Integrated Systems and Technologies

Genomic Landscape of Somatic Alterations inEsophageal Squamous Cell Carcinoma andGastric CancerNan Hu1, Mitsutaka Kadota2, Huaitian Liu2, Christian C. Abnet1, Hua Su1, Hailong Wu2,Neal D. Freedman1, Howard H. Yang2, Chaoyu Wang1, Chunhua Yan2, Lemin Wang1,Sheryl Gere2, Amy Hutchinson1,3, Guohong Song2, Yuan Wang4, Ti Ding4, You-Lin Qiao5,Jill Koshiol1, Sanford M. Dawsey1, Carol Giffen6, Alisa M. Goldstein1, Philip R. Taylor1, andMaxwell P. Lee2

Abstract

Gastric cancer and esophageal cancer are the second andsixth leading causes of cancer-related death worldwide. Multi-ple genomic alterations underlying gastric cancer and esoph-ageal squamous cell carcinoma (ESCC) have been identified,but the full spectrum of genomic structural variations andmutations have yet to be uncovered. Here, we report the resultsof whole-genome sequencing of 30 samples comprising tumorand blood from 15 patients, four of whom presented withESCC, seven with gastric cardia adenocarcinoma (GCA), andfour with gastric noncardia adenocarcinoma. Analyses revealedthat an A>C mutation was common in GCA, and in addition tothe preferential nucleotide sequence of A located 5 prime to themutation as noted in previous studies, we found enrichment ofT in the 5 prime base. The A>C mutations in GCA suggested

that oxidation of guanine may be a potential mechanismunderlying cancer mutagenesis. Furthermore, we identifiedgenes with mutations in gastric cancer and ESCC, includingwell-known cancer genes, TP53, JAK3, BRCA2, FGF2, FBXW7,MSH3, PTCH, NF1, ERBB2, and CHEK2, and potentially novelcancer-associated genes, KISS1R, AMH, MNX1, WNK2, andPRKRIR. Finally, we identified recurrent chromosome altera-tions in at least 30% of tumors in genes, including MACROD2,FHIT, and PARK2 that were often intragenic deletions. Thesestructural alterations were validated using the The CancerGenome Atlas dataset. Our studies provide new insights intounderstanding the genomic landscape, genome instability,and mutation profile underlying gastric cancer and ESCCdevelopment. Cancer Res; 76(7); 1714–23. �2016 AACR.

IntroductionGastric cancer and esophageal cancer cause an estimated

783,000 and 407,000 deaths, respectively, each year, and repre-sent the second and sixth leading causes of cancer-related deathworldwide (1). In China, gastric cardia adenocarcinoma (GCA)and esophageal squamous cell carcinoma (ESCC) occur togetherin the Taihang Mountains of north central China, includingShanxi andHenan Provinces, at some of the highest rates reportedfor any cancer (2), and historically over 20% of all deaths in this

region have been attributed to these diseases (3). However, thecause of the high rates and geographical overlap of these twoanatomically adjacent but histologically distinct tumors has notbeen determined. Gastric cancers in this area occur primarilyin the uppermost portion of the stomach and are referred to asGCA, whereas those in the remainder of the stomach are referredto as gastric noncardia adenocarcinoma (GNCA). In addition tobeing anatomically adjacent, GCA and ESCC share many of thesame etiologic risk factors, and before the widespread use ofendoscopy and biopsy, they were diagnosed as a single diseasereferred to as "esophageal cancer" (4). The reason for the highrates ofGCAandESCC in this geographic area and their relation toeach other remains unclear, but there are almost certainly com-mon etiologically important environmental exposures, and arecent genome-wide association study of germline DNA foundthat the same SNPs in the PLCE1 gene had the strongest associa-tions with risk for both GCA and ESCC (5). This led to ourconcurrent examination of these two cancers plus GNCA in thecurrent study.

Recent advances in next-generation sequencing technologyhave revolutionized how we study cancer genomes. The iden-tification of IDH1/2 mutations, initially in glioma (6, 7) andmore recently in many other cancers such as AML (8), hastransformed our understanding of cancer by relating mutationsto metabolic control and epigenetic regulation (9). IDH1/2encodes isocitrate dehydrogenases (IDH), which convert

1Division of Cancer Epidemiology andGenetics, National Cancer Insti-tute, NIH, Bethesda, Maryland. 2Center for Cancer Research, NationalCancer Institute, NIH, Bethesda, Maryland. 3Cancer GenomicsResearch Laboratory, Leidos,Gaithersburg, Maryland. 4Shanxi CancerHospital, Taiyuan, Shanxi, PR China. 5Cancer Institute, Chinese Acad-emyofMedical Sciences,Beijing, PRChina. 6InformationManagementServices, Inc., Silver Spring, Maryland.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

CorrespondingAuthors:Philip R. Taylor, Genetic EpidemiologyBranch, Divisionof Cancer Epidemiology and Genetics, National Cancer Institute, 9609 MedicalCenter Drive, Rm 6E444 MSC 90892, Rockville, MD 90892-9769. Phone: 240-276-7235; Fax: 240-276-7832; E-mail: ptaylor@mail.nih.gov; andMaxwell P. Lee,leemax@mail.nih.gov

doi: 10.1158/0008-5472.CAN-15-0338

�2016 American Association for Cancer Research.

