A Proinflammatory Role of Type 2 Innate Lymphoid Cells in Murine Immune-Mediated Hepatitis

This article is featured in In This Issue, p.1

Autoimmune hepatitis (AIH) describes a loss of immune tolerance to liver self-antigens and is characterized by progressive necroinflammation and destruction of the hepatic parenchyma triggered by immune-mediated processes that are not fully understood (1, 2). Con A–induced liver inflammation is a well-established mouse model of immune-mediated hepatitis, which shares various similarities with AIH (3–5). After administration of the mitogen Con A, activated CD4⁺ T cells and NKT cells produce inflammatory cytokines such as IFN-γ and TNF-α, thereby causing hepatocyte death by necrosis (6–8). However, activated T cells alone are not sufficient for the development of chronic liver inflammation. Thus, cell populations of the innate immune response also play an important role in the pathogenesis of autoimmune liver diseases (9, 10).

Innate lymphoid cells (ILC) belong to the innate immune system and contribute to the pathology of inflammatory diseases such as atopic dermatitis, asthma, Crohn’s disease, and pulmonary inflammation (11). ILCs produce effector cytokines that match those of certain T helper subsets and are classified into three major groups based on their cytokine expression profile (12, 13). Type 2 ILCs (ILC2) express the cytokines IL-5 and IL-13 upon activation by IL-33 or IL-25. ILC2s participate in immune responses against helminths (14), and mediate airway hyper-responsiveness (15) and lung inflammation (16). ILC2s also function in tissue remodeling and mediate hepatic fibrosis (17). However, the precise immunomodulatory role of ILC2s in liver disease is still largely unknown.

IL-33 functions as an alarmin that is released during necrotic cell death associated with tissue damage. IL-33 binds to a heterodimeric receptor comprising ST2 and the IL-1 receptor accessory protein (18). ILC2s express ST2 and respond to IL-33–mediated signaling by strong expression of IL-5 and IL-13. ILC2–derived IL-5 controls eosinophil maintenance and infiltration (19, 20), whereas IL-13 induces airway hyper-responsiveness (15) and contributes to parasite expulsion (21). Previous studies have implicated IL-33 in the pathogenesis of liver diseases such as chronic viral infection (22), primary biliary cirrhosis (23), and liver fibrosis (17, 24). IL-33 is expressed by hepatocytes during immune-mediated hepatitis (25), but whether IL-33–elicited ILC2s contribute to disease pathogenesis has not been studied yet.

CD4⁺ Foxp3⁺ regulatory T cells (Treg) are crucial for the control of inflammatory immune responses and contribute to the regulation of immune-mediated hepatitis (26). Recently, the IL-33/ST2 pathway has been shown to promote Treg function at sites of inflammation (27), but so far it is not known whether IL-33 influences the regulatory T cell response during immune-mediated hepatitis.

In the current study, we identified hepatic ILC2s as a new effector cell population of the innate immune system that responds to IL-33 release upon liver tissue damage through expression of IL-5 and IL-13, thereby driving pathogenesis of immune-mediated hepatitis. We further showed that IL-33 not only activates hepatic ILC2s but also expands ST2⁺ Tregs, which might contribute to the regulation of immune-mediated hepatitis.

Male C57BL/6 and Rag1^−/− mice were obtained from the University Medical Center Hamburg-Eppendorf animal facility (Hamburg, Germany). B6-THY1aPL mice were a kind gift from Prof. H.W. Mittrücker (Department of Immunology, University Medical Center Hamburg-Eppendorf, Germany). Mouse experiments were conducted according to the German animal protection law and approved by the institutional review board (G48/11 and G44/15; Behörde für Gesundheit und Verbraucherschutz, Hamburg, Germany). Mice received humane care according to the national guidelines of the National Institutes of Health in Germany.

Hepatitis B patients treated at the hepatitis outpatient department of the University Medical Center Hamburg-Eppendorf were enrolled in the study for which all patients gave written consent and which was approved by the Ethics Committee of the Hamburg Chamber of Physicians (PV3941). Patients were stratified according to their clinical course into three groups: acute hepatitis B (AHB, n = 8), chronic hepatitis B (CHB, n = 13) or resolved hepatitis B (RHB, n = 3). Healthy donors of the blood transfusion service at the University Medical Center Hamburg-Eppendorf were anonymously enrolled in the study as healthy controls (HC, n = 35).

