Genomic profiling of 766 cancer-related genes in archived esophageal normal

and carcinoma tissues

Jing Chen1, Liping Guo2, Daniel A. Peiffer1, Lixin Zhou1, Owen Tsan Mo Chan3, Marina Bibikova1,Eliza Wickham-Garcia1, Shih-Hsin Lu2, Qimin Zhan2, Jessica Wang-Rodriguez3, Wei Jiang2,4* and Jian-Bing Fan1*

1Illumina Inc., San Diego, CA 921212Cancer Institute, Chinese Academy of Medical Sciences, Beijing, China3The VA Medical Center, UCSD, San Diego, CA 921614The Burnham Institute for Medical Research, La Jolla, CA 92037

We employed the BeadArrayTM technology to perform a geneticanalysis in 33 formalin-fixed, paraffin-embedded (FFPE) humanesophageal carcinomas, mostly squamous-cell-carcinoma (ESCC),and their adjacent normal tissues. A total of 1,432 single nucleo-tide polymorphisms (SNPs) derived from 766 cancer-related geneswere genotyped with partially degraded genomic DNAs isolatedfrom these samples. This directly targeted genomic profiling iden-tified not only previously reported somatic gene amplifications(e.g., CCND1) and deletions (e.g., CDKN2A and CDKN2B) but alsonovel genomic aberrations. Among these novel targets, the mostfrequently deleted genomic regions were chromosome 3p(including tumor suppressor genes FANCD2 and CTNNB1) andchromosome 5 (including tumor suppressor gene APC). The mostfrequently amplified genomic region was chromosome 3q (con-taining DVL3, MLF1, ABCC5, BCL6, AGTR1 and known onco-genes TNK2, TNFSF10, FGF12). The chromosome 3p deletionand 3q amplification occurred coincidently in nearly all of theaffected cases, suggesting a molecular mechanism for the genera-tion of somatic chromosomal aberrations. We also detected signifi-cant differences in germline allele frequency between the esopha-geal cohort of our study and normal control samples from theInternational HapMap Project for 10 genes (CSF1, KIAA1804,IL2, PMS2, IRF7, FLT3, NTRK2, MAP3K9, ERBB2 andPRKAR1A), suggesting that they might play roles in esophagealcancer susceptibility and/or development. Taken together, ourresults demonstrated the utility of the BeadArray technology forhigh-throughput genetic analysis in FFPE tumor tissues and pro-vided a detailed genetic profiling of cancer-related genes in humanesophageal cancer.' 2008 Wiley-Liss, Inc.

Key words: genomic profiling; chromosomal aberration; FFPE;esophageal cancer; cancer susceptibility

Cancer is the result of a series of genetic or epigenetic changes,1

including aneuploidy, multiple gene amplifications, deletions andtranslocations.2 These genetic instabilities are caused by eitherinherited mutations in genes that monitor genome integrity ormutations that are acquired in somatic cells during tumor develop-ment. Environmental risk factors and individual cancer geneticsusceptibilities could contribute to tumor development and pro-gression by facilitating the inactivation or loss of tumor suppressorgenes and by favoring the activation or amplification of onco-genes.3 Thus, comprehensive analysis of genetic alterations intumors and identification of genes involved in tumorigenesis hasbeen a major focus of cancer research.

Human esophageal cancer is one of the most common fatal can-cers worldwide with a 5-year survival rate of less than 10%.4 Thehigh incidences of the disease have been reported in certain areasof China, Japan, Iran, France, Italy and South African. In LinxianCounty, Henan province of China, for instance, the age-adjustedmortality rates for esophageal cancer have been reported as 150/100,000 for men and 115/100,000 for women.5 Although environ-mental and nutritional factors as well as cultural habits are thoughtto play important roles in esophageal carcinogenesis, multiplegenetic alterations associated with the disease have beendescribed. These include frequent amplification and over-expres-sion of the cellular proto-oncogenes encoding epidermal growthfactor receptor (EGFR), c-MYC and cyclin D1, loss of and/or

mutation in tumor suppressor genes (e.g., p53, Rb and p16) anddeath pathway genes (e.g., FAS and FAS ligand).6–9 A recentreport discussed frequent copy number aberrations in esophagealsquamous cell carcinomas10,11; and although several genomicregions were identified, further work is needed to reveal the directcorresponding gene(s).

We have developed a flexible, accurate and high-throughputsingle nucleotide polymorphism (SNP) genotyping system forlarge-scale genetic analysis.12 It includes a miniaturized BeadAr-ray platform and a highly multiplexed SNP genotyping assay(GoldenGate1 assay).13 Since the GoldenGate assay interrogatesonly �40–50 bp of sequence surrounding a SNP of interest, theassay can tolerate a certain degree of DNA degradation and allowsreliable genotyping and targeted genomic profiling with partiallydegraded, low-quality DNA from FFPE tissues.14,15 Formalin-fixed archival tissues represent an invaluable resource for geneticanalysis in cancer, as they are the most widely available materialsfor which patient outcomes are known. The ability to performgenetic analysis in these samples will enable both prospective andretrospective studies, and should greatly facilitate research in cor-relating genetic profiles with clinical outcomes. In our study, weused the GoldenGate technology to analyze genomic DNAs(gDNA) extracted from 33 pairs of FFPE archived cancerous andmatched normal adjacent tissues with 1,432 SNPs derived from766 well-characterized cancer-related genes. We detected bothnovel and previously reported gene deletions/amplifications. Fur-thermore, we found SNPs showing significantly different allelefrequency distribution between the esophageal cohort of our studyand normal control samples from the International HapMap Pro-ject, suggesting that they may play a role in esophageal carcino-genesis.

