The common stress responsive transcription factor ATF3 binds genomic sites enriched with p300 and H3K27ac for transcriptional regulation
Dysregulation of the common stress responsive transcription factor ATF3 has been causally linked to many important human diseases such as cancer, atherosclerosis, infections, and hypospadias. Although it is believed that the ATF3 transcription activity is central to its cellular functions, how ATF3 regulates gene expression remains largely unknown. Here, we employed ATF3 wild-type and knockout isogenic cell lines to carry out the first comprehensive analysis of global ATF3-binding profiles in the human genome under basal and stressed (DNA damage) conditions.
Although expressed at a low basal level, ATF3 was found to bind a large number of genomic sites that are often associated with genes involved in cellular stress responses. Interestingly, ATF3 appears to bind a large portion of genomic sites distal to transcription start sites and enriched with p300 and H3K27ac. Global gene expression profiling analysis indicates that genes proximal to these genomic sites were often regulated by ATF3. While DNA damage elicited by camptothecin dramatically altered the ATF3 binding profile, most of the genes regulated by ATF3 upon DNA damage were pre-bound by ATF3 before the stress. Moreover, we demonstrated that ATF3 was co-localized with the major stress responder p53 at genomic sites, thereby collaborating with p53 to regulate p53 target gene expression upon DNA damage.
These results suggest that ATF3 likely bookmarks genomic sites and interacts with other transcription regulators to control gene expression.
KeywordsATF3 ChIP-seq Enhancer p300 H3K27ac p53
The development of human diseases is often accompanied by changes in the gene expression landscape. Regulated mainly at the transcription level, gene expression is tightly controlled by transcription factors (TF) that bind not only promoters proximal to transcription start sites (TSS), but also distal cis-regulatory elements (i.e., enhancers) that are far removed from TSS [1, 2]. Genome-wide profiling studies using chromatin immunoprecipitation coupled with sequencing (ChIP-seq) have identified thousands of functional/active enhancers that are either bound by the transcriptional co-activator p300, or characterized by their association with high levels of H3 K27 acetylation (H3K27ac) [3, 4, 5]. These enhancers often carry binding sites for more than one TF, which interact with the basal transcription machinery associated with core promoters to regulate gene transcription . Very often, TFs also recruit chromatin-modifying enzymes to convert the chromatin to a state permissive for transcription. Pioneer transcription factors (e.g., FoxA1, PU.1), for example, are often the first to engage in a regulatory chromatin region upon stimulation, and enhance transcription by remodeling the local chromatin to make it competent for other TFs to bind . While global profiling of genomic sites competent for TF binding is imperative for the understanding of TF functions, such work has also become increasingly important for defining disease etiologies, as mutations in cis-regulatory elements are frequently found to be associated with human diseases (e.g., cancer) by whole-genome sequencing studies .
Activating transcription factor 3 (ATF3) is a member of the ATF/CREB family of transcription factors involving in many important human diseases including cancer [8, 9, 10, 11], atherosclerosis , infections , cardiac hypertrophy , and hypospadias . The contributions of ATF3 to these diseases are often owing to its rapid induction by a wide-range of cellular stresses (e.g., DNA damage, oxidative stress, and endoplasmic reticulum (ER) stress), leading to activation of cellular signaling required for the maintenance of cell homeostasis. Indeed, while it binds and activates the tumor suppressor p53 in response to oncogenic challenges (e.g., DNA damage and Pten inactivation) [11, 16], ATF3 also engages in the immune response by interacting with NF-κB and repressing expression of proinflammatory cytokines induced by the toll-like receptor 4 . Similarly, ATF3 induced by reactive oxygen species causes high susceptibility to secondary infections by repressing interleukin 6 (IL-6) expression during sepsis-associated immunosuppression . Like other ATF/CREB transcription factors, ATF3 regulates transcription by binding the canonical ATF/CRE cis-regulatory element (5’-TGACGTCA-3’) or the similar AP-1 site (5’-TGA(C/G)TCA-3’) via its basic region-leucine zipper domain (bZip) . Although an over-simplified model suggests that ATF3 homodimers and heterodimers (with other bZip proteins) repress and induce gene expression, respectively , the mechanism by which ATF3 regulates transcription remains largely unknown. Interestingly, although the structures of the bZip domains are highly similar allowing the largely diversified ATF/CREB proteins to bind the same cis-regulatory elements, the genes regulated by ATF3 are distinct from those controlled by its family members. ATF3 and ATF6, for instance, regulate expression of proapoptotic genes and genes involved in protein folding and ER quality control upon ER stress, respectively . As recent evidence supports that ATF3 engages in a complex protein-protein interaction network involving many TFs and transcription co-regulators [16, 21, 22], it is likely that the interactions with other nuclear proteins define the genomic sites where ATF3 binds and the transcription programs that ATF3 regulates. Characterization of genome-wide ATF3 binding sites would thus lead to further elucidation of the ATF3 interaction network and a better understanding of how ATF3 regulates expression of disease-associated genes.
