IL-33 Released in the Liver Inhibits Tumor Growth via Promotion of CD4⁺ and CD8⁺ T Cell Responses in Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer, representing 85% of liver malignancies and ranks as the third most common cause of cancer-related death worldwide (1). The major reason for the high mortality in HCC is tumor progression and metastasis; however, the mechanisms that underlie tumor initiation, progression, and metastasis in the case of HCC are still poorly studied (2). Many studies have reported that in HCC patients, the cancer-associated microenvironment components, including immune cells, fibroblast cells, endothelial cells, and extracellular matrix can support the neoplastic cells to proliferate, grow, and invade (3–5). Chronic inflammation predisposes individuals to various types of cancer, including HCC (6). Tumor microenvironment with the recruited inflammatory immune cells prepares a niche for the neoplastic cells and facilitates cancer angiogenesis, metastasis, and invasion (4, 7). Cytokines are important mediators of this process, possessing both protumor and antitumor roles (8–12). Understanding their regulatory functions during tumor development can shed light on the immune interactions in the tumor microenvironment and provide therapeutic strategies for HCC.

IL-33, a member of the IL-1 family, interacts with a heterodimeric receptor comprising of suppression of tumorigenicity 2 (ST2) and IL-1R accessory protein (IL-1Racp). It is widely expressed in many tissues such as liver, lung, CNS, and many types of cells, including epithelial cells, smooth muscle cells, fibroblasts, and macrophages (13, 14). IL-33 is constitutively and abundantly expressed in normal tissues and can localize to the nucleus (15). It is thought to be released and function as an alarmin during cellular stress or death (16, 17). ST2 is expressed mainly in Th2, mast cells, but can also be found in regulatory T cells (Treg), group 2 innate lymphoid cells (ILC2), CD8⁺ T cells, and NK cells (18, 19). The interaction between IL-33 and ST2 leads to cell proliferation, survival, amphiregulin expression, and production of inflammatory mediators, including IL-1β, IL-3, IL-6, TNF, IL-5, and IL-13 (20, 21). IL-33 was first found to support type 2 immune responses, activating Th2 cells, Treg, and ILC2s (22, 23). Recent evidence also supports its role in type 1 immune responses involving IFN-γ production by invariant NKT cells, NK cells, and CD8⁺ T cells (24–27).

IL-33 released because of cell death or stress can lead to inflammatory responses. IL-33 administration to mice caused airway inflammation by activating ILC2s (28, 29), mobilizing eosinophils (30), and polarizing M2 macrophages (31). IL-33 levels were found to be increased and correlated with disease severity in Crohn disease and inflammatory bowel disease (32, 33). However, whether IL-33 plays a protective or pathological role in inflammatory bowel disease is still controversial. IL-33 contributes to Con A–induced hepatitis, and blockade of IL-33 ameliorates liver injury by impairing activation of T cells and NKT cells and reducing production of proinflammatory cytokines (13).

Serum IL-33 levels are increased in many cancers such as gastric cancer, HCC, head and neck cancer, breast cancer, and lung cancer (34–38). Both protumor and antitumor functions of IL-33 have been shown in previous studies. IL-33 may also function as an immune adjuvant for antitumor immune response. It has been shown that IL-33 can be used as an adjuvant for DNA vaccine to promote tumor-specific effector and memory T cell response (39). In contrast, IL-33 was involved in tumoral oncogenesis in myeloproliferative neoplasms and colorectal cancer (40, 41). Although IL-33 can be expressed in both cytoplasm and nucleus, HCC tissues only showed cytoplasmic IL-33 staining, which suggests that cytoplasmic or released IL-33 might play a role in HCC development (42). Moreover, tumor-infiltrating, IL-33–producing effector memory CD8⁺ T cells have been shown to be associated with prolonged survival of HCC patients (43), suggesting IL-33 produced in the tumor microenvironment could be an important mediator of antitumor immune response. However, the exact function of IL-33 in HCC, both intracellular and released, is still not known.