Materials and methods

Tissue sample acquisition, medical data collection and DNAextraction

The study was approved by the Institutional Review Board ofthe Cancer Institute/Hospital, Chinese Academy of Medical Scien-ces. A total of 66 tissue samples from 33 patients diagnosed withesophageal cancer from 1998 to 2000 in Beijing city and Henanprovince, China, were entered into the study (see Supplementary

This article contains supplementary material available via the Internet athttp://www.interscience.wiley.com/jpages/0020-7136/suppmat.Grant sponsor: National Key Basic Research Program of China; Grantnumber: 973-2002BC513101; Grant sponsor: NIH; Grant number:GM67859.*Correspondence to: The Burnham Institute for Medical Research, La

Jolla, CA 92037, USA. Fax: +858-713-6274. E-mail: wjiang@burnham.org or Illumina Inc., San Diego, CA 92121, USA. Fax: +858-202-4680.E-mail: jfan@illumina.com.Received 17 September 2007; Accepted after revision 22 November

2007DOI 10.1002/ijc.23397Published online 1 February 2008 in Wiley InterScience (www.interscience.

wiley.com).

Int. J. Cancer: 122, 2249–2254 (2008)' 2008 Wiley-Liss, Inc.

Publication of the International Union Against Cancer

Table I). This included the 33 esophageal tumor samples (mostlysquamous cell carcinoma) and the 33 matched adjacent normal tis-sues. The samples were FFPE and stored at least for 4 years beforeuse. The patients’ medical data and lifestyle cancer risk factors(e.g., smoking, alcohol drinking and family history of cancer)were documented with informed consent.

To extract gDNA from these FFPE samples, deparaffinizationwas done by adding 800 ll Xylene to each reaction tube containingmultiple 5-lm tissue sections; inverted several times, and centri-fuged at 4,000 rpm for 3 min; removed the supernatant from thetube; added 800 ll Xylene and 400 ll EtOH (100%) to each tube,inverted several times, centrifuged at 4,000 rpm for 3 min; removedthe supernatant and added 1 ml ETOH to each tube, inverted severaltimes, centrifuged at 4,000 rpm for 3 min; removed the supernatantand centrifuged at 14,000 rpm for 1 min; carefully removed all thesupernatant from the tube. Proteinase K treatment was done after the‘‘high pure RNA paraffin kit’’ procedure (Roche, cat no. 3270289).DNA purification was done after the ‘‘high pure PCR template prep-aration kit’’ procedure (Roche, cat no. 1796828).

SNP genotyping on Illumina universal bead arrays

We used a high-throughput SNP genotyping assay describedpreviously.13 Assay probes were designed for 1,432 SNPs from766 cancer related genes (Supplementary Table II). On average,this panel contains a relatively even distribution of genes perChromosomes (n 5 33), with the greatest number of genes resid-ing on Chromosome 1 (n 5 61) and the fewest on Chromosome21 (n 5 11). For each SNP locus, 3 probes were designed: 2 al-lele-specific oligos (ASO) and 1 locus-specific oligo (LSO). TheASOs consisted of 2 parts: the locus-specific sequence and a uni-versal PCR primer sequence at the 50-end. The LSOs consisted of3 parts: the locus-specific sequence, a unique address sequence,which is complementary to a capture sequence immobilized onthe array, and a universal PCR primer sequence (P3) at the 30-end.Assay oligos corresponding to the 1,432 SNPs were pooled andhybridized to the gDNA template. Hybridized ASOs wereextended and ligated to their corresponding LSO to create a PCRtemplate that was amplified subsequently with universal primers(P1, P2 and P30). The PCR products, which were fluorescently la-beled by incorporation of 50-labeled primers P1 (Cy3) and P2(Cy5), were hybridized to capture probes on the beads in the array.The ratio of the fluorescent signals from 2 allele-specific ligation

products was used to determine the genotype. All of the SNPswere assayed on 1 array and 500 ng of gDNA from each individ-ual sample was used for each array experiment. Each individualsample was assayed twice by the array. Standard software devel-oped at Illumina were used for automatic image registration andintensity extraction.16

Copy number detection

Genotyping data consists of 2 channel intensity data corre-sponding to the 2 alleles. Data is generated as rectangular coordi-nates of the raw A versus raw B allele intensities. After normaliza-tion, using Illumina BeadStudio 2.0, the genotyping data weretransformed to a polar coordinate plot of normalized intensity R 5Xnorm 1 Ynorm and allelic intensity ratio u 5 (2/p)* arctan (Ynorm/Xnorm), where Xnorm and Ynorm represent transformed normalizedsignals from alleles A and B for a particular locus. The observednormalized intensity of the subject sample (Rsubject) was comparedto the expected intensity (Rexpected) computed from a linear inter-polation of the observed allelic ratio (usubject) with respect to thecanonical genotype clusters.17 This transformed parameter, thelog2 R ratio [log2(Rsubject/Rexpected)] was analyzed along the entiregenome for all SNPs on the array. In some cases, the log2 R ratiois shown for a given gene and this was calculated by averaging thelog2 R values of all SNPs within the gene of interest.

