Overrepresentation of transcription factor families in the genesets underlying breast cancer subtypes
© Joshi et al.; licensee BioMed Central Ltd. 2012
Received: 14 October 2011
Accepted: 17 February 2012
Published: 22 May 2012
The human genome contains a large amount of cis-regulatory DNA elements responsible for directing both spatial and temporal gene-expression patterns. Previous studies have shown that based on their mRNA expression breast tumors could be divided into five subgroups (Luminal A, Luminal B, Basal, ErbB2+ and Normal-like), each with a distinct molecular portrait. Whole genome gene expression analysis of independent sets of breast tumors reveals repeatedly the robustness of this classification. Furthermore, breast tumors carrying a TP53 mutation show a distinct gene expression profile, which is in strong association to the distinct molecular portraits. The mRNA expression of 552 genes, which varied considerably among the different tumors, but little between two samples of the same tumor, has been shown to be sufficient to separate these tumor subgroups.
We analyzed in silico the transcriptional regulation of genes defining the subgroups at 3 different levels: 1. We studied the pathways in which the genes distinguishing the subgroups of breast cancer may be jointly involved including upstream regulators (1st and 2nd level of regulation) as well as downstream targets of these genes. 2. Then we analyzed the promoter areas of these genes (−500 bp to +100 bp relative to the transcription start site) for canonical transcription binding sites using Genomatix. 3. We looked for the actual expression levels of the identified TF and how they correlate with the overrepresentation of their TF binding sites in the separate groups. We report that promoter composition of the genes that most strongly predict the patient subgroups is distinct. The class-predictive genes showed a clearly different degree of overrepresentation of transcription factor families in their promoter sequences.
The study suggests that transcription factors responsible for the observed expression pattern in breast cancers may lead us to important biological pathways.
Previous studies have shown that breast tumors can be divided into five subgroups (Luminal A, Luminal B, Normal-like, ErbB2 over-expressing, and Basal-like) based on their mRNA expression patterns . These patterns have been validated in independent datasets representing different laboratories, platforms and different patient cohorts . Survival analyses on a sub-cohort of patients with locally advanced breast cancer showed a significant difference in outcome of the patients in the various expression subgroups, with poor prognosis for the ErbB2+ and basal-like subtypes . The expression of 552 genes, the intrinsic gene list, has been suggested to be sufficient to separate breast carcinomas into the five distinct subgroups. What mechanisms of common regulation make these genes cluster together? We have previously shown that we can separate the patient clusters based only on the promoter composition of single binding sites in the promoters of the genes from the intrinsic gene list . However, regulation of gene expression in eukaryotes is highly complex and depends on sets of TFs rather than individual TFs  and in this study we attempt to characterize the overrepresentation of entire TF families. The promoter composition of the genes is one of the major determinants of gene regulation including multiple transcription binding sites that interact with a specific combination of transcription factors (TF). Eukaryotes achieve this diversity by combining a small number of transcription factors whose activities are modulated by diverse sets of conditions . Different functionalities can be conferred on one TF by its association with different co-factors. These factors may act as global TFs that assist their gene-specific partners in their function, and may thus activate or repress transcription depending on the partner motif and the condition . Analyzing transcription network dynamics in yeast, Luscombe et al. showed that, in response to diverse stimuli, transcription factors may alter their interaction patterns to varying degree, thereby rewiring the network . While few transcription factors serve as permanent hubs, most of them act transiently during certain conditions. Exogenous processes like environmental responses facilitated fast signal transductions to multiple genes with short regulatory cascades, whereas endogenous processes needed to progress through multiple stages with a complex combination of TFs to fewer target genes . The same TFs may act both in endogenous and exogenous processes. Regulatory hubs targeting disproportionately large numbers of genes and thereby representing the most influential components of a network- have been described. Both Pilpel  and Luscombe  concluded that precise regulation of a condition cannot arise from the specificity of individual TFs, therefore combinatorial TF usage seems to be the key. The NF-κB family of TFs is an example of transcription regulators that are activated by both intra- and extra-cellular stimuli such as cytokines, oxidant-free radicals, ultraviolet irradiation, and bacterial or viral products . Aberrant NF-κB activity has been implicated in carcinogenesis and in the control of cellular response to anti-cancer agents. Activated NF-κB was detected predominantly in ER-negative breast tumors, and mostly in the ErbB2over-expressing tumor subgroup .
