-Author name in bold denotes the presenting author
-Asterisk * with author name denotes a Non-ASH member
Clinically Relevant Abstract denotes an abstract that is clinically relevant.

PhD Trainee denotes that this is a recommended PHD Trainee Session.

Ticketed Session denotes that this is a ticketed session.

4174 Identifying Targeted Therapies for CBFA2T3-GLIS2 Acute Myeloid Leukemia

Program: Oral and Poster Abstracts
Session: 604. Molecular Pharmacology and Drug Resistance: Myeloid Neoplasms: Poster III
Hematology Disease Topics & Pathways:
Research, Fundamental Science, Acute Myeloid Malignancies, AML, Combination therapy, pediatric, Diseases, neonatal, Therapies, Myeloid Malignancies, Study Population, Human
Monday, December 11, 2023, 6:00 PM-8:00 PM

Fanny Gonzales, MD, PhD1,2, Shan Lin, PhD1,2*, Delan Khalid1,2*, Jana M Ellegast, MD1,2, Gabriela Alexe, PhD1,3* and Kimberly Stegmaier, MD4,5

1The Broad Institute of MIT and Harvard, Cambridge, MA
2Department of Pediatric Oncology, Dana-Farber Cancer Institute and Boston Children’s Hospital, Boston, MA
3Pediatric Oncology, Dana-Farber Cancer Institute, Boston
4Department of Pediatric Oncology, Dana-Farber Cancer Institute, Boston, MA
5The Broad Institute of MIT and Harvard, Cambridge

CBFA2T3-GLIS2 fusion positive pediatric acute myeloid leukemia (AML) remains one of the worst prognostic AML subgroups. Children with this subtype of AML frequently experience upfront chemoresistance with a high cumulative incidence of relapse (42-83%) and a dismal overall survival (~20%) even with intensive therapy. New treatment approaches are urgently needed.

By interrogating the Broad Institute’s Cancer Dependency Map (DepMap), a data set composed of genome-scale CRISPR-Cas9 screens in over 1000 cancer cell lines, we discovered that the CBFA2T3-GLIS2 AML model M-07e was highly dependent on the kinase JAK2. Indeed, M-07e was one of the most JAK2 dependent models in the data set. Using the same genome-scale library, we next performed CRISPR-Cas9 screens in WSU-AML and CMS, two additional AML cell lines harboring this fusion, and revealed JAK2 as a strong dependency shared by the 3 cell lines.

Using a doxycycline-inducible knockout (KO) system, we validated JAK2 KO by western blot in three AML cell lines and cells from a patient-derived xenograft (PDX) harboring the fusion. Using two different sgRNAs targeting JAK2 (sgJAK2) and a non-targeting (sgNT) control, we validated JAK2 dependency in all 4 AML models in competitive growth assays. We also observed impaired proliferation and induced apoptosis upon JAK2 KO.

To evaluate a role for JAK2 dependency in vivo, we intravenously injected NSGS mice with WSU-AML either transduced with doxycycline-inducible sgJAK2 or sgNT. After confirmation of engraftment by bone marrow evaluation, doxycycline chow was initiated. After 2 weeks of doxycycline treatment, we observed a significantly decreased tumor burden in mice with JAK2 KO AML compared to sgNT as measured by a decrement in human CD45% cells: peripheral blood (2.8% versus 12.9%), bone marrow (37.4% versus 56.3%,) and spleen (19% versus 79.5%).

We next assessed chemical inhibitors of JAK2. Both the type I JAK2 inhibitor ruxolitinib and the type II JAK2 inhibitor CHZ868 exhibited strong activity in vitro in CBFA2T3-GLIS2 cell lines and PDX cells, many resistant to cytotoxic chemotherapy. Because ruxolitinib is FDA-approved for other indications, we evaluated this drug in a cell line and PDX model of CBFA2T3-GLIS2 AML in vivo. As previously reported in myeloproliferative neoplasms, JAK2 target inhibition with ruxolitinib treatment was incomplete, and accordingly, in vivo response was modest with leukemia ultimately progressing. Evaluation of CHZ868 in vivo is ongoing.

Because single agent targeted therapy is typically insufficient for durable response in the treatment of most cancers, we utilized CRISPR-Cas9 anchor screening to identify candidate resistance mechanisms and potential synergistic combinations with JAK2 inhibitors in CMS and WSU-AML cells. We compared the changes in abundance of sgRNAs in the cells either treated with ruxolitinib (IC50) or vehicle for 14 days. The correlation of hits between the two cell lines was strong, and we found that sgRNAs targeting multiple negative regulators of the MAPK pathway were significantly enriched in the ruxolitinib-treated arm. In parallel, we generated resistant CMS sublines by chronic exposure to ruxolitinib. Ruxolitinib-resistant CMS cells were found to have a pathogenic NRAS missense mutation and an associated increase in levels of phosphorylated MEK. Accordingly, they were responsive to the MEK inhibitor trametinib. Both approaches converge on activation of the MAPK pathway as a resistance mechanism to JAK2 inhibitors in CBFA2T3-GLIS2 AML.

We next hypothesized that targeting the MAPK pathway could overcome resistance to JAK2 inhibitors in this disease context. Indeed, treatment with ruxolitinib and the MEK inhibitor trametinib exhibited a synergistic effect when used in combination in cell lines and PDX cells harboring the fusion in vitro, with in vivo testing as a next step.

In summary, we validated JAK2 as a strong dependency in CBFA2T3-GLIS2 fusion positive pediatric AML. We demonstrated that JAK2 KO or small molecule inhibitors impaired cell viability and induced apoptosis in vitro, and JAK2 KO was highly efficacious in decreasing leukemia burden in vivo. Activation of the MAPK pathway renders resistance to JAK2 inhibitors and the combination of a JAK2 with MEK inhibitor is highly synergistic.

Disclosures: Stegmaier: Kronos Bio: Research Funding; Novartis: Research Funding; Auron Therapeutics: Current holder of stock options in a privately-held company.

*signifies non-member of ASH