CancerResearch

Cancer Res; 76(7) April 1, 20161714

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

isocitrate to 2-oxoglutarate. But mutant IDHs produce 2-hydro-xyglutarate, which inhibits the methyl cytosine hydroxylaseTET2 as well as H3K36 demethylases, thus changes the globalepigenetic landscape. Whole genome sequencing (WGS) isparticularly useful for elucidating complex genomic changes,including translocations, inversions, tandem duplications, andlarge deletions. The importance of these structural changes hasbeen well documented in the case of BCR-ABL in leukemia,TMPRESS2-ERG in prostate cancer (10), and EML4-ALK in lungcancer (11).

For gastric adenocarcinoma and ESCC, several publicationshave reported genomic scale analyses of cancers using exomeor WGS technology. Wang and colleagues performed exomesequencing of 22 gastric cancer samples and identified fre-quent mutations in ARID1A (12). Mutations in ARID1A wereparticularly high in gastric cancers with microsatellite insta-bility (MSI; 83%) or with Epstein-Barr virus (EBV) infection(73%). Exome sequencing of 15 gastric adenocarcinomas andtheir matched normal DNAs by Zang and colleagues alsoidentified frequent mutations of ARID1A (13). In addition,they found 5% of gastric cancer contained FAT4 mutations. Astudy by Agrawal and colleagues reported exomic sequencingof 11 esophageal adenocarcinomas (EAC) and 12 ESCCsand found frequent NOTCH1 mutations in ESCC (14).Nagarajan and colleagues performed WGS analysis for twogastric cancer samples and found three mutational signatures(15). Dulak and colleagues did exome sequencing for 149EAC tumor-normal pairs and WGS for 15 EACs and matchednormals. They found a high prevalence of A>C transversions atAA dinucleotides (16). A recent study by Wang and colleaguesanalyzed 100 tumor–normal pairs of gastric cancer withWGS and identified MUC6, CTNNA2, GLI3, RNF43, andRHOA as significantly mutated driver genes. They found thatRHOA mutations are specific for diffuse-type tumors (17).Recently, the International Cancer Genome Consortium re-search team published a study involving WGS of 17 ESCCcases and exome sequencing of 71 ESCC cases, which iden-tified two novel cancer driver genes, ADAM29 and FAM135B(18). Lin and colleagues performed exome sequencing on20 ESCC cases and targeted sequencing of 139 ESCC casesand identified several mutated genes that were previouslyunknown, including FAT1, FAT2, ZNF750, and KMT2D (19),and Gao and colleagues did exome sequencing on 113 ESCCcases and noted that histone modifier genes were frequentlymutated, including KMT2D, KMT2C, KDM6A, EP300, andCREBBP. (20). Zhang and colleagues recently reported exomesequencing of 90 ESCC and WGS of 14 ESCCs and foundthat an APOBEC-mediated mutational signature in 47% oftumors (21).

Despite this progress, the full spectrum of genomic altera-tions in gastric cancer and ESCCs, particularly genomic struc-tural variations (SV) and intergenic and intronic mutations,remains to be characterized. To discover and characterize geno-mic alterations in GCA, GNCA, and ESCC, we analyzed whole-genome sequences of tumors and matched blood DNA samplesgenerated by Complete Genomics Inc. (CGI). We present hereour findings of novel mutation substitution pattern, drivermutations, and recurrent SVs. Complete characterization of thegenomic landscape of these cancers will hopefully provide newstrategies for early diagnosis and therapy for these deadlydiseases.

Materials and MethodsStudy population

This study was approved by the Institutional Review Boards ofthe Shanxi Cancer Hospital, the Cancer Institute and Hospital ofthe Chinese Academy of Medical Sciences (CICAMS, Shanxi, PRChina), and the U.S. National Cancer Institute (NCI, Bethesda,MD). Fifteen caseswere analyzedwithWGS. These cases came froma larger study sample andwere selected on the basis of high-quality,sufficiently large amount of DNA (requiring at least 10 mg) avail-able forWGS, and patients being deceased. Three cases with ESCC,seven cases with GCA, and four cases with GNCA diagnosedbetween 1998 and 2001 in the Shanxi Cancer Hospital in Taiyuan,Shanxi Province, PR China, were recruited to participate in thisstudy.One ESCC case fromYaocunCommuneHospital in Linxian,Henan Province was also recruited. None of the cases had therapybefore their surgical resection. After obtaining informed consent,cases were interviewed to obtain information on demographics,cancer risk factors (e.g., detailed family history of cancer), andclinical information. Only deceased cases were selected for study.Clinical data are described in Supplementary Table S1.

