Cancer immunotherapy has shown great promise as a new standard therapeutic strategy against cancer. However, the response rate and survival benefit remain unsatisfactory because most current approaches, such as the use of immune checkpoint inhibitors, depend on spontaneous antitumor immune responses. One possibility for improving the efficacy of immunotherapy is to promote antitumor immunity using adjuvants or specific cytokines actively. IL-33 has been a candidate for such cytokine therapies, but it remains unclear how and in which situations IL-33 exerts antitumor immune effects. In this study, we demonstrate the potent antitumor effects of IL-33 using syngeneic mouse models, which included marked inhibition of tumor growth and upregulation of IFN-γ production by tumor-infiltrating CD8+ T cells. Of note, IL-33 induced dendritic cells to express semaphorin 4A (Sema4A), and the absence of Sema4A abolished the antitumor activity of IL-33, indicating that Sema4A is intrinsically required for the antitumor effects of IL-33 in mice. Collectively, these results not only present IL-33 and Sema4A as potential therapeutic targets but also shed light on the potential use of Sema4A as a biomarker for dendritic cell activation status, which has great value in various fields of cancer research, including vaccine development.
Cancer immunotherapy with immune checkpoint inhibitors (ICIs) has significantly advanced clinical cancer management, but this approach is not effective in all cases (1). One reason for treatment failure is the lack of a priming phase of the cancer-immune cycle (2) (i.e., the lack of sufficient CTL activation by APCs, such as dendritic cells [DCs]). Therefore, augmentation of DC-mediated CTL activation either by adjuvants or cytokines is expected to provide more efficacious cancer immunotherapy (3). Although several cytokines are candidates for such DC-promoting therapies, few of them have been approved for clinical application.
IL-33 activates Th2 and group 2 innate lymphoid cells and induces allergic diseases such as allergic rhinitis, spontaneous dermatitis, and asthma (4–8). Recent studies have described novel pleiotropic functions of IL-33, in particular the enhancement of type 1 immunity mediated by Th1 or type 1 CD8+ CTLs (Tc1) (4, 9–11), which led to the identification of IL-33–related diseases, such as infection, inflammation, and metabolic and degenerative disorders. In addition, cumulative studies reported that IL-33 is involved in antitumor immunity (12–14). Notably, although the relevant antitumor immune cells differ among studies, the involvement of IFN-γ is consistent, supporting the importance of IL-33 in polarization to type 1 immunity (15, 16). Recent findings have indicated that IL-33 activates DCs and exerts antitumor effects (17). In addition, IL-33 promoted maturation and activation of DCs, driving antitumor effects by inducing Tc9 cells (18). However, the therapeutic potential of IL-33 remains controversial (4, 9–11); indeed, several studies have reported that IL-33 has protumor functions (19–23). The direct protumor effects of IL-33 include the promotion of proliferation, invasion, and metastasis of cancer cells (19–21). Another important protumor function is enhancing immunosuppression by activating immunosuppressive cells, including regulatory T (Treg) cells and M2-polarized macrophages (22), and suppressing NK cells (23). It seems that these discrepancies are attributable to the pleiotropic functions of IL-33 and the variability of target cells in each experimental system. In any case, it remains unclear how and under which conditions IL-33 exerts antitumor activities.
Semaphorins, a family of proteins that include membrane-bound or secreted proteins, are divided into eight classes based on sequence similarity and distinct structural features. Semaphorins play roles in development, immunity, and various diseases, including inflammation and cancers (24–27). Of these, semaphorin 4A (Sema4A) is involved in multiple physiological and pathological situations, including angiogenesis, immune response, and cancer (28). Our previous studies revealed that Sema4A expressed on DCs activates T cells and contributes to T cell priming (29, 30). In addition, expression of Sema4A is induced on Th1 cells, and this expression is relevant to the balance between Th1 and Th2 responses (29, 31). Moreover, we previously showed that the CD8+ T cell activation is regulated by the Sema4A–Plexin B2 axis that induced IFN-γ production through the mTORC1 pathway (32). However, the functional importance of Sema4A in antitumor immunity remains unclear.
In this study, we identified the antitumor effects of IL-33 using syngeneic mouse models. The effects were mediated by an increase in the level of IFN-γ produced by CD8+ T cells, whereas the involvement of NK cells could not be verified. IL-33 caused DCs to express Sema4A, thereby promoting CD8+ T cell activation and IFN-γ upregulation. The antitumor effects were completely abolished in Sem4A knockout (KO) mice. Taken together, these results indicated that IL-33 and Sema4A are potent and attractive mediators of antitumor immunity, allowing the development of novel strategies for cancer immunotherapy.
