ZNF521 Enhances MLL-AF9-Dependent Hematopoietic Stem Cell Transformation in Acute Myeloid Leukemias by Altering the Gene Expression Landscape

Abstract

Leukemias derived from the MLL-AF9 rearrangement rely on dysfunctional transcriptional networks. ZNF521, a transcription co-factor implicated in the control of hematopoiesis, has been proposed to sustain leukemic transformation in collaboration with other oncogenes. Here, we demonstrate that ZNF521 mRNA levels correlate with specific genetic aberrations: in particular, the highest expression is observed in AMLs bearing MLL rearrangements, while the lowest is detected in AMLs with FLT3-ITD, NPM1, or CEBPα double mutations. In cord blood-derived CD34⁺ cells, enforced expression of ZNF521 provides a significant proliferative advantage and enhances MLL-AF9 effects on the induction of proliferation and the expansion of leukemic progenitor cells. Transcriptome analysis of primary CD34⁺ cultures displayed subsets of genes up-regulated by MLL-AF9 or ZNF521 single transgene overexpression as well as in MLL-AF9/ZNF521 combinations, at either the early or late time points of an in vitro leukemogenesis model. The silencing of ZNF521 in the MLL-AF9 + THP-1 cell line coherently results in an impairment of growth and clonogenicity, recapitulating the effects observed in primary cells. Taken together, these results underscore a role for ZNF521 in sustaining the self-renewal of the immature AML compartment, most likely through the perturbation of the gene expression landscape, which ultimately favors the expansion of MLL-AF9-transformed leukemic clones.

Keywords: acute myeloid leukemia (AML); chromosomal translocations; cord blood-derived hematopoietic stem cells (CB-CD34+); fusion gene MLL-AF9; gene expression; human zinc finger protein 521 (hZNF521); mixed lineage leukemia gene (MLL) AF9 (MLLT3 or LTG9).

Figure 1

ZNF521 gene expression levels according to cytogenetic groups and FAB classification. (A) ZNF521 mRNA expression levels were analyzed in different cytogenetic groups compared to normal karyotype AMLs from the MILE data set (GSE13204). Significant p values: MLL rearrangement vs. non leukemic bone marrow, p = 8.85 × 10^–3; MLL vs. t(15;17), p = 3.97 × 10^–4; MLL vs. t(8;21), p = 7.01 × 10^–4; MLL vs. complex aberrant karyotype, p = 2.381 × 10^–9; and MLL vs. inv(16)/t(16;16), p = 2.611 × 10^–3. (B) den Boer data set (GSE17855): Significant p values: MLL vs. t(15;17), p = 1.15 × 10^–18; MLL vs. cn–AML, p = 1.31 × 10^–14; MLL vs. t(8;21), p = 3.73 × 10^–13; 16)/t(16;16), p = 5.03 × 10^–8. (C) ZNF521 mRNA expression analyzed with respect to FAB subgroup classification of AMLs (GSE6891). Significant p values: MO vs. M3, p = 6.02 × 10^–3 and MO vs. M5 p= 4.47 × 10^–2. ZNF521 mRNA expression level in 460 pediatric and adult AML samples (Delwel) stratified with respect to molecular lesions: (D) FLT3, (E) CEBPα, (F) NPM1, (G) IDH1 mutations, (H) EVI1 (MECOM) aberrant expression and (I) RAS oncogenes. (cn: cytogenetically normal). Significant p values: * < 0.05, ** < 0.01, *** < 0.001.

