Cooperative activation of NF-κB by inv(3)/t(3;3) and NRAS mutations drives myeloid leukemogenesis

Abstract In acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS), inversion/translocation of chromosome 3 (inv(3)/t(3;3)) is a recurrent poor prognostic karyotype (J Clin Oncol. 2010; 24: 3890-3898). Previous studies have shown that GATA2 distal hematopoietic enhancer located on chromosome 3q21 region is brought into close proximity to EVI1 oncogene located on chromosome 3q26 region in inv(3)/t(3;3)-positive leukemia cells, which activates EVI1 oncogene expression and leukemic transformation (Cell. 2014; 157: 369-381). Consistent with this notion, 3q21q26 transgenic mice are reported to develop full blown AML in vivo (Cancer Cell. 2014; 25: 415–427). Whole exome sequencing analysis of AML or MDS patients with inv(3)/t(3;3) showed that somatic mutations of signaling factors coexisted in 98% of cases, with NRAS mutations being the most common among the mutated genes encoding signaling factors which presented in 29% of these cases (Blood. 2015; 125: 133-139). Based on these observations, we sought to explore functional cooperativity between inv(3)/t(3;3) and NRAS mutations in myeloid transformation, its molecular basis and potential therapeutic target. To this end, we crossed the above mentioned 3q21q26 mice and previously reported Mx-Cre+NrasG12D mice, which drive hematopoietic tissue-specific expression of the most frequently observed NRAS mutant allele in human MDS/AML from the endogenous locus leading to the development of myeloproliferative tumors in vivo, to generate 3q21q26/NrasG12D mice (Blood. 2011; 117: 2022-2032). At 2-4 weeks post-recombination, 3q21q26/NrasG12D mice showed an increase in white blood cells (mean at 2-4 weeks 123.6×103/μL, p < 0.01; WT 18.1×103/μL, NrasG12D23.2×103/μL, 3q21q26-EVI1 21.4×103/μL) and developed anemia (mean at 2-4 weeks 9.94 g/dL, p < 0.0001; WT 15.1 g/dL, NRASG12D15.3 g/dL, 3q21q26-EVI1 15.1 g/dL). Peripheral blood (PB) smears showed increased immature blast-like cells in 3q21q26/NrasG12D mice. Flow cytometric analysis of PB revealed significant expansion of CD11b+ Gr1+ myeloid population and cKit+ immature fraction in 3q21q26/NrasG12D mice compared to WT or single mutant mice. 3q21q26/NrasG12D mice developed marked splenomegaly. Pathological examination exhibited robust infiltration of blast-like myeloid cells in bone marrow (BM), liver and lung in 3q21q26/NrasG12D mice. Importantly, recipient mice of 3q21q26/NrasG12D BM cells also developed lethal AML within 4 weeks from transplantation in vivo. Furthermore, 3q21q26/NrasG12D mice had a significantly worse overall survival (OS) than wild type and single mutant mice (median survival 55 days, p<0.01; WT undefined, NrasG12D254 days, 3q21q26-EVI1 317.5 days). These results clearly indicated that inv(3)/t(3;3) and NrasG12D allele functionally cooperate to develop murine myeloid leukemia in vivo. Flow cytometry of BM showed a significant increase of Lin- Sca1+ cKit+ cells (LSKs) and recipients of 3q21q26/NrasG12D BM LSKs recapitulated similar AML-like disease as primary mice in vivo, indicating that these mice developed AML with leukemic stem cell (LSC) fraction enriched in LSKs. To elucidate molecular basis which characterizes LSCs of 3q21q26/NrasG12D AML, we explored transcriptome of LSK fraction. RNA-seq of LSKs suggested activated nuclear factor-kappa B (NF-κB) pathway in 3q21q26/NrasG12D mice. Phospho-flowcytometry confirmed increased phosphorylation of p65 and IκB in 3q21q26/NrasG12D BM LSKs. We next investigated whether NF-κB pathway is a potential therapeutic target of 3q21q26/NrasG12D AML. In line with our hypothesis, an inhibitor of NF-κB kinase (IKK) inhibitor, BMS-345531, which blocks the NF-κB pathway, suppressed cell growth and colony formation of 3q21q26/NrasG12D BM cells. Importantly, treatment of 3q21q26/NrasG12D BM recipients with BMS-345531 (administration of the drug starting 2 weeks post-transplant for 4 weeks) significantly improved blood cell counts (mean at 5 weeks after BM transplantation, WBC; BMS-345531 9.6×103/μL, vehicle 92.3×103/μL, p<0.01, Hb; 14.8 g/dL, 13.0 g/dL, p<0.05), decreased leukemic burden in BM and led to a trend towards improved OS compared to vehicle-treated arm. Collectively, these data suggest that inv(3)/t(3;3) and NRAS mutations functionally cooperative to drive myeloid leukemogenesis. Notably, cooperatively activated NF-κB pathway by inv(3)/t(3;3) and NRAS mutations is a potential therapeutic target of AML with these high-risk genetic lesions.