Published OnlineFirst June 20, 2012; DOI: 10.1158/1541-7786.MCR-11-0488
Molecular Cancer Cancer Genes and Genomics Research
Identification of Genomic Targets of Transcription Factor Aebp1 and its role in Survival of Glioma Cells
Jayashree Ladha, Swati Sinha, Vasudeva Bhat, Sainitin Donakonda, and Satyanarayana M.R. Rao
Abstract A recent transcriptome analysis of graded patient glioma samples led to identification of AEBP1 as one of the genes upregulated in majority of the primary GBM as against secondary GBM. Aebp1 is a transcriptional repressor that is involved in adipogenesis. It binds to AE-1 element present in the proximal promoter of aP2 gene that codes for fatty acid binding protein (FABP4). A comprehensive study was undertaken to elucidate the role of AEBP1 overexpression in glioblastoma. We employed complementary gene silencing approach to identify the genes that are perturbed in a glioma cell line (U87MG). A total of 734 genes were differentially regulated under these conditions ( 1.5-fold, P 0.05) belonging to different GO categories such as transcription regulation, cell growth, proliferation, differentiation, and apoptosis of which perturbation of 114 genes of these pathways were validated by quantitative real time PCR (qRT-PCR). This approach was subsequently combined with ChIP-chip technique using an Agilent human promoter tiling array to identify genomic binding loci of Aebp1 protein. A subset of these genes identified for Aebp1 occupancy was also validated by ChIP-PCR. Bioinformatics analysis of the promoters identified by ChIP-chip technique revealed a consensus motif GAAAT present in 66% of the identified genes. This consensus motif was experimentally validated by functional promoter assay using luciferase as the reporter gene. Both cellular proliferation and survival were affected in AEBP1-silenced U87MG and U138MG cell lines and a significant percentage of these cells were directed towards apoptosis. Mol Cancer Res; 10(8); 1039–51. 2012 AACR.
Introduction (5). However, secondary GBM often exhibits P53 muta- tions, PDGF/PDGFR overexpression, RB loss, and CDK4 Glioblastoma multiforme (GBM) is the most common fi and malignant form of primary tumor of CNS in adults, ampli cations (6). Recent studies have shown, however, that which is characterized by a median survival of less than a year. there is an overlapping spectrum of mutations in these 2 types of GBM (7, 8). In one of our earlier studies we had found The prognostic behavior of GBMs is rather poor and hence AEBP1 there have been efforts to identify molecular signatures and expression to be upregulated in primary GBMs as opposed to progressive secondary GBMs (9). Aebp1 was also to discover new biomarkers for characterizing different fi types and stages of GBMs (1–4). GBM is broadly classified originally identi ed as a transcriptional repressor that binds to into primary and secondary GBM (WHO), each one arising adipocyte enhancer 1 (AE-1 element) located in the proximal de promoter region of the adipose P2 gene, which codes for through distinct genetic pathways. Primary GBM arises fi novo and is frequently associated with amplification and/or adipocyte speci c fatty acid binding protein 4 (FABP4; EGFR PTEN ref. 10). Aebp1 is also overexpressed in transgenic mouse overexpression of and deletion combined with ERBB2 INK4A/ARF and CDKN2A losses and MDM2 amplification probasin-Neu ( ) induced advanced prostate cancer (11). However, the exact role of AEBP1 in tumorigenesis is not clear and hence we set out to identify the genomic targets Authors' Affiliation: Chromatin Biology Laboratory, Molecular Biology and Genetics Unit, Jawaharlal Nehru Centre for Advanced Scientific of this transcription factor to understand its biology in the Research, Bangalore, India cellular context. Toward this direction we have undertaken a Note: Supplementary data for this article are available at Molecular Cancer detailed study to analyze the Aebp1 genomic targets by Research Online (http://mcr.aacrjournals.org/). transcriptome profiling of AEBP1 downregulated U87MG cells and its role in cell proliferation, growth, and survival. J. Ladha and S. Sinha have made equal contributions to this article.
