ETS1 and SP1 drive DHX15 expression in acute lymphoblastic leukaemia

Abstract DHX15 plays a role in leukaemogenesis and leukaemia relapse. However, the mechanism underlying the transcriptional regulation of DHX15 in ALL has not been elucidated. Our present study aimed to explore the functional promoter region of DHX15 and to investigate the transcription factors controlling the transcription of this gene. A luciferase assay performed with several truncated constructs identified a 501‐bp region as the core promoter region of DHX15. Site‐directed mutagenesis, electrophoretic mobility shift and chromatin immunoprecipitation assays showed that ETS1 and SP1 occupied the DHX15 promoter. Furthermore, knockdown of ETS1 and SP1 resulted in suppression of DHX15, whereas the overexpression of these genes led to up‐regulation of DHX15. Interestingly, in samples obtained from patients with ALL at diagnosis, both ETS1 and SP1 correlated positively with DHX15 expression. Additionally, differences in methylation of the DHX15 core promoter region were not observed between the patients and controls. In conclusion, we identified the core promoter region of DHX15 and demonstrated that ETS1 and SP1 regulated DHX15 expression in ALL.

chromosome abnormalities in RNA helicases have been identified in haematological malignancies. DDX3X mutations have been identified in chronic lymphocytic leukaemia 6 and Burkitt's lymphomas, 7,8 although the exact significance of these mutations is not clear. The DDX10 gene fuses with the NUP98 gene to form the chimeric gene NUP98-DDX10, 9 which is involved in de novo or secondary myeloid malignancies 10,11 as well as imatinib resistance. 12 The multiplication and self-renewal of primary human CD34+ cells can also be highly accelerated by NUP98-DDX10. 13 Dysregulation of DDX32 expression has been demonstrated in lymphoid neoplasms, 14,15 suggesting that this gene may contribute to carcinogenesis. These limited studies reveal that some members of the human RNA helicase family may play diverse biological roles in haematologic malignancies.
The DHX15 gene (alias PRP43) is a member of the DEAH-box family and is located on the minus strand of chromosome 4 (4p15.3). 16 Recent evidence has suggested that DHX15 may contribute to carcinogenesis, and overexpression of DHX15 has been observed in lung adenocarcinoma samples. 17,18 Semiquantitative RT-PCR analysis showed up-regulation of DHX15 in breast cancer cells.
RNAi-mediated DHX15 suppression inhibited the proliferation of MCF-7 and T47D human breast cancer cells. 19 The DHX15 p.R222G mutation has been identified in de novo or relapsed acute myeloid leukaemia (AML) and myelodysplastic syndrome (MDS) patients and has been implicated as a potential new AML driver gene. [20][21][22][23] Using real-time qRT-PCR and Western blotting, we observed higher DHX15 mRNA and protein levels, respectively, in human ALL samples than in normal bone marrow (BM) cells. Knockdown of DHX15 in Jurkat cells leads to impaired cell proliferation and increased apoptosis. 24 The TSS of the DHX15 gene was identified in a previous study. 16 However, the transcriptional regulatory mechanism of DHX15 in ALL remains unknown. Our present study explored the putative promoter region of the DHX15 gene to characterize the transcriptional regulatory mechanisms of DHX15 expression in ALL.
The results revealed a 501-bp functional promoter region of DHX15 that harboured binding sites for ETS1 and SP1, which regulate DHX15 expression.   years) following the ethical guidelines of our institution and in accordance with the Declaration of Helsinki. All patient samples were obtained prior to the initiation of any therapy. Informed consent was obtained for the procurement and analysis of these specimens.

