RESEARCH PAPER
RESEARCH PAPER
Epigenetics 5:6, 547-556; August 16, 2010; © 2010 Landes Bioscience
Ischemia dysregulates DNA methyltransferases and p16INK4a methylation in human colorectal cancer cells Karolina Skowronski,1 Sonam Dubey,1 David Rodenhiser2 and Brenda L. Coomber1 Department of Biomedical Sciences; Ontario Veterinary College; University of Guelph; Guelph, ON CA; 2Departments of Biochemistry, Oncology and Paediatrics; University of Western Ontario; London Regional Cancer Centre and Children’s Health Research Institute; London, ON CA
1
Key words: DNA methylation, DNA methyltransferases, colorectal carcinoma, ischemia, p53, hypoxia, hypoglycaemia Abbreviations: DNMT, DNA methyltransferase; CRC, colorectal carcinoma; CpGi, CpG islands; 5-mc, 5-methylcytosine; CA IX, carbonic anhydrase IX; HRE, hypoxia response element; NC2, negative cofactor 2; H3K9me2, histone H3 lysine 9 dimethylation; LOI, loss of imprinting
Epigenetic modifications are involved in the initiation and progression of cancer. Expression patterns and activity of DNA methyltransferases (DNMTs) are strictly controlled in normal cells; however, regulation of these enzymes is lost in cancer cells due to unknown reasons. Cancer therapies which target DNMTs are promising treatments of hematologic cancers, but they lack effectiveness in solid tumors. Solid tumors exhibit areas of hypoxia and hypoglycaemia due to their irregular and dysfunctional vasculature, and we previously showed that hypoxia reduces global DNA methylation. Colorectal carcinoma (CRC) cells (HCT116 and 379.2; p53+/+ and p53-/-, respectively) were subjected to ischemia (hypoxia and hypoglycaemia) in vitro and levels of DNMTs were assessed. We found a significant decrease in mRNA for DNMT1, DNMT3a and DNMT3b, and similar reductions in DNMT1 and DNMT3a protein levels were detected by western blotting. In addition, total activity levels of DNMTs (as measured by an ELISA-based DNMT activity assay) were reduced in cells exposed to hypoxic and hypoglycaemic conditions. Immunofluorescence of HCT116 tumor xenografts demonstrated an inverse relationship between ischemia (as revealed by carbonic anhydrase IX staining) and DNMT1 protein. Bisulfite sequencing of the proximal promoter region of p16INK4a showed a decrease in 5-methylcytosine following in vitro exposure to ischemia. These studies provide evidence for the downregulation of DNMTs and modulation of methylation patterns by hypoxia and hypoglycaemia in human CRC cells, both in vitro and in vivo. Our findings suggest that ischemia, either intrinsic or induced through the use of anti-angiogenic drugs, may influence epigenetic patterning and hence tumor progression.
Introduction Cancer is a genetic disease initiated by modifications to gene expression via alterations to DNA, such as point mutations or changes in copy number and by non-coding modifications to chromatin.1 Non-coding modifications, or epigenetic alterations,2 include post-translational modifications to histones and DNA methylation. These epigenetic alterations cause modifications in chromatin configuration, leading to changes in gene expression. DNA methylation is the covalent addition of a methyl group to cytosine bases which are 5' to guanine bases—5'CpG 3',4 resulting in gene silencing.1,3 Concentrated regions of CpGs, referred to as CpG islands (CpGi), are generally unmethylated and located in the 5' untranslated region, promoters and first exons and are found in approximately 60% of genes.5 Exceptions to this general pattern include imprinted genes and inactivated X
chromosomes, which are generally hypermethylated. CpGs not located in islands are generally methylated and make up 3–4% of the cytosines in a genome.4 Non-coding regions of DNA, including repeat elements, viral sequences and transposons are heavily methylated to prevent the transcription of these sequences.4 DNA methylation is catalyzed by enzymes known as DNA methyltransferases (DNMT). There are three main members of the DNMT family: DNMT1, DNMT3a and DNMT3b.5 DNMT1, the maintenance methyltransferase is present in high amounts in somatic tissue.6 During DNA replication, DNMT1 copies the methylation pattern from the template strand to the newly synthesized strand.6 DNMT3a and 3b are the de novo methyltransferases and have a preference for unmethylated DNA. These enzymes are active post-replication and establish methylation patterns during gametogenesis and embryogenesis, but have low expression in normal somatic tissue.5
*Correspondence to: Brenda L. Coomber; Email:
[email protected] Submitted: 01/29/10; Accepted: 05/19/10 Previously published online: www.landesbioscience.com/journals/epigenetics/article/12400 DOI: 10.4161/epi.5.6.12400 www.landesbioscience.com Epigenetics
547
Epigenetic regulation plays a significant role in tumor formation and progression. Cancer cells, when compared to normal cells, are generally characterized as having global hypomethylation combined with selective promoter hypermethylation.4 Loss of normal DNA methylation marks lead to cancer development by inducing re-expression of viral genes, activation of oncogenes, loss of imprinting, X chromosome activation and genomic instability.4 More commonly noted are the effects of DNA hypermethylation, which leads to the silencing of cell cycle control and other tumor suppressor genes.7 Abnormal epigenetic alterations in cancer are hypothesized to arise via genetic disruptions in the enzymes which catalyze chromatin modifications, but this remains unclear. Overexpression of DNMT1, 3a and 3b resulting in DNA hypermethylation has been reported in many different cancer types, such as breast, hepatocellular, stomach and colorectal carcinoma (CRC).8-10 It has been hypothesized that more genes are disrupted in cancer cells by epigenetic modifications than by genetic mutations.11 Angiogenesis, defined as the growth of new blood vessels from pre-existing vasculature, is essential in reproduction, development and wound healing.12 However, there is also a broad