Differentiated embryo chondrocyte plays a crucial role in DNA damage response via transcriptional regulation under hypoxic conditions

Tumor hypoxia contributes to a biologically aggressive phenotype and therapeutic resistance. Recent studies have revealed that hypoxia reduces expression of several DNA damage recognition and repair (DRR) genes via both hypoxia-inducible factor (HIF)-independent and -dependent pathways, and this induced genomic instability in cancer cells. We show here that one of the HIF-target genes—differentiated embryo chondrocyte (DEC)—plays a role in DNA damage response via transcriptional repression. Comprehensive gene expression and database analyses have revealed systemic repression of DNA-DRR genes in cancer and non-cancer cells under hypoxic conditions. Hypoxic repression in typical cases was confirmed by quantitative RT-PCR and promoter reporter experiments, and knockdown experiments indicated the critical role of DEC2 in such repression. Assessment of histone H2AX phosphorylation revealed that recognition and repair of DNA double-strand breaks (DSBs) induced by bleomycin or γ-ray irradiation were attenuated; moreover, Cleaved Caspase-3 levels were decreased with pre-conditioning under hypoxia: opposing phenomena were ascertained by knockdown of DEC2. Finally, pre-conditioning under hypoxia decreased the sensitivity of cancer cells to DSBs, and knockdown of DEC2 increased γ-ray sensitivity. These data imply that a critical reduction of DNA-DRR occurs via DEC-dependent transcriptional repression and suggest that DEC is a potential molecular target for anti-cancer strategies.

Introduction Hypoxia is one of the hallmarks of growing solid tumors. The tumor microenvironment is now known to play a critical role in the development of biologically aggressive tumor phenotypes, and also their resistance to chemo-and radio-therapy, through altered expression of various genes [1][2][3][4]. Hypoxia-inducible factor-1α (HIF-1α) is a key transcription factor that controls dozens of target genes in a variety of cancers [5][6][7][8][9]. Among the protein products of these genes, differentiated embryo chondrocyte (DEC) 1 and 2 have been reported to regulate expression of PPARG, PER, STAT1, SREBF1, MITF, BHLHE40, and BHLHE41 in adipogenesis, circadian rhythm, immune system function, cell differentiation, and carcinogenesis [10][11][12][13][14]. Previously, we and others reported that the HIF-DEC signaling pathway -whereby HIF-1 induces DEC1 (also called BHLHE40, STRA13, or SHARP2) and DEC2 (also called BHLHE41 or SHARP1)-plays an important role in transcriptional repression of the mismatch repair gene MLH1 under hypoxic conditions [15,16].
Repressed DNA repair functions and mutator phenotypes under hypoxic conditions were first documented some time ago, but details of molecular mechanisms were not clear [17][18][19][20]. Recent publications have shown that expression of several DNA damage recognition and repair (DRR) genes -MLH1, RAD51, BRCA1, MSH2 and NBS1-decreases under hypoxic conditions via HIF-1-dependent and -independent mechanisms, and that this repression of DNA-DRR leads to genomic instability, carcinogenesis, tumor progression, and treatment failure [21][22][23][24][25][26][27]. On the other hand, the precise functional role of the HIF-DEC signaling pathway, especially in hypoxic repression of DNA-DRR gene expression, is not yet fully understood.
In this study, we examined changes in the expression of numerous genes related to DNA-DRR and also the molecular mechanisms of transcriptional regulation under hypoxic conditions, by focusing on the HIF-1-induced transcriptional repressors DEC1 and DEC2 and the influence on DNA damage of cultivation under hypoxic conditions. We demonstrate for the first time that DEC1 and DEC2 are involved in the transcriptional repression of many DNA-DRR genes under hypoxic conditions. In addition, the HIF-DEC signaling pathway impaired the recognition and repair of double-strand breaks (DSBs)-as well as apoptosiscaused by various kinds of DNA-damage-inducing treatments via effective repression of transcription in cells subjected to hypoxia. To understand the mechanisms by which malignant phenotypes and genomic instability are acquired in cells exposed to low oxygen is important for developing new diagnostic and therapeutic targets.

Chemicals
Most of chemicals were analytical grade and were purchased from Wako Pure Chemicals (Osaka, Japan) or Sigma (St. Louis, MO). HIF-prolyl hydroxylase inhibitors, FG-4592 and DMOG, were purchased from AdooQ BioScience (Irvine, CA). Bleomycin was kindly provided from Nippon Kayaku Co Company.

