RAS-ON inhibition overcomes clinical resistance to KRAS G12C-OFF covalent blockade

Abstract

Selective KRAS^G12C inhibitors have been developed to covalently lock the oncogene in the inactive GDP-bound state. Two of these molecules, sotorasib and adagrasib, are approved for the treatment of adult patients with KRAS^G12C-mutated previously treated advanced non-small cell lung cancer. Drug treatment imposes selective pressures leading to the outgrowth of drug-resistant variants. Mass sequencing from patients' biopsies identified a number of acquired KRAS mutations -both in cis and in trans- in resistant tumors. We demonstrate here that disease progression in vivo can also occur due to adaptive mechanisms and increased KRAS-GTP loading. Using the preclinical tool tri-complex KRAS^G12C-selective covalent inhibitor, RMC-4998 (also known as RM-029), that targets the active GTP-bound (ON) state of the oncogene, we provide a proof-of-concept that the clinical stage KRAS^G12C(ON) inhibitor RMC-6291 alone or in combination with KRAS^G12C(OFF) drugs can be an alternative potential therapeutic strategy to circumvent resistance due to increased KRAS-GTP loading.

Fig. 1. Characterization of LUAD disease progression upon sotorasib (AMG-510) treatment.

A Timeline of the patient clinical record. B Computed tomography (CT scan) images showing the evolution of the patient lesions along the sotorasib treatment, including baseline (left), 6 weeks (stable disease, SD, center) and 12 weeks (progressive disease, PD, right). The sequential scans demonstrate an initial stable disease on target lesions: Fourth right rib (top), left pectoral lesion (center), and umbilical lesion (bottom). The latter progressed with right pleural effusion and ascites appearance (white arrows), while the remaining target lesions persisted as stable disease (yellow arrows). C Scatter plot (D) describing the acquired mutations and their allelic frequencies shared across biopsies on treatment (SD, 6 weeks) and at progression (PD, 12 weeks).

Fig. 2. Genomic characterization of longitudinal samples from sotorasib-resistant KRAS^G12C LUAD.

A Schematic overview of the longitudinal samples. B Heatmap showing the variant allelic frequency (VAF) of somatic mutations present in at least two of the four longitudinal samples. The top annotation heatmap depicts the type of mutation, the cognate damaging effect according to CancerVar (i.e., tier), and the classification of the corresponding gene as cancer genes of lung or other tumors according to the Network of Cancer Genes. C Scatter plots show the VAF of selected mutations across the longitudinal samples. D View of chromosomal amplifications and losses in the longitudinal samples. Oncogenes (red) and tumor suppressor genes (blue) undergoing amplifications and losses in at least one sample are shown. Loss of heterozygosity (LOH) events are depicted in yellow. E Reconstruction of clonal evolution across the longitudinal samples. The evolution of the major clones and minor subclones are shown as colored branches. Tree nodes show samples containing the corresponding clones. F Fish plot visualizes clonal evolution across longitudinal samples. Color codes represent clones detected as in (E). Source data are provided as a Source Data file.

Fig. 3. The acquisition of TGFBR2^C421Y and the increased allelic frequency of KEAP1^G476W do not drive resistance to sotorasib in PDOX-derived cells.

A Cell viability assay (MTT) was performed with H358, H23, H1792, H2030, H2122, and PDOX-derived (DSFC3B/C-R) cells treated with the indicated concentrations of sotorasib for 72 h. B Western blot analysis of TGFβ signaling activation following treatment with TGFβ (5 ng/mL) for 1 h. Extracts were obtained from parental PDOX cells as well as those overexpressing exogenous TGFBR2 WT. KRAS^G12C cell lines (H23, H2030, and H358) were used as control. C Synergy matrix comparing parental and PDOX cells overexpressing TGFBR2 WT after 72 h in the presence of the indicated concentrations of sotorasib and TGFβ. D Proliferation curves of PDOX cells overexpressing exogenous TGFBR2 WT and treated with TGFβ (5 ng/ml) for the indicated time points. E Cell viability assay (MTT) was performed with PDOX cells overexpressing exogenous TGFBR2 WT and treated with the indicated concentrations of sotorasib in the presence or absence of TGFβ (5 ng/ml) for 72 h. F Western blot analysis of KEAP1 expression in PDOX cells infected with the indicated inducible constructs (KEAP1 WT or G476W) and treated with doxycycline 1 mg/ml for 7 days. G Proliferation curves of PDOX cells infected with the inducible KEAP1 WT or G476W construct and treated with doxycycline (1 mg/ml) in the presence or absence of sotorasib (1 µM) for the indicated time points. H Cell viability assay (MTT) performed with PDOX cells infected with the inducible KEAP1 WT or G476W construct and treated with the indicated concentrations of sotorasib in presence of doxycycline (1 mg/ml) for 72 h. All data were presented as mean ± SEM from three independent experiments. Representative western blots of two independently performed experiments are shown. Source data are provided as a Source Data file.

Fig. 4. MAPK adaptive activation in patients resistant to KRAS inhibitors.

A Volcano plot showing differentially expressed genes in pre- versus post-treatment samples in patients who developed resistance to KRAS inhibitors. Genes of interest are color-coded by pathway according to the Molecular Signatures Database (MSigDB) collections. Differential expression across sample groups was analyzed using the R package edgeR. The threshold for calling a gene as differentially expressed was at least a twofold change (log2FC ≥1 or log2FC ≥−1) with an FDR <0.05. B Gene set enrichment analysis showing pathways that are significantly (Q < 0.05) up and downregulated in post versus pre-treatment samples. NES normalized enrichment score. C Single sample gene set enrichment analysis (ssGSEA) showing enrichment scores of selected pathways involved in the RAS/ERK (upper) and in PI3K/AKT (lower) pathways in paired samples from patients with pre and post-KRAS inhibition tumor samples. Source data are provided as a Source Data file.

