multi-color immunofluorescence analysis revealed that IL-4⁺ Tfh cells were expanded in patients with IgG4-RD ⁵. However, our previous studies were not performed at the single-cell level. Furthermore, Tfh cells expressing other cytokines have not been investigated in secondary lymphoid organs (SLOs) and tertiary lymphoid organs (TLOs) from IgG4-RD patients.

Kimura’s disease (KD) is a rare, chronic inflammatory disorder that is characterized by subcutaneous eosinophilic lymph follicular granuloma in affected tissues and high serum IgE levels. Other similar manifestations have been observed in KD and IgG4-RD. Ectopic germinal center formation ¹¹, infiltrating tissue eosinophilia, peripheral eosinophilia, or high serum IgE levels are often found in IgG4-RD patients, especially among those with manifestations such as sialoadenitis, dacyoadenitis, and orbital disease. From these clinicopathologic observations, Th2 cells and allergic triggers were once hypothesized to be important in the pathogenesis of both diseases, although there is little evidence for Th2 cells accumulating in tissues in IgG4-RD ^{8, 9, 12, 13}. Gowthaman et al. showed that IL-4⁺ Tfh cells induce direct switching of B cells to low-affinity IgE during a subset of type 2 immune responses, but IL-13-secreting Tfh (Tfh13) cells produce additional signals that regulate high-affinity IgE during allergic responses ¹⁴. Tfh13 cells are distinguished by their cytokine products (e.g., IL-4, IL-5, and IL-13) and expression of GATA3. Tfh13 cells were found within germinal centers in mice and were also expanded in blood from human patients with allergies ¹⁴.

Both KD and IgG4-RD lesions are characterized by fibrosis ^{6, 11}. Fibrosis is the end result of chronic inflammatory reactions induced by various stimuli including persistent infections, autoimmune reactions, allergic responses, radiation, and tissue injury. Fibrotic diseases likely have many different etiologies, and they may not all be driven by CD4⁺ T cells. In our previous study examining the etiology of tissue fibrosis in patients with IgG4-RD, we found that recurrent apoptotic cell death, driven by the recognition of self-peptides by autoreactive CD4⁺ cytotoxic T lymphocyte (CTL) and CD8⁺ CTL clones, may contribute to cell loss and subsequent extensive tissue remodeling, leading to fibrosis and organ dysfunction ¹⁵. In contrast, type 2 immune cells and their cytokines (IL-4, IL-5, and IL-13) are critical in the pathogenesis of allergic inflammation and fibrosis ¹⁶. Type 2 immunity induces a complex inflammatory response characterized by eosinophils, mast cells, basophils, type 2 innate lymphoid cells, Th2 cells, and specific IgE antibody subclasses that are crucial to the pathogenesis of many allergic and fibrotic disorders.

This study investigated two chronic human diseases with polarized isotype switching to obtain insights regarding unique Tfh cell subsets. We found that SLOs and TLOs in patients with elevated tissue expression levels of IgE and IgG4 represent useful substrates for examining distinct cytokine expression by Tfh and Tfh-like cells in the context of KD and IgG4-RD. We also identified distinct disease-specific Tfh cell subsets expressing IL-10 in IgG4-RD patients. Furthermore, we characterized differences in the etiologies of inflammatory and allergic fibrosis and identified two different types of fibrosis: allergic fibrosis (driven by type 2 cells) and inflammatory fibrosis (driven by cytotoxic T cells).

Materials and methods

This study included 11 KD patients, 25 IgG4-RD patients, 16 Sjögren’s syndrome (SjS) patients, and 10 chronic sialoadenitis (CS) patients. CS is a non-specific inflammatory disease of the salivary glands linked to sialolithiasis. Patients were followed up between 2010 and 2021 at the Department of Oral and Maxillofacial Surgery of Kyushu University Hospital, a tertiary care center. Open submandibular gland (SMG) biopsies were obtained from IgG4-RD patients, while CS patients underwent submandibulectomy. Lip biopsies were obtained from SjS patients. IgG4-RD was diagnosed according to previously established criteria ¹⁷. SjS was diagnosed as previously described ¹⁸. Each patient exhibited objective evidence of salivary gland involvement based on the presence of subjective xerostomia and a decreased salivary flow rate, abnormal findings on parotid sialography, and focal lymphocytic infiltrates in labial salivary glands. None of the patients had a history of treatment with steroids or other immunosuppressants or infection with HIV, HBV, or hepatitis C virus; no patient had sarcoidosis or evidence of lymphoma at the time of the study. All patients had strong lymphocytic infiltration in these tissues. This study protocol was approved by the Institutional Review Board of the Center for Clinical and Translational Research of Kyushu University Hospital (approval no. 834-00) and followed the tenets of the Declaration of Helsinki. All participants provided written informed consent. Infiltrating immune cells and sorted CD3⁺ T cells and CD19⁺B cells from four IgG4-DS patients were used for scRNA-seq. Experimental procedures for scRNA-seq followed established techniques using the Chromium Single Cell 5′ Library V2 kit (10x Genomics). To obtain the T-cell and B-cell receptor repertoire profile from three IgG4-RD patients, VDJ enrichment for B-cell receptors was carried out with Chromium Single Cell V(D)J Enrichment Kit, Human T Cell (10× Genomics). Paraffin-embedded salivary gland sections from 21 IgG4-DS patients and 11 KD patients were used for multi-color immunofluorescence staining, performed as previously described ¹⁹. Images of tissue specimens were acquired using the TissueFAXS platform (TissueGnostics). Cells of a particular phenotype were identified and quantified using the TissueQuest software (TissueGnostics) ²⁰. See online supplemental patients and methods for details.

