Tyrosine Phosphorylation of CARM1 Promotes Its Enzymatic Activity and Alters Its Target Specificity in Myeloid Neoplasms with Hyperactivated JAK2

Introduction Coactivator-associated arginine methyltransferase 1 (CARM1) catalyzes the asymmetric dimethylation of arginine residues in histone and non-histone substrates, including transcription factors, RNA splicing factors, and the SWI/SNF chromatin remodeling complex. CARM1 has emerged as an important target in acute myeloid leukemia.

We have previously shown that Janus kinase 2 (JAK2) affects the activity of epigenetic modifiers. A V617F mutation in JAK2 genes leads to its hyperactivation. This mutation is a common hallmark of the BCR/ABL1-negative myeloproliferative neoplasms, and also found in patients with acute myeloid leukemia and myelodysplastic syndrome. Here, we find that hyperactivated JAK2, mediated by biallelic V617F mutation, triggers CARM1 phosphorylation, increasing its methyltransferase activity and altering its target specificity.

Methods We generated two phospho-tyrosine specific antibodies against CARM1 tyrosine-149 (Y139) and -334 (Y334), and two asymmetric dimethylation-arginine specific antibodies against RUNX1 arginine-223 (R223) and -319 (R319). We also generated isogenic HEL cells carrying homozygous CARM1 non-phosphorylatable mutation (Y149F or Y334F single, or Y149F/Y334F double mutation), using the CRISPR/Cas9 nuclease system.

Results We first performed cell-free in vitro kinase assays using recombinant CARM1 protein and JAK2 kinase, and mass spectrometry analysis, which identified Y139 and Y334 as the sites of JAK2 phosphorylation in CARM1. We next compared the levels of CARM1-Y149 and -Y334 phosphorylation in 15 myeloid leukemia cell lines and normal human CD34+ cord blood cells. HEL and UKE-1 cells, that harbor bi-allelic JAK2-V617F mutation, showed the highest level of phosphorylated CARM1-Y149 and -Y334, while SET2 cells with mono-allelic JAK2-V617F mutation and the other cell lines showed abundant CARM1 protein but a lower relative amount of phosphorylated CARM1 protein. We found that the JAK2 activation, through transphosphorylation by JAK2 and TYK2, enhances tyrosine kinase activity against CARM1 using HEL cells with knockout of either JAK1 or TYK2.

Having previously shown that CARM1 methylates RUNX1 at R223, in an immunoprecipitation assay, WT-CARM1 was able to pull down RUNX1, while the non-phosphorylatable CARM1 mutants were not. Using mass spectrometry and an in vitro methylation assay, we confirmed R223 and R319 as CARM1 target sites. We examined non-phosphorylatable CARM1 mutation knock-in HEL cells and found that mutation of either site reduces the asymmetric dimethylation of RUNX1 as well as known substrates (BAF155 and PABP1).

We observed that CARM1 Y149F/Y334F mutant-expressing HEL cells reduced cell proliferation with G2/M cell cycle arrest and an increase in apoptosis. To identify the transcription networks regulated by CARM1 phosphorylation, we performed RNA-seq on WT-CARM1 and double CARM1 Y149F/Y334F mutant-expressing HEL cells. Gene set enrichment analysis identified that gene sets associated with G2/M cell cycle progression and anti-apoptosis were significantly downregulated in Y149F/Y334F mutation knock-in HEL cells. We next performed ChIP-seq analyses using antibodies against total RUNX1 and against asymmetrically dimethylated R319-RUNX1 in HEL cells expressing knock-in of CARM1 non-phosphorylatable mutation. CARM1 non-phosphorylatable mutation decreased the signals of dimethylated RUNX1-R319 within 5 kb of the transcription start sites for RUNX1 target genes, with less effect on total RUNX1.

We next assessed the efficacy of the combination with CARM1 inhibitor (EPZ025654) and JAK2 inhibitor (ruxolitinib). Dual targeting inhibition of JAK2 and CARM1 showed a synergistic inhibition effect on HEL and UKE-1 cells (with abundant phosphor-CARM1), but only an additive effect on SET2 cells (with less amount of phosphor-CARM1).

Conclusion CARM1 phosphorylation mediated by hyperactivated JAK2 regulates its methyltransferase activity and is required for the maximal proliferation of myeloid neoplasms. We also demonstrated that dual targeting of JAK2 and CARM1 is more effective than monotherapy in cell lines with phosphor-CARM1. Thus, our results reveal a novel role for CARM1 phosphorylation in myeloid neoplasms.

No relevant conflicts of interest to declare.