C/EBPα and MYB regulate FLT3 expression in AML

Abstract

The interaction between the receptor FLT3 (FMS-like tyrosine kinase-3) and its ligand FL leads to crucial signalling during the early stages of the commitment of haematopoietic stem cells. Mutation or over-expression of the FLT3 gene, leading to constitutive signalling, enhances the survival and expansion of a variety of leukaemias and is associated with an unfavourable clinical outcome for acute myeloid leukaemia (AML) patients. In this study, we used a murine cellular model for AML and primary leukaemic cells from AML patients to investigate the molecular mechanisms underlying the regulation of FLT3 gene expression and identify its key cis- and trans-regulators. By assessing DNA accessibility and epigenetic markings, we defined regulatory domains in the FLT3 promoter and first intron. These elements permit in vivo binding of several AML-related transcription factors, including the proto-oncogene MYB and the CCAAT/enhancer binding protein C/EBPα, which are recruited to the FLT3 promoter and intronic module, respectively. Substantiating their relevance to the human disease, our analysis of gene expression profiling arrays from AML patients uncovered significant correlations between FLT3 expression level and that of MYB and CEBPA. The latter relationship permits discrimination between patients with CEBPA mono- and bi-allelic mutations, and thus connects two major prognostic factors for AML.

Figure 1. FMH9 cells and AML cells flow cytometric characterisation.

(a) (b) Staining with labelled antibodies against stem cell-related antigens (KIT, SCA-1, CD34, FLT3) and lineage specific markers (GR-1, CD11b) are overlaid against matched isotype controls. (c) FLT3 mRNA quantification by quantitative-PCR, normalised against the 18S house keeping gene results.

Figure 2. Flt3 regulatory modules.

(a) Schematic representation of the Flt3 promoter and first intron (filled boxes indicate exon1 and 2). The level of sequence conservation between seven mammalian species and the position of repeated sequences are indicated below. The global illustration is aligned to the lower (b) and bottom (c) plots. (b) DNaseI HSS mapping at the Flt3 locus. Ratios between quantitative-PCR results from partially digested and untreated samples (2^–ΔCt) reflect the extent of nuclease sensitivity across regions covered by PCR amplicons. (c) Histone mark distribution at the Flt3 locus in the murine FMH9 AML model. Relative enrichments are determined against the control immunoglobulin G X-ChIP material.

Figure 3. Detection of transcription factor binding at the sites of nuclease hypersensitivity.

Relative enrichments were determined against the immunoglobulin G control material by quantitative-PCR at the location of the hypersensitive regions (HSA to C at and – 1.45, – 0.15 + 7.5 kb relative to the Flt3 initiating codon) and normalised against two internal control regions (at 3.5 and – 0.27 kb from the ATG). Error bars represent the s.e. of – the mean. All plots are representative of a minimum of three independent experiments used to determine the two-tailed P-value by paired t-test: ***00.001 and **00.01.

Figure 4. Expression codependency between FLT3 and its potential transcriptional regulators in CN-AML measured by the Pearson correlation test.

The correlation coefficients were determined using log expression values from two sets of arrays summing 46 and 58 CN-AML patients without CEBPA mutations. Statistical significance ***<0.001 **<0.01 and *<0.05.

Figure 5. C/EBPα and MYB bind the FLT3 locus in human AML cells.

(a) Schematic comparison of the murine and human FLT3 promoter and first intron (filled boxes represent exons) with indication of sequence conservation and repeated sequence locations. (b) DNaseI hypersensitivity mapping at the human FLT3 locus in CN-AML patients. Plotted for the human sequence orthologs of the murine HSA, HSB, HSC and a control region, ratios between quantitative-PCR results from partially digested and untreated samples are representative of four experiments. (c) Detection of MYB and C/EBPα in vivo binding at the sites of nuclease hypersensitivity, located at – 15.3 kb (hA), –0.37 kb (hB) and + 6.2 kb (hC1/C2) from the human FLT3 initiating codon. Relative enrichments were determined against the immunoglobulin G control material and normalised against the control region (at –6.95 kb from the ATG). Error bars represent the s.e. of the mean. C/EBPα ChIP plots are representative of three independent experiments used to determine the two-tailed P-value by paired t-test. MYB ChIP plots represent patient-specific experiments for which the P-values were calculated using replicate measurements. P-value indications ***<0.001, **<0.01 and *<0.05.

Figure 6. Silencing and ectopic expression of MYB and C/EBPα in FMH9 cells.

Transcript abundances were determined by quantitative-PCR in cells transfected with shRNA and expression vectors, normalised against the Gapdh house keeping gene PCR result and compared with the correspondent empty vector controls. Results are representative of three or four independent experiments. Statistical significance ***<0.001 and **<0.01.

Figure 7

(a) Pearson correlation test between FLT3 and CEBPA transcript levels applied to CEBPAWT, CEBPASM and CEBPADM patient arrays. Statistical significance: ***<0.001 and *<0.05. (b) Scatter plots and trend line fitting between FLT3 and CEBPA expression in the different subgroups (c) FLT3 mRNA levels in AML patients with wild-type mono- or bi-allelic CEBPA mutations. Median values are indicated. Differential statistical significances are indicated through P-values or non-significance sign NS.

Figure 8. Patients with wild-type and bi-allelic CEBPA mutations differentially express genes that are part of the signature of the constitutively activated mutant receptor FLT3-ITD.

Bars represent the average expression in AML patients with wild-type and bi-allelic CEBPA mutations and error bars depict the s.e. of the mean. Tested genes were previously shown to be up-(left panel) or downregulated (right panel) in response to FLT3-ITD signalling (a) or are elements of the FLT3-ITD signature in CN-AML (b). Statistical significance: ***<0.001, **<0.01 and *<0.05.