Therapeutic KRASG12C inhibition drives effective interferon-mediated antitumor immunity in immunogenic lung cancers

Abstract

Recently developed KRAS^G12C inhibitory drugs are beneficial to lung cancer patients harboring KRAS^G12C mutations, but drug resistance frequently develops. Because of the immunosuppressive nature of the signaling network controlled by oncogenic KRAS, these drugs can indirectly affect antitumor immunity, providing a rationale for their combination with immune checkpoint blockade. In this study, we have characterized how KRAS^G12C inhibition reverses immunosuppression driven by oncogenic KRAS in a number of preclinical lung cancer models with varying levels of immunogenicity. Mechanistically, KRAS^G12C inhibition up-regulates interferon signaling via Myc inhibition, leading to reduced tumor infiltration by immunosuppressive cells, enhanced infiltration and activation of cytotoxic T cells, and increased antigen presentation. However, the combination of KRAS^G12C inhibitors with immune checkpoint blockade only provides synergistic benefit in the most immunogenic tumor model. KRAS^G12C inhibition fails to sensitize cold tumors to immunotherapy, with implications for the design of clinical trials combining KRAS^G12C inhibitors with anti-PD1 drugs.

Fig. 1.. Oncogenic KRAS regulates immune gene expression in cell lines.

(A) Cytokine array of cell culture supernatant from KRAS^G12V-ER pneumocytes treated with 500 nM 4-OHT or ethanol control for 24 hours. Graph shows secreted protein relative to control spots on the array for each condition. (B) Log₂Fold change of selected cytokine genes from RNA-seq data in 4-OHT (500 nM, 24 hours)–treated KRAS^G12V-ER pneumocytes. (C) MSigDB Hallmarks gene set enrichment analysis (GSEA) plots of IFNα and IFNγ pathways in 4-OHT–treated versus control samples. (D) Log₂Fold change of selected cytokine genes from RNA-seq data in ARS-1620 (2 μM, 24 hours)–treated NCI-H358 cells versus dimethyl sulfoxide (DMSO) control. (E) MSigDB Hallmarks GSEA plots of IFNα and IFNγ pathways in ARS-1620–treated versus control samples. (F) Same analysis as (D) of RNA-seq from 3LL ΔNRAS cells treated with 100 nM MRTX1257 (24 hours, n = 3). (G) Same analysis as (E) of RNA-seq from 3LL ΔNRAS cells. All statistics represent false discovery rate (FDR)–adjusted P values (q < 0.05).

Fig. 2.. KRAS signaling down-regulates IFN pathway gene expression via MYC.

(A) qPCR analysis of IFN-induced genes in MRTX1257-treated (100 nM, 24 hours) 3LL ΔNRAS cells (2^−ΔΔCT, normalized to control sample for all genes, n = 6, unpaired t test, mean + SEM). (B) Same as (A) using KPB6^G12C (n = 4) and CT26^G12C (n = 3) cell lines. (C) qPCR showing KRAS-dependent regulation of Myc in 3LL ΔNRAS cells (n = 3) after treatment with MRTX1257 and KRAS^G12V-ER pneumocytes (n = 4) after treatment with 4-OHT (unpaired t test, mean + SEM). (D) RNA-seq mRNA counts of 3LL ΔNRAS lung tumors treated with vehicle or MRTX1257 (50 mg/kg) for 28 hours or 8 days (each dot represents a tumor, n = 6 per group, FDR P adjusted value). (E) Western blot showing MYC knockdown and STAT2 up-regulation of 3LL ΔNRAS cells treated with 100 nM MRTX1257 (24 hours), Myc small interfering RNA (siRNA; 48 hours), or both. Quantification for two independent experiments is shown on the right (mean + SEM). (F) qPCR analysis of IFN-induced genes in 3LL ΔNRAS cells treated with 100 nM MRTX1257, Myc siRNA, or both (2^−ΔΔCT, normalized to control sample for all genes, n = 3, paired t tests, siMyc versus Mock, mean + SEM). (G) Same analysis as (F) in KPB6^G12C (n = 4) and CT26^G12C (n = 3) cells.