In the current study, we examined the function of released IL-33 in the liver during tumor development using murine HCC models. We demonstrated that IL-33 inhibited HCC growth in an ST2-dependent manner. IL-33 release recruited large numbers of lymphocytes to the liver and significantly increased percentages and numbers of activated CD4⁺ and CD8⁺ T cells as well as IFN-γ production and cytotoxicity of the CD8⁺ T cells. Our in vitro experiments further showed that IL-33 could directly activate CD4⁺ and CD8⁺ T cells. Depletion of CD4⁺ and CD8⁺ T cells in vivo diminished the antitumor activity of IL-33. In summary, our study demonstrated that IL-33 could inhibit HCC development via promoting antitumor CD4⁺ and CD8⁺ T cell responses.

Six- to eight-week-old specific pathogen-free male C57BL/6 mice and BALB/c mice were purchased from Shanghai Laboratory Animal Center (Shanghai, China). ST2^−/− mice were authorized by Dr. A.N.J. McKenzie of Medical Research Council Laboratory of Molecular Biology, Cambridge, U.K. and gifted by Dr. Q. Wang of Zhejiang University. All mice were housed in specific pathogen-free facilities. Animal studies, including mouse care and experimental procedures, were approved by the Institutional Laboratory Animal Care and Use Committee of Soochow University.

The mature peptide (S109 to I266) sequence of murine IL-33 was amplified from cDNA generated from hepa1-6 cell line. The IL-33 expression construct was generated by fusing the nucleotide sequence-encoding Igκ signal sequence to the 5′ end of IL-33 (S109 to I266) sequence and then inserted between sites of BamH I into minicircle (MC) plasmid (pMC.EF1; SBI, Palo Alto, CA) and lentiviral plasmid. MC plasmids were purified (44) and injected by hydrodynamic gene transfer (HGT) technique to express IL-33 in vivo. Lentiviral plasmid was used to generate IL-33 stable expressing cell line.

Human embryonic kidney cell line 293T and murine HCC cell line hepa1-6 were obtained from American Type Culture Collection (Manassas, VA) and cultured in complete DMEM (Life Technologies, Carlsbad, CA) containing 10% FBS. The murine hepatoma cell line H22 was gifted by Dr. Z. Fang of Shanghai University of Traditional Chinese Medicine and maintained in complete RPMI 1640 (Life Technologies) medium containing 10% FBS. The IL-33–expressing lentivirus generated from 293T cells was used to infect hepa1-6 cells. Empty vector was used as vector control. Yellow fluorescent protein-positive monoclonal cells were sorted into 96-well plates by FACSAria III flow cytometer (BD Biosciences, San Jose, CA) and amplified to stable clones.

Representative samples of liver were obtained from mice that received MC plasmids by HGT technique and conserved in 4% formalin, paraffin embedded, sectioned, and stained with H&E. The paraffin sections of liver were stained with an anti–IL-33 Ab (Abcam, Cambridge, U.K.) by immunohistochemistry at concentration of 1:1000 as previously described (45). Sections were observed and imaged under an optical microscope (Nikon, Tokyo, Japan).

Cell proliferation was detected by CCK-8 assay (Beyotime Biotechnology, Shanghai, China). Hepa1-6 cells expressing IL-33 or vector control (3000 per well) were plated to 96-well plates and then detected by CCK-8 assay after 48 h at 450 nm using a microplate reader (BioTek, Burlington, VT). Cell apoptosis was detected by PE–annexin V and PerCP–Cy5.5–7AAD staining and flow cytometry (BD Biosciences).

