Host IL11 Signaling Suppresses CD4⁺ T cell–Mediated Antitumor Responses to Colon Cancer in Mice

Colon cancer remains as one of the leading causes of cancer-related death globally and frequently arises from a combination of genetic determinants and environmental factors (1). Although immunotherapy has shown unprecedented clinical success as a treatment for mismatch repair–deficient colon cancer (2), the much more common mismatch repair–proficient colon cancer is largely unresponsive to immune checkpoint inhibition (3). In patients who have mismatch repair–proficient colon cancer, several limiting factors contribute to poor responses toward immunotherapy including insufficient, or loss of, tumor antigen presentation via MHC, absence of tumor-infiltrating lymphocytes (TIL) associated with the presence of immunosuppressive cytokines, and other factors (4). Because of their soluble nature, cytokines can be readily exploited for cancer therapies including for the purpose of enhancing immune-mediated antitumor responses and to overcome poor patient responses toward immune checkpoint inhibition (5).

IL11 is a member of the IL6 family of cytokines that plays roles in thrombopoiesis (6), erythropoiesis (7), megakaryocyte development (8), decidua formation (9, 10), tissue fibrosis (11), and cancer (12). IL11 signaling occurs when the ligand simultaneously binds to its cognate ligand-binding receptor subunit IL11RA and the glycoprotein 130 (GP130) signal-transducing receptor subunit that is shared with the receptor for IL6 (13). Like IL6, IL11-mediated intracellular signaling converges on the transcription factor STAT3 and less prominent pathways including the RAS/ERK and PI3K pathways. Tumor intrinsic and extrinsic STAT3 activity contributes to many hallmarks of cancer, providing a fitness advantage for tumor cells while concurrently suppressing immune-mediated antitumor responses (13).

Tumors share hallmarks of unresolved wounds characterized by a subclinical, “para inflammatory” state (14) that fuels tumor growth and simultaneously dampens the immune response in the tumor environment (15). IL6 is well understood as a major regulator of wound-healing and anti-inflammatory responses (16), acting in concert with TGFβ to favor the development of “regenerative” CD4⁺ T helper 17 (Th17) cells (16). Meanwhile, IL11 is emerging as a cytokine with biologically similar, albeit more localized, function during wound healing and ensuing tissue repair (11, 17, 18) that favors neoplastic epithelial cell growth in gastrointestinal cancers (19, 20). However, it remains unclear whether IL11 signaling also contributes to cancer-associated immune suppression, and specifically whether IL11 directly shapes T cell–dependent antitumor immune responses.

In this study, we describe a hitherto unrecognized immune suppressing role of IL11 during colon cancer development by discovering that genetic deficiency of IL11 signaling in the host inhibited tumor growth across numerous mouse models of colon cancer. We found that IL11 suppressed expression of the antitumor effector molecules IFNγ and TNFα specifically in CD4⁺ T cells, and that the antitumor benefit conferred by deficiency of IL11 responses in the host is abrogated upon CD4⁺ T-cell depletion. We correlated these functional observations from mouse models with findings in human colorectal cancer biopsies where elevated IL11RA⁺ expression in intratumoral CD4⁺ T cells corresponded with lower expression levels of Ifng. Collectively, our data ascribe an immunosuppressive role to IL11 and suggest interference with IL11 signaling may provide an attractive immunotherapeutic approach for the treatment of colon cancer.

Murine MC38 colon cancer cells were a gift from Fred Hollande (University of Melbourne, Melbourne, Victoria, Australia) in 2016 and CT26 colon cancer cell lines were a gift from Joan Heath at the Walter and Eliza Hall Institute (WEHI, Melbourne, Victoria, Australia) in 2015. Both cell lines were cultured in DMEM-F12 medium (catalog no. 11320082; Thermo Fisher Scientific) supplemented with 10% FCS (Moregate Biotech), 100 U/mL penicillin and 100 μg/mL streptomycin (catalog no. 15140122, Thermo Fisher Scientific), and 2 mmol/L GlutaMAX-1 (catalog no. 35050061; Thermo Fisher Scientific). Murine BA/F3 pro-B cells were transfected with 1 μg plasmid DNA (pCMV6-AC-IL11RA, catalog no. SC319416, Origene) using the Amaxa Nucleofector (Lonza) as described previously (21). BA/F3 cells were cultured in RPMI medium (catalog no. 61870036; Thermo Fisher Scientific) supplemented with 10% FCS, 100 U/mL penicillin and 100 μg/mL streptomycin, and 10 ng/mL IL3 (catalog no. 213-13, Peprotech). The cells were grown at 37°C in a humidified atmosphere of 5% CO₂ and were passaged once every 3 to 4 days up to a maximum of eight passages. MC38 and BA/F3 cell lines were Mycoplasma tested in 2016. CT26 cells were Mycoplasma tested in 2015. Cell lines were not authenticated.

