Prognostic Potential of DNA Methylation and Transcript Levels of … · the EPAS1 transcript, as...

Chromatin, Gene, and RNA Regulation

Prognostic Potential of DNA Methylation and TranscriptLevels of HIF1A and EPAS1 in Colorectal Cancer

Agnieszka Anna Rawłuszko-Wieczorek1, Karolina Horbacka2, Piotr Krokowicz2, Matthew Misztal1, andPaweł Piotr Jagodzi�nski1

AbstractHypoxic conditions during the formation of colorectal cancer may support the development of more aggressive

tumors. Hypoxia-inducible factor (HIF) is a heterodimeric complex, composed of oxygen-induced HIFa andconstitutively expressed HIFb subunits, which mediates the primary transcriptional response to hypoxic stress.Among HIFa isoforms, HIF1a (HIF1A) and endothelial PAS domain–containing protein 1 (EPAS1) are able torobustly activate hypoxia-responsive gene signatures. Although posttranslational regulation of HIFa subunits iswell described, less is known about their transcriptional regulation. Here, molecular analysis determined thatEPAS1 mRNA was significantly reduced in primary colonic adenocarcinoma specimens compared with histo-pathologically nonneoplastic tissue from 120 patients. In contrast, no difference in HIF1A mRNA levels wasobserved between cancerous and noncancerous tissue. Bisulfite DNA sequencing and high-resolution meltinganalysis identified significant DNA hypermethylation in the EPAS1 regulatory region from cancerous tissuecompared with nonneoplastic tissue. Importantly, multivariate Cox regression analysis revealed a high HR forpatients with cancer with low EPAS1 transcript levels (HR, 4.91; 95% confidence interval, CI, 0.42–56.15; P ¼0.047) and hypermethylated EPAS1 DNA (HR, 33.94; 95% CI, 2.84–405.95; P ¼ 0.0054). Treatment with aDNA methyltransferase inhibitor, 5-Aza-20-deoxycytidine (5-aza-dC/Decitabine), upregulated EPAS1 expressionin hypoxic colorectal cancer cells that were associated with DNA demethylation of the EPAS1 regulatory region. Insummary, EPAS1 is transcriptionally regulated by DNA methylation in colorectal cancer.

Implications:DNAmethylation andmRNA status of EPAS1 have novel prognostic potential for colorectal cancer.Mol Cancer Res; 12(8); 1112–27. �2014 AACR.

IntroductionImmense proliferation of tumor cells and their inadequate

perfusion results in hypoxia, which is a hallmark of manysolid tumors, including colorectal cancer (1). Mechanismsby which tumor cells alter their expression profile to adjust tolow oxygen tension involve hypoxia-inducible factor (HIF;ref. 1). HIF is a heterodimeric transcription factor assembledfrom a and b subunits. It recognizes the hypoxia responseelement (HRE) and promotes expression of many genesinvolved in glucose metabolism, angiogenesis, or metastasis(1). The b subunit of HIF is constitutively expressed,whereas HIFa is mainly controlled by a posttranslational

mechanism (1, 2). In normoxic conditions, HIFa is hydrox-ylated at specific residues, which results in proteasomaldegradation (2). There are three a isoforms referred to as:HIF1a, HIF2a (officially designated as endothelial PASdomain–containing protein 1—EPAS1) and HIF3a, whichare encoded by the HIF1A, EPAS1, and HIF3A genes,respectively (2). Among them, HIF1a or EPAS1 may bindtogether with b and other coactivators to HRE and activateHIF-dependent gene transcription (2). Many articlesdescribe aberrant HIF1a or EPAS1 protein levels and theirassociation with colorectal cancer prognosis (3–10). How-ever, there are only few articles about the same issues on theHIF1A and EPAS1 mRNA level (8, 11, 12). Moreover,relatively little is known about transcriptional regulation ofHIF1A and EPAS1. It should be noted that both of thempossess a CpG island in their promoter region. DNAmethylation within the CpG island associates the genetranscriptional repression, and aberrant DNA methylationpatterns are observed during colorectal tumorigenesis (13).To date, only one article indicates DNA methylation ofHIF1A in colorectal cancer (14), and there is no scientificreports about this type of epigenetic regulation of EPAS1expression. Therefore, we aimed to examine DNA methyl-ation and mRNA levels of the HIF1A and EPAS1 genes in

Authors' Affiliations: Departments of 1Biochemistry and Molecular Biol-ogy and 2General and Colorectal Surgery, Pozna�n University of MedicalSciences, Poland

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Corresponding Author: Agnieszka Anna Rawłuszko-Wieczorek, Pozna�nUniversity of Medical Sciences, S´wiecickiego 6 Street, Pozna�n 60-781,Poland. Phone: 486-1854-6516; Fax: 48618546510; E-mail:[email protected]

doi: 10.1158/1541-7786.MCR-14-0054

�2014 American Association for Cancer Research.

MolecularCancer

Research

Mol Cancer Res; 12(8) August 20141112

on October 3, 2020. © 2014 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst May 13, 2014; DOI: 10.1158/1541-7786.MCR-14-0054

http://mcr.aacrjournals.org/

primary cancerous and histopathologically unchanged colo-rectal tissues from the same 120 patients. Moreover, weevaluated the impact of EPAS1 DNA methylation andtranscript level in colorectal cancer tissue with respect topatient survival. We also assessed the effect of 5-Aza-20-deoxycytidine (5-dAzaC), an inhibitor ofDNAmethyltrans-ferases, onDNAmethylation level of theEPAS1 gene and onthe EPAS1 transcript, as well as protein level in HCT116and DLD-1 colorectal cancer cells under hypoxic andnormoxic conditions.

Materials and MethodsAntibodies and reagentsRabbit polyclonal (Rp) anti-HIF1a (NB100-449) and

anti-EPAS1 (NB100-122) Abs were provided by NovusBiologicals. Rp anti-GAPDH Ab (FL-335) and goat anti-rabbit horseradish peroxidase (HRP)–conjugated Ab wereprovided by Santa Cruz Biotechnology. 5-dAzaC was pur-chased from Sigma-Aldrich Co.

Patient materialPrimary colonic adenocarcinoma tissues were collected

between June 2009 and March 2013 from 120 patients whounderwent radical surgical resection of the colon at theDepartment of General and Colorectal Surgery, Pozna�nUniversity of Medical Sciences, Poland (Table 1). Thehistopathologically unchanged colonic mucosa located atleast 10 to 20 cm away from the cancerous lesions wasobtained from the same patients. One set of samples wasimmediately snap-frozen in liquid nitrogen and stored at�80�C until DNA/RNA isolation. The other set of sampleswas directed for histopathologic examination. Histopatho-logic classification was performed by an experienced pathol-ogist. No patients received preoperative chemo- or radiother-apy. An informed consent was obtained from all participatingindividuals. The procedures of the study were approved bythe Local Ethical Committee of Pozna�n University of Med-ical Sciences.

Measurement of overall and disease-free survivalFollow-up data were available for 80 patients, who were

observed from August 11, 2009, until death or October 15,2013, whichever came first. Nine patients were excludedfrom further analysis because they did not fulfill criteria givenbelow. Disease-free survival (DFS) is defined as the timeelapsed from surgery to the first occurrence of any of thefollowing events: recurrence or distant metastasis of colo-rectal cancer, development of a second noncolorectal malig-nancy. In overall survival (OS) analysis, deaths from anycause without clinical documentation of cancer related eventwere excluded from the study.

Cell cultureDLD-1 colon cancer cells were obtained from the ATCC,

andHCT116 cells were kindly provided by the Departmentof Experimental and Clinical Radiobiology, Maria Skło-dowska-Curie Cancer Center, Institute of Oncology

Branch, Gliwice, Poland. These cells were cultured inDMEM GibcoBRL containing 10% heat-inactivated FBSand 2 mmol/L glutamine. To determine the effect of 5-dAzaC on DNA methylation, transcript and protein level oftheHIF1A and EPAS1 genes, theHCT116 andDLD-1 cellswere cultured for 24 hours in DMEM GibcoBRL supple-mented with 10% FBS from Sigma-Aldrich Co. Cells werethen cultured under normoxic or hypoxic (1% O2) condi-tions, either in the absence or in the presence of 5-dAzaC. 5-dAzaC was at concentrations of 1.00 and 5.00 mmol/L for 6,24, and 48-hour time frames. Hypoxic conditions wereachieved using a MCO-18M multi-gas cell culture incuba-tor, Sanyo, modified to permit flushing the chamber with ahumidified mixture of 5% CO2, 94% N2. These cells wereused for DNA and RNA isolation, quantitative real-timePCR (qRT-PCR), Western blotting, and high-resolutionmelting (HRM) analysis.