Materials and Methods
Cell lines and cell culture
Lewis lung carcinoma (LLC) and B16 cells were purchased from the American Type Culture Collection (Manassas, VA). Platinum-GP cells were purchased from CosmoBio (Tokyo, Japan). LLC and B16 cells were cultured in complete RPMI (cRPMI; RPMI 1640 [Nacalai Tesque, Kyoto, Japan] supplemented with 10% heat-inactivated FBS [Sigma-Aldrich, Tokyo, Japan], 100 U/ml penicillin, and 100 mg/ml streptomycin [Nacalai Tesque]).
Mouse IL-33 expression vector was purchased from Open Biosystems (GE Dharmacon, Tokyo, Japan). This vector was subcloned into the murine stem cell virus retroviral vector (Clontech Laboratories, Shiga, Japan). The resultant vector was cotransfected with pVSV-G (Riken, Osaka, Japan) into Platinum-GP packaging cells using Lipofectamine 2000 DNA Transfection Reagent (Thermo Fisher Scientific, Tokyo, Japan), yielding a pantropic retroviral vector. The IL-33 vector was transduced into LLC cells as previously described, with slight modifications (27). Clones stably expressing IL-33 were established by selection in puromycin (2 µg/ml; Sigma-Aldrich) for 4 d.
Animal studies were approved by the ethical review board of the Osaka University Graduate School of Medicine (approval no. 28-008-033 and 28-031-006), and protocols for animal experiments were approved by the Institute of Experimental Animal Sciences of Osaka University Medical School. Six-week-old male C57BL/6J mice were purchased from CLEA Japan (Tokyo, Japan). C57BL/6J Sema4A−/− mice were established previously (30) and bred at the Animal Resource Center for Infectious Diseases, Research Institute for Microbial Diseases and Immunology Frontier Research Center, Osaka University.
Abs and reagents
Abs used for depleting specific immune cells or cytokines in tumor-challenged mice were as follows: anti-NK1.1 (136; Bio X Cell, Lebanon, NH), anti-CD8 Ab (53-6.7; Bio X Cell), and anti–IFN-γ (XMG1.2; BioLegend, San Diego, CA).
Abs for flow cytometry (FCM) were purchased from the indicated suppliers: PE-conjugated anti-mouse ST2 (U29-93; Becton Dickinson, Franklin Lakes, NJ), PE-conjugated rat IgG2a isotype control (R35-95; Becton Dickinson), PE-conjugated anti-mouse Plexin B2 (3E7; BioLegend), PE-conjugated American hamster IgG isotype control (HTK888; BioLegend), PE-conjugated anti-mouse Foxp3 (MF-14; BioLegend), APC-conjugated anti-mouse CD8a (53-6.7; BioLegend), APC-conjugated anti-mouse H-2kb bound to SIINFEKL (25-D1.16; BioLegend), APC-conjugated anti-mouse Sema4A (5E3; BioLegend), APC-conjugated anti-mouse IgG1 κ isotype control (MOPC-21), PE/Cy7-conjugated anti-mouse CD3ε (145-2C11; BioLegend), PE/Cy7-conjugated anti-mouse/human CD11b (M1/70; BioLegend), PE/Cy7-conjugated anti-mouse CD69 (H1.2F3; BioLegend), FITC-conjugated anti-mouse CD49b (DX5; BioLegend), FITC-conjugated anti-mouse I-A/I-E (M5/114.15.2; BioLegend), FITC-conjugated anti-mouse CD40 (HM40-3; BioLegend), FITC-conjugated mouse IgG1 κ isotype control (MOPC-21; BioLegend), PerCP/Cyanine5.5-conjugated anti-mouse CD4 (RM4-5; BioLegend), PerCP/Cyanine5.5-conjugate anti-mouse/human CD45R/B220 (RA3-6B2; BioLegend), PerCP/Cyanine5.5-conjugated anti-mouse CD103 (2E7; BioLegend), PerCP/Cyanine5.5-conjugated anti-mouse Ki67 (16A8; BioLegend), PerCP/Cyanine5.5-conjugated anti-mouse Ly-6G (1A8; BioLegend), Brilliant Violet 421–conjugated anti-mouse CD11c (N418; BioLegend), APC/Cy7-conjugated anti-mouse/human CD11b (M1/70; BioLegend), APC/Cy7-conjugated anti-mouse CD86 (GL-1; BioLegend), and APC/Cy7-conjugated mouse IgG1 κ isotype control (MOPC-21; BioLegend).
Abs used for coating plates in the CD8+ T cell activation assay were anti-CD3ε (clone 145-2C11; BD Pharmingen, Tokyo, Japan) and anti-CD28 (clone 37.51; BD Pharmingen). Anti–IL-4 Ab (clone 11B11, eBioscience, San Diego, CA) was added to the supernatant for inducing the Tc1 condition.
The following reagents were used in this study: recombinant murine IL-33 (BioLegend), recombinant murine IL-2 (R&D Systems, Minneapolis, MN), and recombinant murine IL-12 (R&D Systems).