V. Bhat and S. Donakonda have made equal contributions to this article. Materials and Methods Corresponding Author: Satyanarayana M.R. Rao, Chromatin Biology Cell culture and AEBP1 silencing Laboratory, Molecular Biology and Genetics Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore, India. Phone: +91 U87MG and U138MG cells (ATCC) were grown in 9886233032; Fax: +91-80-22082766/23602468; E-mail: Eagle's Minimal Essential Medium supplemented with [email protected] 10% FBS (Sigma-Aldrich). Cells were transfected with doi: 10.1158/1541-7786.MCR-11-0488 100 nmol/L siRNA pool targeted against AEBP1 (Dharma- 2012 American Association for Cancer Research. con Inc.). Quantitative real-time PCR (qRT-PCR) was done
www.aacrjournals.org 1039
Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2012 American Association for Cancer Research. Published OnlineFirst June 20, 2012; DOI: 10.1158/1541-7786.MCR-11-0488
Ladha et al.
using Eva Green (Biorad) on a Biorad iQ5 cycler. Down- intraarray median normalization to remove dye bias and regulation of AEBP1 was assessed by qRT-PCR and Western interarray median normalization to bring the distribution blot analysis. All the primers used in this study are listed in uniform across replicates. The significantly enriched genes Supplementary Table S1. were detected based on the statistical p-value using White- head Per Array Neighborhood Model. False discovery rate Global gene expression analysis and ChIP-chip promoter analysis (13) was then applied to 11,659 enriched genes for tiling array Aebp1 promoter occupancy using Bioconductor R software Total RNA was isolated from 4 independent scrambled (14). Peaks were considered significant at a P-value 0.05. siRNA and 5 test siRNA–AEBP1 transfected U87MG cells Two biological replicate experiments were carried out for and hybridized to Affymetrix U133Plus 2.0 gene chip that Aebp1 occupancy analysis. queried 47,000 genes. The results were analyzed initially using Gene-Chip operating software and the data were subsequently Promoter sequences retrieval and motif prediction processed using ArrayAssist (Agilent) to statistically analyze Promoter sequences ( 1 kb) of perturbed genes were changesingeneexpression.qRT-PCRvalidationfor114genes retrieved using 3 major databases viz., transcription reg- was done using Eva Green (Biorad) on a Biorad iQ5 cycler. ulatory element database (TRED), eukaryotic promoter ChIP assays were done according to the previously described database (EPD), and UCSC genome browser (20–22). method (12). Briefly, log-phase U87MG cells were fixed with De novo motif discovery was done using CisFinder algo- formaldehyde and chromatin was sonicated to generate an rithm (23) to identify motifs in most enriched sequences average length of 200 to 800 base pairs. After preclearing, the by ChIP experiments. Position frequency matrices were chromatin solution was incubated with affinity-purified rabbit estimated from counts of n-mer words with and without polyclonal Aebp1 antibody (SantaCruz) or purified rabbit IgG gaps and clustered to generate nonredundant sets of antibody. The abundance of genomic DNA containing a motifs. Web logo was used to construct sequence logos promoter was determined by PCR amplification using (24). To test the validity of motifs predicted from ChIP- sequence-specificprimerpairsflanking Aebp1 binding site chip data, we built control data set of 5810 random identified through position weighted matrix analysis within sequences each of approximately 50mer length from 1 kb promoters. For ChIP-chip analysis, amplified immu- human, using RSAT tools (25) and motif analysis were noprecipitated DNA and input genomic DNA was labeled conducted for these random sequences. with Cy5 and Cy3 fluorophores respectively, using random primer labeling kit (Invitrogen Corp.). Five micrograms each Correlation analysis of immunoprecipitated and genomic DNA was combined Pearson's correlation coefficients were calculated between along with human Cot-1 DNA and hybridized to each of the all replicates in gene expression and promoter tiling arrays Agilent human promoter tiling array (2 224) containing using R statistical computing (14). 