CAR-T Cell Therapy Patents

CAR-T cell therapy patents cover CAR-construct (chimeric antigen receptor) architecture innovations; binding-domain and target innovations; manufacturing and gene-delivery innovations; and allogeneic/next-generation cell innovations — with IP held by approved-product owners, academic licensors, and allogeneic-platform startups. MAJOR CAR-T PATENT HOLDERS: NOVARTIS (Kymriah/tisagenlecleucel): the first approved CAR-T (CD19, 4-1BB costimulation), built on University of Pennsylvania / Carl June foundational patents (exclusively licensed). GILEAD / KITE (Yescarta/axi-cel, Tecartus/brexu-cel): CD19 CAR with CD28 costimulation, traced to NCI/Rosenberg work. BRISTOL MYERS SQUIBB / JUNO / CELGENE (Breyanzi/liso-cel CD19; Abecma/ide-cel BCMA): Juno held foundational costimulatory-domain patents (Sloan Kettering/Sadelain lineage) and famously won, then lost on appeal, a $1.2B judgment against Kite over a CAR costimulatory-domain patent (Federal Circuit invalidated it for lack of written description). JOHNSON & JOHNSON / LEGEND BIOTECH (Carvykti/cilta-cel): BCMA CAR with dual-epitope binders. OTHERS: Allogene, Caribou Biosciences, Cellectis (TALEN allogeneic), Precision BioSciences, 2seventy bio, Arcellx, and the foundational academic holders — UPenn (June), Memorial Sloan Kettering (Sadelain), NCI (Rosenberg), and St. Jude (4-1BB).

Chimeric-antigen-receptor architecture innovations; costimulatory-domain innovations; binding-domain and antigen-target innovations; and signaling and hinge/transmembrane innovations represent core CAR-T patent domains — and the costimulatory domain has been the single most litigated CAR-T patent feature. CAR-ARCHITECTURE PATENTS: the chimeric antigen receptor itself — an extracellular antigen-binding domain (usually an scFv), a hinge/spacer, a transmembrane domain, one or more intracellular costimulatory domains, and the CD3ζ activation domain — claimed across generations (1st: CD3ζ only; 2nd: one costimulatory domain; 3rd: two; 4th/'armored': plus cytokines/switches; 5th: cytokine-receptor domains). COSTIMULATORY-DOMAIN PATENTS: the choice and incorporation of CD28 versus 4-1BB (CD137) costimulation drives persistence and potency and is heavily patented — this is exactly what Juno asserted against Kite (the $1.2B verdict was later invalidated on written-description grounds, a landmark lesson that even foundational CAR patents can fall under §112). BINDING-DOMAIN / TARGET PATENTS: the scFv or binder against a specific antigen (CD19, BCMA, CD22, GPRC5D), humanized/affinity-tuned binders, and dual/tandem-targeting to prevent antigen escape. SIGNALING / TUNING PATENTS: hinge and transmembrane optimization, signaling-domain engineering, and tonic-signaling reduction. The CAR construct (especially its costimulatory and binding domains) is the core composition-of-matter, and §112 written description is the key validity battleground.

Cell-manufacturing and gene-delivery innovations; autologous-process innovations; allogeneic and gene-editing innovations; and next-generation safety/control innovations represent additional CAR-T patent domains. MANUFACTURING / GENE-DELIVERY PATENTS: T-cell engineering via lentiviral or gammaretroviral vector transduction (and the vector designs themselves), transposon (Sleeping Beauty/piggyBac) and mRNA/electroporation non-viral delivery, T-cell selection/activation (anti-CD3/CD28 beads), and ex vivo expansion and formulation. AUTOLOGOUS-PROCESS PATENTS: the patient-specific workflow (apheresis → transduce → expand → cryopreserve → infuse), closed/automated manufacturing systems, vein-to-vein time reduction, and rapid/short-duration manufacturing. ALLOGENEIC PATENTS: off-the-shelf donor-derived CAR-T made universal by gene-editing out the endogenous T-cell receptor (TRAC knockout, to prevent graft-versus-host disease) and HLA/CD52 (to evade host rejection) using CRISPR, TALEN (Cellectis), base editing, or ARCUS meganuclease (Precision) — these editing strategies and the resulting universal cells are key allogeneic IP. SAFETY / CONTROL PATENTS: suicide/safety switches (iCasp9), logic-gated and switchable CARs (ON/OFF, AND/NOT gating to spare normal tissue), armored CARs secreting cytokines, and dominant-negative/checkpoint-resistant designs. Allogeneic gene-editing is the highest-value frontier because it would convert CAR-T from a bespoke per-patient product into a scalable off-the-shelf drug.

