Sct promoter methylation is a highly discriminative biomarker for lung

Received 24 April 2015; accepted 26 September 2015. Date of publication 7 October 2015;date of current version 4 November 2015.

Digital Object Identifier 10.1109/LLS.2015.2488438

SCT Promoter Methylation Is a HighlyDiscriminative Biomarker for Lung

and Many Other CancersADWAIT SATHE1, YU-AN ZHANG2, XIAOTU MA1, PRADIPTA RAY1, DANIELA CADINU1,

YI-WEI WANG2, XIAO YAO1, XIAOYUN LIU2, HAO TANG3, YUNFEI WANG1, YING HUANG1,CHANGNING LIU1, JIN GU4, MARTIN AKERMAN5, YIFAN MO5, CHAO CHENG6,

ZHENYU XUAN1, LEI CHEN7, GUANGHUA XIAO3, YANG XIE3, LUC GIRARD2,HONGYANG WANG7, STEPHEN LAM8, IGNACIO I. WISTUBA9,

LI ZHANG1, ADI F. GAZDAR2, AND MICHAEL Q. ZHANG1,41Center for Systems Biology and Department of Molecular and Cell Biology, The University of Texas at Dallas, Richardson, TX 75080 USA

2The Hamon Center for Therapeutic Oncology Research, Department of Pathology, University of Texas SouthwesternMedical Center, Dallas, TX 75390 USA

3Department of Clinical Science, University of Texas Southwestern Medical Center, Dallas, TX 75390 USA4Division of Bioinformatics, Center for Synthetic and Systems Biology, TNLIST, Tsinghua University, Beijing 100084, China

5Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724 USA6Department of Genetics, Geisel School of Medicine at Dartmouth, Hanover, NH 03755 USA

7Laboratory of Signal Transduction, Eastern Hepatobiliary Surgery Hospital, Second Military Medical University, Shanghai 200438, China8BC Cancer Research Center, BC Cancer Agency, Vancouver, BC V52 1L3, Canada

9Department of Translational Molecular Pathology, Thoracic/Head and Neck Medical Oncology,The University of Texas MD Anderson Cancer Center, Houston, TX 77030 USA

CORRESPONDING AUTHORS: M. Q. ZHANG ([email protected]; [email protected]) and A. F. GAZDAR([email protected])

(Adwait Sathe, Yu-An Zhang, and Xiaotu Ma contributed equally to this work.)

This work was supported in part by the Texas Specialized Program of Research Excellence in Lung Cancer under Grant P50CA70907 and the EarlyDetection Research Network under Grant U01CA084971, in part by the National Cancer Institute, and in part by the Canary Foundation, Palo Alto,

CA, USA. The work of M. Q. Zhang was supported by the National Institute of Health under Grant HG001696, by the National BasicResearch Program of China under Grant 2012CB316503, and by the National Natural Science Foundation of China

under Grant 91019016 and Grant 31061160497.All supplementary files can be accessed at http://biorxiv.org/content/early/2015/08/05/018515.

ABSTRACT Aberrant DNA methylation has long been implicated in cancers. In this letter, we present ahighly discriminative DNA methylation biomarker for non-small cell lung cancers (NSCLCs) and 14 othercancers. Based on 69 NSCLC cell lines and 257 cancer-free lung tissues, we identified a CpG islandin SCT gene promoter, which was verified by qMSP experiment in 15 NSCLC cell lines and threeimmortalized human respiratory epithelium cells. In addition, we found that the SCT promoter was methy-lated in 23 cancer cell lines involving >10 cancer types profiled by ENCODE. We found that the SCTpromoter is hypermethylated in primary tumors from TCGA lung cancer cohort. In addition, we foundthat SCT promoter is methylated at high frequencies in 15 malignancies and is not methylatedin∼1000 non-cancerous tissues across>30 organ types. This letter indicates that SCT promoter methylationis a highly discriminative biomarker for lung and many other cancers.

INDEX TERMS Biomarker, DNA methylation, lung cancer, secretin SCT gene.

