And NK-Cell Activation

Home , IL17A, IL1A

Published OnlineFirst March 29, 2016; DOI: 10.1158/0008-5472.CAN-15-2658 Cancer Microenvironment and Immunology Research

Membrane IL1a Inhibits the Development of Hepatocellular Carcinoma via Promoting T- and NK-cell Activation Dandan Lin1, Lei Lei1, Yonghao Liu1, Yinsheng Zhang1,BoHu1, Guangming Bao1, Yuan Song1, Ziqi Jin1, Chunliang Liu2, Yu Mei1,3, Dedy Sandikin3,YanWu1, Lixiang Zhao1, Xiao Yu1, and Haiyan Liu3

Abstract

Hepatocellular carcinoma is a worldwide health problem with membrane IL1a promoted T- and natural killer (NK)–cell acti- þ limited treatment options and poor prognosis. Inflammation vation in vivo. IFNg production by CD8 T and NK cells was also associated with liver injury and hepatocyte regeneration can lead increased as a result of membrane IL1a expression. Moreover, the to fibrosis, cirrhosis, and eventually, hepatocellular carcinoma. cytotoxicity of the CTL and NK cells was also enhanced by IL1a is one of the most important inflammatory cytokines membrane IL1a expression. Furthermore, in vitro studies dem- involved in inflammation and tumor development. IL1a presents onstrated that membrane IL1a could directly activate T cells and as multiple forms in vivo, including precursor, propiece, mem- NK cells in a cell contact–dependent manner. Conversely, deple- þ brane, and secreted forms, and their functions have been thought tion of both CD8 T and NK cells suppressed the antitumor to be different. The role of membrane IL1a in hepatocellular activity of membrane IL1a. Our studies demonstrated that mem- carcinoma tumorigenesis is still not clear. Here, we examined the brane IL1a could promote antitumor immune responses through functions of membrane IL1a in murine hepatocellular carcinoma activation of T and NK cells. Thus, our findings provide new models. We found that membrane IL1a potently inhibited hepa- insights of IL1a functions during hepatocellular carcinoma devel- tocellular carcinoma tumor growth. Further studies showed that opment. Cancer Res; 76(11); 1–10. 2016 AACR.

Introduction important roles in liver tumor initiation, progression, and metastasis (6). Numerous studies have demonstrated that chronic Hepatocellular carcinoma is the fifth most common liver inflammation can lead to carcinogenesis (7–9). Recently, many malignancy, and the third leading cause of cancer-related death studies focused on the inflammation in the tumor microenviron- worldwide (1). Hepatocellular carcinoma is caused by multiple ment, which can describe and predict the phenotypic character- risk factors, such as infection with hepatitis virus (HBV and HCV; istics of cancer (10). Hence, understanding the role of inflamma- ref. 2), nonviral causes (alcohol abuse, nonalcoholic steatohepa- tory components and its signaling pathways in the tumor micro- titis, type 2 diabetes mellitus; ref. 3), and environmental and environment could lead to the discovery of novel molecular dietary carcinogens (aflatoxin B1, nitrosamines; ref. 4). The therapeutic targets for hepatocellular carcinoma. molecular mechanisms of hepatocellular carcinoma develop- The IL1 cytokine family consists of two important agonistic ment are complex and not fully understood (5). Inflammation, subtypes, IL1a and IL1b, which are both synthesized as precursor tumor microenvironment, oxidative stress, and other events play proteins. Inactive IL1b precursor can be cleaved by caspase-1 into an active cytokine, whereas IL1a precursor (proIL1a)canbe 1Institute of Blood and Marrow Transplantation, Cyrus Tang Hematol- cleaved into mature form (secreted form) and N-terminal propiece ogy Center, Department of Hematology, Collaborative Innovation (ppIL1a) by calpain. IL1a can also be myristoylated and expressed Center of Hematology, Jiangsu Institute of Hematology, Key Labora- on the cell membrane possibly via a mannose-like receptor (11, tory of Thrombosis and Hemostasis, Ministry of Health, the First Affiliated Hospital of Soochow University, Suzhou, P.R. China. 2Depart- 12). Membrane IL1a is constitutively expressed in resting cells ment of Biochemistry and Molecular Biology, Medical College of under homeostatic conditions, whereas secreted IL1a cannot be Soochow University, Suzhou, P.R. China. 3Immunology Programme, detected in healthy human body fluids (13). Studies have sug- Life Sciences Institute and Department of Microbiology and Immu- nology, National University of Singapore, Singapore, Singapore. gested that membrane IL1a is immunostimulatory and may activate immune cells expressing IL1R1 including T cells and natural Note: Supplementary data for this article are available at Cancer Research killer (NK) cells (14–16). However, the role of membrane IL1a in Online (http://cancerres.aacrjournals.org/). inflammation and tumor development has not been well studied. D. Lin and L. Lei contributed equally to this article. Fibrosarcoma cells expressing IL1a intracellularly and on the Corresponding Author: Haiyan Liu, Immunology Programme, Life Sciences cell membrane lost their tumorigenicity (14, 16). On the contrary, Institute and Department of Microbiology and Immunology, National University in smooth muscle cells, intracellular IL1a was shown to stimulate of Singapore, Singapore 117456, Singapore, Phone: 656-516-6661; Fax: 656-778- cell proliferation (17). Besides, in pancreatic ductal adenocarci- 2684; Email: [email protected] noma, tumor-associated IL1a was identified as the initiator of doi: 10.1158/0008-5472.CAN-15-2658 tumor growth by promoting the production of inflammatory 2016 American Association for Cancer Research. factors (18). Overexpression of IL1a propiece in tumor cells

www.aacrjournals.org OF1

Lin et al.

