Immunotherapy has transformed the treatment of many cancers, but one of its biggest hurdles remains on-target toxicity—the unintended destruction of healthy cells that share surface markers with tumors. This challenge is especially acute in acute myeloid leukemia (AML), where therapies directed at the myeloid antigen CD33 can effectively eliminate malignant cells. Unfortunately, CD33 is also present on normal blood-forming hematopoietic stem and progenitor cells (HSPCs), meaning patients treated with CD33-directed drugs risk prolonged myelosuppression, hypogammaglobulinemia, and life-threatening infections or bleeding. The problem isn’t a lack of effective therapies—several CD33-targeted approaches exist, including experimental CD33-directed CAR T cells and the FDA-approved antibody-drug conjugate gemtuzumab ozogamicin (GO, developed in the Bernstein lab here at Fred Hutch). Rather, the barrier is that these powerful drugs cannot discriminate between leukemia and the healthy stem cells required for lifelong blood regeneration.
A promising solution is to engineer donor HSPCs so that they no longer express CD33, rendering them resistant to CD33-targeted drugs while preserving their ability to regenerate the entire blood system. Prior work using CRISPR/Cas9 showed that ablating CD33 did not impair hematopoietic repopulation and conferred protection from GO and CD33-CAR T cells. However, traditional Cas9-based editing introduces double-stranded breaks in the DNA (DSBs), which can have negative effects such as triggering p53 activation, generating large unintended insertions or deletions (indels), and cause chromosomal translocations where a piece of one chromosome breaks off and attaches to a different chromosome. These risks are particularly concerning for HSPCs, where genomic stability is essential for long-term engraftment of cells which retain the ability to generate all the blood cells we need to be healthy.
A recent Nature Communications study from the Kiem lab explores a safer alternative to Cas9 nucleases: base editors. Unlike nucleases, which create double-strand breaks (DSBs), base editors can precisely rewrite single nucleotides without cutting DNA. The work, led by staff scientist Dr. Olivier Humbert and Dr. Hans-Peter Kiem in collaboration with Dr. Siddhartha Mukherjee’s group at Columbia (yes, the Emperor of All Maladies author), used an adenine base editor (ABE8e) to mimic a naturally occurring CD33 single-nucleotide polymorphism. This variant abolishes full-length surface expression of CD33, effectively “erasing” the protein without indels. By leveraging a change already tolerated in humans, the team demonstrated a strategy that minimizes safety risks while preserving therapeutic potential.
The researchers showed that base-edited human HSPCs retained normal functionality and could make all the necessary blood cells both in mouse models and nonhuman primate transplantation studies, the latter being the gold standard for preclinical validation of HSPC therapies. Importantly, the edited cells resisted killing by GO in vivo, causing the CD33-negative cells to become selectively enriched. This enrichment demonstrates a powerful proof-of-concept: engineered resistance can serve not only as a safety mechanism but also to select for gene-edited cells after transplantation, addressing one of the biggest hurdles in the HSPC gene therapy field—achieving and maintaining a high proportion of modified cells.
To test the potential of multiplex editing, the Kiem lab paired CD33 ablation with a second edit in the γ-globin promoters that reactivates fetal hemoglobin (HbF), a strategy of high interest for treating hemoglobinopathies such as sickle cell anemia. Using ABE8e, they achieved >70% dual editing efficiency, and the resulting cells showed durable engraftment in primates with persistent HbF upregulation. This finding illustrates how CD33 editing could function as a broadly applicable “selector” locus to enrich therapeutic edits at other genes, all while maintaining safety and hematopoietic integrity.
Together, these results highlight the potential of base editing as a safer, more versatile alternative to nuclease-based gene editing for HSPC therapies. For AML, CD33 editing may enable the use of highly effective immunotherapies without the dangerous collateral damage to healthy stem cells. More broadly, the work suggests a generalizable strategy for in vivo enrichment of gene-edited cells, opening new doors for safer conditioning regimens and durable gene therapies for both cancer and inherited blood disorders.
Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium Members Drs. Roland Walter and Hans-Peter Kiem contributed to this research.
The spotlighted research was funded by the National Institutes of Health, the Helen Hay Whitney Foundation, and the Government of India Department of Biotechnology.
Borot F, Humbert O, Ehmsen JT, Fields E, Kohli S, Radtke S, Swing K, Pande D, Enstrom MR, Laszlo GS, Mayuranathan T, Ali AM, Weiss MJ, Yen JS, Newby GA, Walter RB, Liu DR, Mukherjee S, Kiem H. 2025. Multiplex base editing to protect from CD33 directed drugs for immune and gene therapy. Nature Communications. https://doi.org/10.1038/s41467-025-59713-2