In this study, we present the first comprehensive analysis of ATF3 binding profiles in the human genome. We show that ATF3 bound a large portion of active enhancers characterized by p300 binding and enriched with K27 acetylated histone H3 (H3K27ac) under the basal condition where ATF3 was expressed at a very low level. While the expression of genes proximal to these enhancers tended to be regulated by ATF3, ATF3 was co-localized with p53 and regulated p53-target gene expression in response to DNA damage. Our results thus suggest that ATF3 likely bookmarks genes for transcriptional regulation under basal and stressed conditions.
Genome-wide mapping of ATF3 binding sites using isogenic cell lines
We thus subjected the wild-type (ATF3-WT) cells and the knockout cells to chromatin immunoprecipitation using an ATF3 antibody. Precipitated DNAs were then labeled and subjected to next-generation sequencing and sequencing reads were mapped to human genome and analyzed for enrichment. Although ATF3 was expressed at a low level (Fig. 1b, lane 1), we identified 33,681 high-confident ATF3-binding peaks in the sample derived from ATF3-WT cells (Fig. 1c). Out of them, a majority of peaks (32,058) were ATF3 specific, as they were not found in the ATF3-KO cells (Fig. 1c and d). A few examples of ATF3 peaks were shown in Fig. 1e. Of note, like a majority of identified sites, these ATF3 peaks were found only in the ATF3-WT sample but not in the ATF3-KO sample (Fig. 1e). Consistent with an early result that ATF3 represses its own expression , we found that ATF3 strongly bound its own promoter (Fig. 1e). Using quantitative PCR (qPCR) to examine samples from an independent ChIP experiment, we confirmed that ATF3 bound to all of the tested genomic sites identified by ChIP-seq (Fig. 1f). Again, ATF3 bound to its own promoter in the ATF3-WT cells but not in the knockout cells (Fig. 1g). The strengths of ATF3 binding to these sites measured by ChIP-qPCR were well correlated with the ChIP-seq scores (R = 0.6328), demonstrating high reproducibility and reliability of our ChIP-seq data.
Global ATF3-binding profile and motif analysis
We also searched the ATF3 binding sites for known TF binding motifs using the Homer de novo motif discovery software. A total of 140 motifs were identified with a p value smaller than 0.01. With only one exception (CEBP:AP1 motif), the top 12 identified motifs contained either the canonical ATF/CRE sequence (i.e., 5’-TGACGTCA-3’) or the AP-1 sequence (i.e., 5’-TGASTCA-3’, S = C/G) (Fig. 2c). As C/EBP harbors a bZip domain that can mediate dimerization with other bZip proteins including ATF3 , it might be that ATF3 bound the CEBP:AP1 motif through dimerization with C/EBP. Overall, 81.7 % of ATF3 binding sites contain an element predicted to be bound by ATF3 - collectively referred to as the ATF3 motif hereafter (Fig. 2d) - suggesting that ATF3 directly binds genomic DNA in most cases. Interestingly, although more ATF3 peaks contained the AP-1 motif (Fig. 2e), the binding affinity of ATF3 to the canonical ATF/CRE element appeared to be higher than that for ATF3 binding to the AP-1 element (Fig. 2e). In addition to these known ATF3 binding motifs, other top ATF3 binding motifs (i.e., Enrichment > 2) include GATA3, CTCF, TEAD, and Sp1, which was presented in 7.7 %, 5.6 %, 14.8 %, and 6.5 % of ATF3 peaks, respectively (Fig. 2f). Although these ATF3-binding peaks often contain a known ATF3 motif (Fig. 2g), ATF3 might also bind these motifs through interacting with corresponding TFs. Indeed, ATF3 has been shown to interact with Sp1 .