induced malignant transformation and increased tumor invasive- Confocal microscopy ness to visceral organs (19). IL1a could induce expression of Hepa1-6–transfected cells were plated and observed in boro- proteases in cancer cells (20, 21) and adhesion molecules in silicate cover glass (Thermo Fisher Scientific). Cells were fixed with vascular endothelial cells (22, 23). In primary gastric carcinoma, 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, IL1a expression was correlated with a higher incidence of liver and stained with rabbit anti-mouse IL1a antibody (Abcam) and metastasis (24). In addition, secreted IL1a plays a key proin- PE goat anti-rabbit IgG secondary antibody (MultiSciences Bio- flammatory role in models of many chronic inflammatory con- tech). DAPI (Sigma) was used to detect nuclei. Microscopy was ditions (13, 25, 26). These reports suggest that IL1a plays an performed with a Nikon A1 confocal microscope spectral detector important role in tumor development. However, the functions of (Nikon). ppIL1a, membrane IL1a, and secreted IL1a cannot be distin- guished in most of these studies. Cell proliferation assay Overexpressing IL1a precursor should result in the overexpres- Cell proliferation was determined by using CCK-8 and Ki-67 sion of all forms of IL1a. In the current study, we mutated the assay. Hepa1-6–transduced cells (3,000 per well, 96-well plate) nuclear localization signal (NLS) and calpain cleavage site of were plated and subjected to CCK-8 assay after 48 hours. CCK-8 IL1a. Therefore, the IL1a precursor cannot either translocate to solution was added to each well and incubated for 4 hours. The the nucleus or be cleaved by calpain to be secreted. This meth- optical density value was detected at 450 nm in a microplate reader odology resulted in mainly membrane IL1a expressions, without (BioTek). For Ki-67 assay, hepa1-6–transduced cells (3 105 per the nuclear or secreted forms. We found that overexpression of well, 6-well plate) were seeded and incubated for 48 hours. Cells membrane IL1a significantly inhibited hepatocellular carcinoma were harvested for Ki-67 staining. tumor growth. Membrane IL1a greatly increased numbers of þ þ activated CD4 T, CD8 T, and NK cells, as well as IFNg produc- Murine hepatocellular carcinoma models þ tion and cytotoxicity in CD8 T and NK cells. Moreover, NK and We generated three murine hepatocellular carcinoma models. In T cells could be activated by membrane IL1a in a cell contact– the subcutaneous hepatocellular carcinoma tumor model, hepa1-6 þ dependent manner. Depletion of both CD8 T and NK cells cells (1 106) were injected subcutaneously into C57BL/6 mice in vivo suppressed the antitumor activity of membrane IL1a. and tumor growth was monitored every 3 days. Mice were sacrificed Taken together, our study demonstrated that membrane IL1a 3 weeks after tumor inoculation. In the orthotopic hepatocellular could inhibit hepatocellular carcinoma development by promot- carcinoma model, hepa1-6 cells (1 106) were injected into ing T- and NK-cell activation. C57BL/6 mice by hydrodynamic cell delivery method. In brief, cell suspension in 2 mL PBS was injected into the tail vein within 5 Materials and Methods to 8 seconds. Mice were sacrificed 3 weeks after tumor injection and the tumor nodules in the liver were counted. In a diethylmitrosa- Experimental animals mine (DEN)-induced hepatocellular carcinoma model, 14-day-old fi – Speci c pathogen-free male C57BL/6 mice (ages 6 8 week) C57BL/6 mice were injected with 25 mg/kg DEN (Sigma-Aldrich). were purchased from Shanghai Laboratory Animal Center After 8 months, mice were sacrificed, and their livers were removed fi (Shanghai, China). All mice were housed in speci c pathogen- and subjected to assessment of tumor growth. Mice were hydrody- free facilities and in accordance with the National Animal Care namically injected with the MC plasmids every month in the last and Use Committee. All animal studies were approved by the 4 months. Briefly, 50 mg/mouse DNA was diluted in 2.5 mL PBS Institutional Laboratory Animal Care and Use Committee of and injected into the tail vein using a 25-gauge needle and syringe Soochow University. within a time period of 5 to 8 seconds.