Allele frequency analysis

The ‘‘fisher test’’ function from R (Version 2.3.0 on i686-red-hat-linux-gnu) was used to perform Fisher’s exact test (for eachSNP) on the allele frequency distribution between the esophagealcancer cohort of our study and the ethnically matched normal pop-ulation (the Han Chinese population from the International Hap-Map study;18). The null hypothesis is that rows and columns in thecontingency table are independent, implying the same AA, ABand BB genotype frequency distribution between the 2 popula-tions. The default parameters for R’s ‘‘fisher test’’ were used toreport p-values.

Results and discussions

Genotyping of exonic SNPs with genomic DNAs extracted fromFFPE human esophageal tissue samples

We employed the highly multiplexed GoldenGate SNP geno-typing assay13 to identify both somatic DNA changes and germ-line genetic loci (i.e., cancer susceptibility genes) from archived

TABLE I – TOP 19 GENES MOST FREQUENTLY DELETED IN THE33 ESCC SAMPLES

Gene Chr PositionNumber of

measured SNP(s)per gene

No. ofindividuals withLRR <20.3

RAP1A 1 111967431 1 7COL4A3 2 228002406 1 7GPX1 3 49370761 1 10FANCD2 3 10115671 2 7EPHA3 3 89611638 2 6IL17RB 3 53858762 3 6CTNNB1 3 41255811 2 5CCNA2 4 123099822 1 7FGF2 4 124172505 3 5ISL1 5 50725852 1 11APC 5 112190753 4 9XRCC4 5 82684733 2 7TGFBI 5 135427272 1 6FER 5 108161866 3 5CDKN2A 9 21957952 3 8CDKN2B 9 21993223 4 8CYLD 16 49385019 1 6FVT1 18 59149381 1 8DCC 18 48121222 2 7

We used a LRR < 20.3 cutoff to call a deletion event and requiredboth technical replicates to meet this cutoff, to make a call for a givenindividual.

TABLE II – TOP 18 GENES MOST FREQUENTLY AMPLIFIED IN THE33 ESCC SAMPLES

Gene Chr PositionNumber of

measured SNP(s)per gene

No. ofindividuals with

LRR > 0.3

DVL3 3 185372039 2 16TNK2 3 197094329 1 12MLF1 3 159771737 1 9ABCC5 3 185120547 3 8BCL6 3 188928913 2 8AGTR1 3 149940340 3 6RARRES1 3 159897868 3 5SKIL 3 171560415 1 4FGF12 3 193343237 3 3MDS1 3 170349822 2 3TNFSF10 3 173706628 3 3TERT 5 1306744 1 3LRRC14 8 145721314 1 3CCND1 11 69172037 3 9PTPN6 12 6926121 1 9JARID1A 12 276553 1 5E2F1 20 31728336 1 6CTAG2 X 153444534 1 5

We used a LRR > 0.3 cutoff to call an amplification event andrequired both technical replicates to meet this cutoff, to make a callfor a given individual.

2250 CHEN ET AL.

human esophageal cancer samples. To this end, we compiled1,432 SNPs from the exonic regions of 766 cancer-related genes(Supplementary Table II). These included (but are not limited to):tumor suppressor genes (e.g., CDKN2A, CDKN2B, BRCA1, APC);oncogenes (e.g., CCND1, ERBB2, EGFR, FGF12, VEGF); genesregulating cell growth and differentiation; and genes locatedwithin published genomic regions subject to deletion or amplifica-tion in cancer. We parsed the latest NCBI RefSeq database andretrieved mapped SNPs in these genes. The SNP collection con-tains both synonymous and nonsynonymous changes, as well asthose that affect splicing.

All 1,432 SNPs were analyzed simultaneously by the Golden-Gate assay using 500 ng gDNA isolated from each of the 33paired FFPE esophageal cancer and adjacent normal tissue sam-ples. The gDNAs isolated from these esophageal FFPE tissueblocks were partially degraded with an average size of 500–700bp (data not shown). We also genotyped gDNAs isolated from 4human lymphoblastoid cell lines as intact gDNA controls. Thegenotyping calls were made automatically, and each call wasassigned a quantitative score that reflects quality. On average,we obtained high call rates in the FFPE cancer samples (97.2%),the FFPE adjacent normal tissue samples (98.6%) and the lym-phoblastoid cell lines (99.7%) (Supplementary Table III); thedifference between FFPE cancer and FFPE normal tissues likelyreflect the fact that more chromosomal aberrations occurred inthe cancer samples, making the determination of a genotype inthose regions difficult. We also measured the concordance of ge-notype calls made from technical replicates, and obtained 98.3,98.0 and 99.9% for these 3 sample groups, respectively (Supple-mentary Table III). These results, together with other recentreports,15 demonstrated that highly accurate genotyping resultscould be obtained from partially degraded gDNAs such as thosederived from FFPE tissue blocks, using the GoldenGate assay.