ATF3 globally binds active enhancers enriched with p300 and H3K27ac
ATF3-regulated gene expression correlates with ATF3 enhancer binding
DNA damage alters the ATF3-binding landscape for gene regulation
We next addressed the question as to what changes in gene expression the altered ATF3-binding would cause under the DNA damage condition. Treating HCT116 cells with CPT for 4 h resulted in an increase in expression of 733 genes and a decrease in expression of 1095 genes (fold > 1.5, p < 0.05) (Fig. 5f). 1,300 (71.1 %) of these altered genes were bound by ATF3 after DNA damage (Fig. 5g), and thus were more likely regulated by ATF3. Interestingly, 82.9 % of ATF3-bound, CPT-regulated genes were also bound by ATF3 before the cells were treated with CPT (Fig. 5g), suggesting that ATF3 were pre-loaded on the genomic sites for gene regulation under stressed conditions. However, stressed ATF3 appeared to bind these sites more strongly than the quiescent protein (Fig. 5h). Given that CPT equally increased or decreased ATF3 binding on the “shared” sites (see above), these results indicate that DNA damage selectively promoted ATF3 to bind to genomic sites associated with regulated genes.
To further determine the relationship between ATF3 binding and gene regulation under the stressed condition, we analyzed the gene expression data for ATF3 knockout cells, and generated a curated list of 93 genes that were judged, with high confidence, as ATF3-regulated genes in response to DNA damage (Additional file 1: Figure S2D), based on (1) that fold changes before and after the CPT treatment were significantly different (p < 0.05, paired t-test) between ATF3-wildtype and -knockout cells, (2) that the genes bound by ATF3 with a small binding score (<10) and thus more likely to be derived from experimental errors were excluded. Once again, while 82 (88.2 %) of these genes had already been bound by ATF3 under the basal condition, CPT further increased ATF3 binding to these regulated genes, regardless whether their expression was induced or repressed by CPT (Fig. 5i and j). Interestingly, about a half (43, or 46 %) of these genes contained one or more active enhancers that were bound by both p300 and ATF3 (Fig. 5i), consistent with our previous conclusion that ATF3 can bind to active enhancers to regulate gene expression.
ATF3 collaborates with p53 in regulating target gene expression
It is often shown that ATF3 binds the ATF/CRE cis-acting element localized in gene promoters and regulate expression of genes associated with human diseases [12, 13, 14]. We carried out this study in light of the fact that a genome-wide ATF3-binding profile in the human genome was lacking. Employing engineered ATF3-knockout cells as the specificity control, we identified 33,681 specific ATF3-binding sites across the human genome under the basal condition. Although this number was surprisingly large given that the basal ATF3 expression level was low, it was comparable to 22,521 sites identified in mouse dendritic cells . As 81.7 % of the ATF3-binding sites contained a known ATF3 motif (Fig. 2e) , ATF3 might directly bind a majority of these sites. It was thus likely that the low level of constitutively-expressed ATF3 was sufficient to bind most of available sites in the genome. Interestingly, ATF4, a family member sharing the same binding motif with ATF3, binds only 1,210 sites in the mouse genome . While this difference might be owing to different DNA-binding affinity, interactions with other transcription regulators could poise ATF3 for a higher level of genome binding. The latter possibility is supported by the fact that ATF3 differs from ATF4 in its ability of interacting with other proteins . It is worth noting that the ATF3 genome-occupancy level is lower than that of pioneer factors, which often bind more than 50,000 genomic sites , but significantly higher than that of most of gene-specific TFs (e.g., p53) that generally occupy a few thousands of genomic sites (Fig. 6a). It is thus tempting to hypothesize that ATF3 serves as a molecular beacon, or “primer factor”, that binds genomic sites subsequent to binding of pioneer factors, and directs other TFs or transcription co-regulators to appropriate genomic sites upon stimulation . This hypothesis was partly supported by the findings that ATF3 directly interacts with many TFs (e.g., p53, p63, AR, Sp1) [16, 21, 26, 41] and histone modifying enzymes (e.g., Tip60 and HDAC) [17, 22]. Importantly, while the GO analysis revealed that the ATF3-bound genes were associated with cellular response to stress under