AUTHORS
Edoardo Campodonico, MD, Saksham Garg, M.Sc.2, Marta Freschi3, Giada Zambonini, M.Sc.4, Sofia Salamanca Torres, B.Sc.5, Pietro Genovese, PhD6, Gabriele Casirati, M.D., Ph.D.7
AFFILIATIONS
1 Boston Children's Hospital, Boston, United States, Harvard Medical School, Boston, United States
2 Ludwig Maximilian University of Munich, Munich, Germany
3 Boston Children's Hospital, Boston, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States
4 Boston Children's Hospital, Boston, United States
5 Boston Children's Hospital, Boston, United States, Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, Germany
6 Boston Children's Hospital, Boston, United States, Harvard Medical School, Boston, United States
7 Boston Children's Hospital, Boston, United States, Harvard Medical School, Boston, United States
ABSTRACT
Despite recent advances, high-risk and relapsed/refractory (R/R) acute myeloid leukemia (AML) remains a major clinical unmet need. One of the central obstacles to the successful deployment of targeted immunotherapies rests in the paucity of antigens (Ag) that combine broad leukemia coverage with an acceptable safety profile. CD244, a SLAM-family receptor implicated in leukemia-initiating cell maintenance, has been found expressed on AML leukemia stem cells (LSC) in 96.7% of cases at diagnosis and 97.1% at relapse. Its concurrent expression by hematopoietic stem and progenitor cells (HSPCs) and selected mature immune populations creates a predictable toxicity liability. We hypothesized that CD244 and its expression pattern could be harnessed to eradicate LSCs and achieve pre-transplant non-genotoxic conditioning, as an “all-in-one” curative strategy for patients affected by high-risk or R/R AML. This approach would integrate remission induction and immune-based myeloablation before hematopoietic stem cell transplantation (HSCT). To this end, we sought to develop a CD244-directed CAR platform together with genome-editing strategies to control CAR fratricide effects and shield the hematopoietic graft.
We sequenced the variable regions of 2 independent CD244 mAbs and generated 2nd-gen. CARs combining scFvs with IgG4 or CD8α hinges, and an EGFRt safety switch. All constructs achieved efficient expression (>85%) and triggered Ag-dependent NFAT and NF-κB activation in dual-reporter Jurkat cells. In primary T cells, CD244 CARs selectively killed CD244+ MOLM-13, Kasumi-1, and patient-derived AML xenografts (PDX) while sparing Ag-negative K562, MM.1S, and NALM-6 cells. Across co-cultures at different effector-target ratios, elimination exceeded 99% and was accompanied by robust T cell activation, degranulation and cytokine secretion, with IgG4-hinge constructs outperforming other designs.
As CD244 is upregulated in a subset of exhausted CD8+T cells, CAR expression caused selective fratricide and generated a CD4-skewed product, with effects modulated by each binder’s affinity and epitope localization. To abrogate fratricide killing and eliminate a potential checkpoint vulnerability, we optimized a base editing (BE) strategy to achieve functional knock-out of CD244 through splice-site disruption. CD244-KO CAR-T retained expansion and Ag-specific cytotoxicity comparable to unedited products, establishing a practical strategy to preserve therapeutic-cell composition and fitness.
We next harnessed our advanced in vivo humanized models to simultaneously capture antileukemic activity and hematopoietic toxicity. NBSGW mice were humanized with CD244-KO or AAVS1-edited control human CD34+ HSPCs, engrafted with CD244+ PDX cells, and treated with both CD244 scFv-based CARs or untransduced T cells. Both CARs demonstrated potent in vivo antileukemic activity, resulting in near-complete eradication at 2 weeks post-CAR infusion. As expected, CD244 CAR expansion was accompanied by depletion of healthy hematopoietic lineages, with near-complete elimination of HSCs and early and lineage-primed progenitors. CD244 disruption in donor HSPCs conferred protection from CAR-mediated depletion, demonstrating that genetic shielding can uncouple therapeutic targeting from hematopoietic injury. To develop a function-preserving epitope-editing shielding strategy, we generated saturation-mutagenesis libraries spanning the CD244 extracellular domain. NGS sequencing identified candidate substitutions that abrogated binding by therapeutic mAbs, nominating variants for base- or prime-editing of donor HSPCs. This approach could enable long-term post-HSCT anti-leukemia maintenance, without the need for CAR-T depletion.
In conclusion, these results establish CD244 as an actionable, broadly shared AML antigen, converting a potential safety liability into a powerful therapeutic opportunity. This platform supports innovative therapeutic paradigms in which fratricide-resistant CD244 CAR-T cells both eliminate R/R or measurable residual AML and deplete recipient HSPCs, thus creating marrow space for transplantation of an allogeneic graft. By combining leukemia targeting and immune conditioning, CD244-directed therapy could enable HSCT with reduced reliance on genotoxic chemotherapy or irradiation, therefore expanding its applicability to patients in advanced disease stages or affected by comorbidities.
DISCLOSURES
Edoardo Campodonico, MD: No relevant conflicts of interest to declare.