All animal experiments were conducted in accordance with, and with the approval of, the Animal Ethics Committee of Austin Health and La Trobe University (Victoria, Australia). Homozygous Il11ra^−/− C57BL/6 mice were generated as described previously (22). Homozygous Il11ra^−/− BALB/c mice were generated by backcrossing BALB/c wild-type (WT) mice with Il11ra^−/− C57BL/6 mice for 12 generations. WT C57BL/6 and BALB/c mice, Rag1^−/− C57BL/6 mice, and Ly5.1 C57BL/6 mice were sourced from WEHI (Melbourne, Victoria, Australia). All mice were housed under specific pathogen–free conditions and 6- to 8-week-old mice were used for all experiments.

For the induced sporadic colon cancer model, mice received intraperitoneal injections of the alkylating agent azoxymethane (10 mg/kg; catalog no. A5486, Sigma) once a week for 6 consecutive weeks (12). To longitudinally monitor the growth of tumors in the distal colon of mice, we performed endoscopy (KARL STORZ-ENDOSKOPE) once a fortnight for 14 weeks following the last administration of azoxymethane as described previously (12).

Syngeneic allografts were established using MC38 (1 × 10⁶) or CT26 (5 × 10⁵) cells injected subcutaneously in 100 μL of PBS (catalog no. 14190144; Thermo Fisher Scientific). For in vivo T-cell depletion experiments, mice were administered 200 μg of CD4- or CD8-depleting antibodies (clones GK1.1 and YTS169, respectively) or rat IgG2b isotype control (clone LTF-2; catalog no. BP0090; BioXCell) intraperitoneally 1 day before MC38 cells were engrafted subcutaneously followed by injections once every 3 days until the ethical endpoint was reached. All depleting antibodies were a gift from Louis Boon (Polpharma Biologics).

Bone marrow chimeras were generated using 6-week-old congenic Ly5.1 recipient WT mice, which were lethally irradiated (2 × 4.5 Gy) and injected with bone marrow cells (5 × 10⁶ cells) from either WT or Il11ra^−/− donor Ly5.2 mice via tail vein injection. The animals were allowed to recover for 2 months of reconstitution before MC38 subcutaneous allografts were established and grown as described above.

Whole spleens were placed in FACs buffer (2.5% FCS in PBS), minced, and filtered through a 70 μm cell strainer using a syringe plunger. The cell suspension was then pelleted, lysed in red cell lysis buffer (150 mmol/L NH₄Cl, 10 mmol/L KHCO₃, 0.1 mmol/L EDTA; WEHI) for 5 minutes on ice, and washed in FACs buffer (21). Allograft tumors were minced and incubated in digestion buffer [2.5% FCS, 1 mg/mL collagenase III (catalog no. 11097113001; Sigma), 2 U/mL DNAse I (catalog no. 10104159001; Sigma) in Hank's Balanced Salt Solution (catalog no. 14170161; Thermo Fisher Scientific)] at 37°C for 30 minutes on a rotating shaker. The released cell suspensions were then filtered through a 70 μm cell strainer, pelleted and washed in FACs buffer. Single-cell suspensions were stained with fluorophore-conjugated primary antibodies diluted in FACs buffer for 25 minutes on ice in the presence of 1% Fc blocking solution (catalog no. 14016186; Thermo Fisher Scientific). The following fluorochrome-conjugated antibodies were used for flow cytometric analysis: CD45.2 [clone S450-15-2; WEHI, Melbourne, Australia (23)], TCRβ (clone H57-597; Thermo Fisher Scientific), CD4 (clone GK1.5; eBioscience), CD8 (clone 53-6.7; BD Biosciences), FOXP3 (clone FJK-16S; BioLegend), IFNγ (clone XMG1.2; BioLegend), granzyme B (clone GB11; BioLegend), CD19 (clone eBio1D3; BioLegend), CD11b (clone M1.70; BD Biosciences), Ly6C (clone HK1.4; eBioscience), Ly6G (clone 1A8; BD Pharmingen), F4/80 (clone BM8; eBioscience), and MHCII (clone M5/114.15.2; Miltenyi Biotec). Dead cells were stained with either SYTOX Blue Dead Cell Stain (catalog no. S11348; Thermo Fisher Scientific) or Fixable Viability Dye (catalog no. 65086618; eBioscience), and excluded from the analysis using a BD FACSCanto II flow cytometer (BD Biosciences). Data analysis was performed using FlowJo V10 (BD Biosciences). Cell sorting was performed using the BD FACSAria III cell sorter (BD Biosciences).

