Elimination of oncogenic KRAS in genetic mouse models eradicates pancreatic cancer by inducing FAS-dependent apoptosis by CD8+ T cells

Abstract

Oncogenic KRAS^G12D (KRAS^∗) is critical for the initiation and maintenance of pancreatic ductal adenocarcinoma (PDAC) and is a known repressor of tumor immunity. Conditional elimination of KRAS^∗ in genetic mouse models of PDAC leads to the reactivation of FAS, CD8⁺ T cell-mediated apoptosis, and complete eradication of tumors. KRAS^∗ elimination recruits activated CD4⁺ and CD8⁺ T cells and promotes the activation of antigen-presenting cells. Mechanistically, KRAS^∗-mediated immune evasion involves the epigenetic regulation of Fas death receptor in cancer cells, via methylation of its promoter region. Furthermore, analysis of human RNA sequencing identifies that high KRAS expression in PDAC tumors shows a lower proportion of CD8⁺ T cells and demonstrates shorter survival compared with tumors with low KRAS expression. This study highlights the role of CD8⁺ T cells in the eradication of PDAC following KRAS^∗ elimination and provides a rationale for the combination of KRAS^∗ targeting with immunotherapy to control PDAC.

Keywords: CD8(+) T cells; FAS-FASL; KRAS(G12D); oncogenic KRAS; pancreatic cancer.

Figure 1.. Kras* drives T cell suppression and myeloid infiltration in the PDAC TME irrespective of other genetic mutations

(A) Timeline of the experiment. Kras* was induced in PiKP and PiKT mice at 18 weeks of age. After 6 weeks of induction, Kras* was turned off and tumors were analyzed at 5 days, 2 weeks and 4 weeks after removal of Dox. (B-D) Intra-tumoral immune composition of PiKP (Kras* On: n=11, Kras* 5d off: n=7) and PiKT (Kras* On: n=13, Kras* 5d off: n=5) tumors. (B) T cell populations determined as a percentage of CD45⁺ cells. T cells (CD45⁺ CD3⁺), CD4⁺ T cells (CD45⁺ CD3⁺ CD4⁺), CD8⁺ T cells (CD45⁺ CD3⁺ CD8⁺), CD4⁺ Foxp3⁺ cells and CD4⁺ Foxp3⁻ cells of indicated groups. (C) Myeloid populations determined as a percentage of CD45⁺ cells. CD45⁺ CD11b⁺ (Myeloid cells), CD45⁺ CD11b⁺ Ly6C⁺ Ly6G⁻ (Mo-MDSCs) and CD45⁺ CD11b⁺ Ly6G⁺ Ly6C⁻ (Gr-MDSCs) of indicated groups. (D) Fold change of immune cells in Kras* 5d off PiKP and PiKT tumors, increase or decrease is indicated by arrows. (E-F) Representative images (E), and quantification (F) of CD4⁺ T cells (CD4), CD8⁺ T cells (CD8), and myeloid cells (CD11b⁺) by immunolabeling in Kras* On, Kras* off (5 days, 2 weeks and 4 weeks) and never on dox PiKP tumors (n = 2–6 per group). Scale bars, 100 μm; inset 50 μm. (G) Pie chart of intra-tumoral immune composition of PiKP (Kras* On: n=11, Kras* 5d off: n=7) and PiKT (Kras* On: n=13, Kras* 5d off: n=5) tumors as a percentage of CD45⁺ cells by flow cytometry. (H-I) Flow cytometry analysis of immune cells in PiKP tumors with Kras* On (n=6) and Kras* 5d off (n=6). Shown are percentages of TIM3⁺ T cells (H) and T-bet⁺ T cells (I). (J) qPCR analysis of Ifng expression on tumor infiltrating T cells (n=5 biological replicates). (K) Flow cytometry analysis of immune cells in PiKP tumors with Kras* On (n=6) and Kras* 5d off (n=6). Shown are percentages of CD40⁺ and CD86⁺ macrophages (CD11b⁺ F4/80⁺) and dendritic cells (CD11b⁻ CD11c⁺DCs and CD11b⁺ CD11c⁺ DCs). In (B, C, F, H, I, J and K), data are presented as the mean ± SD, as mean values in (G) and as mean fold change in (D). In (B, C and D), significance was determined by unpaired T-test for comparison of PiKP (CD3⁺ T cells, CD8⁺ T cells, CD4⁺ T cells, CD4⁺ Foxp3⁻ T cells, CD11b⁺ Ly6C⁻ Ly6G⁺ cells, CD19⁺ B cells), and PiKT (CD3⁺ T cells, CD4⁺ T cells, CD8⁺ T cells, CD4⁺ Foxp3⁺ T cells, CD11b⁺ Ly6C⁻ Ly6G⁺ cells) and by Mann-Whitney test for all other comparisons. Significance was determined by unpaired T-test (CD4⁺ T cells quantification) or Mann-Whitney test (CD8⁺ T cells quantification) or one-way ANOVA with Dunnett’s multiple comparisons test (CD11b⁺ cells quantification) for (F), by unpaired T-test in (H, I, J and K. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001, ns: not significant. Note: CD11b⁺ F4/80⁺ and CD45⁺ Ly6G⁻ Ly6C⁻ cells have been used interchangeably to identify macrophages due to overlap of fluorophores conjugated to antibodies.