Tumor challenge and treatments
LLC cells (2 × 106) or B16 melanoma cells (5 × 105) suspended in PBS were inoculated s.c. into the right flanks of the mice. There were six mice in each group. From day 9 after tumor challenge, 1 µg rIL-33 was administered i.p. every 3 d (days 9, 12, 15, 18, 21, 24, and 27). Tumor size was monitored every 3 d. Engrafted tumors were obtained on day 15 or 27. Tumor appearance, tumor diameter, and tumor volume were monitored. Tumor volume was determined as width2 × length/2 (20). Depletion of NK cells, CD8+ T cells, and IFN-γ was achieved using anti-NK1.1, anti-CD8, and anti–IFN-γ Abs, respectively. The mice were divided into three groups, each of which received a different Ab. Anti-NK1.1 Ab was administered at a dose of 500 µg/mouse on day 1 and at 250 µg/mouse on days 3, 7, and 14. Normal mouse IgG2a (Santa Cruz Biotechnology, Dallas, TX) was administered in equal doses as a control. Anti-CD8 Ab was administered at 200 µg/mouse on days 1, 8, and 15. Normal rat IgG2a (Fuji, Osaka, Japan) was administered in equal doses as a control. Anti–IFN-γ Ab was administered at 500 µg/mouse on days 9, 12, 15, and 18. Normal Rat IgG2 (BioLegend) was administered in equal doses as a control.
Preparation of tumor-infiltrating leukocytes
LLC syngeneic graft were obtained after administration of PBS or rIL-33 to C57BL/6J wild-type (WT) or Sema4A KO mice and shredded into small pieces, followed by incubation in collagenase-containing buffer (100 U/ml collagenase type IV [Thermo Fisher Scientific], 100 ng/ml of DNase I [Roche Diagnostics, Basel, Swiss], and 10% FBS in RPMI 1640 medium) for 45 min. After incubation, cells were passed through a cell strainer to remove debris. RBCs were destroyed by incubation with RBC Lysis Buffer (eBioscience) for 15 min at 20°C. The cell pellet was dissolved in 2% FBS for FCM analysis or processed for mRNA extraction, followed by RT-PCR and RNA sequencing (RNA-seq). CD8high and CD11chigh cells were obtained by autoMACS (Miltenyi Biotec, Cologne, Germany).
Cells were stained for FCM analysis of cell-surface markers with the indicated fluorochrome-conjugated Abs diluted in PBS (Nacalai Tesque) containing 2% FBS. Cells were blocked with Fc block (1:10; Miltenyi Biotec) for 15 min on ice and then stained with each Ab or isotype control for 20 min at 20°C. After staining of cell-surface molecules, the expression level of each molecule was determined for the 7-aminoactinomycin D (1:200; BioLegend)–negative population. We considered CD45+ CD3+ CD4+ cells as CD4+ T cells, CD45+ CD3+ CD8+ cells as CD8+ T cells, CD45+ CD3− CD19− DX5+ cells as NK cells, CD45+ CD11b+ Ly-6G+ cells as neutrophils, CD45+ CD11b+ Ly-6G− cells as monocytes, and CD45+ CD11c+ cells as DCs. For intracellular staining, cells were incubated in the presence of brefeldin A (GolgiPlug; BD Biosciences) for 5 h. After staining of cell-surface molecules, intracellular staining was performed using the Cytofix/Cytoperm kit (BD Biosciences). Stained cells were analyzed on a BD FACSCanto II flow cytometer equipped with the Diva software (BD Biosciences). The data were analyzed, and figures were generated using FlowJo (Tree Star, Ashland, OR).
CD8+ T cell culture
Lymphocytes were collected from spleens obtained from C57BL/6 WT and Sema4A KO mice. CD8+ T cells were purified by FCM or a magnetic bead–based method and cultured in cRPMI. For in vitro generation of effector cells, CD8+ T cells were cultured in a 96-well, round-bottom microplate and then activated with soluble anti-CD3ε (1 μg/ml) and anti-CD28 (1 μg/ml) in cRPMI for 4 d in the presence of different cytokine mixtures. To generate Tc0 cells, CD8+ T cells were differentiated for 4 d in the presence of recombinant mouse IL-2 (20 U/ml). To generate Tc1 cells, CD8+ T cells were differentiated for 4 d in the presence of recombinant mouse IL-2 (20 U/ml), recombinant mouse IL-12 (3.4 ng/ml), and anti–IL-4 Ab (10 μg/ml). CD8+ T cells under the Tc0 and Tc1 conditions were stimulated with PBS or rIL-33 (10 ng/ml) for 24 h.
