Distinct roles for interleukin-23 receptor signaling in regulatory T cells in sporadic and inflammation-associated carcinogenesis

Abstract

Introduction: The pro-inflammatory cytokine interleukin-23 (IL-23) has been implicated in colorectal cancer (CRC). Yet, the cell-specific contributions of IL-23 receptor (IL-23R) signaling in CRC remain unknown. One of the cell types that highly expresses IL-23R are colonic regulatory T cells (Treg cells). The aim of this study was to define the contribution of Treg cell-specific IL-23R signaling in sporadic and inflammation-associated CRC.

Methods: In mice, the role of IL-23R in Treg cells in colitis-associated cancer (CAC) was investigated using azoxymethane/dextran sodium sulphate in wild-type Treg cell reporter mice (WT, Foxp3 ^YFP-iCre), and mice harboring a Treg cell-specific deletion of IL-23 (Il23r ^ΔTreg). The role of IL-23R signaling in Treg cells in sporadic CRC was examined utilizing orthotopic injection of the syngeneic colon cancer cell line MC-38 submucosally into the colon/rectum of mice. The function of macrophages was studied using clodronate. Finally, single-cell RNA-seq of a previously published dataset in human sporadic cancer was reanalyzed to corroborate these findings.

Results: In CAC, Il23r ^ΔTreg mice had increased tumor size and increased dysplasia compared to WT mice that was associated with decreased tumor-infiltrating macrophages. In the sporadic cancer model, Il23r ^ΔTreg mice had increased survival and decreased tumor size compared to WT mice. Additionally, MC-38 tumors of Il23r ^ΔTreg mice exhibited a higher frequency of pro-inflammatory macrophages and IL-17 producing CD4⁺ T cells. The decreased tumor size in Il23r ^ΔTreg mice was macrophage-dependent. These data suggest that loss of IL-23R signaling in Treg cells permits IL-17 production by CD4⁺ T cells that in turn promotes pro-inflammatory macrophages to clear tumors. Finally, analysis of TCGA data and single-cell RNA-seq analysis of a previously published dataset in human sporadic cancer, revealed that IL23R was highly expressed in CRC compared to other cancers and specifically in tumor-associated Treg cells.

Conclusion: Inflammation in colorectal carcinogenesis differs with respect to the contribution of IL-23R signaling in regulatory T cells.

Keywords: AOM-DSS; colorectal carcinoma; inflammation-associated cancer; interleukin-23; macrophages; orthotopic MC-38; regulatory T cells; sporadic cancer.

Figure 1

Cell-specific ablation of Il23r in Treg cells increases tumor volume and dysplasia in inflammation-associated carcinogenesis. (A) Schematic of chronic DSS model. (B) Composite colon histology score for mice on chronic DSS or water only control. (C) RT-qPCR of sorted colonic FOXP3⁺ cells from mice with IL-23R-sufficient Treg cells (Il23r ^flox/+ Foxp3 ^YFP-Cre ) or IL-23R-deficient Treg cells (Il23r ^ΔTreg) after chronic DSS or water only control. ANOVA with post-hoc Tukey. (D) Weight curves and (E) histology of WT (Foxp3 ^YFP-Cre) and Il23r ^ΔTreg mice during chronic DSS treatment. Data are representative of two independent experiments. (F) Schematic depicting mice AOM/DSS model of inflammation-associated carcinogenesis. (G) RT-qPCR for Il23p19 expression in tumor and adjacent normal tissue in wild-type mice. Paired t-test. Data are representative from two independent experiments. (H) Weight curves of wild-type and Il23r ^ΔTreg mice subjected to AOM/DSS. Data are representative of five independent experiments. (I) Kaplan-Meier survival curves of AOM/DSS-treated mice. Data are pooled from five independent experiments, N=52-57. (J) Representative endoscopic images of tumors induced by AOM/DSS. (K) Quantification of tumor number per mouse and volume per tumor. Data are pooled from three independent experiments. (L) H&E-stained colon section depicting colon tumor with (M) grade of tumor dysplasia assessed by a pathologist. The grade shown is the highest grade of dysplasia per mouse. Chi-square among mice that developed dysplasia/tumors: X² (2, N=56) = 6.78, p=0.03. Data are pooled from three independent experiments. (N) Treg cell frequency in spleen, mesenteric lymph node (MLN) and (O) tumor with flow cytometry plots on the left and quantification on the right. Data are pooled from three independent experiments. (N-O) Unpaired T-test. *p<0.05; **p<0.01; ***p<0.001.

