Phenotype, function and clinical significance of innate lymphoid cells in immunoglobulin G4–related disease

Abstract

Objectives

The innate immune system participates in immunoglobulin G4–related disease (IgG4-RD). While the role of innate lymphoid cells (ILCs) in IgG4-RD remains to be elucidated, we aimed to evaluate the phenotype, function and clinical significance of ILCs in IgG4-RD patients.

Methods

Sixty-seven untreated IgG4-RD patients and 44 age- and sex-matched healthy controls (HCs) were enrolled. Circulating and tissue infiltration of ILCs were detected by flow cytometry. Serum suppression of tumorigenicity 2 (sST2) was detected by ELISA and membrane-bound ST2 (ST2L) was detected by flow cytometry. Tissue infiltration of IL-33 was measured by immunohistochemistry staining. Real-time quantitative PCR was performed to analyse the expression pattern of ILC2-associated genes between HCs and IgG4-RD patients. In addition, correlation analysis was performed in order to evaluate the clinical significance of ILCs in IgG4-RD.

Results

The frequency of circulating pan ILCs in IgG4-RD patients was lower than in HCs. ILC2s were higher in IgG4-RD compared with HCs, whereas ILC1s were lower in IgG4-RD. sST2 and ST2L were higher in IgG4-RD than in HCs. Infiltration of ILC1s in the submandibular glands of IgG4-RD patients was more prominent than ILC2s. Intracellular secretion of IL-9 was increased in ILC2s of IgG4-RD patients than in HCs. Circulating ILC2s correlated positively with Treg cells and the surface expression of CD154, PD-1 and CXCR5 in ILC2s correlated positively with CD19⁺ B cells, serum IgG4 levels and serum IgE, respectively.

Conclusion

ILCs and their subsets were significantly altered in IgG4-RD. We demonstrated the dysfunction of ILC2s in IgG4-RD by phenotype, correlation analysis and function investigation, revealing ILC2s participated in the pathogenesis of IgG4-RD.

We propose a pathogenic model that ILC2s induce fibrosis and abnormal adaptive immune response in IgG4-RD. After simulation through the IL-33–ST2 axis, hyperactive ILC2s secrete more IL-9, which may participate in fibrosis directly and/or promote fibrosis by activating Treg and TGF-β secretion. Activation of ILC2s may require Treg cells accumulating via an ICOSL–ICOS costimulatory signal. Furthermore, PD-1 and CXCR5 in ILC2s may play a role in IgE and IgG4 production, respectively, leading to adaptive immune response in IgG4-RD. ILC, innate lymphoid cells; Treg, regulatory T cells; IL, interleukin; PD-1, programmed cell death protein 1; CXCR5, CXC chemokine receptor type 5; ICOS, inducible T cell costimulator; ICOSL, inducible T cell costimulator ligand; TGF-β, transforming growth factor β; IgE, immunoglobulin E; IgG4, immunoglobulin G4.

Open in new tabDownload slide

Introduction

Immunoglobulin G4–related disease (IgG4-RD) is a chronic multiorgan fibro-inflammatory disease with tumefactive swelling of affected organs and elevation of serum IgG4 levels, abundant infiltration of lymphocytes and IgG4-positive plasma cells in affected organs [1–3]. The pathogenesis of IgG4-RD is unknown, although evidence has indicated that innate immunity plays an important role [4]. M2 macrophages may promote the activation of Th2 immune responses via IL-33 secretion in IgG4-RD [5]. In addition, γδ T cells, eosinophils, basophils and the complement system participate in IgG4-RD pathogenesis [6–8]. Until now there has been no report on the role of innate lymphoid cells (ILCs) in IgG4-RD.