Figure 2

CB-CD34⁺ transduced cell expansion in vitro.(A) Purified cord blood hematopoietic progenitor cells were characterized by FACS for the CD34 antigen. (B) FACS analysis of transduced cells was performed to evaluate the percentage of E-GFP positive cells at 6 days after transduction. The transduction efficiency of CD34⁺ cells was evaluated by Q-RT-PCR analysis: mRNA levels of (C) ZNF521 and (D) the MLL-fusion gene are significantly enhanced in transduced cells compared to control cells. (E) In four different MLL-AF9 transductions of CB-CD34⁺ cells, the amounts of ZNF521 mRNA are shown (analyzed at 37 days). (F) CB-CD34⁺ cells (2 × 10⁴) were plated in MyeloCult™ complete with cytokines. Cumulative cell expansion is shown, and the p-value is calculated compared to control vector. All assays were performed in triplicate. (G) Colony-forming cell (CFC) assays were performed using StemMACS HSC-CFU medium complete with cytokines. Colonies were scored after 2 weeks, and the number of progenitors was normalized to the total cell number in the relevant culture at the time of plating. Data are represented as mean ± SD (* p < 0.05, ** p < 0.01, *** p < 0.001).

Figure 3

Gene expression analysis was performed with the ClariomD array. (A) Clustering on the top 1% (1358/135750) of transcripts with the highest variance across the entire dataset. Average clustering method was applied. Blue–red color scale was used to set rows with a mean of zero and a standard deviation of one. (B) Venn diagrams illustrating the significantly up- or down-regulated genes at FDR <10% by RankProduct analysis in ZNF521 and MA9 at 23 days or after their combination at 58 days compared to control cells. The corresponding transcript lists are reported in Supplementary File S2. (C) Selected significant Hallmark and Kegg pathways determined by GSEA analysis with corresponding normalized enrichment score (NES) and p-value parameters.

Figure 4

Heatmaps of selected transcripts from differentially expressed transcript lists at FDR < 10% of ZNF521, MA9, and ZNF521-MA9 (ZM) at 23 and 58 days compared to the control condition at 23 days. Known or other target genes that are up-regulated in all comparisons (A) at early (B,D) or late time points (C,E) are depicted. Blue–red color scale was used to set rows with a mean of zero and a standard deviation of one. (F) Confirmation of the array by Q-RT-PCR for MEIS1, MEF2C, HOXA10, HOXA9, and PBX3.

Figure 5

Heatmaps of selected differentially expressed transcripts in ZNF521, MA9, and ZNF521-MA9 (ZM) at 23 days and/or in ZNF521-MA9 (ZM) at 58 days compared to the control condition. (A) Genes involved in epigenetic functions. (B) Genes involved in cell cycle function. (C) Transporter genes. Blue–red color scale was used to set rows with a mean of zero and a standard deviation of one.

Figure 6

Effect of ZNF521 silencing in CD34⁺ hematopoietic stem/progenitor cells. CB-MA9 transduced cells cultivated for 30 days were silenced for ZNF521 with specific shRNA transduction, and the levels of ZNF521 (A) and the MLL-AF9 rearrangement (B) were analyzed by RT-Q-PCR. Expression was normalized by GAPDH mRNA expression, and the data represent the means ± SD of three replicates. (C) Immunoblotting analyses with nuclear extracts showed ZNF521 silencing in CB-MLL-AF9 cells; the signal was normalized to the nuclear protein HDAC1. (D) Cell viability was tested using the CellTiter 96^® Aqueous One Solution Assay: 1 × 10³ cells were plated in 96-well plates in complete medium, and absorbance at 490nm was measured after 2 h. Each point represents the mean ± SD of four replicates. (I) Effect of ZNF521 silencing in AML cell lines: THP-1 cells highly expressing both MLL and ZNF521 and were transduced with shRNAs for ZNF521 or with a non-silencing control shRNA. Protein lysates, which were prepared five days post-infection, were analyzed for ZNF521 expression by immunoblotting. (F) The proliferation of these cells was investigated by MTT assay; data are represented as mean ± SD. (G–I) Methylcellulose colony-formation assay was used to evaluate the effect of ZNF521 silencing on THP-1, MOLM-13, and HL60 cell lines. ZNF51 silencing resulted in a significant decrease in the number of colonies formed in cells expressing the MLL-AF9+ fusion protein, THP-1, and MOLM-13. The number of CFU did not change in HL60 cells negative for MLL rearrangements. The effects of ZNF521 silencing were compared to non-target shRNA control cells. Data are represented as mean ± SD (*** p < 0.001).