The IL-33/ST2 axis augments effector T-cell responses during acute GVHD

Abstract

Interleukin (IL)-33 binding to the receptor suppression of tumorigenicity 2 (ST2) produces pro-inflammatory and anti-inflammatory effects. Increased levels of soluble ST2 (sST2) are a biomarker for steroid-refractory graft-versus-host disease (GVHD) and mortality. However, whether sST2 has a role as an immune modulator or only as a biomarker during GVHD was unclear. We show increased IL-33 production by nonhematopoietic cells in the gastrointestinal (GI) tract in mice post-conditioning and patients during GVHD. Exogenous IL-33 administration during the peak inflammatory response worsened GVHD. Conversely, GVHD lethality and tumor necrosis factor-α production was significantly reduced in il33(-/-) recipients. ST2 was upregulated on murine and human alloreactive T cells and sST2 increased as experimental GVHD progressed. Concordantly, st2(-/-) vs wild-type (WT) donor T cells had a marked reduction in GVHD lethality and GI histopathology. Alloantigen-induced IL-18 receptor upregulation was lower in st2(-/-) T cells, and linked to reduced interferon-γ production by st2(-/-) vs WT T cells during GVHD. Blockade of IL-33/ST2 interactions during allogeneic-hematopoietic cell transplantation by exogenous ST2-Fc infusions had a marked reduction in GVHD lethality, indicating a role of ST2 as a decoy receptor modulating GVHD. Together, these studies point to the IL-33/ST2 axis as a novel and potent target for GVHD therapy.

Figure 1

IL-33 expression in the murine small intestine increases upon TBI and in the colon of stage IV GVHD patients. (A-B) After TBI of BALB/c mice with 9 Gy, RNA expression levels of multiple genes in the small bowel were analyzed using microarray-based analysis. This experiment was performed once. (A) Shown is the tile display for the most significantly regulated genes encoding for cytokines, expressed by Robust Multichip Average (RMA). Signal values of 4 individual samples in the following groups are shown: untreated, 24 hours, and 48 hours after TBI. Green rectangle: IL-33 is the second most significantly regulated gene. (B) Each RMA value for IL-33 is shown, including the P values. (C-D) Gut tissue from control (day 0), lethally-irradiated (day 2 post-irradiation), and WT or C57BL/6 recipients, that had received either 2.5 × 10⁶ T cells and 1 × 10⁷ BM cells, or BM only from BALB/c donors in GVHD studies (day 7 or day 14) were stained for IL-33, CD45, vimentin, and 4,6 diamidino-2-phenylindole (DAPI). This experiment was performed once. Images are representative of n = 3 mice/group and graphs were generated by averaging 2 to 3 fields (IL-33/DAPI–20×; IL-33/CD45, or IL-33/vimentin–20×) per animal. Fluorescence area/intensity reported as arbitrary units. *P < .05; ***P < .001. This experiment was performed once. (E) Gut tissue from BALB/c mice 7 days after transplantation with 5 × 10⁶ BM cells and 3 × 10⁵ T cells from C57Bl/6 mice. Prior to transplantation on day 0, mice received either TBI (2× 4,5 Gy), busulfuan/cyclophosphamide (Bu/Cy: busulfan day −7 to day −4 [dose: 10 mg/kg]), cyclophosphamide day −3 and day −2 (dose 100 mg/kg), treosulfan/cyclophosphamide (Treo/Cy: treosulfan day −6 to day −4 [dose: 1.5 g/kg]), cyclophosphamide day −3 and day −2 [dose 100 mg/kg]), or thiotepa/cyclophosphamide (Thio/Cy: thiotepa day −6 to day −4 [dose: 10 mg/kg], cyclophosphamide day −3 and day −2 [dose 100 mg/kg]) conditioning regimens. Fluorescence area/intensity reported as in (C-D). (Ei) Bar diagram for the indicated groups. (Eii) Representative tissue sections. Images are representative of n = 2 to 3 mice per group and graphs were generated by averaging 3 to 6 fields (IL-33/DAPI–20×) per animal. This experiment was performed once. (F-G) The amount of IL-33 in human colon biopsies was quantified by immunohistochemistry as shown for one representative section per group ([Fi] no GVHD; [Fii] GVHD IV° of the intestines) and for multiple patients (G). Samples were taken at different time points after allo-HCT in patients without GVHD or with GVHD grade 3/4. Patients’ characteristics, time of biopsy, conditioning, and immunosuppression data are shown in supplemental Table 1.

Figure 2

IL-33 deficiency in the host nonhematopoietic cells decreases GVHD severity. (A) Survival of C57BL/6 mice (WT or IL-33 deficient) after TBI (1000 cGy) and transplantation of 8 × 10⁵ T cells and 5 × 10⁶ BM from BALB/c mice. Data pooled from 3 independent experiments. (B) Serum was taken from recipient mice on day 7 post–allo-HCT, and TNF-α production was measured by enzyme-linked immunosorbent assay (ELISA). Data from a single experiment.