Fig. 3.. KRAS^G12C inhibition enhances tumor cell–intrinsic IFN responses.

(A) qPCR analysis of IFN-induced genes in MRTX1257 (100 nM, 24 hours) and/or recombinant IFNγ (100 ng/ml)–treated 3LL ΔNRAS cells (2^−ΔΔCT, normalized to control sample for all genes, n = 6, paired t test, mean + SEM). (B) Protein validation of IFN response regulation by KRAS. Left: Percentage of CXCL9-positive cells as measured by flow cytometry on 3LL ΔNRAS cells after treatment with MRTX1257 and/or IFNγ. Right: Concentration of CXCL9 secreted to the medium of 3LL ΔNRAS cells after treatment with MRTX1257 and/or IFNγ, measured by enzyme-linked immunosorbent assay (normalized to control sample, n = 3, paired t test, mean + SEM for both). (C) Same as (A) using KPB6^G12C (n = 4) and CT26^G12C (n = 3) cell lines. (D) qPCR analysis of IFN-induced genes in 3LL ΔNRAS cells treated with IFNγ only or IFNγ and MRTX1257 in the presence of 48 hours of Mock or Myc siRNA (2^−ΔΔCT, normalized to IFNγ only–treated sample for all genes, n = 3, paired t test, mean + SEM). (E) qPCR analysis of IFN pathway genes in human KRAS^G12V-ER pneumocytes after treatment with 4-OHT and/or recombinant IFNγ for 24 hours (normalized to control sample, n = 4, paired t test, mean + SEM).

Fig. 4.. KRAS^G12C inhibition remodels the immunosuppressive TME of 3LL ΔNRAS lung tumors.

(A) Immunophenotyping of dissected lung tumors obtained by intravenous administration of 3LL ΔNRAS cells (n = 5 mice) versus healthy lung tissue (n = 6 mice) obtained by flow cytometry. (B) Whole-exome sequencing SNV analysis of two NRAS CRISPR-edited 3LL clones. (C) Posttreatment tumor volume change as measured by μCT scanning of 3LL ΔNRAS lung tumors after 1 week of treatment with vehicle control or MRTX1257 (50 mg/kg) (each bar represents one tumor, Mann-Whitney test). (D) Summary of significantly (FDR q < 0.05) up- and down-regulated pathways in MRTX- versus vehicle-treated lung tumors (MSigDB Hallmarks). (E) Hierarchical clustering of relative frequencies of tumor-infiltrating cell types in MRTX- and vehicle-treated tumors obtained by IMC. (F) Percentage of neutrophils (gated as CD45⁺ CD11b⁺ Ly6C⁺ Ly6G⁺) and monocytes (gated as CD45⁺ CD11b⁺ Ly6C^hi Ly6G⁻) in vehicle-treated (n = 5) and MRTX-treated (n = 8) lung tumors measured by flow cytometry (each dot represents a mouse, unpaired t test). (G) mRNA counts for Ccl2 gene in MRTX-treated 3LL ΔNRAS tumors (n = 6 per group, left) and cells (n = 3, right) obtained by RNA-seq (FDR-adjusted P value). (H) Live cell count (by flow cytometry) of bone marrow–derived monocytes that have migrated through a transwell in the presence of conditioned medium from 3LL ΔNRAS cells, MRTX-treated cells, or two clones from Ccl2 CRISPR knockout [n = 3 independent experiments, one-way analysis of variance (ANOVA)].

Fig. 5.. KRAS^G12C inhibition promotes APC activation.

