SMARCB1 Deficiency Integrates Epigenetic Signals to Oncogenic Gene Expression Program Maintenance in Human Acute Myeloid Leukemia

Published OnlineFirst February 26, 2018; DOI: 10.1158/1541-7786.MCR-17-0493

Chromatin, Epigenetics, and RNA Regulation Molecular Cancer Research SMARCB1 Deﬁciency Integrates Epigenetic Signals to Oncogenic Gene Expression Program Maintenance in Human Acute Myeloid Leukemia Shankha Subhra Chatterjee1, Mayukh Biswas1, Liberalis Debraj Boila1, Debasis Banerjee2, and Amitava Sengupta1

Abstract

SWI/SNF is an evolutionarily conserved multi-subunit chroma- AML blasts, and loss-of-function studies confirmed transcriptional tin remodeling complex that regulates epigenetic architecture and regulation of Rac GTPase guanine nucleotide exchange factors cellular identity. Although SWI/SNF genes are altered in approx- (GEF) by SMARCB1. Mechanistically, loss of SMARCB1 increased D imately 25% of human malignancies, evidences showing their recruitment of SWI/SNF and associated histone acetyltransferases involvement in tumor cell–autonomous chromatin regulation (HAT) to target loci, thereby promoting H3K27Ac and gene and transcriptional plasticity are limiting. This study demonstrates expression. Together, SMARCB1 deficiency induced GEFs for Rac that human primary acute myeloid leukemia (AML) cells exhibit GTPase activation and augmented AML cell migration and sur- near complete loss of SMARCB1 (BAF47 or SNF5/INI1) and vival. Collectively, these findings highlight tumor suppressor role D D SMARCD2 (BAF60B) associated with nucleation of SWI/SNF . of SMARCB1 and illustrate SWI/SNF function in maintaining an D SMARCC1 (BAF155), an intact core component of SWI/SNF , oncogenic gene expression program in AML. colocalized with H3K27Ac to target oncogenic loci in primary AML cells. Interestingly, gene ontology (GO) term and pathway Implications: Loss of SMARCB1 in AML associates with SWI/ D analysis suggested that SMARCC1 occupancy was enriched on SNF nucleation, which in turn promotes Rac GTPase GEF expres- genes regulating Rac GTPase activation, cell trafficking, and AML- sion, Rac activation, migration, and survival of AML cells, D associated transcriptional dysregulation. Transcriptome profiling highlighting SWI/SNF downstream signaling as important revealed that expression of these genes is upregulated in primary molecular regulator in AML. Mol Cancer Res; 1–14. 2018 AACR.

Introduction not always inform transcriptional dependencies embedded in tumorigenesis. SWI/SNF (BAF) chromatin remodelers are evolutionarily con- Emerging evidences indicate that SWI/SNF subunits critically served, large (2 MDa) multi-protein complexes, which utilize regulate murine hematopoiesis. Recent studies have shown that energy derived from ATP hydrolysis to mobilize nucleosomes (1). SMARCD2 mediates granulopoiesis through CEBPe-dependent SWI/SNF core components include SMARCB1 (BAF47, SNF5 or mechanism (5, 6). Actl6a (BAF53a) plays essential role in hemato- INI1), SMARCC1/SMARCC2 (BAF155 and BAF170), and one of poietic stem cell (HSC) function (7). Mutant allele of Arid1a the mutually exclusive ATPase subunits, SMARCA4 (BRG1) and (BAF250a) determines pool size of fetal liver HSC populations SMARCA2 (BRM). SWI/SNF complexes often include cell type– (8). In addition, SWI/SNF was also implicated in murine leuke- specific, lineage-restricted subunits, and play important roles in mia development. SMARCA4 was shown to regulate proliferation pluripotency and cellular reprogramming (1, 2). Cancer genome of murine leukemic cells (9, 10). SMARCB1 plays tumor sup- sequencing studies have identified SWI/SNF complexes as one of pressor role in several cancers, and frequent deletion of SMARCB1 the most commonly mutated (25%) chromatin modulators in is observed in chronic myeloid leukemia patients (11). Loss of human cancer (3, 4). However, mutational profiling alone may Smarcb1 in vivo leads to fully penetrant malignant rhabdoid tumors (12, 13). Rac GTPases belong to small Rho GTPase family and are 1 Stem Cell & Leukemia Lab, Cancer Biology & Inflammatory Disorder Division, involved in regulation of a diverse array of cellular functions CSIR-Indian Institute of Chemical Biology, Translational Research Unit of including cell proliferation, survival, adhesion, migration, actin Excellence (TRUE), Salt Lake, Kolkata, West Bengal, India. 2Clinical Hematology, Park Clinic, Gorky Terrace, Kolkata, West Bengal, India. assembly, and transcriptional activation (14, 15). Similar to Ras superfamily proteins, Rac GTPases cycle between inactive GDP- Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/). bound and active GTP-bound conformations, regulated by specific guanine nucleotide exchange factors (GEF), to transduce S.S. Chatterjee and M. Biswas contributed equally to this article. signals to effector proteins (14). Recent studies have suggested Corresponding Author: Amitava Sengupta, CSIR-Indian Institute of Chemical that Rac GTPases play integral roles in myeloid leukemia cell Biology, Kolkata 700091, India. Phone: 9133-2473-0492; Fax: 9133-2473-5197; homing, engraftment, survival, and trafficking within the bone E-mail: [email protected] marrow microenvironment (15–18). Attenuation of Rac GTPase doi: 10.1158/1541-7786.MCR-17-0493 signaling in synergy with Bcl-2 inhibition has been shown as a 2018 American Association for Cancer Research. modality for combination targeted therapy in MLL-AF9 leukemia

www.aacrjournals.org OF1

Chatterjee et al.

