Various companies are exploring ways to manufacture allogeneic chimeric antigen receptor T cells, which could then be administered off the shelf to multiple recipients, thereby expanding access to this immunotherapy. However, challenges, such as finding ways to address the twin complications of graft-versus-host disease and host rejection of the therapeutic product, remain.

One of immuno-oncology's most vibrant fields is that of chimeric antigen receptor (CAR) T-cell therapy, with the first two such treatments, tisagenlecleucel (Kymriah; Novartis) and axicabtagene ciloleucel (Yescarta; Gilead), gaining FDA approval in 2017. Even as companies tinker with autologous, or patient-derived, T cells—to shorten the manufacturing process and enhance therapeutic efficacy—some have begun probing ways and means of engineering an allogeneic version from unrelated donors.

The latter represents “the ultimate possibility, in that one could easily manufacture CAR T cells from a particular source and administer the treatment, off the shelf, to multiple recipients,” says Michel Sadelain, MD, PhD, of Memorial Sloan Kettering Cancer Center in New York, NY.

“It would change market access entirely for CAR T-cell therapy,” adds André Choulika, PhD, of the biotech Cellectis—based in Paris, France, and New York, NY—which paved the way for allogeneic therapy and is the furthest ahead clinically.

Autologous CAR T cells are still a boutique therapy, notes Devon Shedlock, PhD, of San Diego, CAbased Poseida Therapeutics. As well, “the raw material harvested from each patient is typically in pretty bad shape,” he explains, “whereas on the allogeneic side, you could screen donors and select the healthiest ones with easily modifiable, good-quality T cells.”

Steven Rosenberg, MD, PhD, of the NCI, agrees that “an off-the-shelf approach would be a major simplification,” saving both money and time. Ongoing research in this area is “something I completely applaud,” he adds, and despite the challenges involved, getting to FDA-approved allogeneic CAR T cells is, he thinks, an achievable goal.

There are two inherent issues when it comes to allogeneic CAR T-cell therapy, Sadelain says: Donor cells could attack the recipient, causing graft-versus-host disease (GVHD); conversely, a patient's own T cells could reject the infused product, resulting in a lack of therapeutic impact.

Companies pursuing an off-the-shelf strategy have dealt with the first problem by using gene-editing tools such as TALEN or CRISPR/Cas9 to remove the native T-cell receptor (TCR) from donor cells. “I'd say this mostly resolves the issue, but not completely,” observes Marcela Maus, MD, PhD, of Massachusetts General Hospital Cancer Center in Boston, “because the infused product is usually not 100% pure, and we don't yet know what threshold [of purity] is acceptable.”

Even a tiny percentage of T cells retaining their TCR risks the development of potentially lethal GVHD, explains Eric Ostertag, MD, PhD, also of Poseida. As well, the gene editing necessary for allogeneic CAR T cells “has the FDA quite concerned about the possibility of unwanted, off-target mutations that could be pretty detrimental.”

Laurence Cooper, MD, PhD, of Ziopharm Oncology in Boston, MA, notes that although the focus has mostly been on targeting the αβ TCR—found in 95% of human T cells—the minority of T cells expressing the γδ TCR may emerge as a candidate for off-the-shelf therapy. There are “intriguing hints in preclinical and bone marrow transplant data,” he says. “It's thought to be unlikely that this receptor will trespass and attack something vital like a recipient's heart or lungs.”

Meanwhile, to keep rejection at bay, patients are first heavily immunosuppressed through lymphodepleting chemotherapy. Choulika points out that currently, even with autologous CAR T-cell therapy, “you don't get strong, durable antileukemic activity without prior lymphodepletion. You have to create space for the [therapeutic] T cells to expand.”

However, “I don't really see a future for allogeneic CAR T-cell therapy in which every patient has to be severely immunocompromised, and exposed to the consequences thereof, in order to receive this treatment,” Sadelain remarks. “This problem is still awaiting an effective solution; it will require creative immunology and sophisticated engineering.”

One approach, which both Cellectis and Poseida are evaluating, is to disrupt not only TCR expression, but also that of MHC I, in donor T cells. “We're trying to make a product that's fully rather than semi-allogeneic; the removal of MHC I helps our cells better resist host rejection,” Ostertag says.

Yet, Cooper cautions that “if you eliminate all MHC I molecules, you've now removed the ligand for KIR [killer immunoglobulin-like receptor],” which keeps endogenous natural killer (NK) cells in check. Without this inhibitory signal, the infused therapy would be vulnerable to NK cellmediated destruction.

These are complex issues, Maus agrees. In the lab, she and her team are probing MHC I expression levels low enough to mitigate host rejection, yet avoid CAR T-cell elimination by NK cells. The hope is to eventually figure out “a happy medium,” she says.

Essentially, “there's this yin and yang with allogeneic CAR T-cell therapy,” Cooper says. “You need to simultaneously prevent toxicity and improve engraftment; the two coexist on this knife edge.”

“We're still figuring out exactly how our candidate therapies work in patients, to better understand what would be helpful, versus too much, on the gene editing front,” Choulika adds. “I favor moving stepwise—after learning to crawl, we can walk, then start running.”

Sadelain and his group have used CRISPR/Cas9 to create CD19-recognizing TRAC-CAR T cells that are semi-allogeneic—the native TCR is disrupted in the TRAC locus—and, preclinically, much enhanced in potency. “We hope to open a clinical trial by the end of this year,” he says. “It would be an important advance to demonstrate that CRISPR/Cas9-edited cells are at least as good as, if not better than, conventionally engineered CAR T cells.”