DNA isolation and bisulfite modificationGenomic DNA from tissues of patients with colorectal

cancer and cell lines were isolated using the DNA Mam-malian Genomic Purification Kit purchased from Sigma-Aldrich Co. A total of 500 ng of genomic DNA wassubjected to bisulfite conversion of cytosine to uracil,according to the EZ DNA Methylation Kit procedure fromZymo Research Corporation. The position of the CpGislands and binding sites of transcription factors located intheHIF1A and EPAS1 promoter were determined by onlineprograms (15–17).

Table 1. Demographic and histopathologicclassification of patients with colorectal cancer

FeaturesNumber ofpatients

Total number of patients 120Gender (female/male) 54/66Mean (�SD) age at radical surgicalresection of colon (y)

67.94 � 12.45

Colorectal cancer localizationProximal colon (cecum totransverse)

45

Distal colon (splenic flexureto sigmoid)

22

Rectum 53Histologic gradeG1 7G2 77G3 36

TNM classificationI 17IIA 47IIC 6IIIA 3IIIB 34IIIC 13

DNA Methylation and mRNA Level of HIF1A and EPAS1 in Colorectal Cancer

www.aacrjournals.org Mol Cancer Res; 12(8) August 2014 1113




DNA methylation evaluation by bisulfite sequencingTheDNA fragments containing CpGdinucleotides locat-

ed in the promoter region of the HIF1A and EPAS1 geneswere amplified from the bisulfite-modified DNA by theprimer pairs (Supplementary Table S1A) complementary tothe bisulfite DNA-modified sequences in 5 patients. Clin-icopathologic parameters for these patients are given inSupplementary Table S1B. PCR amplification was per-formed by FastStart Taq DNA Polymerase from RocheDiagnostic GmbH. The PCR products were purified usingthe Agarose Gel DNA Extraction Kit, Roche DiagnosticGmbHwith subsequent cloning into pGEM-T Easy VectorSystem I, Promega and transformation into TOPO10 E. colistrain cells. The plasmid DNA isolated from five positivebacterial clones was used for commercial sequencing of thecloned fragment of DNA. The results of bisulfite sequencingwere assessed and presented using BiQ analyzer software andBisulfite sequencing Data Presentation and Compilation(BDPC) web server, respectively (18, 19).

DNA methylation assessment by HRM analysisMethylation level of DNA fragments located within the

CpG island of theHIF1A and EPAS1 genes was determinedby RT-PCR amplification of bisulfite-treated DNA, fol-lowed by HRM profile analysis by Light Cycler480 Real-Time PCR System, Roche Diagnostics GmbH. For PCRamplification, 1 mL of the bisulfite-treated DNA frompatients, HCT116, DLD-1 cells, or standards, and primers(Supplementary Table S1A) was added to 19 mL of 5 X HotFIREPol EvaGreen HRM Mix, Solis BioDyne Co. Stan-dardized solutions of DNA methylation percentage wereprepared by mixing methylated and nonmethylated bisul-fite-treated DNA from Human Methylated/Non-methylat-ed DNA Set, Zymo Research Corp. in different ratios. Todetermine the percentage of methylation, the HRM profilesof patients DNA PCR products were compared with HRMprofiles of standard DNA PCR products (20, 21). HRMmethylation analysis was performed using Light Cycler480Gene Scanning software, Roche Diagnostics GmbH. EachPCR amplification andHRMprofile analysis was performedin triplicate. The HRM results were compared with thoseobtained from bisulfite sequencing for analyzed genes inreconstituted samples. A similar pattern of DNA methyla-tion was observed between these two methods. The meth-ylation for each patient was presented as a percentage ofmethylation in amplified fragments located in the CpGisland of HIF1A and EPAS1. Because low level of methyl-ationmay not demonstrate significant biologic effects andwewere not able to quantify all the CpG dinucleotides withinthe analyzed CpG island, the percentage results were dividedinto three groups: 0% to 1% methylation, 1% to 10%methylation, and 10% to 100% methylation for statisticalanalysis (22–25).

Reverse transcription and quantitative real-time PCRanalysisTotal RNA from tissues of patients with colorectal cancer

and cell lines were isolated according to the method of

Chomczy�nski and Sacchi (26). The RNA samples werequantified and reverse-transcribed into cDNA. qRT-PCRwas carried out in the Light Cycler480 Real-Time PCRSystem, Roche Diagnostics GmbH using SYBR Green I asthe detection dye. The target cDNA was quantified by therelative quantification method using a calibrator for theprimary tissues. The calibrator was prepared as a cDNAmix from all of the patients' samples, and successive dilutionswere used to create a standard curve as described in RelativeQuantification Manual Roche Diagnostics GmbH. Foramplification, 1 mL of (total 20 mL) cDNA solution wasadded to 9 mL of IQ SYBR Green Supermix, Bio-RadLaboratories Inc. with primers (Supplementary TableS1A). To prevent amplification of sequences from genomicDNA contamination, primers and/or amplicons weredesigned at exon/exon boundaries and covered all gene splicevariants. The quantity of HIF1A and EPAS1 transcripts ineach sample was standardized by the geometric mean of twointernal controls: porphobilinogen deaminase (PBGD) andhuman mitochondrial ribosomal protein L19 (hMRPL19;Supplementary Table S1A). The selection of internal controlgenes was made as previously (27). The HIF1A and EPAS1transcript levels in the patients' tissues were expressed asmultiplicity of the cDNA concentrations in the calibrator. InHCT116 and DLD-1 cells, transcript levels were presentedas multiplicity of the respective controls.

Western blotting analysisHCT116 and DLD-1 cells were treated with lysis RIPA

buffer, and proteins were resuspended in the sample bufferand separated on 10% Tris-glycine gel using SDS-PAGE.Gel proteins were transferred to a nitrocellulose membrane,which was blocked with 5%milk in Tris/HCl saline/Tweenbuffer. Immunodetection of bands was performed with Rpanti-HIF1a and -EPAS1 Ab, followed by incubation withgoat anti-rabbit HRP-conjugated Ab. To ensure equalprotein loading of the lanes, the membrane was strippedand incubatedwith Rp anti-GAPDHAb (FL-335), followedby incubation with goat anti-rabbit HRP-conjugated Ab.The bands were revealed using SuperSignal West FemtoChemiluminescent Substrate, Thermo Fisher Scientific, andBiospectrum Imaging System500,UVPLtd. The amount ofanalyzed proteins was presented as the protein-to-GAPDHband optical density ratio. For HCT116 and DLD-1 cellscultured in the absence of 5-dAzaC, the ratio of EPAS1 toGAPDH was assumed to be 1.

Statistical analysisThe normality of the observed patient data distribution

was assessed by the Shapiro–Wilk test, and the Mann–Whitney U test was used to compare the median values.The c2 test was used to examine the significance in DNAmethylation. Fisher exact probability test was used for datathat do not fulfilled criteria for Cochran theorem. Toevaluate the association between different ranges of DNAmethylation (0%–1% methylation, 1%–10% methylation,and 10%–100% methylation) and the ratios of canceroustissue EPAS1mRNA level to histopathologically unchanged

Rawłuszko-Wieczorek et al.

Mol Cancer Res; 12(8) August 2014 Molecular Cancer Research1114




EPAS1 mRNA level, the nonparametric Kruskal–Wallis testwas used. Survival curves were plotted using the Kaplan–Meier method, and survival differences were achieved usingthe log-rank test. Multivariate Cox proportional hazardmodel was used to estimate the adjusted HR. Data groupsfor cell lines were assessed by ANOVA to evaluate if therewere significant differences (P < 0.05) between the groups.For all experimental groups that fulfilled the initial criteria,individual comparisons were performed, by post hoc Tukeytest with the assumption of two-tailed distribution. Statis-tically significant results were indicated by P < 0.05. Statis-tical analysis was performed with STATISTICA 10.0software.