Total RNA was isolated from whole tissues or cultured cells using the Qiagen RNeasy Plus Mini Kit (catalog no. 74134; Qiagen) and 1,000 ng of RNA was reverse transcribed into cDNA using SensiFast cDNA Reverse Transcription Kit (catalog no. QT605; Bioline) as per manufacturer's instructions. Total RNA isolated from FACs sorted lymphocytes was purified using the Qiagen RNeasy Plus Micro Kit (catalog no. 74004; Qiagen) and 20 ng of RNA was reverse transcribed into cDNA using the SuperScript IV First-Strand Synthesis System (catalog no. 8090050; Thermo Fisher Scientific) as per manufacturer's instructions.

qRT-PCR analysis was performed on biological triplicates using SensiFAST SYBR Kit (catalog no. QT605-20; Bioline) and the Viia7 Real-Time PCR system (Thermo Fisher Scientific). Expression data were normalized against HPRT gene expression using the 2^–ΔΔC_t method (24). Primers were sourced from Integrated DNA Technologies and the sequences are as follows: mHprt (5′-primer, GGG GGC TAT AAG TTC; 3′-primer, TCC AAC ACT TCG AGA), mIl11 (5′-primer, TGT TCT CCT AAC CCG; 3′-primer, CAG GAA GCT GCA AAG ATC CCA), mIl11ra (5′-primer, TGG AAG TCC ACC TGA GGA ATG TGT; 3′-primer, AGA CCG CAC ACA CTC TCC AAT CAT), mGzmb (5′-primer, ACT CTT GAC GCT GGG ACC TA; 3′-primer, AGT GGG GCT TGA CTT CAT GT), mPrf1 (5′-primer, TGC TAC ACT GCC ACT CGG TCA; 3′-primer, TTG GCT ACC TTG GAG TGG GAG), mIfng (5′-primer, TCA AGT GGC ATA GAT GTG GAA GAA; 3′-primer, TGG CTC TGC AGG ATT TTC ATG), mTnfa (5′-primer, ACC CTC ACA CTC AGA TCA TC; 3′-primer, GAG TAG ACA AGG TAC AAC CC).

Cytokine analysis of cell culture supernatants was performed in technical duplicates using the Mouse 6-plex Th1/2 Luminex kit. This analysis was performed by Crux Biolabs.

Formalin-fixed paraffin-embedded (FFPE) blocks were cut into 4-μm–thick sections onto SuperFrost Plus microscope slides, which were then heated in a microwave pressure cooker with 0.1% citrate buffer for 15 minutes. Sections were treated with 3% hydrogen peroxide for 20 minutes at room temperature. Tissue sections were blocked in 10% [volume for volume (v/v)] goat serum for 1 hour at room temperature in a humidified chamber. Sections were incubated with primary antibodies diluted in 10% (v/v) goat serum (catalog no. PCN5000; Thermo Fisher Scientific) in a humidified chamber overnight at 4°C. Sections were stained with antibodies against CD4 (Clone 4SM95; catalog no. 14976682; Thermo Fisher Scientific), CD8 (Clone 4B11; catalog no. MA180231; Thermo Fisher Scientific), and FOXP3 (Clone FJK-16s; catalog no. 14577382; Thermo Fisher Scientific). The sections were then incubated in biotinylated secondary antibodies (Avidin Biotin Complex ABC-kit; catalog no. VEPK6100; Vector Laboratories) as per manufacturer's instructions, counterstained with hematoxylin and mounted using DPX mounting solution (catalog no. 06522; Sigma).

Tissue or cells were lysed in RIPA buffer containing protease (catalog no. 11873580001; Sigma) and phosphatase inhibitors (catalog no. 4906837001; Sigma), clarified by centrifugation (13,000 × g for 15 minutes at 4°C), and protein concentrations were measured with the Pierce BCA Protein Assay Kit (catalog no. 23225; Thermo Fisher Scientific). Protein lysates (30 μg) were subjected to Western blotting and immunoreactive bands were visualized using the Odyssey Imaging system (Li-COR Biosciences) for anti-mouse pTyr705 STAT3 (Clone D3A7; catalog no. CS9145) and total STAT3 (Clone 79D7; catalog no. CS4904).

Multiplex IHC using the Opal 7-color IHC kit (catalog no. NEL811001KT; Akoya Biosciences) was performed on FFPE MC38 tumors using an optimized serial protocol. FFPE tissue sections were heated in a microwave with Tris-EDTA pH 9.0 (10 nmol/L Tris Base, catalog no. BIO3094T, Astral; 1 mmol/L EDTA, catalog no. 798681, Sigma) for 1 minute at 100% power followed by 10 minutes at 10% power. Slides were treated with PeroxAbolish (catalog no. PXA969; Biocare Medical) for 30 minutes at room temperature and blocked with Opal antibody blocking solution. Slides were then serially stained with CD4 (Clone 4SM95; catalog no. 14976682; Thermo Fisher Scientific), CD8 (Clone 4SM15; catalog no. 140808; Thermo Fisher Scientific), and pTyr705 STAT3 (Clone D3A7; catalog no. CS9145) followed by Opal TSA amplification and DAPI counterstain. Stained slides were imaged using the Vectra 3 automated quantitative pathology imaging system 3.0.5 (Akoya Biosciences) and analyzed using the inform 2.4.1 software (PerkinElmer).

Tissue microarrays (TMA) were prepared from a cohort of 182 patients with stage III colorectal cancer diagnosed between 2001 and 2015 at Austin Health (Victoria, Australia). Resected tumors included those from the proximal colon, distal colon, and rectum, and fixed in 10% neutral buffered formalin prior to making FFPE blocks. The FFPE blocks were stored in archives at room temperature prior to retrieval for TMA creation. The original diagnostic hematoxylin and eosin slides were reviewed by a pathologist (David Williams) and archived FFPE tumor blocks were retrieved. TMAs were prepared using a manual tissue arrayer (MTA-II, Beecher Instruments) including three 1-mm–diameter cores per tumor, randomly sampled from central regions of tumor. This study was approved by the Austin Health Human Research Ethics Committee (H2013/05077) with a waiver of consent and conducted in accordance with the declaration of Helsinki.