Somatic chromosomal alterations detected in humanesophageal cancer

We examined the genomic profiles of the cancer-related genesin the human esophageal cancer and the matched adjacent nor-mal tissue samples in detail. We compared each of the 1,432SNPs across all samples and calculated the allele intensities inthe 2 channels (Cy3 and Cy5 channels) to derive both DNAcopy number and allele ratio information using Illumina Bead-

Studio software.17 The log R ratio (LRR) measurement was usedas an indicator of DNA copy number change to detect homozy-gous/hemizygous gene deletions and gene duplication/amplifi-cation in the tumor samples. For instance, a hemizygous dele-tion (loss of 1 copy) would be manifested as a decrease in theLRR from �0 to 20.55 although this depression is typicallyattenuated to approximately 20.4. One copy duplications aremanifested as an increase in the LRR from �0 to 10.40 and thissignal is attenuated as well. Because of the low density of thisarray (average �2 SNP probes per gene), we were unable to av-erage over a large number of probes, which can improve preci-sion of the measurement.17

We detected a wide spectrum of chromosomal aberrations withseveral notable patterns across the esophageal cancer samples. Forexample, we found frequent homozygous/hemizygous deletions ofthe genes on chromosome 3p, including the known tumor suppres-sor genes FANCD2 and CTNNB1, and on chromosome 5, includ-ing the known tumor suppressor gene APC and ISL1, FER,ERCC4 and TGFBI (Table I). We also detected frequent duplica-tion/amplification of the genes on chromosome 3q, including theknown oncogenes (TNK2, TNFSF10 and FGF12) and other genes(BCL6, DVL3, ABCC5, AGTR1, MDS1 and MLF1) (Table II).There was a clear reciprocal correlation between gene deletionson chromosome 3p and gene duplications/amplifications on chro-mosome 3q in most of the affected esophageal cancer samples(Fig. 1), suggesting a molecular mechanism for the generation ofsomatic chromosomal aberrations. Deletion of chromosome 3p orduplication of chromosome 3q, or coupled 3p deletion and 3qduplication have also been observed in lung cancer,19–21 suggest-ing that genetic alterations of genes on chromosome 3 might be acommon molecular mechanism for human epithelial cell carcino-genesis.

High-levels of gene amplification of CCND1 (cyclin D1 gene),ranging from 3–10 fold or higher, were detected in �30% (9/33)of the human esophageal cancer samples (Table II). The amplifica-tion of CCND1 in these tumor samples was further validated byquantitative real-time PCR (qPCR) using a set of CCND1 specificprimers (data not shown). Previously, we and others showed thatCCND1 was frequently amplified in �30–40% of human esopha-geal cancer cell lines and primary tumors using southern blot and/or qPCR analysis.6–8 Our results are consistent with the previousfindings, further demonstrating that amplification of CCND1 is acommon genetic alteration in human esophageal cancer. The

TABLE III – ALLELE FREQUENCY DISTRIBUTION BETWEEN THE ESCC AND THE NORMAL (HAPMAP) POPULATIONS

SNP name Gene AA AB BB Function Protein residueHapMap (normal

controls)ESCC patient

samplesFisher’s test

AA AB BB AA AB BB p-values

rs1058885 CSF1 CC CT TT Nonsynonymous Leu [L]�Pro [P] 0 0 45 7 21 3 9.02 E218rs963982 KIAA1804 AA AG GG Synonymous Ser [S] 0 2 43 2 13 18 2.13 E205rs3087209 IL2 GG GT TT Nonsynonymous Leu [L]�Arg [R] 0 0 45 7 20 6 1.65 E215rs1805321 PMS2 CC CT TT Nonsynonymous Pro [P]�Ser [S] 12 22 9 33 0 0 3.15 E211rs1061501 IRF7 AA AG GG Synonymous Arg [R] 40 5 0 20 8 5 0.002425115rs1933437 FLT3 CC CT TT Nonsynonymous Thr [T]�Met [M] 7 13 25 0 20 13 0.003952709rs2289657 NTRK2 GG GT TT Synonymous IIe [I] 40 4 1 10 17 5 3.36 E207rs3829955 MAP3K9 CC CT TT Synonymous Asn [N] 41 0 1 15 14 4 4.28 E208rs1058808 ERBB2 CC CG GG Nonsynonymous Pro [P]�Ala [A] 31 14 0 8 17 7 2.86 E205rs8080306 PRKAR1A AA AC CC Untranslated N/A 44 1 0 10 22 1 5.32 E211

SNP name Gene Chr SNP seq (NCBI)

rs1058885 CSF1 1 CCAGGCTCTCCCAGGATCTCATCAC[C/T]GCGCCCCCAGGGCCTCAGCAACCCCrs963982 KIAA 1804 1 ATGGCAGGCTTGTGGGAGGACTGGC[A/G]CTTCTGCGGGACTTGGAAGCCCGCTrs3087209 IL2 4 ACACAGCTACAACTGGAGCATTTAC[G/T]KCTGGATTTACAGATGATTTTGAATrs1805321 PMS2 7 CATCTCTGACAAAGGCGTCCTGAGA[C/T]CTCAGAAAGAGGCAGTGAGTTCCAGrs1061501 IRF7 11 GCACGCGTCGCTTCGTGATGCTGCG[A/G]GATAACTCGGGGGACCCGGCCGACCrs1933437 FLT3 13 AAAGTGCTTCATGAATTATTTGGGA[C/T]GGACATAAGGTGCTGTGCCAGAAATrs2289657 NTRK2 13 TTTAACAGACTTGTTTAATCTTCTC[G/T]ATGTAGATGTTTATGTAGGTACTTCrs3829955 MAP3K9 14 CTCCCCTGAGTCCATGTACCCACAA[C/T]CCCCTGGTCAATGTCCGAGTAGAGCrs1058808 ERBB2 17 ACCTGCTGGTGCCACTCTGGAAAGG[C/G]CCAAGACTCTCTCCCCAGGGAAGAArs8080306 PRKAR1A 17 GCGCGGAGAGAGAGCGAAGAGCAGG[A/C]GGAGGAACAAAGGCGACCCAAGACA

2251GENOMIC PROFILING OF ARCHIVED ESOPHAGEAL TUMORS

results also indicate that genomic profiling of partially degradedDNAs from FFPE tissues is highly reliable and accurate using theBeadArray technology.