Biological specimen collection and processingVenous blood (10mL) was taken from each case before surgery

and germline DNA from whole blood was extracted and purifiedusing the standard phenol/chloroformmethod. Tumors obtainedduring surgery were snap-frozen in liquid nitrogen and stored at�130�C until used. The specimens were chosen for this studybased on two criteria: (i) histologic diagnosis of ESCC or gastriccancer confirmed by pathologists at the Shanxi CancerHospital orCICAMS, and the NCI; (ii) availability of high purity tumor tissue(at least >75%).

Tissue DNA isolationDNA from frozen tumors was extracted using AllPrep DNA/

RNA/Protein Mini Kit (Qiagen, Inc.). DNA was dissolved in 100mL Buffer BE. Concentrations of DNA were measured with theNanoDrop 2000 Spectrophotomer (Thermo Fisher Scientific)according to the manufacturer's instructions. DNAs were run on0.7% agarose gel (UltraPure Agarose Powder, Invitrogen) in 1�TAE buffer to identify high-molecular weight genomic DNA (>20Kb single band) and photographed using AlphaImager EC system(Biosciences, Inc.). DNA was quantified using Quant-iT Pico-Green Kit (Invitrogen).

Whole genome sequencingOf note, 10 mg DNA was sent to CGI for WGS. CGI delivered

data for SVs, DNA copy number variations, and single nucle-otide variations (SNV). A detailed description of WGS data canbe found in the company's website http://www.completege-nomics.com/. Summary of the WGS data is described in Sup-plementary Table S2. We focused on somatic alterations, that is,tumor-specific changes present in tumor but absent in blood.We used cgatools, CallDiff, and JunctionsDiff to generatesomatic SNVs and somatic SVs, respectively. Somatic copynumber alterations (CNA) data came from CGI reports. Thesethree types of somatic changes for each patient are shown inCircos plots in Fig. 1.

Genotyping with the Affymetrix SNP5.0 arrayWe performed genotyping experiments for 12 samples using

the Affymetrix GeneChip Human Mapping 5.0 arrays. Briefly,

Somatic Alterations in Esophageal and Gastric Cancer

www.aacrjournals.org Cancer Res; 76(7) April 1, 2016 1715

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

250 ng of DNA was digested with Nsp I or Sty I, ligated to anadaptor, and amplified by PCR. PCR products were processedfor fragmentation and labeling; labeled DNA was hybridizedonto the chips. The chip was scanned with the AffymetrixGeneChip Scanner 7G Plus using Affymetrix GeneChip Com-mand Console software, and the data files were automaticallygenerated. Genotype calls were generated by GenotypingConsole v4.1 software (Affymetrix). SNP calls generated bythe SNP array and WGS agreed very well, with a median of99.65% of the time across all 12 genomes (Supplementary

Table S3). We noted that agreement was slightly lower intumors (median 98.85%); this reduced concordance was dueto genomic regions containing SVs and CNAs. Concordancesfor CC0996, EC1475, and EC8413 were below 99%.

GEO accession number for the SNP array is GSE43470.

PCR and Sanger sequencingTo validate somatic SNVs and SVs identified from the WGS

data, we used PCR to amplify genomic DNA spanning SNVs or SVjunctions. In addition, we also amplified SV junction sequences

Figure 1.Circos plots of somatic single nucleotide variants, copy number alterations, and structural variations in the 15 cancer genomes. The inner ring displays SVs: blackfor intrachromosomal SVs and red for interchromosomal SVs. The second ring next to SVs is CNA, shown in gray. The third ring is SNV, shown in green. Theoutside ring is the chromosome ideogram. The sample description can be found in Supplementary Table S1.

Hu et al.

Cancer Res; 76(7) April 1, 2016 Cancer Research1716

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

from cDNAs. Target regions spanning SNVs and SVs were ampli-fied by PCR using genomic DNA or cDNA as a template. Theprimers were designed using Primer3 (http://www.broadinsti-tute.org/genome_software/other/primer3.html), and they aresummarized in Supplementary Tables S7 and S8. PCR productswere purified using QiaQuick PCR Purification Kit (Qiagen),and sequenced using ABI BigDye Terminator BDT 3.1 (AppliedBiosystems). Sequencing reactions were carried out at 96�C for10 seconds, 50�C for 5 seconds, and 60�C for 2 minutes for 25cycles, and the reaction product was analyzed in 3730 XL DNAAnalyzer (Applied Biosystems). We carried out validation for asubset of somatic missense mutations in subjects CC0996 andEC0379 using Sanger sequencing. High-quality sequencingdata were obtained for 62 mutations from CC0996; 51(82%) validated and 11 did not. For EC0379, we examined30 mutations, and 17 (63%) were validated. Mutations calledby WGS but not validated by Sanger sequencing often had a lowfraction of the mutant allele (below 15% of total sequencereads). The discrepancy between the two methods is partlyattributable to the stringent criteria of mutation calling weused for the Sanger sequencing data, which usually requiresa minimal of 15% of the mutant allele. Validated mutations aresummarized in Supplementary Table S4.