The in silico analysis of the transcriptional regulation of genes defining the subgroups was performed at three different levels: (1) Study of the pathways in which the genes distinguishing the subgroups of breast cancer may be jointly involved including upstream regulators (1st and 2nd level of regulation) as well as downstream targets of these genes. (2) Then we analyzed the promoter areas of these genes (−500 bp to +100 bp relative to the transcription start site) for canonical transcription binding sites using Genomatix. (3) We looked for the actual expression levels of the identified TF and how they correlate with the overrepresentation of their TF binding sites in the separate groups.
Selection of genes
The expression of 552 genes, the intrinsic gene list, which has been suggested to be sufficient to separate breast carcinomas into the five distinct subgroups defined in  and [2, 9] was used for the pathway analysis in this study (referred to as full list). A subset consisting of 197 genes  that best represented the classification scheme in breast cancer (referred as top list) were selected from the intrinsic list, and used in the promoter analysis part (Additional file 1: Table S1).
Pathway analysis was performed using Pathway Studio  from Ariadne Genetics. Two network prediction algorithms were used that allow to discover the patterns of gene expression inherent in the experimental data: Pearson Correlation and Auto Net Finder network prediction algorithm. Pathway Studio’s text mining tools were applied to extract biological associations by mining PubMed to build pathways from extracted facts using data from recent publications and public and commercial databases such as KEGG, BIND, GO, and the PathArt database of curated signaling and disease pathways. The algorithm for building Correlation Network in Pathway Studio is based on Pearson Correlation. Genes with similar expression profiles are connected with edges indicating the significance of the correlation. The group of tightly correlated genes form cluster in the correlation network. The algorithm can be used for clustering genes according to their expression profiles across multiple samples. The tool calculates correlation coefficients between all pairs of gene expression profiles measured in the experiment and outputs clusters of highly correlated genes. Identified gene clusters can be further validated and analyzed using relations from the database that have been extracted from the literature by Ariadne Genetics. Auto Net Finder is a network estimation system that combines hierarchical clustering and Graphical Gaussian Modeling and is used for distinguishing direct and indirect relationship among variables. Bibliosphere pathways (release 7.1)  (http://www.genomatix.de, Genomatix Software GmbH) was used for extracting the associations between gene, transcription factor and proteins corresponding with the genesets defining each molecular subtype of breast cancer. Genomatix Bibliosphere is a knowledge database consisting of manually curated co-cited genes in PubMed, which additionally provides information about the presence of TFBS in their promoters, using in silico tool- MatInspector, interactions and associated pathways from Molecular Interactions database-NetPro and BioCyc, respectively.
Analysis of overrepresentation of TFBS families in the promoter sequences
In terms of fold factor of overrepresentation compared to the backgroundFold factor of TFBS overrepresentation was calculated by a formula as mentioned below:(1)
- 2.As z-scores that provide a measure of the distance of sample from the reference population mean. Here sample refers to the number of observed hits of any particular TFBS in a given input set of sequences and reference refers to the number of hits of the same TFBS in equally sized human genomic promoter sequence population.(2)
z(X) is a z-score of overrepresentation of a transcription factor binding site family (X);
n obs (X) is a number of observed hits of X in an input promoter sequences;
n exp (X) is expected number of hits of X in an equally sized sample sequences in human genomic promoter background;
S(X) is a population standard deviation of number of hits of X
We used Genomatix RegionMiner tool (Genomatix Software GmbH, http://www.genomatix.de) in order to evaluate the degree of TFBS family overrepresentation. The histogram of z-scores of each TFBS motif families in each subtype-specific promoter sequences is shown in the Additional file 2: Figure S1. Histograms like this indicate that choosing the cut-off level of 2.0 allows identifying TFBS families that are overrepresented. However, z-score cut-off level of 2.0 does not provide a precise measure of significance, because of the disparity of sample size between sample and reference. Due to the copyright and technical limitations in accessing the Transfac database, further statistical testing of over-representation could not be performed within that tool.