Aside from CCND1, high-levels of gene amplification of DVL3,TNK2, PTPN6, MLF1, BCL6, ABCC5, AGTR1 and E2F1 werealso frequently detected in the esophageal cancer samples (Table

II). The BCL6 has been shown to be frequently over-expressed inseveral types of human cancers.22 Expression of DVL3 wasreported to be upregulated in human head and neck squamous cellcarcinomas (HNSCC).23 TNK2 (ACK1) has also been found to beamplified in primary tumors and correlated with poor prognosis.24

PTPN6 is a member of the protein tyrosine phosphatase (PTP)

FIGURE 1 – A genome-wide (a) or chromosome-wide (b) LRR profile of esophageal samples. X-axis: assayed genomic loci (SNPs) on allchromosomes (a; 1,432 SNPs/766 genes) or chromosome 3 (b; 3p: 47 SNPs/26 genes; 3q: 46 SNPs/24 genes). Y-axis: all samples, including rep-licates. The LRR is the log2 of the SNP intensity from the tumor sample divided by the intensity from the matched normal sample. An increase(red) in LRR indicates an increase in copy number whereas a decrease (green) in copy number is represented as a decrease in LRR. Black indi-cates no changes. Note the dramatic loss in intensity on chromosome 3p and gain in intensities on chromosome 3q.

FIGURE 2 – Average LRRs across all cancer samples for the 3 CCND1 SNPs, 3 CDKN2A SNPs and 4 CDKN2B SNPs, respectively. Samples,shown on the X-axis, are clustered based upon their LRR profile using a Manhattan hierarchical clustering metric, available within BeadStudio.Note the common pattern within samples on the right, which show an increase in the copy number of CCND1 and a concomitant loss of copynumber on CDKN2A and CDKN2B. Bright green indicates a homozygous deletion.

2252 CHEN ET AL.

family, which is known to be signaling molecules that regulatecell growth and differentiation.25 Over-expression of ABCC5 hasbeen frequently detected in several types of human cancers andfound to be associated with chemotherapeutic resistance.26

Frequent deletions of tumor suppressor genes CDKN2B (p15)and CDKN2A (p16) were observed in a large proportion of theesophageal cancer samples (Fig. 2). These genes locate on chro-mosome 9p21, a locus that is frequently deleted in many humancancers including esophageal carcinomas.27,28 Previously, weshowed that there was a mutual exclusion between CCND1 ampli-fication and RB gene inactivation in human esophageal tumorsindicating that cell cycle control could be abrogated either by lossof Rb or by increased expression of cyclin D1 during esophagealtumor development.7

To determine if there was also any correlation between amplifica-tion of CCND1 and deletions of CDKN2B and CDKN2A, we ana-lyzed the genomic profiles of CCND1 (3 SNPs), CNDN2A (3 SNPs)and CDKN2B (4 SNPs) across the esophageal cancer samples. Asshown in Figure 2, while 5 samples with high-levels of CCND1amplification displayed low levels of gene deletion of CDKN2B and/or CDKN2A, 4 samples with high-levels of CCND1 amplificationshowed homozygous deletions of CDKN2B and/or CDKN2A. Theseresults, in sharp contrast to the reciprocal correlation betweenCCND1 amplification and RB gene inactivation, indicated thatCCND1 amplification and CDKN2B/A deletion in human esopha-geal tumors were not mutually exclusive. CCND1 amplification andCDNK2B/A deletion, which could cause high Cdk4/6 kinase activ-ities, might provide greater growth advantage to tumor cells as previ-ously suggested.29 Coincident amplification of CNND1 and deletionof CNDK2B/A were also observed in HNSCC.30,31

Frequent deletion of ISL1, GPX1, APC, FVT1, DCC, FANCD2,RAP1A, COL4A3, CCNA2 and XRCC4 were also detected in a sig-nificant proportion of esophageal cancer samples (Table I). Amongthem, APC had been implicated in esophageal cancer develop-ment.32 CCNA2 (cyclin A2) promotes both cell cycle G1/S and G2/M transitions; it may have a role in progression of Barrett’s esopha-gus to esophageal adenocarcinoma.33 The detection of frequent dele-tion of FANCD2 gene in the esophageal cancers was also interestingsince it is mutated in the inherited genetic disease, Fanconi anemia(FA). FA is a rare chromosome instability syndrome characterizedby aplastic anemia, cancer susceptibility and cellular hypersensitiv-ity to interstrand DNA crosslinking agents.34 Functional defects inFA pathway were detected in many human cancers and deletion ofFANCD2 in knockout mice caused epithelial cancer.35 Thus, geneticalterations of FANCD2 and FA pathway could play a critical role inhuman esophageal carcinogenesis.

