Drug Targets Don't Pass Muster with CRISPR

The presumed targets of many anticancer drugs may be wrong, concludes a recent study that used the latest CRISPR gene-editing technologies to probe the mechanisms of action of 10 small-molecule drugs in or near clinical testing (Sci Transl Med 2019;11:eaaw8412). With all 10, the putative targets proved nonessential both to cell growth and to drug activity—a finding that calls into question some of the common techniques used for target validation in oncology.

“There are some fundamental issues in how new drug targets are studied and how new drugs are characterized,” says Jason Sheltzer, PhD, of Cold Spring Harbor Laboratory in New York, who led the research.

RNAi assays, Sheltzer points out, are prone to off-target effects, and this older approach to evaluating gene dependencies may have frequently led drug developers astray. With more precise CRISPR-based methods, he says, “you do a better job of finding cancer-essential genes, you do a better job validating a drug's on-target mechanism of action, and that kind of preclinical validation will help clinicians design better clinical trials to decrease the failure rate of new drugs.”

“The depth and scope of this analysis highlights the potentially widespread problem of off-target pharmacology for small-molecule inhibitors,” says Jeff Settleman, PhD, head of oncology research at Pfizer in San Diego, CA. He notes, however, that the study focused only on investigational agents, most of which either failed early testing or had yet to reach the clinic. “For the vast majority of approved oncology drugs,” he emphasizes, “there are abundant data supporting the on-target nature of the observed clinical benefit.”

Sheltzer and his colleagues had previously shown that MELK, a protein once thought critical for the growth of multiple cancer types, was not needed for cell proliferation, and that a putative MELK inhibitor worked through some other pathway (eLife 2017;6:e24179). To test whether similar problems were widespread, his team considered 10 more drugs that collectively target six different proteins.

The researchers selected five drug targets reported to underpin cancer “addictions”—proteins without which the cells cannot survive—as well as one target reported to induce cell death when activated. None of the targets had a known resistance mutation that would definitively demonstrate the drug had on-target activity.

Using gene-editing tools to mutate or silence each protein in a variety of cell lines, Sheltzer and his colleagues showed that the cells' proliferative capacity was unscathed by CRISPR-induced perturbations. What's more, the 10 drug inhibitors all continued to block cancer growth in the gene-edited cells, indicating that the molecules “must necessarily be killing the cancer cells through some other unknown, off-target effect,” Sheltzer says.

Follow-up experiments with CRISPR involving one of the drugs helped pinpoint its true function. The preclinical-stage agent OTS964 didn't work through TOPK inhibition, as its manufacturer OncoTherapy Science had reported (Sci Transl Med 2014;6:259ra145). Instead, it proved to be a CDK11 blocker. Several approved breast cancer drugs target other members of the CDK protein family, but OTS964 is one of the first directed against CDK11, a protein vital for cell-cycle control and RNA transcription regulation.

The paper “definitely speaks to the irreproducibility of science,” says Sourav Bandyopadhyay, PhD, of the University of California, San Francisco, who was not involved in the study. However, the problem is much bigger than any one assay: “You need to do rigorous science, rather than just throw out RNAi and pop in CRISPR,” he says. “CRISPR is not quite a panacea in rescuing every drug target.” –Elie Dolgin

For more news on cancer research, visit Cancer Discovery online at http://cancerdiscovery.aacrjournals.org/CDNews.