Under-representations or absence of TFBS family motifs in sub-type specific genes may occur due to a fewer number of subtype-representative genes and subsequently a smaller number of promoter sequences used for any particular subtype. This can be a source of false positivity. Therefore we have not taken into account the under-representations of TFBS family motifs in this analysis.
Principal component analysis to identify TFBS with maximum variance between subtypes
Principal component analysis (PCA)  was performed for ranking the TFBS families with respect to the variance of fold-factor overrepresentation contributed by them between five subtypes. We prepared a matrix of TFBS fold-factors for subtypes, with subtypes as columns and TFBS families as rows. We performed PCA on this matrix using the princomp function of Matlab. Subtracting each data point from the column mean represents a center of this matrix. Hotelling’s T 2 statistic was used as a measure of multivariate distance of each TFBS family from the center of the TFBS fold-factor matrix as described in http://www.mathworks.com/help/toolbox/stats/princomp.html.
Gene expression data
mRNA expression of the studied TF
Transcription factor families with overrepresentation z-score >2.0 were mapped to their corresponding probes in the mRNA expressions dataset. By applying multiclass SAM, we extracted 120 TF genes with significantly different (at the FDR <0.1) expression between the five subtypes. Pearson’s correlation between the subtype-specific geometric mean expression of this subset of transcription factor genes and fold overrepresentation was computed. The justification of using geometric mean instead of arithmetic mean is that typically mRNA expression values are log-normally distributed.
Results and discussion
Pathway analysis of the genes that define the five breast cancer subgroups
Using Pathway Studio from Ariadne Genetics, we studied the direct interactions between the genes with distinguished gene expression pattern in the breast cancer subgroups as described in Materials and Methods, selection of genes. Most profound direct interactions were observed for the genes defining the luminal A group with protein-protein interactions between XBP1 and ESR1 and CCND1 (Additional file 3: Figure S2). Trefoil (TFF3) has been functionally coupled to CCND1 through angiotensin receptor 1 (AGTR1). Angiotensin II is converted from its precursor by angiotensin I-converting enzyme (ACE) and has been shown to mediate growth in breast cancer cell lines via ligand-induced activity through the angiotensin II type 1 receptor (AGTR1). We also searched for upstream regulators as well as downstream targets of these genes. Downstream targets could be observed centered at the ESR1, MYC, NFKB1, GATA3, CCND1, TP53 and MSX2/FOXC1 (Additional file 4: Figure S3).
A somewhat less organized pathway structure is observed in the luminal B subclass. The ESR1 node was not observable and the TP53 network was more sparse with fewer partner genes. Novel nodes were centered at NRG1, GSTP1 and CUL1 (Additional file 5: Figure S4), CUL1 has homology to yeast Cdc53, which is part of a complex known as SCF that mediates the ubiquitin-dependent degradation of G1 cycles and cyclin-dependent kinase inhibitors, while NRG1 contains a domain related to the epidermal growth factor family of ligands and can act as receptor agonists. The direct interactions between genes highly expressed in Luminal B subtype were observed between GSTP1 and CDK2AP1, S100A10 and S100A11 and PPP1R13B and TP53BP2. The latter protein interacts with TP53 to specifically enhance p53-induced apoptosis but not cell cycle arrest.
