Q&A: Philip Greenberg on T Cells, IO's Future

As personalized license plates go, “DRTCELL” is both unique and unsurprising, given that its owner, Philip Greenberg, MD, has lived and breathed immunology at Fred Hutchinson Cancer Center in Seattle, WA, since the mid-1970s.

“When I joined the Hutch, it was still the early days of bone marrow transplantation” for hematologic malignancies, Greenberg recalls. Graft-versus-host disease was a pervasive toxicity, “but it became clear there was also a graft-versus-leukemia effect mediated by T cells. I saw no reason why we couldn't eliminate the toxicity by making therapeutically beneficial, cancer-specific T cells within the next few years. I was off by several decades.”

Now a reality, adoptive transfer of antigen-specific T cells remains a personal favorite on Greenberg's considerable list of accomplishments. Another is watching his mentees blossom “now that immuno-oncology is really happening after such a long, bumpy road.”

In April, Greenberg became president of the American Association for Cancer Research (AACR) for 2023–2024. He spoke with Alissa Poh for Cancer Discovery, discussing some of the initiatives he hopes to pursue during his term and where the immuno-oncology (IO) field is headed.

I'd like to help our members more effectively communicate what we do. Virtually all of us are passionate about research; we just often have trouble telling others why we're excited. There are communications professionals out there to create educational programs that we could make accessible to watch through AACR. The idea is to train us to explain our work better—especially to those from underrepresented communities, so they're inspired to consider science a feasible, rewarding career option.

The pandemic made us painfully aware how poorly we've communicated the value of science. We'd labored under the assumption that people trusted us, believed in what we did, and it was a real eye-opener—truly terrifying—to discover quite the opposite. I think our attitude has been somewhat naive. We should take responsibility for what we haven't done well and fix it.

Yes, to improve access to the latest technologies—our capacity for high-throughput sequencing and proteomics, for instance, has never been greater, yet not all institutions, especially smaller centers, can afford the necessary equipment and facilities. Establishing “bio-hubs” throughout the country that all scientists can access would help democratize research. This will take longer than a year, but what I hope to do is create enough momentum that others can sustain until it's finally achieved.

We're advancing a new iteration of our WT1-targeting TCR [T-cell receptor] therapy beyond acute myeloid leukemia to solid malignancies. We're also exploring a couple of TCR therapies for pancreatic cancer. One is mesothelin-specific, currently in the dose-escalation trial phase. It's safe so far, and there's tumor infiltration even at the lowest dose, which is encouraging. The other targets KRAS G12V. We're pretty excited about it, because we've come up with a way to engineer coordinated CD4⁺ and CD8⁺ T-cell responses.

Unlike MHC-independent CAR [chimeric antigen receptor] T cells, with our type of therapy, a given TCR that's MHC class I restricted won't work on CD4⁺ T cells. Yet this subset is really important for keeping CD8⁺ T-cell exhaustion at bay—their absence in many TCR therapy trials has been a real limitation in terms of lasting efficacy. We've now engineered these cells so they can function even under class I restriction. Such coordinated CD4⁺/CD8⁺ responses will almost certainly improve durability.

The capacity is incredible—the things we can do with it and synthetic biology are exciting. To keep tumors with upregulated FAS ligand from triggering T-cell apoptosis, we've changed this death signal to one that's prosurvival by swapping, on T cells, the FAS receptor's cytoplasmic tail with costimulatory 4-1BB. Likewise, we've been able to turn TGFβ into a positive signal, as opposed to one that is inhibitory.

In terms of improving efficacy, if we generate an effector T cell and it doesn't eradicate the tumor, understanding why is crucial: This is a living reagent—what kept it from finishing the job? How can these obstacles be turned into advantages? Carrying out high-dimensional multiomic analyses on our patients should help provide clarity.

It depends whether you're a “lumper” or a “splitter”—the number of subsets you define can get infinitely large or infinitesimally small. The reality, in terms of true phenotypes, is probably somewhere in the middle. And there are lots of different tools, including RNA velocity, to track cellular transitions and whether a subset is in fact relevant to response/resistance. That's what's really changed in the last decade—the speed and depth with which we can interrogate a potential issue, then apply our knowledge to address it, is breathtaking.

Much of it will be about figuring out how we combine different components in our IO toolbox to create benefit without overwhelming toxicity. High-dimensional analyses will keep uncovering new targets, and the ways we modulate different pathways will change. Our understanding of cancer's vulnerabilities is shifting, and instead of merely trying to capitalize on a small therapeutic index, we're going after far more powerful, highly specific drugs. We'll also look for ways to engage innate and adaptive immune responses that are additive, accomplishing the end goal of tumor eradication.