The gene-editing tool known as CRISPR may eventually enable oncology investigators to create methods for killing tumors directly, but thus far, the focus is on enhancing forms of immunotherapy already in clinical practice.
Marcela V. Maus, MD, PhD
The gene-editing tool known as CRISPR may eventually enable oncology investigators to create methods for killing tumors directly—or even for preventing people from ever developing cancer—but thus far, the focus is on enhancing forms of immunotherapy already in clinical practice.
Investigators around the world are using CRISPR in hopes of improving existing immunotherapies and creating new ones, and this revolutionary technology has produced exciting results in vitro and in animal trials. Human trials are underway in China and could begin at several cancer centers in the United States this year.
“We have been using genetic manipulation to redirect T cells for about 15 years now, but it has always been a slow and expensive process,” said Marcela V. Maus, MD, PhD, director of cellular immunotherapy at Massachusetts General Hospital Cancer Center in Boston. “CRISPR is so cheap, so easy, and so fast that it has democratized the field. Many researchers are using it to test multiple theories, and the findings are coming very quickly. The biggest obstacle to a major surge in clinical trials of redirected T cells— and it’s an obstacle researchers and regulators are diligently working through—is the natural caution about using new technology in humans.”
Clustered regularly interspaced short palindromic repeats (CRISPRs) were discovered in bacterial DNA 31 years ago, but no one understood their function until 2007, when researchers at a yogurt company figured out that they help bacteria defend themselves against viruses. The first time a bacterium fights off a particular virus, it cuts up the virus’ RNA and stores little chunks of it in its CRISPRs. If the virus returns, the bacterium produces special attack enzymes such as CRISPR-associated protein 9 (Cas9) and attaches a CRISPR containing the viral RNA. If this “guide RNA” in the CRISPR matches any in the invading virus, the enzyme starts cutting the virus’ DNA to ribbons (FIGURE).
This interesting discovery became a practical tool in 2012, when a team led by Jennifer A. Doudna, PhD, and Emmanuelle Charpentier, PhD, reported that investigators could make Cas9 chop up the DNA of any molecule they wanted, in a precise location, by deliberately inserting RNA into a CRISPR-carrying enzyme.1 The potential to dramatically change medical research bercame evident just 1 year later, when teams led by Feng Zhang, PhD, and George M. Church, PhD, discovered how to use CRISPR to exchange any DNA researchers wanted for any that they needed to chop out.2,3
Existing genetic manipulation techniques could be used to accomplish most of the same tasks as CRISPR, which is often referred to as CRISPR-Cas9 to acknowledge both of its working halves. However, as Maus said, CRISPR was so much faster and easier and it has generated a flood of research. Last year, just 4 years after publication of the papers that described findings involving CRISPR as a practical gene-editing tool, approximately 20,000 articles indexed by Google Scholar reported use of the technology or at least mentioned it. Agricultural researchers have already used CRISPR to genetically improve various qualities in experimental versions of tomatoes, wheat, tobacco, and many other crops.4 Scientists in China have reported using it to produce more muscular dogs.5 A team led by investiagors at Oregon Health & Science University in Portland described using CRISPR-Cas9—based techniques to correct a germline mutation in human preimplantation embryos that affects a DNA repair response.6
Such studies have prompted speculation that editing cancer-related genes such as BRCA1 and BRCA2 could prevent the development of malignancies.7 For now, investigators have focused instead on using CRISPR as a practical tool for improving immunotherapy. In January 2018, a potential hurdle surfaced when researchers from Stanford University in California reported that the 2 bacteria strains from which Cas-9 proteins are most frequently derived exhibited a preexisting adaptive immune response in experiments on human serum.8 The authors said their findings raise important considerations that should be investigated as the research moves forward.Even without CRISPR, efforts to target carcinogenic activity by adding chimeric antigen receptors (CARs) to T cells extracted from patients have produced exciting clinical trial results. In August 2017, tisagenlecleucel (Kymriah) became the first CAR T-cell therapy to be approved by the FDA after a trial in 63 children and young adults with B-cell acute lymphoblastic leukemia (ALL) produced an overall remission rate of 83%.9 Axicabtagene ciloleucel (Yescarta) gained FDA approval 2 months later after it demonstrated a complete remission rate of 51% and a partial remission rate of 21% in 101 patients with large B-cell lymphomas treated during clinical trials.10
Those medications, along with other experimental CAR T-cell therapies that have produced dramatic results in smaller trials, use retroviruses to insert the desired CAR into the DNA of patients’ T cells, which are then grown in the lab and reinfused. The retroviruses are effective at transferring the CAR to the T cell, but they have several drawbacks, according to Michel Sadelain, MD, PhD, director of the Center for Cell Engineering at Memorial Sloan Kettering Cancer Center (MSK) in New York, New York.