ILCs are a family of lymphocytes involved in the initiation, regulation and resolution phases of inflammatory processes [9]. Three distinct groups of ILCs (ILC1s, ILC2s, ILC3s) have been described that functionally correspond to innate counterparts of CD4⁺ Th cells [10]. ILC1s express the transcription factor (TF) T-BET and produce the Th1-associated cytokines IFN-γ and TNF. ILC2s secrete Th2-associated cytokines via a GATA-3 and ROR-α-dependent pathway [11]. Human ILC2s were first characterized in the foetal intestine and were defined as Lin⁻CD127⁺CD161⁺ cells expressing the chemoattractant receptor CRTH2, which has been shown to mark Th2 cells [12]. ILC2s have been reported to play an important role in type 2 immunity and in allergic inflammatory disease [13]. ILC3s use ROR-γt to drive production of the Th17-associated cytokines IL-17 and IL-22. ILC disturbance contributes to immune disturbance in multiple immune diseases, including SLE, RA and SpA, among others [14–18].

Activation of Th2 and ILC2 pathways is at the core of type 2 inflammation. The main type 2 cytokines of ILC2s are IL‐5, IL‐9 and IL‐13 [19]. IgG4-RD is associated with aberrant production of Th2 cytokines [20], thus the investigation of ILCs remains of great significance in understanding the pathogenesis of IgG4-RD. In this study we aim to evaluate the phenotype, function and clinical significance of ILCs in IgG4-RD.

Materials and methods

Patients enrolment

Sixty-seven newly diagnosed patients with IgG4-RD fulfilling the 2011 comprehensive IgG4-RD diagnostic criteria [21] and 44 age- and sex-matched healthy controls (HCs) were enrolled in this study. Patients with other autoimmune diseases, cancer or lymphoma were excluded. Patients’ clinical data, laboratory parameters and organs involved are shown in Supplementary Table S1, available at Rheumatology online. The study was approved by the ethics committee of Peking Union Medical College Hospital. Written informed consent was obtained from all patients and HCs. Sixteen fresh submandibular glands (SMGs) of biopsy-proven IgG4-RD were obtained for detection of ILC infiltration in affected organs of IgG4-RD patients. Five paraffin-embedded biopsy-proven SMGs of IgG4-RD patients and three labial gland tissue samples from patients with primary SS (pSS) were obtained for immunohistochemistry (IHC) examination. Mononuclear cell preparation in SMGs of patients with IgG4-RD and pathohistological fibrosis evaluation are described in the supplementary material, available at Rheumatology online.

Flow cytometry analysis of ILC subsets

Separated by Ficoll gradient centrifugation, cryopreserved peripheral blood mononuclear cells (PBMCs) from IgG4-RD patients and HCs stored in −80°C were used for ILCs and surface marker staining. Fresh isolated mononuclear cells from SMGs of IgG4-RD patients were obtained and stained for flow cytometry. In order to remove dead cells, a Dead Cell Removal Kit (#130-090-101, Miltenyi Biotec, Bergisch Gladbach, Germany) was added in PBMCs before flow cytometry staining according to the manufacture’s procedures. The representative flow cytometry plots before and after dead cell removal are shown in Supplementary Fig. S1, available at Rheumatology online. Live cells were then used for flow cytometry analysis.

ILCs were stained with allophycocyanin (APC)-conjugated anti-CD127, FITC-conjugated anti-lineage, PE-cyanine 7 (Cy7)-conjugated anti-CRTH2, APC-Cy7 conjugated anti-c-Kit and isotype-matched controls. PE-conjugated anti-IL-4, anti-IL-5, anti-IL-9 and anti-IL-13 were used for intracellular staining of ILC2s. PE-conjugated IFN-γ and PerCP-Cy5.5-conjugated TNF-α were used for intracellular staining in ILC1s. PE-conjugated anti-human ST2 (IL-33R), anti-PD-1, anti-CCR-10, anti-CD200R and PerCP-Cy5.5-conjugated anti-CD154, anti-CXCR5 were used for cell surface marker staining of ILC subsets. APC-conjugated anti-CD25, FITC-conjugated anti-CD4 and PE-conjugated anti-Foxp3 were used for Treg cells. Lineage and ST-2 were from eBioscience (San Diego, CA, USA), other antibodies were from BioLegend (San Diego, CA, USA). After incubation for 30 min at 4°C, PBMCs were washed and resuspended in PBS.

