GATA2 regulates the erythropoietin receptor in t(12;21) ALL

Abstract

The t(12;21) (p13;q22) chromosomal translocation resulting in the ETV6/RUNX1 fusion gene is the most frequent structural cytogenetic abnormality in children with acute lymphoblastic leukemia (ALL). The erythropoietin receptor (EPOR), usually associated with erythroid progenitor cells, is highly expressed in ETV6/RUNX1 positive cases compared to other B-lineage ALL subtypes. Gene expression analysis of a microarray database and direct quantitative analysis of patient samples revealed strong correlation between EPOR and GATA2 expression in ALL, and higher expression of GATA2 in t(12;21) patients. The mechanism of EPOR regulation was mainly investigated using two B-ALL cell lines: REH, which harbor and express the ETV6/RUNX1 fusion gene; and NALM-6, which do not. Expression of EPOR was increased in REH cells compared to NALM-6 cells. Moreover, of the six GATA family members only GATA2 was differentially expressed with substantially higher levels present in REH cells. GATA2 was shown to bind to the EPOR 5'-UTR in REH, but did not bind in NALM-6 cells. Overexpression of GATA2 led to an increase in EPOR expression in REH cells only, indicating that GATA2 regulates EPOR but is dependent on the cellular context. Both EPOR and GATA2 are hypomethylated and associated with increased mRNA expression in REH compared to NALM-6 cells. Decitabine treatment effectively reduced methylation of CpG sites in the GATA2 promoter leading to increased GATA2 expression in both cell lines. Although Decitabine also reduced an already low level of methylation of the EPOR in NALM-6 cells there was no increase in EPOR expression. Furthermore, EPOR and GATA2 are regulated post-transcriptionally by miR-362 and miR-650, respectively. Overall our data show that EPOR expression in t(12;21) B-ALL cells, is regulated by GATA2 and is mediated through epigenetic, transcriptional and post-transcriptional mechanisms, contingent upon the genetic subtype of the disease.

Keywords: EPOR; GATA2; acute lymphoblastic leukemia; epigenetic; transcriptional regulation.

Figure 1. EPOR and GATA family members are differentially expressed between ETV6/RUNX1 positive and ETV6/RUNX1 negative ALL cell lines

(A) The expression of EPOR was analyzed in REH (ETV6/RUNX1 positive), NALM-6 (ETV6/RUNX1 negative) and UT-7 (EPOR positive control) cells in triplicate by Q-PCR. Expression values were corrected to 18S ribosomal RNA levels. Mean corrected Ct values (±SD) are shown and statistical differences to NALM-6 were detected by one-way ANOVA and are indicated by *** (p < 0.001). (B) Western blot analysis of EPOR expression in protein extracted from REH, NALM-6 and UT-7 cells. GAPDH was used as a loading control. EPOR expression levels were calculated relative to NALM-6 by densitometric analysis using GAPDH as a normalization factor. (C) The expression of each GATA family member (GATA1-6) was analyzed in REH, NALM-6 and UT-7 cells in triplicate by Q-PCR. Expression values were corrected to 18S ribosomal RNA levels. Mean corrected Ct values (±SD) are shown and statistical differences to NALM-6 were detected by one-way ANOVA and are indicated by *** (p < 0.001). (D) Western blot analysis of GATA2 expression in protein extracted from REH, NALM-6 and UT-7 cells. GAPDH was used as a loading control. GATA2 expression levels were calculated relative to NALM-6 by densitometric analysis using GAPDH as a normalization factor.