474,393 probes excluding control features. Transcription factor Network Analysis Microarray data analysis The DNA binding sites of 25 transcription factors, which Gene expression data was normalized using PLIER algo- were perturbed upon AEBP1 gene silencing and also vali- rithm in ArrayAssist (Agilent) and expression changes were dated by qRT-PCR, were mined from the literature. These filtered at >1.5-fold between experiments. Genes were binding sites were searched in 1 kb promoters of each of considered significantly perturbed at a p-value of 0.05. the transcription factor genes and transcription factor gene The method of Benjamini and Hochberg (13) for false network among these 25 transcription factors was generated. discovery rate was set to 0.05 using R software (14). These The heat maps were constructed using Java Tree view genes were then subjected to an unsupervised 2-way average software (16). linkage hierarchical cluster analysis with uncentered corre- lation as similarity metric using Cluster 3.0 software (15). Identification of Aebp1 binding site by functional Java Tree view version 1.1.4 was used to visualize structure of promoter assay the data (16). Functional annotation was done using Gene FABP4 promoter ( 200 to þ21) was amplified from Ontology database and DAVID Ease software (17–19) on genomic DNA and cloned into the XhoIandHindIII sites differentially regulated genes. Pathway enrichment analysis of basic pGL3-promoter vector (Promega Corp.). Mutant was done using Genotypic Technologies Biointerpretor tool. motif promoters were generated by substituting G for A and C A p-value cut-off of 0.05 was used to identify significant for T and vice versa (Supplementary Table S1). Two micro- enrichment pathway categories. grams of various reporter constructs were cotransfected in U87MG cells with 200 ng of pCMVb (that expresses the Promoter tiling array analysis b-galactosidase gene under the control of CMV promoter) as Raw intensity data were generated using Feature extrac- transfection control. After 24 hours of transfection relative tion software v 10.5.1.1. Feature extracted data were ana- light units was measured in a Luminometer (Berthold lyzed using DNA Analytics software from Agilent (hg18 detection systems). b-Galactosidase activity was measured build). Data were normalized using Median Blanks subtrac- by fluorometric assay and used to normalize transfection tion to exclude the probes having negative intensities, efficiency.
1040 Mol Cancer Res; 10(8) August 2012 Molecular Cancer Research
Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2012 American Association for Cancer Research. Published OnlineFirst June 20, 2012; DOI: 10.1158/1541-7786.MCR-11-0488
Genomic Targets and Role in Cell Survival of AEBP1
Proliferation, growth suppression, and apoptosis assays further analysis. To identify targets of Aebp1, we started U87MG and U138MG cells were plated in 96-well plates with Aebp1 DNA binding site based on the previous and were transfected every 60 and 36 hours, respectively, literature (26). Aebp1 binds to AE-1 sequence ( 168 to with siRNA pool designed against AEBP1 or nontargeting þ21) of aP2 gene that was originally identified by Hunt and scrambled siRNA. To assess the effect of the AEBP1 gene colleagues (27) (Supplementary Fig. S2A). Ro and Roncari silencing, cells were treated with MTT (3-[4-5 dimethyl (26) also showed the importance of (G-139) in the aP2 thaizole-2-yl]2-5 diphenyltetrazolium bromide; Sigma- promoter for binding of positive and negative factors to the Aldrich) for 4 hours and the formazan crystals formed by AE-1 element. PPARg and LXRa are also reported as targets metabolically active cells was solubilized and measured in a of AEBP1 (28). This was done on the sequence length of 35 spectrophotometer at 550 nm. Colony suppression was done base pairs. When we reanalyzed these sequences, we found on U87MG or U138MG cells (0.5 106) by transfection that nucleotides matching with AE-1 element were scattered with 2 mg of control shRNA vector (Open Biosystems) or in the case of PPARg. The nucleotides that matched most shRNA designed against AEBP1. Forty-eight hours post- with PPARg promoter were "AGAA" starting at 644 to transfection, puromycin selection was done for >2 weeks. 