range of diseases dependent on pathological angiogenesis such as autoimmune disorders, age-related macular degeneration, atherosclerosis and cancer.13 Angiogenesis is a fundamental step in tumor progression, as a solid tumor cannot grow larger then 2–3 mm3 without a blood supply.14 Both the structure and function of tumor vessels are abnormal and characterized as disorganized, tortuous, dilated, of uneven diameter and with excessive branching and shunts.15 As a result, blood flow in tumors is chaotic and variable and leads to heterogeneous tissue microenvironments with areas of hypoxia, acidity and low nutrient availability.16 Previous results from our laboratory showed that hypoxia decreased the global levels of 5-methylcytosine in certain tumors and cancer cell lines.17 Here we further pursue the mechanisms behind this finding by determining the expression patterns and activity levels of DNMTs under hypoxia and hypoglycaemia, as models of tumor ischemia. Tumor protein p53, a critical tumor suppressor, is mutated in approximately 50% of all primary human cancers.18 In CRC, about 40–50% of all cases have a mutation in the p53 gene.19 Deletion of p53 in HCT116 CRC cells resulted in increased transcription and translation of DNMT1, presumably through p53 mediated gene repression.20 Therefore, loss of p53 may significantly contribute to aberrant genomic methylation through dysregulation of DNMT expression. Interestingly, mice that lack functional p53 have increased DNMT1 and DNMT3b message and protein in their somatic tissues.21 We therefore hypothesized that expression of DNMT enzymes under ischemic conditions might be affected by p53 status, and investigated this in a paired set of p53 +/+ and p53-/- human colorectal cancer cells. Results Cellular responses to ischemia. BrdU incorporation into DNA was used to assess the proliferation status of HCT116 cells in ischemic conditions. Both p53 +/+ and p53-/- cells continue to
548
proliferate following ischemic exposure (Fig. 1). The proportion of proliferating cells (compared to total number of cells) did not change significantly between control, no oxygen or no glucose conditions in p53 +/+ cells (p ≥ 0.05; Fig. 1A). In p53-/- cells, nearly 100% of cells were proliferating in all three conditions (p ≥ 0.05; Fig. 1B). Apoptosis was measured through caspase-3 cleavage after ischemic exposure. p53 +/+ cells did not show any detectable caspase-3 cleavage following no oxygen or no glucose treatment (Fig. 1C). Similarly, p53 -/- cells did not show any detectable caspase-3 cleavage following no oxygen, although some evidence of cleaved caspase-3 was detected following the no glucose treatment of these cells when the western blot was exposed for a prolonged length of time. Hypoglycaemia downregulates transcription of DNMTs. To determine if the transcription of DNMTs is affected by ischemic conditions, quantitative real-time RT-PCR was performed. DNMT1 and DNMT3a transcription levels were significantly repressed in low glucose compared to control conditions, in p53 +/+ but not p53-/- HCT116 cells (p ≤ 0.05; Fig. 2A and B). In p53-/- cells, DNMT3a message was significantly reduced under low glucose conditions when compared to message levels in low oxygen; low oxygen alone did not alter DNMT1 and DNMT3a message significantly. DNMT3b transcription levels were altered the most with significant reductions from control levels in both p53 +/+ and p53-/- cells under low glucose and low oxygen conditions (Fig. 2C). Variable effects of ischemic conditions on DNMT1 and DNMT3a protein levels. Next, we investigated how ischemic conditions affected DNMT expression at the protein level. DNMT1 protein levels in p53 +/+ cells were significantly reduced (p ≤ 0.05) as compared to control by both low oxygen and low glucose, however, DNMT1 protein in p53-/- cells remained unchanged (Fig. 3A and B). The effects of ischemic conditions on DNMT3a protein levels were very different from those observed for DNMT1 (Fig. 3C and D). In low glucose, DNMT3a protein levels were greatly reduced in both p53 +/+ cells and p53-/- cells, and in some replicates, there were negligible levels of the protein (Fig. 3D). Low oxygen exposure had no significant impact on DNMT3a protein levels in either p53 +/+ or p53-/- CRC cells. As seen in Figure 3D, p53-/- cells express more DNMT3a then p53 +/+ cells. DNMT activity declines in ischemic conditions. When total DNMT activity was assessed using an ELISA-based assay, a reduction in enzyme activity was seen in low oxygen and low glucose conditions in the p53-/- cells, compared to control (p ≤ 0.05; Fig. 4). Though DNMT activity was also reduced in p53 +/+ cells grown in ischemic conditions, this was not statistically significant (p ≥ 0.05). Qualitative in vivo evidence for downregulation of DNMT1 in hypoxia. Immunofluorescent staining performed on tumor xenografts revealed a heterogeneous population of cells in both p53 +/+ and p53-/- HCT116 tumors. Some carbonic anhydrase IX (CA IX) positive cells (i.e., cells in hypoxic regions of tumor), showed robust DNMT1 staining, while others showed a clear absence of expression (Fig. 5D). No quantitative differences in the number of hypoxic, DNMT1 negative cells were detectable
Epigenetics
Volume 5 Issue 6
