SOS1 inhibition enhances the efficacy of and delays resistance to G12C inhibitors in lung adenocarcinoma

Abstract

Clinical effectiveness of KRAS G12C inhibitors (G12Cis) is limited both by intrinsic and acquired resistance, necessitating the development of combination approaches. We found that targeting proximal receptor tyrosine kinase (RTK) signaling using the SOS1 inhibitor (SOS1i) BI-3406 both enhanced the potency of and delayed resistance to G12Ci treatment, but the extent of SOS1i effectiveness was modulated by both SOS2 expression and the specific mutational landscape. SOS1i enhanced the efficacy of G12Ci and limited rebound RTK/ERK signaling to overcome intrinsic/adaptive resistance, but this effect was modulated by SOS2 protein levels. Survival of drug-tolerant persister (DTP) cells within the heterogeneous tumor population and/or acquired mutations that reactivate RTK/RAS signaling can lead to outgrowth of tumor initiating cells (TICs) that drive therapeutic resistance. G12Ci drug tolerant persister cells showed a 2-3-fold enrichment of TICs, suggesting that these could be a sanctuary population of G12Ci resistant cells. SOS1i re-sensitized DTPs to G12Ci and inhibited G12C-induced TIC enrichment. Co-mutation of the tumor suppressor KEAP1 limits the clinical effectiveness of G12Cis, and KEAP1 and STK11 deletion increased TIC frequency and accelerated the development of acquired resistance to G12Ci in situ. SOS1i both delayed acquired G12Ci resistance and limited the total number of resistant colonies regardless of KEAP1 and STK11 mutational status. These data suggest that SOS1i could be an effective strategy to both enhance G12Ci efficacy and prevent G12Ci resistance regardless of co-mutations.

Figure 1.

SOS1i synergizes with G12Ci to drive transcriptional changes regulating MAPK and hypoxia pathways. A. Heat map of cell viability (top) and excess over Bliss (EOB, bottom) for H358 cells treated with increasing (semi-log) doses of the indicated G12Ci [ARS-1620 (10^−9.5 – 10⁻⁶ M), sotorasib (10^−10.5 – 10⁻⁷ M), or adagrasib (10^−10.5 – 10⁻⁷ M)], the SOS1i BI-3406 (10⁻¹⁰ – 10^−6.5 M) the combination of G12Ci + SOS1i in a 9 × 9 matrix under 3D spheroid culture conditions. Data are the mean from three independent experiments, each experiment had three technical replicates. B-C. The sum of excess over Bliss values over the 9 × 9 treatment matrix (C) or quantitation of the fold change in G12Ci potency (log α₂) versus % change in G12Ci efficacy (β_obs) (D) for H358 cells treated with a 9 × 9 matrix of G12Ci + SOS1i from B. D. MPAS score (top) and heat map of changes in expression of individual MAPK-regulated genes (bottom) from RNA sequencing of 3D spheroid-cultured H358 cells treated with G12Ci (10 nM) ± SOS1i (100 nM) for 6h or 72h (left) or NT or SOS2^KO H358 cells treated with G12Ci (10 nM) ± SOS1i (100 nM) for 6 h (right). Data are averaged from four biologic replicates. * p < 0.05, *** p< 0.001 for G12Ci treated vs. untreated; ### p < 0.001 for SOS1i treated vs. untreated. E-H. Venn diagram from GVSA analysis of 50 Hallmark MsigDB gene sets that identified gene sets that were differentially downregulated (p < 0.05) by G12Ci ± SOS1i after 6h (E) or 72h (F) treatment. GSEA (G) and GVSA analysis (H) for hypoxia associated genes in H358 cells left untreated (open black square) or treated with a SOS1i for 6h (open red triangle) (E), G12Ci ± SOS1i for 6h (closed black square [G12Ci] or red triangle [G12Ci + SOS1i]), or 72h (half-filled black square [G12Ci] or red triangle [G12Ci + SOS1i]) (F). I. Venn diagram from GVSA analysis of 50 Hallmark MsigDB gene sets that identified nine gene sets that were differentially downregulated (p < 0.05) by SOS1i ± SOS2^KO in H358 cells. J. GVSA analysis of MsigDB gene sets that identified hypoxia signatures from RNA sequencing of 3D spheroid-cultured H358 cells treated with G12Ci ± SOS1i for 6h or 72h. Data are averaged from four biological replicates. * p < 0.05 vs. untreated controls. K. Heat map of Z-scores for the 14 PROGENy pathway gene sets from RNA sequencing of 3D spheroid-cultured H358 cells treated with G12Ci ± SOS1i for 6h or 72h (left) or NT or SOS2^KO H358 cells treated with G12Ci ± SOS1i for 6 h (right). Data are averaged from four biological replicates. * p < 0.05 vs. untreated controls. L. VIPER score for ETS1, EGR1, MYC, JUN, HIF1α, and E2F1 targets from RNA sequencing of 3D spheroid cultured H358 cells untreated (open black square) or treated with a SOS1i for 6h (open red triangle), G12Ci ± SOS1i for 6h (closed black square [G12Ci] or red triangle [G12Ci + SOS1i]) or 72h (half-filled black square [G12Ci] or red triangle [G12Ci + SOS1i]). Data are presented as mean ± from four biologic replicates. * p < 0.05, ** p < 0.01, *** p < 0.001 vs untreated cells; ^ p < 0.05 for G12Ci vs. G12Ci + SOS1i treated for 72h.