Cell lines and treatment
The oral squamous cell carcinoma cell line, HSC-2, and a human hepatoma cell line, HepG2 (The Japanese Cancer Research Resource Bank), were cultured in DMEM or RPMI 1640 medium (Sigma-Aldrich Japan, Tokyo, Japan) supplemented with 10% fetal bovine serum (BioWhittaker, Verviers, Belgium) and Kanamycin (100 μg/ml). For gene expression analyses, cells were cultured under normoxic (21% pO 2 ) or hypoxic (1% pO 2 ) conditions for the indicated length of time in a hypoxic chamber. For knockdown analyses, siRNA against HIF1A, These funders were not involved in study design, data analysis, decision to publish, or preparation of the manuscript. Nippon Kayaku Co., Ltd provided support in the form of salary for one of the authors (T. Nakanishi), but did not have any additional role in the study as described above. The specific role of this author is articulated in the 'author contributions' section. DEC1, DEC2, or non-specific siRNA (QIAGEN, Inc., Valencia, CA) were transfected with TransIT 1 -siQUEST™ Transfection Reagent (Mirus Corporation, Madison, WI) in HSC-2 for 12 hours. Then the cells were incubated under normoxic or hypoxic conditions for 24 hours, and cells were then harvested and stored at -80˚C until use.

RNA preparation
Total RNA was prepared from frozen cell pellets with a QIAGEN RNeasy 1 mini kit (QIA-GEN) according to the manufacturer's instructions.

Oligonucleotide array analysis
The CodeLink Expression Bioarray System (GE Healthcare, Tokyo, Japan) was used as previously described [28]. Briefly, from 1 μg of total RNA, double-strand cDNA and subsequent biotin-labeled cRNA were synthesized using a CodeLink TM Expression Assay Reagent Kit (GE Healthcare). Ten μg of fragmented cRNA was hybridized to a CodeLink TM Uniset Human 20 K I Bioarray (19,981 probes). After hybridization, the arrays were scanned with an Agilent DNA Microarray Scanner TM (Agilent, Santa Clara, CA), then analyzed with CodeLink TM Expression Analysis Software (GE Healthcare). Expression levels were normalized to the median of the entire spot array using GeneSpring TM GX (Agilent Technologies). The data were submitted to the Gene Expression Omnibus (GSE52795). Gene-annotation enrichment analysis was performed using DAVID Bioinformatics Resources 6.8 (http://david.ncifcrf.gov/ home.jsp)).

Database analyses
The NCBI Gene Expression Omnibus (GEO) [29] database now stores over 2 million records of samples, so we attempted a collective-intelligence approach to assess gene expression data related to hypoxic responses in order to identify hypoxia-related genes in a data-driven way. To normalize the microarray data, we used data from the Affymetrix Human Genome U133 Plus 2.0 Array (GEO accession: GPL570), which is the most often used platform in human studies, and limited our search to "human". The pair of experiments performed in normoxic or hypoxic conditions was undertaken after careful curation of dataset descriptions and corresponding papers in the citation metadata. As a result, we identified 29 hypoxia-normoxia pairs from among 9 data series (GEO accessions: GSE19197, GSE29641, GSE4086, GSE9234, GSE39042, GSE58049, GSE52315, GSE46054, and GSE53012). After Robust Multichip Average (RMA) normalization [30] by the affy package [31] in R/BioConductor (version 2.15.0), up/down regulated genes were selected from all pairs with a 1.5-fold threshold. The number of experiments in which the gene was up-or down-regulated was counted for all genes.

Reverse Transcription-PCR
Two μg of total RNA extracted from each cell line were reverse-transcribed using a High-Capacity cDNA Archive TM Kit (Applied Biosystems, Foster City, CA). A two-hundredth aliquot of the cDNA was subjected to real-time RT-PCR using primers (final concentration of 200 nM each) and MGB probe (final concentration of 100 nM, the Universal Probe Library: UPL, Roche Diagnostics, Tokyo, Japan) (shown in S1 Table) sets for HIF1A, DEC1, and DEC2, or the TaqMan TM Gene Expression Assay (Applied Biosystems) for MLH1, MSH2, MBD4, MRE11A, BRCA1, and RAD51, with Pre-Developed TaqMan TM Assay Reagents (Applied Biosystems) for ACTB as an internal control. PCR reactions were carried out using 7500 Real-Time PCR System (Applied Biosystems) under the following standard conditions. At least three independent measurements were averaged, and relative gene expression levels were calculated using ACTB expression as the denominator for each cell line.