641 and "AGAAATTT" at ( 631 to 624). We checked Resistant colonies were stained with crystal violet solution the presence of this motif within the ChIP DNA obtained and photographed. Apoptosis was assayed using FITC- from Aebp1 chromatin immunoprecipitation using the Annexin V-PI (Invitrogen Corp.) and APO-BrdU kit (Bec- flanking primer pairs to GAAAT present in the FABP4 ton Dickinsion) following manufacturers protocol. promoter. We could observe enrichment of this fragment in the ChIP DNA of FABP4 (Supplementary Fig. S2B). Results Analysis of the promoter sequence ( 1 kb) of brain-specific RNA interference of AEBP1 and gene expression FABP7 and its subsequent enrichment in ChIP PCR con- profiling firmed that it also contains the GAAAT sequence (Supple- We had observed earlier that AEBP1 was upregulated in the mentary Fig. S2C and S2D). Based on all these observations majority of primary GBM tumor samples (9). Here, we have we predicted that "GAAAT" is the probable Aebp1 binding used a complementary approach wherein we have suppressed site. Using this information we searched for the presence of endogenous AEBP1 expression in U87MG cells, an astrocy- this motif in the 1 kb upstream sequences of the tran- toma cell line, to gain an insight toward understanding the scription start sites in all the 669 genes. Among these, 442 biological role of AEBP1. We found that 100 nmol/L of genes had this predicted AE-1 element, whereas 227 genes siRNA pool brought about significant downregulation of did not have this element (Fig. 1C). There were a total of 863 AEBP1 (>90%) as against mock (scrambled siRNA) treated predicted motifs in these 442 promoters. cells without affecting the expression of human b-actin and GAPDH (Fig. 1A). Downregulation of AEBP1 was also Promoter occupancy of Aebp1 using ChIP-chip observed at the protein level (Fig. 1B). Global gene expression Our next effort was to experimentally show the occupancy profile of U87MG cells after mock transfection or transient of Aebp1 in the promoter sequences. To address this silencing of AEBP1 was determined by using human U133 question we employed the ChIP-chip technique using Agi- plus 2 array from Affymetrix. The correlation coefficient lent human promoter tiling array. The Aebp1 bound immu- analysis of the expression data revealed that results are com- noprecipitated DNA was hybridized to the tiling array in parable between replicates (Supplementary Fig. S1A). A flow replicates. Peak detection algorithm of DNA Analytics diagram of different steps of our analysis is shown in Fig. 1C. detected robust peaks of probe signal corresponding to the We observed perturbation of expression in 734 genes at more binding events. The correlation analysis showed that results than 1.5-fold change at a P-value of 0.05 (Supplementary are reproducible between replicates (Supplementary Fig. Table S2) of which 326 genes were upregulated and 408 genes S1B). The chromosome wise occupancy of Aebp1 is shown downregulated. These genes were sorted by expression ratios; in Supplementary Fig. S3(A–X). We detected 11,659 genes median centered and then subjected to hierarchical cluster as target sites of Aebp1 occupancy. These genes were further analysis (Fig. 1D, downregulated and Fig. 1E, upregulated). subjected to FDR analysis (13) to minimize false positives. A Functional categorization revealed a diverse set of GO bio- total of 5810 genes were subsequently identified following logical processes that were statistically significant. Enriched this exercise (Fig. 1C). Binding sites predicted using CisFin- categories included cell proliferation, cell cycle, cell differen- der algorithm for these genes are documented in Supple- tiation, apoptosis, transcription, protein and ion binding, mentary Table S3 along with their frequencies and enrich- signaling, and ubiquitin related (Fig. 1F). A list of the most ment ratios. Further to rule out any nonspecificity, we altered genes based on gene ontology is given in Table1. retrieved random sequences of 5810 genes using RSAT tools (25). Motif analysis of these random sequences did not Bioinformatic analysis of the promoter sequences of predict real motifs (GAAAT/TTTCT) as shown in Supple- perturbed genes mentary Table S4. Of the 669 genes that were modulated To elucidate transcriptional targets of Aebp1, we retrieved upon silencing of AEBP1 gene, 185 genes overlapped with well-annotated and characterized promoter sequences of the ChIP-chip enriched gene list. The list of these congruent these 734 genes. Among them 65 genes were unannotated genes seen both in microarray and tiling arrays are given in leaving behind 669 genes, which formed the basis for our Supplementary Table S5 and their location on individual
www.aacrjournals.org Mol Cancer Res; 10(8) August 2012 1041
Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2012 American Association for Cancer Research. Published OnlineFirst June 20, 2012; DOI: 10.1158/1541-7786.MCR-11-0488
Ladha et al.