Figure 2.

SOS2 expression determines the extent of G12Ci:SOS1i synergy. A. Heat map of cell viability (top) and excess over Bliss (EOB, bottom) for the indicated KRAS^G12C-mutated LUAD cell lines treated with increasing (semi-log) doses of the G12Ci adagrasib (10^−10.5 – 10⁻⁷), the SOS1i BI-3406 (10⁻¹⁰ – 10^−6.5) or the combination of G12Ci + SOS1i under 3D spheroid culture conditions. Data are the mean from three independent experiments, each experiment had three technical replicates. B-C. The sum of excess over Bliss values over the 9 × 9 treatment matrix (B) or quantitation of the fold change in G12Ci potency (log αα₂) versus % change in G12Ci efficacy (β_obs) (C) for the indicated NT (squares) or SOS2^KO (blue circles) cells treated with a 9 × 9 matrix of G12Ci + SOS1i at 10% serum (filled) or 2% serum (open), or treated with a 9 × 9 matrix of G12Ci + SHP2i at 10% serum (purple diamonds). *** p < 0.001 vs. NT cells treated with G12CI + SOS1i in 10% serum. D-E. Western blots of whole cell lysates (WCLs) from the indicated LUAD cell lines for SOS1, SOS2, tubulin, and β-actin (D) and plot of the relative ratio of SOS1:SOS2 versus the sum of the EOB values for cells treated with a 9 × 9 matrix of G12Ci + SOS1i in 10% serum conditions (E). Red circles indicate cell lines showing EOB significantly > 0; purple diamonds indicate cell lines showing EOB ≤ 0. F-G. Western blots of WCLs of 3D spheroid-cultured H358 or H1373 cells (F) or NT vs. SOS2^KO H1792 or H2030 cells (G) treated with G12Ci (10 nM) ± SOS1i (100 nM) for the indicated times. Western blots are for pERK, ERK, and β-actin. Western blots are representative of three independent experiments.

Figure 3.

Isolated ALDH^high populations are resistant to G12Ci but sensitive to combined G12Ci:SOS1i. A. Aldefluor staining for ALDH enzyme activity in the indicated KRAS^G12C-mutated cell lines for DEAB negative control (DEAB), untreated cells, or cells treated with 100 nM adagrasib or sotorasib for 72 hours. B. Aldefluor staining for ALDH enzyme activity in H358 cells and gaiting strategy for isolating ALDH^low (green) versus ALDH^high (dark blue) populations. C. G12Ci dose response curves for unsorted (grey), ALDH^low (green), and ALDH^high (dark blue) H358 cells treated with increasing doses of adagrasib (left) or sotorasib (right). Data are the mean ± sd from three independent experiments, each experiment had three technical replicates. D. Heat map of cell viability (top) and excess over Bliss (EOB, bottom) for the ALDH^high H358 cells treated with increasing (semi-log) doses of the G12Ci adagrasib (left) or sotorasib (right) (10⁻¹¹ – 10^−7.5), the SOS1i BI-3406 (10⁻¹⁰ – 10^−6.5) or the combination of G12Ci + SOS1i under 3D spheroid culture conditions in ALDH^high sorted H358 cells. Data are the mean from three independent experiments, each experiment had three technical replicates.

Figure 4.

