Fig 2: Regulatory mutations at HMX2/3 in EOL-1.The illustration of RNA-seq data from EOL-1 shows two different mutations at the HMX2/3 locus. (A) An A-to-G mutation in the 5´-region of HMX3 generates a consensus ETS-site. (B) A T-to-C mutation in the 5´-region of HMX2 transforms an NFkB-site into a SP1-site. (C) Quantification of ETS1 (above) and ELK1 (below) in selected AML cell lines by RQ-PCR. (D) Forced expression of ETS1 (above) and ELK1 (below) in EOL-1 cells resulted in respectively reduced and increased expression levels of both HMX2 and HMX3. Asterisks indicate calculated p-values obtained by t-Test analysis of controls (vector) and factor expression. (E) Reporter-gene assay in NIH-3T3 cells using a fragment containing a normal (obtained from MV4-11) or mutated (obtained from EOL-1) ETS-site. Forced expression of ELK1 (left) and ETS1 (right) resulted in elevated reporter-gene activity which was higher using ELK1 for the mutated fragment. (F) Reporter-gene assay in HELA cells using a fragment containing a normal (MV4-11) or mutated (EOL-1) NFkB-site. Treatment with NFkB-activator TNFa (left) resulted in reduced reporter-gene activity which was more pronounced using the normal fragment. Forced expression of SP1 (right) resulted in elevated reporter-gene activity which was higher using the mutated fragment.

Fig 3: Regulation of HMX2 and HMX3 by KMT2A and ERK-signalling.(A) RQ-PCR analysis of EOL-1 (left) and MV4-11 (middle) treated for siRNA-mediated knockdown of KMT2A demonstrates raised expression levels of HMX2 and HMX3, indicating a repressive impact of KMT2A. This treatment resulted in enhanced expression of HMX3 in HL-60 cells which did not express HMX2 (right). Asterisks indicate calculated p-values obtained by t-Test analysis of controls (siCTR) and siRNA-targeted knockdown. (B) RQ-PCR analysis of KMT2A of EOL-1 cells treated for siRNA-mediated knockdown of HMX2 (left) and HMX3 (right) showed unaltered KMT2A expression levels, discounting any regulatory impact. (C) Quantification of KMT2A transcripts in AML cell lines and primary granulocytes showed low expression levels in EOL-1, MV4-11 and in eosinophils. (D) RQ-PCR analysis of HMX2 showed elevated expression levels in EOL-1 cells treated with PDGFRA-inhibitor dasatinib. Asterisks indicate calculated p-values obtained by t-Test analysis of controls (DMSO) and pharmacological treatments. (E) Treatment of EOL-1 cells with ERK-inhbitor PD98059 resulted in reduced levels of phospho-ERK as analyzed by Western blot (left), and in elevated transcript levels of HMX2 (middle) and HMX3 (right) as analyzed by RQ-PCR. (F) Treatment of EOL-1 cells with ERK-activator FLT3LG resulted in elevated levels of phosphorylated ERK (above) and in reduced transcript levels of HMX2 and HMX3 (below).

Fig 4: Gene regulatory network for HMX2 and HMX3 in AML.This diagram summarizes the results of this study, showing a gene regulatory network. HMX2 and HMX3 are located centrally, activating and repressing factors and pathways are indicated above, and the target genes EPX and HTR7 below. Deregulation of EPX and HTR7 indicates aberrant effects for differentiation and proliferation. Genomic rearrangements targeting KMT2A or PDGFRA are indicated.

Fig 5: Identification of upstream factors regulating HMX2 and HMX3.(A) Treatment of EOL-1 cells with IL7 resulted in increased transcript levels of HMX2 and HMX3 as analyzed by RQ-PCR. Asterisks indicate calculated p-values obtained by t-Test analysis of medium controls and interleukin stimulation. (B) Treatment of EOL-1 and MV4-11 cells with WNT3A and WNT5B resulted in increased expression levels of HMX2 (left) and HMX3 (right). (C) Treatment of EOL-1 cells with NFkB-activator TNFa resulted in reduced expression levels of HMX2 and HMX3 (left). Treatment of EOL-1 cells with NFkB-inhibitor resulted in increased expression of HMX2 and HMX3 (middle). STAT5A served as marker for NFkB-activity which showed low expression levels and increased activity after stimulation with TNFa (right). (D) A genomic map of the locus for HMX2/3 was obtained from the UCSC genome browser, showing potential transcription factor binding sites including those for IRF1, IRF2 and NKX2-5. (E) RQ-PCR analysis of EOL-1 cells treated for siRNA-mediated knockdown of IRF8 resulted in elevated expression levels of HMX2 (left) and HMX3 (right), indicating an activating impact of IRF8. (F) Quantification of IRF8 by RQ-PCR in primary granulocytes and monocytes (left) and selected AML cell lines (right) demonstrated significant levels in EOL-1 and MV4-11. (G) RQ-PCR analysis of EOL-1 (left) and MV4-11 (right) treated for siRNA-mediated knockdown of HMX3 and HMX2 demonstrated slightly reduced expression levels of HMX2 and HMX3, respectively, indicating mutual activation.