between the p53 +/+ and p53-/- tumors as determined by area measurements counts (data not shown). Recovery time for DNMT1. HCT116 p53 +/+ cells grown in no oxygen required between 8 and 12 h return to normoxia to recover DNMT1 to preischemic levels. Cells grown in no glucose required between 4 and 8 h return to normal glucose media to recover DNMT1 to pre hypoglycaemic levels (Fig. 6). p16INK4a promoter methylation altered by ischemia. A 269 bp region of the human p16INK4a promoter located between positions -601 and -332 from the start codon, containing two transcriptionally essential GC boxes27 was assessed via bisulfite sequencing following 48 h exposure to no glucose or no oxygen. In control conditions, approximately 50% of clones assessed were fully methylated in both p53 +/+ and p53-/- HCT116 cells. Following exposure to no oxygen, p53 +/+ cells had a statistically significant increase (p ≤ 0.05) in methylation while p53-/- cells had no change in their methylation patterns in the region of interest. Following exposure to no glucose, both p53 +/+ and p53-/- cells showed significant decreases in methylation compared to control (p ≤ 0.05; Fig. 7). Discussion This study demonstrates that ischemic conditions, specifically hypoxia and hypoglycaemia, have significant affects on DNMT expression and activity in human colorectal cancer. The p53 status of cells plays a significant role in this observed dysregulation of DNMT expression and varies in outcome depending on the enzyme examined. DNMT1 protein was only downregulated in p53 +/+ cells, by both hypoxia and hypoglycaemia. Hypoglycaemia significantly reduced DNMT3a protein in both p53 +/+ and p53-/- cells, while DNMT3b mRNA was reduced by both hypoxia and hypoglycaemia, regardless of p53 status. Although a minimal amount of caspase-3 Figure 1. Cellular proliferation demonstrated by BrdU incorporation into HCT116 p53+/+ cleavage was seen in p53 -/- cells under hypoglyce(A) and p53-/- (B) cells grown under control, no oxygen and no glucose conditions. mia, cell cycling was not affected by our ischemic Arrows indicate cells which are proliferating (BrdU positive), while arrow heads indicate conditions thus the expression patterns we report BrdU negative cells. Scale bar represents 100 µm. (C) We examined apoptosis under ischemic conditions via western blotting for caspase-3 cleavage. Top: blot shows here are likely due to other factors besides differno detectable cleaved caspase-3 in any of the samples. Bottom: the same blot after ential cell survival. Overall, we find that ischemic prolonged exposure reveals low amounts of cleaved (activated) caspase-3 in p53-/- cells conditions modify DNMT expression profiles, and exposed to no glucose conditions. may account for the irregular methylation patterns seen in solid tumors.28 When HCT116 cells lacking a functioning p53 were exposed that hypoxia can also elevate the expression of DNMT3a in to hypoxic conditions, a modest (although not statistically colorectal cancer cells, which could potentially lead to abnormal significant) increase in DNMT3a protein levels was seen. hypermethylation in hypoxic regions of a tumor. We also found Interestingly, Park et al. reported approximately 85% increased the combination of hypoxia and p53 deletion/mutation to have DNMT3a expression in the liver of pre-malignant, p53 -/- even greater effects on this enzyme’s expression then either mice when compared to p53 +/+ mice.21 We demonstrate here stress alone. Studies have examined DNMT transcriptional
www.landesbioscience.com Epigenetics
549
Figure 2. Quantitative real-time reverse transcription PCR. To represent mRNA, cDNA levels of DNMT1 (A), DNMT3a (B) and DNMT3b (C) from p53+/+ and p53-/- HCT116 cells, after 24 h in control (C), low oxygen (LO) and low glucose (LG). mRNA levels were normalized to β-actin expression. DNMT expression of each cell line was normalized to the control condition. Averages of 5 independent experiments with standard error are shown. *Indicates p ≤ 0.05 relative to cell type respective control. **Indicates p ≤ 0.05 relative to low oxygen.
550
Epigenetics
Volume 5 Issue 6
regulation by identifying several transcription factor binding sites in the promoter region of DNMT1, such as for activator protein 1 (AP-1), E2F, 29 Sp1 and Sp3,30 supporting the concept that DNMT1 expression is cell cycle dependent. However, our ischemic conditions did not significantly alter cell proliferation in HCT116 cells and we therefore conclude that any ischemia-mediated changes in DNMT levels reported here are likely due to other pathways. A possible mechanism for the observed modest increase in DNMT3a expression could be the hypoxia inducible factor (HIF1) pathway, which acts through the binding of HIF to the hypoxia response element (HRE) DNA recognition sites.31 Database searches of 1,500 bp promoter region upstream of the start site for all three DNMTs revealed that they each contain at least one putative conserved core region (A/GCGTG) of the HRE (data not shown). However, we did not see significant increases in gene expression, thus these sites are unlikely to be active in our system. This is consistent with the fact that the flanking regions of HRE are highly variable and critical for HRE function.31 Regarding DNMT1 and DNMT3b hypoxia-mediated repression, repressors such as negative cofactor 2 (NC2) have been reported to be stimulated by hypoxia and could be a possible mechanism for DNMT downregulation in ischemic conditions.32 To evaluate the relationship between DNMT1 and hypoxia in vivo, we performed dual immunofluorescence for DNMT1 and carbonic anhydrase IX (CA IX). CA IX is a transmembrane protein that catalyzes the hydration of carbon dioxide to carbonic acid, and has been shown to be overexpressed in hypoxia and therefore a useful endogenous marker of tissue ischemia.33 Our analysis of the immunostaining showed that while some cells stained positive for both hypoxia and DNMT1, many regions within these tumors showed an inverse relationship with increased hypoxia resulting in decrease DNMT1. Evaluation of biopsies from colorectal adenocarcinoma patients found a range of pimonidazole staining (an exogenous marker of hypoxia) of 2.2–37.8% of hypoxic cells, with hypoxic regions spread heterogeneously throughout the tumor.34 Therefore, with hypoxia being prevalent in human clinical CRC, our findings suggest DNMT expression, and hence DNA methylation, could be regionally disrupted throughout a solid tumor by hypoxia. We previously demonstrated that in vitro ischemia leads to reduced levels of 5-methylcytosine in HCT116 cells and that CRC xenografts contain less 5-methylcytosine in ischemic regions than in normoxic regions,17 consistent with the downregulation in DNMT activity we report here.