Comparison of somatic mutation call using blood controlversus adjacent normal tissue control

In addition to blood DNA controls, we had WGS data for sixadjacent normal tissues. We compared somatic variant callusing blood DNA as control versus adjacent normal tissue ascontrol. Two tumors with very low somatic mutation rates wereexcluded from this analysis because they had higher false-positive calls. We found that the concordance between bloodDNA controls versus adjacent normal tissue controls rangedfrom 86% to 95%. The tumors with higher somatic mutationrates also had higher concordance. This is also consistent withthe validation data from Sanger sequencing described in theprevious section.

Statistical analysisWhole-genome sequence analysis. We used cgatools (http://www.completegenomics.com/sequence-data/cgatools/) and custom-built Perl scripts to analyze WGS data. The calldiff and junction-diff (cgatools) were used to identify somatic SNVs and SVs,respectively; the junctions2events method (cgatools) was usedto identify genomic events such as deletions, duplications, inver-sions, and complex events that underlie observed SV junctionsequences. The Circos plots were generated as described (http://circos.ca/).

Ingenuity variant analysis. For functional interpretation ofsomatic mutations, we used Ingenuity Variant Analysis (IVA).IVA uses a cascade of filters to identify potential cancer drivermutations; these filters consist of selecting variants based oncommon variants database (1,000 genome, CGI database, andNHLBI ESP exomes), predicted deleterious effects (literature,SIFT, phyloP, etc.), genetic analysis (gain or loss function,number of alleles, absence of variant in the control), and can-cer driver variants [literature, Catalogue of Somatic Mutationsand Cancer, and The Cancer Genome Atlas (TCGA)]. (For de-tails, see http://www.ingenuity.com/products/ingenuity_variant_

analysis.html). We used the following cascade of filters toidentify potential cancer driver mutations: (i) selection ofvariants based on common variants database (1,000 genome,CGI database, and NHLBI ESP exomes); (ii) selection of vari-ants based on predicted deleterious effects (literature, SIFT,phyloP, etc.). Variants were selected on the basis of phenotypes(pathogenic or unknown significance), gain of function (liter-ature, Ingenuity Knowledge Base, or BSIFT), or loss of function(non-synonymous, damaging by SIFT, splicing sites, or affect-ing microRNA); (iii) selection of variants based on geneticanalysis (gain or loss of function, number of alleles, absenceof variant in the control). We excluded variants found in anyblood sample. We included homozygous, hemizygous, orcompound heterozygous; (iv) selection of variants were basedon cancer driver variants.

Miscellaneous analyses. All statistical analyses and plots weregenerated using the R environment, and we used shell scriptsand Perl scripts for general bioinformatics data processing.

ResultsWe analyzed 30 genomes fromDNAs isolated from two tissues

(tumor and blood) from 15 patients, seven with gastric cardiaadenocarcinoma (GCA), four with gastric noncardia adenocarci-noma (GNCA), and four with ESCC, by WGS, generated by theCompleteGenomics Inc. (CGI). Patient clinical data are describedin Supplementary Table S1. Total sequences generated for eachgenome ranged from193.7Gb to 363.4Gb (median¼321.7Gb).Details of our WGS data for each genome are described inSupplementary Table S2.

We focused on identification and characterization of somaticchanges, that is, changes in tumors that were absent in matchedgermline (blood) DNA. Figure 1 shows the Circos plots of threetypes of somatic changes of interest (SNVs, SVs, andCNAs) for the15 cancer genomes Gastric cancer, in particular GNCA, had morestructural alterations than ESCC.

Characterization of somatic single nucleotide substitutionsFigure 2A shows somatic mutation rates per million bases.

Mutation rates ranged between 12.7 and 70.9 per million bases(median¼33.4) for intergenic regions, between 10.2 and 41.5(median¼21.7) for introns, and between 9.4 and 27.8 (med-ian¼17.3) for exons. These mutation rates were similar to thosereported for colon cancer (22), pancreatic cancer (23), liver cancer(24) gastric cancer (12, 13), and esophageal cancer (14).Mutationrates were highest in intergenic regions, followed by introns andexons. This is consistent with the role of a transcription-coupledrepair process in reducing intragenic mutations. The codingmutation rates ranged from 3.9 to 14.0 per million bases (med-ian¼8.6) for missense mutations, and from 0.13 to 0.64 (med-ian¼0.36) for nonsense mutations. The mutation rates of syn-onymous, missense, nonsense, and other changes are summa-rized in Fig. 2B.

We noted that BC5439, CC1649, CC1730, and EC1475 hadmuch higher mutation rates, exceeding 40 million per Mb inintergenic regions. To investigate a potential mutagenic mech-anism for the higher mutation rates, we calculated mutationrates for each type of single nucleotide substitution (Fig. 2C).Transition rates, A>G and C>T, were high as expected. How-ever, we saw high A>C mutation rates in six of the 11 gastric

Somatic Alterations in Esophageal and Gastric Cancer

www.aacrjournals.org Cancer Res; 76(7) April 1, 2016 1717

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

cancers. This increased A>C mutation was also observed inseveral recent studies involving EAC (14, 16) and gastriccancer (17). We also calculated the nucleotide substitutionrates with their dependency on the flanking nucleotide

sequences, and displayed the result in a heatmap (Fig. 2D).We found that the A>C mutation showed a preference ofnucleotide sequence of A located 5 prime to the mutation(Fig. 2D).