(19). Myeloid leukemia cells are characterized with elevated Rac and the integrity was assessed on a 0.8% Agarose Gel. Genomic GTP level; however, molecular regulation of Rac activation in DNA with OD260/OD280 >1.8 and OD260/OD230 1.3 was leukemia pathophysiology remains incompletely understood. used for microarray experiments. DNA was considered to be of Here we identify that in human primary acute myeloid good quality when a single clear band was seen when run against a leukemia (AML) cells, SMARCB1 deficiency associates with reference ladder. A total of 0.5 mg of DNA in 10.1 mL was taken into D nucleation of SWI/SNF . SMARCC1, an intact core component a microfuge tube and digestion master mix containing restriction D of SWI/SNF , colocalized with H3K27Ac to target tumor onco- enzymes (Alu I, 5U and Rsa I, 5U) was added. The samples were genic loci, including Rac GTPase GEFs,inAMLcells.Lossof incubated at 37C for 2 hours followed by heat inactivation of D SMARCB1 induced recruitment of SWI/SNF and associated enzymes at 65C for 20 minutes. Samples were labeled using histone acetyltransferases (HAT) to target GEFs for Rac GTPase Agilent sure tag DNA Labeling Kit (catalog no: 5190-3399). activation and promoted AML cell migration. Collectively, Control samples were labeled with Cy3 and test sample with these findings highlight tumor suppressor role of SMARCB1 Cy5. The labeled samples were cleaned up using Amicon Ultra D and illustrate SWI/SNF function in maintaining an oncogenic columns 30-kDa size exclusion filter. DNA yield and incorpo- gene expression program in AML. ration of labeled dye (specific activity) was measured using NanoDrop spectrophotometer. One micrograms each of Cy3- Materials and Methods and Cy5-labeled sample was added with human cot-1 DNA (catalog no. 5190-3393), Agilent aCGH/CoC Blocking agent (part Patient cohort number: 5188-6416), and hybridization buffer (part number: Human AML (n ¼ 67) bone marrow aspirates (1–2 mL each) 5188-6420). The labeled samples in above hybridization mix were obtained from Park Clinic, Kolkata from untreated, freshly were denatured at 95C for 3 minutes and were incubated at 37C diagnosed patients after written, informed consent according to for 30 minutes. The samples were then hybridized at 65C for 24 Institutional Human Ethics Committee (HEC) approval and hours. After hybridization, the slides were washed using Agilent following Indian Institute of Chemical Biology (CSIR-IICB) Insti- aCGH Wash Buffer1 (Agilent Technologies, part number 5188- tutional Review Board (IRB) set guidelines. Sample collection was 5221) at room temperature for 5 minutes and Agilent aCGH Wash part of routine diagnosis and the inclusion criterion for this study Buffer 2 (Agilent Technologies, part number 5188-5222) at 37C was histopathologic confirmation of bone marrow aspirates or for 1 minute. The slides were then washed with acetonitrile (part biopsies, karyotyping, and immunophenotypic analyses (20). number: A2094) for 10 seconds. The microarray slides were Bone marrow aspirates were also collected from age-matched scanned using Agilent Scanner (Agilent Technologies, part num- normal individuals (n ¼ 6) after informed consent, who turned ber G2600D). Image analysis was performed using Agilent Fea- out to be pathologically negative for AML (20). Individual case ture Extraction software, feature extracted raw data was analyzed information is presented in Supplementary Tables S1 and S2. using Agilent Genomic Workbench 7.0 software. The data were Umbilical cord blood samples (40 mL each) were obtained from normalized using Lowess normalization. Significant regions hav- Deb Shishu Nursing Home (Howrah, West Bengal, India) from ing amplification and deletions were identified among each of the term pregnancies after written, informed consent according to samples. Genomic view and chromosome view of the amplifica- CSIR-IICB HEC approval and following IRB set guidelines. Low tion and deletion region for each sample were generated. Graph- density (1.077 gm/cc) nuclear cells from AML bone marrow, ical representation has been done using Human UCSC Genome normal bone marrow, or cord blood samples were isolated by Browser by loading the data in wiggle file format. Details are Ficoll (Sigma) separation and cryopreserved in liquid nitrogen. included as Supplementary array CGH Files. Normal blood specimens were obtained from age-matched healthy volunteers (n ¼ 3) after written, informed consent and Methylation-specific PCR nucleated cells were isolated using RBC lysis (BD Pharmingen). Genomic DNA was isolated using the QIAamp DNA Blood Mini Kit (Qiagen, catalog no. 51104). Isolated genomic DNA was Reverse transcription and quantitative PCR then bisulfite converted, that is, unmethylated cytosines were Total RNA was isolated by using TRIzol (Life Technologies) converted to uracil using EpiTect Bisulfite Kit (Qiagen, catalog according to manufacturer's recommendation. RNase-free DNase no. 59104) according to the manufacturer's protocol. Methyla- treatment were carried out to remove any genomic DNA contam- tion-specific primers for the gene of interest were made using ination using DNase I recombinant, RNase free kit (Roche). RNA MethPrimer. Bisulfite converted DNAs were amplified using fi amount was quanti ed and cDNA was prepared using TaqMan methylated DNA (M pair)-specific primers. Methylation-specific Reverse Transcription Reagents (Applied Biosystems). Gene PCR at SMARCB1 promoter loci in primary AML blasts was expression levels were determined by quantitative PCR performed compared with normal BMNC. The fold change levels of the using cDNA with SYBR Select Master Mix (Applied Biosystems) on methylated DNA were calculated with respect to GAPDH (unre- the 7500 Fast Real-Time PCR System (Applied Biosystems). lated control). Relative methylation levels were plotted after GAPDH was used as a housekeeping gene. Relative expression DD normalizing it with GAPDH. qMSP primer details are available Ct – levels were calculated using the 2 method (21 24). qRT-PCR in Supplementary Table S3. primer details are available in Supplementary Table S3. Coimmunoprecipitation, histone acid extraction, and Array Comparative Genomic Hybridization and analysis immunoblotting Array Comparative Genomic Hybridization (CGH) and anal- Nuclear extracts for immunoprecipitation experiments were ysis were carried out by Genotypic Technology Private Limited. prepared using NE-PER Nuclear and Cytoplasmic Extraction Genomic DNA quantity and purity was assessed by the NanoDrop Reagents (Thermo Fisher Scientific) and diluted in 1 RIPA (Cell ND-2000 UV-Vis Spectrophotometer (NanoDrop Technologies) Signaling Technology) containing protease and phosphatase

OF2 Mol Cancer Res; 2018 Molecular Cancer Research

SWI/SNF Regulates Oncogenic Signaling in AML

inhibitor cocktails. About 300 mg extracts were incubated with 2.0 chromatin elution buffer at 65C with gentle vortexing. The eluted mg of respective antibodies against SMARCC1 (clone R-18, sc- chromatin was treated with RNAse for 30 minutes at 37C. Reverse 9746, Santa Cruz Biotechnology), p300 (A300-358A, Bethyl cross-linking was performed by treating the eluted chromatin with Laboratories), CBP (clone D6C5, 7389S, Cell Signaling Technol- Proteinase K (Sigma) at 65C for 2 hours. The DNA was finally ogy), BRD4 (clone E2A7X, 13440S, Cell Signaling Technology), or precipitated by phenol–chloroform extraction; precipitated DNA rabbit IgG (P120-101; Bethyl Laboratories), and incubated over- was dissolved in TE buffer, and subjected to ChIP-seq analyses. þ night at 4C with gentle rocking. Fifty microliters of protein A/G For ChIP-qPCR experiments, 4–5 106 normal CD34 cells or agarose beads (Cell Signaling Technology) were added and incu- 293T cells and 2 mg of antibody, or ChIP DNA obtained from bated for 3–4 hours at room temperature. The beads were then primary AML BMNCs were used. ChIP-qPCR primer details are washed 6 times with 1 RIPA supplemented with 300 mmol/L available in Supplementary Table S5. NaCl and resuspended in 1 SDS gel loading buffer. The proteins Size distribution of the ChIP-enriched DNA was checked using were separated in SDS-PAGE and transferred to PVDF membrane High Sensitivity chips in 2100 Bioanalyzer (Agilent) for each (Millipore) and subsequently probed with respective antibodies. sample and quantitation was performed in Qubit Fluorometer Detailed list of antibody is included in Supplementary Table S4. (Invitrogen) by picogreen method. ChIP-seq library preparation All antibodies were used at a dilution of 1:1,000 unless specified was performed using TruSeqChIP Sample Prep Kit (Illumina) otherwise. Total cell lysate for immunoblotting was prepared by according to the manufacturer's instructions. Ten nanograms of incubating cells in 1 RIPA for 15 minutes followed by brief input ChIP-enriched DNA was used for ChIP-seq library prepa- sonication. Supernatants were collected following centrifugation ration. Final libraries were checked using High Sensitivity chips in at 16,000 g for 15 minutes at 4C. Protein concentration was 2100 Bioanalyzer (Agilent). Average fragment size of final librar- determined using Pierce BCA Protein assay kit (Thermo Fisher ies was found to be 280 8 bp. Paired-end sequencing (2 100 Scientific). SDS-PAGE was used to separate proteins, transferred to bp) of these libraries were performed in HiSeq-2500 (Illumina). PVDF membrane, and probed using respective antibodies. Den- Quality control analysis of the raw data using NGS-QC ToolKit sitometry analyses were performed using NIH Image J software. was done and HQ reads with filter criteria of bases having 20 Phred score and reads with 70% were filtered. Paired-end reads Sucrose density gradient centrifugation (.fastq format) were aligned with Bowtie software using –best and A total of 700 mg–1.0 mg nuclear extracts isolated from pooled –m 2, that is, mismatches against reference genome Ensembl (n ¼ 5–7) primary AML BMNCs were prepared and diluted in 300 build GrCh37/hg19 (considering 2% input as the baseline) and mLof1 RIPA. The extracts were overlaid on a 10 mL 20%–50% saved in SAM format, which was then converted to sorted BAM file sucrose gradient (in 1 RIPA) in 13 89 mm polyallomer tubes using SAMTOOLS. PCR duplicates were removed using SAM- (Beckman Coulter). The tubes were then centrifuged in a SW-41 Ti TOOLS rmdup. Peak calling was performed using MACS14 model swing out rotor at 30,000 rpm for 12 hours at 4 C. A total of 500- building with P value cutoff of 0.05. Annotation of the identified mL fractions were collected and separated in SDS-PAGE, trans- peaks was performed with PeakAnalyzer. ferred to a PVDF membrane, and subsequently probed with Functional enrichment analysis (Gene Ontology and Pathway) specific antibodies. was done using The Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.8. Gene list were uploaded and ChIP-seq, ChIP-qPCR, and analyses converted to its respective ENTREZ Gene ID. The converted gene ChIP-seq experiments were carried out at Core Technologies list was submitted to DAVID and Functional Annotation Clus- Research Initiative (CoTeRI), National Institute of Biomedical tering was carried out which comprises of Gene Ontology and Genomics (NIBMG, University of Kalyani, Kolkata, West Bengal, Pathway analysis. R bioconductor package ChipSeeker was used India). A total of 1 107 primary bone marrow nuclear cells to generate heatmap, average profile distibutions, and pie charts. (BMNC), isolated from three independent (biological replicates) Unique gene names were used to plot the Venn diagram repre- AML patients (Supplementary Table S1 case nos. 82, 83, and 80 as sented by peaks in the respective samples either upstream or AML 01, AML 02, and AML 03, respectively), per chromatin downstream or overlap to the genetic region. Bigwig/bed files immunoprecipitation (ChIP) set were crosslinked with formal- were imported into Integrated Genome Viewer (IGV) and snap- dehyde (Sigma) in culture media. After crosslinking, chromatin shots of particular genomic loci were captured. For motif analysis, was extracted and sonicated to fragment lengths between 150 and the MEME ChIP V4.10 was used to analyze the motif using the 900 bp in chromatin extraction buffer containing 10 mmol/L Tris peak sequences, with default parameters. and for transcription pH ¼ 8.0, 1 mmol/L EDTA pH ¼ 8.0, 0.5 mmol/L EGTA pH ¼ 8.0. encoding motif, Jaspar database (Jaspar) was used using MEME Chromatin was incubated with ChIP-grade antibodies to ChIP tool. Details are included as Supplementary ChIP-seq Files. SMARCC1 [sc-9746 (R-18), Santa Cruz Biotechnology), H3K27Ac (ab4729, Abcam), H3K27Me3 (07-449, Millipore), or RNA-seq and analyses rabbit IgG (clone P120-101; Bethyl Laboratories) or mouse IgG RNA-seq experiments were performed by Bionivid Technology. (clone G3A1; 5415S; Cell Signaling Technology) or goat IgG (sc- Total RNA was isolated from BMNCs from the identical AML 2028, Santa Cruz Biotechnology) during overnight at 4C with cohort (n ¼ 3) and age-matched normal (n ¼ 2) hematopoietic rotation. Six micrograms of antibody was used per immunopre- cells using TRIzol (Life Technologies) according to manufacturer's cipitation. All antibodies were used at 1:1,000 dilution. Detailed instruction. DNase treatment was carried out to remove any list of antibody is included in Supplementary Table S4. Protein genomic DNA contamination using DNase I recombinant, RNase A/G Agarose beads (Cell Signaling Technology) were then added free kit (Roche). RNA amount was quantified and the sequencing and incubated for 2 hours at 4C. The beads were washed with library prepared using TruSeq RNA Sample Prep Kit v2 (Illumina). chromatin extraction buffer and by increasing salt concentration Paired-end sequencing was performed on HiSeq 4000 using for four times. The chromatin was eluted from the beads in TruSeq 3000 4000 SBS Kit v3 (Illumina).