the basal condition (Fig. 2c), we found that the genes whose expression was regulated by DNA damage were often pre-bound by ATF3 (Fig. 5g). Thus, like the transcription factor p63 , ATF3 might bookmark genes for transcriptional regulation. In this regard, it is likely that ATF3 recruits diverse sets of TFs to genomic sites pre-bound by ATF3 upon varying stimuli, thereby regulating gene expression and mounting rapid, appropriate responses to varying cellular stresses. However, DNA damage-induced changes in ATF3 binding were more dynamic than what the “primer-factor” hypothesis suggests . DNA damage not only increased the number of ATF3 binding sites by 1 fold, but abolished up to one-fifth of the basal binding events (Fig. 5b). In addition, CPT increased ATF3 binding to some genomic sites but decreased its binding to almost equal numbers of other sites. While stress-induced loss of genomic binding has also been reported for other stress-inducible TFs (e.g., JunB) , the decrease in ATF3 binding to a substantial number of genomic sites argues against the notion that the dynamic changes in ATF3 binding was a mere consequence of elevated ATF3 expression induced by DNA. As DNA damage can alter chromatin structure [43, 44, 45], it might allow access of some genomic sites to, while shielding other sites from, ATF3. Interestingly, the CPT treatment appeared to promote ATF3 to bind to sites distal to TSS (Fig. 5d). While the exact mechanism remains elusive, it might be that the epigenetic environments where the distal sites reside are favorable for TF binding. Indeed, these distal sites often coincide with p300/H3K27ac-enriched active enhancers (Fig. 3d), which are known to have lower nucleosomal density .
Like other TFs [36, 42], binding of ATF3 to the regulatory region of a gene did not always result in a change in gene expression. Indeed, although ATF3 bound more than 10,000 genes, a complete loss of ATF3 expression only altered expression of a small number of genes under both the quiescent and the stressed condition. While RNA-based assays (e.g., microarray and RNA-seq) may not serve as accurate measurements of transcription activity , other TFs capable of binding the same motifs (e.g., JunB)  might compensate for ATF3 loss. Interestingly, the ATF3-binding sites often contained motifs of other TFs in addition to the ATF3 motif (Fig. 2), suggesting that ATF3 might act in concert with other TFs to regulate gene expression. Our results also indicate that ATF3 can activate or repress gene expression depending on gene context. While the location and motif composition of the ATF3-binding site did not appear to determine whether ATF3 activates or represses gene expression (Additional file 1: Figure S1D and S1E), it is very likely that the epigenetic environment surrounding the ATF3-binding sites determine the availability of transcription co-activators (like Tip60), or transcription co-repressors (e.g., HDAC), which consequently transactivate or repress expression of ATF3-bound genes. Thus, the early notion that ATF3 homodimers and heterodimers respectively repress and activate transcription appears oversimplified and misleading.
An important finding from this study is that ATF3 bound to 37 % of genomic sites that were bound by p300 and characterized by high levels of H3K27ac under the basal condition (Fig. 3b). These genomic sites are defined as active enhancers and have been shown to contain functional regulatory elements that drive proximal gene expression during embryonic development [3, 4]. Interestingly, DNA damage increased the percentage of active enhancers bound by ATF3 to 57.6 %. Moreover, although ATF3 binding alone was not sufficient to regulate transcription, most of genes regulated by ATF3 appeared localized proximal to ATF3-bound active enhancers (Fig. 4). This strong correlation between TF binding to active enhancers and the regulation of gene expression was not without precedent. The transcription factor p63, for instance, was recently shown to bind H3K27ac-enriched active enhancers, and the binding correlates with dynamic gene expression regulated by p63 during epidermal differentiation . As active enhancers often contain a cluster of motifs allowing for binding by multiple TFs, it is likely that these TFs collaboratively interact with the basal transcription machinery in core promoters to regulate gene expression. Therefore, the observed correlation between enhancer binding and transcriptional regulation is consistent with our notion that ATF3 needs to cooperate with other TFs to regulate gene expression.