Lymphocytes were collected from spleens and tumors obtained from C57BL/6 WT and Sema4A KO mice. For DC selection, spleen cells were suspended in cold PBS at a concentration of 1 × 108 cells/ml and incubated with anti-mouse CD11c beads (Miltenyi Biotec) for 30 min at 4°C. CD11c+ cells were purified by FCM or a magnetic bead–based method and cultured in cRPMI. DC cells from WT, Sema4A KO, and MyD88 KO mice were stimulated with rIL-33 (10 ng/ml) for 24 h and analyzed by FCM. MyD88 Inhibitor Peptide (NBP-2; Novus Biologicals, Centennial, CO) was also used to examine downstream signaling mechanisms.
Coculture of DCs and CD8+ T cells
Coculture experiments containing DCs and CD8+ T cells were performed as described previously (33) with slight modifications. Briefly, DCs from WT and Sema4A KO mice were stimulated with OVA (100 ng/ml) for 24 h, followed by the addition of PBS or IL-33 (10 ng/ml) and incubation for the next 24 h. Splenic lymphocytes were obtained from OT-1 mice containing transgenic inserts of mouse Tcra-V2 and Tcra-V5 genes to recognize OVA residues 257–264 in the background of H-2Kb. CD8+ T cells were collected and purified by FCM or a magnetic bead–based method, followed by incubation in cRPMI. These preconditioned DCs and CD8+ T cells were cocultured for an additional 24 h.
IL-12 concentration in culture supernatants was detected by ELISA as previously described (34). Capture/detection Abs for IL-12 and recombinant mouse IL-12, used as an ELISA standard, and avidin-HRP were purchased from R&D Systems.
Total RNA was extracted using the RNeasy Mini Kit (QIAGEN, Tokyo, Japan), and subsequently, cDNA was synthesized using the SuperScript IV cDNA synthesis kit (Thermo Fisher Scientific). Quantitative RT-PCR analysis was performed on a 7900HT Fast Real-Time PCR system (Thermo Fisher Scientific) using the TaqMan protocol. TaqMan real-time probes for mouse Sema4a (Mm00443140_m1), Actb (Mm00607939_s1), Il33 (Mm00505403_m1), Il12b (Mm01288989_m1), and Ifng (Mm01168134_m1) were purchased from Applied Biosystems. Samples were run in triplicate using Actb as an internal control, and the ΔΔ threshold cycle method was used to calculate relative mRNA levels.
The extracted total RNA was treated with DNase I (Thermo Fisher Scientific) and cleaned up again with the RNeasy Mini Kit. The processed RNA was subjected to RNA-seq analysis as described previously (35). Briefly, library preparation was performed using a TruSeq Stranded mRNA Sample Prep kit (Illumina, Tokyo, Japan). Sequencing was performed on an Illumina HiSeq 2500 platform in 75-base, single-end mode. CASAVA 1.8.2 software (Illumina) was used for base calling. Sequenced reads were mapped to the mouse reference genome sequence (mm10) using TopHat v2.0.13 in combination with Bowtie2 ver. 2.2.3 and SAMtools ver. 0.1.19. Fragments per kilobase of exon per million mapped fragments were calculated using Cuffnorm version 2.2.1.
Modified MTT assay
Cell proliferation was measured by a modified MTT assay (Cell Counting Kit-8; Dojindo, Kumamoto, Japan). Briefly, cells were plated on 96-well plates, and cell viability was assayed after incubation for 72 h. Each assay was performed in quadruplicate, and means and SD were calculated.
Genomic DNA from all cell lines was obtained using the AllPrep DNA/RNA Mini kit (QIAGEN). Mycoplasma detection was performed using the e-Myco Plus Mycoplasma PCR Detection Kit (iNtRON, Jungwon-Gu, South Korea). Only cells that were confirmed to be negative for mycoplasma were used in this study.
All statistical evaluations for experiments using cell lines and mice were repeated at least three times. Data are expressed as mean ± 2 SD or mean ± 2 SEM. For FCM, quantitative RT-PCR, and animal experiments, differences were evaluated using the Wilcoxon or Steel test. All p values <0.05 were considered significant. All statistical analyses were performed using the JMP Pro 14.0.0 software (SAS Institute Japan, Tokyo, Japan).
IL-33 suppresses in vivo tumor growth
To examine the roles of IL-33 in antitumor immunity in vivo, we conducted experiments using a syngeneic mouse tumor model consisting of LLC and B16 cells in C57BL/6 mice. When IL-33 administration was initiated 9 d after LLC inoculation, tumor growth was significantly suppressed (Fig. 1A) even though in vitro proliferation of LLC cells was not affected by IL-33 (data not shown). In vivo experiments using B16 cells also revealed significant tumor-suppressive effects of IL-33 (Fig. 1B).
