Abstract

MLL-rearranged leukemias are found in 5-10% of adult leukemias and over 70% of infant leukemias and are associated with intermediate to poor prognosis. We have recently shown that JMJD1C, a Jumonji domain containing lysine demethylase (KDM), is important for leukemia stem cell (LSC) function in MLL-AF9 and HOXA9 leukemias but dispensable for normal hematopoietic stem cell function, therefore a potential therapeutic target for acute myeloid leukemia (AML).

To determine which domains within JMJD1C are essential for its function in AML, we performed a CRISPR/Cas9 negative selection screen against its coding sequence using guide RNAs (gRNA) tiled at ~ 30bp intervals. Specifically, gRNAs were cloned into a lenti-viral vector with a TdTomato marker gene, transduced into a clonal Cas9 expressing mouse MLL-AF9 leukemia and percentage of gRNA positive cells was analyzed by flow cytometry in a time course. JMJD1C Jumonji and Zinc finger domain (ZFD) and surrounding regions exhibited the most depletion at day 14 compare to day 2, comparable or higher than that of the positive control. To investigate the effect of mutating Jumonji and ZFD on MLL-AF9 leukemogenesis, we subject gRNA transduced MLL-AF9 Cas9 cells to colony forming cell assay. Mutating Jumonji or ZFD impaired colonegenic potential of MLL-AF9 Cas9 cells as evident by a significant reduction in the number of colonies. The resulting colonies appeared smaller and diffused, indicative of cell differentiation. Consistently, we observed impaired proliferation, increased apoptosis and increased expression of the myeloid differentiation marker Mac1 upon mutating these domains. Moreover, mutating Jumonji and ZFD prolonged survival of recipient mice compare to control in transplantation experiment. Finally, we performed chromatin-immunoprecipitation followed by sequencing (ChIP-seq) on H3K27, K36 and K9 methylation in JMJD1C knockout versus floxed MLL-AF9 leukemia cells to identify histone marks that correlate with JMJD1C activity. We observed increased H3K36 trimethylation (me3) level on JMJD1C regulated genes loci in the knockout cells compare to controls. Similar increase was observed upon mutating Jumonji and ZFD using CRISPR/Cas9 in ChIP assay. Using Gene Set Enrichment Analysis (GSEA), we found enrichment of JMJD1C regulated genes in the knockout cells in H3K36me3 ChIP-seq data but not H3K27me3 or H3K9 methylation data. Together, these data suggest that Jumonji and ZFD of JMJD1C are most critical for its function in AML and that methylated H3K36 can be used as a marker for JMJD1C activity at gene loci.

Next, we performed single cell RNA-seq to examine the transcriptome change upon mutating the catalytic Jumonji domain. Our analysis showed a hierarchy within control leukemia cells with one cluster bearing more LSC property than the other. Upon mutating Jumonji domain, leukemia cells are driven into differentiated states that bear either more granulocyte or monocyte gene signatures. Using MAST, a single cell transcriptome GSEA tool, we observed enrichment of KRAS signatures in Jumonji mutated cells. This observation in addition to the increased expression of the IL-3 receptor subunit genes Csf2ra/b led us to examine whether the IL-3 signaling pathway is affected in these cells. Western blotting of IL-3 down signaling molecules ERK1/2 and STAT5 showed elevated phosphorylation in Jumonji mutated cells compare to controls both at basal level and in response to IL-3 stimulation. Consistently with this finding, Jumonji mutated cells showed increased sensitivity to inhibitors against RAS (Tipifarnib), MEK (Selumetinib) and JAK (Tofacitinib and Ruxolitinib). Finally, by mining published shRNA and gRNA screen, we observed that mouse as well as human MLLr leukemia cell lines with activating RAS mutations showed decreased or diminished response to knockdown or mutation of JMJD1C suggesting that the presence of RAS mutations render them resistant to JMJD1C modulation.

Overall, we show that enzymatic domain of JMJD1C, the Jumonji domain, mediates its function in AML leukemogenesis, therefore can be targeted for potential therapeutic intervention. In addition, we uncovered a novel interplay between JMJD1C and IL-3 signaling pathway.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.