Rockefeller University published this work in Nature (doi:10.1038/s41586-026-10367-0).
Cancer cells are supposed to die when you hit them with the right drug. Melanoma cells treated with vemurafenib, an approved melanoma drug, do die at first. Then they adapt. They find a workaround. The treatment stops working. This is called acquired drug resistance, and oncologists have been losing this arms race for decades.
The standard explanation was complexity: too many genes can mutate, too many escape routes exist, so resistance is essentially inevitable. A team at Rockefeller University just published evidence that this explanation might be wrong. The escape routes look different at the surface, but underneath, they all funnel into the same molecular bottleneck.
"If you disrupt different parts of this same complex, you trigger drug resistance through routes that ultimately converge on the same survival signal," said Xin Cao, the Rockefeller geneticist who led the work, speaking to GEN News. "This suggests that you don't have to target each resistance mechanism individually. Instead, you could target this shared node."
The findings come from a platform called PerturbFate, which the team developed to map what happens inside cells when you knock out drug resistance genes one by one. They profiled 318,032 cultured melanoma cells treated with vemurafenib, targeting 143 resistance-associated genes, as the paper describes. The data showed something striking: wildly different genetic perturbations all pushed cells into the same drug-resistant state. The individual mutations were many. The destination was one.
CRISPR was central to the method. The team used CRISPR interference to systematically disable each candidate gene and watch what happened. This kind of large-scale, single-cell CRISPR experiment is expensive and technically demanding, which is probably why nobody had done it at this scale before for resistance mapping.
Once they found the convergence point, the researchers tested two interventions: combined inhibition of the cooperative transcription factor hubs, and blocking the downstream survival program directly. Both worked. Resistant cells could no longer proliferate after either intervention.
What makes this immediately relevant to drug developers is that the downstream node the paper identifies is not obscure. One existing approach targets a VEGFC-driven survival signal, and a VEGFR3 inhibitor called EVT801 entered clinical trials for lymphatic metastasis in 2021 (ClinicalTrials.gov NCT05114668), as a 2022 paper in Cancer Gene Therapy noted. The drug already exists. PerturbFate's map suggests it may also work against the resistance route that makes melanoma drugs fail.
The academic literature had anticipated convergence-based frameworks for drug resistance as early as 2018, as a review in Advances in Cancer Research noted, identifying three categories: pathway reactivation, pathway bypass, and pathway indifference. What nobody had done was the systematic map of where those paths actually converge inside a living cell.
The Rockefeller team has made both the experimental and computational tools behind PerturbFate openly available, meaning any lab with CRISPR access can replicate or extend the approach. Three team members are listed as inventors on a related patent application (US provisional 63/789,011). The tools are free; what they reveal is not.
The team is now applying the same approach to aging and Alzheimer's disease in living systems, not just cultured cells. Whether the same convergence logic holds in tissues, where drug resistance involves the immune system and the tumor microenvironment, is an open question.
This work was funded by the NIH (grants DP2HG012522 and RM1HG011014) and the Mathers Foundation.
The immediate implications are for drug discovery. Any oncology program where acquired resistance is a known problem, and that is most of them, now has a systematic tool to ask whether their resistant cases route through a shared node. If they do, a combination therapy targeting that node from the start could sidestep the resistance problem rather than chasing it after it emerges. EVT801 is not the only VEGFC-targeting approach in development, but PerturbFate gives the field a map to look for others.
Whether that map holds in patients, rather than cells in a dish, is what the next five years of clinical research will determine.