Figure 2. EPOR and GATA2 are differentially expressed between ETV6/RUNX1 positive and ETV6/RUNX1 negative ALL patients

(A) The expression of EPOR was analyzed in hyperdiploid (N=10) and t(12;21) translocated (N=10) ALL patients by Q-PCR. Expression values were corrected to 18S ribosomal RNA levels. (B) EPOR probe intensities (probe ID: 37986_at) of hyperdiploid (N=40) and t(12;21) translocated (N=58) MILE study ALL patients extracted after normalization of expression files. (C) The expression of GATA2 was analyzed in hyperdiploid (N=9) and t(12;21) translocated (N=10) ALL patients by Q-PCR. Expression values were corrected to 18S ribosomal RNA levels. (D) GATA2 probe intensities (probe ID: 209710_at) of hyperdiploid (N=40) and t(12;21) translocated (N=58) MILE study ALL patients extracted after normalization of expression files. Whiskers indicate Tukey minimum and maximum values; boxes indicate inter-quartile range, with the median marked. Significantly different expression was detected by Student’s t-test with Welch’s Correction and indicated by *** (p < 0.001). (E) Correlation between EPOR and GATA2 mRNA expression as measured by Q-PCR in patient samples. NOTE: High ΔCt values correspond to low gene expression. Correlation coefficient (r = 0.714) and associated p-value (p < 0.001) were calculated by Pearson’s correlation test. (F) Correlation between EPOR (probe ID: 37986_at) and GATA2 (probe ID: 209710_at) intensity values in MILE study patient samples. NOTE: High Probe Intensity values correspond to high gene expression. Correlation coefficient (r = 0.614) and associated p-value (p < 0.001) were calculated by Pearson’s correlation test.

Figure 3. GATA2 binds to the 5′ UTR region of the EPOR gene in REH, but not NALM-6, cells

(A) Schematic of the EPOR genomic locus showing the relative positions of predicted GATA2 binding sites (↓), the EPOR transcription start site (TCSS), the EPOR translation start site (TLSS) and the amplicon targets in ChIP experiments. All genomic coordinates are given relative to the TCSS. (B) ChIP assays were performed on formaldehyde-fixed chromatin prepared from REH and NALM-6 cells. Enrichment of GATA2 binding to EPOR 5′ DNA was determined by comparison to a non-specific binding region and input chromatin controls. Two independent GATA2 antibodies were used. GATA2 binding enrichment was assessed at four genomic loci (Amplicon 1-4). Significant enrichments were detected by two-way ANOVA with Bonferroni’s post-hoc test and are indicated by *** (p < 0.001).

Figure 4. The EPOR and GATA2 5′ DNA is highly methylated in NALM-6, but not in REH cells

(A) Schematic of the EPOR genomic locus showing the relative positions of CpG dinucleotides (), the EPOR transcription start site (TCSS), the EPOR translation start site (TLSS) and the CpG sites included in the pyrosequencing assay (grey dashed box). All genomic coordinates are given relative to the TCSS. (B) Schematic of the GATA2 genomic locus showing the relative positions of CpG dinucleotides (), the GATA2 transcription start site (TCSS) and the CpG sites included in the pyrosequencing assay (grey dashed box). (C) EPOR 5′ DNA specific pyrosequencing assays were performed on bisulphite converted DNA prepared from REH, NALM-6 and UT-7 cells. DNA methylation was assessed at 18 CpG sites in triplicate. Whiskers indicate Tukey minimum and maximum CpG methylation values; boxes indicate inter-quartile range, with the median marked. Significant enrichments were detected by one-way ANOVA and are indicated by *** (p < 0.001). (D) GATA2 5′ DNA specific pyrosequencing assays were performed on bisulphite converted DNA prepared from REH, NALM-6 and UT-7 cells. DNA methylation was assessed at 5 CpG sites in triplicate. Whiskers indicate Tukey minimum and maximum CpG methylation values; boxes indicate inter-quartile range, with the median marked. Significant enrichments were detected by one-way ANOVA and are indicated by *** (p < 0.001).

Figure 5. Decitabine causes demethylation of both EPOR and GATA2 but only increases expression of GATA2

EPOR and GATA2 expression in REH and NALM-6 cells after treatment with 50 to 500 nM Decitabine. Expression values were corrected to 18S ribosomal RNA levels. Mean corrected Ct values (±SD) are shown. Statistical differences compared to the control (0 nM Decitabine) are indicated by * (p < 0.05), ** (p < 0.01), or *** (p < 0.001) were calculated with the two-way ANOVA with Holm Sidak correction.