www.aacrjournals.org Mol Cancer Res; 2018 OF3

Chatterjee et al.

Raw data resulted in an average of 35.39 106 reads in normal Cells were seeded in T-225 flasks at 70% confluence and trans- hematopoietic cells and 36.86 106 reads of 101-bp length in fected with the target plasmid DNA, PAX2 (Addgene; 12260), and primary AML cells. Quality control using NGSQC toolkit yielded pMD2.G (Addgene; 12259) using calcium-phosphate transfection around 34.28 106 HQ reads in normal hematopoietic cells and method (23, 25). After overnight incubation, butyrate induction 36.05 106 HQ reads in primary AML cells. Around 88.2% of the was given for 8 hours. Supernatant containing lentiviral particles HQ reads from normal and 85.12% HQ reads from primary AMLs were collected after 36–40 hours of incubation at 37C with 5% could be mapped to the Homo sapiens (hg 38) genome reference CO2, and ultracentrifuged at 25,000 rpm for 90 minutes at 4 C sequence using TOPHAT, suggesting a good quality of RNA using (Sorvall, Thermo Scientific). Virus pellet was resuspended in sequencing. Transcripts were given a score for their expression X-VIVO (Lonza) and aliquoted in several tubes and stored at by Cufflinks-based maximum likelihood method. A total of 80C. HL60 and U937 cell lines were stably transduced with 24,784 transcripts were identified using Cuffdiff validation to be lentiviral particles containing mock (sh-Control) or sh-SMARCB1 expressed in either normal or primary AML cells representing plasmids in a U-bottom 96-well nontissue-culture–treated plate. 13,847 genes. Transcript type analysis revealed 95.5% of the A total of 1 105 cells were incubated overnight with virus transcripts to be of "Full Length" or "Known Transcripts" and particles at a MoI of 5 and polybrene was added at a final 4.5% as "Potentially Novel Isoforms" Transcripts as per Cufflinks concentration of 8 mg/mL to initiate lentiviral infection. Class Code distribution. This indicates a largely complete transcription machinery activity in both normal and AML cells. Cells, drug treatments and survival, and proliferation assays Significant Biology for Differentially Expressed Transcripts was Human AML cell lines (obtained from Dr. Jose Cancelas, performed with GO-Elite_v.1.2.5 Software. A cutoff of P value less Cincinnati Children's Hospital, Cincinnati, OH) were maintained than 0.05 was considered for filtering the significantly enriched in Iscove's modified Dulbecco's medium (IMDM) supplemented GO pathways. In testing for differential expression, we consider with 10% FBS, 100 U/mL penicillin, 100 mg/mL streptomycin, L log2FC > 2 (upregulation) and log2FC 2 (downregulation). For and 2 mmol/L -glutamine (all from Gibco) at 37 C with 5% CO2. gene-set enrichment analysis (GSEA), differentially expressed 293T cells (obtained from Dr. Jose Cancelas, Cincinnati Chil- genes from individual comparisons were preranked on the basis dren's Hospital, Cincinnati, OH; refs. 15, 23) were maintained in of fold change such that maximally upregulated genes fall top- DMEM supplemented with 10% FBS, 100 U/mL penicillin, 100 most in the list. This was used as an input to perform GSEA mg/mL streptomycin, and 2 mmol/L L-glutamine (all from Gibco) (GeneSpring). GSEA was performed on "H: Hallmark gene set" at 37 C with 5% CO2. Adherent cells were transfected at 70% representing well defined biological states or processes available confluency using the calcium phosphate transfection method (16, on Molecular Signature Database. Details are included as Sup- 22, 23, 25). Cell lines have been freshly authenticated using STR plementary RNA-seq Files. profiling (Supplementary Cell Lines_STR Profiles). JQ1 (cat SML0974) was purchased from Sigma. For calcu- Gene enrichment and functional annotation analysis lating IC50, parental or lentivirus transduced AML cell lines Gene ontology (GO) analysis of the shared gene set (Supple- were treated with varying doses of JQ1 from 0.1 to 50 mmol/L. mentary ChIP-seq Shared GENELIST) was carried out using Viable cell counts were taken after 48 hours of drug treatment þ DAVID v6.8 (https://david.ncifcrf.gov/). The P value used in the and the total number of GFP cells were analyzed by flow analysis is a modified one, termed as EASE score threshold cytometry. Cell counts were normalized and plotted against (maximum probability). The threshold of EASE Score is a mod- logarithm of the inhibitor concentration using GraphPad fi i ed Fisher exact P value used for gene enrichment analysis. It Prism5 to measure the IC50. For proliferation assay, lentivirus ranges from 0 to 1. Fisher exact P value ¼ 0 represents perfect transduced cell lines were grown in triplicate in regular media enrichment. Usually P value is equal or smaller than 0.05 to be for 6 days. Trypan blue–negative cell numbers were determined þ considered strongly enriched in the annotation categories. at respective time points and the total number of GFP cells were analyzed by flow cytometry. Plasmids AML BMNCs were grown in regular media supplemented with shRNA-expressing lentiviral constructs targeting against cytokines in presence of 500 nmol/L 5-azacytidine (Sigma, catalog SMARCB1 (pLKO-shSMARCB1, 39587) and SMARCB1-overex- no. A1287) or DMSO (vehicle) for 72 hours. Media was calibrated pression vector HA_INI1/BAF47 was a gift from Dr. Olivier with fresh 5-azacytidine and cytokines after every 24 hours. Delattre (Institut Curie, Paris, France). shRNA-expressing lenti- Posttreatment, total RNA was isolated and gene expression levels viral construct targeting against SMARCB1 (pLKO.1-puro-CMV- were determined by qRT-PCR. TGFP, TRCN0000295966) was purchased from Sigma. Empty vector SHC003 was purchased from Sigma. Lentiviral packaging PAK1 pulldown and Rac GTPase activation assay constructs PAX2 (Addgene; 12260) and pMD2.G (Addgene; Cells were lysed using 400 mL of MLB buffer (1) with repeated 12259) were purchased from Addgene. pipetting, centrifuged at 14,000 g for 5 minutes at 4C and supernatants were used for PAK1 pull down assay (Upstate, HSPC isolation, lentivirus preparation, and transduction Millipore) as described earlier (15, 16, 18, 26, 27). To the þ CD34 HSPCs were isolated from freshly collected normal supernatant, 10 mL of Rac1-conjugated agarose beads were added BMNCs and cord blood nuclear cells or from cryopreserved and incubated for 45 minutes at 4C with gentle rocking. The specimens using CD34 Microbead positive selection kit (Miltenyi beads were centrifuged at 14,000 g for 10 seconds at 4C. After Biotec) following manufacturer's protocol (21, 23). For lentivirus removing the supernatants the beads were washed three times preparation, 293T cells were maintained in DMEM supplemented with MLB buffer (1) and resuspended in 40 mL of protein with 10% FBS, 100 U/mL penicillin, 100 mg/mL streptomycin, loading buffer, boiled for 5 minutes, separated in 12% polyacryl- and 2 mmol/L L-glutamine (all from Gibco) at 37 C with 5% CO2. amide gel, and transferred to PVDF membrane and probed using

