Identification of Genomic Targets of Transcription Factor Aebp1 and Its Role in Survival of Glioma Cells

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 mutations, 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

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 correlation 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

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

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

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

Denotes perturbed gene in primary GBM (9) Denotes perturbed gene in secondary GBM (9)

GAAAT motif. Hence when we scanned the enriched peaks tained part (TTTCT) of the alternate consensus motif, more of the congruent target genes and queried for highly repre- so in FABP4 where both motifs were present in tandem. It sented binding site using CisFinder algorithm, GAAAT was noteworthy that some congruent target genes also had containing motif was enriched in these genes. The resulting this motif present along with GAAAT, albeit noncontigu- position weight matrix (Supplementary Fig. S5A) was used ously in their 1 kb promoter. The distribution of these 2 to generate sequence logo (Fig. 2C). Some of the genes like enriched motifs in the promoters of the congruent 185 genes PCDHA2, zinc finger proteins, EPHA4, CA3, and is represented in Fig. 2E. CSRP2BP had this site represented more than 4 times within The motifs identified above by bioinformatics approach their 1 kb promoters. Those perturbed genes derived from were further experimentally confirmed by Luciferase report- the microarray data that did not have predicted AE-1 er assay with wild-type and mutant Aebp1 binding motifs in elements in their proximal promoters [227] was further the FABP4 promoter region containing a continuous stretch compared with congruent targets [185]. This comparison of GAAAT and TTTCT (Fig. 2F). U87MG cells were yielded 63 genes for which we further analyzed for an cotransfected with b-galactosidase construct to normalize alternate consensus motif.We found a TTTCTTTAT motif for transfection efficiency. This analysis showed that to be highly enriched in these genes, the position weight GAAAT site in particular was responsible for AEBP1 pro- matrix and logo of which is represented in Supplementary moter function. Mutant GAAAT motif showed considerable Fig. S5B and S2D, respectively. Interestingly, AE-1 element reduction in luciferase activity in comparison to TTTCT in the promoter region of both FABP4 and FABP7 con- motif mutation. A double mutant motif of GAAATTTCT

www.aacrjournals.org Mol Cancer Res; 10(8) August 2012 1043

Ladha et al.

AB C 100 A. Inside (85) 80 B. Promoter (92) 60 B A C. Downstream (8) 40 20 No. of probes No. 0 −6 −5 −4 −3 −2 −11TSS 23456 Position of probes relative to TSS (in kb) C D 2 2

Figure 2. Pie chart distribution of 1 congruent targets showing the bits 1 bits number of Aebp1 binding peaks. The annotation is defined by the 0 0 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 location of the enriched probe ′ ′ ′ ′ 5 3 5 3 relative to transcription start site as E F being inside, in the promoter and 60 10,000 downstream of the gene (A). B, GAAAT 9,000 depicts binding region of enriched TTTCT 50 8,000 best probes for congruent target 7,000 genes based on their relative 40 6,000 distance from transcription start 5,000 site (6kbtoþ6 kb). X-axis 30 4,000 denotes genomic region relative to Fold change in Fold 3,000 start site, whereas Y-axis shows relative light units/s relative 2,000 20 the number of probes enriched in a Number of genes 1,000 1 kb window. C, D, the top-scoring 0 10 motifs enriched in targets. The bar graph represents Aebp1 target genes showing either or both of 0 Wild type 123456789 pGL3 Basic TTTCT Motif these enriched motifs. X-axis Untransfected GAAAT Motif Number of motifs SV40 Promoter represents number of times each GAAATTTCT Motif Plasmid transfected motif is present in 1 kb promoter G and Y-axis represents number of Transcription Transcription Signaling Transferase activity Negative control genes having these motifs (E). IgG IP IgG IP IgG IP IgG IP IgG IP Functional promoter assay 24 hours posttransfection as shown in TP53INP1 FOXJ3 KLF4 MAPK13 HS3STB1 F. Chromatin immunoprecipitation was done (G) using either purified RIT2 LCMT2 ERBB2IP NPAS3 ATF6 rabbit IgG (lane IgG) as control or SIVA-1 with Aebp1 polyclonal antibody GAP43 Receptor activity ZMNYD11 PPARA (lane IP) from U87MG cells. ChIP- HTR3C DMTF PCR was done to validate Aebp1 MS4A7 CAMTA1 TCF12 Lipid metabolism targets using primers flanking AE-1

PLA2G12A element within -1Kb promoters of TCEB3 PIK3CB NFATC1 the genes. Genes were categorized Protein binding based on functional classes. GABRE RAD54B Genes (TP53INP1, DMTF, ASCL1 SDC1 ERBB2IP SIVA ITPR2 , and ) that did not MDM2 have AE-1 element in 1kb Cell growth SNX12 promoters were also assessed by NCKIP SD TNFAIP8 ARNT Immune response ChIP-PCR.