Figure 2.. Kras* suppresses Fas expression on pancreatic cancer cells

(A) Schematic representation of isolation of PiKP-785 cell line from advanced PiKP tumor. Cell lines were cultured in the presence (red, Kras* On) or absence of Dox (blue, Kras* off). (B) qPCR analysis of relative Kras* expression in PiKP-785 cells with Kras* on and Kras* 3 and 5 days off (n=3 biological replicates per group). (C) Seurat UMAP clustering of PiKP-785 cells in the presence or absence of Dox. (D) scRNA seq analysis of Kras expression level in PiKP Kras* On, 3d off, 5d off and On-off-on cells. (E) Pathway enrichment analysis of the differentially expressed genes by scRNA seq: Kras* 5d off vs. Kras* On PiKP-785 cells. Top 10 enriched immune related pathways were selected based on the FDR. (F) scRNA seq analysis of Fas expression level of PiKP-785 cells in the presence or absence of Dox. (G) Immune qPCR microarray was performed on PiKP-785 cells cultured with Kras* on and Kras* 3 and 5 days off (n=1 biological replicate per group). Heat maps depicts the fold change in gene expression 3d and 5d following Kras* elimination. The values indicate the relative fold change. (H) qPCR analysis of Fas expression in PiKP-785 cells with Kras* on and Kras* 3 and 5 days off (n=3 biological replicates per group). (I-J) Flow cytometry analysis of Fas surface expression on PiKP-785 cell line. Cells were cultured with Kras* on; Kras* off for 3 days, 5 days, 14 days, 3 weeks and 4 weeks, and Kras* off for 14 days and then Kras* on for 10 days (n=3 biological replicates per group). Histograms showing Fas expression (I) and percentage Fas⁺ cells calculated by overtone method (J). (K) Apoptosis in PiKP-785 cells with and without Kras* expression treated with anti-Fas agonist antibody (Jo2) (n=3 biological replicates per group). (L) Caspase 8 activity in PiKP-785 cells which were cultured with Kras* off for 5 days with Jo2 antibody treatment. Data are presented as the mean ± SD in (B, H, J, K and L) and as violin plots with normalized gene expression levels of indicated genes in (C and F). Significance was determined by one-way ANOVA with Dunnett’s multiple comparisons test (B, H and J), one way ANOVA with Sidak’s multiple comparisons test (K) and by unpaired T-test (L). ** P<0.01, *** P<0.001, **** P<0.0001, ns: not significant.

Figure 3.. Suppression of Fas on cancer cells reverses Kras* elimination mediated eradication of PDAC.