ResultsDNA hypermethylation of EPAS1 regulatory region isassociated with a decrease in EPAS1 mRNA level inprimary cancerous tissue compared withhistopathologically unchanged tissue from patients withcolorectal cancer, whereas there is neither DNAmethylation nor transcript changes of HIF1ATo compare the HIF1A and EPAS1 transcript and DNA

methylation levels in the HIF1A and EPAS1 promoterregions in cancerous and histopathologically unchangedtissues from 120 patients with colorectal cancer, we usedRQ-PCR and bisulfite DNA sequencing followed by HRManalysis, respectively. We found significantly lower levels ofthe EPAS1 transcript (P¼ 0.000011) in primary cancerousthan in the histopathologically unchanged tissues in patientswith colorectal cancer (Fig. 1A). Moreover, we observedsignificantly lower levels of the EPAS1 transcript in cancer-ous tissues in different age groups, genders, colorectal cancerlocalizations, histologic grades, and tumor–node–metastasis(TNM) stages (Supplementary Table S2). There was nosignificant difference in the level of the HIF1A transcriptbetween primary cancerous and histopathologicallyunchanged tissues in 120 patients with colorectal cancer(P¼ 0.87; Fig. 1A). We also undetected DNA methylationwithin theHIF1A promoter in the analyzed regions (chr14:62 161 804-62 162 333 and chr14: 62 162 250-62 163 074using bisufite sequencing; chr14: 62 161 655-62 161 825and chr14: 62 162 301- 62 162 427 using HRManalysis; Fig. 1B and C). Moreover, we did not observeDNA methylation in the regulatory region of the EPAS1gene in cancerous and histopathologically unchanged tissuesin regions chr2: 46 524 336-46 524 767 and chr2: 46 524751-46 525 189 using bisulfite sequencing; chr2: 46 524636-46 524 769 and chr2: 46 524 969-46 525 075 usingHRM analysis (Fig. 1D and E). However, study of EPAS1gene regulatory region chr2: 46 526 521-46 527 161revealed significant DNA hypermethylation in canceroustissue compared with histopathologically unchanged tissue,using bisulfite sequencing in 5 patients (Fig. 1F). In keepingwith the bisulfite sequencing data, we observed significantlyhigher DNA methylation within EPAS1 regulatory regionchr2: 46 526 762-46 526 905 in cancerous compared withhistopathologically unchanged tissue from 120 patients with

colorectal cancer (P < 0.00001; Fig. 1F and Table 2).Patients were also stratified by gender, age, histologic grades,and TNM stages for DNA methylation analysis. Weobserved higher DNA methylation within the analyzedregion of the EPAS1CpG island in primary cancerous tissuefor a majority of subgroups and no distinctive subgroupbiased DNAmethylation (Supplementary Table S3). More-over, we observed that an increase in DNAmethylation levelof EPAS1 in region chr2: 46 526 762-46 526 905 correlatedto a decrease in the ratio of cancerous to histopathologicallyunchanged tissue EPAS1 mRNA level (P¼ 0.0036; Fig. 2).

DNA hypermethylation and low mRNA level of theEPAS1 gene are prognostic factors for patients' OS withcolorectal cancerTo investigate the effect of transcript and DNA methyl-

ation level of EPAS1 on patients' survival, we carried outretrospective clinical analysis of 71 patients. The mediansurvival was 36 months (range, 9–51 months). On the basisof RQ-PCR data, the EPAS1 mRNA level in histopatho-logically unchanged and cancerous tissue was subdividedinto three groups: low, intermediate, and high EPAS1transcript levels. Univariate analysis of OS revealed thatpatients with low mRNA expression level of EPAS1 inhistopathologically unchanged tissue had a significantincrease in risk of death compared with patients with anintermediate and/or high expression level (Fig. 3A). Thisrelated to survival: 33 months in low EPAS1 mRNAsubgroup versus 36 in intermediate and high EPAS1mRNAsubgroups (Fig. 3A). Moreover, the Kaplan–Meier analysisrevealed benefit of a high EPAS1 transcript level inhistopathologically unchanged tissue of a 7-month medianincrease in survival compared with the intermediate and 14-month increase compared with the low EPAS1 mRNAsubgroup in patients not treated with postoperative chemo-therapy (Fig. 3B). Analysis of cancerous tissue disclosed lackof impact of EPAS1 mRNA level on OS (Fig. 3A and B).Furthermore, there was no evidence of impact of EPAS1mRNA level on DFS in both the histopathologicallyunchanged and cancerous tissues (Supplementary Fig.S1). Impact of DNA methylation of EPAS1 regulatoryregion was done by comparison of two groups: absent DNAmethylation and present DNA methylation in the EPAS1gene regulatory region. Of note, 1% to 10% and 10% to100% subgroups were merged into one because of limitednumber of patients in 10% to 100% subgroup (n¼ 4). Wefound that patients with DNA hypermethylation of EPAS1in cancerous tissue compared with histopathologicallyunchanged had shorter OS rate compared with patientswith no changes in DNA methylation status (Fig. 3C).Although, result was statistically insignificant, it suggeststhat there may have been a reduction in the risk of deathfor patients with the hypomethylated EPAS1 gene regu-latory region. The analysis in the group of patients withoutpostoperative chemotherapy and analysis of impact ofEPAS1 DNA methylation on DFS did not reveal statis-tically significant data (Fig. 3C and D; Supplementary Fig.S1). However, multivariate Cox regression analysis with






5.0

4.5

4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

A

B

C

7

6

5

4

3

2

1

0

Log

[HIF

1A m

RN

A]

Log

[EPA

S1

mR

NA

]

P = 0.87 P = 0.000011

Median25%–75%Min-MaxHistopathologically unchanged tissue

Histopathologically unchanged tissue


Temperature (ºC)

Primary cancerous tissue Histopathologically unchanged tissue Primary cancerous tissue

Cancerous tissue

HIF1A.1

HIF1A.1

chr14: 62 161 804

chr14: 62 161 655 chr14: 62 161 825

chr14: 62 162 333CpG dinucleotides

CpG dinucleotides

P1P2P3P4P5

P1P2P3P4P5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

Normalized melting curves


100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Rel

ativ

e si

gnal

(%

)


Cancerous tissue

chr14: 62 162 250

chr14: 62 162 301 chr14: 62 162 427

HIF1A.2

HIF1A.2

chr14: 62 163 074CpG dinucleotides

CpG dinucleotides

P1P2P3P4P5

P1P2P3P4P5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

74 74.5 75 75.5 76 76.5 77 77.5 78.5 79.579 80 80.5 81 81.578

100%

50%

10%Cancerous tissue


Cancerous tissue

1% 0%

Temperature (ºC)

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Rel

ativ

e si

gnal

(%

)

75 75.5 76 76.5 77 77.5 78.5 79.579 80 80.5 81 81.5 82 82.5 83 83.578

100%

50%

10%

1% 0%

Figure 1. DNAmethylation of promoter region and transcript levels ofHIF1A andEPAS1 in primary cancerous and histopathologically unchanged tissues frompatients with colorectal cancer. The cancerous and histopathologically unchanged tissues from 120 patients with colorectal cancer were used for RNA andDNA isolation. A, total RNAwas reverse-transcribed, and cDNAswere investigated by RQ-PCR relative quantification analysis. The HIF1A and EPAS1mRNAlevels were corrected by the geometric mean of PBGD and hMRPL19 cDNA levels. The amounts of mRNA were expressed as the decimal logarithm ofmultiples of these cDNA copies in the calibrator. The P value was evaluated by the Mann–Whitney U test. (Continued on the following page.)








Cancerous tissue

CpG dinucleotides

D

E

chr2: 46 524 336

chr2: 46 524 751 chr2: 46 524 189EPAS1.2

chr2: 46 525 075chr2: 46 524 969EPAS1.2

chr2: 46 524 767

chr2: 46 524 769chr2: 46 524 636EPAS1.1

EPAS1.1

CpG dinucleotides

CpG dinucleotides

CpG dinucleotides

P1P2P3P4P5

P1P2P3P4P5

P1P2P3P4P5

P1P2P3P4P5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

Cancerous tissue1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44



100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Temperature (ºC)

Rel

ativ

e si

ngna

l (%

)

7473.57372.5 74.5 75 75.5 76 76.5 77 77.5


Cancerous tissue


Cancerous tissue

100%

50%

10%

1%0%

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Temperature (ºC)

Rel

ativ

e si

ngna

l (%

)

74.5 75 75.5 76 76.5 77 77.5 78 78.5 79 79.5 80 80.5 81 81.5

100%

50%

10%

1%

0%

Figure 1. (Continued. ) B to F, primary cancerous and histopathologically unchanged tissues from the same patients with colorectal cancer (P1–P5;Supplementary Table S1B) were used for genomic DNA isolation followed by bisulfite conversion of cytosine to uracil. The HIF1A regions containing49 CpG dinucleotides (chr14: 62 161 804-62 162 333; B) and 70 CpG dinucleotides (ch14: 61 162 250-62 163 074; C) as well EPAS1 regionscontaining 49 CpG dinucleotides (chr2: 46 524 336-46 524 767; D), 44 CpG dinucleotides (chr2: 46 524 751-46 525 189; E), and 37 CpG dinucleotides(chr2: 46 526 521-46 527 161; F) were then amplified by a pair of primers complementary to the bisulfite DNA-modified sequence (Supplementary Table S1A).(Continued on the following page.)






respect to age, gender, and postoperative chemotherapystatus revealed that mRNA level and DNA methylationare both independent prognostic factors for patient'ssurvival (Table 3). Low EPAS1 mRNA level in histopath-ologically unchanged tissue and DNA hypermethylationin cancerous tissue compared with histopathologicallyunchanged have significant HR equal to 4.91 and33.94, respectively (Table 3). Neither EPAS1 mRNAstatus nor DNA methylation were associated with DFSin multivariate analysis (Table 3).

EPAS1 gene regulatory region is hypermethylated inHCT116 colorectal cancer cells in normoxic and hypoxicconditionsTo evaluate DNAmethylation and expression level of the

HIF1A andEPAS1 genes inHCT116 andDLD-1 colorectalcancer cells, we performed HRM analysis, RQ-PCR, andWestern blotting. We observed no DNAmethylation of theHIF1A promoter region in the analyzed regions using HRManalysis under hypoxic and normoxic conditions inHCT116 andDLD-1 cells (Fig. 4A).Moreover, we detectedDNAhypomethylation in the EPAS1CpG island inDLD-1cells (Fig. 4A). Nonetheless, we detected a high level ofDNA methylation in HCT116 in the chr2: 46 526 762-

46 526 905 and no DNA methylation in chr2: 46 524636-46 524 769 and chr2: 46 524 969-46 525 075 (Fig.4A). We revealed a lower level of the HIF1A and EPAS1transcript in HCT116 cells compared with DLD-1 cells inboth hypoxic and normoxic conditions (Fig. 4B). TheHIF1A transcript level was not induced upon hypoxia inboth cell lines (Fig. 4B). However, we observed a signif-icant induction of the EPAS1 transcript level upon hyp-oxia in DLD-1 cells, with no changes in HCT116 cellsunder the same conditions (Fig. 4B). In both analyzed celllines, hypoxic conditions induced HIF1a and EPAS1protein level (Fig. 4B).

5-dAzaC induced DNA demethylation of EPAS1 generegulatory region, EPAS1 transcript, and proteincontents in HCT116 cells; did not affect EPAS1 DNAmethylation or expression level in DLD-1 cells underhypoxic conditionsTo assess the effect of 5-dAzaC on DNAmethylation and

the EPAS1 gene expression level, we used HRM analysis,RQ-PCR, and Western blotting. We observed no effect of5-dAzaC treatment on DNA methylation status in theanalyzed region of the EPAS1 promoter in DLD-1 cellsunder hypoxic and normoxic conditions (Fig. 4C). On the

Histopathologically unchanged tissueEPAS1.3

Cancerous tissue

CpG dinucleotides

Fchr2: 46 526 521

chr2: 46 526 762 chr2: 46 526 905EPAS1.3

chr2: 46 527 161

CpG dinucleotides

P1P2P3P4P5

P1P2P3P4

P5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37


100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Temperature (ºC)

Rel

ativ

e si

ngna

l (%

)

76 76.5 77 77.5 78 78.5 79 79.5 80 80.5 81 81.5


Cancerous tissue

100%

50%

10%

1%

0%

Figure 1. (Continued. ) The PCR products were purifiedwith subsequent cloning into a plasmid vector. PlasmidDNA isolated from five positive bacterial cloneswas used for commercial sequencing. The results of bisulfite sequencing were assessed and presented using BiQ analyzer software and BDPC web server(18, 19). Black and gray boxes, methylated and unmethylated CpG dinucleotide, respectively. Black rectangles, regions amplified in HRM analysis byspecific primers (Supplementary Table S1A). Bottom panels in B to F, HRM profiles of standard and example of patient DNA (patient P1 from bisulfitesequencing). Methylation percentage of three DNA fragments within the HIF1A and EPAS1 CpG island was determined by RT-PCR amplification ofbisulfite-treated standard and patient DNA, followed by comparison of their HRM profiles. DNA standards were prepared by mixing different ratios ofmethylated and nonmethylated bisulfite-treated DNA. HRM methylation analysis was performed using Light Cycler480 Gene Scanning software, RocheDiagnostics GmbH. Each PCR amplification and HRM profile analysis was performed in triplicate.






contrary, using HRM analysis, we noticed significant DNAdemethylation in chr2: 46 526 762-46 526 905 region of theEPAS1CpG island inHCT116 cells cultured for 48 hours inthe presence of 5.00 mmol/L 5-dAzaC under both hypoxicand normoxic conditions (Fig. 4C). The changes in DNAmethylation level were accompanied by 5-dAzaC–inducedexpression of EPAS1 in HCT116 cells. We observed that 5-dAzaC resulted in a progressive increase in the EPAS1transcript level in HCT116 cells. For HCT116, we foundapproximately a 2.11- and 2.88-fold increase in the EPAS1transcript level at 48 hours of incubation under normoxicand hypoxic conditions, respectively (Fig. 4D). Despite theabsence of DNA methylation in DLD-1, we noticed pro-gressive increase in mRNA level in DLD-1 cells undernormoxic conditions (Fig. 4D). Alterations in the EPAS1transcript level in HCT116 cells were associated with anincreased EPAS1 protein level in hypoxic conditions (Fig.4D). Densitometric analysis of Western blotting bandsindicated an approximately 2.31-fold increase in EPAS1protein level in HCT116 cells, incubated with 5.00 mmol/L5-dAzaC for 48 hours as compared with the respectivecontrols under hypoxic conditions (Fig. 4D). These changeswere not observed for the HCT116 cells under normoxicconditions (Fig. 4D). Incubation of DLD-1 cells with 5-dAzaC at various concentrations for different time periodsdid not significantly increase EPAS1 protein content undereither hypoxic or normoxic conditions (Fig. 4D).

DiscussionHIFa initiates adaptive responses that maintain proper

metabolism and pH homeostasis, thereby reinforcing tumorgrowth and metastasis (1). Many immunohistochemicalstudies reported a correlation between the HIF1a or EPAS1protein and patient survival, response to chemotherapy,expression of oncogenes and genes involved in angiogenesisin breast, head and neck, cervix, gastric, hepatocellular, andglioma cancers (28–32). The level of HIF1a and EPAS1protein was also determined in colorectal cancer, but theresults are inconclusive. HIF1a was correlated with poor

patient prognosis by three independent studies, but threeconsecutive articles showed lack of such association incolorectal cancer (3–9). Moreover, Yoshimura and collea-gues demonstrated strong positive immunohistochemicalstaining of EPAS1 in advanced colorectal cancer comparedwith low-grade tumors (6). Nonetheless, a study conductedby two other research teams described the opposite results(9, 10). The HIF1A and EPAS1 transcript level was notanalyzed extensively. In the esophageal squamous cell car-cinoma, pancreatic, gastric, cervical and colon cancers,

Table 2. DNA methylation level of the EPAS1 gene regulatory region in primary cancerous tissue andhistopathologically unchanged tissue sample from patients with colorectal cancer

EPAS10% to 1%methylation

1% to 10%methylation

10% to 100%methylation Pa

Histopathologically unchanged tissue 113 5 2 <0.00001Primary cancerous tissue 77 29 14

NOTE: The primary cancerous and histopathologically unchanged tissue samples from the same patients were used for genomic DNAisolation, followed by bisulfite conversion of cytosine to uracil. The DNA fragments of the CpG island were then amplified pairs ofprimers complementary to the bisulfite DNA-modified sequence (Supplementary Table S1A). To determine the percentage ofmethylation, the HRM profiles of the patients' DNA PCR products were compared with HRM profiles of the prepared standard PCRproducts (Fig. 1F). DNA methylation of the EPAS1 regulatory region for each patient was calculated as a percentage of methylation inamplified fragment.ac2 test.

1.6

1.5

1.4

1.3

1.2

1.1

1.0

0.9

0.8

0.7

0.6

0.50%–1% 1%–10%

Methylation

10%–100% Median25%–75%Min–Max

P = 0.036

Rat

io o

f can

cero

us E

PAS

1 m

RN

A le

vel t

ohi

stop

atho

logi

cally

unc

hang

ed ti

ssue

EPA

S1

mR

NA

leve

l

Figure 2. Ratio of cancerous to histopathologically unchanged tissueEPAS1mRNA level in three rangesofEPAS1DNAmethylation status: 0%to 1%, 1% to 10%, and 10% to 100%. Methylation percentage of threeDNA fragmentswithin theEPAS1CpG island (Supplementary Table S1A)was determined by RT-PCR amplification of bisulfite-treated standardand patient DNA, followed by comparison of their HRM profiles. DNAmethylation for each patient was calculated as a percentage ofmethylation in amplified fragments. The samples were divided into threegroups for statistical analysis: 0% to 1% methylation, 1% to 10%methylation, and 10% to 100% methylation (27). To evaluate thestatistically significant difference in the ratio of cancerous EPAS1 mRNAlevel to histopathologically unchanged tissue EPAS1 mRNA levelbetween the three DNA methylation ranges (0%–1% methylation,1%–10% methylation, and 10%–100% methylation), the nonparametricKruskal–Wallis test was used.






Histopathologically unchanged tissueA

B

C D

Event Censored

Cancerous tissueEvent Censored

Histopathologically unchanged tissueEvent Censored

Cancerous tissueEvent Censored

Event CensoredEvent Censored

1.0

0.8

0.6

0.4

0.2

0.00 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55 60

Time (months) Time (months)

Sur

viva

l pro

babi

lity

1.0

0.8

0.6

0.4

0.2

0.0

Sur

viva

l pro

babi

lity

P = 0.022 P = 0.32 Low EPAS1 mRNAlnt EPAS1 mRNAHigh EPAS1 mRNA

0 5 10 15 20 25 30 35 40 45 50 55Time (months)

1.0

0.8

0.6

0.4

0.2

0.0

Sur

viva

l pro

babi

lity

P = 0.31

0 5 10 15 20 25 30 35 40 45 50 55Time (months)

1.0

0.8

0.6

0.4

0.2

0.0

Sur

viva

l pro

babi

lity

P = 0.42

0 5 10 15 20 25 30 35 40 45 50 55Time (months)

1.0

0.8

0.6

0.4

0.2

0.0

Sur

viva

l pro

babi

lity

P = 0.098

Low EPAS1 mRNAlnt EPAS1 mRNAHigh EPAS1 mRNA

Low EPAS1 mRNA 15 4 36

39 2 36

17 1 33

lnt EPAS1 mRNA

High EPAS1 mRNA

N Events Median OS

Low EPAS1 mRNA 9 3 24

22 1 31

11 1 38

lnt EPAS1 mRNA

High EPAS1 mRNA

N Events Median OSLow EPAS1 mRNA 11 1 35

21 4 21

10 0 36.5

lnt EPAS1 mRNA

High EPAS1 mRNA

N Events Median OS

Low EPAS1 mRNA 18

36

17 0 37

5 34.5

2 39

lnt EPAS1 mRNA

High EPAS1 mRNA

N Events Median OS

5 10 15 20 25 30 35 40 45 50 55Time (months)

1.0

0.8

0.6

0.4

0.2

0.0

Sur

viva

l pro

babi

lity

P = 0.87DNA methylationof EPAS1 present

DNA methylationof EPAS1 absent

DNA methylation ofEPAS1 present

DNA methylation ofEPAS1 absent

DNA methylationof EPAS1 present

DNA methylationof EPAS1 absent

N

47 25 3 333 36

24 4 33 17 2 31

Events Median OS

DNA methylation ofEPAS1 present

DNA methylation ofEPAS1 absent

N Events Median OS






HIF1A mRNA levels were increased in cancerous comparedwith noncancerous tissue (8, 11, 12, 33, 34).However, otherstudies reported a constant HIF1A mRNA level in tumorcells and suggested mainly posttranslational regulation ofHIF1A expression (35–37), which is consistent with ourobservations. We have not detected significant HIF1Atranscript changes between cancerous and histopathologi-cally unchanged tissues isolated from 120 patients. Only onepublication described a correlation of elevated EPAS1mRNA level in blood plasma with poor outcome for thepatients with colorectal cancer (38). We found reducedEPAS1 mRNA level in primary cancerous comparedwith histopathologically unchanged tissues. Moreover, weobserved a higher risk of death of patients with colorectalcancer with low EPAS1 mRNA level in histopathologicallyunchanged tissues. In keeping with our results, reducedEPAS1 transcript level was observed in a non–small celllung carcinoma relative to normal tissue (39).Data suggest that the importance ofHIF1a and EPAS1 in

response to hypoxia may differ among tumor types anddifferent stages of tumor progression. Moreover, recentarticles indicate that these two HIFa subunits exhibitdistinct roles in hypoxic conditions (31). HIF1a and EPAS1may regulate the expression of many of the same targetgenes, but each has unique responsive genes as well (40). Themechanism responsible for the activation of HIF1a- or

EPAS1-specific target genes seems to be the interaction ofN-transactivation domain of HIF protein with differentcoactivators (41, 42). In renal cell carcinoma, tumor pro-gression and metastasis were predominantly dependent onEPAS1 (43). However, an in vitro study in colorectal cancercells revealed the induction number of genes associated withglycolysis and angiogenesis byHIF1a and tumor-suppressorgenes such as cyclin G2 or angiopoietin-like 4 by EPAS1 (10).Xenograft studies support the hypothesis of a protectivefunction of EPAS1 in colorectal cancer. Silencing of EPAS1expression in a mouse model is associated with a moreintensive development of colorectal cancer (10). In aKRAS-driven non–small cell lung carcinoma mouse model,the loss of EPAS1 expression resulted in increased tumorgrowth and progression (39). Moreover, siRNA knockdownof EPAS1 reduced apoptosis in glioblastoma cells (44). Ourobservation illustrated an increased risk of death in patientswith low EPAS1 mRNA level in histopathologicallyunchanged tissues, which supports the idea of EPAS1 ashaving a tumor-protective role. Obviously, large multicenterstudies on various patient populations with extended follow-up need to confirm these results.For the first time, we examined a relationship between

epigenetic silencing of EPAS1 and clinical prognosis ofpatients with colorectal cancer. We found that a reducedEPAS1 mRNA level was associated with DNA

Figure 3. The Kaplan–Meier survival analysis among patients with colorectal cancer according to the EPAS1 mRNA level and DNA methylation of the EPAS1promoter. A and B, patients were subdivided into three groups: low, intermediate, and high EPAS1 transcript levels in histopathologically unchangedandcancerous tissue.CandD,patientswere subdivided into twogroups: absentDNAmethylation andpresentDNAmethylation in theEPAS1gene regulatoryregion in cancerous tissue compared with histopathologically unchanged. The analysis was performed with the division into groups: A and C: all patientsincluded in the analysis (n ¼ 71); B and D: patients without postoperative chemotherapy (n ¼ 42). N, number of patients.

Table 3. Multivariate analysis of DNA methylation and transcript level of EPAS1 in patients with colorectalcancer

OS DFS

Variable HR (95% CI) P HR (95% CI) P

EPAS1 mRNAHigh 1 1Intermediate 0.65 (0.049–8.64) 0.21 0.39 (0.083–1.86) 0.09Low 4.91 (0.42–56.16) 0.047 1.24 (0.28–5.59) 0.24

EPAS1 DNA methylationAbsent 1 1Present 33.94 (2.84–405.95) 0.0054 1.62 (0.32–8.36) 0.56

GenderFemale 1 1Male 27.89 (1.56–497.24) 0.022 1.98 (0.38–10.40) 0.42

Age (y)Below 60 1 1Above 60 0.86 (0.079–9.23) 0.89 1.42 (0.29–6.82) 0.66

TherapyNo 1 1Yes 0.19 (0.027–1.33) 0.095 1.39 (0.44–4.37) 0.57






hypermethylation in cancerous tissues, and EPAS1 DNAmethylation was a prognostic factor for patient survival inmultivariate Cox regression analysis. An investigated frag-

ment of the EPAS1 CpG island is the region of a transcrip-tion factor binding and different types of epigenetic mod-ifications such as histone acetylation (17). Hence, it may be


A





74 74.5 75 75.5 76 76.5 77 77.5

100%

DLD-1 HypoxiaDLD-1 Normoxia

HCT116 HypoxiaHCT116 Normoxia








HCT116 Hypoxia

HCT116 Normoxia

50%

10%

0%1%

100%

50%

10%

0%1%

100%

50%

10%

0%

1%

100%

50%

10%

0%1%

100%

50%

10%0%

1%

Temperature (°C)7473.57372.5 74.5 75 75.5 76 76.5 77 77.5

Temperature (°C)78 78.5 79 79.5 80 80.5 81 81.5

74.5 75 75.5 76 76.5 77 77.5Temperature (°C)

78 78.5 79 79.5 80 80.5 81 81.5

76 76.5 77 77.5Temperature (°C)78 78.5 79 79.5 80 80.5 81 81.5

75 75.5 76 76.5 77 77.5Temperature (°C)

78 78.5 79 79.5 80 80.5 81 81.5 82 82.5 83 83.5

Rel

ativ

e si

gn

al (

%)

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Rel

ativ

e si

gn

al (

%)

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Rel

ativ

e si

gn

al (

%)

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Rel

ativ

e si

gn

al (

%)

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Rel

ativ

e si

gn

al (

%)

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Figure 4. DNAmethylation, expression level of theHIF1A andEPAS1genes aswell effect of 5-dAzaConDNAmethylation andexpression ofEPAS1 inHCT116andDLD-1 colorectal cancer cells. HCT116andDLD-1 cellswere culturedunder normoxic or hypoxic (1%O2) conditions for 48hours.Cellswere then used forDNA isolation followed by bisulfite modification, RNA and protein isolation. A, methylation percentage of DNA fragments within the HIF1A and EPAS1 CpGisland (Supplementary Table S1A) in HCT116 and DLD-1 cells under hypoxic and normoxic conditions was determined by RT-PCR amplification of bisulfite-treated standard and cell line DNA, followed by comparison of their HRM profiles. (Continued on the following page.)






1.00

B

C

0.75

0.50

0.25

0.00

DLD-1 HCT116

DLD-1 HCT116

HIF

1A m

RN

A le

vel

of th

e re

spec

tive

cont

rols

EPA

S1

mR

NA

leve

lof

the

resp

ectiv

e co

ntro

ls

DLD-1 HCT116

HIF1α (92 kDa)

GAPDH (36 kDa)

1 45.9 1.13 78.2

N H N H

1 4.3 1.13 2.53

N H N H

DLD-1 HCT116

EPAS1 (97 kDa)

GAPDH (36 kDa)

1.75

1.50

1.25

1.00

0.75

0.50

0.25

0.00

EPAS1.3chr2: 46 526 762 chr2: 46 526 905

100,000

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0,000

Temperature (ºC)

Rel

ativ

e si

ngna

l (%

)

76 76.5 77 77.5 78 78.5 79 79.5 80 80.5 81 81.5

100%

HCT 116 Hypoxia control

HCT 116 Normoxia control

HCT 116 Normoxia 5-dAzaC

HCT 116 Hypoxia 5-dAzaC

50%

1% 10%

0%

DLD-1 Hypoxia 5-dAzaC

DLD-1 Hypoxia control

DLD-1 Normoxia 5-dAzaC

DLD-1 Normoxia control


Figure 4. (Continued. ) B, cells were cultured in DMEMeither in hypoxic (1%O2; H) or normoxic (N) conditions for 48 hours. After incubation, the cells were usedfor total RNA isolation followed by reverse transcription and protein isolation. The HIF1A and EPAS1 cDNA levels were determined by RQ-PCR relativequantification analysis. RQ-PCR results were standardized by the geometric mean of PBGD and hMRPL19 cDNA levels. HIF1A and EPAS1 cDNAlevels are expressed as a multiplicity of these cDNA copies in the cell line's calibrator. Proteins were separated by 10% SDS-PAGE, and transferred to amembrane that was then immunoblotted with specific primary and secondary Ab. The band densitometry readings were normalized to GAPDHloading control. The ratio of HIF1a or EPAS1 to GAPDH for DLD-1 in normoxic conditions was assumed to be 1. C, HCT116 and DLD-1 cells werecultured under normoxic or hypoxic (1% O2) conditions either in the absence or in the presence of 5-dAzaC at a concentration of 5.00 mmol/L for 48 hours.Cells were then used for DNA isolation followed by bisulfite modification. Methylation percentage of DNA fragment within the EPAS1 CpG island (chr2:46 526 762-46 526 905) in HCT116 and DLD-1 cells under hypoxic and normoxic conditions was determined by RT-PCR amplification of bisulfite-treatedstandard and cell line DNA, followed by comparison of their HRM profiles. (Continued on the following page.)






3

2

1

06 24 48

Time (h)6 24 48

Time (h)

HCT116D DLD-1

EPA

S1

mR

NA

leve

lsof

the

resp

ectiv

e co

ntro

ls

EPA

S1

mR

NA

leve

lsof

the

resp

ectiv

e co

ntro

ls

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Normoxia

Hypoxia

HCT116

5-dAzaC concentration (μmol/L) 0.0 1.0 5.0 0.0 1.0 5.0 0.0 1.0 5.0

0.0 1.0 5.0 0.0 1.0 5.0 0.0 1.0 5.0

1.00 0.85 1.32 1.00 1.21 1.98 1.00 1.11 2.31

1.00 1.001.21 1.01 0.850.77 1.00 0.99 0.91

1.00 1.000.96 0.87 0.870.92 1.00 0.94 1.04

1.00 1.09 1.07 1.00 1.21 1.39 1.00 1.23 1.33

Hypoxia

Normoxia

Hypoxia

Normoxia

EPAS1 (97 kDa)

EPAS1 (97 kDa)

GAPDH (36 kDa)

GAPDH (36 kDa)

5-dAzaC concentration (μmol/L)

EPAS1 (97 kDa)

EPAS1 (97 kDa)

GAPDH (36 kDa)

GAPDH (36 kDa)

6 h 24 h 48 h

6 h 24 h 48 h

DLD-1

Figure 4. (Continued. ) D, HCT116 and DLD-1 cells were cultured in DMEM for 6, 24, and 48 hours either in the absence or in the presence of 5-dAzaC at aconcentration of 1.00 or 5.00 mmol/L under hypoxic or normoxic conditions. After incubation, the cells were used for total RNA isolation andprotein isolation. Total RNA was reverse-transcribed, and EPAS1 cDNA levels were determined by RQ-PCR relative quantification analysis. EPAS1cDNA levels are expressed as a multiplicity of the respective controls. Each sample was determined in triplicate, and results are presentedas the mean� SE from three experiments ��, P < 0.01; ��, P < 0.001. The cell protein was separated by 10%SDS-PAGE, and transferred to a membranethat was then immunoblotted with specific primary and secondary Ab. The band densitometry readings were normalized to GAPDH loading control.The ratio of EPAS1 to GAPDH for control was assumed to be 1.






