Interferon-γ Induces Distinct mRNA and Protein Profiles in Acute and Chronic Myeloid Leukemia

Leukemia cells promote tumor survival and progression by avoiding host immune surveillance. A key pathway involves indoleamine 2,3-dioxygenase-1 (IDO1), an interferon (IFN)-γ-inducible molecule mediating feedback inhibition and hindering anti-tumor immunity. We have previously shown thattreatment of primary acute myeloid leukemia (AML) cells with IFN-γ in vitro induces functional IDO1 in 50% of unselected cases ('IFN-γ responders'), an effect which was abrogated by STAT3 inhibitors. With a median follow-up of 8 years, these IFN-γ responders experienced a shorter event-free survival and overall survival. The current study was undertaken to dissect the intracellular signaling pathways that are activated in leukemia cells in response to IFN-γ.

We employed an alpha version of the Vantage 3D™ DNA:RNA:Protein Heme Assay (NanoString Technologies Inc., Seattle, USA) to profile DNA mutations, RNA, protein and phosphoprotein simultaneously. The assay profiles 180 mRNA species involved in hematology/oncology pathways, 38 total and phosphorylated proteins (including JAK-STAT, Src, BTK, FLT3, PI3K-Akt, NF-kB), and 124 single nucleotide variants (SNVs) relevant to hematological malignancies (IDH, FLT3, KRAS, NRAS, DNMT3, JAK3 and others). K562 and TCC-S cells (chronic myeloid leukemia [CML] in blast crisis), KG1a cells (leukemia stem cell [LSC]-like AML) and Kasumi-1 cells [AML with t(8;21)] were stimulated with 100 ng/mL IFN-γ then harvested and lysed after 30m, 1h, 2h, 12h and 24h for processing in parallel workflows (protein/mRNA and DNA SNV assays) prior to overnight hybridization with NanoString reporters. Pooled RNA/DNA/protein samples were analyzed using the nCounter® FLEX platform and the nSolver® software (alpha version 4.0). CCRF-CEM and HUT78 cell lines, with known DNA variants and expression of protein targets represented in the 3D Heme panel, were used as controls.

K562, TCC-S and Kasumi-1 cells, but not KG1a cells, up-regulated IDO1 mRNA after IFN-γ challenge. IFN-γ-sensitive CML cells (K562 and TCC-S) expressed high levels of mRNA transcripts for JAK-STAT intermediates and genes implicated in apoptosis, DNA damage and cytokine signaling (Fig. 1). IFN-γ-sensitive t(8;21) AML cells displayed a distinct transcriptomic profile with an up-regulated subset of IFN-inducible genes, including IRF1 , STAT1, BCL3, SOCS1 , PD-L1 , ISG15 , OSM , HLA class I and genes involved in antigen processing ( CIITA ). In contrast, the transcriptome of IFN-γ-insensitive LSC-like KG1a cells consisted of high levels of genes implicated in epigenetics and transcription, NF-κB, MYC, B-cell receptor and T-cell receptor signaling, and low expression levels of genes mediating JAK-STAT signaling, apoptosis, DNA damage and cytokine signaling. Overall phosphorylation of intracellular proteins was largely unchanged after IFN-γ stimulation of KG1a cells. In contrast, IFN-γ activation of Kasumi-1 AML cells was associated with increased phosphorylation of Akt(S473), STAT3(Y705) and proline-rich Akt substrate of 40kD (PRAS40) (T246). Similar phosphorylation changes were apparent in TCC-S CML cells, although with lower magnitude. Interestingly, HOXA9 transcription factor, which is commonly dysregulated in acute leukemias, was highly expressed, both as mRNA and protein (r²=0.912; p<0.001), in LSC-like KG1a cells, but not in t(8;21) AML and CML cell lines. SNV analysis identified all expected gene mutations in the control cell lines and TP53(R248Q) in Kasumi-1 cells.

The use of artificial neural network algorithms and a public collection of AML transcriptomic data (641 patients; E-MTAB-3444) allowed us to identify enriched functional-related gene groups ('immune response', 'antigen processing and presentation', 'MAPK signaling pathway'; DAVID v6.8), the expression of which was predicted by the 3 most up-regulated transcripts in our IFN-γ-sensitive leukemia cells ( IRF1, STAT1 and BCL3 ).

In conclusion, the combined analysis of SNVs, mRNA and protein expression on the nCounter® platform is a powerful tool to characterize the molecular landscape of leukemias. Our analysis of IFN-γ-stimulated leukemia cells has identified molecular pathways that could be targeted in patients with 'inflamed' tumors, in which microenvironmental production of IFN-γ is heightened due to ongoing T-cell responses.

Grant support: Roger Counter Foundation, UK; Qatar National Research Fund (#NPRP8-2297-3-494).

View largeDownload slide

View largeDownload slide

Close modal