Plasmid construction Flow cytometry Membrane IL1a (wherein the NLS KK was mutated to RR, Splenocytes and intrahepatic leukocytes were harvested from and S was mutated to L at the calpain cleavage site) was mice and analyzed by flow cytometry (27). The antibodies used amplified from cDNA generated from lipopolysaccharide-stim- for FACS staining including anti-mouse CD3-PE/CF594, CD44- ulated mouse peripheral blood mononuclear cells (PBMC), PE, CD8-APC, CD4-APC/H7, CD69-FITC, CD86-FITC, CD11c- and then inserted into lentiviral plasmid and minicircle (MC) PE/Cy7, Gr1-PE, CD11b-APC, CD19-PE/Cy7, NK1.1-PerCP/ plasmid (pMC.EF1; SBI). MC plasmids were purified and Cy5.5, were purchased from BD Biosciences, and CD4-FITC, delivered in vivo by hydrodynamic gene transfer (HGT) tech- CD62L-PerCP/Cy5.5, NKG2D-FITC, F4/80-PerCP/Cy5.5 IFNg- nique for in vivo expression. PE, and TNFa-PE/Cy7 were purchased from Biolegend. Anti- mouse CD16/32 FcR block antibody was purchased from Biole- Cell culture gend. The rabbit anti-mouse IL1R1 antibody and the chromeo The murine hepatocellular carcinoma cell line hepa1-6 and 488 goat anti-rabbit IgG secondary antibody were purchased from human embryonic kidney cell line293Twereobtainedfrom Abcam. Flow cytometric analyses were performed using a FACS- the ATCC and cultured in 10% FBS DMEM (Gibco). And the Canto II flow cytometer (BD Biosciences) and analyzed using human 293T cell line was recently authenticated by short FlowJo software (Tree Star). tandem repeat profiling analysis. The membrane IL1a-expressing lentivirus generated from 293T cells was used to infect Cytokine analysis hepa1-6 cells. Empty vector was used as vector control. The Serum cytokine levels were analyzed using BD Cytometric Bead YFP-positive monoclonal cells were sorted by FACSAria III flow Array (CBA) Mouse Soluble Protein Master Buffer Kit and CBA cytometer (BD Biosciences) into 96-well plate and then mouse soluble protein flex set (BD Pharmingen) on a FACSCanto screened for stable clones. II Cytometer (BD Biosciences) and analyzed by BD FCAP Array

OF2 Cancer Res; 76(11) June 1, 2016 Cancer Research

Membrane IL1a Inhibits HCC Development

Figure 1. Membrane IL1a expression in hepatocellular carcinoma cells. A, membrane IL1a construct is a mutant of IL1a precursor in which NLS sequence KK (85–86 aa) was mutated to RR, and S (115aa) was mutated to L at the calpain cleavage site. B, IL1a expression (red) was detected in hepa1-6–transfected cells (green). DAPI staining (blue color) was used to detect nuclei. Scale bar, 50 mm. C, cell viability was determined by CCK-8 assay. D, the proliferation of hepa1-6–transfected cells was measured by Ki-67 assay. The data shown are representative of three experiments. , P < 0.01; , P < 0.001.

software (BD Biosciences). The ﬂex sets including mouse IL1a, 4-hour cytotoxicity assay using YAC-1 and hepa1-6 cells as target IL1b, IL4, IL6, IL17A, TNF, and IFNg were used for cytokine cells, respectively. þ detection. Effector CD8 T cells and NK cells were added to target cells in different effector:target (E:T) ratio and incubated for 4 hours, Cytotoxicity assay respectively. The cytotoxic activity was determined by the CytoTox Splenocytes were harvested from the tumor-bearing mice 7 96 nonradioactive cytotoxicity assay (Promega). days after tumor inoculation and stimulated with hepa1-6 cell þ lysates with rhIL2 (100 U/mL) for 4 days. CD8 T cells were T-cell activation assay isolated from the cultured splenocytes by negative selection using Na€ve T cells were sorted from splenocytes by FACSAria III Flow immunomagnetic beads according to the manufacturer's protocol Cytometer (BD Biosciences). Plates were coated with 2 mg/mL (StemCell Technologies). CTL activity was assayed in a 4-hour anti-CD3 and 0.4 mg/mL anti-CD28 antibodies (Biolegend) over- cytotoxicity assay using hepa1-6 cells as target cells. night before treatment. T cells and the transduced hepa1-6 cells Splenocytes were harvested from the tumor-bearing mice 7 were cocultured for 48 hours. In coculture system, 2 105 T cells days after tumor inoculation and NK cells were isolated from and 1 105 hepa1-6 cells (irradiated at 50 Gy) were seeded per splenocytes by negative selection using immunomagnetic beads well in 96-well plate. In the transwell system (24-well), 1 105 (StemCell Technologies). NK cells killing activity was assayed in a hepa1-6 cells were placed in the top chamber and 4 105 T cells

www.aacrjournals.org Cancer Res; 76(11) June 1, 2016 OF3

Lin et al.

were seeded in the bottom chamber. After incubation, T cells were analyzed by ﬂow cytometry to determine the activation status of þ þ total T cells, CD8 T, and CD4 T cells.

NK-cell activation assay NK cells were prepared as described previously (27). NK cells and the transduced hepa1-6 cells were cocultured for 24 hours. In coculture system, 2 105 NK cells and 1 105 hepa1-6 cells (irradiated at 50 Gy) were seeded per well in 96-well plate. In the transwell system (24-well), 1 105 hepa1-6 cells were placed in the top chamber and 4 105 NK cells were seeded in the bottom chamber. After incubation, NK cells were analyzed by ﬂow cytometry to determine their activation status.

þ Depletion of CD8 T/NK cells þ To deplete CD8 T and NK cells, the mice were injected intraperitoneally with anti-CD8a mAb (150 mg/mouse) and anti-NK1.1 (CD161) mAb (150 mg/mouse) once every week. þ Depletions of CD8 T and/or NK cells were conﬁrmed by ﬂow cytometric analysis.

Statistical analysis Statistical analyses were performed with Student t test (unpaired, two-tailed) with GraphPad Prism 5 software (Graph- Pad). Data were reported as mean SD. P value less than 0.05 was considered as statistically signiﬁcant.

