Aryl Hydrocarbon Receptor Ligands Drive Pancreatic Cancer Initiation and Progression through Protumorigenic T-cell Polarization

Abstract

Although smoking is a risk factor for pancreatic adenocarcinoma (PDAC), the underlying mechanisms promoting tumorigenesis and progression are unknown. In this study, we show that aryl hydrocarbon receptor (AHR) ligands found in cigarette smoke, like the carcinogen 2,3,7,8-tetrachlorodibenzo-p-dioxin, promote pancreatic dysplasia and PDAC progression in a mouse model of this disease. This effect is mediated by AHR activation in CD4+ T cells, leading to their polarization to IL22-producing TH22 cells and regulatory T cell accumulation, ultimately driving a blunted CD8+ T-cell effector response. Analysis of human pancreata from organ donors revealed that smokers have increased AHR activation relative to nonsmokers. Similarly, PDAC tumors from patients with a history of cigarette smoking presented with increased regulatory T-cell accumulation compared with nonsmokers. These findings support a model whereby AHR ligands (AHRL) in cigarette smoke promote tumorigenesis and progression of PDAC through dysregulation of immune responses.

Significance: Our study investigates the mechanistic link between AHRL and pancreatic cancer. We determined that AHRLs polarize naïve T cells, resulting in increased production of IL22 and immunosuppression. Our findings identify a novel signaling axis linking environmental chemicals to pancreatic tumorigenesis via the immune system. See related commentary by Zhao and Hill, p. 13.

Figure 1.

Cigarette smoke and TCDD promote PDAC tumor progression through immune modulation. A, Experimental design, n = 7–9/cohort, of the orthotopic model with CSE. B and C, Tumor weight and volume (B) and RNA expression of CYP1A1 (C) after 19 days of CSE treatment. D and E, Representative H&E staining (D) and IHC for the AHR (E) of orthotopic tumors after 19 days of CSE treatment. F, Experimental design, n = 4 = 5/cohort, of the orthotopic model with TCDD. G and H, Tumor weight and volume (G) and mRNA expression of CYP1A1 (H) after 19 days of TCDD treatment. I, Representative H&E staining of orthotopic tumors after 19 days of TCDD treatment. J and K, Metastatic burden (J) in the TCDD orthotopic model after 19 days, with representative H&E staining of a liver metastasis (K). L, Trichrome staining of tumors after 19 days of vehicle or TCDD treatment and collagen quantification. M, Experimental design, n = 7–9/cohort, of the orthoptic model with TCDD in NSG mice. N and O, Tumor weight and volume (N) and CYP1A1 RNA expression (O) of orthotopic tumors in NSG mice after 19 days of TCDD. Data are represented as the mean ± SD unless otherwise noted. Scale bars, 100 μm for all H&E and IHC images. NSG, NOD scid gamma.

Figure 2.

Cigarette smoke and TCDD increase IL22 production in an AHR-dependent manner. A − C, In vitro naïve CD4⁺ T-cell polarization schema (A). CYP1A1 mRNA expression, IL22 mRNA expression, and IL22 protein concentration in culture supernatant after in vitro naïve CD4⁺ T-cell polarization with CSE (B and C, n = 3/group), 6-formylindolo(3,2-b)carbazole (FICZ) and TCDD (D, n = 2–3/group), and TCDD with AHR inhibitor CH-223191 (E, n = 3/group). F, Experimental schema of naïve CD4⁺ T-cell polarization with human splenocytes. G, CYP1B1 and IL22 mRNA expression and IL22 protein concentration in the culture supernatant after naïve CD4⁺ T-cell polarization with TCDD and CH-223191 (n = 3 technical replicates/group). H, Experimental schema, wherein the culture supernatant of polarized naïve CD4⁺ T cells was placed upon 7940b PDAC cells for 15 minutes, and protein was then isolated for Western blot. I, Western blot showing pSTAT3, total STAT3, and GAPDH expression among 7940b PDAC cells treated with unpolarized Th0 T-cell media, TCDD-polarized T-cell media, TCDD-polarized media with IL22-binding protein (IL22bp), TCDD alone, negative control, and IL22 alone. J, Schema of orthotopic tumor model with CSE or TCDD. K, IL22 mRNA expression in orthotopic tumors treated with CSE (n = 3–5/group) or TCDD (n = 4/group). Data represented as the mean ± SD unless otherwise noted (portion of the figure created with BioRender.com).