OF4 Mol Cancer Res; 2018 Molecular Cancer Research

SWI/SNF Regulates Oncogenic Signaling in AML

www.aacrjournals.org Mol Cancer Res; 2018 OF5

Chatterjee et al.

respective antibodies. Presence of active Rac GTP was determined logarithm of the respective inhibitor concentration using Graph- by using antibody against Rac GTP in the pulldown fraction and Pad Prism5. Gene expression correlation plots were derived using normalized against total Rac present in the lysate. Densitometry the TCGA datasets of AML cohort available through cBioPortal. analyses were performed using NIH Image J. Densitometry analyses were performed using Image J software (NIH). For all statistical analyses, the level of significance was set Migration assay at 0.05. HL60 cells were transduced with lentiviral particles expressing sh-SMARCB1 or sh-Control in a U bottom 96-well nontissue Database availability culture–treated plate. Cells were incubated overnight with virus All sequencing data have been submitted to the database with particles at a MoI of 5 and polybrene was added at a final accession numbers as follows. ChIP-seq: GSE108976. RNA-seq: concentration of 8 mg/mL. Forty-eight hours posttransduction, SRA accession: SRP127783; BioProject ID: PRJNA428149. 50,000 cells were seeded in duplicate in the top chamber of a 24- well Transwell (Corning Incorporated) in 100 mL of IMDM. After 4 Results hours of incubation at 37 C with 5% CO2, the total migrated cells m Human primary AML cells show loss of SMARCB1 expression were counted from bottom chamber containing 600 L of IMDM D supplemented with 10% FBS, 100 U/mL penicillin, 100 mg/mL and SWI/SNF nucleation streptomycin, and 100 ng/mL CXCL12 (PeproTech). Total num- We set out to identify SWI/SNF contribution to human AML þ fi fi ber of GFP cells was analyzed by flow cytometry. pathogenesis. Gene expression analysis identi ed a signi cant loss of SMARCB1 (SNF5 or BAF47) in human primary AML bone marrow nuclear cells (BMNCs; P < 0.0058, n ¼ 67) compared with Flow cytometric analysis þ age-matched normal bone marrow (NBM) CD34 hematopoietic AML cell lines were transduced with lentiviral particles expres- stem/progenitor cells (HSPC; Fig. 1A; Supplementary Tables S1 sing sh-SMARCB1 or sh-Control and coexpressing GFP at a MoI of and S2). SMARCB1 expression was also downregulated in estab- 5. After 48 hours, cells were harvested at 500 g for 5 minutes and lished AML lines (Fig. 1B). Array CGH analysis did not detect washed two times with cold PBS and resuspended in 500 mLof amplification, deletion, or copy number gain or loss, or any other PBS supplemented with 2% human serum and 7-AAD at a final þ genetic alteration at the SMARCB1 locus in our AML cohort concentration of 1 mg/mL. 7AAD /GFP cells were analyzed (Supplementary array CGH Files). However, compared with in LSRFortessa (Becton Dickinson) using FACSDiva software normal hematopoietic cells, AML blasts showed a substantial (Becton Dickinson). increase in repressive DNA methylation at CpG islands of the SMARCB1 promoter (Fig. 1C), accounting for SMARCB1 down- Correlation and survival analysis of TCGA AML cohort regulation observed in AML. In agreement with this result, inhi- Cross-cancer analysis of SMARCB1 and DOCK5 expression, as bition of DNA methyl transferases in vitro restored SMARCB1 well as SMARCB1 and DOCK5 expression heatmap cluster of levels in SMARCB1lo AML blasts (Fig. 1D). TCGA AML cohort, was derived using cBioPortal for Cancer Apart from SMARCB1, expression of SMARCD2 (BAF60B), Genomics interface (28, 29). For correlation analysis, mRNA SMARCE1 (BAF57), and ARID2 (BAF200) were also significantly þ expression (RNA Seq V2 RSEM) data were obtained from TCGA reduced in AML BMNCs compared with NBM CD34 cells (Fig. database and plotted using GraphPad Prism5. For correlation as 1E and F). Consistent with the mRNA downregulation, SMARCB1 – well as Kaplan Meier survival analysis, mRNA expression (RNA and SMARCD2 protein levels were dramatically lost in primary Seq V2 RSEM) was used for clustering of samples; all samples AML cells (Fig. 1G). Expression of SMARCC1 (BAF155), core showing expression level above mean value for the particular gene subunit, and SMARCA4 (BRG1), ATPase subunit of SWI/SNF were considered "hi," whereas all samples showing expression remained intact in AML (Fig. 1G). Coimmunoprecipitation level below mean value were considered "lo." experiments indicated association of endogenous SMARCC1 with the remaining SWI/SNF complex in AML BMNCs as well as Statistical analyses HL60 cells (Fig. 1H). Sucrose density gradient analysis further Statistical analyses were performed using GraphPad Prism 5. confirmed presence of an endogenous, residual, nuclear SWI/SNF D Statistics were calculated with Student's t test. Quantitative data complex (hereafter called SWI/SNF ) in primary AML cells are expressed as mean SEM. unless specified otherwise. For IC50 (Fig. 1I). Collectively, these data identify loss of SMARCB1 and D calculation, cell counts were normalized and plotted against SWI/SNF nucleation in human AML.

Figure 1. Human primary AML cells show loss of SMARCB1 expression and SWI/SNFD nucleation. A, qRT-PCR expression of SMARCB1 in AML (n ¼ 67) low-density bone marrow nuclear cells (BMNC) compared with age-matched normal bone marrow (BM; n ¼ 6) CD34þ HSPCs (considered as 1-fold). B, RT-qPCR expression of SMARCB1 in established AML cell lines compared to normal BM CD34þ cells (considered as 1-fold). Error bars represent means SD. C, Methylation-specific PCR at SMARCB1 promoter loci in primary AML blasts compared with normal BMNC. The fold change levels of the methylated DNA were calculated with respect to GAPDH (unrelated control). Location of respective qMSP primers are shown in the schema. Error bars represent means SD. D, qRT-PCR expression of SMARCB1 in primary AML (n ¼ 3) cells treated with 5-azacytidine or DMSO (considered as 1-fold). Untreated AML 17, AML 30, and AML 38 had significantly reduced (0.03, 0.21, and 0.13- fold, respectively) SMARCB1 expression. Error bars, means SD. E, qRT-PCR expression analysis of SMARCD2, SMARCE1,andARID2 in AML (n ¼ 67) BMNCs compared with normal bone marrow (n ¼ 6) CD34þ HSPCs (considered as 1-fold). F, qRT-PCR expression analysis of remaining SWI/SNF subunits in AML (n ¼ 67) BMNCs compared with normal bone marrow (n ¼ 6) CD34þ HSPCs (considered as 1-fold). G, Immunoblot analysis of primary AML BMNC and normal (N) hematopoietic cells. H, Coimmunoprecipitation of endogenous SMARCC1 or IgG in nuclear lysate of primary AML BMNCs (left) or HL60 cells (right). I, Sucrose density gradient (20% to 50%) analysis of primary AML (pooled from n ¼ 7) BMNC-derived nuclear lysates and immunoblotted with respective SWI/SNF antibodies. qRT-PCR values were normalized to GAPDH. Statistics were calculated with Student t test; error bars, means SEM (if not specified otherwise). Coimmunoprecipitation and immunoblots are representatives of 2–3 independent experiments with similar results.

OF6 Mol Cancer Res; 2018 Molecular Cancer Research

SWI/SNF Regulates Oncogenic Signaling in AML

Figure 2. SMARCC1 (SWI/SNFD) occupies target oncogenic loci in primary AML cells. A, ChIP-seq (P < 0.05) heatmaps showing occupancy of SMARCC1 or H3K27Ac peaks 2.5 kb upstream or downstream from transcription start site (TSS) in primary AML (n ¼ 3; AML 01, AML 02, and AML 03 as biological replicates) BMNCs. B, Representative ChIP-seq venn diagram analysis showing overlap of genes identified from SMARCC1, and H3K27Ac in primary AML (AML 01) BMNC. Number of cooccupied genes (13,158) is shown in the intersection. C, Pie chart representing ChIP-seq genomic distribution of SMARCC1 (left) and H3K27Ac (right) occupancy in primary AML BMNCs. Data represent average of three biological replicates of AMLs. D, ChIP-seq profile plots showing SMARCC1, H3K27Ac ChIP-seq signal intensities 2.5 kb upstream or downstream from transcription start site (TSS; upper) or 2 kb upstream or downstream form TSS and transcription end site (TES; lower) in primary AML (AML 02) BMNC. E, Venn diagram analysis showing SMARCC1 and H3K27Ac ChIP-seq cooccupied genes that are shared (2660) among the three biological replicates of AML BMNCs. F, ChIP-seq average genomic distribution of SMARCC1 and H3K27Ac on the shared (2,660) gene set in AML. PCR duplicates were removed using SAMTOOLS rmdup. Peak calling was performed using MACS14 model building with P value cutoff of 0.05. Annotation of the identified peaks was performed with PeakAnalyzer. Unique gene names were used to plot the Venn diagram represented by peaks in the respective samples either upstream or downstream or overlap to the genetic region. Error bars, means SEM. www.aacrjournals.org Mol Cancer Res; 2018 OF7

Chatterjee et al.