Figure 3. Protein expression profiles of DNMT1 and DNMT3a. Protein levels of DNMT1 (A and B) and DNMT3a (C and D) in p53+/+ and p53-/- HCT116 cells cells after 24 h in control (C), low oxygen (LO) and low glucose (LG). DNMT1 (200 kDa) and DNMT3a (130 kDa) protein levels were normalized to α-tubulin (50 kDa) to account for any protein loading discrepancies. Averages of three independent experiments with standard error are shown. *indicates p ≤ 0.05 relative to cell type respective control.
www.landesbioscience.com Epigenetics
551
Figure 4. Total DNMT activity. Activity levels of total DNMTs were measured with EpiQuikTM DNA Methyltransferase Activity/Inhibition Assay Kit. Nuclear protein was isolated from p53+/+ and p53-/HCT116 cells after 24 h in control (C), low oxygen (LO) and low glucose (LG). The assay measures the amount of DNA methylation, which is proportional to DNMT activity. The average of four independent experiments with standard error is shown. *indicates p ≤ 0.05 relative to cell type respective control.
Our results also agree with previous studies that DNMT1 expression is mediated by p53,20,35 and we extend this finding by showing that ischemia-mediated downregulation of DNMT1 does not occur in p53-/- cells. However, our study is the first to demonstrate a downregulation in DNMT activity by ischemia, in both p53 +/+ and p53-/- cells. Although we observed mild increases in DNMT3a protein levels in the face of decreased enzyme activity, one explanation for this apparent contradiction is that DNMT1 is the main methyltransferase,6 therefore low levels of active DNMT3a in these cells may not have measurable effects on total DNA methylation activity. The impact of this ischemia-mediated regulation of DNMT activity was further pursued by examining DNA methylation patterns at a known hypermethylated gene, p16INK4a. p16INK4a, a cyclin-dependent kinase inhibitor, is very commonly inactivated or mutated in many types of cancer including CRC, and loss of expression of this tumor suppressor can occur due to promoter DNA hypermethylation.36 In normal somatic tissue, p16INK4a remains unmethylated, allowing for normal cell cycle control by binding cdk4/cdk6.36 However, once p16INK4a is silenced or mutated, cdk4 and cdk6 are available to bind to cyclin D, resulting in loss of cell cycle control.37 In HCT116 cells, one allele of p16INK4a is hypermethylated in the promoter region,38 consistent with our finding that 50% of bisulfite sequencing clones in control cells were fully methylated for the region we evaluated. Therefore, it was of interest to assess the functional consequence of DNMT dysregulation by ischemia in the promoter region of p16INK4a. After cells were grown for 48 h in a glucose-free media, decreases in the amount of methylated clones/cells were seen (in both p53 +/+ and p53-/- cells), matching what was observed in DNMT activity. However, this hypomethylation was not sufficient to induce re-expression of p16 as detectable Figure 5. Detection of DNMT1 and hypoxic regions in HCT116 xenografts by immunofluorescence. Blue channel is nuclear staining by DAPI (A), red channel is DNMT1 (B), green is CA IX/hypoxia marker (C), and overlay of DNMT1 and CA IX (D). Arrows indicate cells which are hypoxic and DNMT1 negative. Asterisks indicate cells which are both hypoxic and expressing DNMT1. Scale bar represents 50 µm. 552
Epigenetics
Volume 5 Issue 6
Figure 6. Western blot showing recovery of DNMT1 protein levels in HCT116 cells after 24 h in no oxygen or no glucose. Protein was collected after cells were restored to normal conditions for 0, 4, 8, 12 and 24 h.
Figure 7. Map of proximal p16INK4a promoter region chosen for analysis; each circle represents a CpG (A). Detection of DNA methylation changes using bisulfite sequencing in human p16INK4a promoter region following exposure to ischemia in vitro for 48 h (B). Open and closed circles represent unmethylated and methylated CpGs, respectively. A dash is an unknown methylation status. In p53+/+ cells, there were significant changes in p16INK4a methylation after both no oxygen and no glucose treatments (p ≤ 0.05). In p53-/- cells, there was a significant decrease in methylation following no glucose treatment (p ≤ 0.05).
by qRT-PCR for mRNA, or western blotting for protein (results not shown), perhaps due to concomitant repressive alterations to histone proteins, an area under study in our laboratory. There is evidence for ischemia mediated regulation of other epigenetic marks such as the histone modification dimethylated histone H3 lysine 9 (H3K9me2), which results in gene repression and silencing. Costa and colleagues showed that hypoxia increased H3K9me2, as well as the histone methyltransferase G9a protein and enzyme activity in human lung carcinoma cells.39 Promoter regions of Mlh1 and Dhfr genes had increased H3K9me2 and decreased expression as a result of the hypoxia.39 A recent study has shown that hypoxia decreased histone H3 acetylation at the Mlh1 promoter in mouse colon cells and resulted in reduced MLH1 expression at mRNA and protein levels.40 Solid tumors undergo angiogenesis to meet the increased demand for oxygen and nutrients. These tumor blood vessels are often abnormal and result in chronic or transient ischemia.16 According to our findings, the ischemia that a solid tumor is commonly subjected to34 is helping to drive epigenetic disruption. Anti-angiogenic drugs designed to cut off blood supply to tumors may also be accelerating ischemia in solid tumors41 and as a result these therapies may unintentionally alter epigenetic processes. If the tumor microenvironment is modifying DNMT expression and activity, therapies targeted at inhibiting DNMT1 may not be effective if DNMT1 levels or activity are already being reduced by ischemic conditions. To date, several clinical trials that explored the use
of DNMT inhibitors reported poor results with low rates of complete remission (CR) and partial response (PR) in solid tumors.42 Clinical success (improved CR and PR percentages and hematologic recovery) has been seen with DNMT inhibitors in hematologic malignancies,42 but extending these results to solid tumors remains a challenge. Therefore, it is of interest to better understand
www.landesbioscience.com Epigenetics
553
Table 1. Primers and sequences used in quantitative real-time PCR and bisulfite sequencing Primers