We established four murine HCC models in the current study. H22 cells (5 × 10⁶) were injected s.c. into BALB/c mice, tumors were measured every 2 d, and tumor volumes were calculated by ab²/2 (a, long diameter; b, short diameter). In one orthotopic HCC model, hepa1-6 cells expressing IL-33 or vector control (1 × 10⁶) were hydrodynamically injected into C57BL/6 mice. Specifically, hepa1-6 cells (1 × 10⁶) were resuspended in 2 ml of PBS and delivered through the tail vein within 5 to 8 s. Three weeks postinjection, mice were sacrificed, and the tumor nodules in the liver were counted. In the other orthotopic HCC model, C57BL/6 mice were injected with MC plasmids expressing IL-33 or vector control by HGT. Seven days later, mice were anesthetized, and an incision below the thoracic diaphragm was made to expose the liver. Hepa1-6 cells (1 × 10⁶) were resuspended in 25 μl of PBS and then slowly injected into the hepatic lobule by an insulin syringe. Fifteen seconds gentle press with a cotton swab was necessary to avoid bleeding and leakage of cells. The incision was sutured with 5-0 suture lines. The mice were returned to the housing facilities after being revived. Two weeks after tumor injection, mice were sacrificed, and the tumor volumes were measured. In a diethylmitrosamine (DEN)-induced HCC model, 14-d-old male C57BL/6 mice were injected i.p. with 25 mg/kg DEN (Sigma-Aldrich, Los Angeles, CA). Mice were injected with MC plasmids expressing IL-33 or vector control every other month by HGT. Eight months later, mice were sacrificed, and livers were removed. Tumor nodules and the maximum tumor sizes were assessed.

Splenocytes and intrahepatic leukocytes were obtained from tumor-bearing mice and analyzed by flow cytometry. The fluorescence Abs used for FACS staining, including Abs against mouse CD3-PE-CF594, lineage-allophycocyanin, Sca-1–PE-CF594, CD8-Pacific Blue, CD4-PE, NK1.1-Pacific Blue, and F4/80-Pacific Blue were purchased from BD Biosciences. CD107a-FITC, CD27-PE, KLRG1-allophycocyanin, CD127-PE, ST2-FITC, CD117-PE-Cy7, CD69-FITC, CD86-FITC, CD11c-PE-Cy7, Gr1-PE, CD11b-PE-Cy7, CD19-allophycocyanin, CD44-PE-Cy7, CD4-FITC, CD62L-allophycocyanin-Cy7, NKG2D-FITC, IFN-γ–PE, IL-9–allophycocyanin, granzyme B–PE-Cy7, perforin-allophycocyanin, and TNF-α–PE-Cy7 were purchased from BioLegend (San Diego, CA). Anti-mouse CD16/32 FcR block Ab was also purchased from BioLegend. Flow cytometric analyses were performed using a FACSCanto II flow cytometer (BD Biosciences) and NovoCyte (ACEA Biosciences, San Diego, CA) and analyzed by FlowJo software (Tree Star, Ashland, OR).

Serum cytokine levels were assessed by BD Cytometric Bead Array (CBA) Mouse Inflammation Kit (BD Pharmingen, San Diego, CA) on a NovoCyte Flow Cytometer (ACEA Biosciences) and analyzed by FCAP Array Software (BD Biosciences). The flex sets include mouse IL-10, IL-6, MCP-1, IL-12p70, TNF, and IFN-γ.

Blood biochemical parameters, including alanine aminotransferase, aspartate aminotransferase, albumin, total bilirubin, alkaline phosphatase, uric acid, triglyceride, cholesterol, and total protein were analyzed by automatic biochemical analyzer (HITACHI, Japan).

CTL activity was detected by cytotoxicity detection kit (LDH) (Roche, Basel, Switzerland). Splenocytes were obtained from the tumor-bearing mice 7 d after tumor implantation and stimulated with hepa1-6 cell lysates and rhIL-2 (100 U/ml; SL PHARM, Beijing, China) for 4 d. After 4-d stimulation, splenocytes were cocultured with hepa1-6 cells at different concentration E:T ratios. CTL activity was assessed according to the manufacturer’s protocol after a 4-h cytotoxicity assay. The percent killing at each E:T ratio was calculated using the following formula: (A490nm[experimental]−A490nm[effector spontaneous] A490nm[target spontaneous]) × 100/(A490nm[target maximum]−A490nm[target spontaneous]).