TMAs constructed from histologically annotated primary human colorectal tumors, were sectioned at 4 μm onto SuperFrost Plus microscope slides and “baked” for 1 hour at 60°C. Singleplex RNAscope (catalog no. 322370; ACDBio) was performed on TMAs as per manufacturer's instructions using the following paired oligonucleotide probes: Hs-IL11 (catalog no. 425289; ACDBio; NM_001267718.1; bp target region 131–1799 spanning exons 1–4) and Hs-IL11RA (catalog no. 568311; ACDBio; NM_001142784.2; bp target region 44–1177; spanning exons 1–11). We used FFPE BA/F3 cells engineered to overexpress the human IL11RA gene [hIL-11RA^OE (21)] as positive controls. FFPE blocks from parental BA/F3 cells and probes against the bacterial DapB gene served as negative controls.

Multiplex RNAscope was performed as per manufacturer's instructions using the following oligonucleotide probes: Hs-IL11RA (catalog no. 568311; ACDBio; NM_001142784.2, bp target region 131–1799 spanning exons 1–4), Hs-IFNG (catalog no. 310501-C2; ACDBio; NM_000619.2, bp target region 80–1152 spanning exons 1–4), and Hs-CD4 (catalog no. 605601; ACDBio; NM_000616, bp target region 1726–2734 spanning exon 10). Following RNAscope, probes were visualized using the TSA Plus System Kits (catalog no. NEL741001KT; PerkinElmer; fluorescein, cyanine 3, and cyanine 5). Stained slides were imaged using the Vectra slide analysis system and analyzed using the inForm software (PerkinElmer). Multiplex analysis was performed on a total of 174 tumor cores from 58 patients. Low-quality sectioned and/or stained sections were excluded from analysis.

IL11 and IL11RA gene expression was accessed from The Cancer Genome Atlas (TCGA) colon adenocarcinoma (COAD) published studies (GSE12945, GSE17536, and GSE41258) via the UCSC Xena platform (http://xena.ucsc.edu/; ref. 25). The TCGA COAD gene expression data were measured with the Illumina HiSeq 2000 RNA platform (IlluminaHiSeq_RNASeqV2). The expression data are representative of a total of 41 normal and 286 primary tumor tissue samples. The dataset displays the gene level transcription estimates in log2(x+1) transformed RSEM normalized count.

Single-cell RNA sequencing (scRNA-seq) gene expression profiles of Il11, Il11ra, Ifng, and Tnfa in FACS-sorted murine splenic CD4⁺ T cells was accessed from Th-Express (https://th-express.org; ref. 26). When comparing the expression profiles of IL11RA, IFNG, and FOXP3 in intratumoral CD4⁺ T cells from human colon cancers, we used scRNA-seq gene expression data that had been previously published and deposited to NCBI's Gene Expression Omnibus database with accession number GSE146771 (27). For this study, we selected the dataset generated using the SMART-seq2 platform which consisted of already normalized gene expression values (TPM) for human colon cancer cells together with their corresponding metadata. The metadata included information on previously identified t-SNE cell clusters and x-y coordinates as well as cell type annotation. The data were first imported into R as a Seurat (28) object after which CD4⁺ T cells were identified on the basis of the cell-type annotation information. t-SNE plots were generated using t-SNE coordinates for each cell type included in the metadata. To compare the differences in the gene expression profile of IL11RA, IFNG, and FOXP3 across different types of CD4⁺ T cells, TPM expression values for IL11RA, IFNG, and FOXP3 were each mean centered across cells within a cluster and scaled across clusters, these were then compared by way of a heatmap.

CD4⁺ T cells were grouped into four different cell types by using a set of known marker genes for each cell type; cytotoxic (GZMB), exhausted (LAG3, TOX), Th17 cells (IL23R) or regulatory T cells (Treg; FOXP3, IL2RA). To annotate cells, we first established a threshold for determining cells that can be said to express a particular marker gene for each of the cell types by looking at the expression profile of each marker gene in all CD4⁺ T cells. A cell was annotated to a cell type if the marker gene(s) for that cell type in that cell were expressed above the established threshold. Cells that could not be uniquely assigned to a cell type were excluded. Student t test was used to test for the statistical significance of the differences in the expression profiles of each gene between clusters.

All animal studies were performed at least twice and all ex vivo experiments were performed at least three independent times. The data are presented as mean ± SEM. Differences between two groups were analyzed by Student t test. For multiple comparisons, statistical analysis was performed using a one-way ANOVA followed by post hoc Tukey analysis. A P < 0.05 was considered to be statistically significant. Statistical analysis was performed using GraphPad Prism 7.0 (GraphPad Software).