Four distinct regulatory nodes were observed in the ERBB2 group: around the ERBB2 itself, TP53, NFKB1 and CTNNB1 (cadherin-associated protein, beta 1) (Additional file 6: Figure S5). NFkB-p65 was shown to repress β-catenin-activated transcription of cyclin D1 . Moreover, a direct interaction is established between ERBB2 and GRB7 (Additional file 3: Figure S2). The solution structure of the Grb7-SH2/erbB2 peptide complex was described and suggested to be involved in cell signaling pathways that promote the formation of metastases and inflammatory responses. PPARBP, which is co-amplified with ERBB2, has in early studies been suggested to play a role in mammary epithelial differentiation and in breast carcinogenesis by its ability to function as ESR1 coactivator. It was shown to contain a typical CCAT box and multiple cis-elements such as C/EBPbeta, YY1, c-ETS-1, AP1, AP2, and NFkappaB binding sites. The 4 different regulatory nodes are connected by FLOT2, the human epidermal surface antigen involved in epidermal cell adhesion. NFKB1 was present in the network for the Basal group, where also the FOX family, a whole family of cyclins and CDK2, and CDK6 and isoforms of protein kinase (RPS6K) were present (Additional file 7: Figure S6). Interestingly, a large number of connections lead to GJA1 (Cap junction protein, alpha, also known as connexin 43). Other distinct nodes around TP53 are those connecting to KRT5, MAPK signalling, E2F1 and NCL. NCL, Nucleolin, one of the most abundant nucleolar proteins, has been recently shown to be involved in the reprogramming of somatic cells for derivation of either embryonic stem (ES) cells, by somatic cell nuclear transfer (SCNT), or ES-like cells, by induced pluripotent stem (iPS) cell procedure. Nucleolar proteins are proposed to be the markers of activation of embryonic genes  and provide mechanism for nucleolar control of progression of cell cycle in stem cells and cancer cells . TP53 was a central node in the regulatory network of the normal-like subgroup, surrounded by JUN, ACSS2, ACSL1, KRT13, PIK3R1 and other nodes some representing glycolysis, energy metabolism, pyruvate metabolism and metabolism of carbohydrate (Additional file 8: Figure S7).
Over-representation of specific transcription factor binding sites in the promoter of the genes that distinguish the subtypes
Presence of promoter modules in genes that define the ErbB2+ subgroup
Over-representation of TP53 mutations in the tumors that belong to the ErbB2+ and basal-like subgroups
In human breast tumors, the two tumor subgroups exhibiting the most prominent activation of putative NF-κB target genes (ErbB2+ and Basal-like) also harbored the highest frequency of p53 mutations. 86% of the patients in the ErbB2+ subgroup had TP53 mutations in their tumors and all the genes that are abnormally expressed in this tumor type have NF-κB binding sites in their promoter (Figure 3C). There is an evidence that NF-κB can regulate TP53 expression and that NF-κB is required for TP53-dependent cell death . In turn, TP53 activates NF-κB through the RAF/MEK1/p90 pathway . The TP53 protein interacts with NF-κB and enhances its transcriptional activity and its anti-apoptotic efficacy. Over-expression of ErbB2 is known to induce the classical NF-κB pathway [31, 32]. The estrogen receptor (ER) can bind physically to NF-κB to inhibit its DNA binding functions, hitherto repressing gene expression . Therefore the NF-κB pathway was shown to be a major stroma-tumor signaling mediator in ER negative tumors with over-expression of ErbB2 . NF-κB signaling has been associated with doxorubicin resistance, and agents blocking NF-κB function have been proven beneficial in the treatment of tumors in combination with standard anti-cancer therapies .
Over-represented transcription factor families within the promoter sequences
By using the Wilcoxon rank sum test, we observed significantly elevated mRNA expressions of ESR1 and PGR in Luminal A or Luminal samples compared to the basal ones (p < 1.0e-6), with non-significant differences in ERBB2 expressions. As expected ERBB2 was significantly upregulated in ERBB2+ tumors along with downregulated ESR1 and PGR, compared to the rest (p < 1.0e-4). Regulation by many transcription factors shown overrepresented here in ER + ve or ER-ve subtypes is not well characterized in context of estrogen and progesterone receptor activity. However, overrepresentation of some of the TFBS, such as GATA, BTBF, NF Kappa B – appear to be consistent with prevailing knowledge about the subtypes and their ER/PR or Her2 status.