Results Membrane IL1a expression inhepatocellular carcinoma hepa1-6 cells To study the role of membrane IL1a, we constructed the membrane IL1a expression plasmid in which the NLS and calpain cleavage sites were mutated (Fig. 1A). Therefore, the precursor IL1a could only be expressed on the cell surface as the membrane form without being able to translocalize into the nuclei or be secreted as the mature form. Membrane IL1a expression was detected and conﬁrmed by confocal microscopy (Fig. 1B). To assess the effect of membrane IL1a overexpression on tumor cells, we examined the proliferation of the hepa1-6 cells stably expressing membrane IL1a. As shown in Fig. 1C and D, cell proliferation was slightly promoted by membrane IL1a expression as determined by CCK-8 assay (Fig. 1C) and Ki-67 staining (Fig. 1D).

Membrane IL1a exerts an antitumor effect in murine hepatocellular carcinoma models To assess the role of membrane IL1a in vivo, we established Figure 2. murine hepatocellular carcinoma models by subcutaneously or Membrane IL1a exerts an antitumor effect in murine hepatocellular hydrodynamically injecting hepa1-6 cells stably expressing mem- carcinoma models. A, hepa1-6 cells expressing membrane IL1a 6 n ¼ brane IL1a or vector control in C57BL/6 mice. In both subcuta- (1 10 ) were injected into mice subcutaneously ( 6 each group). Data shown are representative of tumor morphology, tumor sizes, and tumor neous and orthotopic hepatocellular carcinoma models, tumor weight.B,mice(n ¼ 6 each group) were injected with hepa1-6 cells growth was greatly reduced by membrane IL1a expression (Fig. expressing membrane IL1a by hydrodynamic cell delivery method. Data 2A and B). To further determine the role of membrane IL1a in shown are representative of tumor morphology, the numbers of tumor development, we hydrodynamically injected the mem- tumor nodules, and liver weight. C, mice (n ¼ 6eachgroup)were brane IL1a MC plasmids to overexpress membrane IL1a in administered a single injection of DEN on day 14 after birth. After 4 hepatocytes in a DEN-induced hepatocellular carcinoma model. months, mice were injected hydrodynamically with 50 mg MC plasmids ﬁ every 2 months for 4 months. Data shown are representative of We previously con rmed membrane IL1a expression on the tumor morphology, the numbers of tumor nodules, maximal tumor surface of hepatocytes after hydrodynamic injection and the size, and liver weight. The data shown are representative of three expressed membrane IL1a could not be secreted nor detected in experiments. Values are presented as means SD. , P < 0.05; the peripheral blood (28). The results showed that tumor growth , P < 0.01; , P < 0.001.

OF4 Cancer Res; 76(11) June 1, 2016 Cancer Research

Membrane IL1a Inhibits HCC Development

Figure 3. Membrane IL1a promotes T-cell and NK-cell activation in vivo. Splenocytes, intrahepatic leukocytes, and serum from tumor-bearing mice of an orthotopic hepatocellular carcinoma model were collected 7 days after tumor injection. Splenocytes and intrahepatic leukocytes were isolated for FACS analysis. A, the percent and number of activated (CD69þ), na€ve, effector, and memory CD4þ T cells or CD8þ T cells in spleens are shown. B, the percent and number of activated (CD69þ), na€ve, effector, and memory of CD4þ T cells or CD8þ T cells in livers are shown. The percent and number of NK cells and activated NK cells in spleens (C) and livers (D) are shown. The percent and number of DC and macrophages in spleens (E) and livers (F) are shown. The data shown are the representative of three experiments. Values are presented as means SD. , P < 0.05; , P < 0.01; , P < 0.001.

in mice with membrane IL1a overexpression occurred at a much membrane IL1a was overexpressed (Fig. 3A and B). The percen- þ þ slower rate than in control mice (Fig. 2C). These data demon- tages and numbers of effector CD4 and CD8 T cells were strated that membrane IL1a inhibited hepatocellular carcinoma upregulated in both spleens and livers, whereas the percentages þ þ development in murine hepatocellular carcinoma models. and numbers of na€ve CD4 and CD8 T cells were decreased in membrane IL1a group (Fig. 3A and B). These results suggested Membrane IL1a promotes T-cell and NK-cell activation in vivo that overexpression of membrane IL1a promoted the activation þ þ To understand the mechanisms of antitumor effect of mem- of CD4 and CD8 T cells. brane IL1a, we analyzed the percentage and number of different ThepercentageandnumberofNKcellswerealsohigherin immune cell subsets and their activation status in orthotopic both spleens and livers in the membrane IL1a group (Fig. 3C tumor-bearing mice on day 7 after tumor cell inoculation by flow and D). While the percentage of activated NK cells out of total þ þ cytometry. The percentages and numbers of CD4 and CD8 T NK cells was significantly decreased in the spleens, the numbers cells in spleens and livers were not significantly affected by of activated NK cells were increased in spleens and livers due to membrane IL1a expression (data not shown). However, the membrane IL1a expression (Fig. 3C and D). These data dem- þ þ percentage of CD69 CD4 T cells was increased in both spleens onstrated that NK-cell activation was also enhanced by mem- and livers, with significant increase in numbers in the livers when brane IL1a expression. In addition, the percentages of DCs and

www.aacrjournals.org Cancer Res; 76(11) June 1, 2016 OF5

Lin et al.

