Oncofusion

start / graph / target-cd19-car-secreting-a-clec12a-bridge-against-lineage-switch

Targetmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.

A CD19 CAR that secretes a CLEC12A bridge: aimed squarely at lineage switch, and it inherits both antigens' failure modes

also called bridging protein CAR · CD19-anti-CLEC12A · biparatopic bridge · CAR-T engager

The hardest failure in treating this leukaemia with CD19-targeted cell therapy is that the cancer changes clothes: it stops being a B-cell leukaemia and comes back as a myeloid one, where the CD19 therapy cannot see it. One design tries to cover both at once - a CD19 CAR cell that also secretes a small adaptor protein letting it attack a myeloid marker. It is clever and it is preclinical, and the catch is that the myeloid marker it picks has since failed in patients for a reason this design does not fix.

The same thing, for a scientist

**PMID 34253594** (Rennert/Ambrose, *Mol Cancer Ther* 2021, verified): a CD19-anti-CLEC12A bridging protein binds CAR19 T cells and CLEC12A, so that a **CD19-directed CAR acquires CLEC12A reactivity without re-engineering the CAR**; biparatopic targeting increased potency, with activity against aggressive AML in vitro and in vivo. In KMT2A-rearranged infant ALL, where **lineage switch to myeloid is the characteristic escape from CD19-directed therapy**, covering both antigens with one product is the obvious structural answer. **It is preclinical, and the CLEC12A arm has since read out badly in humans**: the KITE-222 CLL-1 CAR phase 1 saw no clinically meaningful responses, with CLEC12A-negative blasts persisting while CLEC12A-positive blasts cleared.

Go deeper
Why the design is the right shape for this disease

Lineage switch is not a rare curiosity in KMT2A-rearranged disease - it is the characteristic escape route from antigen-directed therapy, and this graph carries it in risk-lineage-switch-cd19-directed-therapy and a dozen related entries. A switched clone is not a new leukaemia; it is the same fusion in a myeloid coat, and a therapy that only sees B-lineage antigens is blind to it by construction.

The bridging-protein architecture addresses that without a second product. The CAR stays CD19-directed; the secreted bridge is what confers CLEC12A reactivity. Practically that means one manufacturing process, one infusion, and a myeloid arm that is present from day one rather than deployed after the switch has been detected - which matters because detection of a switch is itself slow and, as this graph's flow-cytometry entries record, sometimes missed.

What it does not fix, and the read-out that has happened since

The paper is preclinical. In vitro and xenograft activity against AML, published 2021, with no clinical data in the abstract.

The CLEC12A arm has since been tested in patients and failed in a way this design does not address. In the KITE-222 CLL-1 CAR phase 1, no clinically meaningful responses were seen at any dose, and where the CAR expanded, CLEC12A-positive blasts were nearly cleared while CLEC12A-negative blasts persisted with no fall in total blast count. That is antigen-negative escape on the myeloid arm — the same failure the bridge was added to prevent on the lymphoid arm. A product covering CD19 and CLEC12A covers two antigens, not two escape routes, if a blast can be negative for both.

A second inherited problem: on-target myeloid toxicity. CLEC12A is not confined to leukaemic cells, and the clinical CLEC12A experience includes severe pancytopenia and deaths from infection in chronic agranulocytosis. A bridge that is secreted continuously by an engrafted CAR population is harder to stop than an infused antibody, which is a pharmacological difference that cuts against this architecture specifically.

What would make it worth pursuing anyway. The failure above is an argument against CLEC12A as the myeloid partner, not against the bridging architecture. The same CAR19 platform could carry a bridge to a different myeloid antigen, and this graph's immunophenotype beat has candidates the CLEC12A programmes did not test — including the carbohydrate epitopes CD15 and CD65s, which no gene-keyed target-discovery pipeline can even enumerate.

How this could bite us back
  • This is preclinical and should not be read as a candidate for an infant. The myeloid antigen it uses has since failed in a phase 1 with no meaningful responses, and the clinical CLEC12A experience includes severe pancytopenia and deaths from infection in chronic agranulocytosis.
  • A continuously secreted bridge is harder to withdraw than an infused antibody. If on-target myeloid toxicity emerges, the effector population producing the bridge is engrafted and self-renewing - a safety property that is worse than it looks on a mechanism diagram, and one the preclinical work does not test.
  • Covering two antigens is not covering two escape routes. A blast negative for both CD19 and CLEC12A defeats this product, and antigen-negative escape on the CLEC12A arm has already been observed in patients.
Still open
  1. Has any bridging-protein CAR been tested against a KMT2A-rearranged B-lineage model, as opposed to AML? The architecture's rationale here is lineage switch, which is a B-lineage starting point.
  2. Could the same CAR19 platform carry a bridge to a myeloid antigen other than CLEC12A, given the KITE-222 read-out?
  3. How is a continuously secreted bridge stopped if myeloid toxicity develops? That is the pharmacological question this architecture raises and the abstract does not address.
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    A CD19-anti-CLEC12A biparatopic bridging protein binds CAR19 T cells and the antigen CLEC12A, conferring CLEC12A-directed cytotoxicity on CD19-directed CAR T cells; biparatopic targeting increased the potency of bridging-protein-mediated cytotoxicity, with activity reported against highly aggressive acute myeloid leukaemia in vitro and in vivo.

    PMID 34253594DOI
    What this evidence cannot show

    Abstract fetched and read back on 2026-09-20; full text not read. **Moderate because it is entirely preclinical** - cell lines and xenografts, no patient data. The models are AML; **no KMT2A-rearranged infant B-lineage system is named in the abstract**, so its relevance to lineage switch in this disease is inferred from the architecture rather than demonstrated. 'Highly aggressive' is the authors' characterisation of their models.

    Evidence that argues against this (1)
    • The CLEC12A arm has since failed in humans in a way this architecture does not address: antigen-negative escape, with CLEC12A-negative blasts persisting while positive ones cleared. Adding a second antigen protects against loss of the first, not against a clone negative for both - and the switched clone is selected by exactly that pressure. Daver N et al. 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Lineage switch after antigen-targeted immunotherapy is documented across an international series of cases, the majority originating as B-lineage acute lymphoblastic leukaemia, which is the escape route a CD19-directed product faces in KMT2A-rearranged disease.

    PMID 40193715DOI
    What this evidence cannot show

    Carried from this graph's existing lineage-switch entries rather than re-read in full for this entry. **Strong as a statement that the phenomenon is real and recurrent**; the series is a collected case set, so it describes switches that were recognised and reported and cannot give an incidence. This graph separately records that flow cytometry can miss a switch.

    Evidence that argues against this (1)
    • A collected case series over-represents dramatic and well-documented events, so it cannot support a claim about how often switching happens - and the argument for a dual-antigen product depends on frequency, not only on existence. Nothing here establishes that the myeloid arm would be needed in most patients. Rennert PD et al. 2021
Connected to