For intracellular staining, perm and fixation buffers (BD Biosciences, Franklin Lakes, NJ, USA) were used and experiments were performed according to manufactural instructions. All experiments were measured using the FACSAria II system (BD Biosciences). Data were analysed by FlowJo version 10 software (FlowJo, Ashland, OR, USA).

Pan ILCs were defined as Lin⁻CD127⁺ cells. ILC1s, ILC2s and ILC3s were defined as Lin⁻CD127⁺CRTH2-c-Kit⁻, Lin⁻CD127⁺CRTH2⁺c-Kit^±, Lin⁻CD127⁺CRTH2-c-Kit⁺ cells, respectively.

IHC staining of IL-33

For IHC staining, slides were incubated with primary rabbit anti-human IL-33 (Abcam, Cambridge, UK) following the manufacturer’s instructions. IHC images were acquired and processed using a microscope (Olympus, Tokyo, Japan) at ×400 magnification. IHC results were recorded by light microscopy and analysed by Image-Pro Plus 6.0 software (Media Cybernetics, Crofton, MA, USA). The cell nucleus or membrane was stained yellow or brown, suggesting a positive signal. The protein expressions were quantified by integrated optical density per high-powered field. Data were presented as the average result in three randomly selected fields.

ELISA detection of serum ST2 (sST2)

Serum was collected and stored at −80°C until use. The concentrations of sST2 were measured by ELISA (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions.

Real-time quantitative PCR

Total RNA was extracted from PBMCs using Trizol solution (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. Details are shown in Supplementary Data S1, available at Rheumatology online.

In vitro cell culture

Fresh isolated PBMCs from IgG4-RD patients and HCs were resuspended in Roswell Park Memorial Institute 1640 medium supplemented with 10% foetal calf serum and antibiotics (penicillin 100 IU/ml, streptomycin 100 μg/ml), sodium pyruvate (100 mM; Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences Basic Medical Cell Center, Beijing, China) in six-well flat-bottomed plates (Corning-Costar, Corning, NY, USA) in a humidified atmosphere of 5% carbon dioxide CO₂ at 37°C, 5 × 10⁶ in each well. For intracellular staining of TNF-α and IFN-γ in ILC1s, PMA, ionomycin and GolgiStop (BD Bioscience) were added to each well. After 6 h of culture, PBMCs were collected for intracellular staining. For intracellular staining of IL-4, IL-5, IL-9 and IL-13 in ILC2s, 50 ng/ml anti-human recombinant IL-33 (PeproTech, Cranbury, NJ, USA) with and without 10 ng/ml anti-human IL-2 (BioLegend) was added at the beginning. After 2 h of culture, PMA, ionomycin and GolgiStop were added to each well. After 6 h of culture, PBMCs were collected for intracellular staining of Th2 cytokines in ILC2s.

Statistical analysis

Statistical analysis was performed using SPSS Statistics version 24.0 software (IBM, Armonk, NY, USA). P-values <0.05 were considered significant. Details are shown in Supplementary Data S1, available at Rheumatology online.