D SMARCC1 (SWI/SNF ) is involved in maintenance of Rac GTPase GEFs. To validate differential expression of gene sets D oncogenic gene expression program in primary AML cells identified as SWI/SNF targets, we next performed transcriptome Mammalian SWI/SNF complex organizes nucleosome occu- analysis. RNA-sequencing of paired samples identified, among pancy at target promoters and enhancers (30), thereby regulating the SMARCC1 and H3K27Ac cooccupied genes, 280 were signif- gene expression. Recent studies have shown SMARCB1 deficient icantly upregulated in AML, compared with normal (n ¼ 2) SWI/SNF complex to be essential for rhabdoid tumor survival (31, hematopoietic cells (Supplementary RNA-seq Files). Among the D 32). To elucidate the function of SWI/SNF in AML, we investi- upregulated genes were VAV3 and the DOCK family of Rac GEFs. gated genome-wide occupancy of endogenous SMARCC1, Importantly, although these Rac GTPase GEFs show SMARCC1 D which would indicate SWI/SNF binding, in primary AML cells binding also in controls, but the binding sites are significantly using chromatin immunoprecipitation-sequencing (ChIP-seq). distinct from that observed in AML (Fig. 3B; Supplementary Fig. SMARCC1 ChIP-seq identified about 14,000 genes on average S2D). In addition, H3K27Ac did not cooccupy SMARCC1 binding in three independent (biological replicates) AML BMNCs (Fig. 2A sites in normal hematopoietic cells (Fig. 3B; Supplementary Fig. D and B; Supplementary Fig. S1A–S1C). SMARCC1 localized S2D). Together, genes showing SWI/SNF overlap with H3K27Ac approximately 10% at promoters, approximately 41% at gene and resultant upregulation represent putative SMARCB1-depen- body, and approximately 49% at transcription start site (TSS)- dent targets. D distal intergenic regions (Fig. 2C and D). In rhabdoid tumor, it has To further confirm differential SWI/SNF binding due to been shown that SWI/SNF binding at TSS-distal enhancer loci, SMARCB1 loss, ChIP-qPCR analysis was performed using þ marked by H3K27Ac, is essential for its oncogenic role (31). Also SMARCC1 and H2K27Ac antibodies in normal CD34 cells and SMARCB1 deficiency has been shown to regulate H3K27Ac level SMARCB1-deficient primary AML blasts (Fig. 3C; Supplementary at enhancers (32, 33). In general, H3K27Ac is enriched at sites of Fig. S3A). SMARCC1 and H3K27Ac cooccupancy were signifi- active transcription; therefore, SWI/SNF function is typically cantly higher at the GEFs in AML blasts compared with the þ associated with transcription activation. H3K27Ac ChIP-seq anal- identical loci in normal CD34 cells (Fig. 3C; Supplementary ysis indicated SMARCC1 overlapped with H3K27Ac at 2,660 Fig. S3A). Upregulation of the GEFs expression was additionally genes (Supplementary ChIP-seq Files) that are shared among the validated by qRT-PCR analysis, which shows that they are elevated three biological replicates of AML. Detailed statistical analysis is in AML blasts compared to control (Supplementary Fig. S3B). included in Supplementary ChIP-seq Files as "Supplementary Collectively, these results indicate differential locus-specific bind- ChIP-seq_P values of Venn diagram genes." Enrichment of ing of SWI/SNF at target GEFs in AML. SMARCC1 at these shared genes was approximately 10% at promoters, approximately 55% at gene body, and approximately SMARCB1 levels correlate with DOCK expression and AML 35% at TSS-distal intergenic regions (Fig. 2E and F). pathophysiology D To identify the role of SMARCB1 in directing SWI/SNF func- The Cancer Genome Atlas (TCGA) cross-cancer analysis reveals tion, we compared the AML ChIP-seq data with primary, normal SMARCB1 median expression level to be minimum in AML hematopoietic nuclear cells expressing SMARCB1-containing patients, after mesothelioma (Fig. 4A). This as well as our AML intact SWI/SNF complex (Supplementary Fig. S2A). ChIP-seq cohort indicates that SMARCB1 downregulation is a general analysis identified about 16,500 genes occupied by SMARCC1, phenomenon observed in AML. Unlike SMARCB1, median out of which 13,987 also showed H3K27Ac (Supplementary Fig. expression of SMARCD2 was not downregulated, and was at par S2B). The distribution pattern of SMARCC1 was also similar to with other cancers (Supplementary Fig. S3C). Therefore, we that in AML, with 4%, 37%, and 58% being the occupancy at focused our analysis on SMARCB1 for subsequent studies. We promoter, intron, and TSS-distal intergenic region respectively wanted to evaluate whether SMARCB1 levels have any prognostic (Supplementary Fig. S2B). Hence unlike in rhabdoid tumor, significance in AML. To this end we studied the survival trends of SMARCB1 deficiency does not affect the overall occupancy of AML patients corresponding to their expression of SMARCB1 SWI/SNF complex in AML, indicating that the regulation may from TCGA database. In many AML subtypes, well-characterized rather be gene or loci specific. Motif analysis of SMARCC1- oncogenic translocations are enough to drive leukemogenesis. To binding sites in AML identified enrichment of several transcrip- eliminate the effect of complex translocations, only patients with tion factors, notably KLF4, HOXA13, and HOXD13 that are normal karyotype were considered. Patients with lower SMARCB1 implicated in AML (Supplementary Fig. S2C and Supplementary levels showed correspondingly poorer nonsignificant overall (P ¼ ChIP-seq Files). Motif analysis in normal hematopoietic cells 0.561) as well as disease-free (P ¼ 0.230) survival. identified enrichment of MEF2A, MEF2B, and MEF2D (Supple- DOCK family of Rac GTPase GEFs being one of the primary D mentary Fig. S2C and Supplementary ChIP-seq Files). Therefore, targets of SWI/SNF as identified from ChIP-seq and transcrip- SWI/SNF is associated with different sets of transcription factors in tome analysis, we next wanted to check whether there was AML and normal hematopoietic cells. This suggests that although any correlation between SMARCB1 and DOCK members in SMARCB1 loss does not alter overall chromatin affinity of SWI/ AML. TCGA cross-cancer analysis showed that DOCK5 has SNF, it may differentially determine recruitment to and regulation highest expression in AML among multiple cancers (Fig. 4B). of altered gene sets. This is in stark contrast to SMARCB1 expression, which led us D Next, to identify gene sets regulated by SWI/SNF , we per- to hypothesize that SMARCB1 and DOCK5 expression must formed functional annotation clustering of the SMARCC1 and be inversely correlated. Indeed, TCGA database analysis con- H3K27Ac cooccupied genes. Gene Ontology (GO) terms and firmed significant negative correlation between SMARCB1 and pathway analysis showed an enrichment of transcripts associated DOCK5 expression in AML (Fig. 4C). Patients with reduced with Rac GTPase-dependent cell migration, hematopoietic self- SMARCB1 showed significantly upregulated DOCK5 levels renewal, and transcriptional regulation (Fig. 3A). Interestingly, (Fig.4C).LikeDOCK5, the other atypical Rac GTPase GEFs, among these gene sets, we noted SMARCC1 occupancy at several DOCK2, DOCK8,andDOCK10 also showed reciprocal

OF8 Mol Cancer Res; 2018 Molecular Cancer Research

SWI/SNF Regulates Oncogenic Signaling in AML

Figure 3. SMARCC1 is involved in maintenance of oncogenic gene expression program in primary AML cells. A, Pathway (top) and Gene Ontology (GO) term (bottom) analysis of 2,660 gene set in AML. B, Representative ChIP-seq integrated genome browser view (IGV) snapshots showing occupancy of SMARCC1 and H3K27Ac at VAV3 (top) and DOCK5 (bottom) loci in one of the AML BMNCs. C, ChIP-qPCR analyses showing occupancy of SMARCC1 and H3K27Ac on target genomic loci (Region 1) in normal, CD34þ HSPCs and primary AML blasts, ChIP-qPCR values were normalized to percent input. Location of the respective ChIP-qPCR primers are shown in the schema. Statistics were calculated with Student t test; error bars, means SD.

www.aacrjournals.org Mol Cancer Res; 2018 OF9

Chatterjee et al.