Thus functions of the TF genes corresponding to the over-represented TFBS families hint the predominant characteristics of the subtypes. Findings from the above in silico analysis will be further validated in reporter studies and ChIP analyses. The approach of identifying overrepresented TFBS in a set of coordinately expressed genes under a particular disease class or condition can improve the specificity and noise tolerance. . However, its main limitation is that it does not account for the role of local chromatin environment constituted by structural properties, epigenetic modification etc. The local chromatin environment can offer condition-specific functionality to the existing TFBSs in a set of promoters.
Promoter sequences extending from 500 bp upstream to 100 bp downstream relative to TSS typically contain core promoter elements, CpG islands, downstream promoter element and other components of transcriptional machinery. Besides, this region has been demonstrated to have high density of positional as well as comparative TFBS , many of which are typically location sensitive. Thus limiting the analysis to this proximal promoter region, rather than analyzing the broader region (i.e. -1000 bp to +500 bp relative to the TSS) – could reduce false positives in TFBS overrepresentation. However, by that very limitation we may omit important information about second alternative promoters and distant control loci, which are therefore outside the scope of this analysis.
Correlation between actual abundance of TFs and frequency of their BS in the genes defining the clusters
Biological uncertainty in a correlation between the abundance of TFs and frequency of their BS might be attributed to several factors. The most common and obvious reason could be mutant or copy number altered TF. Moreover, here we have not accounted for the expressions of downstream targets of the TFs. It is noteworthy that mutations (point mutation and copy number alteration) in TFs can also have an impact on the level of expression of the downstream genes. For instance, a mutant TP53, which is still highly expressed, may not recognize the original binding sites anymore, leading to a drop in the expression of the target genes.
Here we report that the promoter composition of the genes that strongly predict the patient subgroups is distinct. The gene classes showed a clear separation when based solely on their promoter composition. This finding suggests that studying those transcription factors associated to the observed expression pattern in breast cancers may lead us to important biological pathways responsible for the regulation of gene expression in breast cancer.
HJ is a fellow of the Health Authority of South-East Norway. SHN is a fellow of the Norwegian cancer society (Den Norske Kreftforening).
transcription factor binding site
principal component analysis
Human Epidermal Growth Factor Receptor 2.
This work was supported by grants D-99061 and D-03067 from The Norwegian Cancer Society, grant 152004/150 from The Functional Genomics program (FUGE), The Norwegian Research Council (NFR) and grant 155218/300 from NFR.
- Perou CM, Sørlie T, Eisen MB, Van De Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lønning PE, Børresen-Dale AL, Brown PO, Botstein D: Molecular portraits of human breast tumours. Nature. 2000, 406: 747-752. 10.1038/35021093.View ArticlePubMedGoogle Scholar
- Sørlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, Deng S, Johnsen H, Pesich R, Geisler S, Demeter J, Perou CM, Lønning PE, Brown PO, Børresen-Dale A-L, Botstein D: Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci USA. 2003, 100: 8418-8423. 10.1073/pnas.0932692100.PubMed CentralView ArticlePubMedGoogle Scholar
- Tongbai R, Idelman G, Nordgard SH, Cui W, Jacobs JL, Haggerty CM, Chanock SJ, Børresen-Dale A-L, Livingston G, Shaunessy P, Chiang C-H, Kristensen VN, Bilke S, Gardner K: Transcriptional networks inferred from molecular signatures of breast cancer. Am J Pathol. 2008, 172: 495-509. 10.2353/ajpath.2008.061079.PubMed CentralView ArticlePubMedGoogle Scholar
- Elkon R, Linhart C, Sharan R, Shamir R, Shiloh Y: Genome-wide in silico identification of transcriptional regulators controlling the cell cycle in human cells. Genome Res. 2003, 13: 773-780. 10.1101/gr.947203.PubMed CentralView ArticlePubMedGoogle Scholar
- Pilpel Y, Sudarsanam P, Church GM: Identifying regulatory networks by combinatorial analysis of promoter elements. Nat Genet. 2001, 29: 153-159. 10.1038/ng724.View ArticlePubMedGoogle Scholar
- Luscombe NM, Babu MM, Yu H, Snyder M, Teichmann SA, Gerstein M: Genomic analysis of regulatory network dynamics reveals large topological changes. Nature. 2004, 431: 308-312. 10.1038/nature02782.View ArticlePubMedGoogle Scholar
- Chen F, Castranova V, Shi X: New insights into the role of nuclear factor-kappaB in cell growth regulation. 2001, 159: 387-397.Google Scholar
- Biswas DK, Shi Q, Baily S, Strickland I, Ghosh S, Pardee AB, Iglehart JD: NF-kappa B activation in human breast cancer specimens and its role in cell proliferation and apoptosis. Proc Natl Acad Sci USA. 2004, 101: 10137-10142. 10.1073/pnas.0403621101.PubMed CentralView ArticlePubMedGoogle Scholar
- Sørlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, Van De Rijn M, Jeffrey SS, Thorsen T, Quist H, Matese JC, Brown PO, Botstein D, Lønning PE, Børresen-Dale A-L: Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA. 2001, 98: 10869-10874. 10.1073/pnas.191367098.PubMed CentralView ArticlePubMedGoogle Scholar
- Muggerud AA, Johnsen H, Barnes DA, Steel A, Lønning PE, Naume B, Sørlie T, Børresen-Dale A-L: Evaluation of MetriGenix custom 4D™arrays applied for detection of breast cancer subtypes. BMC Cancer. 2006, 6: 59-10.1186/1471-2407-6-59.PubMed CentralView ArticlePubMedGoogle Scholar
- Nikitin A: Pathway studio–the analysis and navigation of molecular networks. Bioinformatics. 2003, 19: 2155-2157. 10.1093/bioinformatics/btg290.View ArticlePubMedGoogle Scholar
- Scherf M, Epple A, Werner T: The next generation of literature analysis: Integration of genomic analysis into text mining. Brief Bioinform. 2005, 6: 287-297. 10.1093/bib/6.3.287.View ArticlePubMedGoogle Scholar
- Cartharius K, Frech K, Grote K, Klocke B, Haltmeier M, Klingenhoff A, Frisch M, Bayerlein M, Werner T: MatInspector and beyond: promoter analysis based on transcription factor binding sites. Bioinformatics. 2005, 21: 2933-2942. 10.1093/bioinformatics/bti473.View ArticlePubMedGoogle Scholar
- Jolliffe IT: Principal Component Analysis. Chemom Intell Lab Syst. 1986, 2: 37-52.Google Scholar
- Enerly E, Steinfeld I, Kleivi K, Leivonen S-K, Aure MR, Russnes HG, Rønneberg JA, Johnsen H, Navon R, Rødland E, Mäkelä R, Naume B, Perälä M, Kallioniemi O, Kristensen VN, Yakhini Z, Børresen-Dale A-L: miRNA-mRNA Integrated Analysis Reveals Roles for miRNAs in Primary Breast Tumors. PLoS One. 2011, 6: 13-View ArticleGoogle Scholar
