Extrinsic KRAS Signaling Shapes the Pancreatic Microenvironment Through Fibroblast Reprogramming

Abstract

Background & aims: Oncogenic Kirsten Rat Sarcoma virus (KRAS) is the hallmark mutation of human pancreatic cancer and a driver of tumorigenesis in genetically engineered mouse models of the disease. Although the tumor cell-intrinsic effects of oncogenic Kras expression have been widely studied, its role in regulating the extensive pancreatic tumor microenvironment is less understood.

Methods: Using a genetically engineered mouse model of inducible and reversible oncogenic Kras expression and a combination of approaches that include mass cytometry and single-cell RNA sequencing we studied the effect of oncogenic KRAS in the tumor microenvironment.

Results: We have discovered that non-cell autonomous (ie, extrinsic) oncogenic KRAS signaling reprograms pancreatic fibroblasts, activating an inflammatory gene expression program. As a result, fibroblasts become a hub of extracellular signaling, and the main source of cytokines mediating the polarization of protumorigenic macrophages while also preventing tissue repair.

Conclusions: Our study provides fundamental knowledge on the mechanisms underlying the formation of the fibroinflammatory stroma in pancreatic cancer and highlights stromal pathways with the potential to be exploited therapeutically.

Keywords: Fibroblasts; Macrophages; Pancreatic Cancer; Transformation.

Graphical abstract

Figure 1

Oncogenic Kras drives immune cell recruitment during the onset of tumorigenesis. (A) Experimental design. Control (lacking either the Kras or Ptf1a^Cre allele) and iKras mice were given DOX chow to activate oncogenic Kras (Kras^G12D) and pancreata were harvested 3 days (3d), 1 week (1w), or 2 weeks (2w) after induction of Kras. N = 3–5 mice per group. (B) Quantification of ADM in pancreatic tissue from control or iKras mice that received DOX chow for 3 days, 1 week, or 2 weeks. (C) Representative images of H&E staining of control and iKras pancreata at the indicated time points. N = 3–5 mice per group. Scale bar: 100 μm. (D) Representative images of CD3 (green), CK19 (red), and SMA (magenta) co-immunofluorescent staining in control and iKras pancreata at the indicated time points, arrowheads point to CD3⁺ T cells. N = 3 mice per group. Scale bar: 50 μm. (E) Immunostaining for Ki67 (green), F4/80 (red), and SMA (magenta). Scale bar: 50 μm. (F) Representative images of F4/80 (green), p-ERK (red), and SMA (magenta) co-immunofluorescent staining in control and iKras pancreata at the indicated time points. N = 3 mice per group. Scale bar: 50 μm.

Figure 2

Macrophage expansion precedes acinar transdifferentiation. (A) Immunohistochemistry for F4/80. N = 3 mice per group. (B) Representative images of F4/80 (green), CD8 (red), and CK19 (magenta) co-immunofluorescent staining in control and iKras∗ pancreata. Scale bar: 50 μm. (C) Co-immunostaining for F4/80 (green), ARG1 (red), and CK19 (gray) in iKras∗ pancreata at 3 days ON. Note that macrophages surround CK19^- acinar cells. Right: Image shows a CK19⁺ duct in the same section. Scale bar: 50 μm. (D) Representative images of F4/80 (green), p-ERK (red), and SMA (magenta) co-immunofluorescence staining of an iKras pancreas after activating Kras^G12D for 1 week. Regions with high expression of p-ERK are enlarged and single-channel images are included. N = 3 mice. Scale bar: 50 μm. (E) Representative images of PDGFRβ (green), p-ERK (red), and SMA (magenta) co-immunofluorescence staining of an iKras pancreas after activating Kras^G12D for 1 week. Single-channel images are included. N = 3 mice.

Figure 3

Sustained expression of Kras^G12Dis required to maintain PanIN and fibrosis. (A) Experimental design. Wild-type control and iKras mice were given DOX chow to activate Kras^G12D, followed by induction of acute pancreatitis. Mice either remained on DOX chow for 3 weeks and pancreata were harvested (3 weeks [3w] ON), or DOX chow was removed and the pancreata were harvested after 3 days (3d) or 1 week (1w) (labeled 3d OFF or 1w OFF, respectively). N = 8 mice per group. (B) Representative images of H&E staining of control and iKras∗ pancreata at the indicated time points. N = 8 mice per group. Scale bar: 100 μm. (C) Representative images of Gomori trichrome and periodic acid–Schiff Stain (PAS; inset) of control or iKras pancreata at the indicated time points. N = 2–3 mice per group. Scale bar: 100 μm. (D) Immunofluorescent staining for SMA (white) and p-ERK (red). Scale bar: 50 μm. N = 2–3 mice per group. (E) Immunostaining for podoplanin. Scale bar: 100 μm. (F) Immunofluorescent staining for SMA (single channel) used for quantification. Scale bar: 50 μm. N = 2–3 mice per group. Quantification of (G) podoplanin and (H) SMA staining, shown as means ± SEM. Statistical differences were determined by multiple analysis of variance, all compared with the 3w ON group. BV, blood vessels; HPF, high-power field; PDPN, Podoplanin.