HNF4G RELN USP8 IL1F7

CNS development IRS2 ELF5 CDK6 UBE3A Apoptosis PAK2 KIF12 Microtubule development IL17A EPAS DNAH7 TBX15 Ion binding Adhesion

SOX1 CA3 CD93

CUTL1 SLC8A1

1044 Mol Cancer Res; 10(8) August 2012 Molecular Cancer Research

Genomic Targets and Role in Cell Survival of AEBP1

AB 6 Apoptosis 8 Cell cycle 4 6 4 2 2 0 0 −2 −2 Fold change Fold Fold change Fold − −4 4 −6 −6 −8

H1F0 MX1 PAK2 TJP2 APOH ARNTBIRC3BIRC5 CYLD EDN1 HIPK3INFB1 NRAS TP53UBC CUL4A EGLN2 ARL2 CDK6 KIF2C MDM2 PRR5 RHEBRPA4 CIAPIN1 TNFAIP3TNFAIP8 CDC20 CENPL CTCFL FBXO5 SMC1ASPAG5STAG1 USP22 TNFSF14 CDC25C CDKN1A MAPK13 PARD6B PTP4A1 SERPINB2 CAMK2D CDKN2CCSNK2A2 ERBB2IP TNFRSF10D CSRP2BP CD 8 Differentiation 8 Proliferation 6 6 4 4 2 2 0 −2 0 −4 −2 Fold change Fold − change Fold 6 −4 −8 −6

FZD8 RELN ATOH8 ITGB1 NGEF PTBP1 TPD52 IL4RIRS1IRS2 SCIN TES DICER1 GNA13 LAMA4 SPATA6 CTGFDAB2DLG1EGFR ITGB3 MKI67 PURA USP8 SEMA4F BLZF1CD164CDH13 EPS15FABP7GAP43HDAC1 MEF2B PDGFB SESN1 CACNA1A ENTPD5 PRIM2A VEGFC ZMYND11 EFTranscription factors

6 Transcription factors 4

−2 Fold change Fold −4

−6 Genes

−8

ATF6 DLX2 E2F2ELF5 ETV6FOXI3 KLF2KLF4 SOX1 TLX2 ASCL1ASCL2 CUTL1 DMTF1 EPAS1 KLF12 NR2F6PPARA TBX15TCF12 ARNTL2 CLOCKCREB1 NFATC1NFE2L3

Figure 3. qRT-PCR validation of differentially regulated genes belonging to following categories: apoptosis (A), cell cycle (B), differentiation (C), and proliferation (D). Panel E shows qRT-PCR validation of 25 transcription factors affected by AEBP1 downregulation. Binding sites of the same transcription factors were examined in their 1 kb promoters and heat map among these 25 transcription factors was generated (F). showed almost negligible promoter activity compared with Aebp1 regulates the expression of growth-associated the wild-type control. genes To experimentally confirm the validity of our ChIP-chip As discussed earlier, among the 734 differentially regu- data, we carried out ChIP-PCR analysis of randomly selected lated Aebp1 target genes (Supplementary Table S2) we 48 genes belonging to varied gene ontology (Fig. 2G). noticed several of the genes belonged to varied pathways; Flanking primers pairs were designed for the computation- 27 related to cell cycle, 13 to differentiation, 27 to prolif- ally predicted occupancy in the promoters of these genes eration, and 21 to apoptosis. The quantitative expression (Supplementary Table S1). As can be seen in Fig. 2G, the pattern of these genes following siRNA-mediated down- promoter sequences are indeed enriched in the Aebp1 ChIP- regulation of AEBP1 were further validated by real-time DNA as compared with preimmune control sample. We also PCR analysis from 3 independent biological replicates as amplified some of the genes that did not have GAAAT shown in Fig. 3A to D. Some important molecules related to sequence for Aebp1 binding (TP53INP1, ERBB2IP, cell cycle such as CDC20, CDC25C that promote mitosis DMTF, and SIVA-1). These genes showed no amplification (30, 31) are downregulated, whereas CDK6 and MDM2 are in the immunoprecipitated DNA sample. upregulated. It is also noteworthy that TP53, a tumor

www.aacrjournals.org Mol Cancer Res; 10(8) August 2012 1045

Ladha et al.

