Search

Supplementary Table S4. The complete list of identified phosphopeptides after (pY) enrichment and TMT-based quantitative phosphoproteomics analysis in KRASG12C mutant lung cancer cell lines (n=8) with mean-centered log abundance values.

Supplementary Table S1. Relative quantification of peptides representing total KRAS, WT KRAS, free KRAS G12C protein and drug bound KRAS G12C Protein Levels.

Supplementary Table S9. The correlation analysis between ERBB2/ERBB3 and EMT score in SPORE 442 KRASG12C lung cancer mutants.

Supplementary Table S2. The complete list of identified phosphopeptides after (pY) enrichment in ARS-1620 treated cells with fold change quantisation and p-value calculation.

Supplementary Table S5. The complete list of identified proteins after TMT-based quantitative proteomic analysis in KRASG12C mutant lung cancer cell lines (n=8) with mean-centered log abundance values.

Supplementary Fig. S1. Differential response (2D and 3D culture) of KRASG12C specific covalent inhibitors (ARS-1620 and AMG510) in KRASG12C lung cancer cell lines and signaling rebound following KRASG12C inhibition. Supplementary Fig. S2. MetaCore literature network after ARS-1620 treatment in H358 cells. Supplementary Fig. S3. TMT design for expression and pY comparison in 8 KRASG12C lung cancer cell lines and two group comparison of phosphosites Supplementary Fig. S4. Analysis of CCLE data set for EMT markers and ERBB3 gene expression; and PCA representing 105 probeset TGFβ-EMT signatures (Gordian et al. 2019) translated into proteins expression to describe EMT activity in KRASG12C lung cancer cell lines. Supplementary Fig. S5. Impact of SHP2 inhibitors (SHP-099 and RMC-4550) on signaling in H358 cells and cell viability analysis following treatment with SHP2 inhibitors or in the combination with ARS-1620 in H358 and H1792 cells. Supplementary Fig. S6. Literature-based signaling network after KRASG12C inhibition in H1792 cells and cell viability analysis following treatment with ARS-1620 (1µM), AZD-4547 (1µM) and their combination in NSCLCG12C mutant cell lines. Supplementary Fig. S7. Effect of TGFβ stimulation on cell viability in response to KRASG12C inhibition and different combination strategies. Supplementary Fig. S8. Heatmap showing mean-centered log abundance expression values for indicated phosphosites/protein expression in KRASG12C mutant cell lines and E vs M 2-group comparison performed for FGFR1 gene expression in CCLE data set. Supplementary Fig. S9. Literature-based signaling network after KRASG12C inhibition in Calu1 cells and impact on signaling following treatment with ARS-1620 plus drugs known to target AXL Receptor.

Detailed Supplementary Methods

Legends to Supplementary Figures and Tables

Supplementary Table S3. The complete list of identified phosphopeptides after global (pS/T/Y) phosphopeptides enrichment in ARS-1620 treated cells with fold change quantisation and p-value calculation.

Supplementary Table S7. The mean-centered CCLE normalized log2 RNA-seq counts for indicated genes and TGFβ-EMT analysis of 15 KRASG12C NSCLC cell lines.

Purpose:Covalent inhibitors of KRASG12C specifically target tumors driven by this form of mutant KRAS, yet early studies show that bypass signaling drives adaptive resistance. Although several combination strategies have been shown to improve efficacy of KRASG12C inhibitors (KRASi), underlying mechanisms and predictive strategies for patient enrichment are less clear.Experimental Design:We performed mass spectrometry–based phosphoproteomics analysis in KRASG12C cell lines after short-term treatment with ARS-1620. To understand signaling diversity and cell type–specific markers, we compared proteome and phosphoproteomes of KRASG12C cells. Gene expression patterns of KRASG12C cell lines and lung tumor tissues were examined.Results:Our analysis suggests cell type–specific perturbation to ERBB2/3 signaling compensates for repressed ERK and AKT signaling following ARS-1620 treatment in epithelial cell type, and this subtype was also more responsive to coinhibition of SHP2 and SOS1. Conversely, both high basal and feedback activation of FGFR or AXL signaling were identified in mesenchymal cells. Inhibition of FGFR signaling suppressed feedback activation of ERK and mTOR, while AXL inhibition suppressed PI3K pathway. In both cell lines and human lung cancer tissues with KRASG12C, we observed high basal ERBB2/3 associated with epithelial gene signatures, while higher basal FGFR1 and AXL were observed in cells/tumors with mesenchymal gene signatures.Conclusions:Our phosphoproteomic study identified cell type–adaptive responses to KRASi. Markers and targets associated with ERBB2/3 signaling in epithelial subtype and with FGFR1/AXL signaling in mesenchymal subtype should be considered in patient enrichment schemes with KRASi.