Identification of germline esophageal cancer susceptibility genesusing allele frequency analysis

As discussed above, we generated highly accurate genotypingresults with the FFPE ESCC samples. We next sought to determineif any of the cancer-related genes we assayed could be associatedwith the risk of developing human esophageal cancer. To this end,we exported the genotyping data from the International HapMapProject18 and identified 1,032 overlapping SNPs between the Hap-Map data set and our study. We then compared the allele frequencydistribution between the esophageal cancer cohort (i.e., the cases; n5 33) of our study and ethnically matched normal Han Chinese pop-ulation from the HapMap project (i.e., the controls; n 5 45). Weidentified 10 SNPs derived from 10 genes (CSF1, KIAA1804, IL2,PMS2, IRF7, FLT3, NTRK2, MAP3K9, ERBB2 and PRKAR1A) thatshowed significantly different allele frequency distributions in the 2populations (Table III), based on a Fisher’s exact test. These resultssuggested that these genes might play roles in esophageal cancersusceptibility and/or contribute to esophageal cancer development.Consistent with the notion, these genes have been reported to playcritical roles in regulating cell proliferation, such as regulating cellsignaling (CSF1, KIAA1084, IL2, FLT3, NTRK2, MAP3K9, ERBB2and PRKAR1A) and gene expression (IRF7) in human cells.36–41

Of the 10 SNPs, 1 located in the untranslated region of the gene(PRKAR1A) and 9 located within the coding regions of the genes,which led to either synonymous (KIAA1804, IRF7, NTRK2 andMAP3K9) or nonsynonymous (CSF1, IL2, PMS2, FLT3 andERBB2) changes. Notably, more heterozygotes were detected inthe esophageal cancer cohort (than in the Han Chinese normalcontrol population) for 9 of the 10 SNPs/genes (CSF1, KIAA1804,IL2, IRF7, FLT3, NTRK2, MAP3K9, ERBB2 and PRKAR1A) (Ta-ble III), consistent with the Knudson’s 2-hit theory.42

It was of interest that our analysis also detected an exception,the PMS2 gene, in which more heterozygotes were observed in thenormal population whereas exclusive wild-type homozygoteswere detected in the esophageal cancer cohort (Table III). PMS2,reported previously to be involved in human esophageal carcino-genesis,43 is a component of protein complex required for nucleo-tide mismatch repair function and a potential tumor suppres-sor.44,45 Our results implicated that additional factors might influ-ence the function of PMS2 and thus increase esophageal cancersusceptibility and/or development.

It is worth to point out only 33 cases and 45 controls were ana-lyzed in our study; more samples are needed to further validateour findings. In addition, we would have to validate the genotypesextracted from the HapMap data set, even though genotyping ac-curacy of the entire HapMap data was estimated at 99.99%.18

Except for PMS243 and ERBB2,46 the genes we identified have yetbeen implicated in human esophageal carcinogenesis. Thus, morefunctional studies of all the alleles in large number of samples arerequired to test their roles as the risk factors/biomarkers for esoph-ageal cancer in the future.

In summary, we have employed the BeadArray technology toperform high-throughput genetic analyses in 33 FFPE archivedhuman esophageal carcinomas and their marched adjacent normaltissues. Our results indicate that highly accurate and reliable geno-typing and genomic profiling results can be obtained with partiallydegraded gDNA derived from FFPE archived tissues blocks. Thedetailed analysis of the genomic profiling in the archived esopha-geal samples enabled us to identify not only previously reportedgenetic alterations (e.g., CCND1 amplification and CDKN2A/Bdeletion) but also novel genetic changes (e.g., BCL-6, DVL3,TNK2, PTPN6, ABCC5 and AGTR1 amplifications and ISL1,GPX1, DCC, FANCD2 and CCNA2 deletions) in the humanesophageal cancer. Importantly, we found that the most frequentlydeleted genomic regions are chromosome 3p and chromosome 5,and the most frequently duplicated or amplified genomic region ischromosome 3q. The strong reciprocal correlation between genedeletion on chromosome 3p and gene duplication/amplification onchromosome 3q in most of the affected human esophageal cancersamples suggests that genetic alterations of genes on chromosome3 might have crucial roles in esophageal carcinogenesis. We alsodetected significant differences of allele frequency distributionsbetween the esophageal cohort of our study and normal controlsamples from HapMap project for 10 genes (CSF1, KIAA1804,IL2, PMS2, IRF7, FLT3, NTRK2, MAP3K9, ERBB2 andPRKAR1A), suggesting they might play some roles in esophagealcancer susceptibility and/or development. Further validation studyof these alleles (genes) in large number of samples will allow abetter understanding of their roles in the etiology and susceptibil-ity of esophageal cancer. Our study, together with the recentreports of use of the BeadArray technology for FFPE sample anal-ysis,14,15 has paved a way for large-scale genome-wide genomicprofiling in archived patient samples including cancer samples toidentify genetic alterations and risk factors associated with clinicaloutcomes retrospectively and prospectively.

Acknowledgements

This work was supported by grants from National Key BasicResearch Program of China (973-2002BC513101) to S.-H.L., Q.Z.and W.J., and from NIH (GM67859) to W.J.