Results

Aberrant expression of ILC subsets in IgG4-RD

Representative flow cytometry images in peripheral blood are shown in Fig. 1A. The representative gating strategies of ILCs and their subsets in SMGs are shown in Supplementary Fig. S2, available at Rheumatology online. The percentage of circulating pan ILCs in lymphocytes was decreased in IgG4-RD [0.21% (s.d. 0.20]) compared with HCs [0.28% (s.d. 0.19)] (P = 0.0007; Fig. 1B). The proportion of ILC subsets was significantly altered in IgG4-RD. Compared with HCs, the frequency of ILC1s was significantly decreased [50.40% (s.d. 24.34) vs 60.42 (19.57), P = 0.0048] while the frequency of ILC2s was increased [33.36% (s.d. 19.29) vs 23.83 (16.30), P = 0.0163] in IgG4-RD (Fig. 1C and D). ILC3s were comparable between HCs and IgG4-RD patients (Fig. 1E). Moreover, the ILC2:ILC1 ratio was increased in IgG4-RD [1.08 (s.d. 1.11)] compared with HCs [0.59 (s.d. 0.75)] (P = 0.0068; Fig. 1F). As shown in Supplementary Fig. S3 (available at Rheumatology online), the percentage of ILC1s, ILC2s and ILC3s in lymphocytes of IgG4-RD patients and HCs were 0.107% (s.d. 0.018) vs 0.192 (0.025), 0.067% (s.d. 0.008) vs 0.049 (0.004) and 0.027% (s.d. 0.003) vs 0.038 (0.003) (P = 0.0051, P = 0.0389 and P = 0.0224), respectively.

After treatment, pan ILCs in IgG4-RD increased from 0.21% (s.d. 0.20) to 0.25% (s.d. 0.14) (P = 0.0427). The percentages of ILC subsets were comparable between pretreatment and post-treatment.

In order to investigate whether ILC subsets play a role in IgG4-RD tissue damage, the expression of ILC subsets in lymphocytes from 16 IgG4-RD SMGs was detected. Surprisingly, ILC1s [56.55% (s.d. 4.54)] comprised the main component among the infiltrated ILCs, the ratio of tissue-infiltrated ILC2s [12.63% (s.d. 1.86)] in lymphocytes was lower than in peripheral blood (P < 0.0001; Fig. 1D) and ILC3s [27.86% (s.d. 4.74)] were much more prominent in affected organs compared with circulating ILC3s (P = 0.0099; Fig. 1E).

Intracellular cytokines of ILC1s and ILC2s in IgG4-RD

We found that percentages of ILC1s and ILC2s were aberrantly expressed in IgG4-RD. In order to detect whether ILC1s and ILC2s were functioning normally in IgG4-RD, we tested intracellular cytokines of circulating ILC1s and ILC2s in IgG4-RD. Intracellular staining of TNF-α and IFN-γ was measured in ILC1s and the gating strategy is shown in Fig. 2A. Results demonstrated that the frequencies of IFN-γ- and TNF-α-producing ILCs and ILC1s were comparable to those of HCs (Fig. 2B).

We subsequently focussed on the function of ILC2s. Since ILC2s are critical innate sources of type 2 cytokines in allergic inflammation, and we found ILC2s were significantly increased in IgG4-RD patients, we thought it would be interesting to know the expression of Th2-type cytokines in ILC2s. It was reported that ILC2s are activated and expanded by IL-2, IL-33 and IL-25 in vitro [22], thus intracellular staining of Th2 cytokines was measured in ILC2s after stimulation of IL-33 with and without IL-2. Representative flow cytometry pictures of the gating strategy are shown in Fig. 2C and D. After stimulation with IL-33 and IL-2, intracellular IL-9 in ILC2s from IgG4-RD patients [16.87% (s.d. 4.55)] was higher than in HCs [8.09% (s.d. 4.51)] and higher than in IgG4-RD stimulated with IL-33 alone [10.04% (s.d. 4.63)] (P = 0.0073 and P = 0.0276, respectively; Fig. 2E).

A similar trend was seen in mean fluorescence intensity (MFI). The MFI of intracellular IL-9 in ILC2s from IgG4-RD patients [2821 (s.d. 951)] was higher than in HCs [590 (s.d. 98)] (P = 0.026; Supplementary Fig. S4, available at Rheumatology online). Intracellular staining of IL-4, IL-5 and IL-13 in ILC2s was comparable between HCs and IgG4-RD patients stimulated with IL-33 with and without IL-2.

