It had been widely expected that the first human use of CRISPR would be a 2017 clinical trial by Editas Medicine (EDIT), which announced last year that it plans to use CRISPR to try to treat a rare form of blindness called Leber congenital amaurosis. Only a few hundred people in the US have that disease. The possibility that a study siccing CRISPR on cancer will happen first suggests that the revolutionary genome-editing technology might be used against common diseases sooner than once thought.

The experiment, proposed by scientists at the University of Pennsylvania, still needs the approval of the medical centers where it would be conducted, as well as from the Food and Drug Administration, which oversees the use of experimental treatments in people. If the study gets those OKs, it would enroll patients with multiple myeloma, melanoma, and sarcoma, and be funded by the Parker Institute for Cancer Immunotherapy, which was launched this year by tech mogul Sean Parker.

“Our preliminary data suggests that we could improve the efficacy of these T cells if we use CRISPR,” Penn’s Dr. Carl June, a pioneer in the use of T cells against cancer, told the National Institute of Health’s Recombinant DNA Advisory Committee (RAC) on Tuesday.

Members of the committee were almost unanimously enthusiastic about the proposal. Dr. Michael Atkins, an oncologist at Georgetown University School of Medicine, called it “a really exciting first-in-human” study, adding that “we’ll learn a lot” from work that could “hopefully form the basis of new [cancer] therapies.” Biochemist Paula Cannon of the University of Southern California called it “innovative,” and said the Penn scientists had adequately addressed the questions she had about the safety of the procedure, including how they would tell whether CRISPR accidentally cuts the wrong genes, a problem called off-target effects.

The proposed early-stage clinical trial with 15 patients would gauge the safety of the experimental therapy and see how feasible it is to manufacture genetically engineered and CRISPR’d T cells, June said. The scientists would remove T cells, which normally target cells that are “foreign,” like bacteria, from patients with multiple myeloma, melanoma, or sarcoma. They would then use CRISPR to genetically modify the T cells so that, infused back into a patient, they can target and destroy tumor cells.

The trial would be conducted at MD Anderson Cancer Center (enrolling nine patients) in Texas, and the University of California, San Francisco (three), as well as Penn (three). Penn would also produce the genetically modified T cells.

In a technique that several companies are competing to commercialize, traditional genetic engineering alters T cells extracted from patients so that the cells produce a “chimeric antigen receptor,” or CAR.

Once the T cells are infused back into patients, that CAR lets the cells find molecules called antigens, which protrude from tumor cells, like a key fitting a lock. If all goes well, the T cells would destroy the tumors. In particular, the T cells would glom onto the antigen NY-ESO-1. Last year, June and his colleagues reportedthat T cells targeting that molecule safely beat back multiple myeloma in 16 out of 20 patients, each of whom received some 2.4 billion genetically-tweaked T cells, June told the committee.

Unfortunately, although that and other studies have found promising results with genetically-engineered T’s, the cells have been shown to work only on some cancers (mostly leukemias and other blood cancers), with disappointing results in solid tumors. Many patients who respond eventually see their cancer return, possibly because tumors begin repelling the T cells.

The Penn trial would use CRISPR to slice out two genes in T cells to hopefully prevent recurrence and help the treatment last longer. One gene is for PD-1, a “checkpoint” molecule that also sits on the surface of many cancer cells. When it binds to a T cell, PD-1 disables that T cell. Solution: edit out the receptor gene so T cells can’t bind to PD-1 and will keep fighting the cancer.

CRISPR’s other target would be the gene for a T cell’s natural receptors, called endogenous TCR. Studies have shown that “if you remove the TCR you get better functioning” of engineered T cells, June said. In his team’s mouse experiments, T cells CRISPR’d to lack both the PD-1 gene and the natural receptor gene reduced the size of lung tumors much more than non-CRISPR’d T cells did. Since CRISPR is not perfect, some PD-1 and TCR genes remain, but at low enough levels to make the T cells attack cancer cells more effectively, according to lab data presented to the committee.

CRISPR’ing out the two genes — in a process expected to take 35 days — “may also increase the persistence” of T cells, Penn’s Dr. Edward Stadtmauer told the committee. T cell “exhaustion” might explain why the benefits of the therapy often fade.