suppressor protein, is upregulated upon AEBP1 silencing. cognate DNA binding sites present in the promoters of the Again, ITGB1 as well as FZD8 involved in the process of transcription factor genes. Based on the bioinformatics differentiation and cancer are upregulated whereas NGEF is analysis, we constructed a transcription factor network of downregulated. The expressions of IRS1, EGFR, IL4R, the differentially regulated transcription factor genes, PDGFB, and NRAS that are widely implicated in promoting which is represented in Fig. 3F. We found that some of proliferation in cancer cells upon dysregulation of growth these genes also possess binding sites for many of the factor signaling are increased. DNA replication related gene transcription factors. Prominent among them are FOXJ3, PRIM2A is downregulated and so is the growth index gene ASCL2, CLOCK, ELF5, and NFATC1. The functional MKI67. Interestingly, apoptotic regulators TNFAIP3, relevance of this transcription factor network in the biology TNFAIP8, TNFFRSF10D, TNFSF14, and BIRC5 show of AEBP1 and more so in gliomagenesis needs further increase in their expression. Cellular hypoxia is a hallmark investigation. of cancer and ARNT that rescues cells from such a condition is downregulated upon AEBP1 gene downregulation. Cell proliferation and colony suppression assay It is also interesting to note that a large number of Because many of the genes associated with cell growth and transcription factors are differentially regulated upon AEBP1 apoptosis were perturbed in AEBP1 silenced cells, we exam- gene silencing (Fig. 1F) in addition to the genes affecting ined the growth promoting potential of Aebp1in 2 glioma growth phenotype. We also carried out real-time PCR cell lines, U87MG, and U138MG cells. For this purpose, analyses of these 25 transcription factors that were differ- AEBP1 was silenced using 100 nmol/L siRNA pool. Max- entially regulated in our microarray experiment (Supple- imum downregulation was observed at 72 and 48 hours mentary Table S2), which are presented in Fig. 3E. Among posttransfection for U87MG and U138MG cells, respec- these, transcription factors FOXJ3, ASCL2, ELF5, CLOCK, tively (Fig. 4A). Hence, siRNA pool was replenished every NFATC1, EPAS1, NFE2L3, and TBX15 do contain AE 1 60 hours in case of U87MG and 36 hours in the case of binding element in their promoters and therefore are dif- U138MG cell line for over a period of 9 days and assessed for ferentially regulated upon AEBP1 gene silencing. However, cellular proliferation by the MTT assay. There was no we observe alteration in expression of other transcription apparent change in cell viability during the ﬁrst 4 days of factors as well. It is quite likely that these transcription posttransfection. However, from 5th day onward there was a factors might cross-regulate each other by binding to their reduction in cell viability suggesting that silencing of AEBP1

A 120 P ≤ 0.0001 P ≤ 0.0001 B 2 ** 100 1.8 U87MG ** 1.6 80 1.4 ** 60 1.2 1 ** * ** 40

Percentage 0.8 20 *** 0.6 *** 0.4 0 siControl siAEBP1 siControl siAEBP1 Absorbance at 550 nm 0.2 U87MG U138MG 0 C 2 3456789 1.8 Time (d) U138MG 1.6 ** ** siControl siAEBP1 1.4 ** ** D 1.2 ** 1 ** 1234 0.8 0.6 0.4 0.2 Absorbance at 550 nm 0 23456789 U87MG U138MG Time (d) siControl siAEBP1

Figure 4. A, qRT-PCR for AEBP1 in si control and si AEBP1 transfected U87MG and U138MG cell lines. Maximum downregulation was observed at 72 and 48 hours postsilencing in both the cell lines, respectively. Values are the average of three independent experiments and represents P-value 0.0001 between si control and si AEBP1 treated U87MG and U138MG cell lines. B, C, cell proliferation assay, U87MG and U138MG cells were treated either with si AEBP1 or si control every 60 and 36 hours respectively, and scored for viable cells using MTT assay. Values are the average of 3 independent experiments and and indicate signiﬁcantly different levels between si control and si AEBP1 treatment of P 0.05 and P 0.002, respectively. D, colony suppression assay was done in U87MG and U138MG glioma cell lines with either sh control (1 and 3) or sh AEBP1 (2 and 4). After puromycin selection for 2 weeks, colonies were ﬁxed and stained using crystal violet.

1046 Mol Cancer Res; 10(8) August 2012 Molecular Cancer Research

Genomic Targets and Role in Cell Survival of AEBP1

ABU87MG U138MG siControl siAEBP1 siControl siAEBP1 4 4 4 4

10 0.04 5th day 0.36 10 0.0 5th day 2.83 10 1.32 5th day 0.53 10 0.25 5th day 10.54 3 3 3 3 10 10 10 10 2 2 2 2 10 10 10 10 1 1 1 1 10 10 10 10

0 0.43 0 10.83 0 0.57 0 11.91

10 0 101 102 103 104 10 0 1 2 3 4 10 0 1 2 3 4 10 0 1 2 3 4 4 4 10 10 10 10 10 104 10 10 10 10 104 10 10 10 10

10 0.02 7th day 0.1010 0.45 7th day 12.1610 1.26 7th day 0.7810 0.43 7th day 20.20 3 3 3 3 10 10 10 10 2 2 2 2 10 10 10 10 1 1 1 1 10 10 10 10 0.08 15.45 0.69 11.07 0 0 0 0