Figure 3

Administration of IL-33 post–allo-HCT increases the severity of GVHD. (A) C57BL/6 recipient mice of either 5 × 10⁶ BALB/c (allo) or C57BL/6 (syngeneic) BM and 5 × 10⁶ T cells (CD90-sorted) were treated with recombinant IL-33 (from day 3 to day 7 after BMT; 0.4 μg/dose) or phosphate-buffered saline (PBS) as control and monitored for survival. Allo BM+T-IL-33 vs Allo BM+T-PBS. P = .017. (B) Clinical GVHD scores were monitored throughout the experiment (Bi) and statistical differences were assessed on day 8 after allo-HCT (Bii) for the experiment shown in (A). (C) Concentrations of TNF-α were measured on day 8 after allo-HCT in the serum of C57BL/6 allogeneic recipients (received BM or BM plus T cells from BALB/c mice) treated with IL-33 (5 × 0.2 μg) or PBS. (D-F) Number of BALB/c donor T cells (D), a₄β₇⁺ T cells (E), or IFN-γ⁺ T cells (F) isolated from lamina propria of the small intestine of C57BL/6 allogeneic recipients treated with IL-33 (5 × 0.2 μg) or PBS on day 8 after allo-HCT. NS, not significant.

Figure 4

Increased expression of ST2 on T cells augments T-cell function during aGVHD. (A) ST2 was measured by flow cytometry on murine CD4⁺ T cells isolated from naïve mice after in vitro stimulation with allogeneic DC. Data are representative of 2 experiments. (B) ST2 expression on human CD4⁺ IFN-γ⁺ T cells in the peripheral blood analyzed by flow cytometry. Each data point represents the MFI for ST2 of an individual patient or healthy volunteer donor. Myeloablative conditioning (cond.) was performed with busulfan/cyclophosphamide or fludarabin/BCNU/melphalan, and samples were collected within the range of 5 to 10 days after allogeneic HCT. This experiment was performed once. (C) Survival of mice after allo-HCT performed as described for the 129 into BALB/c combination with 5 × 10⁶ BM cells and 3 × 10⁵ WT T cells, or st2^−/− T cells as indicated. The experiment was performed 3 times and the data were pooled. (D-F) Survival (D), clinical GVHD scores (E), and weight (F) of mice after allo-HCT was performed as described for the BALB/c into C57BL/6 combination with 1 × 10⁷ BM cells and 2.5 × 10⁶ WT BALB/c T cells, or st2^−/− T cells as indicated. This experiment was performed twice and representative data shown. (G) Histopathologic GVHD severity of the intestines and liver isolated on day 8 from mice treated as described under (C) when quantified as mean + SEM. Data are pooled from 2 independent experiments. (H) Survival of WT C57Bl/6 mice after allo-HCT with 1 × 10⁷ BM cells and 2.5 × 10⁶ WT BALB/c T cells, or st2^−/− T cells, performed as described under (D) with prior deletion of CD25⁺ cells within the transferred donor T cells. This experiment was performed once (n = 10). MFI, mean fluorescence intensity.

Figure 5

Upregulation of IL-18R and IFN-γ production in response to alloantigen is reduced in st2^−/− T cells. (A-B) Gene expression was quantified in WT and st2^−/− T cells on the RNA level by microarray analysis. WT or st2^−/− T cells were exposed to allogeneic irradiated DC for 48 hours. The tile display for the most significantly regulated genes expressed by RMA signal values of 4 individual samples in each group shown at the RNA level. (C) The values of individual mice for serum IFN-γ is shown on day 8 after allo-HCT. The experiment was performed twice and the data were pooled. (D) WT and st2^−/− CD4⁺/CD8⁺ T cells stimulated with allo–BM-DCs (2:1 ratio). ELISA for IFN-γ was performed after 24 hours of exposure. The experiment was performed twice with similar results.

Figure 6

Increased sST2 levels decrease GVHD-associated mortality. (A) Lethally irradiated C57BL/6 mice (1000 cGy) received 5 × 10⁶ BM cells from BALB/c donors (BM) or 5 × 10⁶ BM cells plus 5 × 10⁶ purified CD90.2⁺ T cells from BALB/c donors (BM+T), and sST2 concentration in the serum was measured on day 7 and day 14 after allo-HCT by ELISA. *P < .05. (B) Survival of WT C57BL/6 mice after allo-HCT with 5 × 10⁶ BM cells and 8 × 10⁵ WT T cells from BALB/c recipients performed as described. Mice received sST2-Fc or isotype control IgG as indicated. The experiment was performed twice and the data were pooled. (C) The mean value ± SEM for (i) serum IL-6 and (ii) IFN-γ detected by ELISA on day 8 after allo-HCT. The experiment was performed twice and the data were pooled. (D) Histopathologic GVHD severity of the intestines (i, ii) and liver (iii) isolated on day 8 from mice treated as described under panel (B) when quantified as mean + SEM. Data are pooled from 2 independent experiments.