CAR-T startup IP strategy must navigate a dense foundational-patent thicket (UPenn/June, MSK/Sadelain, NCI/Rosenberg, St. Jude — most exclusively licensed to Novartis, BMS/Juno, Kite), the costimulatory-domain and binding-domain estates of approved products, gene-editing IP for allogeneic (CRISPR Broad/Berkeley, TALEN, ARCUS), §112 written-description law (Kite v. Juno shows broad CAR claims can be invalidated), and FDA/biologics regulation; understand that 2nd-generation CD19/BCMA CD28-or-4-1BB CARs are heavily patented and often require licenses, that the durable defensible asset is a NOVEL binder/target, a novel construct feature, or a novel manufacturing/allogeneic method, and that written-description support must be robust; identify whitespace in novel targets (solid-tumor antigens), allogeneic off-the-shelf platforms, logic-gated/armored safety designs, and non-viral/rapid manufacturing. CAR-T STARTUP IP STRATEGY: THE FOUNDATIONAL CONSTRUCT IS LICENSED — NOVEL BINDERS, TARGETS, AND METHODS ARE THE IP: 2nd-gen CD19/BCMA CARs with CD28/4-1BB are covered by UPenn/MSK/NCI estates — patent a NOVEL binder, a novel antigen (especially solid-tumor), or a novel construct/manufacturing feature; WRITTEN DESCRIPTION (§112) IS THE VALIDITY BATTLEGROUND: Kite v. Juno invalidated a $1.2B costimulatory-domain patent for inadequate written description — claim what you actually made and characterized, with full support; ALLOGENEIC (OFF-THE-SHELF) GENE-EDITING IS HIGHEST-VALUE WHITESPACE: TRAC/HLA-knockout universal CAR-T (CRISPR/TALEN/base-edit/ARCUS) would make CAR-T scalable — but layer in gene-editing FTO (Broad/Berkeley CRISPR); SOLID-TUMOR TARGETS AND LOGIC-GATED SAFETY ARE OPEN: solid-tumor antigens, dual/tandem targeting against escape, and ON/OFF/AND-NOT logic-gated and armored designs are active terrain; WHEN TO PATENT: NOVEL CAR/CELL WITH MEASURED ACTIVITY: file once a construct shows measured results (antigen binding + in vitro cytotoxicity + in vivo persistence/efficacy + safety profile) with robust §112 support — measured target affinity, cytotoxicity, persistence, and response rate are the critical CAR-T IP metrics; KEY FTO CHECKLIST: UPenn/Carl June 4-1BB CAR (Kymriah); MSK/Sadelain costimulatory CAR; NCI/Rosenberg CD28 CAR (Yescarta); St. Jude 4-1BB; scFv binder CD19/BCMA/CD22/GPRC5D humanized dual-tandem; lentiviral/gammaretroviral transduction, Sleeping Beauty/piggyBac transposon, mRNA electroporation; allogeneic TRAC/HLA/CD52 CRISPR/TALEN/ARCUS knockout; iCasp9 suicide switch logic-gated armored; §112 written-description (Kite v. Juno); CRISPR FTO (Broad/Berkeley).