Bleomycin or γ-ray treatment
Cells were cultured under normoxic or hypoxic conditions for 24 hours until just before treatment. Various doses of bleomycin were applied under normoxic conditions and cytotoxicity was evaluated by conventional MTT dye-reduction assay as previously described [34]. The indicated doses of 137 Cs γ-ray irradiation were applied with a Gammacell 40 Exactor (MDS Nordion, Ottawa, ON, Canada) under normoxic conditions. Bleomycin or γ-ray treated cells were further incubated until the indicated time points and then harvested for analysis. IC 50 value or relative MTT tetrazolium reduction activity (relative viability) was calculated using the value at 24 hours as the denominator.

Immunoblot analysis
For protein-expression analysis, whole-cell extracts were prepared from cultured cells with or without various treatments, as previously described [35]. Twenty-five or fifty μg of extract was blotted onto a PVDF filter following SDS-polyacrylamide gel electrophoresis. Anti-H2AX, anti-γH2AX (Millipore, Temecula, CA), anti-Caspase3, anti-Cleaved Caspase-3 (Cell Signaling Technology, Danvers, MA), anti-HIF-1α (BD Pharmingen, San Diego, CA), anti-DEC1 (Novus Biologicals, Littleton, CO), anti-DEC2 (invitrogen, Rockford, IL), and anti-β-actin (Sigma) were used as primary antibodies. A mouse or rabbit anti-IgG horseradish peroxidase conjugate (Amersham Life Science) was used as a secondary antibody. Immunocomplexes were visualized with the enhanced chemiluminescence reagent ECL Plus (Amersham Life Science). Expression levels of γH2AX or Cleaved Caspase-3 were quantified by measuring densities using ImageJ (public domain open source software; imagej.nih.gov), and at least three independent measurements were averaged. Relative expression levels of γH2AX or Cleaved Caspase-3 were calculated using expression levels of H2AX or Caspase-3 as the denominator for each time point.

Immunostaining analysis
HepG2 cells were cultured on cover slips under normoxic or hypoxic conditions for 24 hours, and re-oxygenated for 30 minutes. The cells were then exposed to 1 Gy of X-ray, and further cultured for 0.5 and 12 hours under normoxic condition after irradiation. The cells on cover slips were fixed with 4% Paraformaldehyde, permeabilized with 0.5% Triton X-100, and stained with anti-γH2AX (1:1000) as primary antibody. And rhodamine-conjugated sheep anti-mouse immunoglobulin G (IgG) (1:1000) (Chemicon, Temecula, CA) was used as secondary antibody. After several washes, nuclei were stained with Hoechst 33342. Subcellular distributions of γH2AX foci were observed, and pictures were taken using a fluorescence microscope (BZ-8000; KEYENCE Corporation, Osaka, Japan). For each condition, γH2AX foci in 8-25 cells were counted and averages per cell were calculated by authors.

Statistical analysis
Statistical tests were performed based on the ANOVA test, Student's t test, Dunnett's test, or Turkey-Kramer HSD test using StatView 1 version 5.0 software (SAS Institute Inc., Cary, NC, USA).