2253GENOMIC PROFILING OF ARCHIVED ESOPHAGEAL TUMORS

References

1. Ponder BA. Cancer genetics. Nature 2001;411:336–41.2. Albertson DG, Collins C, McCormick F, Gray JW. Chromosome

aberrations in solid tumors. Nat Genet 2003;34:369–76.3. Balmain A, Gray J, Ponder B. The genetics and genomics of cancer.

Nat Genet 2003;33 (Suppl):238–44.4. Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide incidence

of eighteen major cancers in 1985. Int J Cancer 1993;54:594–606.5. Yang L, Parkin DM, Li L, Chen Y. Time trends in cancer mortality in

China: 1987–1999. Int J Cancer 2003;106:771–83.6. Jiang W, Kahn SM, Tomita N, Zhang YJ, Lu SH, Weinstein IB.

Amplification and expression of the human cyclin D gene in esopha-geal cancer. Cancer Res 1992;52:2980–3.

7. Jiang W, Zhang YJ, Kahn SM, Hollstein MC, Santella RM, Lu SH,Harris CC, Montesano R, Weinstein IB. Altered expression of thecyclin D1 and retinoblastoma genes in human esophageal cancer.Proc Natl Acad Sci USA 1993;90:9026–30.

8. Shinohara M, Aoki T, Sato S, Takagi Y, Osaka Y, Koyanagi Y,Hatooka S, Shinoda M. Cell cycle-regulated factors in esophagealcancer. Dis Esophagus 2002;15:149–54.

9. Sun T, Miao X, Zhang X, Tan W, Xiong P, Lin D. Polymorphisms ofdeath pathway genes FAS and FASL in esophageal squamous-cellcarcinoma. J Natl Cancer Inst 2004;96:1030–6.

10. Hu N, Wang C, Hu Y, Yang HH, Giffen C, Tang ZZ, Han XY, Gold-stein AM, Emmert-Buck MR, Buetow KH, Taylor PR, Lee MP. Ge-nome-wide association study in esophageal cancer using GeneChipmapping 10K array. Cancer Res 2005;65:2542–6.

11. Hu N, Wang C, Hu Y, Yang HH, Kong LH, Lu N, Su H, Wang QH,Goldstein AM, Buetow KH, Emmert-Buck MR, Taylor PR, et al. Ge-nome-wide loss of heterozygosity and copy number alteration inesophageal squamous cell carcinoma using the affymetrix GeneChipmapping 10 K array. BMC Genomics 2006;7:299.

12. Fan JB, Gunderson KL, Bibikova M, Yeakley JM, Chen J, WickhamGarcia E, Lebruska, Laurent M, Shen R, Barker D. Illumina universalbead arrays. Methods Enzymol 2006;410:57–73.

13. Fan JB, Oliphant A, Shen R, Kermani BG, Garcia F, Gunderson KL,Hansen M, Steemers F, Butler SL, Deloukas P, Galver L, Hunt S,et al. Highly parallel SNP genotyping. Cold Spring Harb Symp QuantBiol 2003;68:69–78.

14. Lips EH, Dierssen JW, van Eijk R, Oosting J, Eilers PH, TollenaarRA, de Graaf EJ, van’t Slot R, Wijmenga C, Morreau H, van WezelT. Reliable high-throughput genotyping and loss-of-heterozygositydetection in formalin-fixed, paraffin-embedded tumors using singlenucleotide polymorphism arrays. Cancer Res 2005;65:10188–91.

15. Oosting J, Lips EH, van Eijk R, Eilers PH, Szuhai K, Wijmenga C,Morreau H, van Wezel T. High-resolution copy number analysis ofparaffin-embedded archival tissue using SNP BeadArrays. GenomeRes 2007;17:368–76.

16. Galinsky VL. Automatic registration of microarray images. II. Hexag-onal grid. Bioinformatics 2003;19:1832–6.

17. Peiffer DA, Le JM, Steemers FJ, Chang W, Jenniges T, Garcia F,Haden K, Li J, Shaw CA, Belmont J, Cheung SW, Shen RM, et al.High-resolution genomic profiling of chromosomal aberrations usingInfinium whole-genome genotyping. Genome Res 2006;16:1136–48.

18. Altshuler D, Brooks LD, Chakravarti A, Collins FS, Daly MJ, Don-nelly P. A haplotype map of the human genome. Nature 2005;437:1299–320.

19. Braga E, Senchenko V, Bazov I, Loginov W, Liu J, Ermilova V,Kazubskaya T, Garkavtseva R, Mazurenko N, Kisseljov F, LermanMI, Klein G, et al. Critical tumor-suppressor gene regions on chromo-some 3P in major human epithelial malignancies: allelotyping andquantitative real-time PCR. Int J Cancer 2002;100:534–41.

20. Massion PP, Kuo WL, Stokoe D, Olshen AB, Treseler PA, Chin K,Chen C, Polikoff D, Jain AN, Pinkel D, Albertson DG, Jablons DM,et al. Genomic copy number analysis of non-small cell lung cancerusing array comparative genomic hybridization: implications of thephosphatidylinositol 3-kinase pathway. Cancer Res 2002;62:3636–40.

21. Tai AL, Mak W, Ng PK, Chua DT, Ng MY, Fu L, Chu KK, Fang Y,Qiang Song Y, Chen M, Zhang M, Sham PC, et al. High-throughputloss-of-heterozygosity study of chromosome 3p in lung cancer using sin-gle-nucleotide polymorphism markers. Cancer Res 2006;66:4133–8.