Con A (7 mg/kg; Sigma-Aldrich, München, Germany) or saline (PBS) were administered i.v. into recipient mice. Mice were sacrificed 8 or 24 h post Con A challenge. Heart blood was drawn from individual mice and liver injury was quantified by automated measurement of plasma activities of alanine transaminase (ALT) using a COBAS Mira System (Roche Diagnostic, Mannheim, Germany).

Liver samples were fixed with 4% formalin and embedded in paraffin. Liver sections of 3 μm were cut, stained with H&E following standard protocols, and analyzed by light microscopy.

C57BL/6 mice received i.p. recombinant murine (rm) IL-33 (0.3 μg; R&D Systems, Wiesbaden-Nordenstadt, Germany) daily on three consecutive days. Hepatic single cell suspensions were stained with the Lineage Ab Cocktail (BD Pharmingen, Heidelberg, Germany), anti-CD127 (A7R34; BioLegend, Fell, Germany), and anti-ST2 (RMST2-2; eBioscience, Frankfurt, Germany) Abs. Lin⁻ cells were enriched by magnetic cell sorting and lin⁻ Sca-1⁺ ST2⁺ ILC2s were isolated with BD FACSAria III sorter (BD Biosciences). C57BL/6 mice were injected i.v. with 2 × 10⁵ ILC2s. Mice were challenged with Con A 1 d later and analyzed 8 h after hepatitis induction.

Rag1^−/− mice were reconstituted with CD4⁺ T cell by adoptive transfer of 1 × 10⁶ CD4⁺ T cells from B6-THY1aPL mice expressing the CD90.1 Ag. Recipient mice were monitored for successful reconstitution after 2 wk by detecting transferred CD4⁺ T cells in blood using anti-CD4 (GK1.5) and anti-CD90.1 (OX-7; all BioLegend) Abs. CD90.2⁺ ILC2s in reconstituted Rag1^−/− mice were depleted using anti-CD90.2 Ab (30H12; BioXCell, West Lebanon, NH). Mice were injected i.p. daily on four consecutive days with 200 μg of anti-CD90.2 Ab. One day later, mice received i.v. Con A and were analyzed 8 h after hepatitis induction.

Cells were incubated with anti-CD16/32 Ab (93; BioLegend) prior to Ab staining to prevent unspecific binding. Fixable Viability Dye eFlour506 (eBioscience) was used to exclude dead cells. Single cell suspensions from hepatic leukocytes and splenocytes were stained with the Lineage Ab Cocktail, anti-CD127 (A7R34), anti-ST2 (RMST2-2), anti-c-Kit (2B8), anti-Sca-1 (D7), anti-CD45 (30-F11), anti-CD11b (M1/70), anti-CD4 (GK1.5), anti-CD90.2 (53-2.1), anti-Foxp3 (MF-14; all BioLegend), or anti-Siglec-F (E50-2440; BD Pharmingen) Abs. Data were acquired using a BD LSRFortessa II (BD Biosciences) and analyzed by FlowJo software (Tree Star, Ashland, OR).

Cells were stimulated with PMA (10 ng/ml) and ionomycin (250 ng/ml) for 5 h with the addition of brefeldin A (1 μg/ml; all Sigma Aldrich) after 1 h. Thereafter, cells were stained for surface Ags and were treated with Fixable Viability Dye eFluor506 before fixation with Fix/Perm solution (eBioscience). Cells were stained in Perm buffer (BioLegend) with anti-IL-13 (eBio13A; eBioscience) or anti-IL-5 (TRFK5; BD Pharmingen) Abs. To analyze GATA3 expression, Fixable Viability Dye eFluor506-labeled cells were fixed with 2% PFA before intranuclear staining with anti-GATA3 (L50-823; BD Pharmingen) Ab using the Transcription Factor Staining Buffer Set (eBioscience).