We previously reported that a lack of IL11 signaling inhibits the growth of colonic tumors in mice in situations of overt inflammation, consistent with a role for IL11 and IL6 in exacerbating tumor-promoting inflammation (29). Here, we assessed whether IL11 signaling had a similar role in the more common situation of sporadic tumor development in the absence of colitis. Accordingly, we challenged WT and Il11ra^−/− mice for 6 consecutive weeks with the colon-specific alkylating agent azoxymethane (Fig. 1A). Our longitudinal assessment of colon tumor development using a miniaturized endoscope revealed significantly lower numbers of tumors in the Il11ra^−/− cohort compared with WT mice (Fig. 1B and C). These observations were confirmed upon autopsy of the mice (Fig. 1D–F). Consistent with our endoscopic assessment, histologic analysis of whole mount colons from Il11ra^−/− mice revealed a relative over-representation of smaller tumors at the expense of larger tumors when compared with WT mice (Fig. 1F). Collectively, these findings indicate that inhibition of IL11 signaling attenuates intestinal tumor development even in the absence of colitis or overt tumor-promoting and excessive STAT3-mediated inflammation.

There has been a large focus on understanding the effects that IL11 exerts on tumor cells, but it remains unclear whether IL11 also enables tumor promotion via the hematopoietic system. To investigate this, we established syngeneic allograft models of colon cancer in WT and Il11ra^−/− mice. To do this, we transplanted MC38 cells, which harbor a high tumor mutational burden akin to DNA mismatch repair–deficient cancers (30), in C57BL/6 hosts and transplanted CT26 cells in BALB/c hosts, which resemble low tumor mutational burden similar to DNA mismatch repair–proficient colon cancers (31). In comparison with WT hosts, we observed that the growth of both MC38 and CT26 allografts was impaired in IL11-unresponsive Il11ra^−/− hosts (Fig. 2A and B). Collectively, these results suggest that IL11 signaling in the host promotes tumorigenesis irrespective of the tumor mutational burden and the underlying immune bias that distinguishes C57BL/6 from BALB/c mice (32).

To characterize how the IL11 response of the host impacts tumor-infiltrating immune cells, we used flow cytometry to immunophenotype MC38 and CT26 tumors from Il11ra^−/− and WT mice (Supplementary Fig. S1). Lack of IL11 signaling within hosts harboring MC38 tumors did not alter the overall distribution of tumor-infiltrating immune cells including CD8⁺ T cells, CD4⁺ T cells, FOXP3⁺ CD4⁺ T cells, CD19⁺ B cells, Ly6c^high Ly6g^low, and Ly6c^low Ly6g^high myeloid cells or macrophages between Il11ra^−/− and WT hosts (Fig. 2C). We made similar observations in the CT26 mouse model, however, we observed a higher frequency of Ly6c^low Ly6g^high myeloid cells in the CD45⁺ immune cell compartment in Il11ra^−/− hosts than in WT hosts (Supplementary Fig. S2A). Although these findings suggest that IL11 signaling does not affect the relative composition of the tumor immune environment, we observed a striking reduction in the absolute numbers of the most abundant effector cells (i.e., CD8⁺, CD4⁺, and FOXP3⁺ T cells) when normalized against the net weight of MC38 tumors excised from Il11ra^−/− hosts (Fig. 2D). The expression of Il11 and Il11ra remained similar in MC38 and CT26 tumors irrespective of whether they grew in Il11ra^−/− or WT hosts (Supplementary Fig. S2B and S2C). Collectively, our results suggest that the lack of IL11 response in the host is sufficient to compromise the growth of tumors in endogenous and allograft colon cancer models in vivo and that this may be attributed to altered immune cell activity rather than a change in their frequencies.

Because the lack of IL11 responsiveness in the host was inconsequential to the distribution of intratumoral immune cells, we hypothesized that IL11 signaling could suppress the activity and other antitumor properties of immune cells, and collectively this could impair tumor growth. We focused on the two most abundant effector cell types, CD4⁺ and CD8⁺ T cells. Intratumoural T cells were activated with PMA/Ionomycin ex vivo for 4 hours and intracellular staining was performed. Il11ra-deficient and WT CD8⁺ T cells produced comparable levels of granzyme B (Fig. 3A). However, CD4⁺ T cells lacking IL11RA produced higher levels of IFNγ following stimulation compared with WT CD4⁺ T cells (Fig. 3B). We then purified CD8⁺ and CD4⁺ T cells by FACS and analyzed the expression of representative effector molecules by qRT-PCR. Levels of Gzmb, Ifng and Tnfa mRNA were similar in CD8⁺ T cells from Il11ra^−/− hosts and IL11RA-proficient WT hosts, but there was a 50% increase in Prf1 expression in cells derived from Il11ra^−/− host (Fig. 3C). In contrast, CD4⁺ T cells from Il11ra^−/− hosts expressed significantly higher mRNA levels of Ifng and Tnfa than their WT counterparts (Fig. 3D), and such differences were not be observed in tumor-associated CD8⁺ T cells (Fig. 3C). Collectively, these findings indicate that IL11 signaling in the host may promote tumor development primarily through a CD4⁺ rather than CD8⁺ T cell–mediated mechanism.