RESULTS

B cells activated to switch to IgG4 are prominent in affected lesions from IgG4-RD patients, whereas IgE⁺ B cells are prominent in KD

Serum IgG4 and IgG levels were higher in IgG4-RD patients than in KD patients. KD patients exhibited an elevated eosinophil count and elevated IgE concentration (Supplemental E Table 1–3). Ectopic germinal centers were frequently observed in affected tissue sites in IgG4-RD and KD patients but were absent or sparse in SjS and CS patients (Fig. 1A, see also Supplemental E Table 4). We next analyzed SLOs and affected TLOs from KD patients and IgG4-RD patients. TLOs with germinal centers ²¹ were identified by multi-color immunofluorescence approaches (CD4, CD19, Bcl6, and DAPI expressions), and ectopic germinal centers were also identified using multi-color immunofluorescence approaches (CD19, CXCR5, and Bcl6 expressions) (Fig. 1B). We found that 7 of 11 (63.6%) KD patients, 18 of 25 (72%) IgG4-RD patients, and 2 of 17 (11.8%) SjS patients had ectopic germinal centers in affected lesions (Supplemental E Table 4). High frequencies and numbers of ectopic germinal centers were observed in affected lesions of KD patients and IgG4-RD patients.

Figure 1. B cells express IgG4 in affected lesions from patients with IgG4-RD in contrast to IgE-expressing B cells in patients with KD.

(A) Ectopic germinal center formation in salivary gland sections from patients with Kimura’s disease (KD), IgG4-related disease (IgG4-RD), Sjogren’s syndrome (SjS), and chronic sialadenitis (CS).

(B) Top: Multi-color immunofluorescence staining of CD4 (red), Bcl6 (green) and CD19 (blue) in a tertiary lymphoid organ (TLO) from a patient with KD. Bottom: Multi-color immunofluorescence staining for CD19 (red), CXCR5 (green), and Bcl6 (magenta) in ectopic germinal centers of salivary glands from a patient with KD.

(C) Immunostaining with IgE and IgG4 monoclonal antibodies in affected salivary glands from a patient with KD and a patient with IgG4-RD.

(D) Multi-color immunofluorescence staining of CD19 (red), IgE (green), IgG4 (magenta), and DAPI (blue) in a draining lymph node from a patient with KD. Quantification of IgE⁺CD19⁺ B cells and IgG4⁺CD19⁺ B cells in draining lymph nodes from five patients with KD, eight patients with IgG4-RD, and six control tonsils. p-value was determined using Mann–Whitney U test.

Germinal center reaction is a multi-step process during which naïve B cells become Ig-producing cells. We thus analyzed tissue infiltrating Ig-producing B cells in TLOs from KD patients and IgG4-RD patients. IgE-positive cells were mainly localized in interfollicular areas in TLOs from KD patients (Fig. 1C). IgE was expressed either on the plasma membrane or in the cytoplasm. In KD patients, substantial numbers of IgE-positive cells were detected in and around germinal centers; however, these cells were absent or sparse in IgG4-RD patients. In contrast, IgG4-positive cells were abundant in IgG4-RD patients but not in KD patients.

We next analyzed draining lymph nodes from KD patients and IgG4-RD patients to evaluate B-cell class switching to IgE and/or IgG4. Lymph nodes were stained using antibodies to IgE, IgG4, and CD19 and DAPI as a nuclear stain (Fig. 1D). IgE⁺CD19⁺ B cells represented fewer than 5% of all CD19⁺ B cells in normal tonsils from healthy controls and lymph nodes from IgG4-RD patients. In contrast, IgE⁺CD19⁺ B cells comprised approximately 30% of CD19⁺ B cells in KD patients. Furthermore, IgG4⁺CD19⁺ B cells represented fewer than 6% of all CD19⁺ B cells in normal tonsils from healthy controls and lymph nodes from KD patients. In contrast, IgG4⁺CD19⁺ B cells comprised approximately 40% of CD19⁺ B cells in IgG4-RD patients.

Overall, elevated serum IgG4 levels and numbers of IgG4-positive B cells in germinal centers were observed in IgG4-RD patients, suggesting that B cells may have been activated by disease-specific Tfh cells. Class switching is linked to Tfh cells in SLOs and TLOs and not to other tissue infiltrating CD4⁺ T cells subsets.

CD4⁺CXCR5⁺ Tfh cells are located near AICDA⁺ B cells

AICDA (also known as AID) encodes an enzyme with roles in class switching, recombination, and somatic hypermutation. To assess spatial gene expression in AICDA⁺CD19⁺ B cells in affected lesions, we performed spatial transcriptomics analysis (10X Genomics) of tissue sections from a patient with IgG4-RD. Transcriptomes from 1,914 spots in a single section were obtained, yielding a median of 2,722 genes per spot (Supplemental E Fig. 1A). These 1,914 spots were divided into seven cluster types using tSNE visualization (Supplemental E Fig. 1A). One spot included approximately 6–10 cells. We initially focused on AICDA⁺CD19⁺ B-cell cluster spots. As shown in Supplemental E Fig. 1B, cells in cluster_6 expressed high levels of SERPINA9, KLHL6, BIK, CD22, RGS13, LRMP, and ELL3; these genes were co-expressed with AICDA. Cells in cluster_6 also expressed high levels of CD4, CXCR5, PDCD1, Bcl6, CD40, and CXCL13; these genes are characteristic of Tfh cells. Notably, CD4⁺CXCR5⁺ Tfh cell spots were located near AICDA⁺ B cell spots.

Potential recruitment mechanisms for infiltrating Tfh and B cells were also assessed by analyzing chemokine and receptor expression by spatial transcriptomics (Supplemental E Fig. 1C). At the T-B interface, transcripts linked to CXCL13-CXCR5 signaling were upregulated in cluster_6, suggesting the induction of the migration of Tfh cells into the follicle. At this T-B interface in cluster_6, significant up-regulation of genes involved in T and B cell receptor signaling pathway was observed (Supplemental E Fig. 1D).

As shown in Supplemental E Fig. 1E, AID-expressing B cells were visualized both inside and outside germinal centers in IgG4-RD sections. Furthermore, T cells expressing ICOS (a marker of Tfh cells) were physically close to AID-expressing B cells in IgG4-RD TLOs.

