RAS genes encode small GTPase proteins that regulate cell proliferation in response to growth factor stimuli1,5. Cancer-associated KRAS mutations are found frequently in NSCLC, colorectal cancers (CRC), and pancreatic ductal adenocarcinoma (PDAC)2, the top three leading causes of cancer deaths in the United States6. These activating mutations drive tumor progression by stabilizing the active, GTP-bound state of RAS proteins and thereby increasing oncogenic flux through downstream effectors7. Analysis of CRISPR-Cas9 functional genetic screening data demonstrated that KRAS-mutated cancer cell lines are highly sensitive to disruption of the KRAS locus (Extended Data Fig. 1a), and KRAS mutation status was the only genetic feature exhibiting a significant correlation with KRAS dependency (Extended Data Fig. 1b). Similar results were observed for the other RAS isoforms, NRAS and HRAS in NRAS- and HRAS-mutated lines, respectively (Extended Data Fig. 1c-f). Furthermore, KRAS mutations at position 12 are both the most frequent KRAS alterations and associated with the highest degree of KRAS dependency as compared to other KRAS mutations (Fig. 1a). This result suggests that while many mutations at KRAS codons 12, 13, or 61 have transforming potential8,9, not all KRAS mutations are associated with equivalent KRAS oncogene dependence. Additionally, these data suggest that KRASG12X mutation is a genetic marker of RAS oncogene addiction (Extended Data Fig. 1g) and highlight a patient population that may derive particularly high benefit from a targeted inhibitor of these oncogenic RAS variants.
RMC-7977 discovery and development
RAS proteins have historically been recalcitrant drug targets2,3, though progress in targeting the inactive, GDP-bound state of KRASG12C has resulted in two regulatory approvals (sotorasib and adagrasib)10,11. We recently described RMC-4998 and RMC-6291, two covalent tri-complex inhibitors designed to target the active-state of KRASG12C 12. These macrocyclic compounds were derived from sanglifehrin A, a natural product that binds to cyclophilin A (CYPA) with high affinity13. Upon binding CYPA, these compounds generate a neomorphic interface with affinity for active KRAS and achieve high selectivity for KRASG12C via covalent modification of the reactive thiol group introduced by the oncogenic mutation. The resulting CYPA:compound:KRAS tri-complex sterically blocks KRAS-effector interactions and disrupts downstream signaling.
Most RAS oncoproteins with missense mutations are not amenable to selective covalent targeting but could be susceptible to non-covalent inhibition by tri-complex formation with CYPA. We postulated that we could use structure-guided design to optimize the neomorphic interface between CYPA and RAS and generate a reversible inhibitor with broad activity against the active states of multiple RAS variants. Tri-complex formation requires two distinct binding events (Fig. 1b). First, compound binds to CYPA to form the binary complex (KD1). The binary CYPA:compound complex then binds to active RAS (KD2) to form a tri-complex structure in which CYPA sterically occludes RAS:effector interactions (Extended Data Fig. 2a-d). Both binding events are essential for tri-complex formation, and we sought to optimize both KD1 and KD2 to increase the potency of RAS inhibition, focusing initially on KRASG12V as a representative oncogenic mutant.
We selected Compound 1 (Fig. 1c) from the previous study12 as a promising starting point for design of reversible and orally bioavailable inhibitors of the active state of RAS. Compound 1 contains a CYPA-binding motif (Extended Data Table 1, KD1 = 862 nM) and forms reversible tri-complexes with KRASG12V (KD2 = 6,550 nM) that weakly disrupt the binding to the RAS-binding domain (RBD) of BRAF (EC50 = 4,400 nM). Compound 1 is cell-active, inhibiting RAS pathway activation (pERK EC50 = 632 nM and proliferation EC50 = 965 nM) in Capan-1 cells (PDAC, KRASG12V; Extended Data Table 1). Introduction of a thiazole moiety and concomitant scaffold rigidification through rotatable bond reduction and control of hydrogen bond donor count in the side chain yielded Compound 2, with improved affinity for CYPA (KD1 = 330 nM) and improved cellular potency (pERK EC50 = 31.6 nM; proliferation EC50 = 149 nM; Extended Data Table 1). Further reduction of peptidic character resulted in Compound 3, with increased binary complex affinity for KRASG12V (KD2 = 818 nM) as well as improved oral bioavailability in mice (m%F = 43.6), but reduced affinity for CYPA (KD1 = 6,270 nM) and reduced cellular potency (pERK EC50 = 124 nM; proliferation EC50 = 618 nM).