Since ILC2s were not the only sources of IL-9, we then detected intracellular IL-9 in CD4⁺ T cells. The expression of intracellular IL-9 in CD4⁺ T cells was comparable in IgG4-RD patients [1.66% (Q1 0.26–Q3 5.7)] and HCs [1.28% (Q1 0.52–Q3 5.03)] (P = 0.815; Supplementary Fig. S5, available at Rheumatology online).

Expression of costimulatory factors and chemokine receptors on ILC2s in IgG4-RD patients

To further evaluate the potential function of ILC2 and explore possible reasons for the disparity of ILC2s in peripheral blood and affected organs, surface markers as costimulatory factors and chemokine receptors, including CD154 [23], PD-1 [11], CCR10 [24], CD200R [25], CXCR5 [26] and ICOS/ICOSL [12], were detected in pan ILCs and ILC2s of IgG4-RD peripheral blood (Fig. 3 and Supplementary Fig. S6, available at Rheumatology online). Results revealed an elevated percentage of CD154 in pan ILCs of IgG4-RD [6.14% (s.d. 7.70)] compared with HCs [1.98% (s.d. 1.51)] (P = 0.0266; Fig. 3A). In addition, the percentage of PD-1 and the MFI of CCR10 were also increased in pan ILCs of IgG4-RD patients compared with HCs (P = 0.0059 and P = 0.02, respectively; Fig. 3B and C).

In terms of ILC2s, the frequencies of CD154, PD-1, ICOS and CCR10 and the MFI of PD-1 were increased in IgG4-RD [21.18% (s.d. 18.58), 20.2 (18.25), 33.46 (13.95), 35.95 (21.33) and 672 (36)] compared with HCs [11.18% (s.d. 9.25), 10.69 (9.22), 24.52 (10.64), 22.45 (15.66) and 554 (40)] (P = 0.0441, P = 0.0489, P = 0.0357, P = 0.031 and P = 0.0331, respectively; Fig. 3D–G, Supplementary Fig. 6B, available at Rheumatology online), whereas CXCR5 MFI was decreased in IgG4-RD [1087 (s.d. 518)] compared with HCs [1501 (s.d. 604)] (P = 0.0251; Fig. 3H). Other surface markers were comparable between HCs and IgG4-RD patients (Supplementary Fig. S6, available at Rheumatology online).

IL-33, IL-9 and ST2 expression in IgG4-RD

IL-33, a member of the IL-1 cytokine family, contributes to multiple pathogenesis in various inflammatory diseases [27]. ST2 is an IL-1 receptor family member with membrane-bound (ST2L) and soluble (sST2) isoforms and is one of receptors of IL-33 [28]. Activating the IL-33–ST2 axis triggers pleiotropic immune functions in multiple ST2-expressing immune cells, such as ILC2s. As we described above, ILC2s in IgG4-RD secreted more IL-9 after stimulation with IL-33 and IL-2. To further understand whether ILC2 hyperactivation in IgG4-RD is related to the IL-33–ST2 axis, IL-33 and ST2 expressions in IgG4-RD were detected.

As we expected, IL-33 was detected in SMGs of IgG4-RD and the infiltration of IL-33 was comparable between patients with IgG4-RD [221.4 (s.d. 3.9)] and pSS [223.1 (s.d. 0.3)] (P = 0.486; Fig. 4A–C). Consistently the mRNA level of IL-33 was increased in PBMCs of IgG4-RD [2.03 (s.d. 1.36)] compared with HCs [1.01 (s.d. 1.10)] (P = 0.043; Fig. 4D).

sST2 has been recognized as a biomarker indicative of IL-33–ST2 pathway activity [29]. Thus we detected sST2 in IgG4-RD. The level of sST2 in IgG4-RD [29.81 ng/ml (s.d. 10.99)] was elevated compared with that of HCs [20.07 ng/ml (s.d. 8.83)] (P = 0.0009; Fig. 4F). In addition, the mRNA level of IL-9 was also elevated in PBMCs of IgG4-RD patients [8.62 (s.d. 12.27)] compared with HCs [1.13 (s.d. 0.83)] (P = 0.0152; Fig. 4E).

