Using genetically-edited cells to supercharge the immune system caused no adverse effects in cancer patients. It’s too soon to tell if it can be a cure.

It’s been over three years since US regulators greenlit the nation’s first in-human test of Crispr’s disease-fighting potential, more than three years of waiting to find out if the much-hyped gene-editing technique could be safely used to beat back tough-to-treat cancers. Today, researchers from the University of Pennsylvania and Stanford finally revealed the first published report describing the trial. The highly anticipated results showed that the procedure is both safe and feasible; the Crispr’d cells went where they were supposed to go and survived for longer than expected. They didn’t cure anyone’s cancer, but they didn’t kill anyone, either, which means the results hold significant promise for the future of Crispr-based medicines.

The trial was small—just three people—and designed only to assess the technique’s safety. Last year, each cancer patient received infusions of about 100 million of their own T cells, which had been genetically modified in a University of Pennsylvania lab. There, researchers equipped the cells with souped up cancer-recognizing receptors and used Crispr to make them more efficient killing machines. These cells successfully joined up with the rest of each person’s immune system and could still be found circulating in patients’ blood nine months later. The researchers presented some of that preliminary data at a conference in December, but didn’t include any information about how well the Crispr’d cells actually performed. That information is among the new details included in the peer-reviewed study published Thursday in Science.

“Before we did this, no one had ever infused Crispr-edited cells into patients, and we’re encouraged by the fact that we could do it safely,” says Edward A. Stadtmauer, an oncologist at the University of Pennsylvania and the study’s principal investigator. “Now we can move on to a whole new frontier of further engineering these cells and expanding the number of patients treated.”

The study was overseen by Carl June, a pioneer of the emerging field of immunotherapy, which involves supercharging patients’ own immune systems to fight cancer through a series of genetic tweaks and pharmaceutical nudges. June’s biggest breakthrough came in 2012, when his UPenn lab inserted a new gene into the T cells of a terminally ill child named Emily Whitehead; imbued with new cancer-recognizing abilities, those cells wiped her leukemia off the map. In June, the now-14-year-old ran her first 5K to raise money for curing children’s cancer.

Whitehead’s miraculous recovery wasn’t exactly a fluke. But she was lucky. The T cells she received triggered a “cytokine storm” that flooded her body with organ-damaging inflammation. June’s team saved her life by administering a newly-approved drug. But other patients haven’t been so fortunate. Reengineered T cells can also go wrong in other ways—natural receptors will sometimes interfere with the designer ones, making them less effective. The goal of the UPenn trial was to see if Crispr could solve some of those issues—without creating a dangerous immune system reaction. Previous research has shown humans to have existing immunity to the bacteria from which Crispr (the original version, which the UPenn team used) is derived.

Joseph Fraietta, who runs his own immunotherapy lab at UPenn’s Center for Advanced Cellular Therapeutics, designed the Crispr systems they used and supervised the editing. After harvesting T cells from three patients, his group made three edits to them. The first was to a gene called PDCD1. It makes a protein that acts like a brake on the immune system. Tumors have ways of turning up the expression of this protein in immune cells to dampen their response to the invading cancer. By using Crispr to turn off PDCD1, the scientists hoped to increase the likelihood that the patient’s new clone army of T cells would all show up to the fight.

In the second two edits, the scientists used Crispr to cripple genes that code for natural T cell receptors—deleting them from the cell’s surface and creating a blank slate. Then, after a few days’ rest, the researchers inserted a new gene into the cells, this one containing the code for their designer receptor. That step armed each cell with a kind of cancer homing device. Scientists next moved the cells into a collection of large bags, each holding several liters of liquid sugars, salts, and other things cells need to grow. For weeks, the bags rocked gently inside incubators, until the cells had multiplied into the many millions, before being cryopreserved and shipped off for infusion into each patient.

The biggest question going into the trial was what would happen when those 100 million cells were plugged into patients’ bodies. Would they settle in? Would they find their way to the cancer? Would they even survive? Or worse, would residual Crispr proteins trigger massive immune reactions?

There wasn’t much international research they could rely on for precedent. Scientists in China were the first to use Crispr to try to treat cancer in humans in 2016. They have since initiated a number of clinical trials, but released very little data about them.