- Parker JS, Mullins M, Cheang MCU, Leung S, Voduc D, Vickery T, Davies S, Fauron C, He X, Hu Z, Quackenbush JF, Stijleman IJ, Palazzo J, Marron JS, Nobel AB, Mardis E, Nielsen TO, Ellis MJ, Perou CM, Bernard PS: Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol. 2009, 27: 1160-1167. 10.1200/JCO.2008.18.1370.PubMed CentralView ArticlePubMedGoogle Scholar
- Hwang I, Choi YS, Jeon M-Y, Jeong S: NF-κB p65 represses β-catenin-activated transcription of cyclin D1. Biochem Biophys Res Commun. 2010, 403: 79-84. 10.1016/j.bbrc.2010.10.118.View ArticlePubMedGoogle Scholar
- Johansson H, Svensson F, Runnberg R, Simonsson T, Simonsson S: Phosphorylated nucleolin interacts with translationally controlled tumor protein during mitosis and with Oct4 during interphase in ES cells. PLoS One. 2010, 5: e13678-10.1371/journal.pone.0013678.PubMed CentralView ArticlePubMedGoogle Scholar
- Tsai RYL, McKay RDG: A nucleolar mechanism controlling cell proliferation in stem cells and cancer cells. Genes Dev. 2002, 16: 2991-3003. 10.1101/gad.55671.PubMed CentralView ArticlePubMedGoogle Scholar
- Kel-Margoulis OV, Romashchenko AG, Kolchanov NA, Wingender E, Kel AE: COMPEL: a database on composite regulatory elements providing combinatorial transcriptional regulation. Nucleic Acids Res. 2000, 28: 311-315. 10.1093/nar/28.1.311.PubMed CentralView ArticlePubMedGoogle Scholar
- Klingenhoff A, Frech K, Quandt K, Werner T: Functional promoter modules can be detected by formal models independent of overall nucleotide sequence similarity. Bioinformatics. 1999, 15: 180-186. 10.1093/bioinformatics/15.3.180.View ArticlePubMedGoogle Scholar
- Fessele S, Maier H, Zischek C, Nelson PJ, Werner T: Regulatory context is a crucial part of gene function. Trends Genet. 2002, 18: 60-63. 10.1016/S0168-9525(02)02591-X.View ArticlePubMedGoogle Scholar
- Oikawa T, Yamada T: Molecular biology of the Ets family of transcription factors. Gene. 2003, 303: 11-34.View ArticlePubMedGoogle Scholar
- Oikawa T: ETS transcription factors: possible targets for cancer therapy. Cancer Sci. 2004, 95: 626-633. 10.1111/j.1349-7006.2004.tb03320.x.View ArticlePubMedGoogle Scholar
- Sampath J, Sun D, Kidd VJ, Grenet J, Gandhi A, Shapiro LH, Wang Q, Zambetti GP, Schuetz JD: Mutant p53 cooperates with ETS and selectively up-regulates human MDR1 not MRP1. J Biol Chem. 2001, 276: 39359-39367. 10.1074/jbc.M103429200.View ArticlePubMedGoogle Scholar
- Dittmer J: The Biology of the Ets1 Proto-Oncogene. Mol Cancer. 2003, 2: 29-10.1186/1476-4598-2-29.PubMed CentralView ArticlePubMedGoogle Scholar
- De Siervi A, De Luca P, Moiola C, Gueron G, Tongbai R, Chandramouli GVR, Haggerty C, Dzekunova I, Petersen D, Kawasaki E, Kil WJ, Camphausen K, Longo D, Gardner K: Identification of new Rel/NFkappaB regulatory networks by focused genome location analysis. Cell cycle Georgetown Tex. 2009, 8: 2093-2100. 10.4161/cc.8.13.8926.View ArticleGoogle Scholar
- Gu L, Zhu N, Findley HW, Woods WG, Zhou M: Identification and characterization of the IKKalpha promoter: positive and negative regulation by ETS-1 and p53, respectively. J Biol Chem. 2004, 279: 52141-52149. 10.1074/jbc.M407915200.View ArticlePubMedGoogle Scholar
- Sementchenko VI, Watson DK: Ets target genes: past, present and future. Oncogene. 2000, 19: 6533-6548. 10.1038/sj.onc.1204034.View ArticlePubMedGoogle Scholar
- Ryan KM, Ernst MK, Rice NR, Vousden KH: Role of NF-kappaB in p53-mediated programmed cell death. Nature. 2000, 404: 892-897. 10.1038/35009130.View ArticlePubMedGoogle Scholar