(A) Normalized percentage of GFP⁺ Mutu DCs that have phagocytosed CTV⁺ 3LL ΔNRAS cells, previously treated with DMSO control, or 100 nM MRTX1257 for 24 or 48 hours, measured by flow cytometry (n = 3 independent experiments, one-way ANOVA, mean ± SEM). (B) Flow cytometry analysis of 3LL ΔNRAS lung tumors treated with vehicle or MRTX1257 (50 mg/kg) for 7 days. Macrophages are gated as Live CD45⁺ CD11b⁺ CD24⁻ CD64⁺, and cDC1s are obtained by Live, CD45⁺, CD11c⁺ CD24⁺ CD103⁺ gating (n = 5 for vehicle, n = 8 mice for MRTX-treated, unpaired t test, mean ± SEM). (C) qPCR data for 3LL ΔNRAS lung tumors treated as in (B) (2^−ΔΔCT, unpaired t test, n = 7 vehicle, n = 8 treated, mean ± SEM). (D) Normalized mean fluorescence intensity of MHCII and CD86 as measured by flow cytometry of DCs cocultured with 3LL ΔNRAS cells previously treated with either DMSO or MRTX for 48 hours (pregated as GFP⁺, unpaired t test, n = 3 independent experiments, mean ± SEM). (E) qPCR data of Cxcl10 gene in 3LL ΔNRAS tumors, analyzed as in (C). (F) Proportion of CXCL9⁺ cells in each population, as detected by IMC, per ROI (Region of Interest) (n = 11 vehicle, n = 9 MRTX, unpaired t tests, mean ± SEM). (G) Normalized percentage of CXCL9⁺ DCs after coculture with 3LL ΔNRAS cells as in (D) (n = 5 independent experiments, unpaired t test, mean + SEM).

Fig. 6.. KRAS^G12C inhibition leads to T cell infiltration and activation.

(A) Summary of T cell infiltration measured by flow cytometry in vehicle versus MRTX-treated lung tumors (n = 5 for vehicle, n = 8 mice for treated, unpaired t tests). (B) qPCR analysis of cytotoxicity genes in 3LL ΔNRAS lung tumor (2^−ΔΔCT, unpaired t test, n = 7 vehicle, n = 8 treated, mean ± SEM). (C) Flow cytometry analysis of CD8⁺ T cell phenotypes. Left: Percentage of CD69⁺ CD8⁺ T cells in both treatment groups (n = 5 vehicle, n = 8 MRTX-treated, unpaired t test, mean ± SEM). Right: Percentage of naïve (CD44⁻ CD62L⁺), effector (CD44⁺ CD62L⁻), and memory (CD44⁺ CD62L⁺) CD8⁺ T cells, same analysis as on the left for each cell population. (D) Contour plot of PD1 and LAG-3 expression on CD8⁺ cells in vehicle- and MRTX-treated 3LL ΔNRAS lung tumor samples (graph shows one representative example for n = 5 vehicle and n = 8 MRTX-treated samples). (E) Visualization of cell outlines as measured by IMC, of CXCL9-negative and CXCL9-positive DCs, PD1⁺ and LAG-3⁺ CD8⁺ T cells, and T_regs in a vehicle- and MRTX-treated tumor. (F) Quantification of occurrence of the different T cell subsets in the neighborhood of CXCL9⁺ and CXCL9⁻ DCs, depicted as the average proportion of that cell type among all neighbors within 100-pixel radius of the DC subset.

Fig. 7.. KRAS^G12C inhibition does not synergize with ICB in immune refractory tumors.

(A) Tumor volume change after 2 weeks of treatment of 3LL ΔNRAS tumor–bearing mice with either MRTX1257 (50 mg/kg) only (n = 9 mice) or MRTX1257 and anti-PD1 (n = 10 mice). Each bar represents a tumor. (B) Flow cytometry analysis of proliferating tumor cells (CD45⁻ Ki67⁺), T_regs (CD3⁺ CD4⁺ Foxp3⁺), and activated T cells (CD69⁺ CD8⁺) in vehicle (n = 10)–treated, MRTX (n = 7)–treated, or MRTX plus anti-PD1 (n = 8)–treated (2-week treatment) 3LL ΔNRAS lung tumors (one-way ANOVA, mean ± SEM). (C) Immunohistochemistry analysis and quantification for CD8 (n = 4 mice per group) and Foxp3 (n = 3 mice per group) in KPB6^G12C tumor–bearing lungs after 7 days of vehicle or MRTX1257 treatment (50 mg/kg; each dot represents one tumor, unpaired t test, mean ± SEM). (D) qPCR analysis of immune genes in vehicle (n = 15 tumors) or MRTX-treated (n = 10 tumors) KPB6^G12C lung tumors (2^−ΔΔCT, unpaired t test, mean ± SEM). (E) Survival of KPB6^G12Clung tumor–bearing mice treated with vehicle [+immunoglobulin G (IgG) control, n = 6 mice], MRTX1257 (+IgG control, n = 6 mice), anti-PD1 (n = 5 mice), or combination (n = 4 mice, log-rank Mantel Cox test).