Beyond Cellectis's medical first with one of their TALEN-edited products, UCART19—two infants with acute lymphoblastic leukemia (ALL) were successfully treated in 2015 and remain cancer-free—this candidate therapy continues to look promising, Choulika says. Data from CALM and PALL, two phase I trials being conducted outside the United States, were presented during the American Society of Hematology (ASH) 2017 Annual Meeting in Atlanta, GA: Among seven adults and seven children with relapsed/refractory ALL who received UCART19, the rate of complete molecular remission, or negative minimal residual disease, was 83%.

Another candidate, UCART123, is the first allogeneic CAR T-cell therapy greenlighted for phase I trials in the United States, in patients with acute myeloid leukemia and blastic plasmacytoid dendritic cell neoplasm. As well, UCART22, aimed at treating ALL and nonHodgkin lymphoma, is likely to enter clinical studies in the coming months.

“Overall, we're very much technology-agnostic,” Choulika adds, noting that Cellectis holds two U.S. patents for using CRISPR/Cas9 in T cells. That said, “we've optimized TALEN for gene editing at an industrial scale, and we tend to favor it in terms of simplicity, precision, and therapeutic safety.”

Poseida has set its sights on allogeneic CAR T cells that recognize BCMA, a key target in multiple myeloma. At ASH 2017, Shedlock highlighted preclinical findings for their product, P-BCMA-ALLO1. These CAR T cells were not only greatly reduced in their alloreactivity, but also showed strong efficacy in vitro and are being tested further in a mouse model of aggressive human multiple myeloma.

Alongside their proprietary gene editing platform, Cas-CLOVER, which Shedlock describes as highly efficient with low to no off-target cutting, the nonviral piggyBac system used for CAR delivery “enables us to do our editing in resting T cells,” he adds. Viral vectors, by contrast, require actively replicating T cells. “The fewer times we have to force cells to divide, the greater our chances are of getting a high proportion of CAR-bearing stem cell memory T cells,” he explains. “We favor this subset because it's long-lived, durable, and multipotent.”

Ultimately, Ostertag says, the bar will be higher for allogeneic CAR T-cell therapy to cross the FDA-approval finish line. Nonetheless, Rosenberg, who in 2010 first reported the eradication of a patient's advanced lymphoma using autologous CD19-targeted CAR T cells, says that “where there's a will, there's a way.” That patient remains cancer-free.

“I had every expectation, even back then, that CAR T-cell therapy would enter modern oncology,” Rosenberg says. “If it prolongs life and has curative potential—which we've already seen—we'll figure out how to overcome the complexities, because cancer patients need options.” –Alissa Poh

Further Food for Thought

When it comes to allogeneic CAR T-cell therapy, graft-versus-host disease and host rejection aside, “another layer to add is, ‘Where should these T cells come from?’” observes Michel Sadelain, MD, PhD, of Memorial Sloan Kettering Cancer Center in New York, NY. Currently, harvesting T cells from suitable donors is the main strategy, but in vitro differentiation of induced pluripotent stem cells (iPSC) is another possibility.

With the latter, “having to provide the FDA proof that the therapeutic product is completely devoid of undifferentiated pluripotent stem cells would be challenging,” notes Devon Shedlock, PhD, of Poseida Therapeutics in San Diego, CA. Adds Marcela Maus, MD, PhD, of Massachusetts General Hospital Cancer Center in Boston, “Other than for bone marrow transplants, where we use hematopoietic stem cells, we don't have a long safety track record of using pluripotent stem cells clinically, much less genetically manipulated ones.” Nonetheless, Sadelain—who is exploring iPSCs in collaboration with La Jolla, CAbased Fate Therapeutics—thinks state-of-the-art technology will successfully address these matters.

Regarding off-the-shelf potential, how might allogeneic CAR T cells stack up against bispecific T-cell engagers (BiTEs) and other T-cell redirection strategies? For instance, blinatumomab (Blincyto; Amgen) brings cytotoxic T cells and tumor cells in close proximity by binding to CD3 and CD19, respectively; no gene editing is required.

“BiTEs are off the shelf, yes, but the amount of T-cell stimulation they deliver is only that of first-generation CAR T cells, which weren't really effective in the clinic until costimulatory domains like CD28 or 4-1BB were added,” Shedlock says.

Eric Ostertag, MD, PhD, also of Poseida, agrees that strategies such as BiTEs have a place in but are unlikely to be the future of immuno-oncology. At the bench, ever more creative approaches are emerging with CAR T-cell therapy, including combinatorial antigen targeting and sequentially activated T-cell “circuitry” akin to the logic gates of digital systems. Such improvements—which probably can't be made with BiTEs, he adds—will be key in expanding the applicability of CAR T cells, whether autologous or allogeneic, to solid tumors.

Despite skepticism around CAR T-cell therapy's effectiveness in this far more complex realm, Poseida has hopes for another product, P-PSMA-101, which can completely eradicate prostate cancer in mice. “We think it has to do with the persistence and multipotency of the stem cell memory T cells we favor,” Ostertag says. “They're able to differentiate into wave after wave of effector CAR-bearing T cells to eliminate tumors.” P-PSMA-101 is currently autologous, Shedlock says, but “all our optimization efforts are directly applicable to an allogeneic version down the road.”

In the end, the top priority is making sure that more patients have access to CAR T-cell therapy and benefit from it. “While it certainly has advantages, the allogeneic strategy is just one way to get there,” Maus says, adding that “patients like the autologous approach, too—the idea that it's their own T cells attacking their disease.”

“I'd remain open-minded about autologous CAR T-cell therapy and not discount the ongoing progress that will allow better and cheaper iterations over time,” Sadelain agrees. “It just shows how dynamic this whole field is, filled with possibilities.” –AP