- Guo G, Wang T, Gao Q, Tamae D, Wong P, Chen T, Chen W-C, Shively JE, Wong JYC, Li JJ: Expression of ErbB2 enhances radiation-induced NF-kappaB activation. Oncogene. 2004, 23: 535-545. 10.1038/sj.onc.1207149.View ArticlePubMedGoogle Scholar
- Pianetti S, Arsura M, Romieu-Mourez R, Coffey RJ, Sonenshein GE: Her-2/neu overexpression induces NF-kappaB via a PI3-kinase/Akt pathway involving calpain-mediated degradation of IkappaB-alpha that can be inhibited by the tumor suppressor PTEN. Oncogene. 2001, 20: 1287-1299. 10.1038/sj.onc.1204257.View ArticlePubMedGoogle Scholar
- Ray P, Ghosh SK, Zhang DH, Ray A: Repression of interleukin-6 gene expression by 17 beta-estradiol: inhibition of the DNA-binding activity of the transcription factors NF-IL6 and NF-kappa B by the estrogen receptor. FEBS Lett. 1997, 409: 79-85. 10.1016/S0014-5793(97)00487-0.View ArticlePubMedGoogle Scholar
- Wang CY, Cusack JC, Liu R, Baldwin AS: Control of inducible chemoresistance: enhanced anti-tumor therapy through increased apoptosis by inhibition of NF-kappaB. Nat Med. 1999, 5: 412-417. 10.1038/7410.View ArticlePubMedGoogle Scholar
- Kantidze OL, Kamalyukova IM, Razin SV: Association of the mammalian transcriptional regulator kaiso with centrosomes and the midbody. Cell cycle Georgetown Tex. 2009, 8: 2303-2304. 10.4161/cc.8.14.8944.View ArticleGoogle Scholar
- Tong G-X, Yu WM, Beaubier NT, Weeden EM, Hamele-Bena D, Mansukhani MM, O’Toole KM: Expression of PAX8 in normal and neoplastic renal tissues: an immunohistochemical study. Modern pathology an official journal of the United States and Canadian Academy of Pathology Inc. 2009, 22: 1218-1227.View ArticleGoogle Scholar
- Liu H, Hew HC, Lu Z-G, Yamaguchi T, Miki Y, Yoshida K: DNA damage signalling recruits RREB-1 to the p53 tumour suppressor promoter. Biochem J. 2009, 422: 543-551. 10.1042/BJ20090342.View ArticlePubMedGoogle Scholar
- Ritchie A, Gutierrez O, Fernandez-Luna JL: PAR bZIP-bik is a novel transcriptional pathway that mediates oxidative stress-induced apoptosis in fibroblasts. Cell Death Differ. 2009, 16: 838-846. 10.1038/cdd.2009.13.View ArticlePubMedGoogle Scholar
- Frank DA: STAT3 as a central mediator of neoplastic cellular transformation. Cancer Lett. 2007, 251: 199-210. 10.1016/j.canlet.2006.10.017.View ArticlePubMedGoogle Scholar
- Felty Q, Xiong W-C, Sun D, Sarkar S, Singh KP, Parkash J, Roy D: Estrogen-induced mitochondrial reactive oxygen species as signal-transducing messengers. Biochemistry. 2005, 44: 6900-6909. 10.1021/bi047629p.View ArticlePubMedGoogle Scholar
- Kunkle B, Felty Q, Trevino F, Roy D: Oncomine meta-analysis of breast cancer microarray data identifies upregulation of NRF-1 expression in human breast carcinoma. Distribution. 2009, 715-719.Google Scholar
- Ho Sui SJ, Mortimer JR, Arenillas DJ, Brumm J, Walsh CJ, Kennedy BP, Wasserman WW: oPOSSUM: identification of over-represented transcription factor binding sites in co-expressed genes. Nucleic Acids Res. 2005, 33: 3154-3164. 10.1093/nar/gki624.PubMed CentralView ArticlePubMedGoogle Scholar
- Tharakaraman K, Bodenreider O, Landsman D, Spouge JL, Mariño-Ramírez L: The biological function of some human transcription factor binding motifs varies with position relative to the transcription start site. Nucleic Acids Res. 2008, 36: 2777-2786. 10.1093/nar/gkn137.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.