The expression of ST2L in ILC2s of IgG4-RD patients and HCs are shown in Fig. 4G. As we described above, sST2 is elevated in IgG4-RD compared with HCs and ST2L in ILC2s of IgG4-RD patients was also evaluated in our study [6.18% (s.d. 4.63) vs 1.61 (1.12), P = 0.0068] (Fig. 4H).

Correlations of pan ILCs/ILC2s and surface markers of ILCs with laboratory parameters and other immune cells

In order to delineate the clinical significance of pan ILCs/ILC2s and interactions of pan ILCs/ILC2s with other immune cells, correlation analysis was performed. Pan ILCs correlated negatively with serum IgG4, IgG and IgG4-RD responder index (RI) (P = 0.0093, P = 0.029 and P = 0.0375, respectively; Fig. 5A–C). The percentage of PD-1 on ILCs was positively corrected with serum IgG4 level (P < 0.0001; Fig. 5D). In addition, the percentage of CCR10 and CD154 in ILCs correlated positively with the percentage of CD19⁺ B cells (P = 0.0445 and P = 0.0093, respectively; Fig. 5E and F).

Interestingly, there was a positive correlation between ILC2s and Treg cells (P = 0.0211; Fig. 5H). Moreover, the percentage of CD154 and CXCR5 in ILC2 correlated positively with CD19⁺ B cells and serum IgE, respectively [P = 0.0084 (Fig. 5I) and P = 0.0002 (Fig. 5K)]. Although there was no correlation between ILC2 and serum IgG4 (Fig. 5G), the percentage of PD-1 in ILC2s correlated positively with serum IgG4 (P = 0.0236) (Fig. 5J).

Correlation of tissue ILC2s with tissue fibrosis

In order to further elucidate the clinical significance between tissue ILC2s and tissue fibrosis, scores of tissue fibrosis were assessed. As shown in Fig. 6A–C, different degrees of fibrosis were found in our studies. Typical IgG, IgG4 and fibrosis staining are shown in Fig. 6D–F, respectively. Sixteen patients were classified into two groups according to the mean percentage of all IgG4-RD patients’ ILC2s (12.63%) in the total ILCs in tissues. Patients with an ILC2 percentage ≥12.63% were defined as group B, while all others were defined as group A. There was an increasing trend of fibrosis score in group B compared with group A (Fig. 6G). Regarding the percentage of fibrosis area, a similar trend was found in group B compared with group A (Fig. 6H). However, there was no statistical significance.

Discussion

ILC subsets have been identified to participate in the pathogenesis of multiple allergic and autoimmune diseases, including asthma, RA, pSS and SSc. However, the role of ILCs and their subsets in the pathogenesis of IgG4-RD needs to be delineated. In this study we found reduced circulating pan ILCs and a significant alteration of ILC subsets with prominent elevation of ILC2s in patients with IgG4-RD, and ILC2s were hyperactivated by the IL-33–ST2 axis, resulting in IL-9 secretion in IgG4-RD. Moreover, pan ILCs correlated with clinical parameters, which might be a biomarker for disease activity. Interestingly, ILC2s correlated positively with Tregs, the surface expression of CD154, and PD-1 and CXCR5 in ILC2s correlated positively with CD19⁺ B cells, serum IgG4 level and serum IgE, respectively, indicating that ILC2s are involved in immune regulation and adaptive immune response in IgG4-RD. Hence we propose a pathogenic model that ILC2s induce fibrosis and cause adaptive immune response in IgG4-RD, which we have detailed in the graphical abstract. To our knowledge, this study is the first to evaluate the role of ILC subsets in IgG4-RD.