To confirm the tumor-suppressive function of IL-33, we generated another model using IL-33–overexpressing (OE) tumor (LLC) cells. The in vitro growth rates of IL-33–OE and empty vector–transfected (EV) LLC cells were equivalent (Fig. 1C, left). However, when inoculated into mice, the IL-33–OE cells showed a significantly reduced tumorigenic capacity (Fig. 1C, right). Taken together, these findings indicate that IL-33 exerts tumor-suppressive properties by modifying components of the tumor microenvironment (TME), such as immune cells, rather than acting on the tumor cells themselves.
Dynamic immune-cell activation profiling in the TME
We next monitored immune-cell infiltration into the TME. For this purpose, we extracted the tumors and isolated immune cells on day 15, when no difference in tumor volume was observed in the presence or absence of IL-33, and analyzed them by FCM. The numbers of CD4+ T and CD8+ T cells within the tumor were not altered by IL-33 (Fig. 1D), whereas neutrophils, identified based on the shape of the nucleus and the particle size of the cytoplasm as well as cell-surface markers, were slightly less abundant in the presence of IL-33. By contrast, the proportions of ST2-positive CD4+ T and CD8+ T cells within the tumor were significantly elevated in mice treated with IL-33, whereas ST2-positive NK cells were not increased. (Fig. 1E, Supplemental Fig. 1). We then measured the expression of IFN-γ, a representative type 1 cytokine, in tumor tissues and found it to be significantly elevated in TME extract from mice treated with IL-33 (Fig. 1F). IL-33 also induced Treg cells in the TME (Fig. 1G).
Then, we examined those immune cells at later phases to observe their activation status over time. On day 27, the numbers of CD4+ T cells, CD8+ T cells, NK cells, and neutrophils within the tumor did not differ significantly (Supplemental Fig. 2A), and immune-cell activation status, based on ST2 expression, was comparable among these cell types, except neutrophils (Supplemental Fig. 2B). Moreover, the expression of IFN-γ was comparable in tumor tissues with or without IL-33 treatment, as determined by quantitative PCR (qPCR) (Supplemental Fig. 2C). Taken together, these findings indicated that IL-33 suppressed tumor growth by activating type 1 immunity at the early, but not later, phases of antitumor immune responses.
IL-33 exerts its antitumor effects through CD8+ T cells as well as IFN-γ
Based on the results described above, we hypothesized that IL-33 promoted type 1 immunity, and CD8+ CTLs and IFN-γ derived from CTLs were key effectors of IL-33 antitumor functions. To identify the effector cells, we depleted IFN-γ, CD8+ T cells, or NK cells using anti–IFN-γ, anti-CD8, and anti-NK1.1 Abs, respectively. Depletion of IFN-γ eradicated the antitumor effect of IL-33 (Fig. 2A). Depletion of CD8+ T cells with anti-CD8 Ab attenuated the effects of IL-33, with a significant reduction of tumor-infiltrating CD8+ T cells (Fig. 2B). However, depletion of NK cells with anti-NK1.1 Ab had no influence, albeit tumor-infiltrating NK cells were reduced (Fig. 2C). These alterations in cell numbers were also observed in splenic samples (Fig. 2D). These results indicated that IL-33 exerted its antitumor effects via activated CTLs and IFN-γ secreted mainly from those cells, rather than by activating NK cells. To determine the function of IL-33 in CD8+ CTL activation, we examined the influence of IL-33 on Tc0 and Tc1 differentiation in vitro. We cultured CD8+ T cells under Tc0 or Tc1 conditions and then added IL-33. Altering conditions from Tc0 to Tc1 upregulated IFN-γ production in CD8+ T cells. However, the activation of CD8+ T cells by IL-33 was obscure in both Tc0 and Tc1 conditions (Fig. 2E), which contrasted with the in vivo results with mouse tumor tissues that showed IL-33–induced IFN-γ production. These findings suggested that IL-33 induced a Tc1-like condition in the TME, resulting in CD8+ T cell activation and IFN-γ production.