Figure 4. SMARCB1 levels correlate with GEFs expression and AML pathophysiology. A, SMARCB1 mRNA expression [RNA Seq V2 (log)] from TCGA pan-cancer dataset. Box indicates second lowest expression in AML. B, DOCK5 mRNA expression [RNA Seq V2 (log)] from TCGA pan-cancer dataset. Box indicates maximum expression in AML. C, SMARCB1 gene expression correlation plots with GEFs in AML cohort (n ¼ 200) from TCGA/cBioPortal dataset.

OF10 Mol Cancer Res; 2018 Molecular Cancer Research

SWI/SNF Regulates Oncogenic Signaling in AML

Figure 5. SMARCB1 deficiency induces GEFs expression by promoting H3K27Ac at target loci. A, qRT-PCR expression of GEFs in normal CD34þ cells transduced with sh- SMARCB1 or sh-Control (considered as 1-fold). B, RT-qPCR expression of GEFs in AML cell lines expressing sh-SMARCB1 compared with sh-Control (considered as 1-fold). C, ChIP-qPCR analysis showing occupancy of SMARCC1, H3K27Ac, p300, CBP, and BRD4 on target GEFs loci (Regions R1 and R2) in 293T cells that were transiently transfected with sh-SMARCB1 or sh-Control. ChIP-qPCR values were normalized to IgG. Location of the respective ChIP-qPCR primers are shown in the schema. qRT-PCR experiments are representatives of atleast two independent biological replicates with similar results. qRT-PCR values were normalized to GAPDH. Statistics were calculated with Student t test; error bars, means SD. , P < 0.05 was considered to be statistically significant. correlation with SMARCB1 (Fig. 4C). Survival analysis further findings provide mechanistic evidence for SMARCB1 loss and D strengthened the SMARCB1-DOCK interdependence, demon- SWI/SNF -mediated transcriptional regulation of GEFs expres- strating that patients with low SMARCB1 and high DOCK sion in AML. expression have poorer nonsignificant survival than patients with high SMARCB1 and low DOCK levels. Loss of SMARCB1 induces Rac GTPase activation Rac GEFs control the activation of Rac GTPase signaling; SMARCB1 deficiency induces GEFs expression by promoting therefore, we asked whether induction in GEFs expression in H3K27Ac at target loci SMARCB1-deficient cells affect Rac activation. Consistent with D SWI/SNF binding analysis indicated that SMARCB1 deficiency the role of SMARCB1 in preferential recruitment of SWI/SNF, HAT is accompanied with elevated H3K27Ac at oncogenic loci in AML and H3K27Ac, and expression of DOCK genes, loss of SMARCB1 cells. In agreement with this, lentivirus-mediated silencing of resulted in approximately 2-fold activation of Rac GTPase (Fig. 6A þ SMARCB1 in normal CD34 cells and in established AML cell and B). This was accompanied with increase in cell migration, lines (Supplementary Fig. S3D–S3G), which still express residual suggesting its tumor suppressor role (Fig. 6C). In our AML SMARCB1, induced GEFs expression (Fig. 5A and B; Supplemen- discovery cohort, Rac GEFs emerged as important downstream D D tary Fig. S4A and S4B). Acetylation at H3K27, which controls candidates of SWI/SNF . Mechanism of SWI/SNF -mediated transcriptional state and facilitates gene expression are mediated oncogene induction is through recruitment of HATs and increased by HATs, and recently it has been shown that SWI/SNF coimmu- H3K27 acetylation. HATs as well as several SWI/SNF subunits noprecipitates with HATs in rhabdoid tumors (33). Coimmuno- contain acetyl-lysine binding bromodomains, which is a target of D precipitation studies indicated interaction of SWI/SNF and HATs BET inhibitors (34). We therefore sought to determine whether in AML cells (Supplementary Fig. S4C). We investigated whether SMARCB1-deficiency would alter sensitivity of AML cells to BET D loss of SMARCB1 can affect recruitment of SWI/SNF and HATs at inhibition. SMARCB1-deficient AML cells were significantly more target GEFs. Although SMARCB1 loss did not apparently affect sensitive to BET inhibition than controls (Supplementary Fig. SWI/SNF interaction with the HATs (Supplementary Fig. S4D and S5C). Collectively, these results indicate that loss of SMARCB1 in D S4E), interestingly SMARCB1 deficiency resulted in an increased AML cells results in increased occupancy of SWI/SNF along with occupancy of SMARCC1, HATs, and H3K27Ac levels at target GEFs HATs to target GEFs (Fig. 6D), which induces GEFs expression, loci (Fig. 5C; Supplementary Fig. S5A and S5B). Collectively, these activation of Rac GTPase signaling and cell migration.

www.aacrjournals.org Mol Cancer Res; 2018 OF11

Chatterjee et al.

Figure 6. Loss of SMARCB1 induces Rac GTPase activation. A, PAK1 pulldown assay in AML cell lines transduced with shRNA- expressing constructs against SMARCB1 or control. Densitometry represents ratio of Rac GTP versus total Rac normalized to sh-C. B, PAK1 pulldown assay of 293T cells transiently transfected with two different shRNA-expressing constructs against SMARCB1 or control. Densitometry represents ratio of Rac GTP versus total Rac normalized to sh-C. Data represent one of two independent experiments with similar results. C, Migration towards CXCL12 of GFPþ HL60 cells expressing sh-C or sh-SMARCB1. Data represent average of two independent experiments with similar results. Statistics were calculated with Student t test; error bars, means SD. D, Schema representing loss of SMARCB1-driven epigenetic signal integration towards maintenance of elevated Rac GTPase signaling in AML cells. , P < 0.01 was considered to be statistically signiﬁcant.

Discussion ration toward maintenance of oncogenic gene expression pro- In this study, we present evidence that in human primary gram precisely in human primary AML cells. It strengthens the AML cells there is loss of SMARCB1, which is associated with importance of epigenetic perturbation of the SWI/SNF complex in D D nucleation of SWI/SNF . Leukemic SWI/SNF retains core com- tumorigenesis. Essentially, our data indicate that in AML, D ponents SMARCC1 and SMARCA4. Recent reports suggest that SMARCB1 deficiency associates with altered SWI/SNF , and leu- loss of Snf5, yeast homolog of mammalian SMARCB1, induces kemic cells depend on SWI/SNF core components for transcrip- formation of aberrant SWI/SNF complex (35), and SMARCB1- tional dysregulation and survival. deficient malignant rhabdoid tumors depend on SMARCA4 for Importantly, in our study among the SMARCC1 and H3K27Ac transformation (36). SMARCB1 has previously been implicated in cooccupied genomic targets, we noted SMARCC1 occupancy at rhabdoid tumor, with biallelic inactivating mutations sufficient to several Rac GTPase GEFs, which play important roles in cell drive malignant transformation (13). A separate study has shown survival, trafficking, and small GTPase signaling (17, 39, 40). that SMARCA4 regulates proliferation of murine leukemic cells SMARCB1 deficiency upregulated expression of the GEFs, and was (9). In addition, Smarcd2-deficient mice fail to generate function- associated with Rac GTPase activation and hypermigration of ally mature myeloid cells (5, 6). Therefore, individual subunits of AML cells. Earlier, we and others have shown that Rac GTPases SWI/SNF complex have been shown to coordinate diverse cellular critically regulate leukemia cell engraftment and survival (15–18). functions regulating disease and development. Transcriptional VAV3, Rho/Rac GTPase GEF, was implicated in leukemogenesis plasticity is an emerging aspect in tumorigenesis (37, 38). Herein, (41, 42). DOCK2 is a noncanonical GEF for Rac GTPases, and our study provides evidence and reinforces the importance of DOCK2 inhibition in vivo attenuates AML cell survival (43, 44). D leukemic SWI/SNF with altered subunit stoichiometry configu- SNF5 was implicated in regulation of RhoA-dependent