The tumor suppressor p53 drives a transcription program for eliciting diverse cellular responses to DNA damage. Previously, we reported that ATF3 can activate p53 by binding and directly blocking its ubiquitination . We also found that ATF3 can induce p53 activation by promoting the activity of a histone acetyltransferase Tip60 and the subsequent activation of ATM . In this study, we revealed an additional mechanism by which ATF3 regulates p53, i.e., co-localization with p53 at genomic sites. Indeed, we found that ATF3 was co-localized with p53 at more than 20 % of p53-binding sites identified by ChIP-seq (Fig. 6a). As ATF3 can interact with p53 , such co-localization might be a consequence of p53-mediated ATF3 recruiting (Fig. 6h), or vice versa (Fig. 6i). On the other hand, some co-localized genomic sites contained both the p53 and the ATF3 motif, and thus could be bound by p53 and ATF3 simultaneously. Regardless, close proximity between ATF3 and p53 at genomic sites might directly alter p53 conformation thereby regulating the p53 transcriptional activity (Fig. 6j). Our results are supported by a recent report, which carried out ATF3 ChIP-chip assays and shows binding of ATF3 to promoters of many known p53 target genes . However, Our study indicates that a large number of co-localized sites were far beyond promoter regions  and were also often bound by p300. Therefore, the genomic co-localization of ATF3 and p53 serves as an additional mechanism for fine tuning p53 activity in the DNA damage response.
Our results indicate that ATF3 likely preoccupies genomic sites regulatory for genes involved in the cellular stress response, and thus bookmarks these sites for transcriptional regulation under basal and stressed conditions.
Cell culture and generation of ATF3-knockout cells
HCT116 wild-type and p53-knockout cells (obtained from Bert Vogelstein) were cultured in McCoy’s 5A medium supplemented with 10 % fetal bovine serum. H1299 cells and 293 T cells were cultured in RPMI 1640 and DMEM medium, respectively. We knocked out ATF3 expression in HCT116 cells using a rAAV-based approach . Briefly, left and right homology arms flanking a small region (22 bp) in the exon 3 of ATF3 were amplified by PCR, and sequentially ligated into pAAV-TK-Acceptor  via restriction enzyme digestion. The resulted plasmid was then transfected into AAV-293 cells for rAAV packaging using the AAV Helper-free System (Agilent) following the manufacturer’s protocol. For viral infections, HCT116 cells in 60 mm dishes were incubated with 2 ml of viral supernatant overnight, followed by re-suspension in medium containing 500 μg/ml of G418 for selection. Genomic DNAs were then prepared from resistant single clones as describe previously , and used for PCR to identify targeted clones. To remove the inserted selection gene, targeted clones in 24-well plates were transfected with a Cre-expression plasmid. Single clones regaining G418 sensitivity were accordingly identified, and subjected to the 2nd round of genome editing to knock out the 2nd ATF3 allele as describe above. The sequences of primers used in this report are available upon request.
Chormatin immunoprecipitation was performed essentially as described previously . Briefly, cells (2 × 107) treated with or without 1.5 μM of CPT for 4 h were cross-linked with 1 mM of di(N-succinimidyl) glutarate (DSG) for 45 min, followed by 1 % formaldehyde for 10 min at room temperature. After treating with 0.125 M of glycine for 5 min, cells were resuspended in 10 ml of Solution I (10 mM Hepes-KOH, pH7.5, 10 mM EDTA, 0.5 mM EGTA, and 0.75 % Triton X-100), and incubated at 4 °C for 10 min. Cells were further incubated with 10 ml of Solution II (10 mM Hepes-KOH, pH7.5, 200 mM NaCl, 1 mM EDTA, and 0.5 mM EGTA) at 4 °C for 10 min before lysed in cold FA lysis buffer (50 mM Hepes-KOH, pH7.5, 140 mM NaCl, 1 mM EDTA, 1 % Triton X-100, 0.1 % sodium deoxycholate, and proteinase inhibitors). Chromatin was sheared by sonication using Bioruptor to an average fragment size of 500 bp, and then incubated with 2 μg of the antibody (ATF3, sc-188; p53, sc-126; p300, sc-585) or normal IgG (rabbit, sc-3888; mouse, sc-2025) purchased from Santa Cruz, at 4 °C overnight. Immunocomplexes were precipitated with 30 μl of ssDNA-protein A/G agarose (Millipore) at 4 °C for 2 h, and sequentially washed with Buffer I (50 mM Tris–HCl, pH8.0, 150 mM NaCl, 1 % SDS, 0.5 % sodium deoxycholate, 1 % NP 40, and 1 mM EDTA), Buffer II (buffer I with 500 mM NaCl), Buffer III (50 mM Tris–HCl, pH8.0, 250 mM LiCl, 0.5 % sodium deoxycholate, 1 % NP 40, and 1 mM EDTA), and TE buffer (50 mM Tris–HCl, pH8.0, and 1 mM EDTA). Bound chromatin was eluted with 0.3 ml of Elution buffer (50 mM Tris–HCl, pH8.0, 1 % SDS, and 1 mM EDTA). After reversal of crosslinking, RNase A and Proteinase K was added, and DNA was purified by phenol extraction and ethanol precipitation. For re-ChIP assays, chromatin immunoprecipitated with the ATF3 antibody was eluted in 0.15 ml of Elution buffer, and then diluted by 20 times with re-ChIP buffer (20 mM Tris–HCl, pH8.0, 150 mM NaCl, 1 % Triton X-100, and 2 mM EDTA), followed by incubation with the p53 antibody as described above.