| Plasmid constructs and luciferase reporter assay
Fragments of the DHX15 promoter truncated at the 5 0 end with common sequences at the 3 0 end were amplified using PCR. These PCR products were cloned into the XhoI and HindIII sites in a luciferase reporter vector (pGL4.10 [luc]2) (Promega, Madison, WI, USA) lacking a promoter. Positive clones were confirmed by Sanger sequencing.
Overlap extension PCR (OE-PCR) 25 was used to mutate the ETS1-and SP1-binding sites. The pGL4.10-345 plasmid was utilized as a template for the first round of OE-PCR. The presence of the expected mutations in the plasmids was confirmed by Sanger sequencing. All fragments were amplified using the primers listed in Table S1. The Jurkat and NALM6 cells were transiently transfected using the Amaxa Cell Line Nucleofector Kit V (Lonza). A total of 2 9 10 6 cells were resuspended in 100 lL of Cell Line Nucleofector Solution V and mixed with 4 lg of the promoter constructs and 400 ng of the internal control (pRL-TK). Nucleofection was performed as previously described. 26 At 24 hour post-transfection, the cells were lysed using passive lysis buffer and analysed for luciferase activity with the reagents provided in the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's instructions. The SpectraMax i3x Multi-Mode detection platform (Molecular Devices, CA, USA) was used to measure the luciferase activity. Firefly luciferase activity was normalized to Renilla luciferase activity and shown as relative luciferase units to reflect the promoter activity.

| ETS1 and SP1 constructs
The full-length cDNA sequence for human SP1 was obtained by PCR using primers (Table S1)

| Electrophoretic mobility shift assay (EMSA)
The EMSA was conducted using a chemiluminescence EMSA kit according to the manufacturer's instructions (Beyotime, Jiangsu, China). Briefly, nuclear extracts were prepared from Jurkat and NALM6 cells using the Nuclear and Cytoplasmic Protein Extraction CHEN ET AL.

| 2613
Kit following the manufacturer's instructions (Beyotime). The 5 0labelled biotin probes corresponding to the putative ETS1 and SP1 sites were synthesized and annealed (Beyotime , Table S1). For regular EMSA, 20 lg of nuclear extract was incubated with the biotinylated probes at 25°C for 30 minutes. To confirm the binding specificity, a 100-fold excess of unlabelled competitive probe (either a cold probe or a mutated cold probe) was used. To further determine the binding specificity, 4 lg of an antibody that recognized ETS1 or SP1 was also added to the reaction mixture and incubated at 25°C for 30 minutes according to the supershift assay protocol.

| Chromatin immunoprecipitation assay (ChIP)
The ChIP assay was performed with the EZ-Magna CHIP ™ A kit (Millipore, MA, USA) according to the manufacturer's instructions.
Briefly, two million Jurkat and NALM6 cells were harvested, crosslinked, lysed and sonicated (7-seconds bursts with a 50-seconds rest period at 20% power using the VibraCell ™ sonicator). A 1% aliquot of the sample was used as the input DNA control. The remaining cell lysates were incubated with an RNA POL II (Millipore), ETS1 (CST), SP1 (CST) or IgG antibody overnight with rotation. Then, the samples were precipitated using protein A magnetic beads (Millipore).
The chromatin-protein complexes were washed and eluted, and the cross-linking was subsequently reversed. After purification of the precipitated DNA, the samples were analysed using PCR. The PCR amplifications were performed with 30 cycles at 95°C for 30 seconds, 60°C for 30 seconds and 72°C for 30 seconds. An unrelated region of GAPDH was amplified to determine the binding specificity.
The DNA primers used for the PCR are listed in Table S1.