Detection of total amounts of IL-13 and IL-33 in human plasma was conducted using ELISA DuoSets from R&D systems. The manufacturer’s Avidin-HRP conjugate was replaced by PolyHRP80 streptavidin conjugate (SDT Reagents, Baesweiler, Germany) to achieve a 10-fold lower limit of detection. Lower limit of detection (background + 3× SD) was generally at 2–5 pg/ml for all analytes. Murine plasma IFN-γ concentrations were determined by ELISA (BioLegend) according to the manufacturer’s protocol. For cytokine detection in liver organ culture supernatants, liver biopsies (∼5 mm³) were placed in 48 flat-bottom well culture plates containing 400 μl serum-free RPMI 1640 medium supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin (PAA Laboratories). After 18 h at 37°C, culture supernatants were tested for IL-33 amounts by ELISA (eBioscience) according to the manufacturer’s protocol and normalized to the weight of the liver biopsy.

Total RNA was isolated from shock-frozen liver tissue using the NucleoSpin RNA Kit (Machery-Nagel, Duren, Germany) according to the manufacturer’s instructions. Then 1 μg RNA was transcribed into cDNA using the Verso cDNA Synthesis Kit (Life Technologies, Carlsbad, CA) on a MyCycler thermal cycler (BioRad, München, Germany). Quantitative real-time PCR (qRT-PCR) was performed using the Absolute qPCR SYBR Green Mixes (Thermo Scientific). The relative mRNA levels were calculated using the cycle threshold (∆∆CT) method after normalization to the housekeeping gene mitochondrial ATP synthase. Quantification was shown in x-fold changes to the corresponding control cDNA. Primers were designed for detection of exon overlapping amplicons and were obtained from Metabion (Martinsried, Germany). Sequences of the primers are listed in the Supplemental Table I.

Data were analyzed using the GraphPad Prism software (GraphPad Software, San Diego, CA). Statistical comparison was carried out using the one-way ANOVA, employing Bonferroni’s post tests. Data were expressed as means ± SEM. A p value <0.05 was considered statistically significant with the ranges *p < 0.05, **p < 0.01, and ***p < 0.001.

Current evidence indicates that the IL-33 pathway is involved in the pathogenesis of certain liver diseases. We analyzed IL-33 expression in the mouse model of immune-mediated hepatitis as well as in patients with acute and chronic immune-mediated hepatitis due to hepatitis B virus infection. We demonstrated strongly elevated hepatic IL-33 expression during immune-mediated hepatitis (Fig. 1A) and significantly increased IL-33 levels in the plasma of patients with acute and chronic hepatitis B virus infection compared with healthy controls and patients with resolved infection (Fig. 1B), further supporting the assumption of an immunomodulatory role of IL-33 in acute and chronic liver diseases.

Because IL-33 activates and expands ILC2s, we asked whether this type of innate effector cell population might contribute to the pathogenesis of immune-mediated hepatitis. We stained single cell suspensions from hepatic leukocytes and splenocytes for lin⁻ c-Kit⁺ CD127⁺ Sca-1⁺ ST2⁺ cells to identify IL-33–elicited ILC2s (Supplemental Fig. 1) (13). Upon Con A challenge, we showed a strong increase in the frequency and number of hepatic ILC2s, whereas splenic ILC2s did not significantly expand (Fig. 1C). ILC2 expansion in the liver not only correlated with hepatic IL-33 expression (see Fig. 1A) but also with severe hepatitis characterized by strongly enhanced plasma ALT activity (Fig. 1D), hepatocyte death by necrosis leading to formation of necrotic lesions (Fig. 1E), and the expression of the inflammatory cytokine IFN-γ (Fig. 1F, 1G). These data indicate that the inflammatory hepatic environment specifically supports expansion of ILC2s in the liver.

We analyzed activation of ILC2s after Con A challenge and found that hepatic mRNA expression of GATA3, a transcription factor essential for ILC2 function (28, 29), was highly upregulated during immune-mediated hepatitis (Fig. 2A). ILC2s from livers of Con A–treated mice expressed IL-13, IL-5, and GATA3 (Fig. 2B), demonstrating activation and type 2 cytokine expression by hepatic ILC2s during liver inflammation.

We further showed that hepatic IL-13 and IL-5 mRNA expression significantly increased during immune-mediated hepatitis (Fig. 2C, 2D). In patients with acute and chronic hepatitis B infection, plasma IL-13 levels were substantially elevated (Fig. 2E) whereas IL-5 was not detectable, suggesting a more prominent role for IL-13 in the pathogenesis of human acute and chronic liver inflammation.