Figure 4

Immune infiltration persists during tissue repair after epithelial Kras inactivation. (A) Flow cytometry gating strategy for myeloid cells in FlowJo. (B) Quantification of flow cytometry results. Immune cell populations are expressed as the percentage of total cells in control or iKras pancreata at the indicated time points, shown as means ± SD. N = 9–11 mice per group. Statistical analysis by multiple comparison analysis of variance and multiple comparison Kruskal–Wallis. Triangles represent females and circles represent males. Representative immunohistochemistry images for (C) CD45 and (D) F4/80 in the pancreatic tissue of control or iKras∗ mice at the indicated time points. Scale bar: 100 μm. N = 3 mice per group. (E) Top: Representative images of F4/80, Arg1, CD3, CD8, and CK19 immunofluorescent multiplex staining at the indicated time points. N = 5–7 mice per group. Bottom: Identification of cell types using inForm Cell Analysis software. T cell, macrophage, and epithelial cells were selected based on single staining of CD3 (Opal 520), F480 (Opal 540), and CK19 (Opal 690), respectively. Acinar cells were labeled based on the lack of listed fluorophores and by tissue morphology. (F) OPAL staining quantification of the percentage of F4/80-positive cells from total cells in the pancreatic tissue of control or iKras mice at the indicated time points. Data are shown as means ± SEM. N = 5–7 mice per group. Statistical analysis was performed by the 2-tailed unpaired t test. FSC-A, forward scatter area.

Figure 5

Oncogenic Kras regulates myeloid cell polarization status in the PME. (A) Experimental design. N = 8–9 mice per group. (B) CyTOF analysis of CD45+ cells from iKras pancreata visualized by t-distributed stochastics neighbor embedding (tSNE) plot for 3 weeks (3w) ON and 3 days (3d) OFF time points. Nineteen distinct cell clusters were identified by FlowSOM (Bioconductor, Buffalo, NY) using 18 markers and 1061 randomly selected cells per group. N = 2–3 mice per group. (C) Heatmap of the median marker intensity generated with FlowSOM of the CyTOF samples (iKras 3w ON and 3d OFF combined). Colors on the left represent the clusters shown in the t-SNE plot. N = 2–3 per group. All population were gated from CD45⁺ cells. (D) tSNE plots showing the expression of different immune lineage markers in the clusters from the CyTOF data. (E) Quantification of identified clusters proportion by CyTOF analysis at the indicated time points. Statistical analysis by Wrapper function. (F) Heatmap showing differential abundance of clusters from CyTOF analysis of iKras pancreata for 3w ON and 3d OFF time points. N = 2–3 mice per group.

Figure 6

Oncogenic Kras inactivation results in transcriptional reprogramming of infiltrating macrophages. (A) Immunostaining for F4/80 (green), ARG1 (red), and E-cadherin (Ecad) (white) in control and iKras pancreata at the indicated time points. N = 3 mice per group. Scale bar: 50 μm. (B) OPAL staining quantification of ARG1+ macrophages in pancreatic tissue of control or iKras mice at the indicated time points. Data are shown as means ± SEM. Statistical analysis by the 2-tailed unpaired t test. (C) Immunostaining for F4/80 (green), p-STAT3 (magenta), and Ecad (white). N = 3 mice per group. Scale bar: 50 μm. (D) UMAP visualization of scRNAseq data showing unsupervised clustering of cells from iKras pancreatic samples (3 weeks [3w] ON, N = 2; 3 days [3d] OFF, N = 3). Each color represents a distinct cellular cluster. (E) Dot plot of the key genes used to identify different cell populations shown in the UMAP, identified by unsupervised clustering of the scRNAseq samples (3w ON and 3d OFF samples combined). Size of the dots reflects the proportion of cells expressing a determined gene and color represents the average expression level. (F) Heatmap showing the averaged scRNAseq expression data (relative to the highest expressor) for genes in macrophages selected from a curated list of macrophage polarization and functional markers.

Figure 7

Fibroblasts express inflammatory cytokines in response to epithelial oncogenic Kras. (A) Interactome analysis showing all predicted ligand-receptor interactions between different cell populations identified in iKras pancreatic scRNAseq analysis. Each line represents a ligand-receptor pair, color-coded by the cell type expressing the ligand. (B) Differential interactions from panel A positively regulated by oncogenic Kras (higher in 3 weeks [3w] ON compared with 3 days [3d] OFF) (adjusted P value < .05). (C) Circos plot showing the average expression of fibroblast ligands connected to their predicted receptors on various cell populations as measured in the pancreatic scRNAseq analysis. Ligands shown are from panel B (adjusted P value < .05). (D) Heatmap showing averaged scRNAseq expression data (relative to the highest expressor) for genes in fibroblasts from a curated list of immunomodulatory factors. (E) Violin plots showing expression of Il33, Cxcl1, Il6, Saa3, and Saa1 and their respective receptors Il1rl1, Cxcr2, Il6ra, P2rx7, and Scarb1 across all identified cell populations in both Kras∗ ON and OFF samples combined. DC, Dendritic cell; NK T, Natural killer T cell.