ABC siControl siAEBP1 siControl siAEBP1 ChIP and AEBP1 AEBP1 silencing Promoter tiling AEBP1 array β-Actin GAPDH GAPDH 1.5 FC, P ≤ 0.05 FDR, P ≤ 0.05 734 5810 DE Promoter Analysis
Congruence with AE-1 PRESENT AE-1 ABSENT PROMOTER NA microarray
Expt2_18.7.7 ctrl Experiment 1_2.7.7 ctrl Expt3_Scsi ctrl Expt4_23.5.8_Scsi ctrl Expt3_si4_100 nmol/L Expt4_23.5.8_si5 Experiment 1_2.7.7 Si-100 nmol/L Expt2_18.7.7 Si-100 nmol/L.3 Expt2_18.7.7 Si-100 nmol/L.2 442 227 65 Expt2_18.7.7 ctrl Experiment 1_2.7.7 ctrl Expt3_Scsi ctrl Expt4_23.5.8_Scsi ctrl Expt3_si4_100 nmol/L Expt4_23.5.8_si5 Experiment 1_2.7.7 Si-100 nmol/L Expt2_18.7.7 Si-100 nmol/L.3 Expt2_18.7.7 Si-100 nmol/L.2 185
F Apoptosis HLA-DPB2 WIPI1 ↑13, ↓13 ATOH8 N4BP1 CCDC35 ADAMDEC1 Ubiquitin Cell differentiation ARNT TBX18 MFNG ERBB2IP ↑2, ↓7 ↑17, ↓24 ZNF582 TNFRSF10D CYP17A1 PCDHA2 ATF6 APOH Cell adhesion LOC729376 C17orf46 ↑14, ↓24 CNTNAP5 DUSP16 MTM1 RNF103 PDE8B TNFAIP3 LOC492311 SLC11A2 LOC728210 RIT2 MARVELD2 AVIL Cell cycle BAI3 CLINT1 ↑12, ↓18 LRRC3B LAMA4 Protein binding OR2L1P CDH13 ↑ ↓ ADAMTS2 28, 43 Cell proliferation ASB4 GSTA4 RHEB TMPRSS11P ↑22, ↓8 MYLIP DMRTB1 TROAP PHF21B ITGA6 C16orf7 Ion binding PGLS KRT35 SIPA1L3 ↑20, ↓30 GPRIN3 Transcription TMEM106C LOC399978 Signaling MYO1E NEGR1 ↑32, ↓38 CREB1 ERG /// TBX ↑8, ↓10 EXOC7 FLJ23588 ARL2 ZNF460 ATXN7L3 LMLN Apoptosis Cell differentiation Cell adhesion EGLN2 FLJ20273 EXOC7 ATBF1 C1orf86 SLC43A2 Cell cycle Cell proliferation Transcription TMUB2 KIRREL TRIP10 PARD6B Signaling Ion binding Protein binding FAM96B OR7E24 SLC25A39 ENTPD5 ATG10 CDK6 Transferase activity Translation Binding ANKRD39 XKRX PQBP1 SLC16A1 Immune response Receptor activity Lipid metabolism MCF2 ADAL CCNF LOC283267 NFATC1 WDR41 Neurotransmitter activity Hormone activity Metabolism LOC645355 SRPX2 FLJ31033 DET1 Ubiquitin CNS development Isomerase activity CCM2 DNAH7 PTBP1 PTBP1 KCNIP2 Ligage activity DNA binding RNA binding PTBP1 PA2G4 STX2 ZXDB MRPL37 OXSR1 RNA splicing Hydrolase Peptidase activity PTRF MX1 PTBP1 AXIN1 Protein transport Microtobule movement Ion transport LOC653464 H2AFJ SCAMP5 TRFP TES ITIH3 Protein modification process ATP binding Cell motion RAD54B TTC19
Figure 1. AEBP1 gene silencing and transcriptome analysis. A, semiquantitative RT-PCR analysis of AEBP1 in control siRNA (si control) and AEBP1 siRNA (si AEBP1) transfected U87MG cells. b-Actin and GAPDH were analyzed for the same samples. B, Western blot analysis of Aebp1 in si control and si AEBP1- treated cells. The same samples were probed for GAPDH. C, workflow of analysis to identify AEBP1 genomic targets. Heat map of genes that are downregulated (D) and those that are upregulated (E) upon AEBP1 silencing. The number of modulated genes in each gene ontology categories is represented in Pie chart (Panel F) wherein " shows upregulated genes and # denotes downregulated genes in the silenced group.