To address the reduced CYPA affinity of Compound 3, we introduced a piperazine moiety on the left-hand side of the scaffold to create a cation-pi interaction with W121 of CYPA. This modification not only enhanced CYPA binding affinity (Compound 4; KD1 = 605 nM), but also affinity for KRASG12V (KD2 = 292 nM) and cellular potency (pERK EC50 = 1.94 nM, proliferation EC50 = 14.2 nM) (Extended Data Table 1). Structure-guided optimization of a water network-mediated interaction with the Y32 backbone carbonyl of KRAS bound to a GTP analog (GMPPNP) (Fig. 1d) resulted in RMC-7977, a potent (KD1 = 195 nM; KD2 = 85 nM; pERK EC50 = 2.20 nM; proliferation EC50 = 0.421 nM) and orally bioavailable (m%F = 62.7) RASMULTI(ON) inhibitor (Extended Data Fig. 2g,h and Extended Data Table 1). RMC-7977 makes a cation-pi interaction between CYPA and the piperazine moiety, and additional hydrophobic and polar interactions are observed, including with the catalytic R55 (Fig. 1e).
Although neither RMC-7977 nor CYPA alone exhibited any measurable binding to GMPPNP-KRAS (Extended Data Fig. 2i,j), RMC-7977 makes multiple pi-pi and hydrophobic contacts with RAS in the SWI and SWII region once the tri-complex is formed (Fig. 1f). All residues in the binding site are identical among HRAS, KRAS, NRAS (Extended Data Fig. 2e), and RMC-7977 is equipotent across these RAS isoforms (Extended Data Table 1). KD2 measurements for all three WT RAS proteins were approximately 100 nM (KD2 for KRAS = 120 nM, NRAS: 98 nM, and HRAS = 90 nM).
A high-resolution co-crystal structure of RMC-7977 bound to CYPA and GMPPNP-bound KRAS was solved and shows a non-covalent tri-complex with an unoccupied groove containing the common oncogenic residues, G12, G13, and Q61, providing a structural basis for the ability of RMC-7977 to bind these variants (Fig. 1g). Further, RMC-7977 exhibited a consistent binding mode across all KRASG12X mutants tested (Extended Data Fig. 2f). KD2 values for the most common oncogenic RAS variants were all within 3-fold of those for WT proteins (Fig. 1h and Extended Data Table 1). The ability of tri-complex formation to sterically disrupt effector binding for the various mutants was also measured, revealing a good correlation between the KD2 measurements and the biochemical EC50 values for RAS-RAF disruption.
We used a live-cell nano-BRET kinetic assay to show that RMC-7977 induced equally rapid (t1/2 < 5 min., Fig. 2a) association between KRAS and CYPA and dissociation of the CRAF RBD from KRAS, consistent with direct targeting of the active-state of RAS in cells accompanied by steric inhibition of protein-protein effector engagement. EC50 measurements in this assay were in the single-digit nanomolar range across a panel of WT, G12, G13, and Q61-mutant KRAS proteins (Fig. 2b). Notably, the cellular potency for WT KRAS was modestly lower than potency for the oncogenic variants. Further, the cellular potencies for KRAS:CYPA association were approximately 5-50x higher than the corresponding biochemical KD2 measurements (Extended Data Table 1). An increase in cellular potency compared to biochemical potency is expected based on the tri-complex mechanism of action, in which binding to abundant CYPA drives high intracellular concentrations of CYPA:RMC-7977 binary complexes, as is evidenced by accumulation of RMC-7977 in a CYPA-dependent manner in AsPC-1 cells (Extended Data Fig. 3a). Furthermore, biochemical and cellular potencies correlate well when adjusted to reflect the estimated intracellular concentration of binary complexes formed in cells (Extended Data Fig. 3b,c).