OF12 Mol Cancer Res; 2018 Molecular Cancer Research

SWI/SNF Regulates Oncogenic Signaling in AML

cytoskeleton organization and migration of malignant rhabdoid Low SMARCB1 corresponds to elevated DOCK family expression. tumor cells (45). In our study, SMARCC1 occupancy was also Moreover survival analysis considering SMARCB1 and DOCK enriched at KIT, ASXL1, SF3B1, TET2 target loci, and silencing of levels reflect the importance of this correlation, with AML patients SMARCB1 induced their expression. Many of these genes are having low SMARCB1 and high DOCK expression displaying somatically mutated at relatively high frequency in myeloid poorer disease-free survival compared with patients with high malignancies with poor prognosis (46, 47), suggesting that SMARCB1 and low DOCK expression. D SMARCC1/SWI/SNF would help sustain expression of mutant To conclude this study, we elucidate that in human primary AML D oncoproteins in AML. Collectively, these findings account for cells SMARCC1, an intact core component of SWI/SNF , coloca- D SWI/SNF involvement in maintenance of AML cell gene expres- lized with H3K27Ac to target oncogenic loci. Loss of SMARCB1 sion program. induced Rac GTPase GEFs expression, Rac activation and promoted We demonstrate that a fraction of SMARCC1 and H3K27Ac AML cell migration and survival. In summary, these findings cooccupied genomic targets in AML cells were enriched in TSS- inform epigenetic signal integration downstream of SWI/SNF distal intergenic regions (30%). SWI/SNF was shown to play toward oncogenic gene expression program maintenance in AML. chromatin remodeling function at both promoters and enhancers (30, 48). Recent reports have demonstrated an interdependency Disclosure of Potential Conflicts of Interest of SWI/SNF and HAT function (31). SMARCB1 levels were shown No potential conflicts of interest were disclosed. to regulate not only expression of p300, BRD4, and mediator, but also control interaction of SWI/SNF with p300 in rhabdoid tumor Authors' Contributions (31). Intact SWI/SNF function, overlapping with H3K27Ac is Conception and design: A. Sengupta needed for the maintenance of lineage-specific enhancers, regu- Development of methodology: S.S. Chatterjee, M. Biswas lating cell fate and differentiation (33). SMARCB1 deficiency, Acquisition of data (provided animals, acquired and managed patients, however, shifts SWI/SNF recruitment from enhancers to onco- provided facilities, etc.): S.S. Chatterjee, M. Biswas, D. Banerjee Analysis and interpretation of data (e.g., statistical analysis, biostatistics, genic super-enhancer regions (31). However, in contrast to some computational analysis): S.S. Chatterjee, M. Biswas, L. Debraj Boila, of these reports where experiments were performed in different A. Sengupta cell types, our results indicate that loss of SMARCB1 induces Rac Writing, review, and/or revision of the manuscript: L. Debraj Boila, GEFs expression that is associated with elevated SMARCC1 and A. Sengupta H3K27Ac occupancy at target loci. This is similar to an earlier Administrative, technical, or material support (i.e., reporting or organizing study demonstrating that Snf5 localizes to Gli1-regulated pro- data, constructing databases): A. Sengupta Study supervision: A. Sengupta moters and that loss of Snf5 leads to activation of the Hedgehog– Gli pathway in malignant rhabdoid tumors (49). Essentially, Acknowledgments these findings indicate that SWI/SNF function and epigenetic fi This study is supported by funding from DST (SB/SO/HS-053/2013; to A. plasticity secondary to absence of speci c SWI/SNF subunits are Sengupta), DBT (BT/PR13023/MED/31/311/2015; to A. Sengupta), DBT Rama- cell type and context-dependent phenomenon. lingaswami Fellowship (BT/RLF/RE-ENTRY/06/2010; to A. Sengupta), and Genetic perturbations are usually attributed to genetic dele- CSIR, Govt. of India (NWP/BIODISCOVERY, BSC 0120; to A. Sengupta). S.S. tions and inactivating mutations. Apart from this DNA methyl- Chatterjee, and M. Biswas acknowledge fellowships from CSIR and UGC, transferases play an important role in silencing expression of key respectively. The authors thank Dr. Prasanta Mukhopadhyay for providing umbilical cord tumor suppressor genes (50). Genotyping of our AML cohort blood samples. We acknowledge Dr. Olivier Delattre for sharing plasmids and indeed corroborates this, showing that SMARCB1 loss seen in Addgene for shipping DNA constructs. We also thank Drs. Arindam Maitra and AML is not due to genetic deletion, but rather increased DNA Subrata Patra, CoTERI, National Institute of Biomedical Genomics (NIBMG) for methylation at CpG islands of proximal promoter region. In conducting ChIP-seq experiments; Dr. Madavan Vasudevan, Madhura and addition, we demonstrate that expression of SMARCB1 corre- Shemi Ramesh, Bionivid Technology for ChIP-seq analysis and RNA-sequenc- fl sponds to prognosis in AML patients. SMARCB1 expression is the ing; Genotypic Technology, Bangalore for array CGH experiments and IICB ow cytometry core for services. lowest in AML among multiple cancers; and also within AML cohort, patients with low SMARCB1 levels to have comparatively The costs of publication of this article were defrayed in part by the payment of shorter overall as well as disease-free survival period compared D page charges. This article must therefore be hereby marked advertisement in with patients with higher SMARCB1 expression. That SWI/SNF accordance with 18 U.S.C. Section 1734 solely to indicate this fact. mediated Rac GEF regulation is indeed important in AML pathogenesis is demonstrated by reciprocal correlation between Received September 8, 2017; revised January 11, 2018; accepted February 20, SMARCB1 and various members of the DOCK family in AML. 2018; published first February 26, 2018.

References 1. Wu JI, Lessard J, Crabtree GR. Understanding the words of chromatin 4. Wilson BG, Roberts CW. SWI/SNF nucleosome remodellers and cancer. Nat regulation. Cell 2009;136:200–6. Rev Cancer 2011;11:481–92. 2. Ho L, Ronan JL, Wu J, Staahl BT, Chen L, Kuo A, et al. An embryonic stem 5. Priam P, Krasteva V, Rousseau P, D'Angelo G, Gaboury L, Sauvageau G, cell chromatin remodeling complex, esBAF, is essential for embryonic stem et al. SMARCD2 subunit of SWI/SNF chromatin-remodeling complexes cell self-renewal and pluripotency. Proc Natl Acad Sci U S A 2009;106: mediates granulopoiesis through a CEBPvarepsilon dependent mecha- 5181–6. nism. Nat Genet 2017;49:753–64. 3.KadochC,HargreavesDC,HodgesC,EliasL,HoL,RanishJ,etal. 6. Witzel M, Petersheim D, Fan Y, Bahrami E, Racek T, Rohlfs M, et al. Proteomic and bioinformatic analysis of mammalian SWI/SNF com- Chromatin-remodeling factor SMARCD2 regulates transcriptional net- plexes identiﬁes extensive roles in human malignancy. Nat Genet 2013; works controlling differentiation of neutrophil granulocytes. Nat Genet 45:592–601. 2017;49:742–52.

www.aacrjournals.org Mol Cancer Res; 2018 OF13

Chatterjee et al.