IL-33 stimulates DCs to express Sema4A
We next investigated the role of IL-33 on DCs, the main source of the Tc1-inducing factor IL-12. We confirmed the cell-surface expression of ST2 on DCs by FCM (Fig. 3A). To examine the role of IL-33 on DC functions in vivo, non–tumor-bearing mice were treated with IL-33, followed by extraction and examination of splenic DCs. RNA-seq analyses of the DCs from mice treated with IL-33 revealed elevation of Il1rl1 (the gene encoding ST2) expression compared with that in PBS-treated mice. However, we observed no significant upregulation of Il-12 in IL-33–treated mice (Fig. 3B); all gene expression data have been submitted to National Center for Biotechnology Information Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE150110). Instead, another candidate effector of activated DCs, Sema4a, was upregulated, which was confirmed by qPCR (Fig. 3C). Sema4A protein expression was also significantly elevated, as demonstrated by FCM (Fig. 3D, 3E). By contrast, we did not observe the upregulation of other costimulatory molecules such as CD40 and CD86 (Fig. 3D). Next, we investigated whether IL-33 could directly induce Sema4A expression on the DC surface in vitro. IL-33 increased expression of Sema4A, as determined by both qPCR and FCM (Fig. 3F, 3G). These findings suggested that IL-33 stimulated DCs to express Sema4A, which in turn results in the activation of CTLs and upregulation of IFN-γ. To investigate the mechanism underlying induction of Sema4A expression by IL-33 on DCs, splenic DCs were collected from nontreated C57BL/6 mice and MyD88 KO mice and cultured with or without IL-33 for 24 h. Sema4A expression on DCs was evaluated by FCM. As shown in (Fig. 3H, Sema4A expression was not detected in MyD88 KO mice. To confirm the involvement of MyD88 signaling, we performed the experiments using an MyD88 inhibitor, which reduced the expression of Sema4A by IL-33 (data not shown). These findings imply that IL-33 induces Sema4A expression on DCs through MyD88-mediating signaling pathways.
Sema4A is implicated in the antitumor effects of IL-33
Next, we evaluated the involvement of Sema4A in IL-33–induced antitumor immunity using Sema4A KO mice. The antitumor effects of IL-33 were significantly abolished in KO mice than in WT mice (Fig. 4A). We extracted the tumors and isolated the immune cells on day 15 when there was no difference in tumor volume in the presence or absence of IL-33. Expression of IFN-γ in tumor tissues and CD8+ T cells was not elevated in KO mice compared with that in WT mice, as determined by qPCR (Fig. 4B, 4C). The number of CD8+ T cells was not altered by IL-33 treatment in both WT and KO mice (Fig. 4D). The proportion of ST2-positive CD8+ T cells was increased by IL-33 treatment in both WT and KO mice (Fig. 4E). We examined Sema4A expression in immune cells within the tumor, draining lymph nodes, and spleen from WT mice. IL-33 induced Sema4A in both DCs and CD8+ T cells, but not in CD4+ T cells (Fig. 5). These data implied that Sema4A expression in host cells was essential for the antitumor effects of IL-33 and that DCs and CTLs are the candidate cells in which Sema4A plays pivotal roles.
We first investigated DCs in which Sema4A might mediate antitumor effects of IL-33. IL-33 increased the number of DCs comparably in the TME in both WT and KO mice (Fig. 6A). Next, we evaluated the expression of representative costimulatory molecules on DCs from tumor-bearing WT and Sema4A KO mice with or without IL-33 treatment by FCM. Expression levels of these molecules were almost similar except for those of MHC class II (MHC II) (Fig. 6B). IL-12 expression was not affected by IL-33 in splenic DCs from either WT or KO mice (Fig. 6C), consistent with the RNA-seq results from WT mice (Fig. 3B). However, increased IL-12 expression was detected in the tumor tissue extract from IL-33–treated WT mice, but not KO mice (Fig. 6D). These data suggested that IL-33 indirectly induced the IL-12 expression in the TME in an Sema4A-dependent manner.
Sema4A induction on DCs is the key function of IL-33
Because Sema4A was expressed and upregulated by IL-33 in CD8+ T cells, we investigated whether its KO was intrinsically involved in the activation of CD8+ T cells. ST2 expression levels in CD4+ T and CD8+ T cells in the static states were comparable between WT and KO mice (Supplemental Fig. 3A). The increase in the ST2-positive population of CD8+ T cells under the Tc1 condition was comparable in WT and KO cells (Supplemental Fig. 3B). IFN-γ, CD69, and Ki67 induction in CD8+ T cells by IL-33 did not differ between WT and KO cells under Tc0 or Tc1 conditions (Supplemental Fig. 4A–D). These results suggested that CD8+ T cells from WT and KO mice were equally capable of being activated by IL-33.
The investigations described above provided evidence that DCs are one of the primary targets of IL-33. Hence, we sought to determine the influence of Sema4A deficiency in DC phenotypes. For this purpose, we treated non–tumor-bearing WT and KO mice with IL-33 and analyzed the phenotypes of their splenic DCs. Although Sema4A expression on DCs from WT mice was markedly induced by IL-33, the expression levels of costimulatory molecules (MHC class I, MHC II, CD40, and CD86) and cross-presentation ability of DCs were comparable between WT and KO mice (Fig. 7A, 7B). Next, splenic DCs from non–tumor-bearing WT mice were stimulated with PBS, IL-33, LPS, and IFN-γ. Sema4A expression was increased with IL-33, LPS, and IFN-γ treatment, as determined by FCM (Fig. 7C). By contrast, IL-12 expression was not induced by IL-33 in both WT and KO cells (Fig. 7D, 7E). Moreover, Sema4A expression on DCs was reproducibly increased by IL-33 in tumor-bearing WT mice (Fig. 7F). These data implied that the cardinal characteristic of KO DCs is the lack of Sema4A induction in response to IL-33.