One mechanism by which ILC2s can amplify inflammation is via IL-5–mediated recruitment of eosinophils. We stained hepatic leukocytes and splenocytes for CD45⁺ CD11b⁺ Siglec-F⁺ cells to identify eosinophils and demonstrated a liver-specific accumulation of eosinophils during immune-mediated hepatitis (Fig. 2F).

To gain further insight into the role of ILC2s in immune-mediated hepatitis, we analyzed their ability to initiate liver inflammation by using Rag1^−/− mice that harbor ILCs but lack T cells and B cells. After administration of Con A to Rag1^−/− mice, hepatic GATA3 expression was not upregulated (data not shown), hepatic ILC2s did not expand (Fig. 3A) and failed to induce liver inflammation and tissue damage (Fig. 3B, 3C). When Rag1^−/− mice were reconstituted with CD4⁺ T cells 2 wk before Con A challenge, mice developed hepatitis (Fig. 3D–F). The CD4⁺ T cell-mediated tissue injury was associated with hepatic IL-33 expression (Fig. 3G) and was found to be the prerequisite for GATA3 upregulation (Fig. 3H) and ILC2 expansion in the liver (Fig. 3I).

To study the functional role of ILC2s during liver inflammation, we assessed the pathogenesis of immune-mediated hepatitis in the absence of ILC2s. Recent studies have used an anti-CD90.2 Ab to deplete CD90.2-expressing ILCs in vivo (30, 31). We generated CD4⁺ T cell-replete chimeras by transfer of CD90.1⁺ CD4⁺ T cells from B6-THY1aPL mice into Rag1^−/− mice expressing the CD90.2 allele. These mice received an anti-CD90.2 Ab or isotype control daily on four consecutive days before Con A challenge. We demonstrated that anti-CD90.2 Ab treatment did not affect the frequency of hepatic CD90.1⁺ CD4⁺ T cells (Fig. 4A), the effector cells that initiate liver inflammation, whereas hepatic ILC2 expansion was significantly reduced (Fig. 4B). Ablation of ILC2s strongly decreased hepatic GATA3 expression (Fig. 4C). Because hepatic ILC2s of isotype-treated mice expressed GATA3 (Fig. 4D), reduced hepatic GATA3 levels can be attributed to ILC2 depletion before hepatitis induction. This correlated with less severe hepatitis because anti-CD90.2 Ab-treated mice had substantially diminished plasma ALT activity (Fig. 4E) and developed smaller necrotic lesions compared with isotype-treated mice (Fig. 4F). Furthermore, expression of the inflammatory cytokines IFN-γ (Fig. 4G), TNF-α, and IL-6 (Fig. 4H) was strongly reduced. In addition, recruitment of eosinophils into the inflamed liver was impaired in Ab-treated mice (Fig. 4I). These data show that ILC2s amplify immune-mediated liver inflammation and tissue damage.

To further investigate the function of IL-33–elicited ILC2s in immune-mediated hepatitis, we performed cell transfer experiments and analyzed disease pathogenesis. We first treated C57BL/6 mice with rmIL-33 on three consecutive days, resulting in a strong induction of hepatic GATA3 expression (Fig. 5A) and highly increased frequency and number of ILC2s in the liver (Fig. 5B), which were sorted by MACS/FACS (Fig. 5C, Supplemental Fig. 2). These IL-33–elicited hepatic ILC2s were functionally active because they strongly expressed GATA3 and IL-5 and, to a much lesser extent, IL-13 (Fig. 5C). Purified hepatic ILC2s were adoptively transferred into recipient mice 1 d before Con A challenge leading to aggravation of immune-mediated hepatitis (Fig. 5D, 5E) and enhanced expression of IFN-γ (Fig. 5F). Hepatic IL-13 expression was not altered in ILC2-treated mice (data not shown) whereas IL-5 levels were elevated (Fig. 5G), most likely due to high IL-5 expression by transferred ILC2s. In summary, these data strongly indicate a proinflammatory role of IL-33–activated ILC2s in immune-mediated hepatitis.