ChIP-seq and data analysis
ChIP-seq libraries were prepared according to standard protocols using Biosicentific’s DNA Sample Kit (cat#514101) . Libraries were sequenced using Illumina Hi-Seq platforms. Sequence reads were aligned to the Human Reference Genome (assembly hg19) using Burrows-Wheeler Alignment (BWA) Tool Version 0.6.1. Peak identification, overlapping, subtraction and feature annotation of enriched regions were performed using Hypergeometric Optimization of Motif EnRichment suite (HOMER). Heatmaps and intensity plots of peaks were generated by Perl script, R and/or java Treeview. HOMER was used to check motif enrichment.
Microarray data analysis
Total RNA was prepared using Agilent Total RNA Isolation Mini Kit (cat# 5185–6000). Microarray expression profiling was performed using HumanHT-12 v 4.0 Expression BeadChip (Illumina). Data were preprocessed and normalized by GenomeStudio. Differentially expressed genes were identified by Bioconductor limma package and GenePattern. Heatmap view of differentially expressed genes was created by Cluster and Java Treeview. GO term enrichment was determined using DAVID.
Western blotting and quantitative PCR
Western blotting assays were performed as described previously . In brief, cells were lysed in RIPA buffer containing 50 mM Tris–HCl, pH 7.4, 1 % Nonidet P-40, 0.25 % sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, and 1 mM NaF, 1 mM Na3VO4, and protease inhibitor cocktail (Roche), and then resolved in SDS-polyacrylamide electrophoresis for immunoblotting. Quantitative PCR assays were carried out using SYBR Green as described elsewhere .
Availability of supporting data
The data sets supporting the results of this article are available in the GEO with the accession number GSE74363 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=yfozycagbfojfkz&acc=GSE74363).
We thank Dr. Bert Vogelstein and Dr. Todd Waldman for providing cell lines and plasmids.
This work was supported by National Institutes of Health grants (R01CA139107, and R01CA164006), and a US Department of Defense award (W81XWH-15-1-0049) to CY.
- 4.Creyghton MP, Cheng AW, Welstead GG, Kooistra T, Carey BW, Steine EJ, Hanna J, Lodato MA, Frampton GM, Sharp PA, Boyer LA, Young RA, Jaenisch R. Histone H3K27ac separates active form poised enhancers and predicts developmental state. Proc Natl Acad Sci USA. 2010;107:21931–6.CrossRefPubMedPubMedCentralGoogle Scholar
- 8.Wolford CC, McConoughey SJ, Jalgaonkar SP, Leon M, Merchant AS, Dominick JL, Yin X, Chang Y, Zmuda EJ, O'Toole SA, Millar EKA, Roller SL, Shapiro CL, Ostrowski MC, Sutherland RL, Hai T. Transcription factor ATF3 links host adaptive response to breast cancer metastasis. J Clin Invest. 2013;123:2893–906.Google Scholar
- 10.Yuan X, Yu L, Li J, Xie G, Rong T, Zhang L, Chen J, Meng Q, Irving AT, Wang D, Williams ED, Liu JP, Sadler AJ, Williams BR, Shen L, Xu D. ATF3 suppresses metastasis of bladder cancer by regulating gelsolin-mediated remodeling of the actin cytoskeleton. Cancer Res. 2013;73:3625–37.CrossRefPubMedGoogle Scholar
- 11.Wang Z, Xu D, Ding H-F, Kim J, Zhang J, Hai T, Yan C. Loss of ATF3 promotes Akt activation and prostate cancer development in a Pten knockout mouse model. Oncogene. 2015;34:4975–84.Google Scholar