Given that IL11 can modulate Ifng and Tnfa expression in tumor-associated CD4⁺ T cells (Fig. 3D), we next studied whether IL11 signaling could dampen cytokine production by these cells. We first confirmed that CD4⁺ and CD8⁺ T cells express Il11ra mRNA (Fig. 3E) and that receptor stimulation by IL11 induces intracellular signaling in vitro, as detected by STAT3 phosphorylation (Fig. 3F). Multiplex IHC analysis confirmed the presence of nuclear pSTAT3-positive CD4⁺ and CD8⁺ T cells in MC38 tumors grown in Il11ra^−/− and WT hosts (Supplementary Fig. S3A). We did not observe any significant differences in the frequency of pSTAT3⁺ cells, pSTAT3⁺ CD4⁺ cells or pSTAT3⁺ CD8⁺ cells in MC38 tumors from Il11ra^−/− hosts compared with WT mice (Supplementary Fig. S3B). We then activated splenic CD4⁺ T cells with PMA/Ionomycin for 4 hours and analyzed the conditioned media by Multiplex ELISA. PMA/Ionomycin treatment upregulated protein levels of Th1 cytokines (IFNγ, TNFα, and the IL12-specific p70 protein subunit) and the Th2 cytokine IL4, which was potently inhibited when cells were concurrently treated with IL11 (Fig. 3G). In contrast, IL11 was unable to inhibit PMA/Ionomycin-induced IFNγ and TNFα production by CD8⁺ T cells, indicating a CD4⁺ T cell–specific effect (Fig. 3H). Likewise, IL11 was unable to suppress PMA/Ionomycin-induced cytokine production in splenic CD4⁺ T cells derived from Il11ra^−/− mice, indicating that IL11 attenuated the cytokine response through a canonical IL11RA-dependent pathway (Supplementary Fig. S3C). Single-cell genomic screening (26) validated the detection of Il11, Il11ra, Ifng, and Tnfa in CD4⁺ murine T cells across Th1, Th2, Th17, and Tregs (Supplementary Fig. S3D). Collectively, our observations suggest that IL11 signaling suppresses the production of cytokines and effector molecules by CD4⁺ T cells, and this supports immune evasion by tumor cells.

Given that emerging literature indicates that IL11 can elicit non-hematopoietic stromal responses, we set out to substantiate our above observation as being dependent on IL11-responsive immune cells by generating bone marrow chimeras. Following adoptive transfer of bone marrow from either WT or Il11ra^−/− mice to lethally irradiated WT mice, we inoculated the mice with MC38 cancer cell allografts. We found that the lack of IL11RA-mediated signaling in the hematopoietic compartment of recipients (Il11ra^−/− → WT) significantly reduced tumor burden compared with hosts reconstituted with IL11RA-proficient hematopoietic cells (WT → WT; Fig. 4A). We then formally determined the contribution of lymphocytes in controlling the growth of MC38 tumors in the absence of IL11 signaling and confirmed observations by others that MC38 allografts show accelerated growth in Rag1^−/− hosts (Fig. 4B; ref. 33). We next assessed the individual contribution of CD4⁺ and CD8⁺ cells by CD4 and CD8 antibody–mediated cell ablation in Il11ra^−/− hosts (Fig. 4C). CD4⁺ T-cell ablation reduced the proportion of tumor-infiltrating CD4⁺ T cells to approximately half of that observed in tumor-bearing Il11ra^−/− mice treated with isotype control (Fig. 4D and E). Importantly, CD4⁺ T-cell ablation reversed the reduced tumor burden we observed in Il11ra^−/− hosts (Fig. 4G). Meanwhile, antibody-mediated CD8⁺ T-cell depletion from MC38 tumor-bearing Il11ra^−/− hosts also reduced the frequencies of tumor-infiltrating CD8⁺ T cells (Fig. 4D) by approximately 50% (Fig. 4F). However, the latter treatment did not alter tumor burden (Fig. 4G). Collectively, these data suggest that CD4⁺ rather than CD8⁺ T cells are the major effector cell type that mediates antitumor immunity in the absence of canonical IL11 signaling in the host.

