Case Study: Cell Therapy Freedom to Operate: Clearing a CD19 CAR-T Candidate

A cell therapy freedom to operate review decides whether a promising CAR-T construct can actually reach patients, because CD19 cell therapy is one of the most densely patented corners of biotechnology. This representative engagement shows how PerspireIP cleared an autologous anti-CD19 CAR-T candidate across its binder, costimulatory, vector and manufacturing layers — and why validity analysis, not just infringement mapping, changed the client’s build decision.

The Challenge

A clinical-stage cell-therapy developer asked us to clear an autologous anti-CD19 chimeric antigen receptor (CAR) T-cell candidate for relapsed or refractory B-cell lymphoma before it committed capital to a commercial manufacturing suite and opened a Series B round. CD19 is the most crowded target in the field: of the seven CAR-T products the U.S. Food and Drug Administration has approved since 2017, four read on CD19, and the two first-movers — Novartis’ Kymriah and Gilead/Kite’s Yescarta — share a lineage of foundational patents that still shape what a newcomer can and cannot build.

The board’s question was blunt: if we scale this construct, whose patents do we walk into, which of those patents would actually survive a validity challenge, and what would a licence cost versus a design-around? None of that could be answered from the scientific literature alone. It required a structured freedom-to-operate (FTO) analysis that separated the patent thicket around the binder, the costimulatory domain, the vector and the manufacturing process, and then pressure-tested each blocking family for enforceability.

Our Approach

We ran the clearance in five layers, because a CAR-T construct is not one invention but a stack of them, and a licence or a design-around is available at a different layer than most developers expect. A disciplined cell therapy freedom to operate review treats each layer as its own search universe rather than collapsing them into a single ‘CAR-T’ query.

• The antigen binder. Most approved CD19 products derive their single-chain variable fragment from the FMC63 clone. We charted the scFv families, their claim scope and their legal status to see whether the client’s binder read on them or fell outside.
• The costimulatory domain. Second-generation CARs add either a 4-1BB or a CD28 signalling domain. Kymriah uses 4-1BB; Yescarta uses CD28. Each route sits behind different owners, so the choice is as much an FTO decision as a biology decision.
• Hinge, transmembrane and spacer elements. Often overlooked, these carry their own narrow but enforceable claims.
• The vector and manufacturing process. Lentiviral and retroviral vector patents and cell-processing method claims form a separate thicket that can block a product even when the receptor itself is clear.
• Methods of treatment and dosing. Indication- and regimen-specific claims that can be designed around through labelling and trial design.

For each layer we built classification searches across the relevant CPC groups (C12N, C07K and A61K35/17), clustered the results by assignee, pulled the live claim sets, and checked legal status and expiry. We then claim-charted the client’s construct against every family that survived the first screen, and graded each blocking family not only for infringement risk but for validity exposure.

What the Cell Therapy Freedom to Operate Search Found

The validity grading is where the cell therapy freedom to operate search earned its keep, and the reference point was a real decision. In Juno Therapeutics v. Kite Pharma, 10 F.4th 1330 (Fed. Cir. 2021), the Federal Circuit erased a $1.2 billion judgment — the largest patent award in the sector — by holding that the asserted claims of U.S. Patent No. 7,446,190 were invalid for lack of written description. The patent, which traces to Memorial Sloan Kettering, claimed a broad genus of scFv binders but disclosed only two working examples, and the court found the specification did not show the inventors possessed the full genus they claimed. The Supreme Court denied review in 2022.

That holding is not trivia; it is a lever. It tells a newcomer that the broadest, most frightening CAR-T claims — the genus claims that appear to cover any binder against a target — are often the most fragile under 35 U.S.C. § 112. A family that looks like an absolute block on a claim chart may be a paper tiger once its written-description support is read against the construct in question.