| Quantitative PCR
Total RNA was extracted using TRIzol (Invitrogen). The concentration and purity of the RNA were determined using a spectrophotometer (NanoDrop 1000, Thermo Scientific). A total of 2 lg of RNA was reverse-transcribed into cDNA using the PrimeScript RT Reagent Kit (Takara, Dalian, China) and the ABI2720 thermocycler (Applied Biosystems, USA). qPCR was performed with the ABI7500 real-time PCR system (Applied Biosystems) and FastStart Universal SYBR Green Master Mix (Roche). The primers used for quantification are listed in Table S1. The mean triplicate cT values of each cDNA sample were calculated and then normalized to GAPDH. The relative gene expression levels were obtained using the 2 ÀDDCt method as previously described. 29 2.9 | Western blotting  CpG sites from À154 to +96 nt, was amplified through two rounds of PCR using primers (Table S1)

| Statistical analysis
The patient samples were stratified into low and high DHX15 expression groups based on the median value. The statistical analyses were performed with analysis of variance (ANOVA) and the t test in GraphPad Prism 6.02 (GraphPad Software, La Jolla, CA, USA). The Spearman's rho test was used to evaluate the correlation between DHX15 and ETS1 or SP1 expression in the clinical samples.
The chi-square test was used to evaluate the correlations between the clinical characteristics and the DHX15 expression level. P values < .05 (two-sided) were considered significant.

| RESULTS
3.1 | The DHX15 core promoter is located within 501-bp upstream of the transcription start site To identify the DHX15 core promoter, we analysed the 2010-bp region (À1854 to +156 bp) around the TSS (+1) ( Figure 1A). Luciferase experiments suggested that transfection of the two constructs (pGL4.10-1854 and pGL4.10-345) into Jurkat cells resulted in stronger luciferase activity than transfection with pGL4.10 alone (both P < .0001) ( Figure 1B). The difference in luciferase activity between these two constructs was not significant (P > .05) ( Figure 1B). Additionally, we transfected the pGL4.10-1854 and pGL4.10-345 constructs into Jurkat and NALM6 cells. No significant difference in promoter activity was observed between the two constructs in either cell line ( Figure 1C). The results suggested that the DHX15 promoter region spanning from À345 to +156 bp contained the fulllength promoter and was crucial for control of basal DHX15 expression. Thus, we selected the 501-bp region for further study.
3.2 | CpG islands and functional ETS1-and SP1binding sites are present in the DHX15 promoter Comparative analysis using three different software programs (Matinspector, Promo Alggen and TFbind) revealed putative binding sites for several transcription factors in the DHX15 core promoter region (Figure 2A). We placed emphasis on the ETS1 and SP1 transcription factor binding sites for the following reasons. The ETS1 and SP1 transcription factors drive genes that contribute to proliferation and differentiation. SP1 interacts with TATA-binding proteinassociated factors (TAFs), which are essential for transcription. 30 Additionally, SP1 has been implicated in epigenetic regulation 31 and chemosensitivity 32,33 in ALL. An ETS1-binding site was also revealed.
ETS1 has been implicated in the pathogenesis of ALL. 34 Site-directed mutagenesis suggested that the ETS1 and SP1 transcription factors were essential for DHX15 promoter activity. A mutation that altered two bases of the ETS1-binding site reduced the DHX15 promoter activity to 69% (P < .0001) ( Figure 2B). Simultaneous mutations of two sites affecting the SP1-binding site reduced the promoter activity to 62% (P < .0001) ( Figure 2B). In addition, mutations in the binding sites for both ETS1 and SP1 (designated the pGL4.10-double mutant) reduced the promoter activity to 34% ( Figure 2B), indicating that these transcription factors had an additive effect. The ChIP assay revealed that RNA polymerase II bound to the DHX15 promoter (À303 to À165) ( Figure 2C). CpG islands often contain potential core promoter elements, [35][36][37] and CpG islands were identified in the DHX15 core promoter using MethPrimer (Figure 2A). 38 Taken To analyse recruitment of ETS1 and SP1 to the DHX15 core promoter, we performed a ChIP assay followed by PCR. A clear band was observed in the Jurkat and NALM6 cells, whereas no band was detected in the negative control ( Figure 3B). The results indicated in vivo occupancy of the DHX15 core promoter by ETS1 and SP1.
The ChIP and EMSA analysis results support the hypothesis that ETS1 and SP1 bind directly to the DHX15 promoter. Taken together, these results indicate that the DHX15 promoter is a eukaryotic promoter harbouring classical promoter elements, such as CpG islands and ETS1-and SP1-binding sites, which together contribute to transcriptional activation of the DHX15 gene. siRNA-mediated knockdown of ETS1 or SP1 reduced the DHX15 mRNA ( Figure 4A, C) and, consequently, the protein levels (Figure 4E, G). The DHX15 mRNA and protein levels were further reduced by simultaneous knockdown of ETS1 and SP1. In contrast, overexpression of ETS1 or SP1 increased the DHX15 mRNA levels ( Figure 4B, D), and consequently, the DHX15 protein levels (Figure 4F, H) in the Jurkat and NALM6 cells. Overexpression of ETS1 and SP1 further enhanced the DHX15 mRNA and protein levels.
Taken together, the current data suggested that ETS1 and SP1 transcriptionally drove DHX15 expression through their occupancy on the DHX15 promoter.