To verify that formation of the CYPA:RMC-7977 binary complex is essential for cellular activity, we used a competitive CYPA inhibitor14 or genetically knocked out PPIA, the gene that encodes CYPA. These studies confirm that CYPA binding is required for inhibition of RAF-MEK-ERK signaling and proliferation by RMC-7977 in NCI-H441 (KRASG12V, NSCLC) and AsPC-1 (KRASG12D, PDAC) cells (Extended Data Fig. 3d-g). As a control, disruption of the PPIA locus did not affect sensitivity to the MEK1/2-selective inhibitor, trametinib, which does not rely on the tri-complex mechanism of action (Extended Data Fig. 3h,i). We further investigated whether exogenous CYPA expression could restore RMC-7977 sensitivity in NCI-H358 (KRASG12C, NSCLC) cells lacking PPIA. We investigated two clones expressing either low or high CYPA levels through a doxycycline-inducible promoter (Extended Data Fig. 4a). Inhibition of pERK and proliferation was CYPA dependent, and CYPA high cells were 3- and 8-fold more sensitive to RMC-7977 inhibition of signaling and cell proliferation, respectively, compared to CYPA low cells (Fig. 2c, Extended Data Fig. 4b). RMC-7977 accumulation was significantly greater in CYPA high compared to CYPA low cells treated with 10 nM RMC-7977, with no difference at 1 µM at which concentration binding to cellular CYPA is estimated to approach saturation (Fig. 2d). Collectively these observations suggest that the cellular potency of RMC-7977 is dependent on intracellular concentration of binary complexes, driven by intracellular CYPA protein expression. CYPA is highly abundant in cells (median concentration = 12.3 µM) as measured across a panel of 15 cell lines (Extended data Fig. 4c), and CYPA expression was higher in cell line-derived xenograft (CDX) tumors in vivo compared to the corresponding cells cultured in vitro (Extended Data Fig. 4d). Finally, CYPA is abundantly expressed across cancer types and exhibits low inter-patient variation in expression12, suggesting that tumor expression of CYPA is unlikely to be limiting for RMC-7977 potency.
Although RMC-7977 exhibited similar activity for WT and mutant RAS variants in biochemical assays and the live-cell nano-BRET assay, many factors can influence the downstream consequences of RAS inhibition in cells. To assess the spectrum of RMC-7977 activity against common KRAS variants in cells, we evaluated a panel of matched mouse embryo fibroblasts (MEFs) null for all three Ras genes (RAS-less) where proliferation was restored with ectopic expression of WT or mutationally-activated KRAS or BRAFV600E (Fig. 2e)15. RMC-7977 suppressed pERK in all KRAS-expressing cells, but not BRAFV600E-expressing RAS-less MEFs, which lack all RAS proteins and are not RAS dependent, indicating that pERK suppression is KRAS-dependent. Interestingly, we noted that pERK suppression by RMC-7977 typically appeared complete across cells expressing various KRASG12X mutants, but consistently reached a plateau in WT KRAS, KRASG12A, KRASQ61L, KRASQ61R, KRASG13D, and KRASA146T cells; in contrast and as expected, trametinib reduced pERK similarly in all cells, including the BRAFV600E MEFs (Extended Data Fig. 5c). These differences indicate that KRAS genotype could affect cellular response to direct RAS inhibition, and that the cellular response to RMC-7977 inhibition is not equivalent to that of MEK inhibition.
We also compared RMC-7977 activity in cancer cells harboring various activating mutations in the RAS pathway, specifically oncogenic variants of KRAS, NRAS, EGFR, or BRAF. RAS-dependent (KRAS, NRAS, or EGFR mutated) cancer cells treated with RMC-7977 exhibited concentration-dependent inhibition of downstream signaling markers, including phosphorylation of RAF, ERK, and the ERK substrate RSK, and inhibition of proliferation in the low nanomolar range (Fig. 2f,g). KRAS mutant cells also showed durable pathway suppression up to 48 h, as well as apoptosis induction indicated by increased PARP cleavage (Extended Data Fig. 5a,b). No inhibition by RMC-7977 was observed in RAS-independent BRAFV600E-mutant A375 cells (Fig. 2f,g).
RMC-7977 is broadly active in RAS-addicted cancers
We next employed a cell viability assay in 818 human tumor cell lines of different genetic and histological subtypes in a pooled, multiplexed format (PRISM assay) to identify genetic features associated with RMC-7977 sensitivity or resistance. Oncogenic KRAS mutation status provided the most significant genetic marker of sensitivity to RMC-7977 (Fig. 3a). Similar results were observed for NRAS mutations, though no correlation with HRAS mutation status was detected due to the low representation of HRAS mutations (Extended Data Fig 6a,b). Unsurprisingly, among cell lines with BRAF mutations, BRAF Class I V600 mutations were the most abundantly represented and clearly associated with resistance. Cell lines with less common Class II or III mutations, which remain somewhat dependent on upstream RAS signaling and frequently co-occur with RAS mutations, were often sensitive to RMC-7977, as were many unclassified BRAF mutations (Extended Data Fig. 6c). As predicted by KRAS gene dependency (Fig. 1a), KRAS mutated cell lines were more sensitive to RMC-7977 compared to cell lines with other RAS pathway aberrations and cells with no identifiable RAS pathway alteration. Moreover, KRASG12X (and NRAS mutant) cells were particularly sensitive to RMC-7977 as compared to other RAS pathway genotypes (Extended Data Fig. 6d).