Hypoxia systemically down-regulates the expression of DNA-DRR genes via DEC
To discriminate gene expression profiles under hypoxic conditions, we performed comprehensive gene-expression analyses of HSC-2 cells cultured under normoxic or hypoxic conditions. Among 19,981 gene probes tested, 2216 were up-regulated (more than 2-fold) and 2488 were down-regulated (less than 2/3-fold) in cells cultured for 24 hours under hypoxic conditions, suggesting that hypoxic stress strongly affects gene expression. As we had expected, many known hypoxia-inducible genes were found among the up-regulated ones (S1 Fig), and subsequent gene-annotation enrichment analysis using DAVID Bioinformatics Resources 6.8 (http://david.ncifcrf.gov/home. jsp)) strongly supported our observation (S2 Table). Interestingly, DAVID also suggested that many genes responsible for DNA-DRR (more than 50 of 130 genes listed in http://sciencepark. mdanderson.org/labs/wood/DNA_Repair_Genes.html) were involved in down-regulated genes, suggesting the occurrence of systemic repression of DNA-damage response (S3 and S4 Tables, and S1 Fig). Analyses of the NCBI Gene Expression Omnibus (GEO) database further supported our observations. It indicated that hypoxic repression of DNA-DRR genes is common among various kinds of cancer and non-cancer cells (S5 and S6 Tables). One-hundred sixty genes in cancer cells and 123 genes in non-cancer cells among 176 genes analyzed were reported to be down-regulated under hypoxic conditions in at least one GEO data set. With some of the down-regulated genes (MLH1, MSH2, MBD4, MRE11A, BRCA1, and RAD51), we confirmed repression under hypoxic conditions through the use of quantitative real-time RT-PCR ( Fig 1A). We also observed that expression levels of these genes were down-regulated by treatment with various HIF-activating reagents (S2 Fig). We further evaluated the role of HIF-1α-DEC pathway by using samples in which HIF1A, DEC1, or DEC2 had been confirmed to be effectively knocked-down (S3 Fig). Comparing expression levels of these genes under normoxic and hypoxic conditions, all of them were significantly down-regulated under hypoxic conditions with non-specific siRNA transfection. Those repressions under hypoxic conditions were clearly prevented, i.e., significant fold-reduction disappeared, by knock-down of DEC2 with specific siRNA (siDEC2) transfection, whereas siDEC1 and siHIF1A showed only slight effects, as previously reported [15]. In fact, expression of MRE11A under hypoxic conditions increased with DEC2 knock-down ( Fig 1B).

DEC represses promoter activities of DNA-DRR genes
To clarify the molecular mechanisms of the hypoxic repression of DNA-DRR genes, we subcloned the 5' regions of human MLH1, MSH2, MBD4, MRE11A, BRCA1, and RAD51 (approximately 200-1600 bp length from the transcriptional start site of each gene) into a luciferase reporter plasmid, pGL3-Basic vector. It is notable that all of promoter reporters showed significant luciferase activities under normoxic condition, and those activities were suppressed under hypoxic conditions, except for MSH2 promoter (Fig 2A). Co-transfection experiments demonstrated that DEC1 significantly repressed all promoter activities tested here. DEC2 also effectively repressed all promoter activities except for those of MRE11A and BRCA1 (Fig 2B). We further performed co-transfection experiments using the promoter reporters, DEC1, and a dominant-negative form of DEC1 that lacks a DNA-binding domain (Δbasic DEC1) [32]. With forced expression, the dominant-negative form of DEC1 acted against the repression activity of DEC1 on the promoters of MLH1, MSH2, MBD4, MRE11A, but not on the promotor of BRCA1. (Fig 2C)

Pre-conditioning under hypoxia decreased cellular sensitivity to bleomycin or γ-ray treatment
To clarify the effects of hypoxic repression of DNA-DRR genes in response to DNA damage, we first assessed, using MTT assay, the viability of cancer cells pre-conditioned under hypoxia. To observe cellular responses to Bleomycin or γ-ray treatment (which induces DSBs effectively [36]) independently of hypoxia, HSC-2 cells were incubated under hypoxic conditions for 24 hours, and re-oxygenated just before treatment. MTT assays performed after 72 hours of continuous treatment with bleomycin (at various concentrations) demonstrated an IC 50 value that was significantly higher for cells pre-conditioned under hypoxia than for cells under normoxia ( Fig 3A). Hypoxic pre-conditioned HSC-2 cells treated with three different doses of 137 Cs γray were also studied with the MTT assay to evaluate their viability after incubation for 96 hours; relative viability was calculated as the ratio of the value to that of the 24 hour-culture. In non-irradiated cells, no remarkable differences in relative cellular viability were observed between normoxic and hypoxic pre-conditioned cells, implying that viability is not affected by pre-conditioning under hypoxia for 24 hours. On the other hand, among γ-ray irradiated cells, viability of hypoxic pre-conditioned cells was significantly higher than that of normoxic cultured cells. Although overall viability was lowest following 10 Gy treatment, the most remarkable difference between normoxic and hypoxic pre-conditioning was observed following 5 Gy irradiation (Fig 3B).