- 12.Gold ES, Ramsey SA, Sartain MJ, Selinummi J, Podolsky I, Rodriguez DJ, Moritz RL, Aderem A. ATF3 protects against atherosclerosis by suppressing 25-hydroxycholesterol-induced lipid body formation. J Exp Med. 2012;209:807–17.Google Scholar
- 13.Hoetzenecker W, Echtenacher B, Guenova E, Hoetzenecker K, Woelbing F, Bruck J, Teske A, Valtcheva N, Fuchs K, Kneilling M, Park J-H, Kim K-H, Kim K-W, Hoffmann P, Krenn C, Hai T, Ghoreschi K, Biedermann T, Rocken M. ROS-induced ATF3 causes susceptibility to secondary infections during sepsis-associated immunosuppression. Nat Med. 2012;18:128–34.CrossRefGoogle Scholar
- 29.Pan Y, Chen H, Siu F, Kilberg MS. Amino acid deprivation and endoplasmic reticulum stress induce expression of multiple activating transcription factor-3 mRNA species that, when overexpressed in HepG2 cells, modulate transcription by the human asparagine synthetase promoter. J Biol Chem. 2003;278:38402–12.CrossRefPubMedGoogle Scholar
- 31.Hackl C, Lang SA, Moser C, Mori A, Fichtner-Feigl S, Hellerbrand C, Dietmeier W, Schlitt HJ, Geissler EK, Stoeltzing O. Activating transcription factor-3 (ATF3) functions as a tumor suppressor in colon cancer and is up-regulated upon heat-shock protein 90 (Hsp90) inhibition. BMC Cancer. 2010;10:668.CrossRefPubMedPubMedCentralGoogle Scholar
- 38.Fan F, Jin S, Amundson SA, Tong T, Fan W, Zhao H, Zhu X, Mazzacuratin L, Li X, Petrik KL, Fornace AJ, Rajasekaran B, Zhan Q. ATF3 induction following DNA damage is regulated by distinct signaling pathways and over-expression of ATF3 protein suppresses cells growth. Oncogene. 2002;21:7488–96.CrossRefPubMedGoogle Scholar
- 39.Garber M, Yosef N, Goren A, Raychowdhury R, Thielke A, Guttman M, Robinson J, Minie B, Chevrier N, Itzhaki Z, Blecher-Gonen R, Bornstein C, Amann-Zalcenstein D, Weiner A, Friedrich D, Meldrim J, Ram O, Cheng C, Gnirke A, Fisher S, Friedman N, Wong B, Bernstein BE, Nusbaum C, Hacohen N, Regev A, Amit I. A high-throughput chromatin immunoprecipitation approach reveals principles of dynamic gene regulation in mammals. Mol Cell. 2012;47:810–22.CrossRefPubMedGoogle Scholar
- 43.Ikura T, Tashiro S, Kakino A, Shima H, Jacob N, Amunugama R, Yoder K, Izumi S, Kuraoka I, Tanaka K, Kimura H, Ikura M, Nishijubbo S, Ito T, Muto A, Miyagawa K, Takeda S, Fishel R, Igarashi K, Kamiya K. DNA damage-dependent acetylation and ubiquitiantion of H2AX enhances chromatin dynamics. Mol Cell Biol. 2011;27:70287040.Google Scholar
- 46.Tanaka Y, Nakumura A, Morioka MS, Inoue S, Tamamori-Adachi M, Yamada K, Taketani K, Kawauchi J, Tanaka-Okamoto M, Miyoshi J, Tanaka H, Kitajima S. System analysis of ATF3 in stress response and cancer reveals opposing effects on pro-apoptotic genes in p53 pathway. PLoS One. 2011;6:e26848.Google Scholar
- 50.Yu J, Yu J, Mani R, Cao Q, Brenner CJ, Cao X, Wang X, Wu L, Li J, Hu M, Gong Y, Cheng H, Laxman B, Vellaichamy A, Shankar S, Li Y, Dhanasekaran SM, Morey R, Barrette T, Lonigro RJ, Tomlins SA, Varambally S, Qin ZS, Chinnaiyan AM. An integrated network of androgen receptor, polycomb, and TMPRSS2-ERG gene fusions in prostate cancer progression. Cancer Cell. 2010;17:443–54.CrossRefPubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.