The Immune Problem Was Always the Problem: How Engineered "Stealth" Beta Cells Finally Address It

June 4, 2026

crispr type-1-diabetes cell-therapy immune-evasion ipsc

Islet transplantation has worked as a diabetes treatment for decades. The procedure itself isn’t the problem — surgeons can take insulin-producing cells from a donor pancreas, infuse them into the hepatic portal vein, and watch a Type 1 diabetic patient achieve near-normal blood glucose without injecting insulin. It’s been demonstrated repeatedly since the Edmonton Protocol in 2000. The reason it never became standard care has nothing to do with the transplant. It’s what comes after: to keep those cells alive in a foreign body, patients need immunosuppression aggressive enough to cause kidney damage, increase cancer risk, and leave them vulnerable to infections that can kill them. You’re trading one chronic disease for a cocktail of others.

That’s the kind of problem that looks like a treatment problem but is actually a framing problem. We’ve been asking “how do we suppress the immune system enough to protect transplanted cells?” when the real question is “how do we make the cells invisible to the immune system in the first place?”

The Biology of Being Seen

The immune system identifies foreign cells through surface proteins called HLA (human leukocyte antigen) molecules. Think of them as name badges — your CD8+ T cells constantly read these badges, and anything displaying the wrong name gets flagged for destruction. Transplanted beta cells from a donor wear the wrong name badge, so the immune system kills them. The traditional answer is to chemically suppress the immune system so it can’t read any name badges at all. This works, but it’s like disabling your entire security system to let one trusted person in.

The CRISPR answer is different: don’t suppress the immune system, reprogram the cells. Strip the name badges off entirely.

That’s roughly what Century Therapeutics is doing with CNTY-813, their iPSC-derived beta islet replacement therapy. Using their Allo-Evasion 5.0 platform, they’re applying multiplexed CRISPR-Cas9 edits to eliminate HLA Class I and Class II expression — removing the surface markers that CD8+ and CD4+ T cells use to identify foreign tissue. Without those markers, the adaptive immune system largely ignores the cells.

But here’s where it gets more elegant. Stripping HLA creates a new problem: Natural Killer cells are specifically tuned to destroy cells that lack normal HLA expression. It’s a failsafe. The immune system assumes that if a cell has hidden its identity, it must be infected or cancerous. NK cells use a simple heuristic — “if you’re not displaying your credentials, I’m killing you.”

The engineering answer to this is overexpressing CD47, a surface protein that functions essentially as a “don’t eat me” signal. CD47 binds to receptors on NK cells and macrophages and sends an inhibitory signal: stand down. The cell is flying a flag that reads “I belong here” even though it came from an iPSC line in a bioreactor. Combined with the HLA deletions, CNTY-813 is attempting to make donor-derived beta cells genuinely invisible — not immunosuppressed-invisible, but architecturally invisible, at the molecular level.

Why This Changes the Calculus

I’ve been thinking about the difference between solving a problem and routing around it. A lot of biotech progress is routing — finding workarounds that manage symptoms well enough that we stop looking for the root cause. Immunosuppression for transplant rejection is a classic reroute. It works well enough that the field moved on, even though the underlying incompatibility was never resolved.

What CRISPR-based immune evasion represents is the first serious attempt to actually solve the compatibility problem. Not suppress the response. Not reroute around it. Solve it.

The preclinical data for CNTY-813 is compelling: rapid reversal of diabetes in streptozotocin-induced diabetic mice, sustained normoglycemia, detectable human C-peptide production (which confirms the transplanted cells are doing the secretion, not the host), and demonstrated resistance to T cell, NK cell, and humoral immune attack. The IND filing is targeted for 2026, backed by $135 million in financing.

That’s not a curiosity anymore. That’s a clinical program.

The Harder Question

There’s a version of this story where everything works and CNTY-813 becomes the first functional cure for Type 1 diabetes — not a management strategy, a cure. Where a one-time infusion of engineered iPSC-derived beta cells restores endogenous insulin production indefinitely, with no systemic immunosuppression required.

I find that outcome genuinely plausible in a way I didn’t two years ago. The engineering is sound, the mechanism is specific, and the preclinical results track. Whether the 2026 IND data replicates in humans is the open question — there are always surprises when you move from mouse models to the complexity of a human immune system, and “robust resistance” in an STZ mouse model is not the same as durable tolerance in a person with decades of T1D and a fully primed autoimmune response.

But here’s what I keep coming back to: if the immune evasion holds in humans, this technology platform doesn’t stop at diabetes. The same multiplexed editing strategy that protects beta cells could theoretically protect any allogeneic cell therapy — engineered islets for other pancreatic conditions, dopaminergic neurons for Parkinson’s, cardiomyocytes for heart failure. The diabetes trial is effectively a proof of concept for a much larger question about whether we can manufacture immune-privileged cells at scale.

That question doesn’t have an answer yet. But for the first time in a long time, it has a credible experimental path to one.