Fig. 5. Sotorasib resistance is characterized by increased expression and activation of KRAS.

A Cell viability assay (CellTiter-Glo) comparing PDOX-derived cell lines permanently maintained in 1 µM sotorasib (DSFC3B/C-R) versus those subjected to a 3-week drug withdrawal period (DSFC3B/C). Both cell lines were treated with the indicated concentrations of sotorasib for 72 h. Data were presented as mean ± SEM from three independent experiments. B Apoptosis of sotorasib sensitive (DSFC3B/C) and resistant (DSFC3B/C-R) PDOX-derived cell lines cells was analyzed by annexin-V staining and FACS analysis following treatment with 10 µM sotorasib for 24, 48, and 72 h. Data were shown as a representative experiment from three independent experiments. C Western blot analyses comparing both KRAS and KRAS-GTP levels in PDOX-derived cells maintained in 1 µM sotorasib (DSFC3B/C-R) and following a 3-week drug withdrawal period (DSFC3B/C). Representative western blots of two independently performed experiments are shown. D Bar plots show VAF of KRAS G12C in PDOX-derived cell lines permanently maintained in 1 µM sotorasib (DSFC3B/C-R) versus those subjected to a 3-week drug withdrawal period (DSFC3B/C) quantified by ddPCR at the RNA (cDNA) level. E Scatter plots depict the B-allele frequency (BAF, i.e., the fraction of the signal coming from the allele labeled as B) and the Log R Ratio (LRR, i.e., the total intensity of signal) measured for SNPs on the p-arm of chromosome 12 by the OncoScan SNP-array on sotorasib sensitive (DSFC3B/C) and resistant (DSFC3B/C-R) PDOX-derived cell lines. F Western blot analysis of RAS expression in DSFC3B/C-R cells after a drug (sotorasib 1 µM) withdrawal period of 24, 48, or 72 h. Representative western blots of two independently performed experiments are shown. Source data are provided as a Source Data file.

Fig. 6. RMC-4998 and BI-2865 exert differential efficacy in sotorasib-resistant PDOX cells in a wild-type KRAS-dependent manner.

A, B Cell viability assay (CellTiter-Glo) and Incucyte proliferation assays performed with PDOX-derived (DSFC3B/C-R) cells treated with 1 μM sotorasib, BI-2865 and RMC-4998 for 96 h. Data were presented as mean ± SD from three independent experiments and were analyzed using two-way ANOVA followed by Tukey’s multiple comparisons post-test. C Cell viability assay (CellTiter-Glo) of PDOX-derived cells infected with wild-type KRAS inducible constructs and treated with increasing doses of doxycycline (up to 1 mg/ml) together with sotorasib, BI-2865 and RMC-4998 for 72 h. Data were presented as mean ± SEM from three independent experiments. Source data are provided as a Source Data file.

Fig. 7. Sotorasib-resistant PDOX cells are sensitive to KRAS(ON) inhibitors in vivo.

A Tumor growth (left) of sotorasib-resistant DSFC3C-R cells following subcutaneous implantation in mice (≥3) and subjected to the indicated treatments (30 and 50 mg/kg qd for sotorasib and RMC-4998 respectively). Each line indicates the volume fold change variation compared to the baseline of each individual tumor. Data were analyzed using one-way ANOVA followed by Tukey’s multiple comparisons post-test on the last time point. Representative tumor images at the end of the experiment are also shown (right, scale bar 1 cm). B Representative tumor sections following hematoxylin/eosin staining as well as immunostaining with the indicated antibodies. Scale bar 100 μm. C Western blot analysis of independent tumor extracts at the end of the experiment shown in (A). Low mobility KRAS bands are caused by covalent drug binding. Source data are provided as a Source Data file.

Fig. 8. Sotorasib + RMC-4998 combination shows superior efficacy compared to a single agent in vitro and in vivo.

A Synergy matrix in resistant PDOX cells in the presence of the indicated concentrations of sotorasib, BI-2865, or RMC-4998. Data were shown as a representative experiment from six replicates in two independent experiments. B Incucyte proliferation assays performed with PDOX-derived (DSFC3C-R) cells treated with 1 μM sotorasib, 0.5 μM BI-2865, and 0.2 μM RMC-4998 as single agents or in combination for 120 h. Data were presented as mean ± SD from three independent experiments and were analyzed using two-way ANOVA followed by Tukey’s multiple comparisons post-test. C Western Blot analysis of PDOX-derived (DSFC3C-R) cells treated with 1 μM sotorasib, 0.5 μM BI-2865, and 0.2 μM RMC-4998 as single agents or in combination for 24 h. Representative western blots of two independently performed experiments are shown. D Violin plot (left) representing tumor volume fold change of sotorasib-resistant DSFC3C-R cells following subcutaneous implantation in mice (≥3) and subjected to the indicated treatments with sotorasib, RMC-4998 (100 mg/kg qd) or the combination for 3 weeks. Representative tumor images at the end of the experiment are also shown (right, scale bar 1 cm). Data were analyzed using one-way ANOVA followed by Tukey’s multiple comparisons post-test. Source data are provided as a Source Data file.