GATA3-expressing CD4⁺CXCR5⁺ Tfh-like and Th2 cells are prominent in KD tissues, while GATA3-negative CD4⁺CXCR5⁺ Tfh-like and CD4⁺ cytotoxic T cells are prominent in IgG4-RD tissues

T cells are implicated in the pathogenesis of KD, IgG4-RD, and other immune-related diseases for multiple reasons and primarily because many CD4⁺ T cells are present in affected tissues (Fig. 2A). To explore the relevance of T and B cells in the pathogenesis of KD and IgG4-RD, we quantified CD3⁺ T cell subsets in affected lesions from KD and IgG4-RD patients (Fig. 2B). We previously found that IgG4-RD patients had expanded infiltrating CD4⁺GZMA⁺ CTLs in affected lesions ²². We observed a striking dominance of CD4⁺GZMA⁺ CTLs in affected lesions from IgG4-RD patients; however, these CTLs were sparse in affected lesions from KD patients (Fig. 2C). Although other CD4⁺ T-cell subsets have been implicated in the pathogenesis of KD, comprehensive tissue quantitative approaches have not been previously reported. We therefore quantified all major CD4⁺ T-cell subsets, including Th1, Th2, Th17, and Tfh cells, as well as Tregs and CD4⁺ CTLs. Most T cells in KD patients were Th2 and Tfh cells (Supplemental E Fig. 2A–C). GATA3-expressing CD4⁺CXCR5⁺ Tfh cells were abundant in tissue lesions of KD patients, but were sparse in IgG4-RD (Fig. 2D). In contrast, CD4⁺CTLs and GATA3-negative Tfh cells were abundant in tissue lesions of IgG4-RD patients. GATA3-expressing Tfh cells might therefore represent important disease-related type 2–Tfh cells in KD patients.

Figure 2. Quantification of CD4⁺ T cell subsets in patients with KD and patients with IgG4-RD.

(A) Immunofluorescence staining of CD4 (red) and CD8 (yellow) T cells and absolute numbers of CD4⁺ T cells and CD8⁺ T cells per mm² in affected lesions from six patients with KD. p-value was determined using Student’s t-test.

(B) Absolute numbers of CD4⁺ T cells and CD8⁺ T cells per mm² in affected tissues from patients with KD (n=6) and patients with IgG4-RD (n=20). p-value was determined using Student’s t-test.

(C) Left: Immunofluorescence staining of CD4 (red), CD8 (magenta), GZMA (green), and DAPI (blue) in affected tissues from patients with KD. Quantification of CD4⁺GZMA⁺ CTL cells in affected tissues from patients with KD (n=6) and IgG4-RD (n=21). p-value was determined using Student’s t-test. Right: Immunofluorescence staining of CD4 (red), CXCR5 (green), GATA3 (white), and DAPI (blue) in affected tissue from a patient with KD.

(D) Multi-color immunofluorescence staining for Th2, GATA3⁺Tfh, GATA3⁻Tfh, CD4⁺CTLs and CD4⁺Other cells were done. Images of tissue specimens were acquired using the Tissue FAXS platform (also see Supplemental E). Left; Absolute numbers (per mm²) of and Relative proportions of CD4⁺GATA3⁺CXCR5-negative Th2 cells (blue), CD4⁺CXCR5⁺GATA3-positive Tfh cells (red), CD4⁺CXCR5⁺GATA3-negative Tfh cells (green), CD4⁺GZMA⁺ CTLs (magenta), and CD4⁺ Other cells (orange) in affected tissues from 8 patients with KD and 20 patients with IgG4-RD. Right; Relative proportions of Th2 cells, GATA3-positive Tfh cells, GATA3-negative Tfh cells, CD4⁺CTLs, and CD4⁺ Other cells between patients with KD (n=8) and patients with IgG4-RD (n=20). Multiple comparisons are controlled for by Kruskal-Wallis test. Data are presented as mean ± SEM.

scRNA-seq of tissue infiltrating CD4⁺CXCR5⁺ Tfh-like cells in IgG4-RD

Although visualization by multi-color staining permits anatomic localization of Tfh cells in tissue, it provides limited information. To better understand the other cytokine-expressing Tfh cells in affected lesions with TLOs, we performed scRNA-seq analysis (10X Genomics) of infiltrating Tfh cells from four IgG4-RD patients. The schematic strategy to study tissue infiltrating cells in IgG4-RD is presented in Figure 3A. We visualized selected marker genes in these gated Tfh cells. As shown in Figure 3B, infiltrating CD4⁺CXCR5⁺ Tfh cells from IgG4-RD tissues expressed MAF, CD40LG, cytotoxic T lymphocyte-associated protein 4 (CTLA4), PDCD1, and ICOS; these cells also expressed IL10, IL21, and CXCL13.

Figure 3. Transcriptomic profiling of infiltrating CD4⁺CXCR5⁺Tfh cells in affected SGs from IgG4-RD.

(A) Sorted CD3+CD19− T cells from affected SGs from patients with IgG4-RD were gated into CD4+CD8A-CXCR5+ Tfh-like cells (n=4).

(B) Single-cell RNA-seq profiling revealed the expression patterns of Tfh-related cytokine-, chemokine-, chemokine receptor-, membrane-, and transcription factor–related genes in CD4⁺CXCR5⁺ Tfh cells in affected SGs from patients with IgG4-RD (n=4).

(C) Six clusters across 1,853 CD4⁺CXCR5⁺ Tfh cells, identified via tSNE visualization. Schematic representation of single-cell transcriptome experiments from submandibular gland sample collection to clustering of CD4⁺CXCR5⁺ Tfh-like cells in a patient with IgG4-RD (Pt4).

(D) Left: Heat map shows top 10 gene expression patterns in CD4⁺CXCR5⁺ Tfh cells among the six clusters in a patient with IgG4-RD (Pt4). Right: Selected gene expression patterns in CD4⁺CXCR5⁺ Tfh cells in the six clusters in a patient with IgG4-RD (Pt4).