Sequences
Annealing temp (°C)
DNMT1 106 bp (NM_001379.2)
F: AGA ACC AAC ACC CAA ACA G R: GCT TCT TCT CAT CTT TCT CGT
60.0
DNMT3a 118 bp (NM_175629.1)
F: GTG AGG ACC ATT ACT ACG AG R: CCA AAT ACC CTT TCC ATT TCA G
60.0
DNMT3b 171 bp (NM_006892.3)
F: CCT TAC CAT CGA CCT CAC AG R: CTC CTT CCC ATC CTG ATA CTC
60.0
β-actin 171 bp (NM_001101.3)
F: AAG ATC AAG ATC ATT GCT CCT C R: CAA CTA AGT CAT AGT CCG CC
60.0
F: GGT GGG GTT TTT ATA ATT AGG AAA G R: AAA CTA AAC TCC TCC CCA CCT AC
60.0
p16INK4a 269 bp (NG_007485)
Materials and Methods
Sizes of PCR products are shown under respective primer names. ccession numbers are shown in parenthesis. A
the dynamics of DNMTs in solid tumors, to improve on this therapeutic approach. Previous studies have shown genetic depletion of DNMT1 resulted in lower maintenance methyltransferase activity, global and gene-specific demethylation, and re-expression of silenced tumor suppressor genes in HCT116 cells.43 Our study demonstrates that ischemic conditions could be endogenously reducing DNA methylation, leading to re-expression of silenced tumor suppressor genes. Another potential result of ischemia-mediated DNMT downregulation could be increased genomic instability due to the movement of transposable elements, activation of viral DNA and/or loss of imprinting (LOI), resulting in a highly mutated genome. For example, knockout of DNMT-1 and -3a in HCT116 cells has been shown to lead to hypomethylation at the IGF2 gene, resulting in LOI.44 As well, mouse embryos deficient in DNMT1 have a 50–100-fold increase in intracisternal A particle retrovirus transcripts,45 demonstrating that a loss of DNMT1 can activate previously silenced retroviral DNA. Erythropoietin (EPO) is a potential target of hypomethylation in ischemic cells. EPO is a hypoxia-induced cytokine that stimulates erythropoiesis, and many studies have shown that EPO protects against tissue ischemia.46 Interestingly, the CpG regions in the HRE in the EPO promoter must remain unmethylated in order for hypoxia-induced transcription to occur.47 Thus the promoter region of EPO is a target for ischemia-mediated hypomethylation as an adaptive mechanism in ischemic tissue. There are several other genes involved in adaptation to ischemia which utilize the HIF1 pathway, such as vascular endothelial growth factor (VEGF) and inducible nitric oxide synthase (iNOS).47 We have demonstrated that the tumor microenvironment may be an important modulator of DNMT expression and the effects of ischemia on global methylation and expression changes in human colorectal cancer are currently being investigated. The impact of agents which modulate the tumor
554
microenvironment (such as anti-angiogenic therapy) on DNA methylation and the effectiveness of DNMT inhibitors also remains unknown, and further studies are needed to answer this important question.
Cell culture. HCT116 (p53 +/+) (ATCC, Manassas, VA, USA) and 379.2 (a derivative of HCT116; p53 -/-)18 cells were cultured in DMEM (Sigma-Aldrich, Oakville, ON, Canada) supplemented with 10% fetal bovine serum (FBS), 50 µg/ml gentamicin, and 1 mM sodium pyruvate. Cells were grown at 37°C in a humidified chamber with 5% CO2. In vitro ischemia. Hypoxic/anoxic conditions were generated using a Modular Incubator Chamber (Billups-Rothenberg Inc., Del Mar, CA, USA) with continuous flushing with a humidified mixture of 95% N2 and 5% CO2. Hypoglycaemic conditions were mimicked with the use of glucose free and pyruvate free DMEM (Invitrogen, Burlington, ON, Canada). Confluent monolayers of cells were trypsinized, and 1 x 106 cells were plated onto 60 mm plates and left in standard culture conditions overnight. The following day cells were washed with PBS (Sigma-Aldrich) and switched to low serum media: DMEM and 2% FBS. 24 hours thereafter, cells were washed with PBS again and assigned to three groups; control, hypoxic, and hypoglycaemic. Cells were incubated for 24 or 48 h, and lysed for protein, DNA or RNA isolation. Each experiment was performed at least in triplicate. For the DNMT1 recovery experiment, cells were exposed for 24 h to ischemic conditions. Afterwards, media was changed to regular DMEM with glucose and cells were placed into an incubator with standard oxygen levels to allow for DNMT1 levels to recover. Protein was collected 0, 4, 8, 12 and 24 h post incubation in normal conditions. Cellular proliferation assay. To confirm that HCT116 cells proliferate under in vitro ischemia, cells grown in hypoxic or hypoglycaemic conditions for 24 h were treated with 10 µM Bromodeoxyuridine (BrdU; BD Biosciences, Mississauga, ON, Canada) for 2 h. Cells were then fixed with methanol and denatured with 4 M HCl twice for 15 min, then stained with FITC conjugated rat anti-BrdU (1:50; Abcam, Cambridge, MA, USA) for 1 h at room temperature and assessed for BrdU incorporation into DNA. Nuclei were counterstained with a 1:3,600 dilution of DAPI (Invitrogen). RNA isolation and real-time PCR. Total RNA was isolated from HCT116 p53 +/+ and p53-/- cells following 24 h in normal (control), hypoxic and hypoglycaemic conditions using TRIPure (Roche Applied Science, Laval, QC, Canada) according to manufacturer’s instructions. Total RNA was DNase treated with Turbo DNase I (Ambion, Austin, TX, USA), quantified with a spectrophotometer (Nanodrop®) and 5 µg was reverse transcribed and treated with RNase H (Invitrogen). Quantitative Real-time PCR was performed as described in manufacture’s instructions for LightCycler® FastStart MasterPlus SYBR Green I (Roche Applied Science) using a 1.5 LightCycler (Roche Applied Science). Reaction conditions were as followed: 95°C for 10 min
Epigenetics
Volume 5 Issue 6