Sotorasib demonstrated efficacy in patients with non-small cell lung cancer (NSCLC). However, a substantial proportion of patients do not respond. There is therefore a clear need for a predictive biomarker to guide treatment decisions. Proximity ligation assay (PLA) is an immunofluorescent-based approach for analysis of endogenous protein-protein interactions, in which the signal can be detected only when two target proteins are within 40 nm. This method measures drug-targetable signaling-associated protein complexes in cells and tissues. KRAS activity has recently been shown to be related to the sensitivity to sotorasib. We therefore conducted a preclinical study to evaluate whether RAS activation inferred by PLA might predict response to KRASG12C inhibitors. We first examined panRAS:CRAF PLA reaction for a KRASG12C-mutant H358 NSCLC cell line, by comparing a complete PLA reaction with both panRAS and CRAF antibodies and an incomplete PLA reaction in the absence of either panRAS or CRAF antibodies. We observed high panRAS:CRAF PLA intensity in H358 with both antibodies, which was abrogated in the absence of either panRAS or CRAF antibodies. We next evaluated the specificity of the PLA reaction by RNA interference. Transfection with sipanRAS or siCRAF resulted in a reduction in panRAS:CRAF PLA signal. We then performed panRAS:CRAF PLA with a panel of KRAS G12C-mutant NSCLC cell lines. The intensity of panRAS:CRAF PLA signal varied between these cell lines, with H358 and LU65 having higher and H1792 and LU99 having lower intensity of panRAS:CRAF PLA signal. To further understand the RAS-RAF interaction, we evaluated a CRAF-RAS binding domain pulldown assay for the cell lines. Different levels of CRAF-bound panRAS were observed in the cell lines, with H358 and LU65 having higher and H1792 and LU99 having lower level of CRAF-bound panRAS. The intensity of panRAS:CRAF PLA strongly correlated with CRAF-bound panRAS (r = 0.70, P = 0.037), suggesting that panRAS:CRAF PLA works sufficiently well to detect the RAS-RAF complex. We investigated the antitumor efficacy of AMG510 in vitro and found that a heterogeneous response to AMG510 was observed in the cell lines, with the 50% inhibitory concentration values of AMG510 for H358 and LU65 below 30 nmol/L and those for H1792, and LU99 above 10 μmol/L. The intensity of panRAS:CRAF PLA was strongly associated with the sensitivity of the G12C-mutated cells to AMG510 (r = 0.74, P = 0.01). We next examined sections of FFPE tumors from NSCLC cell line-derived xenograft models. The panRAS:CRAF PLA signal was higher in mice with sotorasib-sensitive H358 than with resistant H1792 (P = 0.02). Our results suggest that RAS:RAF PLA may have values as a predictive marker to identify which patients can obtain the greatest benefit from sotorasib. Citation Format: Ryoji Kato, Hitendra S. Solanki, Denis Imbody, Anurima Majumder, Yaakov Stern, Liznair Bridenstine, Joseph Johnson, Eric B. Haura. RAS:RAF proximity ligation assay may predict response to KRASG12C inhibitors in NSCLC [abstract]. In: Proceedings of the AACR Special Conference: Targeting RAS; 2023 Mar 5-8; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Res 2023;21(5_Suppl):Abstract nr A001.