I. INTRODUCTION

METHYLATION of the fifth position in the six atomcarbon ring of the cytosine base (5 mC) is one of

the best studied epigenetic modifications and is associatedwith development in mammals [1]. Due to the differenceof biochemical properties between methylated cytosine andunmethylated cytosine, the methylated cytosine is sometimescalled the ‘‘fifth’’ base [2]. DNA methylation occurs almost

exclusively at CpG dinucleotide sites, although non-CpGmethylation was recently found to be abundant in thestem cell genome [3]. Since the methylated cytosine mayspontaneously deaminate to thymine, CpG sites are under-represented in the human genome, possibly due to DNAmethylation during evolution [4]. However, there are stillstretches of DNA sequences with over-represented CpG sites,called CpG islands [5]. CpG islands are, in general, unmethy-

30

2332-7685 2015 IEEE. Translations and content mining are permitted for academic research only.Personal use is also permitted, but republication/redistribution requires IEEE permission.

See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. VOLUME 1, NO. 3, OCTOBER 2015

www.redpel.com +917620593389


Sathe et al.: SCT Promoter Methylation Is a Highly Discriminative Biomarker

lated and overlap the promoter regions of 60%–70% of allhuman genes [5]. Moreover, the methylation status ofcytosine nucleotides is somatically inheritable through main-tenance methyltransferase (e.g., DNMT1), which mostlytargets hemimethylated CpG dyads on the parental strandduring DNA replication [6].

Lung cancer is the leading cause of cancer-related mor-tality worldwide, and many patients are usually diagnosedat advanced stages [7]. Five-year survival of lung cancerpatients after resection strongly depends on tumor stages(67% for stage I versus 23% for stage IV), indicatingthe importance of early diagnosis [8]. Computed tomogra-phy (CT) screening is the most recent state-of-the-art methodfor early detection [9]. However, it is subject to high false-positive rates [10]. As a result, confirming the presence oflung cancers by invasive biopsy is still a key step for earlydetection [10]. Molecular biomarkers offer great promise inearly detection of lung cancers which in turn can reducemortality. For example, DNA methylation has been studiedin squamous cell lung cancer [11] and lung adenocarci-noma [12]–[14], as well as multiple malignancies [15]. DNAmethylation biomarkers for lung cancers are also studied insputum [16] and plasma [17] for non-invasive procedures.In this report, we describe our search for a near universallung cancer methylation marker, utilizing the vast amount ofpublically available data.

II. RESULTA. INITIAL SCREENING FOR HYPERMETHYLATEDBIOMARKERS IN NON-SMALL CELL LUNGCANCER (NSCLC) CELL LINESIn order to discover DNA methylation markers fornon-small cell lung cancers (NSCLCs), we obtained Human-Methylation450 (450k) methylation data for 69 cell lines(Table S1) derived from the tumors of NSCLC patients and257 primary cancer-free lung tissues (Table S2). We appliedt-test on the methylation M -values of the tumor andcancer-free samples for a search restricted on390 920 CpG sites. We choose the CpG sites that are notwithin ten base pairs of known SNPs and are detectablein >95% cell lines. We considered probes with FDRlevel 1e-68 and effect size (i.e., differences of averageM -values between cancer and non-malignant/normal sam-ples) to be higher than +4.5 (a positive marker fortumors), and have at least two probes that are consistentlyhypermethylated in NSCLC cell lines (Fig. S1-2). The toptwo CGIs were chr11:626728-628037 overlapping with thepromoter of secretin gene [SCT; Figs. S1-2 and 1(a), andTable S3] and chr2:63283936-63284147 overlapping withthe 3’UTR of OTX1 (Table S3). Here, we present the resultson probe cg00249511 in the SCT promoter region in mainfigures.1

1The results of probe cg25774643 can be found in supplementarymaterials that can be accessed at http://biorxiv.org/content/early/2015/08/05/018515.