We compared expression values in the CD4⁺CXCR5⁺Tfh cell cluster to selected cluster marker genes. The expression values of selected genes in each CD4⁺CXCR5⁺Tfh cell, based on tSNE analyses, are shown in Figure 3C. The CD4⁺CXCR5⁺Tfh cell phenotype was clustered into six distinct subtypes. Cells in cluster_0 were characterized by high expression levels of CXCL13, PDCD1, IL21, BCL6, and ICOS, suggesting they were “true” GC Tfh cells (Fig. 3D). Cells in cluster_1 were characterized by high expression levels of IL10, LAG3, CTLA4, and PRDM1. Cells in cluster_2 were characterized by high expression levels of FOXP3 and CTLA4, suggesting they were regulatory follicular helper T (Tfr) phenotype cells ^{23, 24}. Cells in cluster_4 and cluster_5 were characterized by high expression levels of GZMA, GZMB, GZMK, GZMH, GZMM, PRF1, CRTAM, and SLAMF7, suggesting that were cytotoxic phenotype cells.

IL-10-expressing Tfh cells are prominent in IgG4-RD patients

We presumed that the IL-10-expressing LAG3⁺ Tfh cells could be a subset distinct from Tfr cells ^{23, 24}. We used scRNA-seq to compare the transcriptomes between IL10⁺CD4⁺ T cells and IL10⁻CD4⁺ T cells in an affected lesion from a patient with IgG4-RD (Fig. 4A). IL10⁺CD4⁺ T cells expressed significantly higher levels of LAG3 and IL21 than did IL10⁻CD4⁺ T cells. Although they did not express FOXP3 and EGR2, IL10⁺CD4⁺ T cells co-expressed ICOS, LAG3, PRDM1, MAF, and CXCR5; these were presumably Tfh-like phenotype cells that mainly produced IL-10 (Fig. 4B). We thus suspected that most IL-10-secreting CD4⁺ cells near germinal centers in tissues from IgG4-RD patients were Tfh-like cells. We also noted that, despite the absence of FOXP3 expression, the IL-10⁺ Tfh cells expressed CTLA4 and exhibited low IL2 expression, which typically suppresses conventional Treg function. CTLA4 is a key transcriptional target of FOXP3 in Tregs; robust CTLA4 expression combined with low IL-2 production is a good marker of Treg cell function, despite the absence of FOXP3 expression ²⁵. These results suggested that IL-10⁺ Tfh cells in IgG4-RD patients exhibit characteristics of Treg cells but are cells of the Tfh lineage.

Figure 4. Transcriptomic profiling of IL10-expressing LAG3⁺ Tfh cells and cytotoxic Tfh cells in affected SGs from IgG4-RD.

(A) Volcano plot showing single-cell RNA-seq analysis of gene expression in IL10-expressing and non-expressing CD4⁺ T cells in affected tissue from a patient with IgG4-RD. Differential expression patterns of 28,000 genes were investigated by scRNA-seq. Violin plot showing LAG3 gene expression in IL10-expressing and non-expressing CD4⁺ T cells in affected tissue from a patient with IgG4-RD.

(B) Single-cell RNA-seq profiling revealed the expression patterns of Tfh-related cytokine-, chemokine-, membrane-, and transcription factor–related genes in IL10⁺CD4⁺CXCR5⁺ Tfh cells in affected tissues from patients with IgG4-RD (n=4).

(C) Left: Absolute numbers of IL-10⁺CD4⁺CXCR5⁺ Tfh cells in affected tissues from patients with KD (n=7) and patients with IgG4-RD (n=7). Right: Relative proportions of CD4⁺CXCR5⁺ Tfh cells expressing IL-10 in affected tissues from patients with KD (n=7) and patients with IgG4-RD (n=7). p-value was determined using Student’s t-test.

(D) Left: Absolute numbers of IL-10⁺CD4⁺CXCR5⁺ Tfh cells in draining lymph nodes from 8 patients with IgG4-RD and tonsils from 10 healthy controls. Relative proportions of CD4⁺CXCR5⁺ Tfh cells expressing IL-10 in draining lymph nodes from 8 patients with IgG4-RD and tonsils from 10 healthy controls. p-value was determined using Student’s t-test.

We thus applied multi-color imaging for the IL10-expressing CD4⁺CXCR5⁺Tfh cells in affected lesions from IgG4-RD patients and KD patients. Approximately 17% of CD4⁺CXCR5⁺ Tfh cells in IgG4-RD patients expressed IL-10, while fewer than 7% of CD4⁺CXCR5⁺ Tfh cells expressed IL-10 in KD patients (Fig. 4C). We next examined whether infiltrating Tfh cells expressed IL-10 in normal tonsils and lymph nodes from IgG4-RD patients. Approximately 20% of CD4⁺CXCR5⁺ Tfh cells expressed IL-10 in IgG4-RD patients (Fig. 4D). The low frequency of IL-10 expression among Tfh cells was confirmed in normal tonsils. Taken together, these results suggested that the expanded IL-10-expressing Tfh cells in IgG4-RD patients exhibit a gene expression profile that allows localization in and around germinal centers; they also support Treg cell function and abundant expression of IL-10.

Cytotoxic Tfh cells are detected in IgG4-RD patients

In our previous studies, we found clonally expanded CD4⁺CTLs in IgG4-RD patients. Here, we found that cytotoxic phenotype Tfh cells in IgG4-RD shared some genes (SLAMF7, CRTAM, NKG7, GZMA, GZMK, CCL4 and CCL5) with cytotoxic cells (Fig. 3D and Fig. 4E). We also noted that that CD4⁺CXCR5⁺GXMK⁺Tfh cells were cell-cell contact with CD19⁺B cells, as confirmed by nuclear distance measurements (Fig. 4F). Approximately 19% of CD4⁺CXCR5⁺Tfh cells expressed GZMK in an IgG4-RD patient.

