Acute myeloid leukemia (AML) is a heterogeneous myeloid malignancy characterized by mutational and clonal heterogeneity. Mutations in isocitrate dehydrogenase 1 and 2 (IDH1/2) are common, occurring in approximately 15-20% of patients, and actionable, with recently approved inhibitors for both mutations. These inhibitors lead to leukemia cell differentiation in vitro, in vivo and in patients. Healthy myeloid differentiation is governed by precise regulation of intracellular signaling, but this regulation is disrupted in AML. Given that signal transducer and activator of transcription (STAT) proteins are involved in both leukemogenesis and myeloid differentiation, we sought to determine the role of phosphorylated STATs and other signaling proteins in mIDH AML and inhibitor-induced differentiation of mIDH2 AML.

To simultaneously dissect single cell signaling, differentiation and epigenetic changes, we employed fluorescence flow cytometry and mass cytometry to study an in vitro model of mIDH2R140Q leukemia, consisting of a parental TF-1 leukemia cell line and a CRISPR/Cas9 gene-edited mIDH2R140Q-mutated cell line. In order to generate provoked signaling profiles, we stimulated both the wild-type IDH2 and mutant IDH2 cells with nine cytokines. Additionally, we treated both cell lines and primary mIDH2 AML samples ex vivo with enasidenib and measured changes in high-dimensional single cell phenotype and phospho-protein expression via mass cytometry.

mIDH2 leukemia cells displayed elevated basal pSTAT1, pSTAT3 and pNFkB-s529 expression, and concomitant low basal IkBa levels when compared to parental cells. Further, mIDH2 cells had increased PMA induced pS6 and IL1β induced pNFkB-s529 than the parental cell line, while wtIDH2 cells had higher levels of cytokine-induced pSTAT1, pSTAT3 and pSTAT5.

After prolonged enasidenib treatment (28 days), model mIDH2 cells expressed lower basal pSTAT1, pSTAT3 and pNFkB-s529 than vehicle-treated mIDH2 cells. Further, when stimulated with GM-CSF, enasidenib-treated mIDH2 leukemia cells showed increased response at pSTAT3 and pSTAT5 as compared to vehicle-treated leukemia cells. We dissected enasidenib-induced differentiation using 18 cell surface markers and visualized results using t-distributed stochastic neighbor embedding (tSNE). Presence of mIDH2 leads to baseline expression differences including higher CD90 and CD71. Following treatment with enasidenib, mIDH2 leukemia cells had increased CD45 and CD11b expression as compared to vehicle-treated controls. Additionally, prolonged treatment with enasidenib increased proliferation as shown by increased Ki67 and decreased histone hypermethylation at suppressive histone marks, H3K27 and H3K9, while 7 days of enasidenib did not result in changes to these marks.

We further explored the impact of mIDH2 in primary leukemia samples. First, we developed an ex vivo stromal co-culture system that allowed for treatment and expansion of four primary mIDH2 AML samples for 16 days. While the basal levels of signaling markers were sample dependent, consistently across all samples enasidenib-treated primary AML was more sensitive to IL-6 and GM-CSF-induced pSTAT1, pSTAT3 and pSTAT5 signaling. The enasidenib-treated AML also showed increased expression of mature myeloid markers, including CD33 and CD11c.

Here we have shown the presence of mIDH2 mutations leads to decreased STAT signaling in response to cytokine stimulation as compared to wtIDH2 AML cells. We have also demonstrated increased response of pNFkB-s529 and pS6 to cytokine stimulation in mIDH2 AML cells as compared to wtIDH2 AML cells. Moreover, in both mIDH2 AML cell line and primary samples, IDH2 inhibitor-induced differentiation restored sensitivity to cytokine responses and reduced histone hypermethylation. Future work exploring these aberrant signaling events may reveal precise connections between mutated IDH, associated epigenetic changes, and intracellular signaling, potentially uncovering synergistic therapeutic strategies for mIDH AML.


Ferrell:Incyte: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.