SOS1i ± SOS2^KO prevents G12Ci-induced TIC outgrowth. A. TIC frequency from in situ extreme limiting dilution assays (ELDAs) of the indicated cell lines pre-treated with 100 nM adagrasib or sotorasib for 72 hours. ## χ² < 0.01 vs. untreated for SIC upregulation. B. TIC frequency from in situ ELDAs of H358 cells treated with the indicated SOS1i doses. * χ² < 0.05, ** χ² < 0.01 vs. NT untreated. C. TIC frequency from in situ ELDAs in unsorted (grey), ALDH^low (green), and ALDH^high (dark blue) H358 cells left untreated or treated with 100 nM BI-3406 (SOS1i). **χ² < 0.01 vs untreated; ## χ² < 0.01, ### χ² < 0.001 vs. unsorted cells; ^^^ χ² < 0.001 vs. ALDH^low cells. D. TIC frequency from in situ ELDAs in the indicated NT or SOS2^KO LUAD cell lines treated with 100 nM BI-3406 (SOS1i). * χ² < 0.05, ** χ² < 0.01 vs. NT untreated. E. TIC frequency from in situ ELDAs of the indicated NT or SOS2^KO cell lines pre-treated with adagrasib or sotorasib for 72 hours to upregulate TICs, and then left untreated or treated with BI-3406. Data are representative of three independent experiments.

Figure 5.

SOS1 inhibition limits the development of acquired G12Ci resistance. Multi-well resistance assay were performed as outlined in the Materials and Methods. A-B. G12Ci resistance to the indicated dose of adagrasib (A) or sotorasib (B) NT or SOS2^KO cells H358, H1373, H1792, or H2030 cells treated with G12Ci alone (NT black; SOS2^KO blue) or G12Ci + SOS1i (NT red; SOS2^KO purple) at 100 nM (lite) or 300 nM (dark). Data are pooled from three independent experiments. *** p < 0.001 vs. G12Ci alone; ### p < 0.001 for cells treated with 100 vs 300 nM SOS1i; ^^^ p < 0.001 for NT vs. SOS2^KO cells treated with SOS1i.

Figure 6.

KEAP1 and STK11 co-mutations regulate resistance to G12Ci + SOS1i. A-D. Heat map of cell viability (top) and excess over Bliss (EOB, bottom) for H358 NT cells or H358 cells where STK11 and/or KEAP1 were deleted treated with increasing (semi-log) doses of the G12Ci adagrasib (10⁻¹¹ – 10^−7.5), the SOS1i BI-3406 (10⁻¹⁰ – 10^−6.5) or the combination of G12Ci + SOS1i for four days under 3D spheroid culture conditions (A), adagrasib dose response curves (B) EC₅₀ values (C) in NT (black closed squares), STK11^KO (dark grey closed circles), KEAP1^KO (grey open squares), and STK11/KEAP1 DKO (light grey open circles) cells in the absence of SOS1i, and sum of excess over Bliss values over the 9 × 9 treatment matrix from A (D). * p < 0.05, p < 0.01 vs. NT cells treated with G12Ci + SOS1i. Data are the mean from three independent experiments, each experiment had three technical replicates. E. TIC frequency from in situ ELDAs in the indicated H358 cell lines left untreated or treated with 100 nM BI-3406 (SOS1i). ** χ² > 0.01, *** χ² < 0.001 for SOS1i treated cells vs. untreated controls for each cell line; ### χ² < 0.001 s vs. NT untreated cells; ^^ χ² < 0.01 for STK11^KO vs. either KEAP1^KO or STK11/KEAP1 DKO cells. F-G. Multi-well in situ resistance assays comparing the development of G12Ci adagrasib (F) or sotorasib (G) resistance between NT (black), STK11^KO (dark grey), KEAP1^KO (grey), and STK11/KEAP1 DKO (light grey) H358 cells. *** p < 0.001 compared to NT cells. H-I. Multi-well in situ resistance assays comparing the development of G12Ci adagrasib (H) or sotorasib (I) resistance in NT (black), STK11^KO (dark grey), KEAP1^KO (grey), and STK11/KEAP1 DKO (light grey) cells treated with G12Ci alone () or in the presence of SOS1i BI-3406 (300 nM, red lines). *** p < 0.001 for G12Ci + SOS1i vs. G12Ci treated cells.

Figure 7.

SOS1i targets the continuum of G12Ci resistant states. A. Intrinsic G12Ci resistance is driven by adaptive reactivation of RTK signaling due to a loss of ERK-dependent negative feedback. SOS1i targets rebound RTK signaling to limit adaptive G12Ci resistance. B. Cancer cells undergo non-genetic adaptation to G12Ci to both alter the redox environment and enhance alternative RTK signaling, both of which allow these ‘drug tolerant persister’ (DTP) cells to survive under therapeutic pressure. Within the DTP population, a subset of ‘tumor initiating cells’ (TICs) are capable of self-renewal and are thought to be the pharmacologic sanctuary driving that ultimately develop acquired resistance. SOS1i re-sensitizes DTPs to G12Ci and reduces TIC frequeny in G12Ci treated cultures. C. Acquired G12Ci resistance is often driven by RTK/RAS pathway reactivation by both genetic and non-genetic mechanisms. SOS1i both delayed the development of and reduced the frequency with which cultures acquired G12Ci resistance.