60

40

20

0

60

40

20

0

60

40

20

0

60

40

20

0

10

5

0

10

5

0

10

5

0

10

5

0

Mut

atio

ns p

er M

b

Mut

atio

ns p

er M

b

Mut

atio

ns

Fiv

e pr

ime

base

Region

Region

Exon

Intergenic

Intron

Type

Type

Type

Three prime base

Type

Missense

Nonsense

Other

Synonymous

60,000

40,000

20,000

0

60,000

40,000

20,000

0

60,000

40,000

20,000

0

60,000

40,000

20,000

0

A>C

A>C

A>G

A>G

A>T

A>T

C>A

C>A

C>G

C>G

C>T

C>T

EC8413

EC1475

EC1415

EC0379

CC5615

CC1791

CC1730

CC1649

CC1553

CC1220

CC0996

BC5439

BC5430

BC1684

BC1586

A C C C CCG T A C G G G G GT T T T TA A AA

Log (mutation)10

8

6

4

Mutation spectrum

A

C D

B

Figure 2.Summary of somatic mutations and mutation spectrum in GNCA, GCA, and ESCC genomes and non-negative matrix factorization analysis of SNV substitutionmatrix. Somatic SNVs are summarized in Figure 2A and B. The numbers of somatic mutations per million bases in intergenic, intronic, and exonic regionsof GCA, GNCA, and ESCC genomes are illustrated with bar graphs. Here, somatic mutations include single nucleotide variations and small indels. A, graphshows mutations in intergenic, intronic, and exonic regions. B, graph shows mutations of various types of amino-acid changes. SNV substitution patternsare summarized in C and D. C, graph shows the numbers of somatic mutations for the six types of base substitutions, summarized on the left. We also consideredthe context of 5 prime base and 3 prime base; there are 96 possible combinations (4 � 6 � 4). D, the heatmap shows the results for the 15 tumors.

Hu et al.

Cancer Res; 76(7) April 1, 2016 Cancer Research1718

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

Characterization of cancer driver mutationsWe characterized somatic mutations to identify potential can-

cer driver mutations by analyzing our WGS data using IVA. Ourinitial preliminary results from IVA analysis identified 200 genes.We further filtered the genes by requiring that they be mutated inat least three tumors. This resulted in a final list of 24 genes.Tumors with mutations in the 24 genes are illustrated in Fig. 3.Half of these genes are well-known cancer genes, including TP53,JAK3, BRCA2, FGF2, FBXW7, MSH3, PTCH, NF1, ERBB2, andCHEK2. We also identified multiple potentially novel cancergenes, including KISS1R, AMH, MNX1, WNK2, and PRKRIR. Thefull list of the 200 genes is summarized in Supplementary TableS5. This list is enriched for genes that affect diseases of the stomach(Supplementary Table S6).

Characterization of DNA structural variationsCompared with exome sequencing, an advantage of WGS is its

ability to identify SVs. We focused our analysis on tumor-specificSVs (somatic events), and they are summarized in Fig. 4. Figure 4depicts the numbers of SVs per tumor summarized with respect togene location (Fig. 4A), relation to repeats (Fig. 4B), interchro-mosomal versus intrachromosomal (Fig. 4C), and genetic events(Fig. 4D), and recurrence in multiple tumors (Fig. 4E). Note thathere the complex type also contained some interchromosomalSVs. We noted a wide range of SVs across the tumors. Suchvariation in frequency of SVs among different tumors is consistent

with the presence of two types of cancers: high and low genomicinstability; an observation we previously reported in anotherseries of ESCC tumors from this same high-risk region evaluatedby LOH and CNV (25).

To validate SVs identified from the WGS data, we selectedSVs with junction sequences derived from two different genes,but without containing DNA repeats. We successfully PCR-amplified 12 SV junction fragments and sequenced all ofthem by the Sanger method. All 12 SV junction sequenceswere confirmed, that is, they are identical to the WGS data(data not shown).

SVs that recur inmultiple cases are of greatest interest. The genesacross breakpoints are summarized in terms of frequency oftumors affected (Fig. 4E). Those affecting at least five tumors arealso shown (Fig. 4E), and there are 14 of those genes with SVsoccurring in at least five tumors. The details of SVs (only deletionsare shown here that span exonic regions) are shown forMACROD2, FHIT, and PARK2 (Fig. 5). Four tumors deleted exon6 of MACROD2 (NM_080676), which removed amino acids140–180 and also caused frame-shift (Fig. 5A). The change likelyresulted in a loss-function mutation ofMACROD2. Seven tumorsdeleted exon 5 of FHIT (NM_002012), which removed the first 35amino acids and also shifted the reading frame (Fig. 5B). Thedeletion likely inactivated the protein. The locations and sizes ofPARK2 deletions varied (Fig. 5C). Two removed exons 3 and 4(NM_004562), two removed exon 2, one deleted exon 4, and one

Figure 3.Summary of potential cancer drivermutations. The potential drivermutations were identified using IVA.We identified 24 genes that werefrequently mutated in ESCC andgastric cancer. Mutations and affectedtumors are shown in the heatmap.

