NK Cells Without a CAR: AB-101 Is Neither CAR-T nor CAR-NK Save

Much of the meeting came for the next CAR T readout. AB-101 drew attention for a different reason. It depletes B cells by a completely different route.

AB-101 is an allogeneic, cord-blood-derived NK cell product. Start with what it is not. It is not a CAR-NK cell, because it carries no engineered receptor. It is not a CAR T cell, because there are no T cells and nothing is modified at all. These are unmodified NK cells, infused alongside an anti-CD20 antibody, killing B cells the way NK cells kill any antibody-coated target: native CD16 engages the antibody Fc and the cell delivers ADCC. The antibody supplies the address. The NK cell makes the delivery. It is an effector boost built on top of rituximab.

The late-breaker (LB0003, presented by Norman Gaylis) reported 31 refractory patients across three diseases, all treated as outpatients. In RA at 6 months (n=7), mean swollen joint count fell 73 percent and 71 percent reached ACR50. In Sjögren disease (n=7), ClinESSDAI improved by 8.6 points. In systemic sclerosis (n=4), mRSS fell by 9.5 points. No CRS, no ICANS, no hypogammaglobulinemia.

The story is in the mechanism, and the supporting poster makes it worth examining. In the characterization and safety study (AB0345, presented by Guillermo Valenzuela), B cells deplete and then reconstitute between 6 and 12 months, repopulating with naive and transitional cells. Serum IgG, IgA and IgM stay normal, and vaccine titers to measles, mumps, rubella and tetanus do not move. Neither study measured a marrow-level reset directly. A single synovial biopsy at day 28 did show B cells cleared from inflamed tissue, so the depletion reaches beyond blood. But preserved immunoglobulins and vaccine titers raise a fair question: if long-lived plasma cells in the marrow were largely spared, this looks more like deep tissue depletion with recovery than the durable, marrow-level aplasia that defines CD19 CAR T. It is a point the data invite, and one worth putting on the table.

Both abstracts frame AB-101 as depletion comparable to CAR T, so the term deserves a careful read. Comparable in peripheral depth at a timepoint, perhaps. In mechanism, no. The honest comparator is rituximab pushed harder. The early curves suggest added depth, but the dataset is small where it counts: single-digit patient counts per disease, open-label single-arm studies, no head-to-head, and the one tissue-level readout rests on a single patient.

None of that sinks the approach. It places it. CAR T and AB-101 are built for different questions: one for a deep immune reset, the other for a hard, repeatable knockdown delivered as an outpatient. Where AB-101 fits, and how durable its responses prove to be, will depend on data we do not yet have: longer follow-up, larger numbers, and ideally a comparison against rituximab itself. For now it reads as an early, distinct entry in the B-cell depletion toolkit, and one worth watching as the numbers mature.