10 0 1 2 3 4 10 0 1 2 3 4 10 0 1 2 3 4 10 0 1 2 3 4 4 10 10 10 10 10 4 10 10 10 10 10 4 10 10 10 10 10 10 10 10 10 10 4 Propidium iodide 10 10 10

0.05 9th day 0.40 1.20 9th day 15.23 1.30 9th day 0.8210 0.81 9th day 38.32 3 3 3 3 10 10 10 10 2 2 2 2 10 10 10 10 1 1 1 1 10 10 10 10 0.43 22.24 0.91 7.22 0 0 0 0

10 0 1 2 3 4 10 0 1 2 3 4 10 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 100 101 102 103 104 Annexin-FITC CD 60 9 days 80 9 days 50 7 days 70 7 days 5 days 60 40 5 days 50 30 40 20 30 20 10 10 % of Annexin V-positive cells V-positive % of Annexin 0 cells V-positive % of Annexin 0 siControl siControl siAEBP1 siAEBP1 siControl siControl siAEBP1 siAEBP1

Early apoptotic Late apoptotic Early apoptotic Late apoptotic Early apoptotic Late apoptotic Early apoptotic Late apoptotic

Figure 5. A, B, apoptosis analysis using Annexin V-FITC/PI was done on 5th, 7th, and 9th day after transfection either with si control or si AEBP1in U87MG and U138MG cell lines. C, D, graphical representation of the number of cells in different phases of apoptosis in U87MG and U138MG, respectively. resulted in loss of proliferative potential or cell death (Fig. 4B serum deprivation and after siRNA mediated downregula- and C). Further colony suppression assay was done by tion, cells were analyzed for Annexin V/PI staining [Fig. 5A transfecting AEBP1-shRNA construct in U87MG and (U87MG) and B(U138MG)] as well as for DNA fragmen- U138MG cell lines. In comparison to mock shRNA treated tation by TUNEL assay [Fig. 6A (U87MG) and B cells, most of the AEBP1 silenced cells do not survive to form (U138MG)]. Annexin staining because of phosphatidyl colonies after >2 weeks of puromycin selection (Fig. 4D). serine externalization was observed in both the cell lines from 5th day of postsilencing. The percentage of cells Loss of AEBP1 function leads to apoptosis showing early apoptosis were 10.83%, 15.45%, and As described earlier, silencing of AEBP1 expression 22.24% in U87MG cells on 5th, 7th, and 9th day, respec- resulted in loss of cellular viability, which prompted us to tively. At same time points, the early apoptotic cells were explore the mechanism of this phenotype. For this purpose, 11.91%, 11.07%, and 7.2% in U138MG cells. Similarly, U87MG and U138MG glioma cells were synchronized by late apoptotic cells were 2.82%, 12.16%, and 15.23% for

www.aacrjournals.org Mol Cancer Res; 10(8) August 2012 1047

Ladha et al.

ABU87MG U138MG siControl siAEBP1 siControl siAEBP1 5th day 0.58 5th day 12.96 5th day 0.22 5th day 15.09

7th day 1.47 7th day 28.08 7th day 0.71 7th day 32.59

Figure 6. A, B, apoptosis analysis BrdUrd-FITC using TUNEL assay was done on 5th, 7th, and 9th day after 9th day 4.0 9th day 44.80 9th day 3.87 9th day 50.33 transfection either with si control or si AEBP1 in U87MG and U138MG cell lines. C, D, shows graphical representation of the same in U87MG and U138MG cell lines, respectively.

DNA area CD 50 60 siControl siControl 45 AEBP1 siAEBP1 50 si 40 35 40 30 25 30 20 20 15 10 10 % of TUNEL-positive cells TUNEL-positive % of

% of TUNEL-positive cells TUNEL-positive % of 5 0 0 5 days 7 days 9 days 5 days 7 days 9 days