FIGURE 1. SCT promoter (cg00249511) methylation as a biomarkerin (a) 69 NSCLC cell lines (red dots) and 257 normal lung tissues(blue dots). The methylation level (β value) is shown in the y-axis. (b) SCTmethylation in cell lines profiled by ENCODE, including 23 cancer cell lines(red shade) and four cell lines immortalized by oncogenes or oncogenicviral agents, such as Epstein–Barr virus, Adenovirus, and ER-Src, whichare shaded in yellow. Cell line names and tissue sources are indicated.

To check SCT methylation in other cancers, we obtainedDNA methylation data in cell lines profiled by ENCODE(Table S1), of which 23 cell lines were the cancer cell lines.As can be seen in Figs. 1(b) and S3-4, SCT promoter wasmethylated in all cancer cell lines, though its methylationlevel in cell lines BE2_C and SK_N_MC appears to bereduced. These data indicate that the SCT promoter might bea highly discriminative DNAmethylationmarker for lung andmany other cancers.

B. VALIDATION OF SCT PROMOTER METHYLATIONBY SCT-SPECIFIC qMSPTo validate the above findings, we developed an SCT-specificqMSP assay (Materials/Methods) by targeting the SCT pro-

FIGURE 2. Determination of SCT methylation in genomic DNAsof 15 NSCLC cancer cell lines and three immortalized human respiratoryepithelial cell lines (HREC-KT) by SCT qMSP assay (y-axis) and 450kmethylation array (x-axis; probe cg00249511). PMR : percent ofmethylated reference with respect to fully methylated DNA controls.

VOLUME 1, NO. 3, OCTOBER 2015 31




moter region and examined 15 NSCLC cancer cell lines andthree immortalized human respiratory epithelial cell (HREC)lines, whose SCT methylation status was also confirmedby 450k methylation array (Table S4). As can be seenin Figs. 2 and S5, the data from both methods consistentlyindicated the SCT promoter being methylated in lung cancercell lines and none or minimal in those immortalized lungcell lines. In addition, the SCT qMSP analysis of nor-mal WBC cells showed none or minimal level of SCTmethylation, which was consistent with the SCT promoterbisulfite sequencing analysis (Zhang et al. manuscript inpreparation).

C. SCT PROMOTER METHYLATION IN PRIMARYTUMORS AND NON-TUMOR SAMPLESWe next checked SCT promoter methylation in30 cancer types and corresponding non-malignant sam-ples (Table S5) profiled by TCGA [18]. As can be seenin Figs. 3 and S6 (top panel), the SCT promoter is methylated(mean β > 0.5) in a wide range of cancers, includingovarian, uterine/uterine corpus, cervical, lung, liver, bladder,esophageal, low grade glioma, prostate, head/neck, breast,kidney, and stomach, which involves 67.23% (5246/7803)patients in our analysis. Interestingly, the SCT promoter isalso methylated, though at a reduced level and frequency, incancers including diffuse large B-cell lymphoma, glioblas-toma multiforme, mesothelioma, skin, sarcoma, pancreas,and kidney, which involves 18.65% (1455/7803) patientsin our analysis. Finally, the SCT promoter does not seemto be highly methylated in cancers including colorectal,AML, testicular germ cell tumors, adrenocortical, rectum,

FIGURE 3. Top: SCT promoter (cg00249511) methylation in TCGA primarytumors (red). Middle: TCGA non-malignant tissues (blue). Bottom:cancer-free tissues (green). The methylation level is shown in y-axis (Betavalue).

TABLE 1. Summary statistics for SCT promoter methylation inTCGA tumor, non-malignant tissue, and cancer-free samples.

kidney chromophobe, uveal melanoma, and thymoma, whichinvolves 14.12% (1102/7803) patients in our analysis. Thedescriptive statistics of SCT promoter methylation in allcancer types are summarized in Table 1.

Since the non-malignant tissues from TCGA demon-strated reduced but significant SCT promoter methylation[Figs. 3 and S6 (middle panel)], we collected theDNA methylation data from cancer-free samples (TableS2) to account for possible field effect. As can be seenin Figs. 3 and S6 (bottom panel), it appears that the SCTpromoter is not methylated in corresponding normal tis-sues/organs, though the sample sizes for a few tissue typesare small. Moreover, the SCT promoter does not appear tobe highly methylated in 98% samples from >30 tissue/organtypes from cancer-free samples (Table S2 and Fig. S7).

