CD33-Deleted Hematopoietic Cell Transplant Shows Potential for Hematologic Protection During Leukemia-Targeted Maintenance
Toxicity to normal blood-forming cells limits post-transplant targeted treatment
All 30 patients with high-risk blood cancers who received CD33-deleted, gene-edited hematopoietic cell transplants achieved neutrophil engraftment within 28 days. Among them, 19 received a CD33-targeted drug as maintenance therapy, with no prolonged high-grade cytopenias observed. However, severe adverse events and deaths were also reported, so these findings cannot be interpreted as eliminating all toxicity or establishing anticancer efficacy.
Acute myeloid leukemia (AML) can relapse even after allogeneic hematopoietic cell transplantation. Maintenance therapies targeting residual leukemia cells are under study, but hematologic toxicity that also damages transplanted normal cells remains a barrier. CD33 is expressed on AML cells as well as normal myeloid cells and progenitors. This explains why post-transplant use of gemtuzumab ozogamicin (GO), a CD33-targeted antibody–drug conjugate (ADC), can cause severe and prolonged cytopenias.
The researchers addressed this problem by removing CD33 from donor cells before transplantation. The resulting product, trem-cel, consists of CD34-positive hematopoietic stem and progenitor cells from human leukocyte antigen (HLA)-matched donors edited with CRISPR–Cas9. It is designed to prevent production of the CD33 protein, with the aim of protecting normal graft-derived hematopoietic cells from subsequent CD33-targeted treatment.
Engraftment in all 30 patients and a protective signal in 19 maintenance recipients
The study, published in Nature Medicine, was a multicenter, nonrandomized, open-label, first-in-human phase 1/2a trial. Twenty-nine adults with AML at high risk of relapse and one patient with myelodysplastic syndrome (MDS) received trem-cel after myeloablative conditioning. The primary endpoint was neutrophil engraftment by day 28 after transplantation, defined as an absolute neutrophil count of at least 500 per microliter for three consecutive days.
All 30 patients met this criterion, and the median time to engraftment was 10 days. However, recovery did not extend to every blood cell lineage without exception. One patient had delayed platelet recovery, and another had secondary graft failure; both received additional hematopoietic cells.
GO maintenance began approximately 60 days after transplantation in patients who met criteria including disease status and blood count recovery. Nineteen patients received 0.5–2.0 mg/m² per dose every 28 days, with a median of four doses. The recommended phase 2 dose was set at 2.0 mg/m². No dose-limiting toxicities occurred during dose escalation, and neutrophil and platelet counts were maintained during maintenance without prolonged high-grade declines. An increase in the proportion of CD33-negative myeloid cells after GO treatment also supported the potential for protection of edited cells.
Hematologic protection is distinct from overall safety and anticancer efficacy
Across the study, grade 2–4 acute graft-versus-host disease (GVHD) occurred in six patients, including two with grade 3–4 disease. Among the 19 GO maintenance recipients, 11 had GO-related adverse events of grade 3 or higher. Transplant-related deaths occurring more than 100 days after transplantation included one each from renal failure, sinusoidal obstruction syndrome/veno-occlusive disease (SOS/VOD), and sepsis. The SOS/VOD death was assessed as related to GO. Protection of hematopoietic cells therefore did not eliminate other risks, including liver toxicity.
Three GO maintenance recipients relapsed. Median follow-up among surviving patients was 7.9 months, and the study ended early for financial reasons. Its small size, varied patient characteristics, and lack of a concurrent control group prevent definitive conclusions about relapse prevention or improved survival. These findings provide early clinical evidence supporting engraftment of edited grafts and hematologic protection during subsequent targeted treatment. Long-term safety and anticancer efficacy require validation in larger studies.
The study design and results were checked against the original Nature Medicine article and the provided publicly available full text from Europe PMC.
Nature Medicine, Published online: 12 May 2026; doi:10.1038/s41591-026-04362-1In a first-in-human trial combining the transplantation of CD33-negative CRISPR-edited hematopoietic cells with the CD33-targeted antibody–drug conjugate gemtuzumab ozogamicin, all transplanted patients achieved primary engraftment, and the treatment was well tolerated.
This study suggests that removing from donor cells a target shared by normal graft cells and leukemia cells could create an opportunity for post-transplant targeted treatment. A specific research application would be to transplant edited cells into high-risk patients with CD33-positive AML/MDS who have an HLA-matched donor, then begin GO maintenance after confirming complete remission and adequate neutrophil and platelet recovery. However, engraftment in every patient does not mean relapse suppression or prolonged survival. Further development must determine whether protection of normal hematopoiesis translates into better treatment outcomes while also evaluating liver toxicity, infections, transplant complications, and the long-term safety of gene editing.