U87MG cells and 10.54%, 20.20%, and 38.32% in identify the targets in an effort to understand the various U138MG cells at same time points postsilencing (Fig. 5A pathways that are intimately influenced by Aebp1. A total and B). Both FITC and propidium iodide staining of late number of 734 genes were found to be perturbed under apoptotic cells can also be visualized on 7th and 9th day, AEBP1-silenced condition in comparison to mock siRNA respectively. In TUNEL assay, we observed considerable treated U87MG cells. A large number of genes that were DNA fragmentation in both the cell lines upon AEBP1 gene differentially regulated belonged to categories such as cell silencing. For U87MG cells, the percentage was 12.96% on cycle, differentiation, proliferation, apoptosis, and tran- 5th day of postsilencing, which increased to 44.80% on 9th scription regulators. We followed up this study with day and in the case of U138MG cells, the percentage of ChIP-chip analysis to determine the occupancy of Aebp1 15.09% on 5th day increased to 50.33% on the 9th day of on the promoter sequences using Agilent human promot- postsilencing (Fig. 6A and B). Thus, these results clearly er tiling array. Initially, 11,659 genes were picked up for show that AEBP1 downregulation drives the glioma cell lines occupancy of Aebp1. However, on application of false toward apoptosis. discovery rate analysis (13), the list was reduced to 5810 genomic loci. Interestingly, only 185 genes were congruent between the gene list of differentially regulated genes Discussion and those with Aebp1 occupancy in their promoter Aebp1 is a transcription factor that has been studied in region. great detail toward its role in regulating adipogenesis by its Another effort of this work was to identify a consensus repressive action on aP2 promoter (26). Our observation Aebp1 binding element in the promoters of the identified showing an upregulation of AEBP1 in primary GBM (9) gene list. A systematic bioinformatics approach has indeed prompted us to undertake a genome wide approach to identified 863 GAAAT motifs present in a total number of

1048 Mol Cancer Res; 10(8) August 2012 Molecular Cancer Research

Genomic Targets and Role in Cell Survival of AEBP1

442 genes. We also found another motif TTTCT highly as AEBP1 targets in this study. Another important obser- enriched in the genes that did not contain the GAAAT motif vation made in this study is that when we compared the within their promoters. Some of the targets (118 genes) perturbed gene list identified upon AEBP1 silencing and containing GAAAT motif also had TTTCT consensus those genes, which are perturbed in human primary and motif. This bioinformatics/in silico identification of these secondary glioma tumor samples (9), we observed that putative Aebp1 binding elements were experimentally con- EGFR, MDM2, B2M, TEGT, UACA (primary), and firmed by functional promoter assays and it is evident that CAMK2D (secondary) to be represented in both the gene GAAAT plays a major role in Aebp1 mediated promoter lists (Table 1). Moreover, EGFR, MDM2, CAMK2D, and activity on FABP4 promoter (Fig. 2F). The sequence UACA have GAAAT motif in their 1 kb promoter regions TTTCT present in tandem to GAAAT seems to have an while B2M and TEGT have TTTCT motif in their –1kb additive role in the promoter function. Mutation in both of promoters. these sites resulted in a drastic abrogation of the promoter To gain an insight into the functional role of Aebp1 function. occupancy on the promoter of differentially regulated genes, To interpret the biological significance of Aebp1 in we compared ChIP-chip data with gene expression data of glioma, we undertook a systematic and comprehensive siRNA-mediated downregulated AEBP1 in U87MG cells. analysis of the gene expression data. Apart from differential Although 669 annotated genes were affected by AEBP1 regulation of genes related to cell cycle, differentiation, downregulation, Aebp1 occupied promoters of only 185 proliferation, apoptosis, and transcription regulators as men- genes. It is important to note that modulation of gene tioned previously; we also retrieved genes belonging to expression may not be because of cis-binding of the signaling pathways such as insulin, MAPK, WNT, TGF- transcription factor but its binding kilo bases away from b that are widely implicated in gliomagenesis. Some of the transcription start site can also affect changes in its important genes that are affected upon AEBP1 downregula- expression (29). Furthermore, the lack of an absolute tion and have established role in glioma are growth factors correlation between gene expression data and ChIP-chip EGFR, PDGFB, (5, 32) and growth factor signaling inter- may be also because of lack of total depletion of Aebp1 by mediates like NRAS, IRS1, and IL4R (33–35). Although, siRNA-mediated silencing. Even low concentrations of we observed enrichment of genes pertaining to activation of Aebp1 can suffice in such downregulated conditions. Also NF-kB and PI3Kinase signaling, upon depletion of AEBP1 there may be multiple sites of other transcription factors as in U87MG and U138MG cell lines, we did not observe any well as coactivators located far apart, which need to apparent change in either localization or expression of NF- coordinate to propel transcription. Hence, all the 5810 kB (data not shown). Aebp1 is also known to interact with genes whose promoters are occupied by Aebp1 may not be PI3Kinase inhibitor Pten and promote the process of dif- poised for transcription. Thus, it is not surprising that ferentiation in preadipocytes (36). PTEN is often deleted in mere promoter occupancy of Aebp1 has not translated primary GBM, both the cell lines employed in this study into transcriptional perturbationatmanyofthesegene were PTEN mutant (data not shown). Therefore, it is loci. A similar interpretation was proposed in the case of unlikely that this interaction has any significance in glio- genomic promoter occupancy of RUNX2 in osteosarcoma magenesis. Apart from signaling, our analysis also yielded an cells (50). array of molecules related to cell cycle. Cell-cycle regulator AEBP1 (/) null mice has been earlier studied in the CDC20 that is often upregulated in several malignancies context of adipose tissue metabolism and it is reported that (30) and indirectly regulated by p53, is downregulated under AEBP1/ mice display slow growth and suppressed AEBP1 downregulated condition, whereas p53 itself is survival with 75% embryonic lethality (36). In this context, upregulated. Interestingly CDC20 has also been associated perturbation of genes mainly involved in growth-related with prognostic behavior of glioma patients (37). Again ontologies under AEBP1 downregulated condition, CDK6, an established marker associated with GBM (38, prompted us to probe the role of AEBP1 in cellular survival. 39), is also affected by AEBP1 downregulation. Other Experiments on cell proliferation and colony suppression established prognostic markers that have been widely used assay provided direct evidence for such a role of Aebp1 in in GBM such as survin/BIRC5 (40) and MKI67 (41) are also U87MG and U138MG glioma cells. Moreover, these cells seen in our list of differentially regulated genes. Genes like were directed toward apoptosis as monitored by annexinV ITGB1, FZD8, and NGEF belonging to the ontology of staining and DNA fragmentation. An important question differentiation and involved in GBM are also targets of that we are presently addressing is what are the key molecule Aebp1. Amplification of MDM2 (5) and hypermethylation (s) and the pathway that leads to loss of cell viability and of genes such as TES, CDH13 have been reported in glioma initiation of apoptosis? In summary, a comprehensive anal- (42, 43). Differential expression of TPD52 and CDKN2C in ysis of the genomic targets of Aebp1 carried out in the glioma is also reported in literature (44, 45). A wide array of present study should form a basic framework for further genes involved in the initiation, progression or maintenance experimentation on the biological role of AEBP1 in of other cancers such as RELN, BIRC3, TNFAIP3, and gliomagenesis. TNFAIP8 (46–49) can also seen to be differentially regulated AEBP1 upon downregulation. Genes associated with hyp- Disclosure of Potential Conflicts of Interest oxia namely ARNT, BIRC3, and CTGF were also identified No potential conflicts of interest were disclosed.