22. Albagli-Curiel O. Ambivalent role of BCL6 in cell survival and trans-formation. Oncogene 2003;22:507–16.

23. Cromer A, Carles A, Millon R, Ganguli G, Chalmel F, Lemaire F,Young J, Dembele D, Thibault C, Muller D, Poch O, Abecassis J,et al. Identification of genes associated with tumorigenesis and meta-static potential of hypopharyngeal cancer by microarray analysis.Oncogene 2004;23:2484–98.

24. van der Horst EH, Degenhardt YY, Strelow A, Slavin A, Chinn L, OrfJ, Rong M, Li S, See LH, Nguyen KQ, Hoey T, Wesche H, et al. Met-astatic properties and genomic amplification of the tyrosine kinasegene ACK1. Proc Natl Acad Sci USA 2005;102:15901–6.

25. Poole AW, Jones ML. A SHPing tale: perspectives on the regulationof SHP-1 and SHP-2 tyrosine phosphatases by the C-terminal tail.Cell Signal 2005;17:1323–32.

26. Assaraf YG. The role of multidrug resistance efflux transporters inantifolate resistance and folate homeostasis. Drug Resist Updat2006;9:227–46.

27. Hu N, Wang C, Su H, Li WJ, Emmert-Buck MR, Li G, Roth MJ,Tang ZZ, Lu N, Giffen C, Albert PS, Taylor PR, Goldstein AM. Highfrequency of CDKN2A alterations in esophageal squamous cell carci-noma from a high-risk Chinese population. Genes Chromosomes Can-cer 2004;39:205–16.

28. Suzuki H, Zhou X, Yin J, Lei J, Jiang HY, Suzuki Y, Chan T, HannonGJ, Mergner WJ, Abraham JM, Melzer SJ. Intragenic mutations ofCDKN2B and CDKN2A in primary human esophageal cancers. HumMol Genet 1995;4:1883–7.

29. Lukas J, Aagaard L, Strauss M, Bartek J. Oncogenic aberrations ofp16INK4/CDKN2 and cyclin D1 cooperate to deregulate G1 control.Cancer Res 1995;55:4818–23.

30. Okami K, Reed AL, Cairns P, Koch WM, Westra WH, Wehage S, JenJ, Sidransky D. Cyclin D1 amplification is independent of p16 inacti-vation in head and neck squamous cell carcinoma. Oncogene1999;18:3541–5.

31. Park HW, Song SY, Lee TJ, Jeong D, Lee TY. Abrogation of the p16-retinoblastoma-cyclin D1 pathway in head and neck squamous cellcarcinomas. Oncol Rep 2007;18:267–72.

32. Tanaka S, Akiyoshi T, Mori M, Wands JR, Sugimachi K. A novelfrizzled gene identified in human esophageal carcinoma mediatesAPC/beta-catenin signals. Proc Natl Acad Sci USA 1998;95:10164–9.

33. Lao-Sirieix P, Lovat L, Fitzgerald RC. Cyclin A immunocytology asa risk stratification tool for Barrett’s esophagus surveillance. ClinCancer Res 2007;13:659–65.

34. Gurtan AM, D’Andrea AD. Dedicated to the core: understanding theFanconi anemia complex. DNA Repair (Amst) 2006;5:1119–25.

35. Taniguchi T, D’Andrea AD. Molecular pathogenesis of Fanconi ane-mia: recent progress. Blood 2006;107:4223–33.

36. Chitu V, Stanley ER. Colony-stimulating factor-1 in immunity andinflammation. Curr Opin Immunol 2006;18:39–48.

37. Nakagawara A. Trk receptor tyrosine kinases: a bridge between can-cer and neural development. Cancer Lett 2001;169:107–14.

38. Naoe T, Kiyoi H. Normal and oncogenic FLT3. Cell Mol Life Sci2004;61:2932–8.

39. Dhillon AS, Hagan S, Rath O, Kolch W. MAP kinase signalling path-ways in cancer. Oncogene 2007;26:3279–90.

40. Huye LE, Ning S, Kelliher M, Pagano JS. Interferon regulatory factor7 is activated by a viral oncoprotein through RIP-dependent ubiquiti-nation. Mol Cell Biol 2007;27:2910–18.

41. Moasser MM. The oncogene HER2: its signaling and transformingfunctions and its role in human cancer pathogenesis. Oncogene2007;26:6469–87.

42. Knudson AG. Chasing the cancer demon. Annu Rev Genet 2000;34:1–19.

43. Uchida N, Kumimoto H, Nishizawa K, Tokumasu S, Harada H, Shi-mada Y, Ishizaki K. Mismatch repair and microsatellite instability inesophageal cancer cells. Int J Cancer 2001;91:687–91.

44. Jiricny J. The multifaceted mismatch-repair system. Nat Rev Mol CellBiol 2006;7:335–46.

45. Kolodner RD, Marsischky GT. Eukaryotic DNA mismatch repair.Curr Opin Genet Dev 1999;9:89–96.

46. Mimura K, Kono K, Hanawa M, Mitsui F, Sugai H, Miyagawa N, OoiA, Fujii H. Frequencies of HER-2/neu expression and gene amplifica-tion in patients with oesophageal squamous cell carcinoma. Br J Can-cer 2005;92:1253–60.

2254 CHEN ET AL.