BCL7A is silenced by hypermethylation to promote acute myeloid leukemia

Abstract

Background: Recent massive sequencing studies have revealed that SWI/SNF complexes are among the most frequently altered functional entities in solid tumors. However, the role of SWI/SNF in acute myeloid leukemia is poorly understood. To date, SWI/SNF complexes are thought to be oncogenic in AML or, at least, necessary to support leukemogenesis. However, mutation patterns in SWI/SNF genes in AML are consistent with a tumor suppressor role. Here, we study the SWI/SNF subunit BCL7A, which has been found to be recurrently mutated in lymphomas, but whose role in acute myeloid malignancies is currently unknown.

Methods: Data mining and bioinformatic approaches were used to study the mutational status of BCL7A and the correlation between BCL7A expression and promoter hypermethylation. Methylation-specific PCR, bisulfite sequencing, and 5-aza-2'-deoxycytidine treatment assays were used to determine if BCL7A expression was silenced due to promoter hypermethylation. Cell competition assays after BCL7A expression restoration were used to assess the role of BCL7A in AML cell line models. Differential expression analysis was performed to determine pathways and genes altered after BCL7A expression restoration. To establish the role of BCL7A in tumor development in vivo, tumor growth was compared between BCL7A-expressing and non-expressing mouse xenografts using in vivo fluorescence imaging.

Results: BCL7A expression was inversely correlated with promoter methylation in three external cohorts: TCGA-LAML (N = 160), TARGET-AML (N = 188), and Glass et al. (2017) (N = 111). The AML-derived cell line NB4 silenced the BCL7A expression via promoter hypermethylation. Ectopic BCL7A expression in AML cells decreased their competitive ability compared to control cells. Additionally, restoration of BCL7A expression reduced tumor growth in an NB4 mouse xenograft model. Also, differential expression analysis found that BCL7A restoration altered cell cycle pathways and modified significantly the expression of genes like HMGCS1, H1-0, and IRF7 which can help to explain its tumor suppressor role in AML.

Conclusions: BCL7A expression is silenced in AML by promoter methylation. In addition, restoration of BCL7A expression exerts tumor suppressor activity in AML cell lines and xenograft models.

Keywords: AML; DNA hypermethylation; SWI/SNF complex; Tumor suppressor.

Fig. 1

BCL7A methylation and expression status in AML patients and cell lines. a Significant inverse correlation (Pearson’s correlation coefficient) between methylation and expression of BCL7A in TCGA-LAML cohort, TARGET-AML and Glass et al. Methylation is expressed as the beta-value of the Illumina HM450 probe (TCGA, TARGET) or genomic position (Glass et al.) most inversely correlated with gene expression. b CpG methylation profile around the transcription start site (TSS) of BCL7A in AML cell lines. Methylation fraction measured by reduced representation bisulfite sequencing data from DepMap Portal (https://depmap.org/portal/) is shown along with our bisulfite sequencing data of the NB4 cell line (top row). CpGs significantly correlated with expression (FDR < 0.05) are highlighted under grey background and their Pearson’s correlation coefficient is shown above. The TSS of BCL7A is indicated by a vertical line. c Methylation fraction 1 kb upstream of the BCL7A TSS and gene expression in 32 AML cell lines from DepMap Portal. NB4 shows the highest methylation level as well as low BCL7A expression

Fig. 2

BCL7A restoration promotes a tumor suppressor-like phenotype. a Differential expression of BCL7A mRNA between acute myeloid leukemia (AML) and diffuse large B-cell lymphoma (DLBCL) cell lines from DepMap. TPM: transcripts per million (Supplementary Table 1). b Methylation Specific PCR (MSP). PCR products generated by the specific primers for the methylated sequence (M) and for the unmethylated sequence (U) are depicted. MSP was carried out using genomic DNA from the NB4 cell line, as well as completely methylated and unmethylated genomic DNAs. In addition, a negative control with no DNA template was included. c Relative BCL7A protein expression differences of the NB4 cell line due to Decitabine (DAC) treatment. BCL7A presents two alternative isoforms that are co-expressed due to an extended exon variant located in exon 5, and due to this, there are two bands for BCL7A. β-actin was used as a loading control and the numeric result was normalized with this loading control signal. d Competition cell growth effect of BCL7A expression restoration on in vitro proliferation. Competition cell growth assay of non-transduced NB4 cells vs NB4 cells transduced with either pLVX-IRES-ZsGreen1 plasmid expressing wild-type BCL7A, mutant BCL7A (∆27-BCL7A), or empty vector. All constructs were evaluated in a time course of ZsGreen1.⁺ % measurements on days 1, 5, 12, 19, and 26. This assay involves the co-culture of an NB4 model transduced with one of the plasmids together with non-transduced NB4 cells. e Bioluminescence signal at day 21 after injection of NB4-luciferase-transduced cells in the empty vector mice group (n = 6) and wild-type BCL7A mice group (n = 6) groups. f Image of the bioluminescence signal of two representative NGS mice in ventral position on day 21 after the tail-injection of NB4-luciferase cells transduced with BCL7A wildtype (up) and empty vector (down)

Fig. 3

RNA sequencing of wild-type BCL7A vs mutant BCL7A NB4 cell models. a Volcano plots highlighting the differentially expressed genes in NB4 cell line when comparing NB4 transfected with BCL7A wildtype and BCL7A mutant. b Enriched terms across the top significant (p adjusted) 400 differentially expressed genes when comparing NB4 transduced with wild-type BCL7A and BCL7A mutant. c Expression values according to our RNA-seq data of three selected genes: HMGCS1, IRF7, and H1-0. Log₂ of variance-stabilized transformed (VST) gene counts was obtained with DESeq2 in R. The red horizontal lines represent the mean