Based on the above data, we concluded that SCT pro-moter methylation is a highly discriminative biomarker forlung and 14 other cancers. Interestingly, the SCT gene doesnot appear to be expressed in the primary tumor and non-malignant samples in our analysis (Table S6). In turn, we didnot detect the SCT differential expression between tumor andnon-malignant samples in our analysis (Fig. S8).

III. DISCUSSIONIn this report, we have discovered that SCT promoter methy-lation is a potential highly discriminative biomarker to sep-arate the tumors from non-malignant tissues for lung and 14other malignancies. Given the great discriminative potentialof SCT promoter methylation in these 15 cancers, the applica-tions for non-invasive procedures, such as detection in urinefor bladder or renal cancers, sputum (or nasal mucosa as asurrogate marker) for lung cancer [19], [20], nipple aspiratesfor breast cancer, and prostatic massage for prostate cancer,may also be feasible.

32 VOLUME 1, NO. 3, OCTOBER 2015




IV. MATERIALS AND METHODSA. CELL LINES AND DNAThe 15 NSCLC lung cancer cell lines (Table S4) and threeCDK4/hTERT-immortalized HREC lines (Table S4; [21])were provided by A. F. Gazdar and J. D. Minna(UT Southwestern Medical Center at Dallas), and theauthenticity of these cell lines was confirmed by the DNAfingerprint genotyping tests. Genomic DNA was isolatedusing QIAamp BloodMini kit (Qiagen). DNA concentrationswere measured using Nanodrop2000 (Thermo Scientific).

B. SCT-SPECIFIC qMSPThe SCT-specific qMSP assay was designed to target theSCT promoter region and the first exon, and was per-formed similarly as described in [22]. The primers andprobes sequences are available upon request. The qMSPPCR assay, including MYOD1 gene as an input DNA ref-erence [22], was carried out by LightCycler 480 system(Roche). CpGenome Universal Methylated DNA (Milli-pore) was used as positive control and methylated refer-ence. The level of SCT methylation of a given test waspresented as a percent of methylated reference (PMR) asdetermined by the method as described similarly in [22]and [23]. The methylation status of SCT in these lungcancer cell lines and HREC lines was also confirmedby Illumina HumanMethylation450 BeadChip (Illumina,San Diego, CA).

REFERENCES[1] Z. D. Smith and A. Meissner, ‘‘DNA methylation: Roles in mam-

malian development,’’ Nature Rev. Genet., vol. 14, no. 3, pp. 204–220,2013.

[2] L. Jiang et al., ‘‘Sperm, but not oocyte, DNA methylome is inher-ited by zebrafish early embryos,’’ Cell, vol. 153, no. 4, pp. 773–784,2013.

[3] R. Lister et al., ‘‘Human DNA methylomes at base resolution showwidespread epigenomic differences,’’ Nature, vol. 462, no. 7271,pp. 315–322, 2009.

[4] P. W. Laird, ‘‘Principles and challenges of genome-wide DNA methy-lation analysis,’’ Nature Rev. Genet., vol. 11, no. 3, pp. 191–203,2010.

[5] R. S. Illingworth and A. P. Bird, ‘‘CpG islands—‘A rough guide,’’’ FEBSLett., vol. 583, no. 11, pp. 1713–1720, 2009.

[6] D. P. Genereux, B. E. Miner, C. T. Bergstrom, and C. D. Laird,‘‘A population-epigenetic model to infer site-specific methylation ratesfrom double-stranded DNA methylation patterns,’’ Proc. Nat. Acad.Sci. USA, vol. 102, no. 16, pp. 5802–5807, 2005.