Although ILCs belong to the innate immune system, accumulating evidence indicates that they also have adaptive immune features [30]. Besides this, ILCs participate in the onset and maintenance of inflammation [31]. Alteration of ILCs has been reported in multiple immune diseases [32]. In our study, the decreased frequency of ILCs in IgG4-RD and negative correlation of ILCs with serum IgG, IgG4 and IgG4-RD RI suggested that ILCs may be associated with disruption of immune homeostasis and circulating ILCs in IgG4-RD could be an indicator of disease activity. PD-1 is recognized as an activation marker of ILCs, as PD-1⁺ ILCs correlated positively with serum IgG4, indicating that PD-1⁺ ILCs may play more important roles than PD-1⁻ ILCs in IgG4-RD. Moreover, some important surface markers, including costimulatory factors and chemokine receptors, expressed on ILCs were correlated with B cells, indicating ILCs played a role in the disturbance of adaptive immunity in IgG4-RD.

ILC1s play an essential role in viral immunosurveillance at sites of initial infection [33]. ILC1s were elevated in inflamed intestine of Crohn’s disease, which contributed to the pathogenesis of gut mucosal inflammation [34]. Moreover, ILC1 frequencies were increased in pSS patients and correlated with disease activity [35]. In our study, the percentage of circulating ILC1s was decreased in IgG4-RD, while the expression of PD-1 and TNF-α on ILC1s was comparable to that of HCs, indicating circulating ILC1s may not be the main participant in the pathogenesis of IgG4-RD.

Based on our findings of prominently increased circulating ILC2s, expression of surface markers, secretion of cytokines and correlation of ILC2s with Tregs, as well as with clinical parameters, we focussed on investigating the role of ILC2s in IgG4-RD. ILC2s play critical roles in asthma, hyperreactive airway and autoimmune diseases through the production of type 2 inflammation [35]. ILC2s have recently gained attention for modulating remodelling and fibrosis in tissues [36]. ILC2s could exacerbate lung inflammation and fibrosis [31, 37] and correlate with the degree of skin fibrosis in SSc, suggesting that ILC2s may be a useful predictive biomarker of fibrotic activity in SSc [38].

As reported, IgG4-RD has a high prevalence of allergy, and fibrosis is one of the most distinct characteristics of IgG4-RD [3]. We inferred that the prominent accumulation of ILC2s in peripheral blood in IgG4-RD might be related to the allergic reaction and fibrosis, therefore we tested ILC2 secretion of cytokines and showed more secretion of IL-9 through the IL-33 and ST2 pathways. IL-9 was originally described in parasitic infections and allergic diseases and has also been shown to participate in fibrosis, where it was found to be elevated in patients with idiopathic pulmonary fibrosis [39]. It is worth noting that in IgG4-RD patients, Treg cells were increased in peripheral blood, which might participate in IgG4 production and fibrosis via secretion of TGF-β [40]. IL-9-producing ILC2s were found to promote Treg cells activation and proliferation in RA [41], suggesting that IL-9 may not only induce fibrosis directly, but may also promote fibrosis through active Tregs in IgG4-RD. We speculate that the positive correlation of ILC2s with peripheral Treg cells in our study probably demonstrated the association of ILC2s with fibrosis of IgG4-RD. Additionally, ILC2 activation was required for Treg cell accumulation via ICOSL–ICOS costimulatory interactions [42]. In our study, ICOS was highly expressed in ILC2s of IgG4-RD patients, which may further indicate pivotal significance of ILC2s in mediating the survival of Treg cells. Thus it is possible that there is an intimate relationship between ILC2 and Treg cells in IgG4-RD. We also investigated the relationship between ILC2s in tissue and fibrosis score. Unfortunately, because of the sample size and the complexity of different immune cells in the tissue homeostasis system, we can only speculate that ILC2s might contribute to tissue fibrosis. Taken together, the aforementioned findings may provide new evidence for participation of ILC2s in the fibrosis pathology of IgG4-RD.