Naive T cells were sorted from splenocytes of tumor-bearing mice by Mouse Pan-Naive T Cell Isolation Kit according to the manufacturer’s protocol (StemCell Technologies, Vancouver, Canada). Plates were coated with 1 μg/ml anti-CD3 and 0.2 μg/ml anti-CD28 Abs (BioLegend) overnight. A total of 1 × 10⁵ naive T cells were cultured for 48 h alone or with 100 ng/ml rIL-33 protein (BD Pharmingen). T cells were then analyzed by flow cytometry to determine the activation status of total T cells, CD4⁺ T cells, and CD8⁺ T cells.

To deplete CD4⁺ and CD8⁺ T cells in vivo, the mice were injected i.p. with anti-CD8α and anti-CD4 mAb (200 μg/mouse; Bio X Cell, West Lebanon, NH) once a week 1 d before tumor implantation. Three weeks later, the efficiency of the depletion was determined by flow cytometry using NovoCyte (ACEA Biosciences).

Student t test (unpaired, two-tailed) and one-way ANOVA analysis with GraphPad Prism 5 software (GraphPad Software, San Diego, CA) were used for statistical analyses. Data were shown as mean ± SD. A p value < 0.05 was considered statistically significant. The significance levels are marked as *p < 0.05, ** p < 0.01, and *** p < 0.001.

To generate an IL-33–rich tumor microenvironment, we generated hepa1-6 cell line stably expressing IL-33 in a secreted form. The expression of IL-33 was confirmed by Western blot (Fig. 1A). Then we examined the effect of IL-33 expression on the proliferation and apoptosis of hepa1-6 cells (Fig. 1B, 1C). IL-33 expression slightly reduced the proliferation of hepa1-6 cells (Fig. 1B) but had no effect on the apoptosis (Fig. 1C). Other than tumoral expression, IL-33 release in the liver was also achieved by hydrodynamically injecting MC IL-33 plasmids. IL-33 expression in the liver was confirmed by immunohistochemistry 7 d after plasmid injection (Fig. 1D).

To examine the role of IL-33 release during HCC development, we established murine orthotopic HCC model by surgically implanting hepa1-6 cells. IL-33 release in the liver was achieved by hydrodynamically injecting MC IL-33 plasmids (Fig. 2A). Two weeks after tumor implantation, IL-33 release in the liver significantly reduced both liver weights and tumor volumes, suggesting IL-33 expression in the liver can inhibit HCC tumor growth. Tumoral IL-33 release was then performed by injecting hepa1-6 cells stably expressing secreted IL-33 or vector control hydrodynamically (Fig. 2B). Three weeks after tumor cell injection, livers were weighed, and tumor nodules were counted. Tumoral release of IL-33 also significantly inhibited tumor growth. To further examine the role of IL-33 release during HCC development, we established the DEN-induced HCC model (Fig. 2C). IL-33 MC plasmids were injected hydrodynamically every other month during the 8-mo period of tumor development. The results showed that the number of tumor nodes was significantly reduced in the IL-33 group, and the maximal tumor volume was decreased in comparison with the control mice.

Because released IL-33 exerts its function through binding to its receptor, ST2, ST2^−/− mice were used to examine whether the effect of the released IL-33 on tumor development was dependent on ST2 (Fig. 2D). H22 cells (5 × 10⁶) were injected s.c. into ST2^−/− or wild type (WT) BALB/c mice. IL-33 was expressed by hydrodynamic injection of MC IL-33 plasmids. Similar as shown in the orthotopic models, IL-33 release inhibited tumor growth in WT mice. However, the antitumor effect of IL-33 was diminished in ST2^−/− mice (Fig. 2D). Then, we analyzed the association between IL-33 expression and patient survival in HCC patients using Kaplan–Meier plotter database (46) (Fig. 2E). The results showed that HCC patients with high IL-33 expression had prolonged survival compared with the patients with low IL-33 expression. Collectively, these results demonstrated that IL-33 release could inhibit HCC tumor development in an ST2-dependent manner.