Sema4A expression on DCs activates CTLs via Plexin B2
We hypothesized that Sema4A expression on DCs was the mediator of CTL activation. The major receptor for Sema4A, Plexin B2, was expressed on CD8+ T cells in the TME of both WT and KO mice (Fig. 8A), enabling them to respond to the Sema4A on DCs. To confirm whether Sema4A expression on DCs was involved in CD8+ T cell activation, DCs and CD8+ T cells were cocultured and characterized. DCs from WT and KO mice were stimulated with OVA, followed by the addition of either PBS or IL-33. The DCs were cocultured with CD8+ T cells from OT-1 mice, and the activation status of CD8+ T cells was analyzed by FCM. As shown in (Fig. 8B, the proportion of IFN-γ–expressing CD8+ T cells increased in WT, but not in KO, cells. These data suggested that IL-33–induced expression of Sema4A on DCs activated CD8+ T cells, probably via Sema4A–Plexin B2 axis.
In this study, we showed that IL-33 activates CD8+ T cells and upregulates IFN-γ, resulting in significant tumor suppression. Sema4A expression in host cells proved to be essential for the tumor-suppressive effect of IL-33 because the effect completely disappeared in Sema4A KO mice. Mechanistically, IL-33 induces Sema4A expression on DCs via MyD88-mediated signaling pathways, thereby activating CD8+ T cells probably via Sema4A–Plexin B2 axis. These data indicate the potential of Sema4A as a therapeutic target or a marker for DC activation status in tumor immunity.
Although the effects of IL-33 on tumor progression remain controversial, our findings demonstrate that IL-33 exerts antitumor effects. IL-33 administration suppressed tumor growth in vivo, although the proliferation of LLC and B16 cells was not affected by IL-33 in vitro, suggesting that the cytokine’s antitumor effects were mediated by components of the TME, including various immune cells and soluble factors. Among these, we identified CD8+ T cells and IFN-γ as the critical effectors of IL-33, which are consistent with previous reports (12–14). In contrast, some studies reported a protumor effect of IL-33, wherein immune-suppressive cells, including Treg cells or myeloid-derived suppressor cells, were identified as the effector cells (22, 23). Interestingly, we observed an IL-33–induced increase in Treg cells in the tumor tissues. This discrepancy is probably due to the heterogeneity and complexity of the TME components (36). Hence, we speculate that immune-suppressive factors, such as Treg cells and myeloid-derived suppressor cells, do not contribute to the effect in our model, resulting in the predominant tumor-suppressive activity of IL-33. With respect to tumor immunity, the TME can be roughly classified as hot tumor and cold tumor (36). The former is characterized by T cell infiltration and immune activation signatures, whereas the latter lacks these features, probably owing to insufficient activation of the cancer-immune cycle during the earlier phase. At present, cancer immunotherapies, including ICIs, are effective for hot tumors in which immune reactions are induced but evaded by the cancer cells. For cold tumors, ICIs are usually ineffective, and the usefulness of various experimental procedures, such as cancer vaccines and DC therapies, have not been established. In this context, IL-33 exhibited the ability to trigger the cancer-immune cycle at an earlier stage, suggesting that it has the potential to be an effective therapeutic modality for cold tumors. A recent study, published after completion of the current study, reported an interesting finding that IL-33 activated CD8+ T cells in the TME and was required for therapy with ICIs (37), supporting our proposal that IL-33 is a key factor responsible for facilitating antitumor immunity. Further investigations are required to fully clarify the roles of IL-33 in altering tumor growth and modifying the TME.
Another major finding of the current study is that Sema4A expression is critical for antitumor immunity mediated by IL-33. Although IL-33 has been reported to directly activate CD8+ T cells (16), we observed little direct action of IL-33 on CD8+ T cells in vitro. This could be explained by the differences in experimental conditions, such as the continued TCR stimulation under IL-33 administration (38). Based on these findings, we hypothesized that an important function of IL-33 in vivo was to create an environment suitable for CD8+ T cell activation, in addition to directly activating CD8+ T cells. We further evaluated IL-12 expression in splenic DCs by RNA-seq, as it is a key factor for CD8+ T cell activation and is mainly produced by DCs (39); however, no increase in its expression was observed. Notably, Sema4A expression was elevated by IL-33 in those DCs, suggesting its potential as a key mediator of IL-33–induced antitumor immunity. We confirmed the role of Sema4A by first verifying that it did not affect the cross-presentation ability and expression of costimulatory molecules. Next, coculture experiments provided evidence that Sema4A on DCs was required to sufficiently activate CD8+ T cells from OT-1 mice. Thus, our study, to our knowledge, reveals the novel findings that Sema4A upregulation in DCs plays a key role in IL-33–induced antitumor immunity. This might be relatively weak, indirect evidence of our claim that Sema4A on DCs activates CD8+ T cells through Plexin B2. However, our findings from a previous study (32) complement these data and support our hypothesis. This study showed that the Sema4A–Plexin B2 axis is required for the optimal activation of CD8+ T cells, which could explain the findings of the current study. Although using tumor Ags for the study would have been ideal because OVA is widely used for studying antitumor immunity, we chose OVA and regarded it as a control.