Except for the possible role of fibrosis, our data indicate that ILC2s may participate in the abnormal development of the IgG4-RD adaptative immune response. ILC2s could stimulate Th2 cells via PD-L1–PD-1 interaction and promote type 2 immunity. Moreover, stimulated ILC2s also induced IgM, IgG, IgA and IgE production by B cells through IL-25–IL-33 stimulation or Toll-like receptor triggering [23]. Our study revealed higher expression of CD154 in ILC2s and correlated positively with B cells. In addition, PD-1 and CXCR5 expressed in ILC2s correlated positively with serum IgG4 and IgE, respectively. Therefore prominent peripheral blood ILC2s may contribute to elevated serum IgG4 and IgE levels in IgG4-RD.

IL-33 is an epithelial-derived cytokine [43], an important initiator of type 2–associated mucosal inflammation and immunity, participating in the maintenance of progressive type 2 inflammation and fibrosis [44]. Abnormality of IL-33 is involved in many immune diseases, including liver fibrosis, allergic disease and SSc [45–47]. The expression of IL-33 and ST2 were elevated in serum and salivary tissues of patients with pSS [48], indicating potential involvement of the IL-33–ST2 axis in the pathogenesis of pSS [49]. In our study, IL-33 was not only elevated in peripheral blood of IgG4-RD, but was also expressed in involved tissue of IgG4-RD, which indicated that the IL-33–ST2 axis might be involved in the activation of ILC2s, thus inducing inflammation and fibrosis in IgG4-RD.

Limitations

There were some limitations in our study. First, as ILC2s are rare in lymphocytes, it is difficult to perform co-culture experiments to prove the relationship between ILC2s and Treg cells, which we only demonstrated by correction analysis. In addition, due to the small number of ILC2s, we could not prove the main source of IL-9 was ILC2s. Next, among ILC subsets, ILC1s are dominant in affected organs of IgG4-RD while ILC2s are dominant in circulation but decreased in affected organs; lower expression of CXCR5 might explain less migration of ILC2s into tissues. However, whether ILC1s participated in the submandibular damage of IgG4-RD needs further investigation.

Conclusions

ILCs and their subsets are significantly altered in IgG4-RD. We demonstrated the dysfunction of ILC2s in IgG4-RD by phenotype, correlation analysis and function investigation, revealing that ILC2s participate in the pathogenesis of IgG4-RD.

Acknowledgements

We are sincerely grateful to Prof. Ruie Feng and Dr Hui Zhang from the Department of Pathology, Peking Union Medical College Hospital, Chinese Academy of Medical Science and Peking Union Medical College for helping us to evaluate the fibrosis of all SMG tissues. We thank all patients and healthy controls in our study.

P.Z. and Z.L. performed all the experiments and statistical analyses and wrote the manuscript. L.P., J.Z., M.W., J.L., H.L., C.H., L.Z., H.Y., Q.W., Y.F., X.Zhang, X.Zeng participated in patient enrolment and clinical data collection. Y.Z. and W.Z. conceived of the study, participated in its design and coordination and helped to draft and revise the manuscript. All authors read and approved the final manuscript.

Funding: This work was supported by the National Natural Science Foundation of China (81771757, 81771780, 82071839), the Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (NWB20203346), Capital’s Funds for Health Improvement and Research (2020-2-4017) and the Beijing Municipal Science and Technology Commission (Z201100005520023).

Disclosure statement: The authors have declared no conflicts of interest.

Data availability statement

The data underlying this article cannot be shared publicly due to the privacy of individuals who participated in the study. The data will be shared on reasonable request to the corresponding author.

Supplementary data

Supplementary data are available at Rheumatology online.

References

Author notes