Figure 8

Fibroblasts secrete immunomodulatory factors with receptors in myeloid cells. Dot plot showing average expression of Il33, Il6, Cxcl1, Saa1, and Saa3 and their respective receptors Il1rl1, Il6ra, Cxcr2, P2rx7, and Scarb1 across cell populations. Size of the dots reflects the proportion of cells expressing a determined gene and color represents the average expression level. NK T, Natural killer T cell.

Figure 9

Fibroblasts are reprogrammed through epithelial Kras^G12D-induced secreted molecules. (A) Experimental design. CM was collected from iKras∗ cancer cells cultured with +DOX to activate Kras^G12D expression (Kras ON) or without -DOX (Kras OFF). The media samples then were boiled to denature protein factors, or left intact and used to culture pancreatic fibroblasts. DMEM was used as control. (B) Quantitative reverse-transcription polymerase chain reaction for Cxcl1, Il33, Il6, and Saa3 expression in fibroblasts (CD1WT) that were cultured with CM either from Kras ON, Kras OFF, or DMEM, either boiled or not boiled. Gene expression was normalized to Peptidylprolyl Isomerase A (Ppia). Data are shown as means ± SD. N = 3 per group. The statistical differences were determined by multiple-comparison analysis of variance. (C) Experimental design. CM was collected from iKras cancer cells cultured with +DOX to activate Kras^G12D expression (Kras∗ ON) or without -DOX (Kras^G12D OFF), and used for metabolomic analysis. (D) Heatmap showing levels of extracellular metabolites in iKras cancer cell CM with Kras^G12D ON or OFF. Both conditions are compared with regular DMEM media. (E) Fibroblasts were exposed to +DOX CM and -DOX CM for 48 hours. The resulting CM was analyzed for metabolites, in parallel with iKras cell CM and DMEM. Relative metabolite abundance of lactate and pyruvate from all conditions are shown. (F) Heatmap showing averaged scRNAseq expression data (relative to the highest expressor) for genes encoding for ligands that are secreted by fibroblasts and epithelial cells. FB, Fibroblast; Rel., relative.

Figure 10

Fibroblast reprogramming requires JAK/STAT signaling. (A) Immunofluorescent staining of p-STAT3 (red), Platelet-derived growth factor receptor beta (PDGFRβ) (green), and SMA (gray). N = 2–3 mice per group. Arrowheads point to p-STAT3–positive fibroblasts and arrow points to positive epithelial cells. (B) Immunofluorescent staining for PDGFRβ (green), SMA (white), and p-ERK (red). Scale bar: 50 μm. N = 2–3 mice per group. (C) Experimental design. Fibroblasts were exposed to iKras∖tumor cell CM either from Kras^G12D ON, Kras^G12D OFF, or DMEM. Fibroblasts cultured with Kras^G12D ON CM media were treated with a JAK/STAT2/3 inhibitor (ruxolitinib) or vehicle. (D) Western blot showing protein levels of total STAT3 or phospho-STAT3 from CD1WT fibroblasts treated as indicated. α-tubulin was used as a loading control. (E) Quantitative reverse-transcription polymerase chain reaction results for Cxcl1, Il33, Il6, and Saa3 expression in fibroblasts (CD1WT) treated with CM as described in panel C. Data are shown as means ± SD. N = 3 per group. The statistical differences were determined by multiple-comparison analysis of variance. (F) Quantitative reverse-transcription polymerase chain reaction for Cxcl1, Il33, Il6, and Saa3 expression in fibroblasts (B6318) treated with iKras cell CM either from Kras^G12D ON, Kras^G12D OFF, or DMEM. Fibroblasts cultured with Kras^G12D ON CM medium were treated with a JAK/STAT2/3 inhibitor (ruxolitinib) or vehicle. Data are shown as means ± SD. The statistical differences were determined by multiple-comparison analysis of variance. mRNA, messenger RNA.

Figure 11

Inhibition of the JAK/STAT pathway mimics Kras^G12Dinactivation-induced tissue repair. (A) Experimental design. iKras mice were given DOX and after 3 days acute pancreatitis was induced. After pancreatitis, Kras^G12D was left ON for 3 weeks and at 3 weeks mice either received ruxolitinib (180 mg/kg) daily for 3 days or were treated with vehicle. N = 2–3 mice per group. (B) H&E, periodic acid–Schiff (PAS), and immunofluorescent staining for p-STAT3 (red), F4/80 (green), and SMA (white), or ARG1 (red), F4/80 (green), and CK19 (white). Scale bar: 50 μm. N = 2–3 mice per group. (C) Quantification for the p-STAT3 staining. (D) Quantification for the PAS staining. (E) Quantification of CK19-positive area. (F) Quantification of SMA-positive area. (G) Quantification of ARG1+-positive area within the F4/80+ macrophages. (H) Quantification of F4/80+-positive area. All data were analyzed using a t test. Blue, male; pink, females. Rux, ruxolitinib; Veh, vehicle.