The race to treat KRAS-driven cancers reached a new milestone this month. Roche announced that its next-generation inhibitor divarasib cut the risk of death by 44% compared with earlier approved drugs in a late-stage study, offering patients a longer median survival of approximately 18 months. The results, presented at the 2026 ESMO Congress and published in The New England Journal of Medicine, mark the latest chapter in a decades-long effort to target a protein once written off as undruggable.
KRAS encodes K-Ras, a GTPase that acts as a molecular on-off switch for cell growth signals. For decades, the protein was considered undruggable because its surface is remarkably smooth and it binds GTP with picomolar affinity, leaving no deep pockets for traditional small molecules to latch onto. The gene was first identified as a viral oncogene in the 1980s, and researchers spent years hunting for ways to block its signal without success.
The breakthrough came with the discovery of the G12C mutation. This specific change substitutes cysteine for glycine at position 12, creating a unique nucleophilic cysteine near the nucleotide-binding site. This cysteine allowed scientists to design covalent inhibitors that chemically lock K-Ras in its inactive GDP-bound state, effectively turning the switch off. Sotorasib and adagrasib were the first approved drugs targeting this pocket, but they faced limitations in potency and durability.
Roche’s divarasib is a next-generation inhibitor built on the same covalent foundation but engineered for greater selectivity and potency. In the phase III Krascendo 1 trial, divarasib demonstrated superior efficacy against first-generation therapies. Patients receiving divarasib lived a median of approximately 18 months, compared to 12 months for those on sotorasib or adagrasib. The hazard ratio for death was 0.56, with a p-value of 0.0005, indicating a statistically significant improvement. The G12C mutation occurs in roughly one in eight non-small cell lung cancer patients, and those patients have historically faced poorer outcomes than those with wild-type disease.
Safety data from the trial showed no new signals. The most common treatment-related adverse events were mainly low grade and manageable, consistent with the known toxicity profile of the class. Roche plans to submit the data for regulatory review, aiming to expand the treatment options available for this molecular subset of lung cancer.
Researchers are now turning their attention to the next hurdles. Acquired resistance remains a challenge, and work is underway on pan-KRAS inhibitors that could target mutations beyond G12C, as well as combination strategies to outsmart the disease’s ability to adapt. For now, divarasib represents a meaningful advance: a newer tool in a growing kit, offering more time for the patients who need it most.
Talk to your doctor or oncologist about whether new targeted options are appropriate for your specific situation.
The Structural Challenge of Targeting the KRAS Pocket
For decades, the KRAS protein evaded therapeutic intervention due to its unique biophysical properties. Structural studies have consistently described the protein surface as remarkably smooth, lacking the deep hydrophobic pockets that most drug designs rely on to achieve high-affinity binding. This flat landscape, combined with the protein’s intrinsic ability to bind guanosine triphosphate (GTP) with picomolar affinity, presented a formidable barrier; any potential inhibitor would have to compete with a ligand that effectively “locks” the protein in its active conformation. As detailed in foundational literature on the K-Ras protein, it functions as a molecular switch, and its natural high affinity for GTP made the development of competitive small-molecule inhibitors seem chemically implausible.
The conceptual breakthrough that made targeted therapy possible came from understanding the protein’s mutational landscape. Research into the specific G12C mutation revealed that while the overall surface remains smooth, this particular amino acid substitution introduces a unique nucleophilic cysteine residue near the nucleotide-binding site. Scientists leveraged this specific chemical feature to design covalent inhibitors. By targeting this cysteine, drugs like sotorasib and adagrasib could chemically lock the protein in its inactive GDP-bound state. This approach sidestepped the need to bind deep pockets on the smooth surface, instead utilizing a transient, mutation-specific vulnerability to achieve the necessary selectivity and potency for clinical efficacy.
Expanding Beyond G12C: The Quest for Pan-KRAS Inhibitors
The clinical success of divarasib against the G12C mutation has highlighted a critical limitation in current oncology: the vast majority of KRAS mutations remain untreatable. While the G12C substitution introduces a unique nucleophilic cysteine near the nucleotide-binding site, other common mutations lack this specific chemical feature. Consequently, drugs like sotorasib and adagrasib, which rely on covalent inhibition of that specific cysteine, cannot target those variants. Researchers are now turning their attention to developing pan-KRAS inhibitors capable of targeting mutations beyond G12C.
This pursuit is essential because the K-Ras protein functions as a molecular switch within the RAS/MAPK pathway, relaying signals that instruct cells to grow and divide. When mutated in cancers such as non-small cell lung cancer or pancreatic cancer, this switch remains stuck in the “on” position. The G12C mutation is found in approximately one in eight people with non-small cell lung cancer, but other mutations affect a significant portion of patients who currently have no targeted options. Unlike individualized vaccines tailored to specific patient profiles, which may take time to manufacture, small molecule inhibitors offer a potential for immediate, mass-produced solutions.
The structural studies describing the protein surface as remarkably smooth have guided these efforts. Since most drug designs rely on deep hydrophobic pockets that KRAS lacks, scientists must find alternative mechanisms to achieve high-affinity binding without competing with GTP or GDP. The intrinsic enzymatic activity of K-Ras, which converts guanosine triphosphate into guanosine diphosphate, presents both a challenge and an opportunity. By understanding these protein dynamics, including the role of accessory proteins like GAPs and GEFs, researchers aim to outsmart the disease’s ability to adapt and restore effective therapies for patients with wild-type or non-G12C mutated disease.
Sources
- Roche’s divarasib demonstrated best-in-class efficacy, advancing the standard of care in previously treated KRAS G12C-mutated lung cancer
- KRAS
- Vaccine Targeting KRAS in Pancreatic Cancer Shows Promise
Photo: Taxiarchos228 (FAL)
