Why the Tri-Complex Mechanism Changes KRAS Targeting
Revolution Medicines’ RAS(ON) platform targets the active GTP-bound state of KRAS — a fundamentally different approach from the GDP-state trapping used by approved inhibitors sotorasib and adagrasib. Rather than waiting for KRAS to cycle into its inactive conformation, RMC-9805 recruits the chaperone protein cyclophilin A (CypA) to form a ternary complex with GTP-bound KRAS G12D, sterically blocking the effector binding sites that drive oncogenic signaling.
The tri-complex mechanism operates through three sequential steps: first, the inhibitor binds CypA to form a binary complex; second, this binary complex docks onto the switch-I and switch-II regions of active KRAS-GTP; third, the assembled ternary structure occludes binding by effector proteins including RAF, PI3K, and RALGDS, suppressing downstream signaling. This approach does not require GDP-state trapping and is therefore mechanistically orthogonal to all approved KRAS inhibitors.
A RAS(ON) inhibitor targets the active, GTP-bound (“ON”) state of RAS, in contrast to RAS(OFF) inhibitors such as sotorasib and adagrasib that trap KRAS in the inactive GDP-bound state. The RAS(ON) approach is relevant for all KRAS mutations, including those — like G12D — that do not offer a covalent handle in the switch-II pocket used by G12C inhibitors.
The platform was first validated with RMC-4998, a tri-complex inhibitor targeting GTP-bound KRAS G12C that established proof-of-concept and demonstrated subnanomolar cellular potency. A second-generation compound, RMC-6291, optimised the morpholine linker and entered Phase I/Ib clinical trials (NCT05462717), also showing activity against sotorasib- and adagrasib-resistant mutations including Y96D and H95Q. RMC-9805 represents the third generation: a hybridised architecture that combines the spirocyclic scaffold of RMC-4998 with the morpholine linker of RMC-6291, then adds an aziridine warhead to achieve covalent targeting of the aspartate residue unique to KRAS G12D.
Revolution Medicines’ tri-complex inhibitors target active GTP-bound KRAS by recruiting cyclophilin A (CypA) to form a ternary complex that sterically blocks effector protein binding at the switch-I and switch-II regions, suppressing downstream RAF, PI3K, and RALGDS signaling without requiring GDP-state trapping.
Structural Architecture and SAR Patterns of RMC-9805
RMC-9805’s molecular architecture is the product of iterative structure-activity relationship (SAR) optimisation across three compound generations, with each structural element traceable to a specific design decision. The compound integrates four distinct pharmacophoric modules into a single molecule capable of engaging two protein partners simultaneously.
The Four Structural Modules
The spirocyclic core — inherited from first-generation RMC-4998 — provides the rigid three-dimensional geometry required for optimal CypA binding. Spirocyclic scaffolds constrain rotational freedom, pre-organising the molecule into its bioactive conformation and reducing the entropic penalty of binding. The morpholine linker, optimised in second-generation RMC-6291, connects the CypA-binding domain to the KRAS-binding region; its geometry is critical for the correct spatial presentation of the warhead relative to Asp12. The aziridine warhead is the defining innovation of RMC-9805: a strained three-membered nitrogen-containing ring that reacts covalently with the carboxylate side chain of aspartate 12 in KRAS G12D. This electrophilic warhead exploits the nucleophilic character of Asp12, a residue not present in wild-type KRAS (which has glycine at position 12). Finally, the piperidine moiety added during RMC-9805 development contributes to selectivity; this motif was subsequently incorporated into the pan-RAS inhibitor RMC-6236.
The piperazine portion of RMC-9805 forms a critical salt bridge with Asp12 in KRAS G12D. This ionic interaction is absent in wild-type KRAS (which has Gly12) and is the primary driver of mutant selectivity. Combined with the covalent aziridine warhead, this dual-engagement strategy provides both potency and selectivity against the G12D variant.
Compound Evolution: From RMC-4998 to RMC-9805
The structural evolution across the RMC platform demonstrates systematic scaffold hybridisation. RMC-4998 established the spirocyclic scaffold and subnanomolar cellular potency for KRAS G12C(ON) inhibition. RMC-6291 optimised the morpholine linker and demonstrated clinical activity against sotorasib- and adagrasib-resistant mutations (Y96D, H95Q). RMC-9805 merged both elements and introduced the aziridine warhead to redirect selectivity from G12C to G12D. The piperidine moiety added in RMC-9805 was then carried forward into the pan-RAS inhibitor RMC-6236, where a thiazole linker replaced morpholine for improved pharmacokinetics and broader isoform coverage (KRAS G12A/D/V/R/S, NRAS, HRAS).
RMC-9805 is a covalent KRAS G12D(ON) inhibitor built from a spirocyclic core derived from RMC-4998, a morpholine linker from RMC-6291, an aziridine warhead that covalently modifies Asp12 in KRAS G12D, and a piperidine moiety that contributes to mutant selectivity via salt bridge formation with Asp12.