Figure 4. þ Membrane IL1a promotes IFNg production in CD8 T and NK cells. Splenocytes and intrahepatic leukocytes were stimulated and ﬁxed for FACS intracellular analysis. þ þ þ A, the percent and number of IFNg-producing CD4 T, CD8 T, and NK cells in spleens are shown. B, the percent and number of IFNg-producing CD4 T, þ þ CD8 T, and NK cells in livers are shown. C, serum levels of IL1a,IL1b, IL4, IL6, IL17A, TNF, and IFNg are shown. D, cytotoxic activity of CD8 T cells and NK cells sorted from the spleens of tumor-bearing mice were analyzed by nonradioactive cytotoxicity assay. Values are presented as means SD. The data shown are representative of three experiments. Values are presented as means SD. , P < 0.05; , P < 0.05.

macrophages were increased in the spleens and their numbers and NK cells isolated from the tumor-bearing mice overexpressing were increased in the livers with membrane IL1a overexpres- membrane IL1a had increased killing capacity compared with sion (Fig. 3E and F). those from control mice (Fig. 4D). While CTLs only exhibited Next, we examined TNFa and IFNg production by T and NK cytotoxicity to hepa1-6 cells, NK cells exhibited killing capacity þ cells. The percentages and numbers of CD8 T and NK cells against both hepa1-6 and YAC-1 targets. These data suggested that producing IFNg were significantly increased in the spleens of the membrane IL1a overexpression promoted antitumor immune tumor-bearing mice overexpressing membrane IL1a (Fig. 4A). responses through T- and NK-cell activation. þ However, TNFa and IFNg production by CD8 T and NK cells in the livers were not significantly affected by membrane IL1a Membrane IL1a directly promotes T- and NK-cell expression (Fig. 4B). Serum IL1a, IL1b, IL4, IL6, IL17A, TNF, and activation in vitro IFNg levels were determined by CBA soluble protein flex sets (Fig. Previous studies have demonstrated that T cells and NK cells 4C). Serum TNF level was elevated, whereas IL6 level was express IL1R1 (29–31). We observed the same results that about decreased because of membrane IL1a expression. Moreover, CTLs 70% of the NK cells and over 50% of the T cells in the liver of the

OF6 Cancer Res; 76(11) June 1, 2016 Cancer Research

Membrane IL1a Inhibits HCC Development

Figure 5. Membrane IL1a directly activates NK and T cells in vitro. A, NK cells were cocultured with irradiated hepa1-6 cells expressing membrane IL1a and the activation status was analyzed by flow cytometry. B, NK cells were cocultured with hepa1-6 cells expressing membrane IL1a in separate chambers of transwell plate, and the activation status was analyzed by flow cytometry. C, na€ve T cells sorted from splenocytes were cocultured with irradiated hepa1-6 cells expressing membrane IL1a in a plate coated with anti-CD3 and anti-CD28. þ Activated (CD69 ), na€ve, effector, and þ þ memory CD4 T cells and CD8 T cells were analyzed by flow cytometry. D, na€ve T cells sorted from splenocytes were cultured with hepa1-6 cells expressing membrane IL1a in separate chambers of transwell plate. þ Activated (CD69 ), na€ve, effector, and þ þ memory CD4 T cells and CD8 T cells were analyzed by flow cytometry. The data shown are representative of three experiments. Values are presented as means SD. , P < 0.05; , P < 0.01; , P < 0.001.

tumor-bearing mice expressed IL-1R1 (Supplemental Fig. S1A). CD69 expressions in the presence of membrane IL1a (Supple- We also found that IL1R1 was expressed on hepa1-6 cells (Sup- mentary Fig. S1C). To further demonstrate whether the promo- plemental Fig. S1B). To investigate whether membrane IL1a tion of T- and NK-cell activation by membrane IL1a required cell– could directly promote T- and NK-cell activation, we stimulated cell contact, tumor cells and na€ve T cells or NK cells were cultured T and NK cells with anti-CD3/anti-CD28 or IL2 and analyzed the in different chambers of transwell plates (Fig. 5B and D). The activation status of T and NK cells cocultured with hepa1-6 cells T-cell and NK-cell activation status was not affected by the expressing membrane IL1a or control vector. Flow cytometry membrane IL1a expression, suggesting that the membrane IL1a analysis revealed that the percentage of activated NK cells was promoted T- and NK-cell activation in a cell contact-dependent elevated by membrane IL1a expression (Fig. 5A), suggesting that manner. membrane IL1a could directly activate NK cells. The percentages þ þ þ þ of both activated (CD69 ) and effector (CD44 CD62L ) CD8 T Depletion both of CD8 T and NK cells diminished the þ and CD4 T cells were also significantly increased by membrane antitumor effect of membrane IL1a IL1a expression (Fig. 5C). Meanwhile, the percentages of na€ve To further evaluate the mechanisms of the antitumor activity of þ þ þ CD8 T and CD4 T cells were decreased. To examine whether the membrane IL1a, we depleted the CD8 T and/or NK cells in tumor-specific T-cell response could also be promoted by mem- orthotopic tumor-bearing mice in vivo. The depletions were con- brane IL1a, we stimulated na€ve T cells with DCs pulsed with firmed by flow cytometry (Fig. 6A), and our results showed that þ hepa1-6 lysates in the presence of irradiated hepa1-6 cells expres- depletions of either CD8 T or NK cells did not affect the sing membrane IL1a or the control cells. The results showed that antitumor activity of membrane IL1a. However, the antitumor þ þ þ both CD4 and CD8 T cells were better activated in terms of effect of membrane IL1a was totally diminished when both CD8

www.aacrjournals.org Cancer Res; 76(11) June 1, 2016 OF7

Lin et al.