In case the stakes weren’t plain enough, it might help to recall that the University of Pennsylvania is the same place where an 18-year-old named Jessie Gelsinger died from a catastrophic immune reaction to an experimental gene therapy in 1999, setting back the whole field for decades. A similar disaster could sink the efforts of the dozens of companies chasing the engineered T cell idea, and the research they support. June holds a number of patents on T cell technology and is a cofounder of Tmunity, an engineered T cell company that provided funding for the trial. Many of his coauthors have received funding or consulting fees from other cell therapy companies with T cell products in the pipeline including Novartis, Gilead, and Arsenal Biosciences. Proving to the public that these cells are safe for people is more than just an academic exercise. Billions of dollars are on the line.

This time around things went much better. The patients’ health either improved or held steady. They tolerated the engineered T cells with only mild adverse effects and no immune response. And when Fraietta’s team sampled their blood every few months, the researchers kept finding cells with the edits they had made. That’s a good sign, because it means the cells weren’t dying, and appeared to be just as fit as the patient’s natural cells. Moreover, when the researchers biopsied bone marrow from the patients, they found the edited T cells there too, at the sites of the cancer, indicating the new cells had migrated to the right spots.

But though the three patients experienced some stabilization of their disease during treatment, and one saw tumor size reduction, the T cells were far from a total fix. One of the patients, a woman with multiple myeloma, died in December, seven months after receiving the treatment. The other two—another woman with multiple myeloma and a man with sarcoma (the one who’s tumor shrank)—have since had their cancer worsen and are now receiving other treatments.

“It’s really hard for us to make any conclusion about the effectiveness of the therapy except to say it’s not 100 percent effective,” says Stadtmauer. “You really need to treat many more patients to get at that question.”\

Originally, the UPenn team’s plan was to move this Crispr technique into a larger trial involving 18 participants, which could start to answer that question. But so far, they have not treated any additional patients. The reason, says Stadtmauer, is that the gene editing field is moving so quickly they’re not sure they want to push forward with what is now considered to be outdated tech. Today, a Crispr system developed in 2015 looks positively prehistoric. In the years since the trial was approved, a suite of new gene-editing tools that promise greater accuracy and more design flexibility have since been developed. “I see this study as the first stepping stone that leads to many more investigations of this approach,” says Stadtmauer.

In fact, he says, a number of such cancer trials at UPenn are slated to begin later this year. “We’re right on the verge,” he says. “This isn’t many years away. There are many more patients who will be receiving edited cells in the year 2020.”

The results will ripple beyond the University of Pennsylvania. A few other US Crispr trials are just getting underway. Last year, doctors began testing Crispr for the blood disorders sickle cell disease and beta thalassemia. Another trial using Crispr to treat an inherited form of blindness is currently recruiting participants.

“Let’s just say this finding will be cited by every academic lab or biotechnology company filing with an investigational new drug application with the FDA for Crispr-edited cells,” says Fyodor Urnov, scientific director of technology and translation at the Innovative Genomics Institute, a joint research center of UC Berkeley and UC San Francisco. He says the young field of gene-editing has been haunted by unknowns, in particular, the potential impact of Crispr’s mistakes. The DNA-slicing tool isn’t perfect. Fraietta’s UPenn team found evidence of mutations in about one percent of cells they infused into their three patients.

And many papers have come out hypothesizing about potential risks; unexpected mutations might disrupt key cell functions or even cause cancer. (One published in 2017 briefly tanked shares of Crispr-based medical companies.) But Urnov says this shows convincingly that such fears are overblown. “What this shows is that you can transplant edited cells that have all sorts of unwanted things happen to their genome and the cells appear to be fine and they don’t have any untoward effects on patients,” he says.

Fraietta is a little more cautious. “We don’t know yet what the significance is of having introduced genomic instability,” he says. “It’s kind of a wait-and-see.” The remaining two patients will be regularly monitored for the next 15 years to assess any such long-term risks. It may be a long time before the field of gene-editing has a definitive answer. But it still has many more answers today than it had yesterday, and all of them point toward a disease-fighting future transformed by Crispr.