The fragility of retroviruses necessitates exacting and costly storage procedures, and the potential that an improperly neutered retrovirus could infect a patient requires biosafety checks before each usage. The biggest challenge, however, is a lack of precision. CAR T-cell therapy developers know the retroviruses will insert the CAR in the T cell’s genes, but they have no control over where the CAR goes. Each time they draw a new patient’s blood and introduce the retrovirus, the CAR ends up in a random spot. Labs can’t even control the chromosome.
The varying locations in the T-cell DNA lead to varying performance. Expression is too high in some positions (which leads to T-cell burnout) and too low in others. Different batches of what is theoretically the same named treatment emerge with a wide variance in potency, so some patients get too much and suffer needless toxicity, whereas others get too little for optimal therapeutic effect.
Sadelain and colleagues have shown that CRISPR may eliminate those problems and produce CAR T-cell therapies that are more effective and less toxic. Rather than using a retroviral vector to insert a CAR gene at some random site in the T-cell genome, the team has used CRISPR in ALL mouse models to eliminate the natural genetic code that told the T cell what kind of antigen to hunt (the T-cell receptor) and replaced it—at exactly the same spot in the T cell’s DNA—with the CAR targeting the antigen expressed in the tumor. Although the switch did not always work perfectly, findings showed that 95% of the engineered cells had a properly located CAR and no natural T-cell receptor. Nearly all the rest had both, likely because T-cell DNA began with 2 T-cell receptors coded in the same spot.11
It was not obvious that CARs placed in 1 of the 2 T-cell receptor alleles (where the normal coding for T-cell targets is situated) would produce the best results, so Sadelain’s team did not assume that would occur. They compared the efficacy of placing the CAR at 11 different positons and found that the T-cell receptor alpha constant (TRAC) was the best location for the CAR. T cells with TRAC-CAR placement attacked tumors more consistently than those with randomly placed CARs or those with CARs placed elsewhere.11
Investigators determined the functional limits of different T-cell populations in a pre—B-cell ALL mouse model by using a CAR “stress test” that gradually lowers the dosage. They found that TRAC-CARs “differed markedly” in antitumor activity when compared with CARs that had been retrovirally transduced or encoded. “TRAC-CAR T cells induced greater responses and prolonged median survival at every T-cell dose,” the investigators wrote in a paper published in Nature that was cited 66 times in the first 10 months after its publication.11 Notably, at the highest dose, every TRAC-CAR T-cell mouse was alive 50 days after injection, whereas every mouse that received another type of CAR T cell was dead. At the lowest dose, half of the TRAC-CAR T-cell mice were alive at 30 days, and all the mice that received a different type of CAR T cell were dead at 20 days.
At the same time, the TRAC-CAR T cells did not attack too vigorously, as traditional CAR T cells often do; just 2% of TRAC-CAR T cells experienced the sort of burn out that affected more than half of the conventional CAR T cells. This tendency to keep working almost certainly had more to do with the prolonged survival than any greater initial activity associated with the exact CAR positioning, the authors concluded.
The published results led some to predict that CRISPR might produce CAR T-cell therapies that are not only safer and better but also less expensive. (The cost savings would come less from the lower cost of using CRISPR than from eliminating the need for autologous transplants and allowing the mass production of CAR T-cell therapy.12) Sadelain naturally hopes that such predictions prove true, but he also cautions against excessive expectations.
“We don’t even know yet whether it’s safe to use therapies that have been manipulated by CRISPR. There’s a tendency when describing it to suggest that it only cuts at a single site, but that’s not always true. The technology keeps improving, but it’s still far from perfect, and we’ll need to look very closely for unintended genetic modifications when we move to human trials,” Sadelain said in an interview with OncologyLive®.
“Even assuming the technology works perfectly, it alone probably won’t lead to an explosion in new CAR T-cell therapies,” he said. “The only way you get a new treatment is to find a new CAR that makes T cells attack a particular tumor type. CRISPR doesn’t help you discover CARs. It’s just a better tool—we hope—for placing those CARs where they’ll be most effective.”Nevertheless, Sadelain said that if CRISPR proves safe and effective as a tool for improving conventional immunotherapies, investigators will begin testing its ability to attack tumors more directly. For now, the focus is on immunotherapy, and much of the activity is taking place in China.
Investigators at Chinese PLA General Hospital in Beijing began recruiting in July 2017 for a trial of CRISPR-manipulated T cells in 80 patients with B-cell leukemia and B-cell lymphoma. The study team will take T cells from healthy donors and use CRISPR both to insert the CAR and to remove genes that would cause graft-versus-host disease (GVHD). The T cells are expanded in a lab and reinfused into patients (NCT03166878).