DHX15 expression in ALL
To assess the correlation between ETS1 and SP1 and DHX15 in a clinical setting, DHX15, ETS1 and SP1 expression was investigated using qPCR. Utilizing Spearman's rho, we observed a positive correlation between DHX15 and either ETS1 (r = .5144) or SP1 (r = .7388), indicating that these two variables were statistically F I G U R E 2 A, The genomic sequence of the DHX15 core promoter is shown. The sequence spans nucleotides À345 to +156 upstream of the DHX15 gene. The DHX15 promoter harbours binding sites for several transcription factors, which are shown underlined, based on predictions from in silico programs. Thirty CpG sites (indicated in red) are present in this region. B, Luciferase assays using constructs with mutations in the 501-bp region with the predicted transcription binding sites in Jurkat cells. The results represent relative firefly/Renilla luciferase activities, with the activity of the WT 501-bp region considered 100%. The values are expressed as the means AE SDs from three independent experiments. C, Jurkat and NALM6 cell chromatin was immunoprecipitated with an RNA Pol II antibody. Reactions with nonimmune IgG, no antibody and input DNA served as the negative and positive controls. After removal of the cross-links, the immunoprecipitated DNA was PCR-amplified using a primer flanking the basal DHX15 promoter region from À303 to À159 bp. The PCR products were subjected to agarose gel electrophoresis dependent on one another ( Figure 5A). We also investigated the biological relevance of ETS1 and SP1 for DHX15. Patients were stratified into groups with low or high DHX15 expression based on the median cut-off value. Remarkably, we observed significant positive correlations of the transcription factor levels with DHX15; specifically, patients expressing low DHX15 levels also showed low expression levels of either ETS1 or SP1, whereas patients with high DHX15 expression demonstrated high levels of ETS1 and SP1 (both P < .0001) ( Figure 5B). These data further indicated that ETS1 and SP1 regulated DHX15 expression in human ALL. Additionally, correlations of clinical characteristics and the DHX15 levels were assessed in the patients with ALL. The DHX15 expression levels were correlated with peripheral blood blasts (Table S2) 4). B, Equal amounts of Jurkat and NALM6 chromatin were immunoprecipitated with antibodies for ETS1 and SP1 and subsequently quantified through agarose gel electrophoresis using a primer set specific for the basal region (À181 to À36 bp). Moreover, immunoprecipitated DNA was amplified using a primer set specific to the off-target region (GAPDH) shown in the lower panel as a negative control CpG islands. 35 The results of our in silico analysis showed the presence of CpG islands and ETS1-and SP1-binding sites in this region.
The core promoter is the site of action of the RNA polymerase II transcriptional machinery. 35 Our ChIP assay showed that RNA polymerase II bound to the DHX15 promoter (À345 to +156). These findings strongly suggested that the core promoter region of DHX15 High hTERT expression has been observed in ALL, 46 48 These studies show that SP1 is an important mediator that exerts its effects in ALL via downstream signalling molecules.
These findings prompted an investigation of the functional relevance of the binding sites for these transcription factors on the DHX15 promoter. Site-directed mutagenesis studies showed that ETS1 and SP1 were putative transcriptional regulators of the DHX15 promoter. The EMSA analysis results confirmed the in vitro interaction of these transcription factors with the DHX15 promoter. Using the ChIP assay, we confirmed that these interactions occurred in vivo. Overexpression and RNAi studies confirmed that the DHX15 gene was transcriptionally regulated by ETS1 and SP1. Dysregulation of DHX15 as an interacting partner with other proteins has been studied in breast cancer 19 and prostate cancer. 49 Here, we contribute to the understanding of the protein-DNA interactions of the ETS1 and SP1 transcription factors with the DHX15 gene promoter.

