Not medical advice. The Journal of Biohacking reports published research and teaching models. Nothing on this page is medical advice, a diagnosis, or a recommendation to use any compound, product, or therapy. Where a dose appears, it is a figure reported in a cited study or by its sponsor, not an instruction. Editorial position on dosing.
The Off-the-Shelf Reset
Autologous CAR-T just proved that deep B-cell depletion can put refractory rheumatoid arthritis into drug-free remission. Nkarta's NKX019 asks the delivery question: can the same reset come out of a freezer, in a community clinic, with the patient home after two hours of observation? The first autoimmune data arrive at ACR Convergence 2026. This dossier verifies what is actually being shown, and what is not.
The verdict
CAR NK is the logistics play on the one autoimmune hypothesis with published human drug-free remissions behind it. Its engineering trade is explicit: it gives up the years-long persistence of CAR-T for a donor product that needs no manufacturing slot, causes no graft-versus-host disease, and can be re-dosed. On evidence today it sits behind autologous CD19 CAR-T and roughly level with CAR-Treg, with zero published autoimmune efficacy data and the weakest durability biology of the cell-therapy tier. November 2026 settles tolerability. It does not settle RA.
First, the fact check
Load-bearing claims were checked against Nkarta's own releases and the trial registry. The shape of that check is below. Unverified items are labelled in the text.
- CONFIRMEDFirst autoimmune CAR NK data at ACR Convergence 2026. Nkarta's 8 September 2026 release: three presentations, 6 to 11 November, Orlando, the company's first overview of NKX019 safety and activity in autoimmune disease.
- CONFIRMEDDonor-derived, off-the-shelf. NKX019 is allogeneic and cryopreserved, built from healthy adult donor blood NK cells. No per-patient manufacturing.
- CONFIRMEDTwo-hour observation. FDA agreed in April 2026 to cut post-infusion monitoring from 24 hours to 2 and enable community-clinic dosing. The caveat the headline skips: chemotherapy lymphodepletion still comes first.
- CONFIRMEDNtrust-2 basket. Systemic sclerosis, myositis, ANCA vasculitis and rheumatoid arthritis, per the current registry record. But RA was added only in April 2026 (see next item).
- CORRECTEDNo RA readout is coming at ACR 2026. The three presentations are myositis activity (oral), pooled safety across indications (oral), and scleroderma activity (poster). Nkarta's CEO calls myositis and scleroderma "our leading indications". The RA cohort is the newest in the basket; Ntrust-2 primary completion is estimated October 2028.
- UNVERIFIABLEThe abstract contents. Not public at verification time. Anything said before 6 November about what the data show, in either direction, is speculation. Nkarta will post slides after the meeting.
The product
NKX019 is a natural killer cell with two pieces of engineering, each aimed at a known weakness of its class.
A CD19 homing head
A humanised chimeric antigen receptor against CD19, the surface marker carried by essentially the whole B-cell lineage: naive, memory and plasmablast. Published construct descriptions give it an OX40 co-stimulatory domain and a CD3zeta signalling domain.
CD19 is the same target as the CAR-T products that produced drug-free remission in lupus and RA. The depletion thesis is shared; the cell carrying it is not.
A built-in life-support signal
A membrane-bound form of interleukin-15 (mbIL15), tethered to the cell surface as an autocrine growth signal. Donor NK cells normally vanish in days without cytokine support; the published programme figure is an in vivo half-life of up to 28 days, with no systemic IL-2 and no supplemental cytokines in the trials.
Because NK cells carry no T-cell receptor, donor cells cannot cause graft-versus-host disease. That single fact is what makes an off-the-shelf product possible.
The trials
Two Phase 1/2 basket studies, one architecture: lymphodepletion with fludarabine and cyclophosphamide (cyclophosphamide alone if significantly cytopenic), then three doses of NKX019, up to 12 patients per dose level per indication, with re-dosing allowed to restore or deepen a response. Primary endpoints are dose-limiting toxicities and adverse events; persistence and anti-drug antibody assays run for up to 2 years.
| Ntrust-1 · NCT06557265 | Ntrust-2 · NCT06733935 | |
|---|---|---|
| Basket | Lupus nephritis, primary membranous nephropathy | Systemic sclerosis, idiopathic inflammatory myopathies, ANCA-associated vasculitis, rheumatoid arthritis |
| Enrolment | 120 estimated | 240 estimated (raised as cohorts were added) |
| Started | 13 June 2024 | 4 November 2024 |
| Primary completion | April 2027, estimated | October 2028, estimated |
| Population | Ages 18 to 75, active biopsy-proven refractory kidney disease | Ages 18 to 75; RA cohort: seropositive, CRP above 3, at least 6 swollen and 6 tender joints, failed a csDMARD plus two biologics of distinct mechanisms (or biologic plus JAK inhibitor) |
| Key endpoints | Renal response (PERR, CRR), SLEDAI-2K, steroid-free proportion | RA: ACR20/50/70, DAS28-CRP remission, CDAI and SDAI remission, steroid-free proportion; SSc: skin score; IIM: Total Improvement Score; AAV: BVAS remission |
Plus two investigator-sponsored trials: SLE at Columbia University (NCT06518668) and generalised myasthenia gravis, per company materials. Current dosing across all indications is at the 4 billion cells per dose level (company claim, Q2 2026 release).
