Molecular Characterization of a t(9;12)(p21;q13)Balanced Chromosome Translocation in Combinationwith Integrative Genomics Analysis Identifies C9orf14as a Candidate Tumor-Suppressor

Miguel Angel Pujana,1* Anna Ruiz,2{ Celia Badenas,3 Josep-Anton Puig-Butille,3,4 Marga Nadal,2

Mitchell Stark,5 Laia Gomez,1 Joan Valls,1 Xavier Sole,1 Pilar Hernandez,1 Celia Cerrato,6

Irene Madrigal,3 Rafael de Cid,6 Helena Aguilar,1 Gabriel Capella,1 Santiago Cal,7 Michael R. James,5

Graeme J. Walker,5 Josep Malvehy,4 Montserrat Mila,3 Nicholas K. Hayward,5 Xavier Estivill,6 and Susana Puig4

1Bioinformatics and Biostatistics Unit,Translational Research Laboratory,Catalan Institute of Oncology,IDIBELL,L’Hospitalet,Barcelona,Spain2Cancer Research Institute,IDIBELL,L’Hospitalet,Barcelona,Spain3Biochemistry and Molecular Genetics Service,Hospital Cl|Łnic,Barcelona,Spain4Departmentof Dermatology,Melanoma Unit,Hospital Cl|Łnic,Barcelona,Spain5Queensland Institute of Medical Research,Brisbane,Australia6Genes and Disease Program,Center for Genomic Regulation (CRG),University Pompeu Fabra (UPF),Barcelona Biomedical ResearchPark,Barcelona,Spain7Departamento de Bioqu|Łmicay Biolog|Ła Molecular,Facultad deMedicina,Instituto Universitario deOncolog|Ła,UniversidaddeOviedo,Oviedo,Spain

A large number of nevi (LNN) is a high risk phenotypic trait for developing cutaneous malignant melanoma (CMM). In this

study, the breakpoints of a t(9;12)(p21;q13) balanced chromosome translocation were finely mapped in a family with LNN and

CMM. Molecular characterization of the 9p21 breakpoint identified a novel gene C9orf14 expressed in melanocytes disrupted

by the translocation. Integrative analysis of functional genomics data was applied to determine the role of C9orf14 in CMM de-

velopment. An analysis of genome-wide DNA copy number alterations in melanoma tumors revealed the loss of the C9orf14

locus, located proximal to CDKN2A, in approximately one-fourth of tumors. Analysis of gene expression data in cancer cell

lines and melanoma tumors suggests a loss of C9orf14 expression in melanoma tumorigenesis. Taken together, our results indi-

cate that C9orf14 is a candidate tumor-suppressor for nevus development and late stage melanoma at 9p21, a region frequently

deleted in different types of human cancers. This article contains Supplementary Material available at http://www.interscience.

wiley.com/jpages/1045-2257/suppmat. VVC 2006 Wiley-Liss, Inc.

INTRODUCTION

Cutaneous malignant melanoma (CMM) is a

potentially fatal type of skin cancer with increasing

incidence and mortality world wide (Rigel, 1996;

Jemal et al., 2003). A major etiological factor in

the development of CMM is sunlight exposure

(Pho et al., 2006). In addition, epidemiological

studies have revealed that of a number of pheno-

typic traits, the highest risk of developing CMM is

conferred by the presence of a large number of

nevi (LNN) (Swerdlow and Green, 1987; Grob et

al., 1990; Bataille et al., 1996; Briollais et al., 2000).

As with CMM, sunlight exposure is also the

major etiological factor in nevus ontogenesis. How-

ever, the association between sunlight exposure,

CMM, and nevus development is relatively com-

plex. In addition to environmental factors, there

are known and unidentified genetic factors that

contribute to both phenotypes either independ-

ently or in association. Thus, total nevi counts and

nevus density show familial aggregation (Goldgar

et al., 1991; Duffy et al., 1992). Familial aggrega-

{Present address: Department of Molecular Haematology, Insti-tute of Child Health, WC1N 1EH London, UK.Received 15 June 2006; Accepted 2 October 2006

DOI 10.1002/gcc.20396

Published online 10 November 2006 inWiley InterScience (www.interscience.wiley.com).

*Correspondence to: Miguel Angel Pujana, Ph.D., Bioinformaticsand Biostatistics Unit, Translational Research Laboratory, CatalanInstitute of Oncology, IDIBELL, L’Hospitalet, 08907 Barcelona,Spain. E-mail: mapujana@ico.scs.es

Supported by: Fondo de Investigaciones Sanitarias (FIS) of theSpanish Ministry of Health; Grant numbers: 1546-01 and 0019-03;Instituto de Salud Carlos III; Grant number: V2003-REDC03/03-07;U.S. National Cancer Institute; Grant number: CA83115; Fundacio\la Caixa"; Grant number: BM05-254-00; Department of Univer-sities, Research and Information Society; Health Department of theGeneralitat de Catalunya; The Ramon y Cajal Program of the Span-ish Ministry of Education and Science.

