Juvenile myelomonocytic leukemia (JMML) is an aggressive overproduction of cells in the myeloid lineage, characterized by hypersensitivity to granulocyte macrophage colony-stimulating factor (GM-CSF). There is a clear need for improved chemotherapies, as the current treatments are ineffective. The most commonly mutated gene in JMML patients is PTPN11, which encodes the protein tyrosine phosphatase Shp2. One of the downstream targets of Shp2 is phosphoinositide 3-kinase (PI3K), and we have previously focused on the hematopoietic-specific catalytic subunit p110δ (PI3Kδ). The PI3Kδ inhibitor idelalisib is FDA-approved for patients with B cell malignancies. However, the effectiveness of PI3Kδ inhibition in JMML and other myeloid malignancies has not been studied. Therefore, following our previous work in vitro which demonstrated reduced proliferation of GOF Shp2-expressing murine cells and primary JMML cells (Goodwin et. al, Blood 2014), we assessed the effect of PI3Kδ inhibition on GOF Shp2-expressing mice in vivo as the next step in exploring PI3Kδ inhibition as a potential treatment in JMML.

We treated Shp2E76K/+;LysMcre+ mice between 12 and 20 weeks of age with 30mg/kg of the PI3Kδ inhibitor GS-9820 (hereafter PI3Kδ inhibitor) or vehicle BID via oral gavage for 21 days. First, we evaluated mice directly after completing 21 days of treatment (16 hours following final vehicle or PI3Kδ inhibitor dose). The PI3Kδ inhibitor-treated mice had significantly reduced spleen-to-body weight ratio in comparison to the vehicle-treated mice. We measured the colony-forming ability of bone marrow low-density mononuclear cells (LDMNCs) in response to GM-CSF and found only a slight reduction in drug-treated mice, indicating that pharmacologic inhibition of PI3Kδ does not permanently correct GM-CSF hypersensitivity. To investigate how PI3Kδ inhibitor treatment induced the functional effect of reduced spleen size, we phenotypically analyzed the bone marrow and peripheral blood for hematopoietic progenitor cells (lineage-Sca1+cKit+, LSK) and for mature myeloid cells (Gr1+Mac1+). The frequency of LSK cells was significantly reduced in the bone marrow of PI3Kδ inhibitor-treated mice. There was no difference in the frequency of Gr1+Mac1+ in the bone marrow compartment, but there was a significant increase in the peripheral blood of PI3Kδ inhibitor-treated mice. Furthermore, the Mac1+ mean fluorescence intensity in the peripheral blood was significantly higher in PI3Kδ inhibitor-treated mice. Collectively, these findings suggest that PI3Kδ inhibition induced stem/progenitor terminal differentiation and reduced the self-renewal and hyperproliferation of immature myeloid cells.

A second cohort of treated mice was followed for overall survival. Survival of the PI3Kδ inhibitor-treated mice was significantly prolonged compared to their vehicle-treated counterparts. Serial peripheral blood WBC counts were similar between the two groups until approximately 20 weeks and 34 weeks after the start of treatment, when WBC counts trended upward in the vehicle-treated and drug-treated animals, respectively. These time points coincided with a sharp decrease in the number of surviving mice in the corresponding group. When moribund, mice were euthanized to assess the cause of death. The PI3Kδ inhibitor-treated animals had no difference in splenomegaly and leukocytosis compared to the vehicle-treated animals at the time of euthanasia, and composition of the bone marrow and spleen in moribund animals was not significantly different, suggesting that mice of the two groups ultimately succumbed to the same disease. Taken together, these findings suggest that time-limited PI3Kδ inhibition delays onset of mutant Shp2-induced leukemia, but does not alter the ultimate course of disease. These findings imply that continued PI3Kδ inhibition may be needed for optimal treatment.

Our results demonstrate that in vivo administration of a hematopoietic-specific PI3Kδ inhibitor increases the survival of GOF Shp2-expressing mice. The more limited side effect profile of PI3Kδ inhibition compared to pan-PI3K inhibition may provide a therapeutic window to simultaneously target both the Ras-Erk and PI3K-Akt pathways in the treatment of JMML. This work supports the rationale of using PI3Kδ inhibitors in the treatment of patients with JMML and possibly other myeloid cell malignancies.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.