To investigate the antitumor mechanism of the released IL-33, we examined the immune cell phenotypes from spleen and liver in the orthotopic HCC model with hepa1-6 cells and IL-33 expression through hydrodynamically injecting MC IL-33 plasmids or control vectors (Fig. 3). Immune cell infiltration significantly increased in both spleen and liver 7 d after tumor implantation as shown by H&E staining (Supplemental Fig. 1A) and flow cytometry (Supplemental Fig. 1B). There was an increase of CD8⁺ T cell percentages and a decrease of CD4⁺ T cell percentages in both spleen and liver (Supplemental Fig. 1C, 1D). Both percentage and number of CD69⁺CD4⁺ and CD69⁺CD8⁺ T cells in liver and spleen were remarkably upregulated because of IL-33 expression (Fig. 3A). The percentage and number of effector (CD44⁺CD62L⁻) CD4⁺ and CD8⁺ T cells were also increased in spleen and liver in the IL-33 group (Fig. 3B). Meanwhile, the percentages of naive (CD44⁻CD62L⁺) CD4⁺ and CD8⁺ T cells in spleen and liver were significantly decreased when IL-33 was expressed (Fig. 3B). Other markers of activation, including CD27 and KLGR1, were not affected by IL-33 expression in both spleen and liver (Supplemental Fig. 1F). Taken together, these results demonstrated that IL-33 release promoted the activation of CD4⁺ and CD8⁺ T cells in vivo.

Next, we assessed the IFN-γ and TNF-α secretion by CD4⁺ and CD8⁺ T cells to further confirm the role of IL-33 in promoting T cell responses. Although the percentages of TNF-α–producing CD4⁺ and CD8⁺ T cells were decreased in spleen and liver, the numbers were increased in liver by IL-33 expression (Fig. 4A, 4B). Moreover, both percentage and number of IFN-γ–producing CD8⁺ T cells were markedly elevated in spleen and liver in the IL-33 group compared with the control group. The number of splenic and hepatic CD4⁺ IFN-γ⁺ T cells was also upregulated by IL-33 expression. Serum levels of IL-10, IL-6, IFN-γ, MCP-1, TNF, and IL-12p70 were determined by CBA assays. Serum IFN-γ and IL-12p70 levels were elevated by IL-33 expression, suggesting an induced type 1 immune response (Fig. 4C). In the meantime, the levels of alanine aminotransferase, aspartate aminotransferase, albumin, total bilirubin, alkaline phosphatase, uric acid, and triglyceride were not affected by IL-33 expression, with a slight decrease of cholesterol and a slight increase of total protein, suggesting minor toxicity induced by IL-33 expression (Supplemental Fig. 2). Moreover, splenocytes isolated from tumor-bearing mice with IL-33 expression had increased killing capacity against hepa1-6 cells compared with those from the control group (Fig. 4D). These results showed that IL-33 release promoted antitumor T cell responses. To determine the killing mechanism, the expressions of CD107a, granzyme B, and perforin by CD8⁺ T cells were examined. They showed no difference between the T cells of IL-33 group and those of the control group (Supplemental Fig. 1G), suggesting the upregulated killing capacity was not mediated by perforin-dependent mechanism.