Applying that lens, we sorted the client’s blocking families into three tiers: narrow, well-supported claims that genuinely read on the construct and could not be avoided without redesign; broad genus claims with the same written-description vulnerability the Federal Circuit punished in Juno; and expired or soon-to-expire families that posed no forward risk. We also mapped the foundational 4-1BB and CD28 families — including the St. Jude construct (U.S. Patent No. 8,399,645, naming Dario Campana) that Juno licensed, and the University of Pennsylvania 4-1BB CAR work licensed to Novartis — so the client could see exactly which costimulatory route carried which owner.

The Outcome

The deliverable was a risk-ranked clearance readout the board could act on in a single sitting, not a literature dump. Every figure below is a scenario figure illustrating the shape of the work, not a measured result for a named client.

• A screened universe of roughly forty patent families across the five construct layers, reduced to a short list of genuine blockers.
• Three high-risk families flagged as narrow and well-supported — the ones that warranted either a licence or a redesign rather than a validity bet.
• Two viable design-around routes, the cleanest of which swapped the contested costimulatory configuration and adopted a binder outside the charted scFv families.
• A licensing shortlist naming the owners and the layer each controlled, so business development could open the right conversations instead of the loudest ones.
• A regulatory-exclusivity timeline mapping the client’s own twelve-year BPCIA runway against the expiry of the blocking families.

With that in hand, the developer narrowed its construct before locking the manufacturing process, folded the clearance readout into its Series B data room, and entered licensing talks from a position of knowing which patents it actually needed. The expensive mistake — scaling a process around a receptor that later proved blocked — was taken off the table before a dollar of capital equipment was ordered.

What This Means for Similar Matters

The lesson that travels beyond this matter is that a cell therapy freedom to operate review is a construct-layer exercise, not a product-name search, and that infringement risk and validity risk have to be scored together. A claim that reads on your construct is only a block if it would survive — and after Juno v. Kite, the broadest claims in this field are frequently the weakest.

The second lesson is sequencing. FTO done before the manufacturing process is frozen is a design input; done afterwards it is a liability report. The developers who clear early buy themselves the cheapest option of all, which is to change the construct on paper instead of in a GMP suite.

The Regulatory Layer: BPCIA Exclusivity on Top of Patents

Patents are only half the exclusivity picture for a biologic, and FTO that ignores the regulatory layer gives a false sense of runway. CAR-T products are licensed under a Biologics License Application, which means the Biologics Price Competition and Innovation Act (BPCIA) applies. The reference product earns twelve years of regulatory exclusivity from the date of first BLA approval, during which the FDA will not approve a biosimilar of that product.

Layered on top is the BPCIA ‘patent dance’ of 42 U.S.C. § 262(l): a biosimilar applicant must hand its application and manufacturing information to the reference-product sponsor within twenty days of acceptance, and the sponsor then lists the patents it could assert within sixty days. For our client, advancing an original construct rather than a biosimilar, the practical point was the reverse — its own twelve-year clock and its own future patent list become assets worth building deliberately, which in turn shaped how aggressively it should prosecute its own filings.

Data Sources

The market and patent data referenced above comes from:

• Juno Therapeutics, Inc. v. Kite Pharma, Inc. (Fed. Cir. 2021) — Federal Circuit opinion reversing the $1.2B judgment; U.S. Pat. 7,446,190 invalid for lack of written description
• FDA — Approved Cellular and Gene Therapy Products — Official list of FDA-approved CAR-T and other cell/gene therapies
• 42 U.S.C. § 262 — Biologics Price Competition and Innovation Act — Statutory 12-year exclusivity and the 262(l) biosimilar patent-dance framework

Discuss a Cell-Therapy FTO Matter

Planning to scale a CAR-T or other cell-therapy construct? We can map the blocking landscape before you commit to manufacturing.

Related PerspireIP work: Freedom-to-Operate Analysis service · Medical-devices FTO case study · How to conduct a freedom-to-operate analysis.

Frequently Asked Questions