We then selected a second, focused panel of 183 individually arrayed human cancer cell lines enriched for RAS and RAS pathway mutations to interrogate RMC-7977 potency. KRASG12X mutant cells were again the most sensitive, followed by NRAS-, HRAS-, and non-G12 KRAS-mutant cell lines (Extended Data Fig. 6e). KRASG12X mutant cell lines were highly sensitive with a median EC50 of 2.40 nM. By comparison, non-G12 mutant KRAS cell lines showed ~10-fold reduced sensitivity (median EC50 = 25.1 nM) (Fig. 3b). Taken together, these data are concordant with our genetic analysis of KRAS dependence and support the on-target pharmacological activity of RMC-7977. In addition, NRAS and HRAS mutant cell lines (median EC50 = 6.76 nM), and cell lines with mutationally activated RTKs also responded to RMC-7977 inhibition, including those with mutations or fusions of EGFR, ERBB3, FGFR1, FGFR2, FGFR3, ROS1, RET, NTRK1, and ALK (median EC50 = 6.14 nM), and cell lines with WT MET gene amplification (median EC50 = 6.61 nM, Extended Data Fig. 6f). Cell lines with NF1 loss-of-function and PTPN11 mutations, which each cause activation of WT RAS signaling, were moderately sensitive (median EC50 = 28.1 nM).
We then assessed the pharmacodynamic and anti-tumor activity of RMC-7977 in vivo in the NCI-H441 CDX model of non-small cell lung cancer (KRASG12V, NSCLC). The relationship between the total tumor concentration of RMC-7977 and inhibition of a RAS pathway transcriptional target, DUSP6, in tumor lysates yielded an EC50 of 130 nM(Extended Data Fig. 7a), consistent with the measured KRASG12V KD2 of 85 nM (Extended Data Table 1), and with the model for tri-complex RAS inhibition discussed above. A single oral dose of 10 mg/kg RMC-7977 was sufficient to maximally suppress tumor DUSP6 levels (91%) at 8 h, which partially recovered over 48 h, concordant with the decrease in RMC-7977 tumor concentrations (Fig. 3c). Prolonged RMC-7977 exposure in tumors was observed in this and other subcutaneously implanted xenograft tumor models, resulting in a ~2- to 3-fold increase in overall exposure (AUC0-last) in subcutaneous tumors compared to blood (Extended Data Fig. 7b). Repeated daily administration of RMC-7977 at 10 mg/kg was well-tolerated based on body weights (Extended Data Fig. 7c) and resulted in 83% mean tumor regression following 28 days treatment in the NCI-H441 model. Notably, we saw similar anti-tumor activity following RMC-7977 treatment across a larger panel of 15 PDAC, CRC, and NSCLC CDX and patient-derived (PDX) models bearing KRASG12X mutations and co-mutations representative of the genomic landscape of patients with KRAS mutant cancers (Fig. 3e). RMC-7977 treatment resulted in mean tumor regression in 9/15 (60%) models after a 4 to 6-week treatment period (Fig. 3e) and was well tolerated in all (Extended Data Fig. 7c). Importantly, when we continued RMC-7977 treatment in these xenograft models for up to 90 days, the anti-tumor activity of RMC-7977 was found to remain significantly durable as the majority of regressions and even cytostatic responses were maintained. While the controls exhibited a short median time to tumor doubling of 7 days, RMC-7977 treated tumors did not reach a median time to tumor doubling (defined as tumor progression) on treatment in a Kaplan-Meier analysis of these results (Fig. 3f; Cox Proportional Hazard Ratio 0.004, 95% interval 0.0011-0.0191, P < 1 × 10−12).
