DNA methylation-mediated silencing of PU.1 in leukemia cells resistant to cell differentiation

Abstract

In mice, the proviral integration of the Friend Spleen Focus Forming Virus (SFFV) within the PU.1 locus of erythroid precursors results in the development of erythroleukemia. SFFV integrates several kilobases upstream of the PU.1 transcription initiation start site leading to the constitutive activation of the gene which in turn results in a block of erythroid differentiation. In this study we have mapped and sequenced the exact location of the retroviral integration site. We have shown that SFFV integrates downstream of a previously described upstream regulatory element (URE), precisely 2,976 bp downstream of the URE-distal element. We have also found that SFFV persists integrated within the same location in resistant cell lines that have lost their differentiation capacity and in which case PU.1 remains silent. We have examined the methylation status of PU.1 and found that in resistant cells the nearby CpG islands remained methylated in contrast to a non-methylated status of the parental cell lines. Treatment with 5-aza-2'-deoxycytidine caused resistant cells to differentiate yet only when combined with HMBA. Altogether these results strongly suggest that methylation plays a crucial role with regard to PU.1 silencing. However, although demethylation is required, it is not sufficient to overcome the differentiation impasse. We have also showed that activation blockage of the Epo/Epo-R pathway remains despite of the absence of PU.1.

Keywords: DNA methylation; Erythroleukemia cells; PU.1; SFFV.

Figure 1

PCR probe to confirm the SFFV integration site within the PU.1 locus. A) Illustration of the multiple verification PCR amplifications using MEL genomic DNA as a template. Black arrows over the SFFV genome (gray box) indicate the pair of primers designed to amplify the SFFV-PU.1 junctions. Odd numbers 1 and 3 represent the primers used to identify the upstream integration junction and even numbers 2 and 4 represent the primers used for the downstream integration. Number 5 represents the long-range PCR (LR-PCR) used as a probe to confirm the complete SFFV integration. Black boxes correspond to the five exons of PU.1; the arrow above exon number one represents the initiation and direction of translation. B) Agarose gel electrophoresis performed using the primers schematized in A). C) Agarose gel electrophoresis of the LR-PCR probe to confirm SFFV integration; both the wild type (564 bp) and the integrated allele (6,859 bp) are visualized.

Figure 2

Map of the PU.1 genomic region showing the location and orientation of the SFFV proviral integration in MEL-DS19 cells. A) Diagram (drawn to scale) of the MEL-DS19 PU.1 genomic locus showing the location of a 3.4-kb HindIII (H) fragment referred to as the kb-14 URE (Upstream Regulatory Element) (Okuno et al. 2005), the SFFV genome is integrated downstream of the URE (gray box flanked by white boxes that symbolize the viral “long terminal repeats” (L)), and the five coding exons (black boxes) of the PU.1 locus. The sequences drawn below represent the upstream and downstream viral–host junctions, respectively. B) and C) Sequence alignment of the SFFV integration site displaying the PU.1 genomic sequences of mouse and of the MEL-DS19 cell line together with the SFFV genomic sequence. The PU.1 genomic sequence of the MEL-DS19 cell line corresponds to the upstream (B) and downstream (C) viral–host junction. Viral-host junctions as indicated in A) are underlined.

Figure 3

Reactivation of the silenced PU.1 locus by 5-azaC treatment of MEL-R cells. Gene expression analysis of PU.1 using quantitative real time PCR (qRT-PCR) of MEL DS19 and MEL-R cell lines after treatment with 5-azaC. A) Non-induced MEL and MEL-R cells. MEL-R cells were routinely cultured in the presence of HMBA (+). B) MEL-R cells cultured for 24 hr in the absence (-) or presence (+) of 0.4 μM 5-azaC. C) MEL DS19 cells cultured during 24 hr in the absence (-) or presence (+) of 0.4 μM 5-azaC and 5 μM HMBA. The magnitude of gene expression was calculated using the ∆∆Ct method. Each value is the average of three readings.

Figure 4

Methylation status of the PU.1 promoter region. A) DNA chromatogram of the PU.1 upstream region using bisulfite-treated genomic DNA derived from MEL and MEL-R cell lines. Red circles highlight the cytosines of the four CpG islands of the MEL-R cell line that changed to thymines in the MEL-DS19 cell line after bisulfite treatment. B) Methylation analysis of the PCR products of the bisulfite-treated genomic DNA from A). The CG sequences were numbered according to (Shearstone et al. 2011) and their methylation status was determined.

Figure 5

DNA methylation status of the PU.1 locus regulatory region. A) A genomic map (not drawn to scale) depicts the regions containing the analyzed CpGs. The CpG colored in blue represent PCR amplified islands which however were not detected by the sequencing primers. CpG colored in red represent sequenced islands. B) Individual CpG methylation percentages of MEL DS19 and MEL-R cell lines.

Figure 6

Treatment with 5-azaC induces MEL-R cells to differentiate in presence of HMBA. Percentage of differentiation of MEL-R cells was determined by the benzidine assay. Cells were cultured in the absence or presence of different concentrations of 5-azaC, with or without 5 mM HMBA. Each bar indicates the mean and standard deviation of three independent experiments.

Figure 7

Ectopic expression of PU.1 restricts the proliferative capacity but is unable to induce MEL-R cell differentiation. A) Western blot analysis of lysates derived from MEL-R PU.1-ER transfectants stimulated with β-estradiol. Equal amounts of protein (50 μg) were fractionated by SDS-polyacrylamide gel electrophoresis and analyzed by immunoblotting with an anti-PU.1 antibody. Numbers above the panel correspond to different clones. B) Percentage of benzidine positive cells (B+) in PU.1 and GATA-1 stable transfectants activated by β-estradiol and cultured in the presence of 5 mM HMBA. Each bar indicates the mean and standard deviation of three independent experiments. C) PU.1-ER transfectants were cultured in the presence or absence of β-estradiol and cell counts were recorded at the indicated times.

Figure 8

STAT1 tyrosine phosphorylation analysis of MEL DS19 and MEL-R cell lines. A) Non-stimulated MEL DS19 and MEL-R erythroleukemia cells (C) and stimulated with erythropoietin (100 U/ml) (Epo) or interferon α (IFN-α) (500 U/ml). Total cell lysates were immunoblotted with either anti-phospho-STAT1 (tyrosine 701) (upper panel) or anti-STAT1 antibody (lower panel). B) Comparison of Shp-1 expression by semi-quantitative RT-PCR of the parental MEL cells untreated (0) or treated with HMBA during 48 or 96 hours (48, 96) and of MEL-R cells (R). PCR products were normalized to GAPDH, separated by electrophoresis in a 1% agarose gel and stained with ethidium bromide.