Somatic Alterations in Esophageal and Gastric Cancer

www.aacrjournals.org Cancer Res; 76(7) April 1, 2016 1719

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

EC8413

EC1475

EC1415

EC0379

CC5615

CC1791

CC1730

CC1649

CC1553

CC1220

CC0996

BC5439

BC5430

BC1684

BC1586

EC8413

EC1475

EC1415

EC0379

CC5615

CC1791

CC1730

CC1649

CC1553

CC1220

CC0996

BC5439

BC5430

BC1684

BC1586

EC8413

EC1475

EC1415

EC0379

CC5615

CC1791

CC1730

CC1649

CC1553

CC1220

CC0996

BC5439

BC5430

BC1684

BC1586

EC8413

EC1475

EC1415

EC0379

CC5615

CC1791

CC1730

CC1649

CC1553

CC1220

CC0996

BC5439

BC5430

BC1684

BC1586

0 50 100 150 200

0 50 100 150

0 50 100 150 200

Tum

orTu

mor

Tum

orTu

mor

Count

Count Count

Count

0 100 200 300 400

Gene.class

Bo1h

None

One

Repeat.class

Bo1h

None

One

InterchromosomalInterchromosomal

Interchromosomal

Inversion

Interchromosomal

Type

Complex

Deletion

Duplication

Recurrent SVs

SV breakpoints contained within genes SV breakpoints contained within repeats

Interchromosomal vs Intrachromosomal SVs Event types of SVs

20

15

10

5

0

3 4 5 7 8 9 106

The number of tumors with SV for the afected gene

The

num

ber

of g

enes

that

con

tain

SV

s in

mul

tiple

tum

ors

A

C

E

D

B

Figure 4.Characterization of somatic SVs in gastric cancer and ESCC genomes. A, bar graph of SVs with respect to whether breakpoints are located with genes or not. Thecount on x-axis refers to the number of SVs. B, bar graphs of SVs with respect to whether breakpoints are located with repeats or not. C, bar graph ofinterchromosomal versus intrachromosomal SVs. D, bar graph of SVs classified by the processes that generated these SVs. E, bar graph of recurrent SVs. Samplecount refers to the number of tumors affected by the SVs, and gene count refers to the number of genes that showed recurrent SVs for a given number ofsample count.

Hu et al.

Cancer Res; 76(7) April 1, 2016 Cancer Research1720

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

Figure 5.IGV views of structural changes of recurrent SVs. IGV views forMACROD2, FHIT, and PARK2. Blue rectangles are exons. Red rectangles are deleted regions, and thenumber below the box refers to SV id. Tumors are labeled on the left. A, MACROD2; B, FHIT; C, PARK2.

Somatic Alterations in Esophageal and Gastric Cancer

www.aacrjournals.org Cancer Res; 76(7) April 1, 2016 1721

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

deleted exons 2–4. The latter three deletions also caused readingframe shift. These deletions likely also inactivated the protein.

To seek additional supporting evidence for these SVs, weanalyzed 39 tumors from the TCGA gastric cancer dataset thathad high coverage WGS data, which is the subset of the 441STAD tumors with copy number data that will be discussedbelow. We used BreakDancer to identify SVs. All three genes,MACROD2, FHIT, and PARK2, contained deletions involvingcoding exons with 16, 9, and 6 tumors having deletions inMACROD2, FHIT, and PARK2, respectively. Since these SVsare caused by deletions, we reviewed the deletion analysisreports from TCGA gastric cancer data performed by the BroadGDAC team, which has 441 STAD tumor samples analyzed.FHIT, MACROD2, and PARK2 were in the 6th, 7th, and 12thmost significantly deleted regions (http://gdac.broadinstitute.org/runs/analyses__2014_10_17/reports/cancer/STAD-TP/CopyNumber_Gistic2/nozzle.html).

DiscussionWGS data provide a unique opportunity to combine SNV with

large SVs and to perform an integrated analysis. There are threemajor findings from this study. First, A>C mutations were com-mon in GCA and GNCA. In addition to the 5 prime A noted inprevious studies of esophageal adenocarcinoma and gastric ade-nocarcinoma (14, 16, 17), we found enrichment of 5 prime T,whichwas weakly associated with ESCC. Second, we identified 24driver mutations, including a subset that has not been previouslyreported. Third, we identified recurrent chromosome alterationsthat occurred in at least 30% of tumors in 14 genes, includingCAMK1D, MACROD2, ANKRD30BL, FHIT, KCNB2, and PARK2.