What November actually shows
ACR Convergence 2026, Orlando, per Nkarta's 8 September 2026 release.
- Sun 8 Nov · Oral 0903Myositis. Preliminary safety, PK/PD and clinical activity of NKX019 in idiopathic inflammatory myopathy. Valenzuela.
- Mon 9 Nov · Poster 1179Scleroderma. Preliminary safety, PK/PD and clinical activity in systemic sclerosis. Khanna, University of Michigan.
- Wed 11 Nov · Oral 2938Pooled safety. Overall, immune-mediated and haematologic safety after Flu/Cy lymphodepletion and NKX019 across B-cell-mediated autoimmune diseases. This is the tolerability verdict for the whole platform, and the one to read first.
Why anyone believes the safety story: the cancer record
What Phase 1 in lymphoma showed
- 19 patients with relapsed or refractory B-cell malignancies, median 4 prior therapy lines (NCT05020678).
- 7 of 10 patients at the two highest dose levels achieved complete response; objective response 80 percent at those levels.
- No dose-limiting toxicity. No ICANS. No graft-versus-host disease. No cytokine release syndrome above grade 2.
- The dominant severe toxicity was myelosuppression, the fingerprint of the chemotherapy, not the cells.
- 40 percent of eligible patients were dosed as outpatients after the first cycle; re-dosing after relapse deepened or restored responses in some.
How to hold those numbers
They come from company releases and conference presentations (EHA and ICML 2023, November 2022 data cut). No peer-reviewed full text was located in verification, so they are graded as presented, not as journal-verified.
They are also cancer numbers: heavily pre-treated patients, different risk tolerance, and no autoimmune efficacy information in them at all. What they genuinely support is narrower and still valuable: this product, in humans, has so far not produced the toxicities that confine CAR-T to specialist hospitals.
The cell-therapy explorer
Four ways to attempt an immune reset, compared on the axes that decide who can actually receive them. Select a modality to focus it. The scores are editorial grades explained on this page, not measurements.
The honest caveats
Weeks of persistence, years of remission needed
NKX019's reported half-life is up to 28 days. CAR-T cells can persist for years, and the leading account of the CAR-T reset is depth plus duration of depletion, including in synovium and lymph nodes. Whether a short-lived NK product depletes tissue B cells as deeply is simply unknown. If it does not, CAR NK risks becoming a chemotherapy-conditioned, freezer-delivered rituximab. Re-dosing is the built-in answer, and every re-dose is another lymphodepletion.
The fine print on "outpatient"
Outpatient describes the cell infusion and the 2 hours of observation after it. Before that comes fludarabine and cyclophosphamide: cytopenias, infection risk, cumulative-dose concerns. The pooled safety oral at ACR is specifically about haematologic outcomes after that chemotherapy, which is where the sponsor expects the toxicity to live. Add host rejection: donor cells are foreign, the trials assay anti-NKX019 and anti-HLA antibodies, and a sensitised patient may clear later doses faster, blunting re-dosing exactly when it is needed.
What to watch
ACR 2026, 6 to 11 Nov
Read the pooled safety oral (2938) for any grade 3 or higher CRS, ICANS or GvHD (any would break the thesis), cytopenia duration, B-cell depletion depth, and persistence curves. Slides go to Nkarta's Publications page after the meeting.
The RA cohort clock
Ntrust-2 enrolment pace through 2027 is the tell for whether RA is a priority or a label expansion. Registry updates are public; primary completion is estimated October 2028.
The tier around it
COMPARE's randomised Phase 2 (CAR-T versus rituximab) validates or deflates the whole depletion tier. BMS's allogeneic CAR-T (NCT07115745) is the off-the-shelf head-to-head. REGULATE-RA's peer-reviewed data set the tolerance side of the fork.
What this dossier does not claim
- NKX019 has shown no efficacy in any autoimmune disease in published data as of this writing. The November presentations will be the first look, in myositis and scleroderma only.
- It has shown nothing in rheumatoid arthritis. The RA cohort opened in 2026 and reads out years from now.
- "Immune reset" is Nkarta's hypothesis for NKX019, labelled a company claim wherever it appears here. For this product it is a rationale, not a result.
- No price for NKX019 has been published. Any figure in circulation is unverified.
- Nothing here is medical advice, and nothing here speaks to any individual's treatment. That conversation belongs with a rheumatologist.