Explore the full patent landscape for KRAS G12D inhibitors and map competitor IP with PatSnap Eureka.
Search KRAS G12D Patents in PatSnap Eureka →Potency Data: Preclinical Efficacy Across KRAS G12D Models
RMC-9805 demonstrates nanomolar to sub-nanomolar potency in KRAS G12D-mutant cancer cell lines, with target engagement confirmed by cellular thermal shift assay (CETSA) and dose-dependent pathway suppression measured by pERK and pAKT inhibition. The compound shows activity in both 2D monolayer and 3D spheroid culture systems, and induces apoptosis in sensitive models.
In Vivo PDX Performance
The most compelling preclinical evidence for RMC-9805 comes from patient-derived xenograft (PDX) studies in pancreatic ductal adenocarcinoma (PDAC), the cancer type with the highest prevalence of KRAS G12D mutations. RMC-9805 achieved objective responses in 7 of 9 PDAC PDX models — a 78% objective response rate — with demonstrated tumor regression at tolerable doses. This performance in a notoriously difficult-to-treat tumour type is notable given that PDAC has historically shown poor response to most targeted therapies.
“RMC-9805 achieved objective responses in 7 of 9 PDAC patient-derived xenograft models — a 78% response rate in one of oncology’s most treatment-resistant tumour types.”
Synergistic activity was demonstrated in combination studies: RMC-9805 combined with anti-PD1 immunotherapy showed synergistic effects in immunogenic preclinical models, while combination with SHP2 inhibitors produced regression in colorectal cancer models non-responsive to monotherapy. These findings informed the combination cohorts included in the Phase I clinical trial design.
RMC-9805, Revolution Medicines’ covalent KRAS G12D(ON) inhibitor, achieved objective responses in 7 of 9 pancreatic ductal adenocarcinoma patient-derived xenograft models (78% objective response rate) and demonstrated synergy with anti-PD1 immunotherapy and SHP2 inhibitors in preclinical studies.
Pan-RAS Platform Clinical Validation: RMC-6236
While RMC-9805-specific Phase I data are still maturing, the closely related pan-RAS inhibitor RMC-6236 provides clinical validation for the tri-complex platform. In its Phase I/Ib trial (NCT05379985), RMC-6236 demonstrated an overall response rate of 20% in heavily pre-treated PDAC patients (n=46) and 38% in non-small cell lung cancer (NSCLC). The disease control rate was 87% across multiple doses, with a median time to response of 1.4 months. ctDNA analysis showed greater than 50% reduction in mutated allele fraction in 12 of 13 PDAC patients, confirming deep on-target activity. The safety profile was manageable, with rash, nausea, and diarrhoea as the principal toxicities. According to ClinicalTrials.gov, RMC-9805 (NCT06040541) is currently in dose escalation and expansion cohorts for advanced KRAS G12D-mutant solid tumours.
RMC-9805 vs. MRTX1133: Mechanistic and Structural Differentiation
Two distinct chemical strategies have emerged for targeting KRAS G12D: Revolution Medicines’ covalent GTP-state tri-complex approach (RMC-9805) and Mirati Therapeutics’ non-covalent GDP-state direct binding approach (MRTX1133). Understanding the structural and mechanistic differences between these compounds is essential for anticipating their respective clinical profiles, resistance liabilities, and combination strategies.
MRTX1133: Non-Covalent GDP-State Inhibition
MRTX1133 features a pyrido[4,3-d]pyrimidine scaffold with a protonated piperazine group that forms an ionic interaction with Asp12 in KRAS G12D. Its binding affinity is exceptionally high — a KD of approximately 0.2 pM (picomolar) — and it demonstrated less than 2 nM IC₅₀ in cellular assays, with 85% tumor regression in HPAC xenografts at 30 mg/kg BID. MRTX1133 binds the switch-II pocket of GDP-bound KRAS, making it mechanistically analogous to sotorasib and adagrasib, though without a covalent warhead. According to ASCO presentations on KRAS inhibitor resistance, GDP-state inhibitors are vulnerable to mutations that alter the switch-II pocket geometry (Y96C, H95Q/R/D, R68S), RTK-driven reactivation, and KRAS amplification.
Head-to-Head Differentiation
RMC-9805 offers several mechanistic advantages relative to MRTX1133. First, GTP-state targeting means RMC-9805 engages the dominant active conformation of oncogenic KRAS G12D rather than the minor GDP-bound fraction, which may translate to more complete pathway suppression in tumours with high RAS cycling rates. Second, the covalent aziridine warhead provides prolonged target engagement independent of drug concentration, potentially enabling less frequent dosing. Third, switch-II pocket mutations that confer resistance to GDP-state inhibitors are less relevant to a molecule that binds the switch-I/II interface in the GTP-bound conformation. Fourth, RMC-9805 demonstrates dual-state activity — active against both GDP- and GTP-bound KRAS G12D — providing mechanistic redundancy.