Previous studies have shown that CD8⁺ T cells express low levels of ST2, and loss of either IL-33 or ST2 impairs CD8⁺ T cell response to LCMV infection (47). IL-33/ST2 signaling has also been shown to enhance CD8⁺ T cell antitumor activity (27). To investigate whether released IL-33 can directly promote T cell activation, we stimulated naive T cells from the spleen with anti-CD3/anti-CD28 in vitro and analyzed the subsets and activation status of T cells in the presence or absence of rIL-33. As shown in Fig. 5, rIL-33 significantly increased the percentage of CD8⁺ T cells (Fig. 5A), which was consistent with our findings in vivo (Supplemental Fig. 1C, 1D). The percentages of both activated (CD69⁺) and effector (CD44⁺CD62L⁻) CD4⁺ and CD8⁺ T cells were substantially increased by rIL-33 (Fig. 5B). Meanwhile, the percentages of naive (CD44⁻CD62L⁺) CD4⁺ and CD8⁺ T cells were decreased by rIL-33 treatment (Fig. 5B). Next, we detected the IFN-γ and TNF-α production by CD4⁺ and CD8⁺ T cells with or without rIL-33 in vitro. The percentages of both IFN-γ– and TNF-α–producing CD4⁺ and CD8⁺ T cells were significantly increased in the presence of rIL-33 (Fig. 5C). These results confirmed that IL-33 could directly enhance T cell activation in vitro.

IL-33 has been shown to promote the function of both NK and ILC2 cells (28, 29, 48), and we observed an increase in both the percentage and number of ILC2 cells in IL-33–expressing tumor-bearing mice (Supplemental Fig. 3A). However, coculture of ILC2 with naive T cells had no effect on T cell activation (Supplemental Fig. 3B). The percentages of NK cells and NKG2D⁺NK cells were decreased in the liver of IL-33–expressing tumor-bearing mice (Supplemental Fig. 1H, 1I). Moreover, the IFN-γ and TNF-α production in NK cells showed no difference between IL-33–expressing and control mice (Supplemental Fig. 1J). The cytotoxic molecules, including CD107a, granzyme B, and perforin expressed by NK cells were not affected by IL-33 either (Supplemental Fig. 1K). Therefore, both NK and ILC2 cells may not be the mediator of antitumor function of IL-33.

Because we have shown that IL-33 can directly enhance T cell activation, we then performed depletion experiments to determine whether CD4⁺ and/or CD8⁺ T cells mediate the antitumor effect of released IL-33. CD8⁺ T cells have cytotoxicity against tumor cells and have been shown to be crucial in IL-33–mediated modification of the tumor microenvironment (27). We depleted CD8⁺ T cells in tumor-bearing mice expressing IL-33 or control vector (Fig. 6A, 6B). CD8⁺ T cell depletion had no significant effect on the antitumor activity of IL-33. CD4⁺ T cell depletion showed a similar result (data not shown). However, when both CD4⁺ and CD8⁺ T cells were depleted in the orthotopic HCC model, the antitumor effect of IL-33 expression diminished (Fig. 6A, 6B). These results demonstrated that IL-33 exerted its antitumor effect through the activation of both CD4⁺ and CD8⁺ T cells.

Serum IL-33 levels are elevated in patients with chronic liver failure, acute hepatitis, and HCC (34, 49). However, the function of released IL-33 in HCC development is not known. In the current study, we demonstrated that released IL-33 could inhibit tumor growth in murine HCC models in an ST2-dependent manner. IL-33 release could promote antitumor T cell responses in vivo. In vitro experiments showed that rIL-33 could directly enhance T cell activation and cytokine production. Further depletion experiments demonstrated that the antitumor effect of released IL-33 was mediated by both CD4⁺ and CD8⁺ T cells.

Because IL-33 functions as an alarmin during stress and cell death (50), IL-33 release is clearly increased during chronic inflammation. It is critical to understand whether this increased IL-33 release has pro- or antitumor function. Released IL-33 exhibited antitumor activity in two murine orthotopic HCC models (Fig. 2), suggesting IL-33 released from both tumor cells and hepatocytes could inhibit tumor development in HCC. We also addressed this question in a DEN-induced HCC model by increasing IL-33 release in the liver. IL-33 significantly prevented tumor development in this model involving progression from chronic inflammation to HCC. Thus, in all the HCC models we have examined, IL-33 inhibited HCC tumor development.