Having shown that 3-d IL-33 treatment strongly expanded ILC2s in the liver (see Fig. 5B), we used this strategy to analyze disease pathology of immune-mediated hepatitis in the presence of endogenously elevated hepatic ILC2 numbers. Therefore, C57BL/6 mice were treated with rmIL-33 on three consecutive days before Con A challenge. Upon IL-33/Con A treatment, the frequency of ILC2s was highly increased in the liver (Fig. 6A). Interestingly, this did not result in aggravated liver inflammation and tissue damage. On the contrary, we found that 3-d IL-33 treatment before Con A challenge suppressed development of hepatitis (Fig. 6B, 6C) and inflammatory cytokine expression, whereas IL-5 expression was strongly enhanced (Fig. 6D, 6E) and resulted in massive accumulation of eosinophils in the liver (Supplemental Fig. 3).

Regarding the observed anti-inflammatory effect of IL-33, we found that the frequency and number of CD4⁺ Foxp3⁺ Tregs were highly increased in the liver of mice that received IL-33 for 3 d compared with PBS-treated mice (Supplemental Fig. 4A). Strikingly, a substantial proportion of these Tregs expressed the IL-33 receptor ST2 (Supplemental Fig. 4B). The same was true for 3-d IL-33 treatment before Con A challenge, where we found highly elevated frequency and numbers of CD4⁺ Foxp3⁺ Tregs (Fig. 6F, Supplemental Fig. 1B) and ST2⁺ Foxp3⁺ Tregs in the liver (Fig. 6G, Supplemental Fig. 1B), indicating that IL-33 expands hepatic Tregs that might counteract inflammatory immune responses and subsequent development of hepatitis.

Previously, we have shown that CD4⁺ Foxp3⁺ Tregs infiltrate the liver 24 h after hepatitis induction and are involved in resolving inflammation and mediating tolerance during immune-mediated hepatitis (26). We asked whether Tregs expressing ST2 also arise in immune-mediated hepatitis where IL-33 is released upon liver tissue damage. We showed a strongly increased frequency of CD4⁺ Foxp3⁺ Tregs in both liver and spleen 24 h upon Con A challenge (Fig. 6H). Interestingly, ST2⁺ Foxp3⁺ Tregs specifically accumulated in the inflamed liver (Fig. 6I), suggesting that this Treg subset might contribute to regulation and termination of acute flares of liver inflammation in immune-mediated hepatitis.

Acute and chronic immune-mediated liver disease as it occurs during infection with hepatitis B virus or sterile immunity, e.g., in AIH, is a life-threatening disease triggered by so far poorly defined mechanisms. Thus, understanding the molecular and cellular networks involved in the pathogenesis of immune-mediated hepatitis is of high clinical relevance. In this study, to our knowledge we were first to demonstrate that ILC2s participate in immune-mediated hepatitis and contribute to disease pathogenesis by a fast response to the local alarmin IL-33.

Recent studies reported a crucial role for IL-33 in the pathology of immune-mediated liver diseases (22–24). During immune-mediated hepatitis, highly elevated IL-33 levels correlated with severe liver inflammation and tissue damage. We identified the innate lymphoid subset ILC2 as one effector cell population that selectively expands in the inflamed liver. ILC2s express the IL-33 receptor ST2 and strictly depend on the transcription factor GATA3 for cytokine production in response to IL-33 (32–34). In line with this, we showed highly upregulated hepatic GATA3 expression as well as expression of GATA3, IL-5 and IL-13 by ILC2s in the inflamed liver, strongly indicating that hepatic ILC2s become activated by IL-33 during immune-mediated hepatitis.

We found that ILC2s do not initiate immune-mediated hepatitis and instead require CD4⁺ T cell-mediated liver tissue damage and subsequent IL-33 release for activation and expansion. Because reduction of ILC2 numbers correlated with less severe hepatitis whereas transfer of IL-33–activated ILC2s aggravated liver damage, we conclude that ILC2s amplify inflammatory immune responses during immune-mediated hepatitis. Together with another study reporting a critical role for ILC2s in the development of hepatic fibrosis (17), these data suggest that ILC2s drive inflammation and tissue damage in the context of IL-33–triggered liver diseases.