Figure 6. Depletion of both CD8þ T and NK cells diminished the antitumor effect of membrane IL1a. Mice (n ¼ 5–7each group) were injected with hepa1-6 cells expressing membrane IL1a by hydrodynamic cell delivery method and administrated with anti-CD8a and/or anti-NK1.1 mAbs intraperitoneally. Splenocytes and intrahepatic leukocytes from tumor- bearing mice of orthotopic hepatocellular carcinoma model were collected 3 weeks after tumor inoculation. Splenocytes and intrahepatic leukocytes were isolated for FACS analysis. A, the percent of þ CD8 T and NK cells in spleens and livers are shown. B, data shown are the numbers of tumor nodules and liver weight. The data shown are representative of three experiments. Values are presented as means SD. , P < 0.05; , P < 0.01; , P < 0.001.

T and NK cells were depleted (Fig. 6B). These results demonstrated hydrodynamically injecting the stable cell line. As hepatocellular þ that membrane IL1a exerted its antitumor effect through CD8 T- carcinoma tumor cells expressing membrane IL1a exhibited or NK-cell activation. accelerated proliferation, the tumor-inhibiting effect of membrane IL1a was not due to cell-intrinsic mechanism. Furthermore, Discussion we demonstrated that membrane IL1a exerted antitumor effects by promoting T- and NK-cell activation, which required cell–cell IL1a can be myristoylated and expressed on the membranes of contact. various cells including monocytes and macrophages, by a lectin- In the tumor microenvironment, cytokines can modulate like association (11, 12, 32). Hepatocytes can also express IL1a antitumor responses, yet they can also stimulate angiogenesis, (15). Cell-associated IL1a expression could be detected in the tumor progression, and metastasis, and maintain tumor-pro- liver samples of hepatocellular carcinoma patients, both in the moting inflammation (34, 35). IL1a has been shown to be tumor and adjacent normal tissues by confocal microscopy (Sup- involved in inflammation and tumor development (36–38). plementary Fig. S2). Although the membrane localization could Previously, IL1a was found to stimulate cell growth of several not be confirmed with the human samples, previous studies have types of malignant cells in some cases without being secreted by shown that approximately 10% to 15% of translated IL1a is the cells (13). Another study also demonstrated that IL1a myristoylated and anchored to the cell membrane (13, 33). released from necrotic hepatocytes promoted the development However, how the expression of membrane IL1a is regulated of hepatocellular carcinoma (15). Interestingly, membrane and its role during the hepatocellular carcinoma development are IL1a is thought to be the major source of the secreted IL1a still not known and need further investigation. In the current from necrotic hepatocytes during chronic inflammation and study, we provided evidence that membrane IL1a inhibited tumor development. However, overexpression of membrane tumor development in murine hepatocellular carcinoma models. IL1a in hepatocytes in a DEN-induced hepatocellular carcino- We established hepatocellular carcinoma cell lines stably expres- ma model did not increase the incidence of hepatocellular sing membrane IL1a by mutating its NLS and calpain cleavage site carcinoma, but rather, prevented tumor development. The and generated orthotropic hepatocellular carcinoma model by serum IL1a levels were not elevated by membrane IL1a

OF8 Cancer Res; 76(11) June 1, 2016 Cancer Research

Membrane IL1a Inhibits HCC Development

overexpression in DEN-induced (data not shown) or orthoto- 41). Therefore, inducing the right inflammation in the tumor pic tumor model (Fig. 4C), suggesting that the release of IL1a microenvironment may be an ideal strategy for cancer immuno- from necrotic cells may be an active process originating from therapy by increasing the immunogenicity of the tumor. The IL1a precursor instead of passive release from membrane IL1a. injection of fibrosarcoma-expressing cell-associated IL1a could Our results certainly suggested that the function of membrane generate long-term tumor-specific memory (14). Therefore, mem- IL1a was different from the IL1a released from necrotic hepa- brane IL1a could potentially be incorporated in certain tumor tocytes during hepatocellular carcinoma development. vaccines to promote antitumor immune responses. Further stud- We showed that membrane IL1a promoted the activation of T ies on the mechanisms of different IL1a molecules on malignant and NK cells. Potentially, membrane IL1a may serve as a cell- processes will be helpful for understanding tumor–host interac- associated costimulatory molecule for activation of T cells and NK tions and generating novel strategies for tumor immunotherapy. cells by binding to IL1R1. Fibrosarcoma cells expressing cell- associated IL1a lose their tumorigenicity and it is dependent on Disclosure of Potential Conflicts of Interest þ fl CD8 T cells (14, 16). NK cells and activated macrophages also No potential con icts of interest were disclosed. play a role in eradication of IL1a-positive fibrosarcomas. We showed in our in vitro studies that the promotion of T-cell Authors' Contributions activation by membrane IL1a required cell–cell contact. This Conception and design: D. Lin, H. Liu Development of methodology: D. Lin, L. Lei, Y. Liu, Y. Zhang, X. Yu, H. Liu result further supports the role of membrane IL1a as a cell- Acquisition of data (provided animals, acquired and managed patients, associated costimulatory molecule for T-cell activation. Many cell provided facilities, etc.): D. Lin, L. Lei, Y. Liu, Y. Zhang, B. Hu, G. Bao, types in the liver express IL1Rs, including DCs and macrophages. Y. Song, Z. Jin, C. Liu, Y. Mei, D. Sandikin, Y. Wu, L. Zhao, X. Yu IL1a could function on DCs and macrophages to upregulate the Analysis and interpretation of data (e.g., statistical analysis, biostatistics, expression of cytokines that promote T-cell functions, including computational analysis): D. Lin, L. Lei, X. Yu, H. Liu IFNg and IL12. Previous studies indicate that tumor cell–associ- Writing, review, and/or revision of the manuscript: D. Lin, L. Lei, H. Liu Administrative, technical, or material support (i.e., reporting or organizing ated IL1a promotes antigen presentation by the tumor cells data, constructing databases): D. Sandikin, X. Yu themselves, possibly through IFNg-induced MHC class II expres- Study supervision: H. Liu sion or cross-presentation by professional antigen-presentation cells (39). We observed increases in percentages and numbers of Grant Support DCs and macrophages in spleen and liver by membrane IL1a This work has been supported by grants from National Natural Science expression (Fig. 3E and F). Therefore, DCs and macrophages may Foundation of China (81273268, 91029703, 81471586), the project funding also be modulated by membrane IL1a to facilitate T-cell from Suzhou city (SWG0904), Priority Academic Program Development of activation. Jiangsu Higher Education Institutions, and start-up grant from National Uni- versity of Singapore. Membrane-associated cytokines expressed on engineered tumor cells, such as IFNg, TNFa, M-CSF, and IL12, have been Received September 23, 2015; revised February 23, 2016; accepted March 17, shown to increase the immunogenicity of malignant cells (33, 40, 2016; published OnlineFirst March 29, 2016.