MEK and ERK inhibitors have undergone extensive clinical testing as monotherapies or in combinations with other RAS pathway inhibitors in patients with KRAS or NRAS mutated cancers16. Despite encouraging preclinical results, these therapeutic strategies have been unsuccessful in the clinic17-19 to date, with therapeutic efficacy likely compromised by dose-limiting toxicities20-22. We compared the anti-tumor activity of RMC-7977 to cobimetinib (MEK inhibitor), RMC-4550 (SHP2 inhibitor), and the combination thereof in three KRASG12X models. At well tolerated doses, RMC-7977 induced deep regressions in all three models, whereas MEK and SHP2 inhibitors alone or in combination achieved only modest tumor growth inhibition at well tolerated and translatable doses (Fig. 3g). In these preclinical models of KRASG12X mutant cancers, direct targeting of active RAS with RMC-7977 elicited a differentiated and superior anti-tumor activity profile as compared to upstream and/or downstream vertical inhibition of the primary oncogenic RAS driver.
There are several potential mechanistic explanations for why RMC-7977 elicits significantly greater anti-tumor activity in KRASG12X-driven cancers compared to agents targeting upstream and/or downstream nodes in the RAS pathway. Directly targeting the RAS oncoprotein itself may exploit the high degree of oncogene addiction of KRASG12X (and NRAS) mutated cancer cells to a greater degree than targeting upstream and downstream signaling proteins (e.g. SHP2, MEK1/2, and ERK1/2). Furthermore, while MEK inhibition did not distinguish between WT and mutant RAS variants (Extended Data Fig. 5c), RMC-7977 exhibited modestly lower potency and incomplete WT RAS suppression compared to KRASG12X in cells (Fig. 2b,e and Extended Data Table 1). Additionally, the slow elimination of RMC-7977 observed in subcutaneous xenograft tumors relative to blood suggests differential effective distribution to tumors, which may contribute to a wider therapeutic index. Of note, CYPA mRNA expression is reportedly induced by hypoxia under control of HIF-1α and plays a critical role in tumorigenesis23,24. Consistent with these reports, subcutaneous xenograft tumors express elevated CYPA protein compared to cells grown in vitro under normoxic conditions (Extended Data Fig. 4d) and CYPA mRNA expression is increased in tumor cells25. Collectively these data support the notion that CYPA is critical for tumor maintenance and could affect tumor distribution and cellular retention of tricomplex inhibitors.
RMC-7977 can overcome clinically relevant resistance mechanisms to KRASG12C(OFF) inhibitors
While inactive-state KRASG12C inhibitors provide short-term therapeutic benefit for some, patients, most eventually relapse through acquired genetic or adaptive mechanisms of resistance26-29. Ryan et al. have reported that adaptive feedback reactivation of upstream RTK signaling through WT RAS limits the activity of KRASG12C inhibitors29. We observed analogous results in KRASG12D mutant PDAC cell lines treated with the KRASG12D selective inhibitor, MRTX1133, where pERK suppression seen at 2 h rebounded by 48 h after treatment. We hypothesized that RMC-7977 could address adaptive RAS signaling mechanisms that rely on increased active-state WT and mutant RAS proteins. Consistent with this hypothesis, RMC-7977 showed sustained pERK suppression in KRASG12D mutated PDAC cells in culture for 48 h, suggesting that RASMULTI inhibition can overcome adaptive feedback observed with mutant-selective inhibitor treatment (Fig. 4a). Similar sustained pERK suppression and induced PARP cleavage were also observed in a KRASG12V and KRASG12C mutant cancer cell lines (Extended Data Fig. 5b). To further address whether RMC-7977 can overcome resistance mediated by WT KRAS signaling, we utilized an engineered KRASG12C/WT loss-of-heterozygosity (LOH) model system in which the endogenous WT Kras copy is selectively deleted under control of tamoxifen and the ectopic KRASG12C allele is constitutively expressed. KRASG12C cells (LOH) were sensitive to adagrasib and KRASG12C/WT cells (non-LOH) were resistant, whereas both cell lines were sensitive to RMC-7977 regardless of WT Kras expression (Extended Data Fig. 8a,b). We therefore hypothesized that the concurrent suppression of WT and mutant RAS signaling could drive durable anti-tumor responses to RMC-7977 treatment in vivo. As described above, a 90-day treatment study in a series of KRASG12X xenograft models demonstrated marked and significant increase in time to tumor doubling from baseline as compared to controls (Fig. 3f).