Recognition and repair of DSBs were attenuated in cells pre-conditioned under hypoxia
To clarify the molecular mechanisms of decreased sensitivity to DSB-inducing treatments in cells pre-conditioned under hypoxia, we evaluated the phosphorylation status of histone H2AX (γH2AX) in the cells as a DSB marker [37,38]. HSC-2 cells were incubated under hypoxic conditions for 24 hours, and then treated with bleomycin for the indicated periods under normoxic conditions. Immunoblotting analyses confirmed that γH2AX level started to increase at 24 hours, and the level continued elevating until 48 hours after bleomycin hypoxic conditions compared with that under normoxia; the value is noted over the bar. Columns are the mean of three independent experiments; bars, SD. P values calculated with Student's t test are: Ã , P < 0.05; ÃÃ , P < 0.01; ÃÃÃ , P < 0.001.
https://doi.org/10.1371/journal.pone.0192136.g001 , and RAD51 were subcloned into a luciferase reporter plasmid, and these were designated as pGL3-reporters. Each reporter plasmid was transfected into HepG2, and incubated for 36 hours under normoxic or hypoxic conditions. Luciferase activities were determined with a dual-luciferase assay system. (B) Each reporter and DEC1 or DEC2 expression plasmid were co-transfected into HepG2, and incubated for 36 hours under normoxic conditions. Luciferase activities were determined with a dual-luciferase assay system. (C) Each reporter and DEC1 or Δbasic DEC1 expression plasmid were co-transfected into HepG2, and incubated for 36 hours under normoxic conditions. Luciferase activities were determined with a dual-luciferase assay system. Columns are the mean of three independent experiments; bars, SD. The differences between means were significant (ANOVA P < 0.005). P values calculated with Dunnett's test are: Ã , P < 0.05; ÃÃ , P < 0.01; ÃÃÃ , P < 0.001.
https://doi.org/10.1371/journal.pone.0192136.g002 treatment. After 48 hours, the level started to decrease or remained at a low level, indicating that DSB repair progressed successfully in normoxic cells (Fig 4A). In contrast, the level of γH2AX in cells pre-conditioned under hypoxia increased at 48 hours and remained at a high level, although it eventually started to decrease (Fig 4A). Similar results were obtained following irradiation with 5Gy of γ-ray. In normoxic cells, the level of γH2AX increased at 48 hours after irradiation treatment, and then started to decrease. In pre-conditioned cells, γH2AX had slightly increased at 48 hours, and it peaked at 72 hours after irradiation treatment (Fig 4B). Similar kinetics of γH2AX were also observed in 1 Gy of X-ray irradiated HepG2 cells by immunostaining analysis (S4 Fig). To clarify the role of DEC2 in this process, we performed knock-down experiments using siRNA for DEC2. Non-specific siRNA or siDEC2 was transiently transfected 24 hours before pre-conditioning under hypoxia, and after that we performed the foregoing analysis. In siNS transfected cells, the level of γH2AX could be slightly detected at 24 hours, and markedly detected at 48 hours after bleomycin treatment. On the other hand, γH2AX after bleomycin treatment in DEC2 knocked-down cells increased at 24 hours ( Fig 4C). Similar results were obtained following irradiation with 5Gy of γ-ray. In siNS transfected cells, the level of γH2AX was slightly detected at 48 hours, but markedly detected at 24 hours after γ-ray irradiation in DEC2 knocked-down cells (Fig 4D). These data indicate that knock-down of DEC2 restored radiation-induced DNA damage responses in hypoxic pre-conditioned cells.

Hypoxic pre-conditioning altered both DNA damage response and subsequent apoptosis signaling via DEC2
To determine whether hypoxic treatment affects the apoptotic response to DNA damage, we evaluated levels of Cleaved Caspase-3 protein. HSC-2 cells were incubated under normoxic or hypoxic conditions for 24 hours, then treated with bleomycin or γ-ray as in the previous experiments. Immunoblotting analyses demonstrated that levels of full-length Caspase-3 protein increased at 24 hours after bleomycin treatment and then decreased until 72 hours in normoxic cells; a similar increase of full-length Caspase-3 protein was observed in hypoxic preconditioned cells, but with lower intensity. Whereas increased levels of Cleaved Caspase-3 were observed in normoxic cells, the increase was barely detectable in hypoxic pre-conditioned cells (Fig 5A). Similarly, increased levels of Cleaved Caspase-3 were observed in normoxic cells following to γ-ray irradiation, but the increase was detected only in hypoxic pre-conditioned cells at 72 hours after irradiation (Fig 5B). Knock-down experiments demonstrated that bleomycin-induced Cleaved Caspase-3 strongly increased at 24 hours after bleomycin treatment in DEC2 inhibited cells (siDEC2), but it became much weaker at 24 hours and then increased until 48 hours (Fig 5C). Similar effects of DEC2 inhibition were observed in γ-ray irradiated cells (Fig 5D). These results suggest that pre-conditioning under hypoxia altered both DNAdamage response and subsequent apoptosis signaling via DEC2.
Finally, we assessed the role of DEC2 in the decline of radiation sensitivity: Non-specific siRNA or siDEC2 was transiently transfected 24 hours before hypoxic pre-conditioning, and after that we performed the previous analysis. In siNS transfected cells, hypoxic pre-conditioning conferred resistance to 5 Gy of γ-ray as shown in Fig 3B. It was striking that siDEC2 transfected cells showed significantly lower viability in both normoxic and hypoxic pre-conditioned cells, but especially in hypoxic pre-conditioned cells (Fig 6). These data indicate that DEC2 plays a crucial role in sensitivity to treatments that induce DNA damage.