Emerging evidence posits IL11 signaling in the tumor microenvironment as an enabler of gastrointestinal cancer in murine models (20). However, comprehensive characterizations of IL11 and IL11RA RNA expression across large cohorts of human patient samples remain scarce. We therefore interrogated TCGA colon adenocarcinoma patient datasets (GSE12945, GSE17536, and GSE41258) and found significantly elevated IL11 expression in primary tumors compared with normal tissue (Fig. 5A). In contrast, IL11RA expression was lower in tumors when compared with normal tissue (Fig. 5A). To gain a better spatial understanding of tumor-associated cells expressing this ligand/receptor pair, we used RNAscope to analyze IL11 and IL11RA expression in tissue microarrays of 182 human colorectal cancer resections. We first confirmed IL11 (Fig. 5B) and IL11RA (Fig. 5C) RNA expression in neoplastic colon epithelial cells (NEC) and fibroblasts (CAF). We detected IL11RA expression in TILs (Fig. 5B), and this coincided with the expression of IL11, CD4, and IFNG in these biopsies (Fig. 5D). To correlate these observations with the functional insights gained from our studies using preclinical models, we stratified individual biopsies based on the level of IL11RA expression and observed that this had no impact on the distribution of CD4 expression (Fig. 5E). However, both total IFNG expression, and IFNG and CD4 coexpression were lower in IL11RA^hi tumors compared with IL11RA^lo tumors (Fig. 5F). Single-cell genomic screening of tumors from patients with colorectal cancer (27) revealed that IL11RA expression was detected in epithelial cells, fibroblasts, and CD8⁺ and CD4⁺ T cells, which corroborated our RNAscope data (Fig. 6A). Moreover, we detected IL11RA and IFNG expression across all subtypes/clusters of CD4⁺ T cells (Fig. 6B–D). Although the hT11_CD4-CTLA4 cluster did not contain cells with the highest expression of IL11RA, these FOXP3 and CTLA4 expressing cells constitute a major subset of IL11RA-expressing cells (Fig. 6B–D). In addition, we generated a heatmap of CD4⁺ T-cell clusters using our curated gene signatures to visualize IL11RA and IFNG expression in cytotoxic T cells (GZMB), exhausted T cells (LAG3, TOX), Th17 (IL23R), and Tregs (FOXP3, IL2RA; Supplementary Fig. S4A). We observed a marked increase in IL11RA expression in Tregs compared with Th17 cells (Supplementary Fig. S4B). In addition, IFNG expression was significantly higher in cytotoxic and exhausted CD4⁺ T cells compared with Th17 and Tregs (Supplementary Fig. S4B). Collectively, our observations in patient samples support the concept that patients with low IL11 signaling may show more efficient antitumor immune responses and that this may be achieved by therapeutic interference with IL11 signaling in the tumor environment.

Chronic inflammation propagated by dysregulated cytokine signaling is now a recognized feature in colon cancer and most solid malignancies, and is a major driver of tumorigenesis in various preclinical mouse models (34). Many cytokine-mediated signaling pathways converge on STAT3. Deregulated STAT3 signaling not only supports tumor cell survival, proliferation, and other cell-intrinsic hallmarks of cancer, but also indirectly supports tumorigenesis by dampening antitumor immune responses (15, 35). Accordingly, IL6 and IL10 family cytokines that act upstream of STAT3 are prototypical wound-healing cytokines and are inextricably linked with many human cancers (36, 37). Although the role of IL6 in T-cell biology is well defined, such roles for the related cytokine IL11 have not been thoroughly investigated (38, 39). In particular, such roles for IL11 have not been studied in the context of T cell–mediated antitumor immune responses.

Here, we attribute an immunomodulatory role to IL11 signaling in the tumor microenvironment, whereby IL11 suppresses the ability of CD4⁺ effector T cells to mount an effective antitumor response. Together with our previous findings, we propose that the “regenerative wound healing” properties of IL11 support colon cancer growth via two complementary mechanisms. First, by promoting survival and proliferation of epithelial cancer cells (12), and second by suppressing CD4⁺ T cell–mediated antitumor immune response. Notwithstanding additional roles of IL11 on cancer-associated fibroblasts (40) and possibly other stromal components, our working model aligns with the immune cell–intrinsic properties described for STAT3 as the major downstream effector molecule of intracellular IL11 signaling (41).

Our findings indicate that IL11 suppresses IFNγ and TNFα production by tumor-infiltrating CD4⁺ T cells. Although not addressed in this study, FOXP3⁺ Tregs may represent a subset of IFNγ-producing CD4⁺ T cells as Tregs have been reported to produce IFNγ (42). Because IL11 is produced by neoplastic epithelium (12), our observations suggest that IL11 is likely to elicit paracrine immunosuppressive effects on cytotoxic CD4⁺ T cells once they have infiltrated tumors. At present, we can not exclude additional immunomodulatory effects of IL11 given that dendritic cells, macrophages, γδ-T cells, natural killer (NK) cells, innate lymphoid cells, and B cells all express Il11ra. Importantly, IL11 inhibits LPS-induced expression of TNFα, IL1β, IL23p40, and NOS in tumor-associated macrophages, thereby skewing their polarization toward a tumorigenic, wound-healing, and immune-suppressive “M2-like” phenotype (43, 44). It remains to be tested whether IL11 signaling elicits additional indirect tumor-promoting effects by acting on dendritic cells and NK cells (41), given the known immunosuppressive outcomes ascribed to STAT3-dependent signaling in these cell types. However, because partial depletion of CD4⁺ T cells abolished the protective effects observed in Il11ra^−/− mice with MC38 allografts, we believe that IL11 primarily suppresses CD4⁺ effector T cells and mediates little or no effect on the protumor function of Tregs. Interestingly, our single-cell sequencing analysis revealed a cluster of IL11RA-expressing CTLA4-positive CD4 T cells, which most likely display a regulatory phenotype because they highly express FOXP3, which is consistent with previous observations of CTLA4-expressing Tregs (45, 46). It will be important to establish whether these findings also translate to the human equivalents of these CD4⁺ T-cell subsets. Our in vivo depletion experiments indicate that MC38 tumors may be more susceptible to killing by CD4⁺ T cells than CD8⁺ T cells. This is consistent with the observation that CD4⁺ T cells, besides recognizing antigens displayed on MHC class II, can also mount effective antitumor responses independent of MHC class I/II molecules and drive antitumor responses independently of CD8⁺ T cells (47–51). One caveat in our depletion model was that the limited efficiency of the CD4-depleting antibody prevented the in depth analysis required to tease out the contribution of IL11 on various CD4⁺ T-cell subsets including Th17 and Tregs.