www.aacrjournals.org Mol Cancer Res; 10(8) August 2012 1049

Ladha et al.

The costs of publication of this article were defrayed in part by the payment of page Grant Support advertisement This work was supported by grants from NMITLI program of Council of Scientiﬁc charges. This article must therefore be hereby marked in accordance with and Industrial Research (CSIR) and Department of Biotechnology (DBT), New 18 U.S.C. Section 1734 solely to indicate this fact. Delhi. MRSR is a JC Bose Fellow of Department of Science and Technology. The gene expression and ChIP-chip data from this study have been submitted to Gene Expression Omnibus (GEO) http://www.ncbi.nlm.nih.gov/geo/under accession no: Received October 10, 2011; revised May 21, 2012; accepted June 13, 2012; GSE18892. published OnlineFirst June 20, 2012.

References 1. Park CK, Jung JH, Park SH, Jung HW, Cho BK. Multifarious proteomic 21. Schmid CD, Perier R, Praz V, Bucher P. EPD in its twentieth year: signatures and regional heterogeneity in glioblastomas. J Neuro Oncol towards complete promoter coverage of selected model organisms. 2009;94:31–9. Nucleic Acids Res 2006;34(Database issue):D82–5. 2. Somasundaram K, Reddy SP, Vinnakota K, Britto R, Subbarayan M, 22. Karolchik D, Hinrichs AS, Furey TS. The UCSC Table browser data Nambiar S, et al. Upregulation of ASCL1 and inhibition of Notch retrieval tool. Nucleic Acids Res 2004;32(Database issue):D493–6. signaling pathway characterize progressive astrocytoma. Oncogene 23. Sharov AA, Ko MS. Exhaustive search for over-represented DNA 2005;24:7073–83. sequence motifs with CisFinder. DNA Res 2009;16:261–73. 3. Li A, Walling J, Ahn S, Kotliarov Y, Su Q, Quezado M, et al. Unsuper- 24. Crooks GE, Hon G, Chandonia JN, Brenner SE. Weblogo: a sequence vised analysis of transcriptomic profiles reveals six glioma subtypes. logo generator. Genome Res 2004;14:1188–90. Cancer Res 2009;69:2091–9. 25. Thomas-Chollier M, Defrance M, Medina-Rivera A, Sand O, Herrmann 4. Kroes RA, Dawson G, Moskal JR. Focused microarray analysis of C, Thieffry D, et al. RSAT: regulatory sequence analysis tools. Nucleic glycol-gene expression in human glioblastomas. J Neurochem 2007; Acid Res 2011;39(Web Server issue):W86–91. 1:14–24. 26. Ro HS, Roncari DA. The C/EBP-binding and adjacent sites regulate 5. Kleihues P, Ohgaki H. Primary and secondary glioblastomas: from expression of the adipose P2 gene in human preadiopocytes. Mol Cell concept to clinical diagnosis. Neuro Oncol 1999;1:44–51. Biol 1991;11:2303–06. 6. Zhu Y, Parada LF. The molecular and genetic basis of neurological 27. Hunt CR, Ro JH, Dobson DE, Min HY, Spiegelman BM. Adipocyte P2 tumors. Nat Rev Cancer 2002;2:616–26. Gene: developmental expression and homology of 50-flanking 7. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, sequences among fat cell-specific genes. Proc Natl Acad Sci U S A et al. Molecular subclasses of high-grade glioma predict prognosis, 1986;83:3786–90. delineate a pattern of disease progression, and resemble stages in 28. Majdalawieh A, Zhang L, Fuki IV, Rader DJ, Ro HS. Adipocyte enhanc- neurogenesis. Cancer Cell 2006;9 157–73. er-binding protein 1 is a potential novel atherogenic factor involved in 8. Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, et al. macrophage cholesterol homeostasis and inflammation. Proc Natl Integrated genomic analysis identifies clinically relevant subtypes of Acad Sci U S A 2006;103:2346–51. glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, 29. Farnham PJ. Insights from genomic profiling of transcription factors. and NF1. Cancer Cell 2010;17 98–110. Nat Rev Genet 2009;10:605–16. 