A>C mutation rates are often low (range 2.6%–6.6% of allmutations; ref.23). However, a recent exome sequencing study ofesophageal cancer found higher A>C mutation rates in EAC thanin ESCC (14). Another recent study of EAC also found a high A>Cmutation rate, particularly with the 5 prime A (16). Our studyextended this observation to gastric adenocarcinoma. We foundhigh A>C mutations in both GCA and GNCA. In addition, wefound the enrichment of A>C mutations with the prime T inESCC. The presence of A>Cmutations in GCA, GNCA, and ESCCsuggests the oxidation of guanine as a potential mutagen forgastric cancer and ESCC. In cancer cells under high oxidativestress, 8-oxo-dGTP accumulating in the DNA can result in G>T,which is equivalent to C>A. It is conceivable that a similarmechanism may contribute to the observed high A>C mutationrate.

A major interest of this study was to identify recurrent SVs. Wefound 14 genes with SVs occurring in at least five tumors (Fig. 4).The most frequent SVs appeared as deletions. Deletions thatremoved coding sequences were of greatest interest. Three exam-ples of these genes are shown in Fig. 5. These deletions wereclustered in small genomic regions. In the case of MACROD2,

exon 6 ofMACROD2 (NM_080676), spanning amino acids 140–180,was deleted. Thedeletion also caused a frame-shift. Similarly,exon 5 of FHIT (NM_002012), containing the first 35 aminoacids, was deleted in multiple tumors. The mutation also shiftedthe reading frame. In the case of PARK2, the deletions were moreheterogeneous, and varied from deletions involving one to threeexons of exons 2, 3, and 4 (NM_004562). These deletions werevery frequent, affecting about 50%of gastric cancer samples. Theywere supported by deletion events observed in gastric tumorsfrom the TCGA dataset. It will be interesting to develop deletion-specific assays to screen a large number of gastric cancers and toassociate the deletions with clinical phenotypes.

Our studies provide new insights into understanding the geno-mic landscape, genome instability, andmutationmechanisms fordeveloping gastric cancer and ESCC. A limitation of the currentstudy is its small sample size. Future studies should includevalidation of these findings in a larger panel of tumors that areselected from multiple tumor types and use combinations ofexome sequencing, RNA-seq analysis, and target-resequencing forboth SNVs and SVs.

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

Authors' ContributionsConception and design: N. Hu, C.C. Abnet, N.D. Freedman, S.M. Dawsey,A.M. Goldstein, P.R. Taylor, M.P. LeeDevelopment of methodology: N. Hu, H. Wu, P.R. Taylor, M.P. LeeAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): N. Hu, M. Kadota, N.D. Freedman, S. Gere,A. Hutchinson, G. Song, T. Ding, Y.-L. Qiao, J. Koshiol, A.M. Goldstein, P.R.Taylor, M.P. LeeAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): H. Liu, C.C. Abnet, H.H. Yang, C. Yan, P.R. Taylor,M.P. LeeWriting, review, and/or revision of the manuscript: N. Hu, C.C. Abnet, H. Su,N.D. Freedman, L. Wang, J. Koshiol, S.M. Dawsey, A.M. Goldstein, P.R. Taylor,M.P. LeeAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): N. Hu, M. Kadota, H. Su, C. Wang, L. Wang,Y. Wang, T. Ding, Y.-L. Qiao, C.G.C. Giffen, P.R. Taylor, M.P. LeeStudy supervision: N. Hu, P.R. Taylor, M.P. LeeOther (validation of the results): H. SuOther (validation the results):L. Wang

Grant SupportThisworkwas supported by the Intramural Research Programof theNIHand

the National Cancer Institute, Division of Cancer Epidemiology and Genetics(DCEG), and Center for Cancer Research (CCR).

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received February 9, 2015; revised October 19, 2015; accepted December 8,2015; published OnlineFirst February 8, 2016.

References1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of

worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer2010;127:2893–917.

2. Ke L. Mortality and incidence trends from esophagus cancer in selectedgeographic areas of China circa 1970–90. Int J Cancer 2002;102:271–4.

3. Li JY. Epidemiology of esophageal cancer in China. Natl Cancer InstMonogr 1982;62:113–20.

4. Liu SF, Shen Q, Dawsey SM, Wang GQ, Nieberg RK, Wang ZY, et al.Esophageal balloon cytology and subsequent risk of esophageal andgastric-cardia cancer in a high-risk Chinese population. Int J Cancer1994;57:775–80.

5. Abnet CC, Freedman ND, Hu N, Wang Z, Yu K, Shu XO, et al. A sharedsusceptibility locus in PLCE1 at 10q23 for gastric adenocarcinoma andesophageal squamous cell carcinoma. Nat Genet 2010;42:764–7.

Cancer Res; 76(7) April 1, 2016 Cancer Research1722

Hu et al.

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

6. Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, et al. Anintegrated genomic analysis of human glioblastoma multiforme. Science2008;321:1807–12.

7. Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA, Yuan W, et al. IDH1and IDH2 mutations in gliomas. N Engl J Med 2009;360:765–73.

8. Mardis ER, Ding L, Dooling DJ, Larson DE, McLellan MD, Chen K, et al.Recurring mutations found by sequencing an acute myeloid leukemiagenome. N Engl J Med 2009;361:1058–66.

9. Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara K, BermanBP, et al. Identificationof aCpG islandmethylator phenotype that defines adistinct subgroup of glioma. Cancer Cell 2010;17:510–22.

10. Tomlins SA, Rhodes DR, Perner S, Dhanasekaran SM, Mehra R, Sun XW,et al. Recurrent fusion of TMPRSS2 and ETS transcription factor genes inprostate cancer. Science 2005;310:644–8.

11. Soda M, Choi YL, Enomoto M, Takada S, Yamashita Y, Ishikawa S, et al.Identification of the transforming EML4-ALK fusion gene in non-small-celllung cancer. Nature 2007;448:561–6.

12. Wang K, Kan J, Yuen ST, Shi ST, Chu KM, Law S, et al. Exome sequencingidentifies frequent mutation of ARID1A in molecular subtypes of gastriccancer. Nat Genet 2011;43:1219–23.

13. Zang ZJ, Cutcutache I, Poon SL, Zhang SL,McPherson JR, Tao J, et al. Exomesequencing of gastric adenocarcinoma identifies recurrent somatic muta-tions in cell adhesion and chromatin remodeling genes. Nat Genet2012;44:570–4.

14. Agrawal N, Jiao Y, Bettegowda C, Hutfless SM, Wang Y, David S, et al.Comparative genomic analysis of esophageal adenocarcinoma and squa-mous cell carcinoma. Cancer Discov 2012;2:899–905.

15. Nagarajan N, Bertrand D, Hillmer AM, Zang ZJ, Yao F, Jacques PE, et al.Whole-genome reconstruction andmutational signatures in gastric cancer.Genome Biol 2012;13:R115.

16. Dulak AM, Stojanov P, Peng S, LawrenceMS, Fox C, Stewart C, et al. Exomeand whole-genome sequencing of esophageal adenocarcinoma identifiesrecurrent driver events and mutational complexity. Nat Genet2013;45:478–86.

17. Wang K, Yuen ST, Xu J, Lee SP, Yan HH, Shi ST, et al. Whole-genomesequencing and comprehensive molecular profiling identify new drivermutations in gastric cancer. Nat Genet 2014;46:573–82.

18. Song Y, Li L, Ou Y, Gao Z, Li E, Li X, et al. Identification of genomicalterations in oesophageal squamous cell cancer. Nature 2014;509:91–5.

19. Lin DC, Hao JJ, Nagata Y, Xu L, Shang L, Meng X, et al. Genomic andmolecular characterization of esophageal squamous cell carcinoma. NatGenet 2014;46:467–73.

20. Gao YB, Chen ZL, Li JG, Hu XD, Shi XJ, Sun ZM, et al. Geneticlandscape of esophageal squamous cell carcinoma. Nat Genet 2014;46:1097–102.

21. Zhang L, ZhouY, ChengC, CuiH, Cheng L, Kong P, et al. Genomic analysesreveal mutational signatures and frequently altered genes in esophagealsquamous cell carcinoma. Am J Hum Genet 2015;96:597–611.

22. Bass AJ, Lawrence MS, Brace LE, Ramos AH, Drier Y, Cibulskis K, et al.Genomic sequencing of colorectal adenocarcinomas identifies a recurrentVTI1A-TCF7L2 fusion. Nat Genet 2011;43:964–8.

23. Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P, et al. Coresignaling pathways in human pancreatic cancers revealed by global geno-mic analyses. Science 2008;321:1801–6.

24. Li M, Zhao H, Zhang X, Wood LD, Anders RA, Choti MA, et al. Inactivatingmutations of the chromatin remodeling gene ARID2 in hepatocellularcarcinoma. Nat Genet 2011;43:828–9.

25. Hu N, Wang C, Ng D, Clifford R, Yang HH, Tang ZZ, et al. Genomiccharacterization of esophageal squamous cell carcinoma from a high-riskpopulation in China. Cancer Res 2009;69:5908–17.

www.aacrjournals.org Cancer Res; 76(7) April 1, 2016 1723

Somatic Alterations in Esophageal and Gastric Cancer

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338

2016;76:1714-1723. Published OnlineFirst February 8, 2016.Cancer Res   Nan Hu, Mitsutaka Kadota, Huaitian Liu, et al.   Squamous Cell Carcinoma and Gastric CancerGenomic Landscape of Somatic Alterations in Esophageal

  Updated version

  10.1158/0008-5472.CAN-15-0338doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://cancerres.aacrjournals.org/content/suppl/2016/02/06/0008-5472.CAN-15-0338.DC1

Access the most recent supplemental material at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/76/7/1714.full#ref-list-1

This article cites 25 articles, 5 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/76/7/1714.full#related-urls

This article has been cited by 2 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

  .pubs@aacr.org

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://cancerres.aacrjournals.org/content/76/7/1714To request permission to re-use all or part of this article, use this link

on July 23, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 8, 2016; DOI: 10.1158/0008-5472.CAN-15-0338