Discussion
Under hypoxic conditions, the cellular DNA-repair function becomes impaired, which causes hyper-mutability after DNA damage [17,39]. This suggests that tumor hypoxia probably  [22,24,[25][26][27]. We also previously demonstrated that two of the HIF-1-target gene products, DEC1 and DEC2, decreased MLH1 transcription under hypoxic conditions [15]. In this study, we aimed to clarify the mechanisms of regulation of DNA-damage response occurring through transcriptional regulation under hypoxic conditions. We found that expression of DNA-DRR genes is systemically down-regulated under hypoxic conditions, and that both DNA-damage response and subsequent apoptosis induced by bleomycin or γ-ray irradiation are to some extent attenuated by DECs.
We first performed a set of comprehensive gene-expression analyses of the head and neck cancer cell line, and found that expression of a wide range of DNA-DRR genes decreased under hypoxic conditions through gene-annotation enrichment analysis using DAVID Bioinformatics Resources (S3 and S4 Tables). It is generally known that the major DNA-repair pathways -mismatch repair, nucleotide excision repair, base excision repair, homologous recombination repair, and non-homologous end joining-involve a large number of molecules [40]. Our results show that most genes involved in these pathways exhibit decreased expression under hypoxic conditions, resulting in systemic repression of DNA-damage response, although expression of some of these genes was not drastically repressed. A collective-intelligence approach to analyzing gene-expression data on hypoxic responses in the GEO database strongly supported our results and even allowed us to extend the significance of our findings to other cancer and noncancer cells (S5 and S6 Tables). This leads to consideration of possible mechanisms of resistance to therapy and acquired mutator phenotypes in various types of cells under hypoxic conditions. Increasing evidence from hypoxic researches has revealed dynamic changes in gene transcription by crucial transcription factors, i.e., hypoxia-inducible factors, under hypoxic conditions [41][42][43][44]. Among the transcription factors involved in those changes, the most prominent primarily activate gene transcription. Although some inhibitory mechanisms involving HIF or other transcription factors, such as REST, have been suggested previously [26,27,45], mechanisms of hypoxia-induced transcriptional repression remained largely unknown. We previously demonstrated the functional significance of the HIF-DEC signal pathway in MLH1 transcription under hypoxia [15]. Other groups further have reported that HIF-1 induces DEC-repressed expression of PPARG, PER, STAT1, SREBF1, and MITF [10][11][12][13][14]. In the present study, we examined the transcriptional mechanisms of various DNA-DRR genes under hypoxia by focusing on the HIF-DEC signal pathway. We first clarified the functional significance of HIF-1α, DEC1, and DEC2 in regard to expression levels of several DNA-DRR genes, using knock-down analyses. As we had previously shown regarding regulation of MLH1 transcription, DEC2 knock-down clearly canceled those hypoxic regulations, and even increased MRE11A expression. However, knock-down of DEC1 did not show any effect on hypoxicrelated decrease in expression of those genes. One possible explanation for the lack of a DEC1 effect is that DEC1 knock-down significantly increases DEC2 expression, resulting in enhanced (B) HSC-2 cells were cultured under normoxic or hypoxic conditions for 24 hours; afterwards, the cells were treated with 500 ng/mL of bleomycin or 5 Gy of γ-ray irradiation, and then the cells were cultured for the indicated lengths of time under normoxic conditions. (C), (D) HSC-2 cells were transfected with either nonspecific (siNS) or DEC2-specific (siDEC2) siRNA, and then incubated under hypoxia for 24 hours before irradiation. After re-oxygenation, cells were treated with 500 ng/mL of bleomycin or 5 Gy of γ-ray irradiation. Thereafter, the cells were cultured for the indicated length of time under normoxic conditions. Anti-γH2AX, anti-H2AX and anti-β-actin were used for specific detection of each protein, and representative images are shown. Relative expression levels of γH2AX were calculated with H2AX expression as the denominator for each time point, and are shown in bottom panels. Columns are the mean of three independent experiments; bars, SD.
