Command Palette

Search for a command to run...

Preprint WatchModerateOctober 23rd, 2026

DIRAS3 broadly suppresses signaling from diverse oncogenic KRAS hotspot variants through an inhibitory interface

Bildik, G.; Yang, H.; Mao, W.; Gannaban, R. B.; Bhardwaj, V. K.; Uddin, M. N.; Jung, S. Y.; Pierobon, M.; Gomeiz, A.; Gorfe, A. A.; Hancock, J. F.; Marshall, C. B.; McCormick, F.; Millward, S. W.; Liu, J.; Bast, R. C.; Lu, Z.

DIRAS3 suppresses signalling from diverse oncogenic KRAS hotspot variants through an inhibitory interface, in an activation-state-selective manner that largely spares wild-type KRAS.

Moderate contradiction

2 prior failures

Two documented clinical failures match this mechanism, or a single Phase 3 failure is on record.

This preprint describes an endogenous inhibitory interface through which DIRAS3 suppresses signalling across major oncogenic KRAS hotspot classes while largely sparing wild-type KRAS. Claidex holds two KRAS programme failures, bgb-53038-kras-solid-tumors-phase1-strategic-reprioritization and sotorasib-durvalumab-kras-g12c-ctdna-stage3-nsclc-phase2-enrollment-collapse, neither of which turned on a failure of target engagement. A pan-hotspot, activation-state-selective mechanism is a different proposition from the variant-specific covalent inhibitors that define the current clinical record, so the documented failures constrain it only weakly. The flag is MODERATE on failure count rather than on direct contradiction, and the preprint is preclinical.

Abstract excerpt

Despite major advances in KRAS-directed therapies, broadly suppressing oncogenic KRAS signaling across diverse hotspot variants remains an unmet challenge. Here, we identify an endogenous inhibitory mechanism that suppresses signaling across the major oncogenic KRAS hotspot classes while largely sparing basal wild-type KRAS, revealing activation-state-selective inhibition. We demonstrate that the imprinted tumor suppressor DIRAS3 directly engages activated KRAS through a functionally defined inhibitory interface that can be molecularly mimicked to broadly inhibit KRAS-driven signaling. Using genetically controlled RASless reconstitution models, complementary cancer cell validation, structure-function analyses, molecular dynamics-based membrane topology modeling, and ultrastructural imaging, we show that dual lipid modification constrains DIRAS3 membrane topology, enabling an 5-encoded inhibitory interface to engage activated KRAS, disrupt signaling-competent membrane assemblies, and suppress downstream signaling. Structure-function analyses establish this interface as a critical determinant of DIRAS3-mediated KRAS inhibition. A conformationally constrained 5 mimetic recapitulates full-length DIRAS3 activity, broadly suppressing MAPK and PI3K signaling across multiple oncogenic KRAS variants, including Q61 mutants with reduced sensitivity to current tri-complex RAS(ON) inhibitors. Mechanistically, immunoprecipitation shows that the 5 mimetic reduces KRAS association with both RAF1 and p110 similarly across G12 and Q61 mutants, whereas RMC-7977 reduces inhibition of p110 association with Q61 KRAS. The 5 mimetic also disrupts KRAS nanocluster organization and inhibits tumor growth in vivo. These findings identify a targetable inhibitory interface within DIRAS3 and establish mimicry of an endogenous KRAS suppressor as a mechanistically distinct strategy for broad KRAS inhibition.

Matching Claidex post-mortems

2 of 2 indexed

This is an automated contradiction flag, not an editorial judgment on the preprint's quality. Flags identify where the preclinical literature and the clinical failure record diverge.