IL-13-expressing GATA3⁺ Tfh (Tfh13) cells are prominent in KD patients and sparse in IgG4-RD patients

Tfh cells provide help to B cells during T cell-dependent immune reactions, and they contribute to isotype switching, somatic hypermutation, memory B cell generation, plasma cell differentiation and germinal center formation ^{1, 2}. To identify potentially pathogenic subsets of CD4⁺ T cells that were clonally expanded in response to a potential antigen, we analyzed T-cell receptor β chain gene rearrangement using next-generation sequencing. We previously reported that IgG4-RD patients exhibit large clonal expansions of CD4⁺ CTLs that are the dominant T cells in diseased tissues ^{13, 22}. As shown in Fig. 2D, Tfh and Th2 cells were the dominant T cell subsets in affected tissues from KD patients. We therefore initially analyzed the T-cell receptor β chain gene repertoire of circulating CD4⁺CXCR5⁺ Tfh (cTfh) cells and CD4⁺CXCR5⁻CCR6⁻CXCR3⁻Th2 (cTh2) cells from the blood of a KD patient (Fig. 5A). We found that cTfh and cTh2 cells from the KD patients were more oligoclonally expanded compared with circulating naïve CD4⁺ T cells. cTfh cells from two IgG4-RD patients were also oligoclonally expanded, while the repertoire of naïve CD4⁺ T cells from two healthy controls was highly diverse (no dominant clones were observed in either individual) (Fig. 5B). We next performed bulk RNA-seq analysis of sorted activated PD1⁺CD4⁺CXCR5⁺ Tfh cells from a KD patient and an IgG4-RD patient (Fig. 5C). The circulating PD1⁺ Tfh cells from the KD patient expressed higher levels of IL4, IL5, IL13, STAT6, and GATA3 (all associated with type 2 immunity) compared with those in the IgG4-RD patient.

Figure 5. IL-13-expressing Tfh cells are abundant in patients with KD.

(A) Stacked bar chart indicating frequencies of top 10 T-cell receptor (TCR)β clones in circulating CD4⁺CXCR5⁺ Tfh cells and circulating CD4⁺CXCR5⁻CXCR3⁻CCR6⁻Th2 cells from a patient with KD and circulating CD4⁺CD27⁻CD62L⁻ effector memory T (T_EM) cells from a healthy control.

(B) Distributions of clone frequencies in circulating CD4⁺CXCR5⁺ Tfh cells from two patients with IgG4-RD (blue and green) and a patient with KD (red) compared with T_EM cell frequencies from two healthy controls (yellow and gray). Minimum numbers of clones comprising 10% (D10) and 50% (D50) of clone diversity are shown.

(C) Heat map depicting differentially expressed Th2- and Tfh-related genes in circulating CD4⁺CXCR5⁺PD-1⁺ Tfh cells from patients with KD and IgG4-RD (n=1 each).

(D) Multi-color immunofluorescence staining of PD-1 (red), CXCR5 (green), ICOS (magenta), and DAPI (blue) in tissue inside and outside a germinal center in a patient with KD.

(E) Relative proportions of ICOS⁺ T cells expressing GATA3 in tissue from patients with KD (n=7) and IgG4-RD (n=10). p-value was determined using Student’s t-test.

(F) Multi-color immunofluorescence staining of ICOS (green), GATA3 (magenta), IL-13 (red), and DAPI (blue) in affected tissue from a patient with KD. Scatter plots depict mean fluorescence intensity per cell for each immunostained molecule in tissue from a patient with KD. Quantification of ICOS⁺GATA3⁺IL-13⁺ T cells and ICOS⁺GATA3⁺IL-5⁺ T cells in tissue from a patient with KD.

(G) Multi-color immunofluorescence staining of CD4 (red), CXCR5 (green), IL-13 (magenta), and DAPI (blue) in tissue from a patient with KD. Relative proportions of CD4⁺CXCR5⁺ Tfh cells expressing IL-13 in tissue from patients with KD (n=7) and IgG4-RD (n=7). p-value was determined using Student’s t-test.

(H) Higher percentage of IL-13⁺ Tfh cells in tissues from patients with KD (n=7) correlated with higher serum IgE concentrations, compared with IgG4-RD (n=7). The correlation coefficients and p-value were determined using Spearman’s rank correlations.

We next performed in situ analyses of TLOs in affected lesions from KD and IgG4-RD patients to identify activated Tfh phenotype cells. Notably, activated PD1⁺ICOS⁺CXCR5⁺ Tfh cells were detected near TLO-like structures containing germinal centers (Fig. 5D). GATA3-expressing ICOS⁺ T cells, presumably Tfh phenotype cells, were also detected near TLO-like structures containing germinal centers. These GATA3⁺ ICOS⁺ T cells comprised approximately 40% of ICOS⁺ T cells in affected tissue lesions from KD patients, while they comprised fewer than 10% of ICOS⁺ T cells from IgG4-RD patients (Fig. 5E). Finally, we examined TLOs in affected lesions from KD patients to quantify Tfh cells that expressed IL-13 in situ. IL-13⁺ICOS⁺GATA3⁺ T cells were abundant in KD patients. In a lesion from a patient with KD, approximately 54% and 12% of ICOS⁺GATA3⁺ T cells expressed IL-13 and IL-5, respectively (Fig. 5F). IL-13-expressing CD4⁺CXCR5⁺ Tfh cells were also abundant in KD tissue lesions, although they were rare in IgG4-RD tissue lesions (Fig. 5G). Most IL-13-expressing Tfh cells were located near TLO-like structures in tissues from KD patients. Higher serum IgE concentrations were present in KD patients who had greater proportions of IL-13⁺ Tfh cells in these tissues (Fig. 5H). We speculate that the expansion of these IL-13-expressing Tfh cells, presumably the IL-13-expressing GATA3⁺ Tfh13 cells associated with high-affinity IgE ¹⁴, might represent an important disease-related Tfh subset in KD patients.