References 1. Yang JD, Roberts LR. Hepatocellular carcinoma: a global view. Nat Rev 12. Kurt-Jones EA, Beller DI, Mizel SB, Unanue ER. Identification of a mem- Gastroenterology Hepatol 2010;7:448–58. brane-associated interleukin 1 in macrophages. Proc Natl Acad Sci U S A 2. Gao J, Xie L, Yang WS, Zhang W, Gao S, Wang J, et al. Risk factors of 1985;82:1204–8. hepatocellular carcinoma–current status and perspectives. Asian Pac J 13. Dinarello CA. Biologic basis for interleukin-1 in disease. Blood 1996;87: Cancer Prev 2012;13:743–52. 2095–147. 3. Blonski W, Kotlyar DS, Forde KA. Non-viral causes of hepatocellular 14. Dvorkin T, Song X, Argov S, White RM, Zoller M, Segal S, et al. carcinoma. World J Gastroenterology 2010;16:3603–15. Immune phenomena involved in the in vivo regression of fibrosarco- 4. El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epidemiology and ma cells expressing cell-associated IL-1alpha. J Leukoc Biol 2006;80: molecular carcinogenesis. Gastroenterology 2007;132:2557–76. 96–106. 5. Farazi PA, DePinho RA. Hepatocellular carcinoma pathogenesis: from 15. Sakurai T, He G, Matsuzawa A, Yu GY, Maeda S, Hardiman G, et al. genes to environment. Nat Rev Cancer 2006;6:674–87. Hepatocyte necrosis induced by oxidative stress and IL-1 alpha release 6. Aravalli RN, Cressman EN, Steer CJ. Cellular and molecular mechan- mediate carcinogen-induced compensatory proliferation and liver tumor- isms of hepatocellular carcinoma: an update. Arch Toxicol 2013;87: igenesis. Cancer Cell 2008;14:156–65. 227–47. 16. Song X, Voronov E, Dvorkin T, Fima E, Cagnano E, Benharroch D, et al. 7. Demaria S, Pikarsky E, Karin M, Coussens LM, Chen YC, El-Omar EM, et al. Differential effects of IL-1 alpha and IL-1 beta on tumorigenicity patterns Cancer and inflammation: promise for biologic therapy. J Immunother and invasiveness. J Immunol 2003;171:6448–56. 2010;33:335–51. 17. Beasley D, Cooper AL. Constitutive expression of interleukin-1alpha 8. Amann T, Bataille F, Spruss T, Muhlbauer M, Gabele E, Scholmerich J, et al. precursor promotes human vascular smooth muscle cell proliferation. Am Activated hepatic stellate cells promote tumorigenicity of hepatocellular J Physiol 1999;276:H901–12. carcinoma. Cancer Sci 2009;100:646–53. 18. Tjomsland V, Spangeus A, Valila J, Sandstrom P, Borch K, Druid H, et al. 9. Bruix J, Boix L, Sala M, Llovet JM. Focus on hepatocellular carcinoma. Interleukin 1a sustains the expression of inflammatory factors in human Cancer Cell 2004;5:215–9. pancreatic cancer microenvironment by targeting cancer-associated fibro- 10. Leonardi GC, Candido S, Cervello M, Nicolosi D, Raiti F, Travali S, et al. The blasts. Neoplasia 2011;13:664–75. tumor microenvironment in hepatocellular carcinoma (review). Int J 19. Stevenson FT, Turck J, Locksley RM, Lovett DH. The N-terminal propiece of Oncol 2012;40:1733–47. interleukin 1 alpha is a transforming nuclear oncoprotein. Proceed Natl 11. Brody DT, Durum SK. Membrane IL-1: IL-1 alpha precursor binds to the Acad Sci U S A 1997;94:508–13. plasma membrane via a lectin-like interaction. J Immunol 1989;143: 20. Mackay AR, Ballin M, Pelina MD, Farina AR, Nason AM, Hartzler JL, et al. 1183–7. Effect of phorbol ester and cytokines on matrix metalloproteinase and

www.aacrjournals.org Cancer Res; 76(11) June 1, 2016 OF9

Lin et al.