followed by 45 cycles of 10 s at 95°C, 10 s at 60°C and 12 s at 72°C; see Table 1 for primer sequences. All primers sets generated one amplification peak/product, which was confirmed with agarose gel electrophoresis and sequencing. Expression of genes of interest was normalized to beta-actin. Western blotting. After incubation, adherent cells were lysed with whole cell lysis buffer (Cell Signaling Technology Inc., Danvers, MA, USA) containing 1.9 ng/ml aprotinin (Sigma-Aldrich) and 10–50 µg total protein as determined by Dc Protein assay (Bio-Rad Laboratories, Hercules, CA, USA) was loaded on a 7.5% SDS-polyacrylamide gel and separated by electrophoresis. Proteins were transferred to a polyvinylidene difluoride membrane and blocked in 5% non-fat milk. All membranes were incubated with primary antibodies in 5% nonfat milk diluted in Tris-buffered saline/Tween 20 overnight at 4°C, and secondary antibodies were applied for 30 min at room temperature. The following antibodies and dilutions were used: mouse anti-DNMT1 antibody (1:400; Imgenex Corp.; San Diego, CA, USA); mouse anti-α-tubulin (1:200,000; SigmaAldrich); rabbit anti-DNMT3a and rabbit anti-caspase-3 (both at 1:1,000; Cell Signaling Technology Inc.). POD conjugated secondary goat anti-rabbit or goat-anti mouse antibodies (Sigma-Aldrich) were used at 1:5,000–40,000. After washing in Tris-buffered saline/Tween 20, membranes were exposed to chemiluminescence solution (Roche Applied Science) as per kit instructions. Proteins were visualized by X-ray film (Konica, Ramsey, NJ, USA) exposure and densitometric analysis was performed using ImageJ software.22 DNMT1/3a levels were normalized to α-tubulin signal. DNMT activity assay. Nuclear protein was isolated with EpiQuikTM Nuclear Extraction Kit I (Epigentek, Brooklyn, NY, USA) from cells exposed to ischemic-mimicking conditions for 24 h, as described above. Following protein quantification (BioRad Laboratories), 12 µg of nuclear protein was used in the EpiQuikTM DNA Methyltransferase Activity/Inhibition Assay Kit (Epigentek) according to the manufacturer’s protocol to measure total DNMT activity. In vivo evaluation. One million HCT116 p53 +/+ or p53-/- cells were injected subcutaneously into the right flank of ten RAG1-/- immunodeficient mice23 from the University of Guelph colony. Six HCT116 p53 +/+ and four p53-/- tumors were grown. All rodent work was done according to the guidelines of the Canadian Council on Animal Care as approved by the University of Guelph Animal Care Committee. Growth was monitored until tumors reached at least 500 mm3 in volume, then mice were euthanized by CO2 asphyxia and cervical dislocation, and tumor tissue was harvested, fixed in 10% buffered formalin, and paraffin embedded. 5 µm sections were cut, placed on slides, and deparaffinised. Antigen retrieval was performed in 10 mM sodium citrate, pH 6, heated to 95°C. Samples were blocked with DakoCytomation serum free protein block (DAKO, Glostrup, Denmark) for 1 h, then incubated with rabbit anti-CA IX primary antibody (1:500; AbCam) in phosphate buffered saline with 0.1% Tween (PBS-T) for 1 h, followed by donkey anti-rabbit Alexa Flour 488 (1:500; Invitrogen) for 30 min at 22°C. After washing in PBS, specimens were incubated with mouse anti-DNMT1
(1:250; Imgenex) over night at 4°C followed by goat anti-mouse Cy3 (1:200; Jackson Immunoresearch, West Grove, PA, USA) for 1 h at 22°C. Nuclei were counterstained with a 1:3,600 dilution of DAPI (Invitrogen), slides were rinsed and mounted using Dako Florescent mounting medium (DAKO) and stored at 4°C in the dark until imaging. Control slides received PBS as a substitute for primary antibody. Images were captured using a Leica Opti-Tech epifluorescence microscope equipped with appropriate excitation and emission filters, and Q-Capture software. Images were merged using Adobe Photoshop 5.0 (Adobe, San Jose, CA, USA). The immunostained samples were randomized and analyzed in a blinded fashion. Tissue regions were classified into three groups: DNMT1 positive, CA IX positive, or positive for both DNMT1 and CA IX staining. Five fields per tumor were imaged and Optimus software (Fort Collins, CO, USA) was used to measure areas for each group. Necrotic areas were determined by nuclear morphology and excluded from analysis. Bisulfite sequencing. DNA was isolated using the DNeasy Blood and Tissue Kit (Qiagen, Mississauga, ON, Canada) from cells grown in ischemia-mimicking conditions for 48 h. A 48 h treatment was used to allow for the ischemia-mediated changes to DNMT enzyme levels to translate to changes in DNA methylation. One µg of DNA was then bisulfite converted using the Epitect Bisulfite Kit (Qiagen) following the manufacture’s protocol. MethPrimer24 was used to design primers for the first CpG island downstream of the transcription start site of the p16INK4a promoter (Table 1). A hot start Taq polymerase (PlantinumInvitrogen) was used in the bisulfite PCR to amplify regions of interest. PCR reaction conditions were 94°C for 2 min, 40 cycles of 30 s at 94°C, 30 s at 60°C, 45 s at 72°C and a final cycle of 72°C for 7 min. PCR reaction product was run on a 2% agarose gel and the single band was excised and purified using the illustraTM GFX TM PCR DNA and Gel Band Purification Kit (GE Healthcare, Baie d’Urfe, QC, Canada). Due to the nature of bisulfite modification, sequencing directly following PCR yields inconclusive results due to varying conversion efficiency and varying amount of methylation.25 To obtain clear sequencing results, cloning was therefore performed as follows: purified PCR products were ligated into pGEM-T easy plasmids (Promega, Madison, WI) for two hours on ice. DH5alpha competent cells (Invitrogen) were transformed by heat shock and grown overnight on agar plates containing 100 µg/ml ampicillin (Sigma-Aldrich). At least 10 colonies/treatment were selected and expanded in LB media containing 50 µg/ml ampicillin and plasmids were isolated the following day using the PureLink plasmid miniprep kit (Invitrogen). Plasmid DNA containing the bisulfite-modified DNA was analyzed by Sanger-sequencing (Laboratory Services, University of Guelph) and BiQ Analyzer software26 was used to analyze methylation patterns and generate diagrams. Human methylated and bisulfite modified DNA (Qiagen) and unmethylated DNA (converted in parallel with samples; Qiagen) were used as controls for bisulfite conversion efficiency and accuracy. Statistical analysis. For the DNA methylation analysis by bisulfite sequencing, a Fisher’s exact test was performed. Oneway ANOVA was performed on all other data. If the p value was less then or equal to 0.05, then the Bonferroni correction was
www.landesbioscience.com Epigenetics
555
performed on pairs of data. Each assay was replicated at least three times. Acknowledgements
This study was supported by grants to B.L.C. from the Cancer Research Society, Inc., and the Canadian Cancer Society Research Institute. We would like to thank Dr. Bert Vogelstein, Johns Hopkins University for kindly providing the p53 -/- 379.2 References 1.