(A) qPCR analysis of relative Fas expression (2^-Δct) in AK14837 parental cells, sh-Scrambled (shScr) and shFas knockdown cells using different shRNAs (shFas#28, shFas#29, shFas#30, shFas#31, shFas#32) with Kras* on, Kras* 3d off and 5d off (n=3–6 biological replicates per group). (B) Western blot analysis of Fas expression level in parental cells, sh-Scrambled (shScr) and shFas#28 cells with Kras* on, Kras* 3d off and 5d off. (C) Timeline of the experiment for C57BL/6J mice orthotopically injected with syngeneic AK14837 cells (parental) and AK14837 Fas knockdown (sh-Fas#28) cell lines. Baseline MRI measurements were done on day 10 in mice prior to Dox withdrawal and subsequent MRI was done on day 20. (D) Baseline (day 10) (top panel) and post dox withdrawal (day 20) (bottom panel) MRI imaging of tumors of indicated groups. (E) Change in tumor volumes (based on MRI measurements) of parental, shScr and shFas#28 mice with Kras* On and Kras* On and off (n=4–5 mice/ group). (F) Kaplan-Meier survival curve of parental Kras* On (n=10), parental Kras* On and off (n=10), shScr Kras* On (n=10), shScr Kras* On and off (n=10), shFas#28 Kras* On (n=10), shFas#28 Kras* On and off (n=9). (G) Tumor/ pancreas weights of indicated groups at sacrifice or endpoint; Parental Kras* On (n=9), Parental Kras* On and off (n=10), shScr Kras* On (n=8), shScr Kras* On and off (n=8), shFas#28 Kras* On (n=9), shFas#28 Kras* On and off (n=9). (H-I) CD4 and CD8 immunolabeling (H), and quantification (I) of Kras* On and Kras* 5d off parental, shScr and shFas#28 tumors (n=3–6 tumors per group). Scale bars, 100 μm. In (A, E, G and I), data are presented as the mean ± SD. Significance was determined by Kruskal-Wallis (Dunn’s multiple comparisons) test (Kras* On comparisons of Parental, shScr and shFas cell lines) and one-way ANOVA with Dunnett’s multiple comparisons test (Kras* 3d off and Kras * 5d off comparisons of Parental, shScr and shFas cell lines) (A), one-way ANOVA with Sidak’s multiple comparisons test (E), Kruskal-Wallis (Dunn’s multiple comparisons) test (G). In (I), significance was determined by unpaired T-test for CD8⁺ T cells comparisons, and for CD4⁺ T cells comparison: parental and shFas#28 panels and by Mann-Whitney test for other panels. In (F), significance was determined by log-rank test. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001, ns: not significant.

Figure 4.. Kras* elimination epigenetically upregulates Fas on cancer cells to facilitate CD8⁺ T cells mediated eradication of PDAC