[7] R. Siegel, E. Ward, O. Brawley, and A. Jemal, ‘‘Cancer statistics, 2011:The impact of eliminating socioeconomic and racial disparities on prema-ture cancer deaths,’’ CA, Cancer J. Clin., vol. 61, no. 4, pp. 212–236, 2011.

[8] H. Tang et al., ‘‘A 12-gene set predicts survival benefits from adjuvantchemotherapy in non–small cell lung cancer patients,’’ Clin. Cancer Res.,vol. 19, no. 6, pp. 1577–1586, 2013.

[9] U. Ahmad and F. C. Detterbeck, ‘‘Current status of lung cancer screening,’’Seminars Thoracic Cardiovascular Surgery, vol. 24, no. 1, pp. 27–36,2012.

[10] J. M. Croswell et al., ‘‘Cumulative incidence of false-positive results inrepeated, multimodal cancer screening,’’ Ann. Fam. Med., vol. 7, no. 3,pp. 212–222, 2009.

[11] P. P. Anglim et al., ‘‘Identification of a panel of sensitive and specific DNAmethylation markers for squamous cell lung cancer,’’ Molecular Cancer,vol. 7, p. 62, Jul. 2008.

[12] K. M. Kerr, J. S. Galler, J. A. Hagen, P. W. Laird, and I. A. Laird-Offringa,‘‘The role of DNAmethylation in the development and progression of lungadenocarcinoma,’’ Disease Markers, vol. 23, nos. 1–2, pp. 5–30, 2007.

[13] J. A. Tsou et al., ‘‘Identification of a panel of sensitive and specific DNAmethylation markers for lung adenocarcinoma,’’Molecular Cancer, vol. 6,p. 70, Oct. 2007.

[14] S. A. Selamat et al., ‘‘Genome-scale analysis of DNA methylation in lungadenocarcinoma and integration with mRNA expression,’’ Genome Res.,vol. 22, no. 7, pp. 1197–1211, 2012.

[15] D. S. Shames et al., ‘‘A genome-wide screen for promoter methylation inlung cancer identifies novel methylation markers for multiple malignan-cies,’’ PLoS Med., vol. 3, no. 12, p. e486, 2006.

[16] S. Leng et al., ‘‘Defining a gene promoter methylation signature in sputumfor lung cancer risk assessment,’’ Clin. Cancer Res., vol. 18, no. 12,pp. 3387–3395, 2012.

[17] C. Kneip et al., ‘‘SHOX2 DNA methylation is a biomarker for the diag-nosis of lung cancer in plasma,’’ J. Thoracic Oncol., vol. 6, no. 10,pp. 1632–1638, 2011.

[18] The Cancer Genome Atlas—National Cancer Institute. TheCancer Genome Atlas Home Page. [Online]. Available:http://cancergenome.nih.gov/, accessed Apr. 2, 2015.

[19] B. Schmidt et al., ‘‘SHOX2 DNA methylation is a biomarker for thediagnosis of lung cancer based on bronchial aspirates,’’ BMC Cancer,vol. 10, p. 600, Nov. 2010.

[20] X. Zhang et al., ‘‘Similarities and differences between smoking-relatedgene expression in nasal and bronchial epithelium,’’ Physiol. Genomics,vol. 41, no. 1, pp. 1–8, 2010.

[21] M. Sato et al., ‘‘Human lung epithelial cells progressed to malig-nancy through specific oncogenic manipulations,’’Molecular Cancer Res.,vol. 11, no. 6, pp. 638–650, Jun. 2013.

[22] N. Shivapurkar et al., ‘‘Application of a methylation gene panel by quan-titative PCR for lung cancers,’’ Cancer Lett., vol. 247, no. 1, pp. 56–71,Mar. 2007.

[23] J. A. Tsou et al., ‘‘Distinct DNA methylation profiles in malignantmesothelioma, lung adenocarcinoma, and non-tumor lung,’’ Lung Cancer,vol. 47, no. 2, pp. 193–204, Feb. 2005.

VOLUME 1, NO. 3, OCTOBER 2015 33



Education

Sct promoter methylation is a highly discriminative biomarker for lung