Grønbaek K, Hother C, Jones PA. Epigenetic changes in cancer. APMIS 2007; 115:1039-59. 2. Jain PK. Epigenetics: the role of methylation in the mechanism of action of tumor suppressor genes. Ann N Y Acad Sci 2003; 983:71-83. 3. Rodenhiser D, Mann M. Epigenetics and human disease: translating basic biology into clinical applications. CMAJ 2006; 174:341-8. 4. Hellebrekers DMEI, Griffioen AW, van Engeland M. Dual targeting of epigenetic therapy in cancer. Biochim Biophys Acta 2007; 1775:76-91. 5. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev 2002; 16:6-21. 6. Turek-Plewa J, Jadodzinski PP. The role of mammalian DNA methyltransferases in the regulation of gene expression. Cell Mol Biol Lett 2005; 10:631-47. 7. Robertson KD. DNA methylation and chromatin— unravelling the tangled web. Oncogene 2002; 21: 5361-79. 8. Girault I, Tozlu S, Lidereau R, Bièche I. Expression analysis of DNA methyltransferases 1, 3A and 3B in sporadic breast carcinomas. Clin Cancer Res 2003; 9:4415-22. 9. Sun L, Hui AM, Kanai Y, Sakamoto M, Hirohashi S. Increased DNA methyltransferase expression is associated with an early stage of human hepatocarcinogenesis. Jpn J Cancer Res 1997; 88:1165-70. 10. Kanai Y, Ushijima S, Kondo Y, Nakanishi Y, Hirohashi S. DNA methyltransferase expression and DNA methylation of CPG islands and peri centromeric satellite regions in human colorectal and stomach cancers. Int J Cancer 2001; 91:205-12. 11. Herman JG, Baylin SB. Gene silencing in cancer in association with promoter hypermethylation. N Engl J Med 2003; 349:2042-54. 12. Alessi P, Ebbinghaus C, Neri D. Molecular targeting of angiogenesis. Biochim Biophys Acta 2004; 1654: 39-49. 13. Folkman J. Angiogenesis: an organizing principle for drug discovery? Nat Rev Drug Discov 2007; 6: 273-86. 14. Masuda H, Kawamura K, Tohda K, Shozawa T, Sageshima M, Kamiya A. Increase in endothelial cell density before artery enlargement in flow-loaded canine carotid artery. Arteriosclerosis 1989; 9:812-23. 15. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000; 407:249-57. 16. Raghunand N, Gatenby RA, Gillies RJ. Microenvironmental and cellular consequences of altered blood flow in tumors. Br J Radiol 2003; 76: 11-22. 17. Shahrzad S, Bertrand K, Minhas K, Coomber BL. Induction of DNA hypomethylation by tumor hypoxia. Epigenetics 2007; 2:119-29. 18. Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature 2000; 408:307-10. 19. Russo A, Bazan V, Iacopetta B, Kerr D, Soussi T, Gebbia N. TP53-CRC Collaborative Study Group. The TP53 colorectal cancer international collaborative study on the prognostic and predictive significance of p53 mutation: influence of tumor site, type of mutation, and adjuvant treatment. J Clin Oncol 2005; 23:7518-28.
556
cells. We would also like to thank Dr. Sirin Yaeesh for her technical assistance with cloning, and other members of the Coomber laboratory for advice and support. Grant information: This study was supported by grants to B.L.C. from the Cancer Research Society, Inc., and to B.L.C. and D.R. from the Canadian Cancer Society Research Institute.