tissue inhibitor of metalloproteinase expression in tumor and normal cell a CD28-dependent mechanism that does not involve increased expression lines. Invasion Metastasis 1992;12:168–84. of CD28 ligands. Eur J Immunol 2004;34:1085–90. 21. Tran-Thang C, Kruithof E, Lahm H, Schuster WA, Tada M, Sordat B. 31. Klarnet JP, Kern DE, Dower SK, Matis LA, Cheever MA, Greenberg PD. Modulation of the plasminogen activation system by inflammatory cyto- Helper-independent CD8þ cytotoxic T lymphocytes express IL-1 receptors kines in human colon carcinoma cells. Br J Cancer 1996;74:846–52. and require IL-1 for secretion of IL-2. J Immunol 1989;142:2187–91. 22. Lafrenie RM, Gallo S, Podor TJ, Buchanan MR, Orr FW. The relative roles of 32. Conlon PJ, Grabstein KH, Alpert A, Prickett KS, Hopp TP, Gillis S. Local- vitronectin receptor, E-selectin and alpha 4 beta 1 in cancer cell adhesion to ization of human mononuclear cell interleukin 1. J Immunol 1987;139: interleukin-1-treated endothelial cells. Eur J Cancer 1994;30A:2151–8. 98–102. 23. Lauri D, Needham L, Martin-Padura I, Dejana E. Tumor cell adhesion to 33. el-Shami KM, Tzehoval E, Vadai E, Feldman M, Eisenbach L. Induction of endothelial cells: endothelial leukocyte adhesion molecule-1 as an induc- antitumor immunity with modified autologous cells expressing mem- ible adhesive receptor specific for colon carcinoma cells. J Natl Cancer Inst brane-bound murine cytokines. J Interferon Cytokine Res 1999;19: 1991;83:1321–4. 1391–401. 24. Tomimatsu S, Ichikura T, Mochizuki H. Significant correlation between 34. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. expression of interleukin-1alpha and liver metastasis in gastric carcinoma. Cell 2010;140:883–99. Cancer 2001;91:1272–6. 35. Zamarron BF, Chen W. Dual roles of immune cells and their factors in 25. Duewell P, Kono H, Rayner KJ, Sirois CM, Vladimer G, Bauernfeind FG, cancer development and progression. Intl J Biol Sci 2011;7:651–8. et al. NLRP3 inflammasomes are required for atherogenesis and activated 36. Lotze MT, Deisseroth A, Rubartelli A. Damage associated molecular pattern by cholesterol crystals. Nature 2010;464:1357–61. molecules. Clin Immunol 2007;124:1–4. 26. Rajamaki K, Lappalainen J, Oorni K, Valimaki E, Matikainen S, Kovanen PT, 37. Rider P, Carmi Y, Guttman O, Braiman A, Cohen I, Voronov E, et al. IL- et al. Cholesterol crystals activate the NLRP3 inflammasome in human 1alpha and IL-1beta recruit different myeloid cells and promote different macrophages: a novel link between cholesterol metabolism and inflam- stages of sterile inflammation. J Immunol 2011;187:4835–43. mation. PLoS One 2010;5:e11765. 38. Voronov E, Dotan S, Krelin Y, Song X, Elkabets M, Carmi Y, et al. Unique 27. Hu B, Bao G, Zhang Y, Lin D, Wu Y, Wu D, et al. Donor NK Cells and IL-15 versus redundant functions of IL-1alpha and IL-1beta in the tumor micro- promoted engraftment in nonmyeloablative allogeneic bone marrow environment. Front Immunol 2013;4:177. transplantation. J Immunol 2012;189:1661–70. 39. Apte RN, Dotan S, Elkabets M, White MR, Reich E, Carmi Y, et al. The 28. Lin D, Lei L, Zhang Y, Hu B, Bao G, Liu Y, et al. Secreted IL-1alpha promotes involvement of IL-1 in tumorigenesis, tumor invasiveness, metastasis and T-cell activation and expansion of CD11b(þ) Gr1(þ) cells in carbon tumor-host interactions. Cancer Metastasis Rev 2006;25:387–408. tetrachloride-induced liver injury in mice. Eur J Immunol 2015;45: 40. Marr RA, Addison CL, Snider D, Muller WJ, Gauldie J, Graham FL. Tumour 2084–98. immunotherapy using an adenoviral vector expressing a membrane- 29. Hunter CA, Timans J, Pisacane P, Menon S, Cai G, Walker W, et al. bound mutant of murine TNF alpha. Gene Ther 1997;4:1181–8. Comparison of the effects of interleukin-1 alpha, interleukin-1 beta and 41. Soo Hoo W, Lundeen KA, Kohrumel JR, Pham NL, Brostoff SW, Bartho- interferon-gamma-inducing factor on the production of interferon-gamma lomew RM, et al. Tumor cell surface expression of granulocyte-macrophage by natural killer. Eur J Immunol 1997;27:2787–92. colony-stimulating factor elicits antitumor immunity and protects from 30. Khoruts A, Osness RE, Jenkins MK. IL-1 acts on antigen-presenting cells to tumor challenge in the P815 mouse mastocytoma tumor model. J Immu- enhance the in vivo proliferation of antigen-stimulated naive CD4 T cells via nol 1999;162:7343–9.

OF10 Cancer Res; 76(11) June 1, 2016 Cancer Research

Membrane IL1α Inhibits the Development of Hepatocellular Carcinoma via Promoting T- and NK-cell Activation

Dandan Lin, Lei Lei, Yonghao Liu, et al.

Cancer Res Published OnlineFirst March 29, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-15-2658

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2016/03/29/0008-5472.CAN-15-2658.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2016/05/20/0008-5472.CAN-15-2658. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.