VVC 2006 Wiley-Liss, Inc.

GENES, CHROMOSOMES & CANCER 46:155–162 (2007)

tion of LNN and CMM has also been described

(Bahuau et al., 1997; Briollais et al., 2000) although

mutations in known melanoma susceptibility genes

do not always co-segregate with LNN (Puig et al.,

1997; Hayward, 2000). In approximately 30% of

cases, a histological association exists between

CMM and a precursor nevus (Seykora and Elder,

1996; Bevona et al., 2003). Accordingly, CDKN2Aand BRAF mutations have been associated with

the development of both nevus and CMM (Pollock

et al., 2003; Kumar et al., 2004; James et al., 2005).

On the basis of these observations, the existence of

diverse but partially overlapping genetic and mo-

lecular mechanisms for the development of nevus

and CMM has been proposed (Rivers, 2004).

Cytogenetic analysis identified a t(9;12) (p21;q13)

balanced chromosome translocation in a family

with LNN and CMM. Fine mapping of the break-

points identified a novel gene (C9orf14) disruptedby the translocation at 9p21. Molecular characteri-

zation of the translocation breakpoint and integra-

tive analysis of functional genomics data or \inte-grative genomics" indicate that C9orf14 is a candi-

date tumor-suppressor for nevus development and

late stage melanoma.

METHODS

Patients

Study protocol was approved by the Institutional

Review Board of the Hospital Clınic of Barcelona.

All family members gave written informed consent

to participate in the study.

Southern and Pulse-Field Gel Electrophoresis

Analyses

Genomic DNA was extracted from whole blood

using standard protocol. High molecular weight

DNA was purified from agarose-embedded lym-

phocytes by proteinase K digestion and electropho-

resed using a CHEF-DRII Mapper (Bio-Rad).

Conventional alkaline transfer of DNA was used.

PCR products were gel-purified using the QIAgen

Gel Band Purification Kit and then labeled using

the Megaprime DNA Labeling System (GE

Healthcare).

Cytogenetic and Fluorescence In Situ

Hybridization Analyses

Blood cells were cultured and harvested accord-

ing to standard cytogenetic protocols. Karyotypes

are described according to ISCN. For fluorescence

in situ hybridization (FISH) analysis, genomic

clones were purchased from CEPH (France), the

Human Genome Mapping Project Resource

Centre (United Kingdom), Research Genetics

(USA), and the BACPAC Resource Center (USA).

YAC’s inter-Alu PCRs and BAC/PAC DNA mini

preparations (QIAgen) were labeled with either

biotin-16dUTP or digoxigenin-11dUTP (Roche

Applied Science) by a standard nick-translation

reaction. Slides were visualized under a fluores-

cence microscope (AH3, Olympus) with appropri-

ate filters. Images were analyzed using the Cytovi-

sion system (Applied Imaging Ltd.).

DNACopy Number Data Analysis

Data for DNA copy number alterations in pri-

mary melanoma tumors was downloaded from the

Gene Expression Omnibus (GEO) record

GSE2631 (Curtin et al., 2005). This data set con-

tains log2 ratios of DNA analysis of tumor versus

normal tissue samples. The Stats and Graphics

packages in R (www.R-project.org) were used for

clustering analysis.

Microarray Gene Expression Data Analysis

The NCI60 cancer cell lines expression data set

analyzed was GeneLogic Affymetrix U95D (Scherf

et al., 2000), which contains a single C9orf14 probe

(69061_at; AI078569). This data set contains nor-

malized and scaled expression values. The Stats

and Graphics packages in R (www.R-project.org)

were used for clustering analysis. Raw expression

data from normal skin, keratinocytes, melanocytes,

nevi, and melanomas was published elsewhere

(Smith et al., 2005) and deposited in GEO record

GSE4587. Short-term cell cultures were used by

the authors to minimize expression artifacts. We

downloaded the raw data and analyzed it using the

Robust Multi-array Average algorithm (Irizarry

et al., 2003) and the LIMMA package (Smyth,

2004). Differential gene expression between classes

was then assessed using the Student’s t test. The

ovarian expression data set was published else-

where (Lu et al., 2004), downloaded from the MD

Anderson Cancer Center website, and analyzed

using the same methodology. Raw expression data

of different normal human tissue samples was

downloaded from the GEO record GSE3526.

C9orf14 Cloning and Genomic Structure Analysis

EST clones were purchased from the Human

Genome Mapping Project Resource Center

(United Kingdom) and the Resource Center within

the German Human Genome Project (RZPD).

The SL.2NbHM normalized normal melanocyte

cDNA library was purchased from Research

Genes, Chromosomes & Cancer DOI 10.1002/gcc

156 PUJANA ETAL.

Genetics. Human Multiple Tissue Northern

(MTN) and Marathon-Ready cDNAs were pur-

chased from Clontech-Takara. Total RNA from

normal skin biopsies and cell lines (normal human

epidermal melanocytes (PromoCell), HeLa,

HepG2, NP29, and SHY) were extracted with Tri-

Pure Isolation Reagent (Roche Applied Science).