9. Reddy SP, Britto R, Vinnakota K, Aparna H, Sreepathi HK, Thota B, 30. Kidokoro T, Tanikawa C, Furukawa Y, Katagiri T, Nakamura Y, Mat- et al. Novel glioblastoma markers with diagnostic and prognostic value suda K, et al. CDC20, a potential cancer therapeutic target, is nega- identified through transcriptome analysis. Clin Cancer Res 2008;14: tively regulated by p53. Oncogene 2008;27:1562–71. 2978–87. 31. Boutros R, Lobjois V, Ducommun B. CDC25 phosphatases in cancer 10. He GP, Muise A, Li AW, Ro HS. A eukaryotic transcriptional repressor cells: key players? Good targets? Nat Rev Cancer 2007;7:495–507. with carboxypeptidase activity. Nature 1995;378:92–6. 32. Calzolari F, Appolloni I, Tutucci E, Caviglia S, Terrile M, Corte G, et al. 11. Li Z, Szabolcs M, Terwilliger JD, Efstratiadis A. Prostatic intraepithelial Tumor progression and oncogene addiction in a PDGF-B-induced neoplasia and adenocarcinoma in mice expressing a probasin-Neu model of gliomagenesis. Neoplasia 2008;10:1373–82. oncogenic transgene. Carcinogenesis 2006;27:1054–67. 33. Jeuken J, van den Broecke C, Gijsen S, Boots-Sprenger S, Wesseling 12. Ren B, Dynlacht BD. Use of chromatin immunoprecipitation assays in P. RAS/RAF pathway activation in gliomas: the result of copy number genome wide location analysis of mammalian transcription factors. gains rather than activating mutations. Acta Neuropathol 2007;114: Methods Enzymol 2004;376:304–15. 121–33. 13. Benjamini Y, Hochberg Y. Controlling the false discovery rate: practical 34. Park S, Zhao D, Hatanpaa KJ, Mickey BE, Saha D, Boothman DA, and powerful approach to multiple testing. J R Stat Soc B (Method- et al. RIP1 activates PI3K-Akt via a dual mechanism involving NF- ological) 1995;57:289–300. kappaB-mediated inhibition of the mTOR-S6K-IRS1 negative feed- 14. R Development Core Team. R: A language and environment for back loop and down-regulation of PTEN. Cancer Res 2009;69: statistical computing. R Foundation for Statistical Computing, Vienna, 4107–11. Austria. ISBN 3-900051-07-0. Available from: www.R-project.org/ 35. Rahaman Shaik O, Vogelbaum Michael A, Haque Saikh J. Aberrant 15. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and Stat3 signaling via IL-4R/IL-13R activation in malignant glioma cells: display of genome wide expression patterns. Proc Natl Acad Sci U S A involvement of IL-13Ra2. Proc Amer Assoc Cancer Res 2005;65: 1998;95:14893–8. 2956–63. 16. Saldanha AJ. Java Treeview—extensible visualization of microarray 36. Ro HS, Zhang L, Majdalawieh A, Kim SW, Wu X, Lyons PJ, et al. data. Bioinformatics 2004;20:3246–8. Adipocyte enhancer-binding protein 1 modulates adiposity and energy 17. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. homeostasis. Obesity (Silver Spring) 2007;15:288–302. Gene ontology: tool for the unification of biology. The Gene Ontology 37. Bie L, Zhao G, Cheng P, Rondeau G, Porwollik S, Ju Y, et al. The Consortium. Nat Genet 2000;25 25–9. accuracy of survival time prediction for patients with glioma is 18. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative improved by measuring mitotic spindle checkpoint gene expression. analysis of large lists using DAVID bioinformatics resources. Nat Plos One 2011;6:e25631. Protoc 2009;4:44–57. 38. Costello JF, Plass C, Arap W, Chapman VM, Held WA, Berger MS, et al. 19. Huang da W, Sherman BT, Lempicki RA. Bioinformatics enrichment Cyclin-dependent kinase6 (CDK6) amplification in human gliomas tools: paths toward the comprehensive functional analysis of large identified using two-dimensional separation of genomic DNA. Cancer gene lists. Nucleic Acids Res 2009;37:1–13. Res 1997;57:1250–54. 20. Jiang C, Xuan Z, Zhao F, Zhang MQ. TRED: a transcriptional regulatory 39. Lam PY, Di Tomaso E, Ng HK, Pang JC, Roussel MF, Hjelm NM. element database, new entries and other development. Nucleic Acids Expression of p19INK4d, CDK4, CDK6 in glioblastoma multiforme. Br Res 2007;35 (Database issue):D137–40. J Neurosurg 2000;28–32.