Figure 2

Il23r ^ΔTreg inflammation-associated colorectal tumors exhibit a pro-inflammatory transcriptional profile. (A) RNA-seq performed on tumors and adjacent normal tissue from AOM/DSS treated mice with heatmap demonstrating significantly upregulated genes in tumors from WT mice that were not upregulated in tumors from Il23r ^ΔTreg mice. Genes of interest are annotated. Z-scores were calculated based on normalized counts. Red indicates high z-scores for upregulated genes while green indicates high z-scores for downregulated genes. (B) Gene set enrichment analysis (GSEA) using Webgestalt. Two comparisons are shown: (1) tumor vs. adjacent normal in WT mice (black circles); (2) tumor vs. adjacent normal in Il23r ^ΔTreg mice (red circles). (C) PERMANOVA was used to identify cell types which were significantly different between tumors and adjacent normal tissue in WT and Il23r ^ΔTreg mice. Shown are the cell type enrichment scores calculated with xCell, with the significance indicated by the result of the SIMPER test for the comparison Il23r ^ΔTreg tumor vs. WT tumor. Only cell types that differed in at least one comparison are shown. (D) Flow cytometry of tumoral myeloid cells or (E) tumoral CD4⁺ T cells stimulated with PMA/ionomycin and stained for intracellular cytokines. (D) Mann-Whitney U-test. Lineage positive cells (CD3⁺, CD19⁺, NK1.1⁺) cells were excluded. (E) ANOVA with Dunnett’s post hoc. Data are pooled from two independent experiments. *p<0.05; **p<0.01; ***p<0.001.

Figure 3

IL-23R-signaling in Treg cells protects against sporadic colorectal cancer in an orthotopic model. (A) MC-38 cells were submucosally injected into the colon of mice with representative photo of colon lumen pre- and post-injection. (B) MC-38 cells were submucosally or subcutaneously injected into Foxp3 ^RFP and Foxp3 ^RFP Il23r ^GFP/+ mice with data pooled from three independent experiments to assess for expression of IL-23R in Treg cells. (C) Kaplan-Meier survival curve for WT and Il23r ^ΔTreg mice submucosally injected with MC-38 cells (as in A). Mantel-Haenszel statistic. (D) Representative macroscopic image and (E) quantification of tumor volume per mouse across multiple independent experiments. Mann-Whitney U-test. (F) Representative flow cytometry plot of intratumoral macrophages (left) with quantification (right). Data are pooled from six independent experiments. Unpaired t-test. *p<0.05. (G) Kaplan-Meier survival curve for WT and Il23r ^ΔTreg mice submucosally injected with MC-38 cells with some mice being depleted of macrophages using clodronate while a separate cohort treated with control liposomes. Data are pooled from two independent experiments, N=4 (clodronate treated mice) and N=6 (control liposome treated mice).

Figure 4

Colonic Treg cells in human colon tumors express high levels of IL23R. (A) IL23R expression among various cancer types. (B) Transcriptional analysis of select genes in matched samples from patients with colorectal cancer obtained from TCGA. (C) RT-qPCR of the IL23 subunits in patients with colorectal cancer. ANOVA with post-hoc Tukey. * p<0.05 (D) Re-analysis of single cell RNA-seq dataset for IL23R expression in various immune cells. (19) Heatmap indicates z-scores for expression with red indicating higher expression. Cluster designations were implemented as described in the published manuscript.