https://doi.org/10.1371/journal.pone.0192136.g004 DEC regulates DNA-DDR genes Fig 5. Cleaved Caspase-3 levels were reduced via DEC2 in hypoxic pre-conditioned cells. Total Caspase-3 or Cleaved Caspase-3 levels after bleomycin or γ-ray irradiation in HSC-2, with or without hypoxic pre-conditioning, was evaluated using imuunoblotting analysis. (A), (B) HSC-2 cells were cultured under normoxic or hypoxic conditions for 24 hours; afterwards, the cells were exposed to 500 ng/mL of bleomycin or 5 Gy of γ-ray irradiation. After that, the cells were cultured for the indicated length of time under normoxic conditions. (C), (D) siRNA was transiently transfected into cells 24 hours before repression of DEC-target genes, as had been previously shown [15]. We also observed that HIF1A knock-down mildly affected expression of DNA-DRR, suggesting that there are other pathways, including HIF-2α (EPAS1). Treatments with HIF-activating reagents, prolylhydroxylase inhibitors, further supported a role of HIF-DEC signal pathway on regulation of DNA-DRR gene expressions (S2 Fig). Furthermore, forced expression of DEC1 or DEC2 reduced cloned promoter activities of DNA-DRR genes, and those reductions were attenuated with the HDAC inhibitor, TSA, in a dose-dependent manner (S5 Fig), although the effect of TSA on reporter plasmid vector was not clear. Other experiments using a mutant type of DEC1 lacking the DNA binding domain demonstrated that DNA binding activity of DEC1 is critical for repression of some, but not all, of those promoter activities, suggesting the existence of other mechanisms. Analyses of ChIP-sequence data using the UCSC Genome Browser (genome.ucsc.edu) have indicated that DEC1 (BHLHE40) binding signals have been observed near the transcription start sites of a large number of genes, including known target genes [11-hypoxic pre-conditioning. Transiently transfected cells were cultured in normoxic or hypoxic conditions for 24 hours. Thereafter, the cells were exposed to 500 ng/mL of bleomycin or 5 Gy of γ-ray irradiation. After that the cells were cultured for the indicated length of time under normoxic conditions. Anti-Caspase-3, anti-Cleaved Caspase-3, and anti-β-actin were used for specific detection of each protein, and representative images are shown. Relative expression levels of Cleaved Caspase-3 were calculated with Caspase-3 expression as the denominator for each time point, and are shown in bottom panels. Columns are the mean of three independent experiments; bars, SD.
https://doi.org/10.1371/journal.pone.0192136.g005 Ionizing radiation produces intracellular reactive oxygen species (ROS) that damage cellular components, including DNA and proteins. Because oxygen is required for ROS generation, tumor hypoxia contributes to acquiring resistance to the cytotoxic effects of radiotherapy [4,48]. In fact, tumor hypoxia is clinically known to promote resistance to both chemo-and radio-therapy, and HIF-1α is recognized as a molecular target to enhance the cytotoxic effects of chemo-or radio-therapy on cancer cells thereby improving prognosis [1][2][3][4]. In this study, we demonstrated that inhibition of DEC2 restored, at least in part, sensitivity of hypoxic cancer cell lines to bleomycin or radiation treatment, suggesting a potential molecular target for modifying DNA damage-inducing treatments. DECs have been reported to participate in the regulation of cell differentiation, cell-cycle control, and apoptosis [49][50][51][52][53]. DEC1 has been reported to reduce apoptosis and promote survival of cancer cells subjected to serum deprivation or photo-chemotherapeutic agents [51,52]. DEC2 has also been reported to protect cancer cells from chemotherapeutic agents, leading to prolonged survival of those cells. Furthermore, DECs have been shown to decrease levels of Cleaved Caspase-3 protein in cells subjected to serum deprivation or 5-FU treatment. In our study, hypoxic treatment actually repressed and delayed the Cleavage of Caspase-3 induced by bleomycin or radiation treatment in HSC-2 cells, and inhibition of DEC2 effectively restored apoptosis signaling and sensitivity to radiation treatment. This may be due not only to direct regulation of apoptotic molecules by DECs as previously shown, but also to systemic repression of DNA-DRR genes by DECs, as we have shown in this report.