Figure 3.

TCDD induced PDAC growth is mediated by IL22 signaling and Treg infiltration. A, Genetic makeup of the IL22-tdTomato reporter mouse (Catch-22) and experimental schema showing 3-week treatment course with TCDD. B and C, Representative mfIHC (B) and quantification of IL22⁺ immune cells (C) in the duodenum of Catch-22 mice treated with vehicle or TCDD for 3 weeks, n = 3 mice/group, n = 25–26 regions of interest/group. D, Genetic makeup of the IL22 eGFP reporter mouse, with quantification of the proportion of T cells, IL22⁺ T cells, ILCs, and IL22⁺ ILCs among pancreata of mice treated with 3 weeks of vehicle or TCDD, n = 3/group. E and F, Experimental schema, n = 6–9/group, and tumor weight and volume for treatment of WT mice with TCDD, IL22^−/− mice with TCDD, and IL22^−/− mice with vehicle for 19 days. G and H, Representative mfIHC (G, scale bars, 50 μm, inlaid white arrows identifying CD3⁺ FOXP3⁺ Tregs), and quantification of T cell, Th cell, CD8⁺ T cell, and Treg infiltration via mfIHC (H), n = 4–5 mice/group, and quantification of 17 regions of interest/group. Phenotypes were identified as follows: T cells (CD3⁺), Th cells (CD3⁺ CD8⁻ FOXP3⁻), CD8⁺ T cells (CD3⁺ CD8⁺), and Tregs (CD3⁺ FOXP3⁺ CD8⁻). I, Quantification of T cells, CD4⁺ T cells, CD8⁺ T cells, and Tregs via flow cytometry of orthotopic tumors of WT mice treated with vehicle or TCDD, n = 3–4/group. J, Experimental design, n = 4–8/group, and tumor weight and volume of the orthotopic model of WT mice or FOXP3-^DTR mice, allowing for inducible depletion of Tregs, treated with vehicle or TCDD for 16 days. Groups: WT treated with vehicle, WT treated with TCDD, and FOXP3-^DTR treated with TCDD. K, Experimental schema and representative gating strategy on CD45⁺ CD3⁺ intratumoral T cells from orthotopic tumors of WT mice treated with vehicle or TCDD for 19 days. L, Quantification of IL22⁺ cells among CD4⁺ T cells, FOXP3⁺ cells among CD4⁺ T cells, and IL22⁺ cells among FOXP3⁺ Tregs. Data are represented as the mean ± SD unless otherwise noted. DAPI, 4′6-diamidino-2-phenylindole.

Figure 4.

TCDD-mediated IL22 production and Treg accumulation is dependent upon CD4⁺ T cell AHR signaling. A, CYP1A1 and CYP1B1 mRNA expression from CD4⁺ TILs isolated from orthotopic tumors of WT mice treated with vehicle or TCDD for 19 days. B, Genetic makeup of the CD4 Cre AHR mice with inducible CD4-specific deletion of AHR signaling. C, IL22 mRNA expression and protein concentration in the culture supernatant of naïve CD4⁺ T cells isolated from WT or CD4 Cre AHR mice and polarized under nonstimulating (Th0) conditions, with TCDD, or with TCDD and CH-223191, an AHR inhibitor. D − F, Experimental design (D), tumor weight and volume (E), and CYP1A1 and IL22 mRNA expression (F) from tumors of an orthotopic model utilizing WT mice and CD4-Cre AHR^−/− mice treated with vehicle or TCDD, n = 4–7 mice/group. G and H, Representative mfIHC (G, scale bars, 50 μm, inlaid white arrows identifying CD3⁺ CD8⁺ T cell), and quantification (H) of CD8⁺ T cell (CD3⁺ CD8⁺ FOXP3⁻) and Treg (CD3⁺ FOXP3⁺ CD8⁻) infiltration among orthotopic tumors of WT or CD4 Cre AHR mice treated with vehicle or TCDD, n = 17–39 regions of interest/group. I, Experimental design and (J) tumor weight and volume for WT mice treated with vehicle, AHR inhibitor CH-223191, TCDD, or TCDD and the AHR inhibitor for 19 days, n = 7–10/group. Data represented as the mean ± SD unless otherwise noted. 4′6-diamidino-2-phenylindole.

