One CRISPR Edit Could Be the Fix CAR T Therapy Has Been Waiting For
If you're building or investing in CAR T pipelines, NFIL3 knockout is now a validated engineering target — one edit that could layer into any construct. Competitors without it on their roadmap face a gap.
CAR T therapy has a durability problem, and it keeps the field from cracking solid tumors.
The cells kill cancer beautifully at first — then tire out. Exhausted T cells lose their drive, stop multiplying, and let tumors regroup. It's the reason approved CAR T therapies work well in blood cancers like leukemia and lymphoma but have stalled everywhere else.
A team from Columbia University and University Hospital Tübingen thinks they have found a fix. In a paper published May 12, 2026 in Cancer Discovery, Nayan Jain and colleagues describe deleting a single gene — NFIL3 — via CRISPR/Cas9 and watching CAR T cells stay functional far longer across every model they tested. Not just one CAR design. Not just one mouse strain. Every variant they threw at it.
"We're talking about a platform-level effect," said Prof. Judith Feucht, a pediatric oncologist at University Hospital Tübingen who co-led the work. "Switching off NFIL3 could be a decisive step toward significantly improving the long-term potency of CAR T cells."
The team screened roughly 400 transcription factors before landing on NFIL3 as the culprit driving T cell exhaustion. Deleting it didn't just delay fatigue — CAR T cells showed stronger proliferation, better tumor control, and a transcriptional profile that the researchers say mirrors what predicts favorable outcomes in human patients, not just mice.
That last point matters. Plenty of CAR T improvements look spectacular in xenograft mouse models and vanish in more physiological settings. The Tübingen-Columbia team tested their edit in both immunodeficient xenograft mice and fully immunocompetent syngeneic models, plus multiple CAR constructs. The results held.
"The transcriptional state predicts clinical outcomes — not just preclinical surrogates," said Prof. Michel Sadelain, the Columbia immunologist who pioneered CAR T engineering and co-directed the study. That's a meaningful claim in a field that's had to contend with a long list of promising mouse results that didn't translate.
Why This Changes the Platform Calculus
CAR T developers have tried many routes around T cell exhaustion: dual-targeted constructs, armored cytokines, allogeneic "off-the-shelf" approaches. Most require significant re-engineering of existing programs. NFIL3 knockout is one CRISPR deletion. It doesn't require a new antigen target, a new signaling domain, or a redesign of the vector — it layers onto whatever construct you're already building.
That changes the calculus for competitive programs. Gilead's Kite, Novartis, and Juno/BMS — the makers of the approved CAR T therapies Yescarta, Kymriah, and Breyanzi — have spent years optimizing their constructs. None of them, as of this writing, have disclosed NFIL3 knockout as a development priority. If the effect translates to larger animal models and eventually humans, that's a competitiveness gap that closes only by adding the edit.
The more important second-order effect is on solid tumors. The field has drifted toward tumor-infiltrating lymphocytes (TILs) and NK cell approaches partly because engineered T cells have struggled to maintain potency in the immunosuppressive solid tumor microenvironment. Stronger, less-exhaustable CAR T cells make the T cell route competitive again — and redirect investment that had been flowing toward alternative cell types.
Feucht brings a bench-to-bedside perspective that's rare in the field. She treats children with cancer and runs a research lab. Whatever she found at the bench, she will eventually test against the clinical reality. That's a tighter feedback loop than most academic CAR programs can claim.
What Could Kill This Story
Two risks bear watching before this becomes standard engineering practice. First, the larger-animal question: mice are mice, and immune-oncology has a long history of mouse-model enthusiasm that hasn't survived contact with canine or primate studies. Second, on-target off-tumor toxicity. NFIL3 presumably does things in normal tissues. A CRISPR deletion is permanent. The safety window matters.
The Cancer Discovery paper (DOI: 10.1158/2159-8290.CD-25-1524) was published May 12, 2026.
Sources: