Induction of DNA Hypomethylation by Tumor Hypoxia

[Epigenetics 2:2, 119-129; April/May/June 2007]; ©2007 Landes Bioscience

Research Paper

Induction of DNA Hypomethylation by Tumor Hypoxia Siranoush Shahrzad Kelsey Bertrand Kanwal Minhas Brenda L. Coomber* Department of Biomedical Sciences; University of Guelph; Guelph, Ontario Canada *Correspondence to: Brenda Coomber; Department of Biomedical Sciences; Ontario Veterinary College; University of Guelph; Guelph, Ontario, N1G 2W1 Canada; Tel.: 519.824.4120; Fax: 519.767.1450; Email: [email protected] Original manuscript submitted: 02/12/07 Revised manuscript submitted: 06/11/07 Manuscript accepted: 06/14/07 Previously published online as an Epigenetics E-publication: http://www.landesbioscience.com/journals/epigenetics/abstract.php?id=4613

Key words 5-methylcytosine, cancer, hypomethylation, hypoxia, tumor microenvironment, ischemia, HPLC

Abstract In cancer, the extensive methylation found in the bulk of chromatin is reduced, while the normally unmethylated CpG islands become hypermethylated. Regions of solid tumors are transiently and/or chronically exposed to ischemia (hypoxia) and reperfusion, conditions known to contribute to cancer progression. We hypothesized that hypoxic microenvironment may influence local epigenetic alterations, leading to inappropriate silencing and re-awakening of genes involved in cancer. We cultured human colorectal and melanoma cancer cell lines under severe hypoxic conditions, and examined their levels of global methylation using HPLC to quantify 5-methylcytosine (5-mC), and found that hypoxia induced losses of global methylation. This was more extensive in normal human fibroblasts than cancer cell lines. Cell lines from metastatic colorectal carcinoma or malignant melanoma were found to be markedly more hypomethylated than cell lines from their respective primary lesions, but they did not show further reduction of 5-mC levels under hypoxic conditions. To explore these epigenetic changes in vivo, we established xenografts of the same cancer cells in immune deficient mice. We used Hypoxyprobe™ to assess the magnitude of tissue hypoxia, and immunostaining for 5-mC to evaluate DNA methylation status in cells from different regions of tumors. We found an inverse relationship between the presence of extensive tumor hypoxia and the incidence of methylation, and a reduction of 5-mC in xenografts compared to the levels seen in the same cancer cell lines in vitro, verifying that methylation patterns are also modulated by hypoxia in vivo. This suggests that epigenetic events in solid tumors may be modulated by microenvironmental conditions such as hypoxia.

Abbreviations 5-mC HPLC DNMT CRC LOI

5-methylcytosine high performance liquid chromatography DNA methyltransferase enzyme colorectal cancer Loss of imprinting

Acknowledgements We are grateful for financial support from the Canadian Cancer Society/National Cancer Institute of Canada (CCS/NCIC) and the Cancer Research Society (CRS) for this study. HPLC facilities and technical support were kindly provided by Investigative Science Incorporated (ISI), Burlington, ON, Canada. We would also like to thank Alyssa Foulkes and Barb Mitchell, University of Guelph, for assistance with mouse husbandry, and Dr. Senji Shirasawa, International Medical Center of Japan, for providing the Dks-8 cell line.

www.landesbioscience.com

Introduction Seventy to 80 percent of all CpG sites in human DNA are methylated at cytosine residues, generating the modified nuclide 5-methylcytosine (5-mC)1 (Fig. 1). The methylation process depends on the sum balance of several factors. These factors include several DNA methyltransferase enzymes (DNMTs), de-methylases, and methylation centers that trigger DNA methylation and are likely related to unusual tertiary DNA structures.2-8 Classical models propose that, once established during development, DNA methylation patterns are fixed and then perpetuated through the maintenance activity of a DNMT enzyme that recognizes newly synthesized DNA as hemi-methylated and rapidly converts it to fully methylated form.9 In cancer, the tight regulation of DNA methylation breaks down, and the distribution of 5-mC changes. Despite the hypermethylation of normally unmethylated promoter regions and the well-documented presence of increased DNMT1 activity and expression in cancer,10-13 all types of cancer examined to date display global hypomethylation.14-16 The processes of CpG island hypermethylation and global genomic hypomethylation seem to be very different. Researchers have characterized the functional and site-specific effects of aberrant methylation at specific CpGs in important cancer-related genes, and reported on the consequences of hypermethylation of promoter regions, however the effect of global hypomethylation and its role in cancer progression remains unclear. A global decrease in 5-mC levels is commonly observed in neoplasia, where it is an early event, as well as in aging cells. Functionally, hypomethylation may contribute to chromosomal instability in cancer and perhaps, to increased expression of selected affected genes such as the cdk inhibitor p16INK4a in primary and metastatic breast carcinoma17 and to ectopic expression of g-synuclein in cancers outside the CNS.18 Hypomethylation has also been found to affect repeat and unique sequences in parallel, and detection of partially hypomethylated XIST alleles in prostate cancer tissues was suggested to be useful for the identification of Epigenetics

119

Tumor Hypoxia Leads to Changes in DNA Methylation

Figure 1. Molecular Structure of 5-methylcytosine paired with guanine, each connected to their backbone made of sugar and phosphate groups.

cases with pronounced hypomethylation, which tend to be more aggressive.19 Decreased methylation of DNA can lead to loss of imprinting (LOI) and this can drive cellular proliferation in cancer. The clearest example of this phenomenon is the LOI at the IGF2/H19 region as a result of hypomethylation at the differentially methylated region of IGF2, an event seen in about 40% of colorectal carcinomas.20.21 The presence of both methylated and unmethylated DNA sequences within the same tumor sample suggests that epigenetic changes in the tumor cells may differ based on exposures to different tumor microenvironmental conditions such as hypoxia/ ischemia.22,23 It has long been known that as solid tumors continue to expand, the rate of angiogenesis is usually inadequate to provide uniform perfusion to the growing mass.24 Tumor vasculature is also abnormal, highly disorganized and irregular, with poor mural cell recruitment, disrupted hierarchical branching, and inconsistent flow in compressed or occluded segments.25,26 The net effect is that regions of solid tumors are transiently and/or chronically exposed to ischemia and reperfusion, with all of its attendant sequelae of hypoxia/anoxia, followed by relative hyperoxia, fluctuation in nutrient (especially glucose) levels, acidosis and other disruptions in pH, and toxic reactive oxygen species generation. As well as being mutagenic,23,27 such microenvironmental conditions may disrupt DNA methylation patterns, providing a potential link between extracellular environment, epigenetic alterations, and cancer progression. We therefore hypothesized that progression of solid cancers is, at least in part, influenced by local epigenetic alterations (change in DNA methylation patterns) related to vascular proximity and tumor perfusion, leading to inappropriate silencing and awakening of genes involved in cancer. Here we employed cell culture and experimental tumor systems and investigated to what extent global hypomethylation is related to the presence of hypoxia in the microenvironment of solid tumors.

MATERIALS AND METHODS Cell lines. Human colorectal cancer (CRC) cell lines, HCT116, DLD-1, SW480, and SW620 and melanoma cell lines, WM115 and WM239 were obtained from ATCC. Dks-8 was a K-ras wildtype subline of DLD-1 [28]. The SW480 cell line was derived from a primary lesion (Dukes stage B colon carcinoma) from a 50-year old Caucasian male patient. The SW620 cell line was cultured from a lymph node metastasis in the same patient at a later time [29,30]. WM239 cell line was established from a metastatic human skin lesion, where as the WM115 cell line was isolated from a late vertical growth phase lesion [31-33]. Human dermal fibroblasts were purchased from ATCC and used as non-cancer control cells. In vitro exposure to microenvironmental conditions. Cells were maintained in standard culture conditions: DMEM (Sigma, Canada) supplemented with 10% heat-inactivated fetal bovine serum, 50 µg/ml gentamicin, 1 mM sodium pyruvate, and 1.25 µg /ml amphotericin B as 5 µl/ml of Fungizone Antimycotic (Invitrogen/Gibco, Canada). Cells were cultured at 37˚C in a humidified atmosphere containing 120

5% CO2. Hypoxic conditions were achieved using a Modular Incubator Chamber (Billups-Rothenberg Inc., CA, USA) modified to permit continuous flushing of the chamber with a humidified mixture of 95% N2 and 5% CO2. The oxygen content in the chamber was kept at less than 0.1% in all hypoxia experiments. Thus, the in vitro experiments were in fact performed in severe hypoxia, which might be considered anoxia. Confluent monolayers were trypsinized, and 2 x 106 cells were seeded into 100 mm plates, which were incubated under normal cell culture conditions overnight. Thereafter, the plates were assigned to two groups, control and hypoxia, and exposed to these conditions for 24 h. After the incubation, cells were trypsinized, pelleted, and washed twice using a large volume of PBS, pelleted and kept at -80˚C until DNA isolation. Each experiment was performed in duplicate and repeated at least three times. In vivo evaluation in human tumor xenograft model. All in vivo procedures were performed according to the guidelines and recommendations of the Canadian Council of Animal Care (CCAC), under the supervision of the local Animal Care Committee. Cell lines HCT116, DLD-1, SW480, SW620, Dks-8, WM115 and WM239 were used for parallel cell culture and subcutaneous implantation of 1–2 x 106 cells (in 100 µl of 0.1% BSA in PBS) into the flank of immune deficient RAG1- mice.34 Tumor growth was monitored using Vernier calipers and volume determined by the standard formula (length x width2 x 0.5);23 tissues were harvested once tumor volume reached approximately 100–300 mm3. Mice were euthanized by CO2 asphyxia followed by cervical dislocation. For hypoxia detection, mice were injected i.p. with 150 mg/kg body weight Hypoxyprobe-TM (Chemicon International Inc., Temecula, CA) one hour prior to euthanasia. Tumors were removed and cut into pieces, which were then snap-frozen in cryomatrix and stored at -80˚C for later sectioning, or snap-frozen in liquid nitrogen and stored at -80˚C for subsequent DNA isolation. Cell sorting. The tumor cells most proximal to perfused vessels were labeled by a cardiac injection of 0.25 mg/mouse Hoechst 33342 dye in 200 µl volume, five minutes before euthanizing the xenograft bearing mice.35,36 This procedure yields highly fluorescent cells immediately surrounding the vasculature and low fluorescence intensity in more distal hypoxic areas. After enzymatic disaggregation of each tumor into a single-cell suspension, samples were sorted by fluorescence-activated cell sorting (FACS), and the cells displaying the 5% highest and 5% lowest Hoechst fluorescence intensities were collected and stained for hypoxia and 5-mC levels as described below. Immunofluorescence double staining for 5-mC and hypoxic adducts. We modified the protocol described by Benchaib et al.37 to immunostain sections of cancer cell xenografts, and isolated suspensions of cells. Six µm cryosections of HCT116, DLD1, WM115 and WM239 tumors were air-dried at room temperature (RT) for 15 minutes, then fixed in 4% paraformaldehyde for 12 minutes at RT

Epigenetics

2007; Vol. 2 Issue 2


and treated with 1% Triton X (in PBS) for 20 min. To ensure that methylated DNA was accessible to antibodies, the DNA was denatured in situ with 4 N HCl at RT for 15 min. After denaturation, the samples were blocked with 0.4% BSA overnight at 4˚C, and then incubated with mouse anti-5-mC primary antibody (Calbiochem, 1:1000) for three hours at 37˚C. This was followed by anti-mouse Cy3 secondary antibody incubation (1:200) (Sigma-Aldrich) for one hour at RT. Anti-HypoxyprobeTM-1 antibody (Chemicon International Corp.) was directly labeled with Alexa Fluor 488 using Zenon Alexa Fluor 488 Mouse IgG1 Labeling Kit, according to the manufacturer’s instructions (Invitrogen Canada Inc., Burlington, ON) and used at 1:25 dilution. Cell suspensions from four sorted populations (as described above, Section 2.4) were used to generate cytospin slides, which were stained using the same protocol as tissue cryosections. Images were captured using a Leica Opti-Tech epifluorescence microscope equipped with appropriate excitation and emission filters. Images were merged using Adobe Photoshop 5.0 (Adobe). Control slides received PBS in place of primary antibody. The immunostained cells were classified into three groups: strong, weak, and no staining for 5-mC, and an average of at least 25 cells from each experiment were scored in a blinded fashion. High performance liquid chromatography. Genomic 5-mC was quantified in the cultured cells or xenografts by reversed –phase high performance liquid chromatography (HPLC). Genomic DNA was isolated from tumor pieces, or cultured cells using DNeasy Tissue Kit from QIAGEN Inc. (ON, Canada). 50 µg of RNA-free DNA was digested with 100 µg/mL DNase I (Sigma, DN25) at 37˚C for 15 h and then with 50 µg/mL nuclease P1 (Sigma) at 37˚C for 6 h and samples were separated on a reverse phase column (Supelcosil LC-18 DB) as described previously [38], monitoring absorbance at 260 nm. Peak assignments were confirmed using the nucleotide monophosphate standards (Sigma and USB Corporation; Fig. 2A). 5-mC content was expressed as a percentage of the total cytosine pool, using peak areas after correction for extinction coefficients. For each sample, at least two separate aliquots of DNA were digested and chromatographed, and each analysis was replicated at least three times. Statistical analysis. Quantitative data are presented as means of three or more independent measurements ± SD. Differences between means within experiments, were evaluated by unpaired t-test with a two-tailed level of significance, or one-way ANOVA, followed by Tukey’s LSD; in all cases, a significance level of 95% (p < 0.05) was used.

RESULTS Changes in 5-mC content after severe hypoxic exposure in vitro. Genomic 5-mC levels were quantified by HPLC in the hypoxiaexposed cells, control cells, and pieces of xenografts. Although different cell lines had different 5-mC content, all cell lines originally derived from primary tumors (HCT116, DLD-1, SW480, WM115) and normal human fibroblasts, showed a reduction in their 5-mC levels when they were exposed to hypoxia (Fig. 2B). In case of SW480 and WM115 cells and fibroblasts, this reduction was statistically different from control cultures (p < 0.05). In fact, in fibroblasts it was found to be greater than the reduction in 5-mC observed for all tested cancer cell lines (40% reduction), suggesting that the cancer cells which have already experienced some hypomethylation as a www.landesbioscience.com

Figure 2. Quantification of 5-methylcytosine content as a percentage of the total cytosine pool in cultured cells. (A) Representative chromatogram obtained by reversed-phase HPLC separation of genomic DNA digests from HCT116 cells. (B) Measurement of changes in 5-mC content in different primary lesion cancer cell lines (colorectal carcinoma: HCT116, DLD-1, SW480; melanoma: WM115) and in normal human dermal fibroblasts, under different conditions (Control, Xenograft, Low O2). * significant difference from respective control (p < 0.05).

consequence of their tumor condition histories, may contain fewer DNA sites sensitive to hypoxia-induced hypomethylation. Changes in 5-mC content after xenograft formation in vivo. Genomic 5-mC was quantified by HPLC in the xenografts and compared to levels in the same cancer cell lines in vivo. All cancer cell lines examined showed a reduction in their 5-mC content when grown as xenografts compared to control culture (Fig. 2B); this decrease was statistically significant (p < 0.05) for all cell lines except HCT116. The 5-mC content reduction observed in xenografts was not significantly different from the reductions observed in vitro using the hypoxia chamber. Changes in 5-mC content of metastatic cell lines. In order to examine the possible alterations in epigenetic events with tumor progression, we examined pairs of human cancer cell lines derived from primary and metastatic lesions from the same patients (CRC SW480 and SW620,29,30 and melanomas WM115 and WM239).31-33 While growth of cells in vivo as xenografts led to 8 and 9% decreases in 5-mC levels in WM115 and SW480, respectively, the same process had no significant effect on 5-mC levels in WM239 and SW620 cells (Fig. 3A). Also, in contrast to the marked decrease of 5-mC in the primary cancer cell lines WM115 and SW480 exposed to 24 hours severe hypoxia, there was no difference in 5-mC levels in

Epigenetics

121


the metastasis derived cells (WM239 and SW620) between hypoxiaexposed and control cultures (Fig. 3B). Nevertheless, 5-mC levels in the metastasis-derived cells were only 60% and 78% of their respective primary cell lines, WM115 and SW480. Dual immunofluorescence staining of 5-mC and hypoxic adducts.We used dual immunofluorescence staining to examine the patterns of 5-mC levels in relation to hypoxic areas of the tumor xenografts in both tissue cryosections, and in cells sorted from hypoxic and well-oxygenated regions. As demonstrated in Figure 4, protein adducts are only detected in patches that correspond to regions of tissue hypoxia. In contrast, 5-mC is detectable in all regions of tumor sections, but shows lower intensity of staining in the hypoxic areas of the tumors. The Hoechst 33342 injection procedure described above produced a highly fluorescent population of cells immediately surrounding the vasculature (Bright cells), and weakly fluorescent population of cells in more distal tumor regions (Dim cells). The cells displaying the 5% highest and 5% lowest Hoechst fluorescence intensities were collected from tumors, digested into single cell suspensions and sorted by FACS (Fig. 5) and then co-stained for hypoxic adducts and 5-mC levels. The overlay of the two images confirmed an inverse relationship between global 5-mC presence and cells positive for hypoxic adducts (Fig. 5 A-D), reflecting the findings of in situ stained tumor pieces and of 5-mC levels in cells exposed to hypoxia in vitro. Quantification of staining patterns in these sorted cells demonstrated a significantly higher proportion of cells with no detectable 5-mC in the Dim sort compared to the Bright sort, and a significantly higher proportion of cells with strong 5-mC staining in the Bright sort compared to the Dim sort (p < 0.05; Fig. 5E). Figure 3. Measurement of 5-methylcytosine content as a percentage of the total cytosine pool in primary cancer cell lines, melanoma WM115 and colorectal SW480 and their metastatic counterparts WM239 and SW620 under different conditions (A) Control and xenograft and (B) Control and Low O2. * significant difference from respective control (p < 0.05).

DISCUSSION Global DNA hypomethylation throughout the cell genome is a common characteristic of cancer cell DNA, with the degree of hypomethylation increasing from normal through benign, primary and secondary malignancy in some types of cancer.39-41 Global DNA hypomethylation in many cases affects the genome at large and often results in a decreased overall content of 5-mC. It is often not clear exactly which sequences are subject to global hypomethylation, nor are the mechanisms that trigger this event well understood. Nutritional status is known to influence global DNA hypomethylation. For example, dietary methyl deficiency causes DNA hypomethylation, which in case of short-term exposures, can be mostly reversed by physiological intakes of a methyl donor such as folic acid.42-44 However, the influence of nutrition on gene-specific DNA methylation in humans is not as well studied. In cancer, the tumor microenvironment promotes increased mutagenesis as well as selecting for cells which are resistant to these physiological stresses, thereby contributing Figure 4. Immunofluorescence staining of cryosections from colorectal carcinoma (Dks-8) and melanoma (WM115) xenografts. (A and D) Immunofluorescence staining of 5-methylcytosine; (B and E) Detection of hypoxic areas (green) by anti-HypoxyprobeTM-1 immunostaining in the same field; (C and F) overlay of the two channels demonstrating an inverse relationship between global 5-methylcytosine presence and hypoxic regions of the tumor; asterisks mark regions of hypoxia within each section.

122

Epigenetics



Figure 5. Hypoxia and 5-methylcytosine in cells sorted from xenografted tumors. I) Histogram of UV Hoechst 33342 intensity FACS analysis of unsorted cell suspension of a Dks-8 xenograft. II) Histogram of sorted “Dim” cells, and III) Histogram of sorted “Bright” cells; A-D) Immunostaining of sorted “Dim” cells for DNA (DAPI) to indicate nuclei (A); hypoxic adducts (green) by anti-HypoxyprobeTM-1 (B); immunofluorescence staining of 5-mC (red) (C); overlay of hypoxic adduct and 5-mC images (D). Arrows indicate hypoxic cells with reduced 5-mC staining. E) Quantification of 5-mC staining (mean ± SD) of counts performed on immunostained “Dim” and “Bright” cells for 4 different sorted tumors. *significantly higher number of cells with undetectable 5-mC in the Hoechst “Dim” population compared to the “Bright” population (p< 0.05). § significantly higher number of cells with high levels of 5-mC in “Bright” population vs “Dim” (p < 0.05).

to tumor progression.23,27 Using Big Blue® rat derived cell lines;45,46 we previously showed that mutation rate in the λ-shuttle cII transgene increased in cells recovered from xenografts, compared to their in vitro counterparts.47 Furthermore, numerous studies have demonstrated that hypoxia is an especially important mutagenic tumor microenvironmental condition.23,47,48 Nevertheless, the influence of tumor microenvironment in general and hypoxia in particular, on DNA methylation patterns is still unknown. This study is therefore the first to examine the linkage between abnormal DNA methylation in solid tumors and under ischemialike conditions in vitro. DNA 5-mC levels were examined in human CRC and melanoma cell lines exposed to hypoxic environments, either in cell culture, or in regions of tumors xenografted to immune deficient RAG-1 null mice. Both growth as xenografts and in vitro exposure to low oxygen led to decreases of 5-mC levels in the primary cancer cells. In agreement with these observations hypoxia also led to decrease of 5-mC level of fibroblasts in culture indicating that this effect is general and independent of other abnormalities of cancer cells. www.landesbioscience.com

Chawla et al.49 tested the effect of hypoxia on hepatic DNA methylation in rat. They maintained young rats in 10% oxygen atmosphere for ten days and measured DNA methylation in terms of incorporation of [3H] methyl of S-adenosyl-L-methionine (SAM). The methylases mediated incorporation of radioactivity (a measure of CpG specificity) in hypoxic DNA was about nine fold more than in the control liver DNA. This finding is in agreement with our results, although the conditions employed in their study for in vivo hypoxia can be considered chronic, mild, hypoxia. Our experiments were acute (24 h) and performed under severe hypoxic conditions (less than 0.1% oxygen) in order to mimic the highly hypoxic conditions present in solid tumors.50-54 There is evidence that substantial yet highly heterogeneous changes in tumor blood flow and pO2 can occur with different periodicities and magnitudes in different tumor types and in different regions of the same tumor. The severity of hypoxia (the absolute level of oxygen) within tumors is highly variable, ranging from normoxic to completely anoxic.55 It has been demonstrated that impairment in the production of SAM, the major methylating agent, also occurs during hypoxia.49,56-58 SAM is an abundant methyl donor in human metabolism, involved in more than 100 methyltransfer reactions, including DNA methylation, and is synthesized from methionine and ATP by the enzyme methionine adenosyltransferase (MAT).59,60 It has been reported that in rat hepatocytes cultured under low oxygen levels, MAT is inactivated in an NO synthase-dependent fashion.57 While this is at best an indirect approach to evaluate the levels of DNA methylation under hypoxia, these findings suggest that a reduction in SAM levels in hypoxic conditions may be at least partly responsible for the significantly lower levels of 5-mC observed in our report. Thus, limited availability of SAM during hypoxia may be a cause of the increase in unmethylated sites on DNA.

Epigenetics

123


Hypoxia exposure increased expression of IGF2 in xenografted applying anticancer therapies involving tumor hypoxia (e.g., antitumors and in cultured cell lines.61 This report and the findings angiogenesis) in cancers. of Cui et al.,20 who demonstrated that the IGF2 gene undergoes References hypomethylation and loss of imprinting in colorectal cancer, may be 1. Ehrlich M, Gama-Sosa MA, Huang LH, Midgett RM, Kuo KC, McCune RA, Gehrke C. Amount and distribution of 5-methylcytosine in human DNA from different types of tislinked through our findings that hypoxia induces DNA hypomethsues of cells. Nucleic Acids Res 1982; 10: 2709-21. ylation. Furthermore, hypomethylation was found to be responsible 2. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev 2002; 16: 6-21. for the increased expression of the cdk inhibitor p16INK4a in 3. Smith SS, Kaplan BE, Sowers LC, Newman EM. Mechanism of human methyl-directed DNA methyltransferase and the fidelity of cytosine methylation. Proc Natl Acad Sci USA primary and metastatic breast carcinoma,17 and hypoxia has also 1992; 89: 4744-8. 62 been reported to stimulate p16 expression. Our results suggest that 4. Xie S, Wang Z, Okano M, Nogami M, Li Y, He WW, Okumura K, Li E. Cloning, expreshypoxia-induced hypomethylation of p16 gene may play a major role sion and chromosome locations of the human DNMT3 gene family. Gene 1999; 236:87-95. 5. Fremont M, Siegmann M, Gaulis S, Matthies R, Hess D, Jost JP. Demethylation of DNA by in cancer progression. purified chick embryo 5-methylcytosine-DNA glycosylase requires both protein and RNA. In our experiments, the primary cancer cell lines show less sensiNucleic Acids Res 1997; 25:2375-80. tivity to low oxygen (in regard to methylation changes) compared to 6. Bhattacharya SK, Ramchandani S, Cervoni N, Szyf M. A mammalian protein with specific demethylase activity for mCpG DNA. Nature 1999; 397:568-9. the normal human fibroblasts. This observation, together with our 7. Mummaneni P, Walker KA, Bishop PL, Turker MS. Epigenetic gene inactivation induced results that hypoxia had no effect on the global methylation status by a cis-acting methylation center. J Biol Chem 1995; 270:788-92. of metastatic cancer cells, suggests that there is a limited pool of 8. Wilkins JF. Genomic imprinting and methylation: epigenetic canalization and conflict. Trends Genet 2005; 21:356-65. CpG sites sensitive to hypoxia induced demethylation. If those sites 9. Turker MS, Bestor TH. Formation of methylation patterns in the mammalian genome. have been already affected during cancer progression (such as in the Mutat Res 1997; 386:103-5. case of primary versus metastatic cancer cells), they would not show 10. Robertson KD, Keyomarsi K, Gonzales FA, Velicescu M, Jones PA. Differential mRNA expression of the human DNA methyltransferases (DNMTs) 1, 3a and 3b during the G(0)/ sensitivity to hypoxia. Dual immunofluorescence staining of frozen G(1) to S phase transition in normal and tumor cells. Nucleic Acids Res 2000; 28:2108sections or suspension sorted cells from xenografts also supported our 13. hypothesis that there is an inverse relationship between hypoxia and 11. Eads CA, Danenberg KD, Kawakami K, Saltz LB, Danenberg PV, Laird PW. CpG island hypermethylation in human colorectal tumors is not associated with DNA methyltransfer5-mC content. ase overexpression. Cancer Res 1999; 59:2302-6. In contrast to the marked decrease of 5-mC content in primary 12. De Marzo AM, Marchi VL, Yang ES, Veeraswamy R, Lin X, Nelson WG. Abnormal regulation of DNA methyltransferase expression during colorectal carcinogenesis. Cancer Res cancer cells after xenograft formation, this condition did not affect 1999; 59:3855-60. the 5-mC content of metastatic cells, perhaps for reasons outlined 13. Issa JP, Vertino PM, Wu J, Sazawal S, Celano P, Nelkin BD, Hamilton SR, Baylin SB. above. Increased cytosine DNA-methyltransferase activity during colon cancer progression. J Natl Cancer Inst 1993; 85:1235-40. Interestingly, there are some cells in our fluorescence-sorted 14. Herman JG, Baylin SB. Gene silencing in cancer in association with promoter hypermethsamples that show robust hypoxic adduct staining despite being ylation. N Engl J Med 2003; 349:2042-54. obtained from the Bright population of tumor cell suspensions. This 15. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet 2002; 3:415-28. is a reflection of the fact that Hoechst staining is a very rapid event, 16. Dunn B K. Hypomethylation: One Side of a Larger Picture. Ann N Y Acad Sci 2003; while formation and destruction of hypoxic adducts takes consider983:28-42. ably longer. Since vascular flow in solid tumors is heterogeneous 17. Van Zee KJ, Calvano JE and Bisogna M. Hypomethylation and increased gene expression of p16INK4a in primary and metastatic breast carcinoma as compared to normal breast tissue, with rapid fluctuations,50-54 re-establishment of flow to a previously Oncogene 1998; 16:2723-27 . ischemic region just prior to injection of Hoechst would have the 18. Yanagawa N, Tamura G, Honda T, Endoh M, Nishizuka S, and Motoyama T. Demethylation of the synuclein gamma gene CpG island in primary gastric cancers and gastric cancer cell apparently paradoxical effect of labeling hypoxic cells brightly with lines. Clin Cancer Res 2004; 10:2447-51. this perfusion marker. Conversely, stasis in a blood vessel during 19. Laner T, Schulz WA, Engers R, Muller M, Florl AR. Hypomethylation of the XIST gene the few minutes of Hoechst circulation could produce the opposite promoter in prostate cancer. Oncol Res 2005;15:257-64. effect- a cell free of hypoxic adduct staining from the dimly labeled 20. 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; (poorly perfused) population. However, these events are rare, and the 62:6442–6. quantification of 5-mC staining in cells from either sorted popula- 21. Wong JJ, Hawkins NJ, Ward RL. Colorectal cancer: a model for epigenetic tumorigenesis. Gut 2007; 56:140-8. tion, combined with the in-situ co-staining of 5-mC and hypoxic 22. Yu JL, Rak JW, Carmeliet P, Nagy A, Kerbel RS, Coomber BL. Heterogeneous vascular adducts reveals that there is a shift in 5-mC content of cells between dependence of tumor cell populations. Am J Pathol 2001; 158:1325-34. hypoxic and normoxic compartments of xenografted tumors. 23. Shahrzad S, Quayle L, Stone C, Plumb C, Shirasawa S, Rak JW, Coomber BL. Ischemiainduced K-ras mutations in human colorectal cancer cells: role of microenvironmental In conclusion, our studies have uncovered evidence that one of the regulation of MSH2 expression. Cancer Res 2005; 65:8134-41. possible mechanisms for regulation of DNA methylation in cancer is 24. Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 1971; 285:1182-6. the ischemic/hypoxic condition found in solid tumors. Since ours is 25. Carmeliet P, Jain R. Angiogenesis in cancer and other diseases. Nature 2000; 407:249-57. the first study to demonstrate that 5-mC levels alter with hypoxia, we 26. Jain RK. Molecular regulation of vessel maturation. Nat Med 2003; 9:685-93. have no information at present on the kinetics of such changes. It is 27. Papp-Szabo E, Josephy PD, Coomber BL. Microenvironmental influences on mutagenesis in mammary epithelial cells. Int J Cancer 2005; 116:679-85. possible that hypomethylation occurs via an active (and presumably 28. Shirasawa S, Furuse M, Yokoyama N, Sasazuki T. Altered growth of human colon cancer cell lines disrupted at activated Ki-ras, Science 1993; 260:85-8. rapid) process or through lack of maintenance methylation (which would presumably be much slower) or some combination of these. 29. Fogh J, Fogh JM and Orfeo T. One hundred and twenty-seven cultured human tumor cell lines producing tumors in nude mice. J Natl Cancer Inst 1977; 59:221-26. This aspect of the epigenetic response is currently under investigation 30. Leibovitz A, Stinson JC, McCombs WB III, McCoy CE, Mazur KC and Mabry ND. Classification of human colorectal adenocarcinoma cell lines. Cancer Res 1976; 36:4562-69. in our laboratory. Extensive demethylation of DNA during tumor progression could be a source of the continually generated cellular 31. Herlyn M, Thurin J, Balaban G, Bennicelli JL, Herlyn D, Elder DE, Bondi E, Guerry D, Nowell P, Clark WH, Koprowski H. Characteristics of cultured human melanocytes isolated diversity associated with cancer. Modulation of these tissue environments from different stages of tumor progression. Cancer Res 1985; 45: 5670-76. may provide a promising approach to controlling epigenetic events 32. Herlyn M, Kath R, Williams N, Valyi-Nagy I, Rodeck U. Growth regulatory factors for normal, premalignant, and malignant human cells in vitro. Adv Cancer Res 1990; 54:213-34. in progression of solid tumors and carries important implications for 124

Epigenetics


Tumor Hypoxia Leads to Changes in DNA Methylation 33. Rodeck U, Herlyn M, Menssen HD, Furlanetto RW, Koprowski H. Metastatic but not primary melanoma cell lines grow in vitro independently of exogenous growth factions. Int J Cancer 1987; 40:687-90. 34. Chen J, Shinkai Y, Young F, Alt FW. Probing immune functions in RAG-deficient mice, Curr Opin Immunol 1994; 6:313-19. 35. Yu JL, Rak JW, Carmeliet P, Nagy A, Kerbel RS, Coomber BL. Heterogeneous vascular dependence of tumor cell populations. Am J Pathol 2001; 158:1325-34. 36. Yu JL, Rak JW, Coomber BL, Hicklin DJ, Kerbel RS. Effect of p53 status on tumor response to antiangiogenic therapy. Science 2002; 295:1526-8. 37. Benchaib M, Ajina M, Lornage J, Niveleau A, Durand P and Guerin JF. Quantitation by image analysis of global DNA methylation in human spermatozoa and its prognostic value in in vitro fertilization: a preliminary study. Fertil Steril 2003; 80:947–53. 38. Ramsahoye BH. Measurement of genome-wide DNA cytosine-5 methylation by reversedphase high-pressure liquid chromatography. Methods Mol Biol 2002; 200:17-27. 39. Gama-Sosa MA, Slagel VA, Trewyn RW, Oxenhandler R, Kuo KC, Gehrke CW, Ehrlich M. The 5-methylcytosine content of DNA from human tumors. Nucleic Acids Res 1983; 11:6883-94. 40. Goelz SE, Vogelstein B, Hamilton SR, Feinberg AP. Hypomethylation of DNA from benign and malignant human colon neoplasms. Science 1985; 228:187-90. 41. Kim YI, Giuliano A, Hatch KD, Schneider A, Nour MA, Dallal GE, Selhub J, Mason JB. Global DNA hypomethylation increases progressively in cervical dysplasia and carcinoma. Cancer 1994; 74:893-9. 42. Pufulete M, Al-Ghnaniem R, Khushal A, Appleby P, Harris N, Gout S, Emery PW, Sanders TA. Effect of folic acid supplementation on genomic DNA methylation in patients with colorectal adenoma. Gut 2005; 54:579-81. 43. Christman JK, Sheikhnejad G, Dizik M, Abileah S, Wainfan E. Reversibility of changes in nucleic acid methylation and gene expression induced in rat liver by severe dietary methyl deficiency. Carcinogenesis 1993; 14:551-7. 44. Pogribny IP, Ross SA, Wise C, Pogribna M, Jones EA, Tryndyak VP, James SJ, Dragan YP, Poirier LA. Irreversible global DNA hypomethylation as a key step in hepatocarcinogenesis induced by dietary methyl deficiency. Mutat Res 2006; 593:80-7. 45. McDiarmid HM, Douglas GR, Coomber BL, Josephy PD. Epithelial and fibroblast cell lines cultured from the transgenic BigBlue rat: an in vitro mutagenesis assay. Mutat Res 2001; 497:39-47. 46. McDiarmid HM, Douglas GR, Coomber BL, Josephy PD. 2-Amino-1-methyl-6phenylimidazo[4,5-b]pyridine (PhIP)-induced mutagenesis in cultured Big Blue rat mammary epithelial and fibroblast cells. Environ Mol Mutagen 2002; 39:245-53. 47. Papp-Szab—, E, Josephy, PD, Coomber, BL. Microenvironmental influences on mutagenesis in mammary epithelial cells. Int J Cancer 2005; 116:679-85. 48. Yuan J, Narayanan L, Rockwell S, Glazer PM. Diminished DNA repair and elevated mutagenesis in mammalian cells exposed to hypoxia and low pH, Cancer Res 2000; 60:4372-76. 49. Chawla RK, Watson WH, Jones DP. Effect of hypoxia on hepatic DNA methylation and tRNA methyltransferase in rat: similarities to effects of methyl-deficient diets. J Cell Biochem 1996; 61:72-80. 50. Kimura H, Braun RD, Ong ET, Hsu R, Secomb TW, Papahadjopoulos D, Hong K, Dewhirst MW. Fluctuations in red cell flux in tumor microvessels can lead to transient hypoxia and reoxygenation in tumor parenchyma. Cancer Res 1996; 56:5522–8. 51. Dewhirst MW, Braun RD, Lanzen JL. Temporal changes in pO2 of R3230AC tumors in Fischer-344 rats. Int J Radiat Oncol Biol Phys 1998; 42:723–6. 52. Braun RD, Lanzen JL, Dewhirst MW. Fourier analysis of fluctuations of oxygen tension and blood flow in R3230Ac tumors and muscle in rats. Am J Physiol 1999; 277:H551-68. 53. Brurberg KG, Graff BA, Rofstad EK. Temporal heterogeneity in oxygen tension in human melanoma xenografts. Br J Cancer 2003; 89:350–6. 54. Brurberg KG, Graff BA, Olsen DR, Rofstad EK. Tumor-line specific pO2 fluctuations in human melanoma xenografts. Int J Radiat Oncol Biol Phys 2004; 58:403–9. 55. Chaplin DJ, Durand RE, Olive PL. Acute hypoxia in tumors: implications for modifiers of radiation effects. Int J Radiat Oncol Biol Phys 1986; 12:1279–82. 56. Perez-Mato I, Castro C, Ruiz FA, Corrales FJ, Mato JM. Methionine adenosyltransferase Snitrosylation is regulated by the basic and acidic amino acids surrounding the target thiol. J Biol Chem. 1999; 274:17075-9. 57. Avila MA, Carretero MV, Rodriguez EN, Mato JM. Regulation by hypoxia of methionine adenosyltransferase activity and gene expression in rat hepatocytes. Gastroenterology 1998; 114:403-7. 58. Blaise S, Alberto JM, Nedelec E, Ayav A, Pourie G, Bronowicki JP, Gueant JL, Daval JL. Mild neonatal hypoxia exacerbates the effects of vitamin-deficient diet on homocysteine metabolism in rats. Pediatr Res 2005; 57:777-82. 59. Kotb M, Geller AM. Methionine adenosyltransferase: structure and function. Pharmacol Ther 1993; 59:125-43. 60. Mato JM, Alvarez L, Ortiz P, Pajares MA. S-adenosylmethionine synthesis: molecular mechanisms and clinical implications. Pharmacol Ther 1997; 73:265-80. 61. Hedborg F, Ulleras E, Grimelius L, Wassberg E, Maxwell PH, Hero B, Berthold F, Schilling F, Harms D, Sandstedt B, Franklin G. Evidence for hypoxia-induced neuronal-to-chromaffin metaplasia in neuroblastoma. FASEB J 2003; 17:598-609. 62. Zygmunt A, Tedesco VC, Udho E, Krucher NA. Hypoxia stimulates p16 expression and association with cdk4. Exp Cell Res 2002; 278:53-60.

www.landesbioscience.com

Epigenetics

125