Purpose: Sotorasib (AMG510) has demonstrated remarkable response in lung cancer patients with tumors driven by oncogenic KRASG12C mutation. However, recently published clinical data identified acquired mutations in RAS and other genomic alterations as potential mechanisms of resistance. The cause of resistance in more than half of the patient cohort was undetermined with genomic sequencing. We hypothesize that rewiring of signaling networks could be the newly acquired vulnerabilities in cells progressing after sotorasib treatment. Experimental Design: We performed whole exome sequencing, transcriptomic profiling, and mass spectrometry-based proteomics/phosphoproteomics of the parental and sotorasib resistant isogenic line of LU65 (LU65-AMGR). The omics analyses were integrated with functional screens to prioritize a set of druggable targets/pathways. Results: We did not observe obvious acquired mutations that drive resistance to sotorasib in LU65-AMGR cells. To investigate compensatory signaling critically important for the survival/growth of LU65-AMGR cells, we studied signaling perturbations using mass spectrometry-based phosphoproteomics approach. We identified 430 and 1,574 phosphosites differentially expressed in resistant cells (± 1.5-fold & p < 0.05) with phosphotyrosine (pY) and global phosphoproteome (pS/T/Y) enrichment, respectively. Analysis of phosphoproteomics data identified increased phosphorylation of multiple RTKs including EGFR, HER2, HER3, IGF1R, AXL and PDGFRA, including others. RAS-GTP pull-down show increased levels of RAS-GTP and NRAS-GTP in the resistant cells. RAS isoform quantification using parallel reaction monitoring (MRM) mass spectrometry further confirmed increased activity of NRAS in pull-down samples. Transcriptomic analysis revealed elevated genes responsible for anti-apoptosis pathways mediated by PI3K/AKT signaling. Enrichment of transcription factors such as AP1 and AP-2A could drive enhanced EGF expression and EGFR phosphorylation in sotorasib resistant cells. Consistent with enhanced RTK and RAS activity, Western blots confirmed higher phosphorylation of ERK and AKT in LU65-AMGR cells. We show higher EGFR phosphorylation in resistant cells when compared to the parental counterpart. While LU65 cells viability is reduced markedly by sotorasib combination with afatinib, an irreversible HER family inhibitor, resistant cells showed lesser effect on cell viability with afatinib combination. Consistent with our in vitro observations, the sotorasib and afatinib combination treatment significantly regressed tumor growth only in LU65 xenografts, but not in LU65-AMGR xenografts. Conclusions: Our data suggest that activation of multiple RTKs maintains RAS activity and PI3K signaling in LU65-AMGR cells. Thus, dual pan-RAS and PI3K inhibition could serve as promising strategy of treatment in relapsed tumors identified with WT RAS and PI3K signaling activation. Citation Format: Denis Imbody, Hitendra S. Solanki, Bina Desai, Paul A. Stewart, Yaakov Stern, Ryoji Kato, Anurima Majumder, Liznair Bridenstine, Aobuli Xieraili, Bin Fang, Lancia Darville, Fumi Kinose, John M. Koomen, Uwe Rix, Andriy Marusyk, Eric B. Haura. Wildtype RAS activity and PI3K signaling as new vulnerabilities in cells with acquired resistance to sotorasib [abstract]. In: Proceedings of the AACR Special Conference: Targeting RAS; 2023 Mar 5-8; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Res 2023;21(5_Suppl):Abstract nr B015.

The development of covalent inhibitors targeting KRASG12C mutations offers a precision medicine approach to a large cohort of lung cancer patients previously lacking opportunities for targeted therapy. Despite promising initial responses ranging from disease control to partial response in most patients undergoing treatment, the efficacy of these inhibitors has been limited by a short duration of response. Several strategies have been advanced to delay or prevent acquired drug resistance in lung cancer patients treated with KRASG12C inhibitors. Many of these involve co-administration of other targeted therapies to prevent activation of Ras-dependent mitogen activated protein kinase (MAPK) signaling, which is a hallmark response to KRASG12C inhibitors. While this approach has proven effective in some patients, combination clinical trials and postmortem analysis of tumors that progressed on KRASG12C inhibitors demonstrate that non-MAPK-dependent mechanisms of resistance remain a major challenge. We have developed a panel of lung cancer cell lines with acquired resistance to the KRASG12C inhibitor sotorasib. While Erk phosphorylation is observed in all resistant models, treatment with a panel of Ras and MAPK pathway inhibitors demonstrates differential dependence on MAPK pathway activity, including complete MAPK independence in a resistant model derived from the H358 cell line. Mutation, transcription and phosphoproteomic analysis revealed that these resistant models exhibit metabolic reprogramming and activation of a variety of DNA damage response pathways. Based on differential phosphorylation of the transcriptional elongation kinases CDK12 and CDK13 in KRASG12C-treated cells, we hypothesized that inhibiting these kinases might prevent or revert acquired drug resistance. Our prior studies demonstrate that CDK12/13 inhibition reduces expression of a broad spectrum of DNA damage response and metabolic genes, and RNA sequencing confirms that these pathways are sensitive to the CDK12/13 inhibitor SR-4835 in both sotorasib-sensitive and -resistant cells. Resistant cells show differential responses to specific DNA damage-directed therapies and cell-line specific metabolic reprogramming, but up-front combinatorial treatment with SR-4835 prevents the development of acquired resistance to sotorasib across these models. Combined treatment with sotorasib and SR-4835 does not prevent Erk phosphorylation rebound; thus, the mechanism of SR-4835 in suppressing acquired resistance is independent of MAPK pathway reactivation. In vivo combination of sotorasib and SR-4835 potentiates drug response and prevents drug resistance in the H358 cell line, which can develop both MAPK-dependent and -independent drug resistance. These results demonstrate that targeting transcriptional elongation kinases in combination with sotorasib may benefit a broader spectrum of patients than MAPK-directed combination therapies and provide rationale for suppressing the DNA damage response and metabolic reprogramming as an approach to extend the efficacy of KRASG12C inhibitors. Citation Format: Yaakov E. Stern, Pompom Ghosh, Hitendra S. Solanki, Denis Imbody, Liznair Bridenstine, Hannah L. Walker-Mimms, John W. Mosior, Andrii Monastyrskyi, Derek R. Duckett, Eric B. Haura. Targeting transcriptional elongation kinases prevents adaptation to KRASG12C inhibitors in both MAPK-dependent and -independent models of acquired resistance [abstract]. In: Proceedings of the AACR Special Conference: Targeting RAS; 2023 Mar 5-8; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Res 2023;21(5_Suppl):Abstract nr B035.