IgG1- and IgG4-expressing B cells in IgG4-RD tissues were oligoclonally expanded

To identify potentially pathogenic subsets of infiltrating B cells that were clonally expanded in response to a potential antigen, we analyzed B cell receptor chain gene rearrangements using single-cell repertoire analysis (10X Genomics). The schematic strategy for B cell receptor repertoire to study tissue infiltrating cells from salivary glands from three IgG4-RD patients is shown in Figure 6A. We analyzed the frequencies of Ig isotypes using B-cell receptor repertoire analysis in affected lesions from a patient with IgG4-RD. Large clonal expansions of IgG1- and IgG4-expressing B cells were particularly prominent (Supplemental E Figure 3A). In affected IgG4-RD lesions with TLOs, both IGHG1 and IGHG4 were abundant among class-switched isotypes, but IGHE was sparse (Fig. 6B). We hypothesized that antigen-driven B cells would demonstrate some degree of B cell receptor sharing and that this may be reflected in oligoclonality. However, we found no evidence for this. The V-J gene segment usage of expanded clones was not identical across three IgG4-RD subjects and there were no clones with shared CDR3 sequences across individuals (Supplemental E Fig. 3B).

Figure 6. Single-cell B-cell receptor analysis of affected lesions from patients with IgG4-RD.

(A) Sorted CD3-CD19+ B cells from affected SGs with IgG4-RD were examined for B-cell receptor repertoire analyses.

(B) Stacked bar chart indicating barcodes frequencies of top 10 B-cell receptor clones of Ig heavy chain in B cells, from 3 patients with IgG4-RD according to scRNA-seq and B-cell receptor profiling, using 10X Genomics Loupe V(D)J Browser.

The frequency of AID⁺ B cells expressing receptors for IL4, IL-10, and IL-21 correlates with IgG4 expression

We compared expression values in the B-cell cluster to selected cluster marker genes. For these selected genes, expression values in each B cell, based on tSNE analysis, are shown in Figure 7A. The B-cell phenotype was clustered into seven distinct subtypes. We next focused on an IgG4-associated subpopulation in the B-cell cluster. As shown in Fig. 7B, cells in cluster_2 were characterized by high expression levels of CD27, CD38, CXCR5, BCL6, MME, CD40, AICDA, IGHG4, IL4R, IL10RA, and IL21RA. These findings suggested that the cluster_2 population comprised activated B cells undergoing class-switching. Furthermore, cells in cluster_2 exhibited low expression levels of IGHD and IGHM. In contrast, cells in cluster_4 highly expressed IGHG4, CD38, CD27, and CD138; thus, cluster_4 population presumably comprised plasma cells. We next analyzed infiltrating AICDA⁺CD19⁺ B cells in affected lesions from three IgG4-RD patients using scRNA-seq. Most infiltrating AICDA⁺CD19⁺ B cells expressed IL4R, IL10RA, and IL21R, but not IL13R (Fig. 7C). Most AICDA⁺CD19⁺IL4R⁺IL10RA⁺IL21R⁺ B cells expressed a comparatively high level of IGHG4 (Fig. 7D).

Figure 7. Single-cell B-cell transcriptome analysis of affected lesions from patients with IgG4-RD.

(A) Seven clusters across 1,245 CD19⁺ cells, identified via tSNE visualization. Schematic representation of single-cell transcriptome experiments from submandibular gland sample collection to clustering of CD19⁺ B cells in a patient with IgG4-RD (Pt3).

(B) Top: selected gene expression patterns in CD19⁺ B cells among the seven clusters in a patient with IgG4-RD (Pt3). Bottom: Violin plot showing gene expression patterns according to tSNE clusters (Pt3).

(C) Single-cell RNA-seq profiling revealed the expression patterns of interleukin receptor genes in AICDA⁺CD19⁺ B cells in affected tissues from patients with IgG4-RD (n=4).

(D) Single-cell RNA-seq profiling revealed the percentages of IGHG1, IGHG2, IGHG3, IGHG4, IGHE, IGHD, and IGHM in AICDA⁺CD19⁺IL4R⁺IL10RA⁺IL21R⁺ B cells in affected tissues from patients with IgG4-RD (n=4).

Allergic fibrosis in KD is linked to activated type 2 immune cells

Eosinophilic microabscesses and eosinophil-infiltrated neural fibers are typical features in KD patients. Eosinophils were expanded in affected tissues from KD patients, compared with affected tissues from IgG4-RD patients (Supplemental E Fig. 5A). Galectin-10 is released by activated eosinophils during type 2 immune reactions ^{26, 27}. Galectin-10-positive cells (i.e., activated eosinophils) were expanded in affected lesions from KD patients compared with that in IgG4-RD patients (Supplemental E Fig. 5B). Osteopontin (OPN) is a glycoprotein that exhibits fibrogenic and angiogenic properties; it has also been implicated in allergic disease. Thus, we explored the major cellular source of OPN in affected lesions from patients with KD and patients with IgG4-RD. OPN⁺Siglec-8⁺ eosinophils were sparse in patients with IgG4-RD (Supplemental E Fig. 5C), but In contrast, they were abundant in patients with KD (Supplemental E Fig. 5C); most OPN-stained cells also expressed Siglec-8 in patients with KD. Furthermore, approximately 60% of Siglec-8⁺ eosinophils expressed OPN in a patient with KD, while fewer than 20% of Siglec-8⁺ eosinophils expressed OPN in a patient with IgG4-RD (Supplemental E Fig. 5C). Finally, amphiregulin (AREG)⁺CD4⁺ T cells were detected and abundant in affected tissue from patients with KD (Supplemental E Fig. 5D).