recognized as a putative enhancer or a promoter region. Dataabout DNA methylation of the HIF1A promoter region areambiguous. The absence of DNAmethylation of theHIF1Apromoter was observed in advanced uterine cervical carci-noma and datasets available at ENCODE project, whereasDNA hypermethylation was observed in an immaturehematopoietic cell line HMC-1 and normal colon tissuesisolated from 20 patients with colorectal cancer (14, 34, 45).In our studies, DNA methylation of the HIF1A promoterwas undetected in a group of 120 patients. The difference toprevious studies in colorectal cancermay result from the usedmethods for determining DNA methylation, and may sug-gest complexity in epigenetic regulation of HIF1A. More-over, many environmental factors like tobacco smoking,diet, and physical activity may affect DNA methylationstatus in colorectal cancer (46). Further experiments needto verify the potential impact of these factors on EPAS1DNA methylation. In addition, even though we have notdetected subgroup-biased DNA methylation in patients, wecannot exclude potential impact of gender, age, and otherclinicopathologic features in different patient populations.The main limitation of our studies is the lack of association

of DNA methylation status and mRNA level with proteinexpression of analyzed genes. However, because the amountof samples was limited, we preferred to evaluate HIF1A,EPAS1 DNA methylation, and transcript level as the prog-nostic significance of transcript, and DNA methylation hasnot been investigated previously. We assessed DNA methyl-ation and expression level of the HIF1A and EPAS1 genesunder hypoxic and normoxic conditions in DLD-1 andHCT116 colorectal cancer cell lines. We undetected changesin DNA methylation of theHIF1A promoter region and theHIF1AmRNA level under normoxic and hypoxic conditionsin both analyzed cell lines. However, an increase of HIF1aprotein in hypoxia was observed when compared with nor-moxic conditions, which indicates that the main regulator ofHIF1A expression in colorectal cancer is oxygen-dependentposttranslational modification. The EPAS1 transcript levelremained stable in normoxic and hypoxic conditions inHCT116 cells, whereas EPAS1 protein level was higher inhypoxic conditions. On the other hand, a significant increasein both the amount of EPAS1 transcript and protein inhypoxic conditions was observed in DLD-1 cells. EPAS1mRNA changes in DLD-1 were associated with DNA hypo-methylation ofEPAS1CpG island.The same region (chr2: 46526 762-46 526 905)was hypermethylated inHCT116 cells,which may explain the lack of mRNA induction underhypoxic conditions. HCT116 and DLD-1 cells were alsoincubated for different time periods with 5-dAzaC underhypoxic and normoxic conditions. 5-dAzaC may induceexpression of many genes and inhibit growth of colorectalcancer cell lines (47, 48). We observed 5-dAzaC-induced

DNA demethylation of the EPAS1 CpG island in HCT116cells, regardless of oxygen concentration. An increase in theEPAS1 transcript and protein levels under hypoxic conditionssuggests the role of EPAS1 DNA methylation in HCT116cells. In normoxic conditions, the increase of EPAS1 proteinwas not observed in both cell lines, probably due to oxygen-dependent degradation. However, an increase of EPAS1mRNA in DLD-1 cells was observed under normoxic con-ditions after 5-dAzaC treatment, despite the lack of DNAmethylation. The absence of mRNA and protein upregula-tion in DLD-1 cells upon hypoxia suggests that other factorsmust be involved in the induction of EPAS-1 gene expressionbesides DNA methylation. Moreover, discrepancies of theresults obtained from the two cell lines may be a result of theirdifferent genetic background.In conclusion, our findings present epigenetic transcrip-

tional downregulation of EPAS1 in patients with colorectalcancer and the HCT116 cell line. In addition, low EPAS1mRNA level in histopathologically unchanged tissues andDNA hypermethylation in cancerous tissues compared withthe histopathologically unchanged might be an independentprognostic factor and potentially useful for selecting patientswith a higher risk of death after resection. The clinical valueof changes in EPAS1 mRNA and DNA methylation levelsneeds to be confirmed by large longitudinal studies as well asverified in other cancer types.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: A.A. Rawłuszko-WieczorekDevelopment of methodology: A.A. Rawłuszko-WieczorekAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): A.A. Rawłuszko-Wieczorek, K. Horbacka, P. KrokowiczAnalysis and interpretation of data (e.g., statistical analysis, biostatistics, compu-tational analysis): A.A. Rawłuszko-WieczorekWriting, review, and/or revision of the manuscript: A.A. Rawłuszko-Wieczorek,M. Misztal, P.P. Jagodzi�nskiAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): K. HorbackaStudy supervision: P.P. Jagodzi�nski

AcknowledgmentsThe authors thank the Institute of Molecular Biology and Biotechnology, Adam

Mickiewicz University, for access to the MCO-18M multi-gas cell culture incubator,Sanyo.

Grant SupportThis work is supported by grant 2012/05/N/NZ5/00844 from the National

Science Center, Poland.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be herebymarked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Received January 29, 2014; revised April 7, 2014; accepted April 25, 2014;published OnlineFirst May 13, 2014.

References1. Webb JD, Coleman ML, Pugh CW. Hypoxia, hypoxia-inducible factors

(HIF), HIF hydroxylases and oxygen sensing. Cell Mol Life Sci2009;66:3539–54.

2. Jokilehto T, Jaakkola PM. The role of HIF prolyl hydroxy-lases in tumour growth. J Cell Mol Med 2010;14:758–70.






3. RasheedS,Harris AL, Tekkis PP, TurleyH, Silver A,McDonald PJ, et al.Hypoxia-inducible factor-1alpha and -2alpha are expressed in mostrectal cancers but only hypoxia-inducible factor-1alpha is associatedwith prognosis. Br J Cancer 2009;100:1666–73.

4. Schmitz KJ, Muller CI, Reis H, Alakus H, Winde G, Baba HA, et al.Combined analysis of hypoxia-inducible factor 1 alpha and metal-lothionein indicates an aggressive subtype of colorectal carcinoma. IntJ Colorectal Dis 2009;24:1287–96.

5. Rajaganeshan R, Prasad R, Guillou PJ, Scott N, Poston G, Jayne DG.Expression patterns of hypoxic markers at the invasive margin ofcolorectal cancers and liver metastases. Eur J Surg Oncol2009;35:1286–94.

6. Yoshimura H, Dhar DK, Kohno H, Kubota H, Fujii T, Ueda S, et al.Prognostic impact of hypoxia-inducible factors 1alpha and 2alpha incolorectal cancer patients: correlation with tumor angiogenesis andcyclooxygenase-2 expression. Clin Cancer Res 2004;10:8554–60.

7. Kuwai T, Kitadai Y, Tanaka S, Onogawa S, Matsutani N, Kaio E, et al.Expression of hypoxia-inducible factor-1alpha is associated withtumor vascularization in human colorectal carcinoma. Int J Cancer2003;105:176–81.

8. Furlan D, SahnaneN, Carnevali I, Cerutti R, Bertoni F, Kwee I, et al. Up-regulation of the hypoxia-inducible factor-1 transcriptional pathway incolorectal carcinomas. Hum Pathol 2008;39:1483–94.

9. Baba Y, Nosho K, Shima K, Irahara N, Chan AT, Meyerhardt JA, et al.HIF1A overexpression is associated with poor prognosis in a cohort of731 colorectal cancers. Am J Pathol 2010;176:2292–301.

10. Imamura T, Kikuchi H, Herraiz MT, Park DY, Mizukami Y, Mino-Kenduson M, et al. HIF-1alpha and HIF-2alpha have divergent rolesin colon cancer. Int J Cancer 2009;124:763–71.

11. Hoffmann AC, Mori R, Vallbohmer D, Brabender J, Klein E, Drebber U,et al. High expression of HIF1a is a predictor of clinical outcome inpatients with pancreatic ductal adenocarcinomas and correlated toPDGFA, VEGF, and bFGF. Neoplasia 2008;10:674–9.

12. Ma J, Zhang L, Ru GQ, Zhao ZS, Xu WJ. Upregulation of hypoxiainducible factor 1alpha mRNA is associated with elevated vascularendothelial growth factor expression and excessive angiogenesis andpredicts a poor prognosis in gastric carcinoma. World J Gastroenterol2007;13:1680–6.

13. Kim MS, Lee J, Sidransky D. DNA methylation markers in colorectalcancer. Cancer Metastasis Rev 2010;29:181–206.

14. Koslowski M, Luxemburger U, Tureci O, Sahin U. Tumor-associatedCpG demethylation augments hypoxia-induced effects by positiveautoregulation of HIF-1alpha. Oncogene 2011;30:876–82.

15. CpGPlot/CpGReport/Isochore. Available from: http://www.ebi.ac.uk/Tools/seqstats/emboss_cpgplot/

16. CpG Island Searcher. Available from: http://cpgislands.usc.edu/17. UCSC Genome Bioinformatics Site. Available from: http://genome.

ucsc.edu/18. Bock C, Reither S, Mikeska T, Paulsen M, Walter J, Lengauer T. BiQ

Analyzer: visualization and quality control for DNA methylation datafrom bisulfite sequencing. Bioinformatics (Oxford, England)2005;21:4067–8.

19. Rohde C, Zhang Y, Jurkowski TP, Stamerjohanns H, Reinhardt R,Jeltsch A. Bisulfite sequencing Data Presentation and Compilation(BDPC) web server–a useful tool for DNAmethylation analysis. NucleicAcids Res 2008;36:e34.