(A) Timeline of the experiment for C57BL/6J mice orthotopically injected with syngeneic PiKP (AK14837) cells. αCD4, αCD8 or isotype antibody treatment was initiated from day 8 in orthotopic PiKP tumor bearing mice with Kras* On and off. (B) Flow cytometry analysis of depletion of intratumoral CD4⁺ or CD8⁺ T cells (gated on CD45⁺ CD3⁺ CD11b⁻ cells) by αCD4 or αCD8 respectively. (C) Intratumoral immune composition of orthotopic PiKP Iso (Kras* On), PiKP Iso (Kras* 5d off), PiKP αCD4 (Kras* 5d off), and PiKP αCD8 (Kras* 5d off) tumors (n=3 per group). (D-E) Tumor/ pancreas weights at sacrifice or endpoint (D) and Kaplan-Meier survival curve (E) of indicated groups. PiKP Iso (Kras* On; n=7), PiKP Iso (Kras* On and off; n=6), PiKP αCD4 (Kras* On and off; n=7) and PiKP αCD8 (Kras* On and off; n=7). (F) Histopathological analyses of H&E stained tumor/pancreas sections of mice in the indicated groups and time points. PiKP Iso (Kras* On; n=4), PiKP Iso (Kras* On and off; n=5), PiKP αCD4 (Kras* On and off; n=6) and PiKP αCD8 (Kras* On and off; n=4). Scale bars, 100 μm. (G) Quantification of relative percentage of tissue phenotype in indicated groups from (F). (H) tSNE plot demonstrating main cell types in murine PDAC tumors from Elyada et al., study (left panel). Cell type labels were obtained from the original study. tSNE plot of T cells extracted from all cells (right panel). (I) Expression levels of Cd3d, Cd8a, Cd4 and Fasl in T cells are plotted onto the tSNE. (J) Violin plots showing distributions of Cd8a and Fasl in all 5 subclusters of T cells. (K) MeDIP analysis of Fas methylation in PiKP-785 cells with Kras* On, Kras* 3d off and 5d off (n=3 biological replicates per group). (L) Surface Fas expression evaluated by flow cytometry. PiKP-785 cells were cultured with Kras* on, Kras* off for 5 days, or Kras* off for 5 days and then on for 5 days. DNA methylation inhibitor 5-AZA (5-azacytidine) was added to cells during the 5-day culture (n=3 biological replicates per group). (M) scRNA seq analysis of Dnmt1 and S100z expression level in PiKP Kras* On, 3d off, 5d off and On-off-on cells. (N) ChIP assays demonstrating the relative binding of DNMT1, EZH2, H3K27me3 and H3K27ac on Fas promoter in Kras* On, 3d off and 5d off cells (n=3 biological replicates per group). (O) Summary of ChIP results. (P) Representative immunolabeling for co-localization analysis of tumor cells (CK19, green), Fas (red) and cleaved caspase 3 (CC3, yellow) on a PiKT tumor tissue. Co-localization of the three markers is indicated in magenta color. Scale bars, 100 μm. (Q) Quantification of Fas, CC3 and CK19 co-localization in PiKP and PiKT tumors with Kras* on (n=7) and Kras* 5d off (n=8). Data are presented as mean ± SD (C, D, G, K, L, N and Q) and as violin plots with normalized gene expression levels of indicated genes in (J and M). Significance was determined by one-way ANOVA with Dunnett’s multiple comparisons test (C, D, K and N), two-way ANOVA with Tukey’s multiple comparisons test (G), log rank test (E), by Mann-Whitney test (Q). In (L), Mann-Whitney test was used for ‘5d off’ comparisons and unpaired T-test for all other panels. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001, ns: not significant.

Figure 5.. KRAS expression inversely correlates with FAS and CD8⁺ T cell infiltration in human PDAC (A-E)

Single cell RNA seq analysis of human PDAC tumors from Peng et. al. Cell type labels were obtained from the original study. (A) tSNE plot denoting cell types in PDAC tumors (left panel). tSNE plot of T cells (right panel) and ductal cells type 1 and type 2 (bottom panel). (B) tSNE plots of CD3D, CD8A, CD4 and FASLG expression levels in T cells. (C) Violin plots showing distributions of CD8A and FASLG in all 13 subclusters of T cells. (D) tSNE plots of FAS and KRAS expression in cancer cell subclusters. (E) FAS expression level in KRAS low and KRAS high cancer cells. From (L). (F-H) CIBERSORTx analysis of bulk gene expression data of TCGA pancreatic adenocarcinoma (PAAD) from the UCSC Xena database (n=179). Median expression value of KRAS (F), and the proportion of indicated cell types in tumors (G). (H) Kaplan Meier survival curves of KRAS high and KRAS low tumors from (N). Data are presented as violin plots with normalized gene expression levels of indicated genes in (C) and as box and whisker plots in (E, F and G). Significance was determined by unpaired T-test in (F), Mann-Whitney test in (E and G), and by log rank test in (H). *P<0.05, ****P<0.0001, ns: not significant. Scale bars, 100 μm.