Type 2 cytokines are produced by activated IgE coated c-kit⁺ mast cells in KD patients

A previous report showed that IL-4-producing cells were abundant in affected tissues from KD patients. To our knowledge, no previous study has used a quantitative approach to describe IL-4-expressing cells in affected lesions from KD patients. Notably, IL-4 is expressed by Th2 and Tfh cells. Mast cells also play central roles in IgE-mediated allergic diseases by releasing IL-4, IL-5, and IL-13. We thus speculated that a subset of IL-4-producing cells might be mast cells in KD patients. We used immunofluorescence to compare the numbers of mast cells (detected by c-kit staining) among KD patients and IgG4-RD. We found that IgE⁺c-kit⁺ activated mast cells expressed IL-4 in affected tissues from a patient with KD (Supplemental E Fig. 5E). Furthermore, IL-4⁺c-kit⁺IgE⁺ mast cells were abundant in KD patients. Some IgG4-RD patients exhibited infiltrating IL-4⁺c-kit⁺IgE⁺ mast cells in their affected lesions. We subsequently analyzed IL-5 and IL-13 production by activated mast cells. Importantly, IL-5-expressing c-kit⁺IgE⁺ mast cells and IL-13-expressing mast cells were abundant in KD patients but sparse in IgG4-RD patients.

Inflammatory-related cytokines and type 2 immunity-related cytokines in IgG4-RD

We applied scRNA-seq to profile tissue infiltrating all cells in fresh affected salivary glands from IgG4-RD patients. Some non-allergic IgG4-RD patients exhibit elevated serum IgE ⁹. We thus obtained affected salivary glands from three non-allergic IgG4-RD patients with high serum IgE. We visualized selected marker genes for inflammatory- and type 2 immunity-related cytokines using t-SNE map (Supplemental E Fig. 6A). Most inflammatory-related genes (GZMA, GZMK, IFNG, PRF1, TNF, and TGFB1) are expressed in T cells from IgG4-RD tissue (Fig. 4D). Gene expression of IL-6, pro-inflammatory cytokine, was sparse in IgG4-RD tissue (supplemental E Fig. 6A). In contrast, type 2 immunity-related genes (IL5, IL13, IL25, IL33, TSLP, IGHE, and IL1RL1) were sparse in IgG4-RD tissue (Supplemental E fig. 6B). Some CD14⁺ monocytes express TSLP and IL33, but at low levels.

DISCUSSION

In T cell-dependent immune responses, Tfh cells are the primary helper T cells responsible for directing the affinity, longevity, and isotype of antibodies produced by B cells. Specific cytokine-producing Tfh cells contribute to distinct isotype switching events ^{1, 2}. Indeed, distinct CD4⁺CXCR5⁺ Tfh cells are abundant in affected lesions from KD patients and IgG4-RD patients, which are diseases in which there is prominent switching of activated B cells to two different Ig isotypes. Here we described abundant infiltration of distinct Tfh cells in these two disorders. Our novel findings were as follows: (i) infiltrating IL-13-expressing Tfh cells were abundant in affected lesions of patients with allergic disorders (e.g., KD patients); (ii) infiltrating IL-13-expressing Tfh cells were sparse in IgG4-RD; (iii) infiltrating IL-10-expressing Tfh cells, but not Tregs or Tfr cells, were abundant in IgG4-RD patients and expressed IL-21 and LAG3; (iv) the frequency of AID⁺CD19⁺B cells expressing receptors for IL4, IL10, and IL21 correlated with IgG4 expression in IgG4-RD; and (v) distinct infiltrating cell types characterize two distinct types of fibrotic disorders: allergic fibrosis (driven by type 2 immune cells) and inflammatory fibrosis (driven by cytotoxic T cells).

Class switching to IgG4 is poorly understood. CD4⁺ T cells can stimulate IgM-positive B cells to switch to IgG4 and IgE in the presence of added IL-4 ²⁸. In vitro findings have suggested that IL-10 indirectly contributes to IgG4 class switching by facilitating IL-4-mediated switching to IgG4 rather than IgE ¹⁰. Previous work distinguished Tfh cells by identifying a distinctive IL4 enhancer locus bound by BATF in Tfh cells that is distinct from the Th2 DNA regulatory element for IL-4, IL-5, and IL-13 bound by GATA3 ²⁹. Subsequent research indicated that IL-4⁺ Tfh cells instruct plasma cells to switch from IgM to IgE via BATF-driven IL-4, thus producing low-affinity IgE antibodies ¹⁴. To our knowledge, Tfh cells expressing other cytokines have not been assessed in IgG4-RD patients at the single-cell level. In this study, we found that IL-10- and IL-21-expressing Tfh cells were abundant in IgG4-RD patients compared with levels in patients with allergic disorders. Using scRNA-seq analysis of infiltrating lymphocytes, we confirmed that the expanded IL-10-expressing CD4⁺CXCR5⁺ Tfh cells in affected lesions from IgG4-RD patients co-expressed IL-21, PDCD1, and ICOS, but not Foxp3; these cells presumably differed from Treg or Tfr phenotype cells. In a study of IL-10 and IL-21 double-reporter mice, Gang et al. revealed that IL-10⁺IL-21⁺CXCR5⁺PD1⁺ Tfh cells were distinct from Foxp3-expressing Tfr cells ³⁰. In contrast, the CD4⁺CXCR5⁺Foxp3⁺ Tfr cell population in lymph nodes from IgG4-RD patients was also abundant compared with that in healthy tonsils (data not shown). thus, these Tfr cells in IgG4-RD patients require further exploration. Notably, we limited our tissue analysis to IgG4 class-switched B cells, but our results suggest that the amounts of IL-4, IL-10, and IL-21 produced by disease-specific Tfh cells may contribute to specific IgG4 class switching in IgG4-RD patients (Fig. 8).

Figure 8. Model of IgG4 class switching by IL10+Tfh cells in IgG4-related disease, which contrasts with IgE class switching by IL13+Tfh cells observed in Kimura’s disease.

Schematic overview of the proposed disease class switching mechanisms in patients with KD and patients with IgG4-RD. Expansion of infiltrating IL-10⁺ Tfh cells, but not Tfh13 cells, might contribute to IgG4 isotype switching in SLOs and TLOs of patients with IgG4-RD. In contrast, expansion of infiltrating Tfh13 cells might contribute to high-affinity IgE secretion by B cells in Kimura’s disease.

We also found abundant cytotoxic Tfh cells and Tfr cells in IgG4-RD patients. However, additional research for these cytotoxic Tfh cells and Tfr cells is required to further elucidate the pathogenesis of IgG4-RD.