20. Wojdacz TK,DobrovicA.Methylation-sensitive high resolutionmelting(MS-HRM): a new approach for sensitive and high-throughput assess-ment of methylation. Nucleic Acids Res 2007;35:e41.

21. Rawluszko AA, Horbacka K, Krokowicz P, Jagodzinski PP. Decreasedexpression of 17-beta-hydroxysteroid dehydrogenase type 1 is asso-ciated with DNA hypermethylation in colorectal cancer located in theproximal colon. BMC Cancer 2011;11:522.

22. Hsieh CL. Dependence of transcriptional repression on CpG methyl-ation density. Mol Cell Biol 1994;14:5487–94.

23. Shao Y, Zhang W, Zhang C, Wu Q, Yang H, Zhang J, et al. High-resolution melting analysis of BLU methylation levels in gastric, colo-rectal, and pancreatic cancers. Cancer Invest 2010;28:642–8.

24. Dimitrakopoulos L, Vorkas PA, Georgoulias V, Lianidou ES. A closed-tube methylation-sensitive high resolution melting assay (MS-HRMA)

for the semi-quantitative determination of CST6 promoter methylationin clinical samples. BMC Cancer 2012;12:486.

25. Oster B, Thorsen K, Lamy P, Wojdacz TK, Hansen LL, Birkenkamp-Demtroder K, et al. Identification and validation of highly frequent CpGisland hypermethylation in colorectal adenomas and carcinomas. Int JCancer 2011;129:2855–66.

26. Chomczynski P, Sacchi N. Single-step method of RNA isolation byacid guanidinium thiocyanate-phenol-chloroform extraction. Anal Bio-chem 1987;162:156–9.

27. Rawluszko AA, Bujnicka KE, Horbacka K, Krokowicz P, Jagodzi SkiPP. Expression and DNA methylation levels of prolyl hydroxylasesPHD1, PHD2, PHD3 and asparaginyl hydroxylase FIH in colorectalcancer. BMC Cancer 2013;13:526.

28. Birner P, Schindl M, Obermair A, Plank C, Breitenecker G, OberhuberG.Overexpressionof hypoxia-inducible factor 1alpha is amarker for anunfavorable prognosis in early-stage invasive cervical cancer. CancerRes 2000;60:4693–6.

29. Birner P, Gatterbauer B, Oberhuber G, Schindl M, Rossler K, ProdingerA, et al. Expression of hypoxia-inducible factor-1 alpha in oligoden-drogliomas: its impact on prognosis and on neoangiogenesis. Cancer2001;92:165–71.

30. Birner P, Schindl M, Obermair A, Breitenecker G, Oberhuber G.Expression of hypoxia-inducible factor 1alpha in epithelial ovariantumors: its impact on prognosis and on response to chemotherapy.Clin Cancer Res 2001;7:1661–8.

31. Keith B, Johnson RS, Simon MC. HIF1alpha and HIF2alpha: siblingrivalry in hypoxic tumour growth and progression. Nat Rev 2012;12:9–22.

32. Semenza GL. Defining the role of hypoxia-inducible factor 1 in cancerbiology and therapeutics. Oncogene 2010;29:625–34.

33. Matsuyama T, Nakanishi K, Hayashi T, Yoshizumi Y, Aiko S, Sugiura Y,et al. Expression of hypoxia-inducible factor-1alpha in esophagealsquamous cell carcinoma. Cancer Sci 2005;96:176–82.

34. Luczak MW, Roszak A, Pawlik P, Kedzia H, Lianeri M, Jagodzinski PP.Increased expression of HIF-1A and its implication in the hypoxiapathway in primary advanced uterine cervical carcinoma. Oncol Rep2011;26:1259–64.

35. Minet E, Ernest I, Michel G, Roland I, Remacle J, Raes M, et al.HIF1A gene transcription is dependent on a core promotersequence encompassing activating and inhibiting sequences locat-ed upstream from the transcription initiation site and cis elementslocated within the 50UTR. Biochem Biophys Res Commun1999;261:534–40.

36. Dery MA, Michaud MD, Richard DE. Hypoxia-inducible factor 1:regulation by hypoxic and non-hypoxic activators. Int J Biochem CellBiol 2005;37:535–40.

37. GalbanS, KuwanoY, PullmannRJr,Martindale JL, KimHH, Lal A, et al.RNA-binding proteins HuR and PTB promote the translation of hyp-oxia-inducible factor 1alpha. Mol Cell Biol 2008;28:93–107.

38. MohammedN, RodriguezM,Garcia V, Garcia JM, DominguezG, PenaC, et al. EPAS1 mRNA in plasma from colorectal cancer patients isassociated with poor outcome in advanced stages. Oncol Lett2011;2:719–24.

39. Mazumdar J, Hickey MM, Pant DK, Durham AC, Sweet-Cordero A,Vachani A, et al. HIF-2alpha deletion promotes Kras-driven lung tumordevelopment. Proc Natl Acad Sci U S A 2010;107:14182–7.

40. HuCJ,Wang LY, Chodosh LA, Keith B, SimonMC. Differential roles ofhypoxia-inducible factor 1alpha (HIF-1alpha) and HIF-2alpha in hyp-oxic gene regulation. Mol Cell Biol 2003;23:9361–74.

41. Hu CJ, Sataur A, Wang L, Chen H, Simon MC. The N-terminaltransactivation domain confers target gene specificity of hypoxia-inducible factors HIF-1alpha and HIF-2alpha. Mol Biol Cell 2007;18:4528–42.

42. Lau KW, Tian YM, Raval RR, Ratcliffe PJ, Pugh CW. Target geneselectivity of hypoxia-inducible factor-alpha in renal cancer cells isconveyed by post-DNA-binding mechanisms. Br J Cancer 2007;96:1284–92.

43. Carroll VA, Ashcroft M. Role of hypoxia-inducible factor (HIF)-1alphaversus HIF-2alpha in the regulation of HIF target genes in response tohypoxia, insulin-like growth factor-I, or loss of von Hippel-Lindau






function: implications for targeting the HIF pathway. Cancer Res2006;66:6264–70.

44. Acker T, Diez-Juan A, Aragones J, Tjwa M, Brusselmans K, Moons L,et al. Genetic evidence for a tumor suppressor role of HIF-2alpha.Cancer Cell 2005;8:131–41.

45. Walczak-Drzewiecka A, Ratajewski M, Pulaski L, Dastych J. DNAmethylation-dependent suppression of HIF1A in an immature hemato-poietic cell line HMC-1. Biochem Biophys Res Commun 2010;391:1028–32.

46. Migliore L,Migheli F, Spisni R,CoppedeF.Genetics, cytogenetics, andepigenetics of colorectal cancer. J Biomed Biotechnol 2011;2011:792362.

47. Momparler RL. Pharmacology of 5-Aza-20-deoxycytidine (decitabine).Semin Hematol 2005;42(3 Suppl 2):S9–16.

48. Ghoshal K, Motiwala T, Claus R, Yan P, Kutay H, Datta J, et al.HOXB13, a target of DNMT3B, is methylated at an upstream CpGisland, and functions as a tumor suppressor in primary colorectaltumors. PloS ONE 2010;5:e10338.






2014;12:1112-1127. Published OnlineFirst May 13, 2014.Mol Cancer Res Agnieszka Anna Rawluszko-Wieczorek, Karolina Horbacka, Piotr Krokowicz, et al. HIF1A and EPAS1 in Colorectal CancerPrognostic Potential of DNA Methylation and Transcript Levels of

Updated version

10.1158/1541-7786.MCR-14-0054doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2014/05/13/1541-7786.MCR-14-0054.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/12/8/1112.full#ref-list-1

This article cites 45 articles, 9 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/12/8/1112.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mcr.aacrjournals.org/content/12/8/1112To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-14-0054

http://mcr.aacrjournals.org/content/suppl/2014/05/13/1541-7786.MCR-14-0054.DC1

http://mcr.aacrjournals.org/content/12/8/1112.full#ref-list-1

http://mcr.aacrjournals.org/content/12/8/1112.full#related-urls

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mcr.aacrjournals.org/content/12/8/1112


Prognostic Potential of DNA Methylation and Transcript Levels of … · the EPAS1 transcript, as...

Documents

Transcript of Prognostic Potential of DNA Methylation and Transcript Levels of … · the EPAS1 transcript, as...