In addition, this study identified a unique cross-talk signal mediated by Sema4A between DCs and CD8+ T cells in the TME. As shown in (Fig. 6, IL-12 expression was not affected by IL-33 in splenic DCs from both WT and KO mice. However, increased IL-12 expression was detected in the tumor tissue extract obtained from IL-33–treated WT mice, but not KO mice. Tumor tissue infiltration and costimulatory molecule expression of DCs by IL-33 were almost similar in WT and KO mice. These data suggested that IL-33 indirectly induced the IL-12 expression in the TME in an Sema4A-dependent manner. In contrast, IFN-γ expression in CD8+ T cells in the TME was upregulated in IL-33–treated WT mice, but not in KO mice. Although we did not perform a detailed examination of CD8+ T cells in tumor-bearing mice, findings from in vitro experiments suggested that CD8+ T cells from WT and KO mice were equally capable of being activated under appropriate conditions. Taken together, establishment of a positive feedback loop consisting of IFN-γ and IL-12 is indicated in the TME (34, 40). In other words, our data highlight the significance of Sema4A on DC as a mediator for CTL activation, suggesting a potential mechanism of signal amplification, including IL-12. This model may be oversimplified because we focused on DCs and CD8+ T cells in the TME without considering other cell types, including myeloid-derived suppressor cells, monocytes, macrophages, fibroblasts, and endothelial cells. We recognize this as one of the limitations of this study, and further investigation is required to clarify how Sema4A impacts the tumor immunity mediated by IL-33. Moreover, this manuscript does not fully support the hypothesis that DCs are targets of IL-33. To clarify the involvement of DCs in detail, more complex experiments are needed using mice in which the ST2 gene is conditionally knocked out only in DCs. This is another limitation of this study.
To our knowledge, this is the first study to show that IL-33–induced Sema4A-expression in DCs is important for antitumor immunity. However, several challenges must be overcome before these findings can be applied in a clinical setting. Further studies in human tissues are required to verify the study results; therefore, the generality of the antitumor effects of IL-33 remains to be determined. Another issue is the possibility of adverse effects of IL-33, especially on proinflammatory or proallergic cells such as innate lymphoid cells. Despite these limitations and challenges, IL-33 is an attractive candidate for DC-activating therapy in antitumor immunity. Sema4A is also a candidate as CTL stimulator in cancer immunotherapy and a biomarker of DC activation that has applications in translational research on cancer vaccines and DC adjuvants. Further investigations are required to elucidate the exact mechanisms underlying Sema4A induction in DCs and the resultant CTL activation.
We thank Editage (www.editage.com) for English language editing.
This work was supported by the Center of Innovation Program, Ministry of Education, Culture, Sports, Science and Technology of Japan (COISTREAM [to A.K.]), Japan Society for the Promotion of Science KAKENHI grants (JP18H05282 [to A.K.] and 19K08602 [to I.N.]), Japan Agency for Medical Research and Development grants (J200705023, J200705710, J200705049, JP18cm016335, and JP18cm059042 [to A.K.]), Mitsubishi Foundation grant (to A.K.), and a grant from the Kansai Economic Federation (to A.K.).
Conception and design: Y.S., I.N., Y.K., and A.K. Development of methodology: Y.S., I.N., Y.K., S.K., D.O, A.O., Y. Naito, H.T., M.N., S.N., D.I., T.T., T.N., Y. Nakanishi, Y.F., T.K., S.S., K.M., K.F., T.S., K.I., H.H., Y.T., and A.K. Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y.S., I.N., Y.K., S.K., A.O., Y. Naito, H.T., M.N., T.T., T.K., Y.H., K.M., Y.T., and A.K. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, and computational analysis): Y.S., I.N., and A.K. Writing, review, and/or revision of the manuscript: Y.S., I.N., and A.K. Administrative, technical, or material support (i.e., reporting or organizing data and constructing databases): Y.S., I.N., and A.K. Revision of article: A.K.
The microarray data presented in this article have been submitted to the National Center for Biotechnology Information’s Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo) under accession number GSE150110.
The online version of this article contains supplemental material.
Abbreviations used in this article
immune checkpoint inhibitor
Lewis lung carcinoma
- MHC II
MHC class II
type 1 CD8+ CTL
The authors have no financial conflicts of interest.