We detected elevated levels of IL-5 and IL-13 in the inflamed liver and showed that hepatic ILC2s expressed both cytokines during immune-mediated hepatitis. After adoptive transfer of IL-33–elicited hepatic ILC2s, IL-5 expression was increased in the liver and correlated with more severe hepatitis, suggesting an inflammatory function for ILC2–derived IL-5 in immune-mediated hepatitis. IL-5 was found to mediate eosinophilia (19, 20) and might be responsible for the observed eosinophil accumulation in the liver during immune-mediated hepatitis. Further evidence for an ILC2-driven recruitment of eosinophils into the liver came from the depletion experiments in which ILC2 depletion was associated with less hepatic infiltration of eosinophils. Because eosinophilia has been shown to contribute to disease pathogenesis (35), cytokine-mediated recruitment of eosinophils might be one mechanism by which hepatic ILC2s exert their proinflammatory function in immune-mediated hepatitis. So far, no function of IL-13 has been described in immune-mediated hepatitis. IL-13 was found to be proinflammatory in the pathogenesis of atopic dermatitis (30), parasite infection (36, 37), and allergy (38, 39). IL-13 also has a profibrotic function thereby driving progression of chronic liver disease (17, 40, 41), strongly indicating a more prominent role of IL-13 in chronic rather than acute diseases. Thus, hepatic ILC2s produce both IL-5 and IL-13 upon IL-33–mediated activation but seem to primarily contribute to the pathogenesis of immune-mediated hepatitis by IL-5–mediated recruitment of eosinophils. However, further work is needed to clearly define the mechanisms by which ILC2s drive liver inflammation.

Interestingly, endogenous expansion of ILC2s by 3-d IL-33 treatment before hepatitis induction did not result in more severe hepatitis. Instead, liver inflammation was suppressed despite strong IL-5 expression and massive accumulation of eosinophils in the liver. In preliminary experiments, we found impaired activation of eosinophils in IL-33–treated mice (data not shown), indicating that IL-33 suppresses inflammatory eosinophil function and pointing to one mechanism by which IL-33 might exert its anti-inflammatory function.

Recent studies reported the accumulation of Tregs expressing ST2 in the thymus and spleen (42) or visceral adipose tissue (43) after sequential administration of IL-33. We found that 3-d IL-33 treatment also expanded ST2⁺ Tregs in the liver, suggesting a previously unknown function for this cytokine in the hepatic regulatory T cell response. Previously, we have shown that CD4⁺ Foxp3⁺ Tregs play a pivotal role in ameliorating immune-mediated hepatitis (26). However, key factors controlling the Treg response in the inflamed liver are insufficiently understood. In this context, the IL-33/ST2 pathway was found to maintain Foxp3 expression under inflammatory conditions to ensure effective Treg function during inflammation (27). In this study we showed selective expansion of ST2⁺ Foxp3⁺ Tregs in the inflamed liver, which might be particularly capable of regulating the inflammatory immune response during immune-mediated hepatitis. The anti-inflammatory effect of IL-33 was further highlighted by the fact that 3-d IL-33 treatment before Con A challenge potently suppressed development of hepatitis, possibly due to the strong expansion of hepatic ST2⁺ Tregs.

Our results identify ILC2s as a new effector cell population of the innate immune system that respond to IL-33 release upon liver tissue damage through expression of type 2 cytokines, thereby contributing to the pathogenesis of immune-mediated hepatitis, probably by recruitment of eosinophils. The inflammatory activity of ILC2s might be counteracted by IL-33–elicited hepatic ST2⁺ Tregs that also arise in immune-mediated hepatitis. The fact that IL-33 induces multiple modules of the immune response that promote both immunity and tolerance provides important, and previously unknown, insights into the cellular and molecular mechanisms through which hepatic immune responses are regulated by distinct cell populations during immune-mediated hepatitis.

The authors gratefully acknowledge the excellent technical assistance of Elena Tasika and Carsten Rothkegel at the Institute of Experimental Immunology and Hepatology, University Medical Center Hamburg-Eppendorf. We also thank Prof. Hans-Willi Mittrücker (Department of Immunology, University Medical Center Hamburg-Eppendorf) for providing B6-THY1aPL mice and all members of the FACS Sorting Core Unit, University Medical Center Hamburg-Eppendorf for cell sorting.

The authors have no financial conflicts of interest.