Oncogenic fusion proteins are known to drive the development of some pediatric cancers, including leukemias. One fusion of note, CBFA2T3-GLIS2 (CBF/GLIS), is associated with very poor survival outcomes in pediatric acute myeloid leukemia (AML) and an AML subtype, acute megakaryoblastic leukemia (AMKL). To further make matters worse, this fusion protein often goes undetected by conventional diagnostics (thus referred to as a cryptic fusion), meaning a clinical diagnosis of CBF/GLIS AMKL is reliant on next generation sequencing or expression of a specific immunophenotype termed RAM (high CD56, dim/negative CD45 and CD38 expression). Studying this disease subtype in in vitro and in vivo models is essential for better understanding disease progression and the ability of the fusion to drive the malignant transformation of hematopoietic stem cells. Members of Dr. Soheil Meshinchi’s Lab in Fred Hutch’s Clinical Research Division previously worked with human cord blood hematopoietic stem cells engineered to express the CBF/GLIS fusion which enabled the partial development of an AMKL phenotype. Their recent study, published in Journal of Clinical Investigation and led by Dr. Quy Le, Dr. Meshinchi and colleagues, aimed to take these results a step further. “Patients who harbor CBF/GLIS oncogenic fusion exhibit aggressive disease and eventually succumb to leukemia despite high intensity chemotherapy and myeloablative stem cell transplant,” explained Dr. Le. “Given limited available model systems and the difficulty of culturing primary AML cells ex vivo for study, there has been slow progress in the therapeutic development for this vulnerable group of patients. To overcome this hurdle, our goal was to generate a robust model system that recapitulates this disease to better understand the underlying biology and to identify targets for immunotherapeutic intervention.”
After confirming that expressing the CBF/GLIS fusion in cord blood hematopoietic stem cells can induce both morphology and an immunophenotype similar to AMKL in xenografts, the authors focused on developing a long-term co-culture model with endothelial cells. They postulated that endothelial cells that normally line blood vessels would provide the necessary microenvironment in co-culture to support the fusion cord blood stem cells. Under co-culture conditions the engineered cord blood cells proliferated, evolved into AML, and exhibited the RAM phenotype distinct to the AMKL disease subtype. Further investigation by the authors of transcriptomic profiles of co-cultured CBF/GLIS transformed cells determined that their gene expression patterns clustered with fusion-positive patient samples, a finding specific to cells cultured with endothelial cells and not control cells, confirming their hypothesis that endothelial cells support the necessary microenvironment for malignant transformation. Summarizing their model system, Dr. Le said “we show that forced expression of CBF/GLIS oncogenic fusion in cord blood stem/progenitor cells promotes leukemic transformation and further implicates the role of the endothelial microenvironment during transformation. This model system recapitulates primary leukemia in every aspect we looked at (i.e., immunophenotype, morphology, transcriptome, pathology, etc.), providing a useful, renewable source of AML cells to study their biology, and more importantly, for discovery of therapeutic targets.”
Having developed a successful long-term model, the authors next focused on identifying potential CAR T cell targets. Transcriptomic profiling uncovered a subset of candidate genes that were upregulated in both their cultured cells and primary CBF/GLIS fusion driven AML samples. The authors focused their efforts on a single target, FOLR1, a therapeutic target that has had previous success in solid malignancies and was not expressed in the course of normal hematopoiesis. The authors developed a chimeric antigen receptor or CAR-T cell targeting FOLR1 and tested the efficacy in vitro in FOLR1 positive and negative cells, noting cytolytic activity only in FOLR1 positive cells, and in vivo in FOLR1 positive xenograft models resulting in depletion of tumor cells and extended survival compared to control xenografts. Moving to patient samples, co-culture of FOLR1 CAR T cells with patient bone marrow cells resulted in sensitivity to CAR T mediated killing, a finding that was consistent in in vivo xenografts where AML engraftment was undetectable after CAR T cell therapy and led to prolonged survival when compared to control models. Lastly, the authors tested the impact of the FOLR1 CAR T cells on normal hematopoietic stem cells and observed no impact of the CAR T cells on self-renewal or multilineage differentiation, suggesting that FOLR1 CAR T cells will not impact normal stem cell function while targeting malignant cells. “We demonstrate the utility of this model to identify targets and validate the therapeutic potential of one of these targets (FOLR1) with CAR T cells. The FOLR1-directed CAR T cells exhibit potent, target-dependent elimination of CBF/GLIS AML cells, providing a promising approach to treat CBF/GLIS-positive patients,” summarized Dr. Le.
Taken together, the findings from the present study outline a robust long-term culture model that aids understanding of oncogenic fusion driven pediatric cancers and next steps for the group are focused on understanding the efficacy of the CAR T therapy in patients. “Although we show the requirement of the endothelial niche to support leukemic transformation, the underlying mechanisms (i.e., cell-cell interactions, soluble factors) will need to be further investigated” stated Dr. Le. Further, high target directed potency and lack of hematopoietic toxicity of the FOLR1 CAR T provided all requisite data to move this CAR T construct towards clinical development through the Fred Hutch Immunotherapy Integrated Research Center infrastructure and the Therapeutic Products (TPP) laboratory. “The clinical vector has been generated and is undergoing process development/engineering at the Hutch TPP with an Investigational New Drug application to follow. We expect the first patient to be enrolled on this infant AML protocol in mid-2023,” noted Dr. Meshinchi.
This work was supported by funding from Project StElla, the National Cancer Institute TARGET and Target 642 Pediatric AML (TpAML) initiatives, Leukemia and Lymphoma Society SCOR and TRP, a COG 643 Translation Pilot Studies for Hematopoietic Malignancies, the 644 Takaeda Science Foundation, The Nakatomi Foundation, and the Kitasato University School of 645 Medicine Alumi Association.
Fred Hutch/University of Washington/Seattle Children's Cancer Consortium members Dr. Brandon Hadland, Dr. Jay Sarthy, Dr. Scott Furlan, Dr. Shivani Srivastava, Dr. Filippo Milano, Dr. James Olson, Dr. Stanley Riddell, Dr. Katherine Tarlock, Dr. Irwin Bernstein, Dr. Keith Loeb, and Dr. Soheil Meshinchi contributed to this research.
Le Q, Hadland B, Smith JL, Leonti A, Huang BJ, Ries R, Hylkema TA, Castro S, Tang TT, McKay CN, Perkins L, Pardo L, Sarthy J, Beckman AK, Williams R, Idemmili R, Furlan S, Ishida T, Call L, Srivastava S, Loeb AM, Milano F, Imren S, Morris SM, Pakiam F, Olson JM, Loken MR, Eidenschink Brodersen L, Riddell SR, Tarlock K, Bernstein ID, Loeb KR, Meshinchi S. CBFA2T3-GLIS2 model of pediatric acute megakaryoblastic leukemia identifies FOLR1 as a CAR T cell target. J Clin Invest. 2022 Sep 22:e157101. doi: 10.1172/JCI157101. Epub ahead of print. PMID: 36136600.