We propose that the persisting “smouldering” subclinical inflammation characteristic of most solid tumors, elicits an IL11-driven response that is phylogenetically selected to maximize the gastrointestinal wound-healing response by simultaneously promoting epithelial cell survival and proliferation, as well as suppressing antiepithelial immune responses. Thus, reliance on the restricted expression of IL11RA implicates activation of STAT3 as the major orchestrator of regeneration that requires the coordinated response of epithelial and immune cells. Indeed, IL11RA expression transiently increases in epithelial progenitor cells following radiation-induced wound-healing response of the intestinal lining (52). Meanwhile, the limited expression of the membrane-associated IL6RA mutes the gastrointestinal lining to the frequent systemic emergency responses characterized by acute systemic increases in IL6. Instead, an IL6-dependent response of the intestinal epithelium (29) together with lamina propria T cells lacking the expression of the membrane-associated IL6RA (53), restricts where IL6 trans-signaling occurs. This inflammation-associated process depends on the proteolytic cleavage of the extracellular portion of IL6RA from other nearby cells by metalloproteinases (54) to allow formation of a soluble IL6/sIL6RA complex that can activate GP130 signaling in cells lacking expression of native IL6RA. Indeed, the survival and proliferation of malignant transformed gastrointestinal epithelium is augmented by IL11 and IL6 trans- but not IL6 conventional signaling (12, 29). Consistent with this, we detected IL11RA expression in malignant epithelium and lymphocytes across a large cohort of patient-derived TMAs, and consolidation of TCGA data reveals significantly elevated IL11 levels in tumor tissues when compared with matching normal tissues.

Many therapeutic approaches have been developed to target the JAK/STAT3 nexus for the treatment of solid cancers, but many have failed, particularly as single agents (13). In light of the collateral side effects observed with inhibition of JAK kinases and the difficulties of specifically targeting intracellular proteins such as STAT3, the prospect of inhibiting upstream soluble factors such as IL11 presents an attractive alternative. Although the therapeutic antibodies directed against IL6 (siltuximab) and IL6RA (tocilizumab) are now been approved for rheumatoid arthritis, Castleman disease and other chronic inflammatory diseases, these antibodies demonstrate only limited effectiveness in phase I–II trials for the treatment of advanced solid malignancies including colorectal cancer (55). Given our previous findings identifying IL11 signaling as a key driver of gastrointestinal cancer in murine models, therapeutic targeting of IL11 signaling may overcome the limited responses to IL6 inhibition in clinical trials for gastrointestinal cancers (56). Our findings provide proof-of-concept evidence that IL11 signaling confers an immunosuppressive and tumor cell–stimulating effect, thus supporting the application of IL11-targeting mAbs, which are currently benig developed to combat fibrotic diseases, to treat cancer (57). Taken together, our study warrants further evaluation of therapeutic inhibition of IL11 signaling to broaden antitumor immune responses for DNA mismatch repair–proficient colon cancer when combined with immune checkpoint inhibitors.

L. Boon reports other support from Polpharma Biologics outside the submitted work. A.L. Chand reports grants from Victorian Cancer Agency and Cancer Council Victoria during the conduct of the study, as well as other support from Lassen Therapeutics outside the submitted work. Lassen Therapeutics develops anti-IL11 receptor antibodies for treatment of cancer No disclosures were reported by the other authors.

J. Huynh: Conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing–original draft, writing–review and editing. D. Baloyan: Investigation. D. Chisanga: Formal analysis. W. Shi: Formal analysis. M. O'Brien: Investigation. S. Afshar-Sterle: Investigation. M. Alorro: Investigation. L. Pang: Investigation. D.S. Williams: Resources. A.C. Parslow: Investigation. P. Thilakasiri: Investigation. M.F. Eissmann: Investigation. L. Boon: Resources. F. Masson: Conceptualization. A.L. Chand: Conceptualization, supervision, writing–review and editing. M. Ernst: Conceptualization, resources, supervision, funding acquisition, writing–original draft, writing–review and editing.

M. Ernst and A.L. Chand are recipients of Investigator and Fellowship support from the National Health and Medical Research Council (NHMRC) of Australia and the Victorian Cancer Agency, respectively. This work was supported in part by project grant 1125951 (to J. Huynh) from the NHMRC.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.