1050 Mol Cancer Res; 10(8) August 2012 Molecular Cancer Research

Genomic Targets and Role in Cell Survival of AEBP1

40. Preusser M, Gelpi E, Matej R, Marosi C, Dieckmann K, Rossler€ K, et al. genes by oligonucleotide microarray and real-time quantitative PCR. J No prognostic impact of survivin expression in glioblastoma. Acta Neuro Oncol 2008;88:281–91. Neuropathol 2005;109:534–8. 46. Sato N, Fukushima N, Chang R, Matsubayashi H, Goggins M. Differ- 41. Johannessen A, Torp S. The clinical value of Ki-67/MIB-1 ential and epigenetic gene expression profiling identifies frequent labeling index in human astrocytomas. Pathol Oncol Res 2006;12: disruption of the RELN pathway in pancreatic cancers. Gastroenter- 143–7. ology 2006;130:548–65. 42. Mueller W, Nutt CL, Ehrich M, Riemenschneider MJ, von Deimling A, 47. Rossi D, Deaglio S, Dominguez-Sola D, Rasi S, Vaisitti T, Agostinelli C, van den Boom D, et al. Downregulation of RUNX3 and TES et al. Alteration of BIRC3 and multiple other NF-kB pathway genes in by hypermethylation in glioblastoma. Oncogene 2007;26:583–93. splenic marginal zone lymphoma. Blood 2011;118:4930–4. 43. Piperi C, Themistocleous MS, Papavassiliou GA, Farmaki E, Levidou 48. Honma K, Tsuzuki S, Nakagawa M, Tagawa H, Nakamura S, Mor- G, Korkolopoulou P, et al. High incidence of MGMT and RARb pro- ishima Y, et al. TNFAIP3/A20 functions as a novel tumor suppressor moter methylation in primary glioblastomas: association with histo- gene in several subtypes of non-Hodgkin lymphomas. Blood pathological characteristics, inflammatory mediators and clinical out- 2009;114:2467–75. come. Mol Med 2010;16:1–9. 49. Dong QZ, Zhao Y, Liu Y, Wang Y, Zhang PX, Jiang GY, et al. Over- 44. Petrova DT, Asif AR, Armstrong VW, Dimova I, Toshev S, Yaramov N, expression of SCC-S2 correlates with lymph node metastasis and et al. Expression of chloride intracellular channel protein 1 (CLIC1) and poor prognosis in patients with non-small-cell lung cancer. Cancer Sci tumor protein D52 (TPD52) as potential biomarkers for colorectal 2010;101:1562–9. cancer. Clin. Biochem 2008;41:1224–36. 50. van der Deen M, Akech J, Lapointe D, Gupta S, Young DW, Montecino 45. Scrideli CA, Carlotti CG Jr, Okamoto OK, Andrade VS, Cortez MA, MA, et al. Genomic promoter occupancy of Runt-related transcription Motta FJ, et al. Gene expression profile analysis of primary glioblas- factors RUNX2 in osteosarcoma cells identifies gene involved in cell tomas and non-neoplastic brain tissue: identification of potential target adhesion and motility. J Biol Chem 2012;287:4503–17.

www.aacrjournals.org Mol Cancer Res; 10(8) August 2012 1051

Identification of Genomic Targets of Transcription Factor Aebp1 and its role in Survival of Glioma Cells

Jayashree Ladha, Swati Sinha, Vasudeva Bhat, et al.

Mol Cancer Res 2012;10:1039-1051. Published OnlineFirst June 20, 2012.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-11-0488

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2012/06/20/1541-7786.MCR-11-0488.DC1

Cited articles This article cites 31 articles, 7 of which you can access for free at: http://mcr.aacrjournals.org/content/10/8/1039.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/10/8/1039.full#related-urls

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 Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/10/8/1039. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.