Page 127 - 2018_12-Haematologica-web

BET inhibition modulates miRNAs in DLBCL
of the miRNAs were obtained using the MicroRNA Target Filter in Ingenuity Pathway Analysis (Qiagen). Functional annotation of the targets was performed with the Gene Set Enrichment Analysis tool for overlap analysis using the hallmarks and the c2.cp of the Molecular Signatures Database 5.232 and hypergeometric P-values after correction for multiple hypothesis testing according to Benjamini and Hochberg.
Publicly available chromatin immunoprecipitation (ChIP) sequencing datasets obtained by ChIP followed by high-through- put DNA sequencing were downloaded and re-analyzed. They comprised datasets obtained in the ABC-DLBCL cell line HBL-1 (SRP043524)23 and in the GCB-DLBCL cell line OCI-LY-1 (SRP022129),22 both treated with the BET inhibitor JQ1 or DMSO alone. Sequence reads obtained from ChIP fragments were aligned to the human reference genome hg19 using Bowtie, allowing up to one mismatch per fragment length. Redundant reads were removed and only reads uniquely mapping to the reference genome were used for further analysis. The peaks that were genomic regions enriched by ChIP, relative to the background reads, were detected using HOMER (v2.6), a suite of tools for Motif Discovery and next-generation sequencing analysis, with a default option (false discovery rate = 0.001 and Poisson P-value cut-off = 1e-04). Differential peaks were defined as having at least a four-fold difference in enrichment within a 200 bp region between the two conditions (DMSO versus JQ1) and a Poisson enrichment P-value less than 1e-04. All discovered putative peaks were ranked by their normalized tag counts (number of tags found at the peak, normalized to 10x106 total mapped tags) and annotat- ed with the annotatePeaks.pl subroutine. We defined miRNA pro- moters using FANTOM533 and the precursors of microRNAs downloaded from miRBase (v20)34 to annotate the BRD4 ChIP sequencing datasets. We defined enriched regions located within 5 kb regions of predicted promoters and pre-miRNAs as candidate BRD4 binding sites. For global ChIP sequencing visualization, we used ngs.plot (https://code.google.com/p/ngsplot/) for inspection of both average and ‘laid out’ coverages as curves or heatmaps.
Chromatin immunoprecipitation
Cells (SU-DHL-2 and HBL-1) were cross-linked with 1% formaldehyde. Crosslinking was quenched with 125 mM glycine. Cells were washed with ice-cold phosphate-buffered saline con- taining 1 x HALT protease inhibitor (Thermo Scientific, Lausanne, Switzerland) and resuspended in sodium dodecylsulfate lysis buffer (ChIP Assay Kit, Millipore, Schaffhausen, Switzerland) before sonication using the Bioruptor Plus. For each immunopre- cipitation reaction, chromatin from 1x106 cells was incubated overnight with anti-PRMT5 (A1520; NeoBiolab), anti-BRD4 (A301-985A; Bethyl) or 3 mg of the negative control antibody, anti- IgG (Millipore). Immune complexes were collected by incubation with 20 mL magnetic protein G beads at 4°C for 1.5 h. Protein G- bound complexes were sequentially washed with Low Salt Wash Buffer, High Salt Wash Buffer, LiCl Wash Buffer and twice with TE Buffer (ChIP Assay Kit, Millipore). Protein/DNA complexes were eluted using 1% sodium dodecylsulfate and 0.1 M NaHCO3. Following reversal of crosslinks (65°C overnight), samples were treated with RNAse A and then Proteinase K. DNA samples were purified using the QIAquick PCR purification kit (Qiagen, Hombrechtikon, Switzerland). Chromatin samples to which no antibody had been added were processed in parallel as input ref- erences. For qPCR analysis of ChIP samples, triplicate wells con- taining 1 mL of purified ChIP DNA plus PCR master mix were pre- pared. Reactions were performed on a StepOnePlus Real-Time PCR system (Applied Biosystems). Standard curves were con- structed using sonicated and purified chromatin. ChIP-qPCR was performed using primers specific for the upstream regulatory
regions of PRMT5 and miR-96-5p. Primer sequences were as fol- lows: PRMT5 forward; 5’-AGCGCGAGGAGAAAGATG-3’, PRMT5 reverse; 5’-CTATTTCGGGGACGCAATTC-3’, miR-96 forward; 5’-AGCTGGGAGACCTTGCTTC-3’, miR-96 reverse; 5’-TCACCCCTCCTAACCCAAAT-3’.
Results
BET inhibition modulates the expression of a subset of microRNAs
We have previously shown that the BET inhibitor OTX015 modulates the expression of multiple coding transcripts in DLBCL cells.15 Here we assessed the effect of OTX015 on global miRNA expression. GCB-DLBCL OCI- LY-1 and ABC-DLBCL HBL-1 cells were treated with 500 nM OTX015 for 4 and 24 h. Total RNA isolated from vehi- cle- and OTX015-treated cells was interrogated with the Nanostring nCounter. Fourteen miRNAs were modulated (5 downregulated, 9 upregulated) by the BET inhibitor in OCI-LY-1 cells and 11 (5 downregulated, 6 upregulated) in HBL-1 cells (Table 1).
Additionally, we used the Agilent Human miRNA microarray v.3. platform to perform miRNA profiling on two more DLBCL cell lines, DOHH-2 (GCB-DLBCL) and SU-DHL-2 (ABC-DLBCL), the same cell lines we had pre- viously used for mRNA profiling of OTX015-treated cells.15 In this case, seven miRNAs (3 downregulated, 4 upregulated) were affected by OTX015 in the GCB- DLBCL, and five (2 downregulated, 3 upregulated) in the ABC-DLBCL cell line (Table 1).
A few miRNAs were affected in more than one cell line, although we could not determine clear subtype-specific differences in miRNA modulation since only one GCB- and one ABC-DLBCL cell line were interrogated on each profiling platform. The oncogenic miR-92a-1-5p,35 belong- ing to the miR-17-92 cluster, was downregulated in three of four cell lines (2 ABC-DLBCL, 1 GCB-DLBCL). miR- 204-5p, involved in BRAF resistance in melanoma,36 was downregulated and miR-487b-3p, expressed at lower lev- els in DLBCL versus follicular lymphoma,37 was upregulat- ed in both cell lines analyzed with the Nanostring plat- form. The tumor suppressor miR-96-5p38,39 was upregulat- ed in HBL-1 and DOHH-2 cells. Besides these, among the miRNAs modulated by the BET inhibitor there were oth- ers known to be involved in lymphomagenesis. The onco- genic miRNAs hsa-miR-21-3p40-44 and miR-15545,46 were downregulated, while, besides miR-96-5p, another miRNA with a tumor suppressor function, miR-16-5p,47 was also upregulated by the BET inhibitor. qRT-PCR was used to validate the expression of two lymphoma oncomiRNAs modulated by BET inhibition: miR-155-5p and miR-92a-1-5p (Online Supplementary Figure S1). The latter also appeared significantly downregulated after in vivo treatment of SU-DHL-2 xenografts (Online Supplementary Figure S2).
microRNAs modulated by BET inhibition control important pathways in diffuse large B-cell lymphoma
Functional annotation analysis identified the p53 path- way, apoptosis, MYC-targets, cell cycle regulation, B-cell receptor signaling, interleukin-6 signaling, the STAT3 pathway, PI3K and nuclear factor-kB signaling among the biological processes significantly associated with the miRNAs that exhibited expression changes in HBL-1 and
haematologica | 2018; 103(12)
2051