Fig. 8.. In an immunogenic model, MRTX-driven immune responses drive complete tumor rejection.

(A) Tumor volume change after 7 days of treatment of KPAR^G12C tumor–bearing mice with either vehicle (n = 3 mice) or MRTX849 (n = 2 mice). Each bar represents a tumor, Mann-Whitney test. (B) Growth of subcutaneously implanted KPAR^G12C tumors treated with either vehicle or MRTX849 (50 mg/kg) for 2 weeks. At day 71, the remaining mice were rechallenged with KPAR^G12C cells in the opposite flank, which did not give rise to tumors. (C) Summary of significantly (FDR q < 0.05) up- and down-regulated pathways in MRTX849 (50 mg/kg, 6 days) versus vehicle-treated KPAR^G12C lung tumors (MSigDB Hallmarks); n = 9 tumors per group (three mice). (D) Heatmap showing mRNA expression from RNA-seq of KPAR^G12C tumors treated for 6 days with MRTX849 (50 mg/kg). Gene expression is scaled across all tumors. (E) Flow cytometry analysis of KPAR^G12C-bearing lungs treated with either vehicle (n = 8 mice) or MRTX849 (50 mg/kg; n = 7 mice) for 6 days, showing increased CD69⁺ CD8⁺ T cells (top left), increased CD44⁺ CD62L⁻ effector CD8⁺ T cells (top right), and increased checkpoint molecule expression on CD8⁺ T cells (below) after KRAS inhibition (all statistics are Student’s t tests, mean ± SEM).

Fig. 9.. Synergy with anti-PD1 requires an intact tumor cell–intrinsic IFN response.

(A) Survival of KPAR^G12C lung tumor–bearing mice after treatment with IgG control (n = 8 mice) or anti-PD1 (10 mg/kg; n = 6 mice). Dotted lines represent start and end of treatment, respectively, log-rank Mantel Cox test. (B) Survival of KPAR^G12C lung tumor–bearing mice after treatment with vehicle (+IgG control, n = 6 mice), anti-PD1 (10 mg/kg; n = 8 mice), MRTX1257 (50 mg/kg; +IgG control, n = 4 mice), or both (n = 6 mice). Dotted line represents end of treatment, log-rank Mantel Cox test. (C) Survival of KPAR^G12C Ifngr2^−/− lung tumor–bearing mice after treatment with vehicle or MRTX1257 (50 mg/kg; n = 7 mice per group). Dotted lines represent treatment start and end, respectively, log-rank Mantel Cox test. (D) qPCR analysis of KPAR^G12C WT or Ifngr2^−/− lung tumors treated with vehicle or MRTX1257 for 4 days (n = 6 tumors per group, mean ± SEM, 2^−ΔΔCt). Each dot represents a lung tumor, one-way ANOVA. (E) Survival of KPAR^G12C Ifngr2^−/− lung tumor–bearing mice after treatment with vehicle (+IgG control, n = 9 mice), anti-PD1 (10 mg/kg; n = 9 mice), MRTX1257 (50 mg/kg; +IgG control, n = 7 mice), or both (n = 7 mice). Dotted line represents start and end of treatment for MRTX (green) and anti-PD1 (orange), log-rank Mantel Cox test.