Despite the presence of the KRASG12C oncogene, not all patients respond to KRASG12C inhibitors (e.g., sotorasib, adagrasib) and duration of response has been variable. We hypothesized that subtypes of KRASG12C lung cancers could be responsible for such differential effects. To define subtypes, we performed targeted exon sequencing, transcriptomics, proteomics, and phosphoproteomics on 46 frozen surgically resected KRASG12C lung adenocarcinomas. We employed unbiased consensus clustering methods, pathway analysis, and PTM-SEA to identify kinase signatures from phosphoproteomics data. Transcriptomics data revealed 4 clusters. Pathway enrichment showed that gene cluster (GC) 1 was enriched for immune pathways including interferon gamma response. Epithelial-mesenchymal transition (EMT), assessed by a TGF-β gene signature, was also enriched in GC1. 100% of GC1 overlapped the Proximal-Inflammatory subtype (Wilkerson et al.) and 53% with the mesenchymal subtype (Daemen et al.). GC2 was enriched for inflammatory response and TNF-alpha signaling. GC3 represented a metabolic phenotype with enriched xenobiotic metabolism and other metabolism pathways. 83% of GC3 overlapped with the Terminal Respiratory Unit subtype (Wilkerson et al.) and 56% with the proliferative subtype (Daemen et al.). GC4 only had 5 samples and did not have any significant enrichment. We next identified four proteomic clusters with distinct biology. Protein Cluster 1 (PC1) was enriched for immune pathways, PC2 was enriched for EMT, and PC3 was enriched for TNF/NF-κβ signaling. PC4 was enriched for xenobiotic and fatty acid metabolism and had significantly more KEAP1 mutations. We observed concordance of protein based clusters with mRNA based clusters, as PC1 largely overlapped with GC1 (70%) and PC4 largely overlapped with GC3 (86%). Cell type inference using xCell largely recapitulated the immune subtypes identified by transcriptomics and proteomics. GC1/PC1 had significantly more CD8+ effector memory T-cells, CD4+ Th2 T-cells, and M1 macrophages, while GC3/PC4 had significantly less of these same cell types. Collectively, GC1/PC1 represents 43% of the cohort and was defined by enrichment of immune pathways, elevated immune populations, and significantly higher PD-L1. Finally, we leveraged the phosphoproteomics data to determine whether signaling activity differs across subtypes. Using PTM-SEA, we identified phosphoproteomic signatures of activity for 50 kinases and pathways. We identified 26% of samples (N=12) with high EGFR activity, which is known to affect sensitivity of cells to KRASG12C inhibitors. The EGFR-high samples were observed across all genomic, transcriptomic, and proteomic subtypes. Our results suggest concordance of some transcriptomic and proteomic subtypes, including an immune subtype, while a metabolic protein subtype and EGFR high signaling subtype may provide additional subclasses with potential relevance for treatment responses. Additional phosphoproteomics signatures such as PLK1 and AKT1 may identify subtypes amenable to combination therapy approaches. Citation Format: Paul A. Stewart, Hitendra Solanki, Eric Welsh, Yaakov Stern, Denis Imbody, Yonghong Zhang, Bruna Pellini, Bin Fang, Sean Yoder, Steven Eschrich, Jamie Teer, John Koomen, Eric B. Haura. Proteogenomic landscape of KRASG12C lung adenocarcinomas reveals new subtypes and potential combination therapies [abstract]. In: Proceedings of the AACR Special Conference: Targeting RAS; 2023 Mar 5-8; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Res 2023;21(5_Suppl):Abstract nr A007.