Another trial at Hangzhou Cancer Hospital is recruiting 21 patients with esophageal cancer for a variant on existing targeting checkpoint inhibitor therapy. The emerging class of PD-1/PD-L1 pathway inhibitors consists of monoclonal antibodies that block immunosuppressive activity. The Chinese investigators are manipulating T cells such that they do not express any PD-1 at all; the cells are expanded in the lab and reinfused into patients. The expected date for primary study completion is February 2018, and the expected date for full completion is December 2018 (NCT03081715).
Investigators at Sichuan University’s West China Hospital in Chengdu, meanwhile, are recruiting 21 patients with metastatic non— small cell lung cancer for another trial of PD-1 knockout T cells (NCT02793856). Another study team, this one at the Comprehensive Cancer Center of Nanjing Drum Tower Hospital, is recruiting 20 patients with a variety of tumor types associated with Epstein-Barr virus to test the safety and efficacy of its PD-1 knockout T-cell formulation (NCT03044743).
A team at the University of Pennsylvania (Penn) had planned to begin human trials of CRISPR-tweaked CAR T cells last year, but the launch was delayed. In January 2018, Penn investigators registered plans for a clinical trial that would recruit 18 patients with melanoma, multiple myeloma, synovial sarcoma, or myxoid/round cell liposarcoma (NCT03399448). Investigators would engineer autologous T cells to express NY-ESO-1, a tumor antigen, and then use CRISPR gene editing to eliminate endogenous T-cell receptors and PD-1 cells. Even if something derails the Penn project, MSK and several other American research institutions hope to begin trials of CRISPR-manipulated treatments this year.
China’s lead stems at least in part from less-intensive regulatory oversight. Chinese investigators need nothing more to begin using CRISPR on humans than approval from the ethics committee at their hospital. A recent Wall Street Journal story, which reported that CRISPR has already been used to treat dozens of Chinese patients, said that one group’s project had been reviewed and approved in a single afternoon.13 The Penn team, on the other hand, worked for more than 2 years to secure approvals for its trial.
On the corporate side, one leader in this arena is Cellectis, a French-American company that uses an alternative gene-editing technology called TALEN (transcription activator-like effector nucleases) to turn T cells from healthy donors into allogeneic and thus universal CAR T-cell therapies (UCARTs) that can be produced in batches and used in many patients with cancer.
The development process has proved challenging. In September 2017, the FDA halted 2 phase I trials into UCART123, created with the TALEN system; 1 patient with blastic plasmacytoid dendritic cell neoplasm died after UCART123 treatment and another participant with acute myeloid leukemia required hospitalization for serious adverse events (AEs). The FDA lifted the holds 2 months later, after Cellectis agreed to implement changes to the clinical trial protocols.
Positive response data were reported for UCART19, another drug developed with the TALEN process, at the 2017 American Society of Hematology Annual Meeting.14 Preliminary results for the CALM study demonstrated that 5 of 7 adults (≥18 years; range, 18-49) with relapsed or refractory (R/R) CD19- positive B-cell ALL achieved negative minimal residual disease (MRD) 28 days after receiving UCART19. One patient treated at the first dosing level in the study died after developing grade 4 cytokine release syndrome (CRS) and neutropenic sepsis. The rest of the patients experienced mild toxicities with manageable CRS. In the phase I PALL study, all 5 pediatric patients (range, 8 months-16.4 years) with R/R B-cell ALL achieved MRD negativity after UCART19 therapy, with only 1 grade 1 cutaneous acute GVHD, no severe neurotoxicity, and manageable CRS.
“CRISPR obviously gets most of the attention, but there are a couple other gene-editing tools out there, and all of them have advantages and drawbacks. Researchers and companies can choose what works for them and that choice will allow the science to progress faster than CRISPR would on its own,” said Laurent Poirot, PhD, who heads the early discovery program at Cellectis.
The 2 UCART19 trials, both of which are ongoing, are taking place in several European nations and in the United States. This might provide some comfort to people who worry that reluctance to countenance the trial of medications engineered by CRISPR might transfer large swaths of medical research from the United States and Europe to China. UCART19 features modifications that are theoretically very similar to those that can be achieved with CRISPR in the Chinese PLA General Hospital trial. They were simply modified with a technology that has been in development long enough to win the trust of American and European regulators.
Poirot, moreover, sees little evidence that either the FDA or its European equivalent is unduly reluctant to approve trials of genetic modifications.
“When you’re working with new technology, the interactions between regulators and individual companies on the road to the clinic are a mutual education,” he said. “Each side brings different experience and different knowledge to the table. They have overseen far more genetic trials than we have. They have observed things we cannot anticipate. They express their potential concerns before we even design the trial and that allows us to design a trial that addresses those concerns.
“Both the US and the EU seem to be striking a pretty good balance between safety and progress, and a lot of the good clinical work that brings these treatments to patients will take place here,” Poirot added.