F I G U R E 4
Influence of ETS1 and SP1 on DHX15 gene transcription and protein expression in Jurkat and NALM 6 cells. A, C, E, G, Knockdown of endogenous ETS1 or SP1 or ETS1 and SP1 together decreased DHX15 gene transcription and protein expression. Jurkat and NALM6 cells were transfected with 100 pmol of siRNAs targeting ETS1, SP1 or ETS1 and SP1 together or a negative control (NC). The cells were harvested 48 h after transfection, and 2 lg of the total RNA was used to detect the DHX15 mRNA level through qPCR. The relative mRNA level was obtained after comparison with the NC, which was set to 1 A, C. Western blotting analysis of total proteins with anti-ETS1, anti-SP1 and anti-DHX15 antibodies was performed for the Jurkat and NALM6 cells; anti-b-actin served as a loading control E, G. B, D, F, H, Overexpression (OE) of ETS1 or SP1 or ETS1 and SP1 together increased DHX15 gene transcription and protein expression. Jurkat cells were transfected with 4 lg of pcDNA3.1(-)/SP1, pEnter-ETS1, pcDNA3.1(-)/SP1 and pEnter-ETS1 together or the empty control pcDNA3.1(-) pEnter. The cells were harvested 48 h after transfection, and 2 lg of total RNA was used to detect the DHX15 mRNA level through qPCR. The relative mRNA level was obtained after comparison with the empty vector, which was set to 1 B, D. Western blotting analysis of total proteins with anti-ETS1, anti-SP1 and anti-DHX15; anti-b-actin served as a loading control F, H We assessed the association of ETS1 and SP1 with DHX15 in clinical samples. A positive correlation between DHX15 and ETS1 and between DHX15 and SP1 was observed in the patients with ALL. Patients with high DHX15 expression also had high ETS1 and SP1 expression levels. These results further support the notion that ETS1 and SP1 regulate DHX15 expression. Additionally, the correlation between peripheral blood blasts and the DHX15 expression level suggests that the DHX15 expression level possibly predicts the leukaemia burden in the peripheral blood. BSP showed that the DHX15 core promoter was hypomethylated in the patients with ALL and in healthy controls. SP1 maintains the hypomethylated status of CpG islands. 36 Therefore, we postulated that the presence of a SP1binding site in the DHX15 core promoter contributed to maintenance of the hypomethylated status of the DHX15 core promoter and active DHX15 expression.
Our present study is the first to report that the human RNA helicase DHX15 may be transcriptionally regulated in ALL by the ubiquitous transcription factors ETS1 and SP1. The two transcription factors are particularly promising therapeutic targets because these proteins mediate signals from multiple pathways. F I G U R E 5 Correlation of the DHX15 levels with the ETS1 and SP1 levels in 121 ALL peripheral blood mononuclear cell (PBMC) samples. A, The qPCR results were evaluated for correlations using Spearman's correlation coefficient, and the correlation coefficient "r" was calculated. B, Box plot analyses comparing the ETS1 and SP1 levels between samples with low and high DHX15 expression levels. All qPCR results were normalized to GAPDH. The samples were divided into low and high DHX15 expression groups based on the median value