Chromatin Architecture Modulation in B-Cell Acute Lymphoblastic Leukemia Carrying DUX4 Fusions

B-cell acute lymphoblastic leukemia (B-ALL) carrying DUX4 fusions is a novel cluster of B-ALL. DUX4 fusions are generated from insertions of wild- type (WT) DUX4, mainly into the IGH locus.The translocation replaces the 3′ end of the WT DUX4 coding region with a fragment of IGH or another gene, producing DUX4 out-of-frame fusion proteins devoid of the C terminus of WT DUX4. Usually, WT DUX4 is expressed in germ cells in testis, while its expression is epigenetically repressed in somatic tissues. Recently, it is identified to plays a critical role in transcriptional programs at the cleavage of human fertilized egg. In B-ALL, DUX4-IGH (D-I) is shown to be essential for leukemic transformation; however, little is known about the mechanistic basis. Here in this study, we extensively investigated the biological effects of D-I.

First, we assessed the role of D-I using in vitro cell culture assays with human cord blood (CB) CD34+ cells. Introduction of D-I significantly caused retention of the CD34+ cell population compared with the mock vector, even though it failed to preferentially promote differentiation toward B cell lineage in vitro.

To analyze the epigenetic and transcription control by D-I, we performed chromatin immunoprecipitation coupled with sequencing (ChIP-seq) using cell lines. In NALM6, a B-ALL cell line carrying D-I, a subset of D-I binding sites is accompanied by H3K4me3 and H3K27ac. We also assessed the histone modification status in Reh cells, a B-ALL cell line without DUX4 fusions, and observed that active histone marks are detected after binding of ectopically expressed D-I. Nevertheless, RNA sequencing of NALM6 and Reh overexpressing D-I showed minimal activation of genes near the D-I binding sites compared with those of NALM6 overexpressing WT DUX4.

WT DUX4 is known to preferentially bind and activate repeat elements, especially human endogenous retroviral (HERV) elements in embryonic cells. NALM6 cells overexpressing WT DUX4 showed a drastic increase in the expression of HERV elements, while NALM6 and Reh overexpressing D-I did not. The expression of HERV elements was not altered by D-I in all the genomic regions, and we did not observe increased expression of HERV elements in patient leukemia samples with DUX4 fusions as well. Furthermore, Assay for Transposase Accessible Chromatin Sequencing (ATAC-seq) showed that chromatin status was not affected by the binding of D-I at the D-I bound HERV element, indicating that transcriptional and insulating ability of WT DUX4 in these areas are lost in D-I.

Next, we performed ATAC-seq using NALM6 cells, comparing the status between pre- and post- D-I knockdown. Genomic areas with decreased ATAC signal after knockdown of D-I are enriched in D-I binding sites, and ATAC signal was increased when we compared the status between pre- and post- induction of D-I in Reh cells. Through the immunoprecipitation of endogenous D-I in NALM6 cells, we identified SWI/SNF complex elements as binding partners of D-I, further highlighting the chromatin opening ability of D-I. Motif analysis of the genomic areas with decreased ATAC signal after knockdown of D-I identified only DUX4 motif as a significant motif, suggesting that D-I is not apparently cooperating with other transcription factors. On the other hand, ATAC signal was increased in substantial genomic areas after knockdown of D-I, and motif analysis identified SPI1, TCF3, and EBF1 motifs. Integrated analysis of transcriptome data also supports the idea that transcription factors related to B cell differentiation are repressed in the presence of D-I, and derepressed after knock down of D-I.

Despite the attenuated transcriptional activity, B-ALL carrying DUX4 fusions manifests a characteristic expression pattern. D-I binding sites are not always relevant to the gene areas with increased transcriptions. Therefore, we compared the genomic areas where ATAC signal is raised by D-I, and genes whose expression is affected by D-I. We identified genes with ATAC signal change both in NALM6 cells with D-I knockdown and in Reh cells with D-I induction. We identified D-I binding in some of these genes, and the pharmacological inhibition of one of the genes caused cell death in NALM6 cells in vitro and in vivo, suggesting that this gene is the genuine target of D-I.

In summary, our study elucidated the detailed difference of function between WT DUX4 and DUX4-IGH, and demonstrated the ability of DUX4-IGH as a chromatin modulator.