More than 50 years after the discovery of RAS family proteins, which harbor the most common activating mutations in cancer, the U.S. Food and Drug Administration approved the first direct allele-specific inhibitor of mutant

Covalent inhibitors selectively targeting the KRASG12C mutation offer a promising new therapeutic opportunity for non-small cell lung cancer (NSCLC) patients, with partial or complete response observed in up to 40% of cancers harboring this lesion. As observed for other precision medicines targeting oncogenic drivers, resistance to these agents develops upon prolonged treatment. We have developed cell-based models of acquired resistance to the KRASG12C inhibitor sotorasib by serial passage of sotorasib-sensitive NSCLC cell lines in the presence of the drug. We have observed dynamic phosphorylation of the transcriptional elongation kinases CDK12 and CDK13 as well as hyperphosphorylation of a network of DNA damage response (DDR) proteins in sotorasib-resistant cell lines by phosphoproteomic profiling of resistant and sensitive cells. This is accompanied by elevated sensitivity to DDR pathway and CDK12/13 (e.g., SR-4835) inhibitors. Combined treatment of NSCLC cell lines with sotorasib and SR-4835 delays or prevents adaptation to sotorasib and the development of acquired resistance. Furthermore, gene expression profiling of SR-4835-treated cells reveals that CDK12/13 inhibition suppresses both DDR gene expression and metabolic pathways associated with resistance to sotorasib. This reflects synergy between SR-4835 and inhibitors targeting the DDR pathway in breast cancer cell lines and demonstrates that transcriptional elongation is a critical vulnerability for cells undergoing adaptation to KRASG12C inhibitors via DDR-dependent mechanisms. Citation Format: Yaakov E. Stern, Pompom Ghosh, Hannah L. Walker-Mimms, John W. Mosior, Denis Imbody, Hitendra S. Solanki, Andrii Monastyrskyi, Derek R. Duckett, Eric B. Haura. CDK12/13 inhibition antagonizes resistance to KRASG12C inhibitors. [abstract]. In: Proceedings of the AACR Special Conference: Cancer Epigenomics; 2022 Oct 6-8; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2022;82(23 Suppl_2):Abstract nr B028.

Purpose: Covalent inhibitors of KRASG12C specifically target tumors driven by this form of mutant KRAS, yet early studies show that bypass signaling drives adaptive resistance. Although several combination strategies have been shown to improve efficacy of KRASG12C inhibitors (KRASi), underlying mechanisms and predictive strategies for patient enrichment are less clear. Experimental Design: We performed mass spectrometry–based phosphoproteomics analysis in KRASG12C cell lines after short-term treatment with ARS-1620. To understand signaling diversity and cell type–specific markers, we compared proteome and phosphoproteomes of KRASG12C cells. Gene expression patterns of KRASG12C cell lines and lung tumor tissues were examined. Results: Our analysis suggests cell type–specific perturbation to ERBB2/3 signaling compensates for repressed ERK and AKT signaling following ARS-1620 treatment in epithelial cell type, and this subtype was also more responsive to coinhibition of SHP2 and SOS1. Conversely, both high basal and feedback activation of FGFR or AXL signaling were identified in mesenchymal cells. Inhibition of FGFR signaling suppressed feedback activation of ERK and mTOR, while AXL inhibition suppressed PI3K pathway. In both cell lines and human lung cancer tissues with KRASG12C, we observed high basal ERBB2/3 associated with epithelial gene signatures, while higher basal FGFR1 and AXL were observed in cells/tumors with mesenchymal gene signatures. Conclusions: Our phosphoproteomic study identified cell type–adaptive responses to KRASi. Markers and targets associated with ERBB2/3 signaling in epithelial subtype and with FGFR1/AXL signaling in mesenchymal subtype should be considered in patient enrichment schemes with KRASi.