For more than a year now, the big names of CAR-T-cell world have been sharing their excitement about the potential of NK cells to be a much better CAR carrier. There are many hypotheses about why that would be, but you can hypothesize all you want until you give the product to humans. PET scans can lie, but not nearly as much as your average transgenic mouse model.

Well, anti-CD19CAR-NKs have now been given to at least 9 humans, as reported by Dr. Katy Rezvani from the MD Anderson Cancer Center at the SITC Annual Meeting a few days ago, and what she reported was as good as it gets for a first-in-human trial of a new gene/cellular therapy. The five takeaways:

  1. Off-the-shelf cord blood-derived NK cells carrying a CAR can safely be given to humans without HLA-matching. This could decrease both cost and time to treatment.
  2. None of the nine patients had cytokine release syndrome or neurotoxicity. After all, NK cells aren’t big cytokine producers.
  3. Six of nine patients had a complete response and one more had a partial response for a 77% response rate.
  4. One of those CRs and one PR were patients with CLL and Richter’s transformation, another CR was in a 70-year-old with double-hit lymphoma, and yet another in a patient with CLL and 17p deletion. These are tough diseases.
  5. Contrary to what was hypothesized about NK cell longevity, the CAR-NKs persisted for up to six months after treatment. We still don’t know what that means for CAR-Ts, let alone CAR-NKs, but I consider it a win each time experiment refutes theory.

So yes, I am also excited about CAR-NKs. However:

  1. This is an unpublished, non-peer-reviewed, single-center experience. There could be research hospitals out there giving CAR-NKs with results too horrible, or just too middle-of-the-road, for early promotion.
  2. CLL is notoriously difficult to stage, and five of the nine patients had CLL. Consider the MURANO trial (Venetclax-Rituximab combo fo relapsed/refractory CLL), in which the investigator-assessed complete response rate of 26.8% turned to 8.2% on independent review committee assessment.
  3. Follow-up is limited and it is too early to know the response duration. Long enough to get to an allo transplant, at least?
  4. The same limitations in target availability we have in CAR-T cells apply to CAR-NKs. It has to be a surface antigen with limited to no expression on normal cells. There aren’t too many of those, particularly in solid tumors.

Still, kudos to Drs. Rezvani, Shpall, and others at MD Anderson. This is as good as it gets at this early stage.