Figure 5.

TCDD induces CD4⁺ T-cell AHR activation and decreases CD8⁺ T-cell effector function. A, UMAP visualization of scRNA-seq data set of orthotopic tumors of WT mice after 19 days of treatment with vehicle or TCDD, split by treatment group, n = 2 tumors/group. B, UMAP visualization of T cells from the orthotopic tumor scRNA-seq data set split by treatment group, vehicle or TCDD, n = 2 tumors/group. C, Histogram depicting the relative frequency of each T-cell population among vehicle- and TCDD-treated orthotopic tumors. D, Gene set enrichment analysis (GSEA) signatures of T-cell exhaustion, activation, memory, and naïvety relative to treatment with vehicle or TCDD. E, Dot plot representation of relative expression of markers of AHR activation (Cyp1a1, Cyp1b1, Ahrr, and Tiparp) among the CD4⁺ T-cell populations from the orthotopic tumor scRNA-seq data set, split by treatment group, vehicle or TCDD. F, Heat map visualization of gene expression of markers of effector function and exhaustion among CD8⁺ T cells from the orthotopic tumor scRNA-seq data set, shown by individual tumors, n = 2 vehicle and n = 2 TCDD. G, Flow cytometry of TIM3 and PD-1 expression in T cells from control and TCDD-treated tumors. H, CXCR3 expression on naïve T cells from control and TCDD-treated tumors.

Figure 6.

AHR activation increases dysplasia and decreases CD8⁺ T-cell engagement in a spontaneous PDAC model. A, Genetic makeup of the KC mouse model with expression of Kras^G12D within the pancreas. B, Proportion and grade of acinar tissue, acinar-to-ductal metaplasia, and PanIN lesions within the pancreata of KC mice treated with vehicle or TCDD for 3 weeks prior to analysis at 22 weeks of age, n = 3 mice/group, n = 10 high powered fields (HPFs)/mouse. C − E, Representative H&E staining (C) and relative Cyp1a1 and Cyp1b1 gene expression (D), and representative mfIHC of pancreata (E) of KC mice treated with vehicle or TCDD for 3 weeks from 19 to 22 weeks of age. Scale bar, 100 μm. Insert and white arrow identifies CD3⁺ CD8⁺ T cells. F and G, Quantification of T cells, Th cells, Tregs, and CD8⁺ T cells (F), and mean distance from Treg to CD8⁺ T cell as well as CD8⁺ T cell to epithelial cell (G). Phenotypes were identified as follows: T cells (CD3⁺), Th cells (CD3⁺ CD8⁻ FOXP3⁻), CD8⁺ T cells (CD3⁺ CD8⁺), Tregs (CD3⁺ FOXP3⁺ CD8⁻), and epithelial cell (CK19⁺) among pancreata of KC mice treated with vehicle or TCDD, n = 3 mice/group, n = 49–50 regions of interest/group. H − J, Representative mfIHC of resected PDAC samples comparing smokers with nonsmokers (H). Insert and white arrows identify Tregs. Quantification of Tregs (I) by smoking status (n = 27–34/group; G) and Tregs by pack year, n = 62 samples. J, CD8⁺ T cell to Treg proximity, n = 22–32/group and CD8⁺ T cell to Treg proximity by pack year, n = 54 tumors. Data are represented as the mean ± SD unless otherwise noted. 4′6-diamidino-2-phenylindole.