20. Peterson EJ, Bogler O, Taylor SM. p53-Mediated repression of DNA methyltransferase 1 expression by specific DNA binding. Cancer Res 2003; 63:6579-82. 21. Park IY, Sohn BH, Choo JH, Joe CO, Seong JK, Lee YI, et al. Deregulation of DNA methyltransferases and loss of parental methylation at the insulin-like growth factor II (Igf2)/H19 loci in p53 knockout mice prior to tumor development. J Cell Biochem 2005; 94:585-96. 22. Abramoff MD, Magelhaes PJ, Ram SJ. Image Processing with ImageJ. Biophotonics International 2004; 11: 36-42. 23. Mombaerts P, Iacomini J, Johnson RS, Herrup K, Tonegawa S, Papaioannou VE. RAG-1-deficient mice have no mature B and T lymphocytes. Cell 1992; 68:869-77. 24. Li LC, Dahiya R. MethPrimer: designing primers for methylation PCRs. Bioinformatics 2002; 18:1427-31. 25. Zhang Y, Rohde C, Tierling S, Stamerjohanns H, Reinhardt R, Walter J, et al. DNA methylation analysis by bisulfite conversion, cloning and sequencing of individual clones. Methods Mol Biol 2009; 507:177-87. 26. Bock C, Reither S, Mikeska T, Paulsen M, Walter J, Lengauer T. BiQ Analyzer: visualization and quality control for DNA methylation data from bisulfite sequencing. Bioinformatics 2005; 21:4067-8. 27. Wu J, Xue L, Weng M, Sun Y, Zhang Z, Wang W, et al. Sp1 is essential for p16 expression in human diploid fibroblasts during senescence. PLoS One 2007; 2:164. 28. Baylin SB, Herman JG, Graff JR, Vertino PM, Issa JP. Alterations in DNA methylation: a fundamental aspect of neoplasia. Adv Cancer Res 1998; 72:141-96. 29. Bigey P, Ramchandani S, Theberge J, Araujo FD, Szyf M. Transcriptional regulation of the human DNA Methyltransferase (dnmt1) gene. Gene 2000; 242: 407-18. 30. Kishikawa S, Murata T, Kimura H, Shiota K, Yokoyama KK. Regulation of transcription of the Dnmt1 gene by Sp1 and Sp3 zinc finger proteins. Eur J Biochem 2002; 269:2961-70. 31. Kaluz S, Kaluzová M, Stanbridge EJ. Regulation of gene expression by hypoxia: integration of the HIFtransduced hypoxic signal at the hypoxia-responsive element. Clin Chim Acta 2008; 395:6-13. 32. Denko N, Wernke-Dollries K, Johnson AB, Hammond E, Chiang CM, Barton MC. Hypoxia actively represses transcription by inducing negative cofactor 2 (Dr1/ DrAP1) and blocking preinitiation complex assembly. J Biol Chem 2003; 278:5744-9. 33. Olive PL, Aquino-Parsons C, MacPhail SH, Liao SY, Raleigh JA, Lerman MI, et al. Carbonic anhydrase 9 as an endogenous marker for hypoxic cells in cervical cancer. Cancer Res 2001; 61:8924-9. 34. Goethals L, Debucquoy A, Perneel C, Geboes K, Ectors N, De Schutter H, et al. Hypoxia in human colorectal adenocarcinoma: comparison between extrinsic and potential intrinsic hypoxia markers. Int J Radiat Oncol Biol Phys 2006; 65:246-54.
Epigenetics
35. Guo Z, Tsai MH, Shiao YH, Chen LH, Wei ML, Lv X, et al. DNA (cytosine-5)-methyltransferase 1 as a mediator of mutant p53-determined p16ink4A downregulation. J Biomed Sci 2008; 15:163-8. 36. Herman JG, Merlo A, Mao L, Lapidus RG, Issa JP, Davidson NE, et al. Inactivation of the CDKN2/p16/ MTS1 gene is frequently associated with aberrant DNA methylation in all common human cancers. Cancer Res 1995; 55:4525-30. 37. Auerkari EI. Methylation of tumor suppressor genes p16(INK4a), p27(Kip1) and E-cadherin in carcinogenesis. Oral Oncol 2006; 42:5-13. 38. Myöhänen SK, Baylin SB, Herman JG. Hypermethylation can selectively silence individual p16ink4A alleles in neoplasia. Cancer Res 1998; 58:591-3. 39. Chen H, Yan Y, Davidson TL, Shinkai Y, Costa M. Hypoxic stress induces dimethylated histone H3 lysine 9 through histone methyltransferase G9a in mammalian cells. Cancer Res 2006; 66:9009-16. 40. Edwards RA, Witherspoon M, Wang K, Afrasiabi K, Pham T, Birnbaumer L, et al. Epigenetic repression of DNA mismatch repair by inflammation and hypoxia in inflammatory bowel disease-associated colorectal cancer. Cancer Res 2009; 69:6423-9. 41. Shahrzad S, Shirasawa S, Sasazuki T, Rak JW, Coomber BL. Low-dose metronomic cyclophosphamide treatment mediates ischemia-dependent K-ras mutation in colorectal carcinoma xenografts. Oncogene 2008; 27:3729-38. 42. Issa JP, Kantarjian HM. Targeting DNA methylation. Clin Cancer Res 2009; 15:3938-46. 43. Robert MF, Morin S, Beaulieu N, Gauthier F, Chute IC, Barsalou A, et al. Dnmt1 is required to maintain CpG methylation and aberrant gene silencing in human cancer cells. Nat Genet 2002; 33:61-5. 44. Cui H, Onyango P, Brandenburg S, Wu Y, Hsieh CL, Feinberg AP. Loss of imprinting in colorectal cancer linked to hypomethylation of H19 and IGF2. Cancer Res 2002; 62:6442-6. 45. Walsh CP, Chaillet JR, Bestor TH. Transcription of IAP endogenous retroviruses is constrained by cytosine methylation. Nat Genet 1998; 20:116-7. 46. Paschos N, Lykissas MG, Beris AE. The role of erythropoietin as an inhibitor of tissue ischemia. Int J Biol Sci 2008; 4:161-8. 47. Wenger RH, Kvietikova I, Rolfs A, Camenisch G, Gassmann M. Oxygen-regulated erythropoietin gene expression is dependent on a CpG methylation-free hypoxia-inducible factor-1 DNA-binding site. Eur J Biochem 1998; 253:771-7.
Volume 5 Issue 6