The exact source of the elevated IgG4 and IgE in the blood/tissue in IgG4-RD patients is unknown. Tissue IgE-positive memory B cells and plasma cells may emerge directly from germinal centers or through indirect class switching from other intermediate antibody isotype such as IgG1 to IgG4 ^{31, 32}. The majority of IgE⁺ cells derive from somatically hypermutated IgG1-expressing cells as demonstrated from analysis of Ig heavy regions in blood of allergy patients and indirect isotype switching from IgG4 to IgE contributes to the IgE pool ³³. In this study, we demonstrated that a rare IL-13-expressing Tfh cell population, which co-expressed GATA3 and Bcl6, was dominant in affected tissues from KD patients; however, this population was sparse in IgG4-RD patients, despite the presence of high serum IgE levels. Human Tfh cells expressing GATA3, IL-13, and IL-4 have been previously identified ¹⁴. Furthermore, Tfh13 cells in mice and humans have an unusual cytokine profile (IL-4^hiIL-13^hi) and co-express the BCL6 and GATA3 transcription factors ¹⁴. These Tfh13 cells are required for production of high-affinity (but not low-affinity) IgE and subsequent allergen-induced anaphylaxis ¹⁴. We speculated that the expansion of these IL-13-expressing Tfh cells might represent an important disease-related Tfh subset, which contributes to specific class-switching and affinity-maturation events in KD patients (Fig. 8).

Fibrosis is the end result of chronic inflammatory reactions such as allergic responses, infections, autoimmune reactions, tissue injury, and radiation. Fibrotic diseases likely have many different etiologies, and they may not all be driven by CD4⁺ T cells. Our recent findings implicated CD4⁺CTLs in the induction of apoptotic cell death and subsequently fibrosis in IgG4-RD ^{13, 15}, systemic sclerosis ³⁴, COVID-19 ³, and fibrosing mediastinitis, a disease linked to Histoplasma capsulatum infection ³⁵.

In contrast, type 2 immune cells and their cytokines (IL-4, IL-5, IL9, and IL-13) represent a central population in the pathogenesis of allergic inflammation and fibrosis ¹⁶. Complex inflammatory reactions involve eosinophils, basophils, mast cells, type 2 innate lymphoid cells (ILC2), Th2 cells, and subclasses of IgE antibodies; these components play important roles in the pathogenesis of many allergic and fibrotic disorders. IL-5 is a key cytokine involved in eosinophil development and activation. Lesional inflammation causes recruitment of eosinophils with inflammatory features ¹⁶. Recently, Morimoto et al. reported that interactions between pathogenic memory Th2 cells and OPN-producing eosinophils may be a potential target for the treatment of fibrosis induced by chronic allergic disorders ³⁶. Some reports suggested that IgG4-RD appears to involve some of the same pathogenic mechanisms observed in allergic disease, such as Th2 and Treg activation, IgG4 and IgE hypersecretion and blood/tissue eosinophilia ³⁷. Multiple cell types associated with a type 2 immune response are found at sites of affected lesions in KD patients compared with IgG4-RD. Amphiregulin has been shown to reprogram the transcriptome of eosinophils toward an inflammatory state that induces secretion of OPN, an extracellular matrix protein associated with fibrotic disorders ³⁶. Charcot–Leyden crystal protein (also known as galectin-10) is required for eosinophile granulogenesis ²⁷. Galectin-10 is a hallmark of eosinophil death and can persist in tissues for months. Recent genome-wide genetic and epigenetic association studies found a strong association of total IgE levels with hypomethylation at the LGALS10 gene locus, suggesting that the eosinophil galectin-10 axis may also be a trigger for IgE synthesis in human type 2 immune disease ^{26, 38}. In a recent report, increased blood ILC2 were linked to blood eosinophilia, elevated IgE, and itching in KD ³⁹. In our staining data, we also noticed that a large number of CD4⁻GATA3⁺ cells infiltrated KD tissues (Supplemental E Fig. 2A), suggesting that these cells may be ILC2s or related cells. These GATA3-positive non-T cell in lesions will be more thoroughly investigated in future studies. We found that type 2 immune cells were sparse in IgG4-RD patients compared with that in KD. Cross-linking of high-affinity IgE on the surface of mast cells leads to the release of chemical mediators that precipitate anaphylaxis ⁴⁰. Studies from patients with food allergies and murine models of allergic disease indicate that high-affinity (but not low-affinity) IgE induces mast cell degranulation and anaphylaxis ⁴⁰. Mast cells capture monomeric IgE on their surface using the high-affinity Fc receptor for IgE (FcεRI), and antigen-mediated crosslinking of FcεRI-bound IgE leads to mast cell activation and the release of allergic mediators. High-affinity (but not low-affinity) IgE might be detected in germinal centers from KD patients. Activated mast cells produce pro-fibrotic factors including TGF-β1, IL-4, IL-13, tryptase, chymase, and chemokines that promote fibroblast activation in human fibrotic disease ⁴¹. Most IL-13 is produced by Th2, Tfh13, and activated mast cells in affected lesions from KD patients. IL-13 can also directly promote fibrosis by stimulating proliferation or collagen production by fibroblasts, as well as fibroblast differentiation into myofibroblasts ⁴². In KD patients, activated eosinophils and mast cells contribute to a chronic inflammatory response ¹¹. Our findings suggest two types of fibrosis, allergic fibrosis (driven by type 2 cells) and inflammatory fibrosis (driven by cytotoxic T cells), in KD and IgG4-RD.

Key messages:

• A novel IL-10-expressing Tfh-like cell subset, but not Tregs or Tfr cells, was detected in affected tissues from patients with IgG4-RD, and these IL-10-expressing Tfh-like cells expressed LAG3 and PRDM1.

• The frequency of AICDA⁺CD19⁺B cells expressing receptors for IL10 and IL21 correlated with IgG4 expression in IgG4-RD.

• An IL-13-expressing Tfh-like cell subset was detected in tissue lesions from patients with KD but not patients with IgG4-RD.