ST2 is primarily expressed on mast cells and Th2 cells, so IL-33 was initially thought to induce type 2 immune responses as well as secretion of associated cytokines, such as IL-4, IL-5, and IL-13 (14, 19, 51). Recent studies have found that ST2 can also be expressed on CD8⁺ T cells, Th1 cells, NK, and NKT cells, suggesting a role for IL-33/ST2 axis in innate and type 1 immune responses (24–26, 52). Many studies have shown that IL-33 can promote type 1 immune responses by enhancing IFN-γ production by Th1, NK, and NKT cells (24, 25, 53). It has been shown that tumoral expression of IL-33 increases IFN-γ production by CD8⁺ T cells and NK cells in tumor tissue in two s.c. murine tumor models (27). In our HCC models, IL-33 release significantly elevated IFN-γ production by both CD4⁺ T cells and CD8⁺ T cells. However, NK cells did not have increased expression of the activation markers or IFN-γ (Supplemental Fig. 1I, 1J). Therefore, IL-33 may induce a Th1- and CTL-dominated type 1 immune response in the liver, whereas NK cells could be more critical in IL-33–induced antitumor immune response in the skin. Ramadan et al. (54) demonstrated that Th9 and Tc9 cells activated by IL-33 exhibited a strong antitumor ability. However, neither Th9 nor Tc9 increased with IL-33 expression in our HCC models, suggesting this could be organ-specific phenomena (Supplemental Fig. 1E).

IL-33 has been shown to alleviate Con A–induced hepatitis by increasing total number of liver CD4⁺Foxp3⁺ cells and IL-4–producing CD4⁺ T cells (55). It can also expand suppressive CD11b⁺ Gr-1^(int) cells and Treg to promote cardiac allograft survival (56). In fact, in our HCC models, both CD11b⁺ myeloid cells and CD4⁺Foxp3⁺ Treg increased in percentages and numbers in spleen and liver of IL-33–expressing tumor-bearing mice (Supplemental Fig. 1L, 1M). These cells could compromise and be targeted to enhance IL-33–mediated antitumor immune response. Treg depletion and IL-33 may work synergistically for tumor therapy (27), which could be further investigated in HCC models.

Many studies have demonstrated that IL-33 can stimulate ILC2s to produce large amounts of IL-5 and IL-13 (57–60). A recent study has suggested that ILC2s activate Th2 cells to initiate type 2 adaptive immune responses (61). IL-33–induced ILC2s can amplify inflammatory immune responses during immune-mediated hepatitis (58). It is also reported that ILC2s play a fundamental role in enhancing anticancer immunity and controlling tumor metastasis (62). We observed significant increase of ILC2s in liver of the IL-33 group during HCC development (Supplemental Fig. 3A). We then sorted lin⁻CD127⁺c-Kit⁺Sca-1⁺ST2⁺ ILC2s and cocultured them with naive T cells. The results showed that ILC2s had no effect on CD4⁺ and CD8⁺ T cell activations (Supplemental Fig. 3B). Therefore, ILC2s may not be the mediator of IL-33–induced immune activation. Further studies are still needed to confirm the role of ILC2s in IL-33–induced antitumor immune response in vivo when the appropriate research tools are available.

The inflammatory cytokines functioning as alarmins are critical in establishing the organ microenvironment. Understanding their roles in tumor development may provide therapeutic strategies for early intervention. Meanwhile, both checkpoint blockade-based immunotherapy and adoptive T cell therapy, including the chimeric Ag receptor T cell therapy, require sufficient immune stimulus from the tumor microenvironment. Therefore, IL-33 released into the extracellular space could function as an adjuvant to enhance the efficacy of immunotherapy for HCC.

We thank Dr. Andrew N. J. McKenzie (Medical Research Council Laboratory of Molecular Biology, Cambridge, U.K.) and Dr. Qingqing Wang (Zhejiang University, Hangzhou, China) for generously providing ST2^−/− mice. We also thank Dr. Zhaoqing Fang (Shanghai University of Traditional Chinese Medicine, Xuhui, China) for providing H22 cells.

The authors have no financial conflicts of interest.