MRTX1133 (Mirati Therapeutics) is a non-covalent GDP-state KRAS G12D inhibitor with a pyrido[4,3-d]pyrimidine scaffold, a binding affinity KD of approximately 0.2 pM, and IC₅₀ below 2 nM in cellular assays, demonstrating 85% tumor regression in HPAC xenografts at 30 mg/kg BID. It differs from RMC-9805, which targets the GTP-bound (ON) state using a covalent aziridine warhead and cyclophilin A recruitment.
The patent landscape reflects these mechanistic differences. Revolution Medicines’ IP centres on the tri-complex mechanism and CypA recruitment strategy — a space not covered by competitor patents. Core patents include WO2022/105185A1 covering macrocyclic RAS inhibitors including RMC-6236 and RMC-7977, WO2023/060253A1 covering Zoldonrasib and related compounds, and WO2021/257736A1 covering combination strategies with mTOR inhibitors. Mirati’s IP focuses on pyrido[4,3-d]pyrimidine scaffolds for non-covalent GDP-state G12D targeting. Amgen has filed broad macrocyclic KRAS inhibitor patents (WO2025/213065A1, WO2024/104364A1) covering G12D, G12V, G13D, and Q61H. The WIPO patent database shows accelerating filing activity in the KRAS G12D space from 2021 onwards, reflecting the competitive intensity of this target class.
Map the full KRAS G12D patent landscape — including Revolution Medicines, Mirati, and Amgen filings — with PatSnap Eureka’s AI-powered patent analysis.
Analyse KRAS Patents in PatSnap Eureka →Clinical Development, Combinations, and Resistance Strategies
RMC-9805 is currently in Phase I clinical trials (NCT06040541) for advanced KRAS G12D-mutant solid tumours, with dose escalation and expansion cohorts evaluating monotherapy and combination regimens. The trial design incorporates lessons from the RMC-6236 Phase I/Ib programme, including biomarker-selected patient enrichment and ctDNA monitoring for early efficacy signals.
Combination Rationale and Validated Regimens
Monotherapy resistance to KRAS inhibitors follows predictable biological patterns: RTK reactivation (EGFR, HER family), PI3K/AKT/mTOR pathway bypass, secondary KRAS mutations, and cell cycle dysregulation. Four combination strategies have been validated preclinically for RMC-9805. First, RMC-9805 combined with an anti-EGFR antibody improved depth of response in colorectal cancer (CRC) models, addressing the dominant RTK reactivation mechanism in that tumour type. Second, RMC-9805 combined with RMC-6236 addresses polyclonal resistance by covering both G12D-specific and pan-RAS resistance mechanisms simultaneously. Third, RMC-9805 combined with anti-PD1 showed synergistic effects in immunogenic models, consistent with evidence that KRAS inhibition can enhance tumour immunogenicity. Fourth, RMC-9805 combined with SHP2 inhibitors produced regression in CRC models non-responsive to monotherapy, blocking the upstream RAS activating signal that drives adaptive resistance.
Resistance Mechanisms and Mitigation
Switch-II pocket mutations (Y96C, H95Q/R/D, R68S) — the primary resistance mechanism for GDP-state G12C inhibitors sotorasib and adagrasib — are less relevant for RMC-9805 because the tri-complex binds a different interface. KRAS amplification, which can overcome non-covalent inhibitors by overwhelming target occupancy, is partially mitigated by the covalent aziridine warhead’s stoichiometric engagement. Alternative RAS isoform activation (NRAS, HRAS) represents a more challenging resistance pathway, addressed by the pan-RAS inhibitor RMC-6236 in combination or sequential settings. The NIH/NCI has documented the complexity of adaptive resistance in RAS-driven cancers, underscoring the importance of combination strategies from the outset of clinical development.
Disease Context: KRAS G12D Prevalence and Unmet Need
KRAS G12D mutations account for approximately 40% of KRAS-mutant PDAC and 10–15% of KRAS-mutant NSCLC — two tumour types with significant unmet need. PDAC in particular has a five-year survival rate below 12% and limited effective therapeutic options, making RMC-9805’s 78% preclinical PDX response rate particularly significant. The broader RMC pipeline addresses additional KRAS mutations: RMC-5127 targets KRAS G12V (the second most common KRAS mutation), RMC-0708 targets KRAS Q61H, and RMC-7977 provides broad-spectrum activity against mutant and wild-type RAS across diverse RAS-addicted preclinical models. Patent analytics resources at PatSnap provide tools for tracking the competitive IP landscape as this field evolves rapidly.
“KRAS G12D mutations account for approximately 40% of KRAS-mutant PDAC — a tumour type where RMC-9805’s tri-complex mechanism may finally deliver durable responses in a patient population with few options.”