7. Krasteva V, Buscarlet M, Diaz-Tellez A, Bernard MA, Crabtree GR, Lessard 29. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. JA. The BAF53a subunit of SWI/SNF-like BAF complexes is essential for Integrative analysis of complex cancer genomics and clinical profiles using hemopoietic stem cell function. Blood 2012;120:4720–32. the cBioPortal. Sci Signal 2013;6:pl1. 8. Krosl J, Mamo A, Chagraoui J, Wilhelm BT, Girard S, Louis I, et al. A mutant 30. Tolstorukov MY, Sansam CG, Lu P, Koellhoffer EC, Helming KC, Alver BH, allele of the Swi/Snf member BAF250a determines the pool size of fetal et al. Swi/Snf chromatin remodeling/tumor suppressor complex estab- liver hemopoietic stem cell populations. Blood 2010;116:1678–84. lishes nucleosome occupancy at target promoters. Proc Natl Acad Sci U S A 9. Buscarlet M, Krasteva V, Ho L, Simon C, Hebert J, Wilhelm B, et al. Essential 2013;110:10165–70. role of BRG, the ATPase subunit of BAF chromatin remodeling complexes, 31. Wang X, Lee RS, Alver BH, Haswell JR, Wang S, Mieczkowski J, et al. in leukemia maintenance. Blood 2014;123:1720–8. SMARCB1-mediated SWI/SNF complex function is essential for enhancer 10. Shi J, Whyte WA, Zepeda-Mendoza CJ, Milazzo JP, Shen C, Roe JS, et al. regulation. Nat Genet 2017;49:289–95. Role of SWI/SNF in acute leukemia maintenance and enhancer-mediated 32. Nakayama RT, Pulice JL, Valencia AM, McBride MJ, McKenzie ZM, Myc regulation. Genes Dev 2013;27:2648–62. Gillespie MA, et al. SMARCB1 is required for widespread BAF com- 11. Grand F, Kulkarni S, Chase A, Goldman JM, Gordon M, Cross NC. Frequent plex-mediated activation of enhancers and bivalent promoters. Nat deletion of hSNF5/INI1, a component of the SWI/SNF complex, in chronic Genet 2017;49:1613–23. myeloid leukemia. Cancer Res 1999;59:3870–4. 33. Alver BH, Kim KH, Lu P, Wang X, Manchester HE, Wang W, et al. The SWI/ 12. Roberts CW, Galusha SA, McMenamin ME, Fletcher CD, Orkin SH. SNF chromatin remodelling complex is required for maintenance of Haploinsufficiency of Snf5 (integrase interactor 1) predisposes to lineage specific enhancers. Nat Commun 2017;8:14648. malignant rhabdoid tumors in mice. Proc Natl Acad Sci U S A 2000;97: 34. Zuber J, Shi J, Wang E, Rappaport AR, Herrmann H, Sison EA, et al. RNAi 13796–800. screen identifies Brd4 as a therapeutic target in acute myeloid leukaemia. 13. Roberts CW, Leroux MM, Fleming MD, Orkin SH. Highly penetrant, rapid Nature 2011;478:524–8. tumorigenesis through conditional inversion of the tumor suppressor gene 35. Sen P, Luo J, Hada A, Hailu SG, Dechassa ML, Persinger J, et al. Loss of Snf5 Snf5. Cancer Cell 2002;2:415–25. induces formation of an aberrant SWI/SNF complex. Cell Rep 2017;18: 14. Schmidt A, Hall A. Guanine nucleotide exchange factors for Rho GTPases: 2135–47. turning on the switch. Genes Dev 2002;16:1587–609. 36. Wang X, Sansam CG, Thom CS, Metzger D, Evans JA, Nguyen PT, et al. 15. Cancelas JA, Lee AW, Prabhakar R, Stringer KF, Zheng Y, Williams DA. Rac Oncogenesis caused by loss of the SNF5 tumor suppressor is dependent on GTPases differentially integrate signals regulating hematopoietic stem cell activity of BRG1, the ATPase of the SWI/SNF chromatin remodeling localization. Nat Med 2005;11:886–91. complex. Cancer Res 2009;69:8094–101. 16. Sengupta A, Arnett J, Dunn S, Williams DA, Cancelas JA. Rac2 GTPase 37. Rathert P, Roth M, Neumann T, Muerdter F, Roe JS, Muhar M, et al. deficiency depletes BCR-ABLþ leukemic stem cells and progenitors in vivo. Transcriptional plasticity promotes primary and acquired resistance to Blood 2010;116:81–4. BET inhibition. Nature 2015;525:543–7. 17. Thomas EK, Cancelas JA, Zheng Y, Williams DA. Rac GTPases as key 38. Liau BB, Sievers C, Donohue LK, Gillespie SM, Flavahan WA, Miller TE, regulators of p210-BCR-ABL-dependent leukemogenesis. Leukemia et al. Adaptive chromatin remodeling drives glioblastoma stem cell plas- 2008;22:898–904. ticity and drug tolerance. Cell Stem Cell 2017;20:233–46. 18. Gu Y, Filippi MD, Cancelas JA, Siefring JE, Williams EP, Jasti AC, et al. 39. Mulloy JC, Cancelas JA, Filippi MD, Kalfa TA, Guo F, Zheng Y. Rho GTPases Hematopoietic cell regulation by Rac1 and Rac2 guanosine triphospha- in hematopoiesis and hemopathies. Blood 2010;115:936–47. tases. Science 2003;302:445–9. 40. Shen X, Li CC, Aponte AM, Shen RF, Billings EM, Moss J, et al. Brefeldin A- 19. Mizukawa B, Wei J, Shrestha M, Wunderlich M, Chou FS, Griesinger A, et al. inhibited ADP-ribosylation factor activator BIG2 regulates cell migration Inhibition of Rac GTPase signaling and downstream prosurvival Bcl-2 via integrin beta1 cycling and actin remodeling. Proc Natl Acad Sci U S A proteins as combination targeted therapy in MLL-AF9 leukemia. Blood 2012;109:14464–9. 2011;118:5235–45. 41. Chang KH, Sanchez-Aguilera A, Shen S, Sengupta A, Madhu MN, Ficker 20. Boila LD, Chatterjee SS, Banerjee D, Sengupta A. KDM6 and KDM4 histone AM, et al. Vav3 collaborates with p190-BCR-ABL in lymphoid progenitor lysine demethylases emerge as molecular therapeutic targets in human leukemogenesis, proliferation, and survival. Blood 2012;120:800–11. acute myeloid leukemia. Exp Hematol 2018;58:44–51. 42. Rinke J, Muller JP, Blaess MF, Chase A, Meggendorfer M, Schafer V, et al. 21. Sengupta A, Banerjee D, Chandra S, Banerji SK, Ghosh R, Roy R, et al. Molecular characterization of EZH2-mutant patients with myelodysplas- Deregulation and cross talk among Sonic hedgehog, Wnt, Hox and Notch tic/myeloproliferative neoplasms. Leukemia 2017;31:1936–43. signaling in chronic myeloid leukemia progression. Leukemia 2007; 43. Wu M, Hamaker M, Li L, Small D, Duffield AS. DOCK2 interacts with FLT3 21:949–55. and modulates the survival of FLT3-expressing leukemia cells. Leukemia 22. Sengupta A, Duran A, Ishikawa E, Florian MC, Dunn SK, Ficker AM, et al. 2017;31:688–96. Atypical protein kinase C (aPKCzeta and aPKClambda) is dispensable for 44. Laurin M, Huber J, Pelletier A, Houalla T, Park M, Fukui Y, et al. Rac-specific mammalian hematopoietic stem cell activity and blood formation. Proc guanine nucleotide exchange factor DOCK1 is a critical regulator of HER2- Natl Acad Sci U S A 2011;108:9957–62. mediated breast cancer metastasis. Proc Natl Acad Sci U S A 2013; 23. Sengupta A, Ficker AM, Dunn SK, Madhu M, Cancelas JA. Bmi1 reprograms 110:7434–9. CML B-lymphoid progenitors to become B-ALL-initiating cells. Blood 45. Caramel J, Quignon F, Delattre O. RhoA-dependent regulation of cell 2012;119:494–502. migration by the tumor suppressor hSNF5/INI1. Cancer Res 2008;68: 24. Chang KH, Sengupta A, Nayak RC, Duran A, Lee SJ, Pratt RG, et al. p62 is 6154–61. required for stem cell/progenitor retention through inhibition of IKK/NF- 46. Patel JP, Gonen M, Figueroa ME, Fernandez H, Sun Z, Racevskis J, et al. kappaB/Ccl4 signaling at the bone marrow macrophage-osteoblast niche. Prognostic relevance of integrated genetic profiling in acute myeloid Cell Rep 2014;9:2084–97. leukemia. N Engl J Med 2012;366:1079–89. 25. Thomas EK, Cancelas JA, Chae HD, Cox AD, Keller PJ, Perrotti D, et al. Rac 47. Shih AH, Abdel-Wahab O, Patel JP, Levine RL. The role of mutations in guanosine triphosphatases represent integrating molecular therapeutic epigenetic regulators in myeloid malignancies. Nat Rev Cancer 2012;12: targets for BCR-ABL-induced myeloproliferative disease. Cancer Cell 599–612. 2007;12:467–78. 48. Hu G, Schones DE, Cui K, Ybarra R, Northrup D, Tang Q, et al. Regulation of 26. Florian MC, Dorr K, Niebel A, Daria D, Schrezenmeier H, Rojewski M, et al. nucleosome landscape and transcription factor targeting at tissue-specific Cdc42 activity regulates hematopoietic stem cell aging and rejuvenation. enhancers by BRG1. Genome Res 2011;21:1650–8. Cell Stem Cell 2012;10:520–30. 49. Jagani Z, Mora-Blanco EL, Sansam CG, McKenna ES, Wilson B, Chen D, 27. Florian MC, Nattamai KJ, Dorr K, Marka G, Uberle B, Vas V, et al. A et al. Loss of the tumor suppressor Snf5 leads to aberrant activation of the canonical to non-canonical Wnt signalling switch in haematopoietic stem- Hedgehog-Gli pathway. Nat Med 2010;16:1429–33. cell ageing. Nature 2013;503:392–6. 50. Lund K, Cole JJ, VanderKraats ND, McBryan T, Pchelintsev NA, Clark W, 28. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio et al. DNMT inhibitors reverse a specific signature of aberrant promoter cancer genomics portal: an open platform for exploring multidimensional DNA methylation and associated gene silencing in AML. Genome Biol cancer genomics data. Cancer Discov 2012;2:401–4. 2014;15:406.

OF14 Mol Cancer Res; 2018 Molecular Cancer Research

SMARCB1 Deficiency Integrates Epigenetic Signals to Oncogenic Gene Expression Program Maintenance in Human Acute Myeloid Leukemia

Shankha Subhra Chatterjee, Mayukh Biswas, Liberalis Debraj Boila, et al.

Mol Cancer Res Published OnlineFirst February 26, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-17-0493

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2018/02/24/1541-7786.MCR-17-0493.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/early/2018/04/02/1541-7786.MCR-17-0493. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.