The interleukin-33 pathway protects mice from immune-mediated hepatitis by inducing AREG-expressing ILC2, anti-inflammatory macrophages and regulatory T cells

Background:

Autoimmune hepatitis (AIH) is a chronic inflammatory liver disease that is due to loss of tolerance towards hepatic self-antigens. Consequently, autoreactive T cells mediate ongoing destruction of liver tissue, thereby causing chronic injury and necroinflammation. The alarmin interleukin (IL)-33 is rapidly released from necrotic hepatocytes and binds to its target receptor ST2, resulting in activation of type 2 innate lymphoid cells (ILC2) and regulatory T cells (Treg). IL-33 was recently shown to ameliorate immune-mediated hepatitis, but the underlying mechanisms remain unknown. Following activation, ILC2 and Treg express the epidermal growth factor amphiregulin (AREG), which has regenerative capacities, but also immune-modulatory functions. This prompted us to investigate if IL-33 responsive immune cells regulate hepatic inflammation in an AREG-dependent manner.

Methods:

Immune-mediated hepatitis was induced in C57Bl/6 (WT), ST2-/- or AREG-/- mice by Concanavalin (Con) A treatment. Additionally, a subgroup of mice was treated with IL-33 on 3 consecutive days prior to Con A. Severity of liver damage was assessed by measurement of serum ALT activities. Frequencies of immune cells were analyzed by flow cytometry and evaluated by an unbiased tSNE-approach. Transcriptional profiles in liver tissue were analyzed by qRT-PCR.

Results:

IL-33 treatment prior to Con A application protected WT mice from immune-mediated hepatitis, as shown by significantly reduced serum ALT activities. Moreover, hepatic ILC2 expanded in response to both, IL-33 and Con A. ILC2 expressed PD-1 in response to Con A, while PD-1 was downregulated in response to IL-33. Additionally, the activation marker KLRG1 was significantly upregulated on ILC2 in response to IL-33 treatment, which correlated positively with AREG expression, indicating that expression of AREG in ILC2 requires full cellular activation. Frequencies of hepatic macrophages were increased in response to all treatment regimens. Con A induced an M1 like Ly6Chi phenotype, while IL-33 treatment favored a Ly6Clo M2 like phenotype. Treg were expanded in response to IL-33 and Con A and showed increased expression of ST2. However, the highest frequencies of ST2+ Treg were only seen in response to IL-33 treatment. To study the relevance of ST2+ Treg in the regulation of immune-mediated hepatitis, ST2-/- mice were treated with Con A. These mice showed an even more severe hepatitis, albeit Treg frequencies were significantly elevated. This indicates that IL-33-responsive Treg are required to regulate liver inflammation. Indeed, adoptive transfer of IL-33 elicited WT Treg to naïve WT recipients suppressed immune-mediated hepatitis, while adoptive transfer of naïve Treg was not protective. Phenotypic analysis of ST2+ Treg, showed higher expression of EGFR, suggesting that AREG might promote suppressive function of ST2+ Treg. In a first attempt to study a role of AREG for regulation of immune-mediated hepatitis, we treated AREG-/- mice with Con A and observed a more severe hepatitis in these animals than in WT mice. Interestingly, frequencies of ST2+ Treg were as high as in WT mice, indicating an impaired function of ST2+ Treg in the absence of AREG.

Conclusion:

IL-33 represents a critical factor that induces AREG+ ILC2 and expands ST2+ Treg to counteract immune-mediated hepatitis.