chromosomes analyzed by human genome tool in UCSC within the coding regions (29). Fig. 2B shows the distribution genome browser (22) is presented in Supplementary Fig. S4. of best probes for top scoring target genes as a function of their The congruent genes between microarray and promoter distance from the transcription start site. We also observed tiling array analysis were divided into 3 categories based on that 227 genes that were differentially regulated upon AEBP1 their probe location as within the promoter, inside the gene or silencing did not possess the predicted binding motifs. downstream from transcription start site (Fig. 2A). A total of 49.73% of these genes showed binding in the proximal Consensus motif in the promoters of Aebp1 target genes promoter region, whereas 50.27% show binding in the genes and functional promoter assay downstream of the transcription start site. It is not uncom- As discussed previously, FABP4 and FABP7 are known mon that transcription factor binding sites are observed targets of Aebp1 containing the AE-1 element that possess
1042 Mol Cancer Res; 10(8) August 2012 Molecular Cancer Research
Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2012 American Association for Cancer Research. Published OnlineFirst June 20, 2012; DOI: 10.1158/1541-7786.MCR-11-0488
Genomic Targets and Role in Cell Survival of AEBP1
Table 1. Significantly altered genes upon AEBP1 silencing
Gene symbol Gene ID Regulation Fold change P-value Function AREG 374 Up 5.75 0.00034 Cell proliferation BLZF1 8548 Up 3.17 0.037 Cell proliferation USP8 9101 Up 2.65 0.028 Cell proliferation ELF5 2001 Up 2.16 0.036 Cell proliferation TNFSF14 8740 Up 4.06 0.026 Cell proliferation, regulation of apoptosis PDGFB 5155 Up 5.38 0.01 Cell proliferation, regulation of cell migration CDKN2C 1031 Up 4.70 0.02 Cell cycle, regulation of apoptosis SCIN 85477 Up 3.11 0.04 Regulation of apoptosis MX1 4599 Up 2.45 0.002 Regulation of apoptosis EDN1 1906 Up 6.86 0.04 Regulation of cell migration APOH 350 Up 6.15 0.019 Regulation of cell migration LAMA4 3910 Up 5.05 0.004 Regulation of cell migration PARD6B 84612 Up 3.54 0.019 Regulation of cell migration EGFR 1956 Up 1.51 0.01 Cell proliferation MDM2 4193 Up 2.63 0.04 Cell cycle, apoptosis RAB54B 25788 Down 2.93 0.001 Cell cycle UBC 7316 Down 2.34 0.03 Cell cycle FBXO5 26271 Down 2.22 0.03 Cell cycle HDAC6 10013 Down 2.22 0.04 Cell Cycle SMC1A 8243 Down 2.10 0.02 Cell cycle KIF2C 11004 Down 2.70 0.001 Cell cycle, cell proliferation E2F1 1869 Down 2.13 0.01 Cell cycle, cell proliferation, apoptosis E2F2 1870 Down 2.61 0.04 Cell cycle, apoptosis CDC25C 995 Down 2.11 0.01 Cell cycle, cell proliferation DLG1 1739 Down 2.20 0.003 Cell proliferation IFNB1 3456 Down 2.19 0.02 Cell proliferation ZMYND11 10771 Down 2.14 0.004 Cell proliferation EPS15 2060 Down 2.07 0.03 Cell proliferation B2M 567 Down 1.56 0.008 Immune response TEGT 7009 Down 1.9 0.001 Regulation of apoptosis UACA 55075 Down 2.8 0.01 Regulation of apoptosis CAMK2D 817 Down 1.69 0.001 Regulation of cell growth