In summary, we demonstrated that the hypoxia-inducible transcription repressors DEC1 and DEC2 play crucial roles in transcriptional repression of numerous DNA-DRR genes under hypoxic conditions, and the HIF-DEC signaling pathway. DEC2, in particular, participates in DNA damage tolerance in cells exposed to low levels of oxygen. Further investigation of this pathway may provide new insights into the mechanisms and efficacy of DNA damageinducing treatments in hypoxic cancer cells, as well as providing strategies for protecting normal tissues from the effect of radiation exposure.
Supporting information S1  Relative mRNA levels were calculated as the ratio to that of ACTB. Columns are the mean of three independent experiments; bars, SD. The differences between means were significant (ANOVA P < 0.005). P values calculated with Dunnet test are: Ã , P < 0.05; ÃÃ , P < 0.01; ÃÃÃ , P < 0.001. (PDF) S3 Fig. Knockdown assays for HIF1A, DEC1, and DEC2 in HSC-2 cells. Non-specific siRNA (siNS) or targeted siRNA was transfected into HSC-2 and cultured under normoxic or hypoxic conditions for 24 hours. Expression levels of HIF1A, DEC1, and DEC2 were evaluated by realtime RT-PCR (left panel). Relative mRNA levels were calculated as the ratio to that of ACTB. Columns are the mean of three independent experiments; bars, SD. The differences between means were significant (ANOVA P < 0.005). P values calculated with Turkey-Kramer HSD test are: Ã , P < 0.05; ÃÃ , P < 0.01; ÃÃÃ , P < 0.001. Expression levels of HIF-1α, DEC1, and DEC2 proteins were evaluated using immunostaining analysis (right panel). (PDF) S4 Fig. Phosphorylate histone H2AX induced by X-ray were inhibited in hypoxic pre-conditioned HepG2 cells. Phosphorylate histone H2AX (γH2AX) levels after X-ray irradiation in HepG2 cells with or without pre-conditioning under hypoxia were evaluated with immunostaining analysis. The cells were stained with primary anti-γH2AX and secondary rhodaminelabeled antibody. Red shows γH2AX foci, and blue is nuclei stained with Hoechst 33342 (left panel). The right-side panel shows the number of γH2AX foci per cell. Total numbers of γH2AX foci and cells were counted in the left panel, and calculated. Columns are the mean of three independent counts; bars, SD. The differences between means were significant (ANOVA P < 0.005). P values calculated with Turkey-Kramer HSD test are: ÃÃ , P < 0.01; ÃÃÃ , P < 0.001. (PDF) S5 Fig. Suppressed promoter activities of DNA-DRR genes by DEC were increased with TSA treatments. The pGL3-promoter reporters and DEC1 or DEC2 expression plasmid were co-transfected into HepG2 and treated with 10 or 100 ng/ml of Trichostatin A (TSA), histon deacetylase (HDAC) inhibitor, 24 hours before harvesting cells. Luciferase activities were determined with a dual-luciferase assay system. Columns are the mean of three independent experiments; bars, SD. The differences between means were significant (ANOVA P < 0.005). P values calculated with Turkey-Kramer HSD test are: Ã , P < 0.05; ÃÃ , P < 0.01; ÃÃÃ , P < 0.001. (PDF)

S6 Fig. High peaks and signals of DEC1 (also known as BHLHE40) in known target genes.
Using the UCSC genome browser database, we analyzed the BHLHE40 ChIP-sequence peaks and signals near the transcription start site of already-reported DEC1 target genes, BHLHE41 (also known as DEC2), DBP, PER1, and PER2. (PDF) S7 Fig. High peaks and signals of DEC1 (also known as BHLHE40) in the genes regulated by HIF-1-DEC pathway. Using the UCSC genome browser database, we analyzed the BHLHE40 ChIP-sequence peaks and signals near the transcription start site of HIF-1-DEC pathway-regulated genes, MLH1, STAT1, and SREBP-1c (also known as SREBF1). (PDF) S8 Fig. High peaks and signals of DEC1 (also known as BHLHE40) in genes related to DNA-DRR. Using the UCSC genome browser database, we analyzed the BHLHE40 ChIPsequence peaks and signals near the transcription start site of genes related to DNA-DRR listed in S3 Table. (PDF)