The activity of RMC-7977 against multiple forms of oncogenic RAS suggests the potential for therapeutic benefit against resistance mechanisms involving secondary RAS mutations. We recently demonstrated that tri-complex KRASG12C inhibitors, such as RMC-4998, bind to RAS through a unique mechanism and a binding site distinct from the switch II pocket occupied by inactive-state KRASG12C inhibitors, such as adagrasib and sotorasib30-32. This enables these inhibitors to overcome resistance mediated by switch II pocket binding mutations such as those occurring at positions Y96 and H9526. Consequently, other tri-complex inhibitors, including RMC-7977, would be expected to overcome these same drug binding site mutations. We then tested whether the broad-spectrum RAS inhibitory activity of RMC-7977 could counter additional genetic resistance mechanisms observed in relapsed patients treated with KRASG12C inhibitors, including secondary oncogenic RAS mutations and RTK activation. Indeed, RMC-7977 inhibited RAS signaling and growth of an NCI-H358 (KRASG12C mutated NSCLC) clone with a concurrent NRASQ61K mutation that emerged in cells grown under continuous exposure to adagrasib in vitro (Fig. 4b,c and Extended Data Fig. 8c). RTK amplification and activating mutations can also cause RAS pathway reactivation through WT RAS proteins. We used an engineered system harboring doxycycline-inducible constructs of full-length and fusion RTKs previously detected in patients who progressed on adagrasib or sotorasib treatment26. Overexpression of WT or mutant RTKs in NCI-H358 cells (KRASG12C, NSCLC) conferred significantly reduced sensitivity to adagrasib (proliferation inhibition EC50 shift for WT EGFR = 42-fold, EGFRA289V = 153-fold, HER2 = 51-fold, FGFR2 = 18-fold, RETM918T = 34-fold) while retaining sensitivity to RMC-7977 (Fig. 4d,e and Extended Data Fig. 8d,e). Similar results were observed when oncogenic RTK fusion proteins (EML4-ALK, FGFR3-TACC3, and CCDC6-RET) were exogenously expressed in MIA PaCa-2 cells (KRASG12C, PDAC) (Fig. 4f,g and Extended Data Fig. 8f,g).
Finally, we examined RMC-7977 treatment in a KRASG12C mutated PDX model derived from a NSCLC patient who had achieved stable disease on sotorasib but quickly relapsed. Genomic alterations in this tumor include amplification of the WT KRAS allele accompanied by increased levels of GTP-KRAS33, which contributes to diminished response to sotorasib treatment at 50 mg/kg qd. RMC-7977 administered daily at 10 mg/kg resulted in significant anti-tumor activity, with 90% inhibition of tumor growth observed at D17 of treatment while sotorasib treatment induced only 47% tumor growth inhibition (Fig. 4i). In sum, these data indicate that both adaptive and acquired mechanisms of resistance to KRASG12C inhibitors that lead to RAS pathway reactivation are susceptible to inhibition by RMC-7977.
RMC-7977 extends the tri-complex inhibitor strategy to non-covalently target the active state of WT and multiple oncogenic RAS variant proteins, with particular activity against the range of common codon 12 mutants, thus offering therapeutic potential for a RASMULTI(ON) inhibitor across a spectrum of RAS-addicted cancers, including NSCLC, CRC, and PDAC. Evidence of robust, durable anti-tumor activity at well-tolerated doses across various RAS mutant xenograft models provides preclinical validation for the direct targeting of active RAS variants as a desirable therapeutic strategy. Furthermore, we demonstrate that concurrent inhibition of multiple oncogenic RAS variants and WT RAS in the same tumor cell with a reversible RASMULTI inhibitor such as RMC-7977 can overcome some of the resistance mechanisms recognized to limit the efficacy and durability for inactive-state KRASG12C inhibitors. The proximity of the RMC-7977 binding site to RAS mutational hotspots (residues G12, G13, and Q61) presents a unique opportunity to expand this approach further by designing additional mutant-selective tri-complex inhibitors. Moreover, RASMULTI inhibitors could also provide therapeutic benefit in combination with mutant selective KRAS inhibitors to improve anti-tumor response by blocking adaptive pathway reactivation and preventing escape through emergence of secondary oncogenic RAS or RTK mutations. RMC-6236, a first-in-class, clinical-stage RASMULTI inhibitor, is currently being evaluated (NCT05379985) in patients with solid tumors harboring KRASG12X mutations.