Figure 6. Expression of microRNAs predicted to target EPOR and GATA2 are increased in NALM-6 cells

(A) Venn diagram showing consensus between an in silico microRNA targeting algorithm [45] and publicly available prediction databases. MicroRNAs predicted to target EPOR were selected and overlapped. (B) Venn diagram showing consensus between an in silico microRNA targeting algorithm [45] and publicly available prediction databases. MicroRNAs predicted to target GATA2 were selected and overlapped. (C) Volcano plot of the differential expression of microRNAs between REH and NALM-6 cells and the associated nominal p-value. The expression of 670 microRNAs was analyzed in REH and NALM-6 cells in triplicate by Q-PCR using multiplex assays. Expression values were corrected to the mean RNU6 and RNU44 levels. Nominal p-values associated with the fold differences compared to NALM-6 were determined using the Bioconductor package ‘limma’. Selection criteria for significantly different expression were an absolute fold change ≥5 and a nominal p-value < 0.05. (D) The differential expression of selected microRNAs (miR-362-5p and miR-650) was validated in REH and NALM-6 cells in triplicate by Q-PCR using single microRNA assays. Expression values were corrected to the mean RNU6 and RNU44 levels. Mean relative expression levels (±SD) compared to REH are shown and statistical differences to REH were detected by two-way ANOVA and are indicated by * (p < 0.05) or *** (p < 0.001).

Figure 7. Forced expression of miR-362-5p and miR-650 reduces EPOR and GATA2 expression

(A) The expression of miR-362-5p and EPOR were analyzed in REH cells in triplicate by Q-PCR 24 hr post-transfection with a miR-362-5p expression vector. Expression values were corrected to 18S ribosomal RNA levels. Mean relative expression levels (±SD) compared to empty vector controls are shown and statistical differences to control were detected by one-way ANOVA and are indicated by *** (p < 0.001). (B) The expression of miR-362-5p and EPOR were analyzed in REH cells in triplicate by Q-PCR 72 hr post-transfection with a miR-362-5p expression vector. Expression values were corrected to 18S ribosomal RNA levels. Mean relative expression levels (±SD) compared to empty vector controls are shown and statistical differences to control were detected by one-way ANOVA and are indicated by *** (p < 0.001). (C) Western blot analysis of EPOR expression in protein extracted from REH cells 24 hr. and 72 hr. post-transfection with a miR-362-5p expression vector. GAPDH was used as a loading control. EPOR expression levels were calculated relative to htR (EV control) by densitometric analysis using GAPDH as a normalization factor. (D) The expression of miR-650, EPOR and GATA2 were analyzed in REH cells in triplicate by Q-PCR 24 hr. post-transfection with a miR-650 mimetic oligo. Expression values were corrected to 18S ribosomal RNA levels. Mean relative expression levels (±SD) compared to scrambled oligo controls are shown and statistical differences to control were detected by one-way ANOVA and are indicated by *** (p < 0.001). (E) The expression of miR-650, EPOR and GATA2 were analyzed in REH cells in triplicate by Q-PCR 72 hr. post-transfection with a miR-650 mimetic oligo. Expression values were corrected to 18S ribosomal RNA levels. Mean relative expression levels (±SD) compared to scrambled oligo controls are shown and statistical differences to control were detected by one-way ANOVA and are indicated by *** (p < 0.001). (F) Western blot analysis of EPOR and GATA2 expression in protein extracted from REH cells 24 hr. and 72 hr. post-transfection with a miR-650 mimetic oligo. GAPDH was used as a loading control. EPOR and GATA2 expression levels were calculated relative to scrambled oligo control by densitometric analysis using GAPDH as a normalization factor.