Total RNA was reverse transcribed using random

hexamers and Superscript II (GIBCO-BRL, Invi-

trogen) according to the manufacturer’s instruc-

tions. RT-PCR reactions were carried out with

Advantage-GC 2 Polymerase Mix (Clontech-

Takara) and did not exceed 40 cycles. All RT-PCR

reactions were performed at least twice for each

cDNA sample using GAPDH as positive control

(primers GGTGAAGGTCGGAGTCAACG and

CAAAGTTGTCATGGATGACC). Purification of

small RNAs was carried out using the mirVana

miRNA Isolation Kit (Ambion). Fifty micrograms

of total RNA were then used for purification and

Northern blotting.

C9orf14 Protein Localization Analysis

Protein localization analysis was carried out

using the Gateway system (Invitrogen). Briefly, the

PCR-amplified open reading frame (ORF)

sequence was cloned by site-directed recombina-

tion into the pENTR201 entry vector and then

transferred into the pT-Rex-DEST31 expression

vector. Telomerase-immortalized human mammary

epithelial cells (HME/TERT) (DiRenzo et al.,

2002) growing on glass cover slips were transfected

with Opti-MEM1 I Reduced Serum Medium and

Lipofectamine 2000 (Invitrogen), and then fixed in

cold methanol for 10 min followed by cold acetone

for 1 min. Staining was performed overnight at

48C using a 1:1,000 dilution of an anti g-tubulin(TUBG1) (Sigma). PBS-Tween 0.02% containing

4% fetal bovine serum was used in antibody incu-

bations and PBS-Tween 0.02% was used for

washes. Alexa-596 conjugated anti-mouse second-

ary antibody (Invitrogen, Molecular Probes) was

used for detection.

RESULTS AND DISCUSSION

Cytogenetic analysis identified a t(9;12)

(p21;q13) balanced translocation in three members

of a two generation family. The translocation was

observed in the father, one of the sons (aged 10

years when included in the study), and the daugh-

ter (newborn) (Fig. 1). The father and the son

carrying the translocation had LNN, with >150

nevi. In addition, the son with the translocation

had lentigines on the face, one cafe-au-lait macule

and despite strictly avoiding exposure to sunlight

developed an in situ CMM at the age of 16 years

(Fig. 1). The daughter developed almost 20 nevi at

the age of 5 years, being phototype IV (dark skin

color), while the father and the affected son had

phototype II (light skin color with difficulties to

tan and easily sunburned). The nevi present in

individuals with the translocation were macular

lesions, dark in color with atypical reticulated pat-

tern under dermoscopy, suggestive of lentiginous

junctional or compound atypical nevi. The mother

had a few moles of a globular pattern under dermo-

scopy, suggestive of compound/dermal nevi, and

multiple lentigines and other signs of photo-aging.

The son without the translocation had fewer than

50 moles at the age of 14 years, most of which were

dermal or compound type, being phototype II.

FISH analysis using genomic clones localized

the translocation breakpoints to within RP1-121K10

and betweenRP11-923I11 andRP11-1100L3 clones

at 9p21 and 12q13, respectively (Supplementary

Fig. 1; Supplementary material for this article can

be found at http://www.interscience.wiley.com/

Figure 1. Familial pedigree with a t(9;12)(p21;q12) translocation,LNN, and CMM. Bottom panels show (A) back of the index case (sonof 16 years old, arrow in pedigree), (B) clinical image, (C) CMM dermo-scopy image, and (D) CMM histological analysis image.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

157C9orf14 AS A CANDIDATE TUMOR-SUPPRESSOR

jpages/1045-2257/suppmat). Thus, the breakpoints

were primarily mapped to between the TEK and

C9orf11 genes, and between the ANKRD33 and

ACVRL1 genes at 9p21 and 12q13, respectively.

To focus the study on the translocation breakpoint

that could potentially be the cause of LNN and sub-

sequently CMM, we examined whether genomic

regions on 9p21 or 12q13 are lost in melanoma. For

this analysis, a publicly available data set of genome-

wide DNA copy number alterations in 126 primary

melanomas was used (Curtin et al., 2005). The

results showed no alterations of 12q13 but common

loss of 9p21 including TEK and C9orf11 loci

(Fig. 2A). A total of 30 out of 126 (23%) tumors

showed relative loss of this region located proximal

to the CDKN2A gene. CDKN2A is a tumor-suppressor

gene mutated in melanoma (Serrano et al., 1996).

Different types of genetic and molecular evidence

suggest the existence of an additional tumor-suppres-

sor at 9p21. Mutational analysis of melanoma-prone

families with linkage to 9p21 did not identify altera-

tions in CDKN2A, CDKN2B, or p14ARF in a signifi-

cant proportion of cases (Liu et al., 1997; Laud et al.,

2006). Loss of heterozygosity of a region proximal to

the CDKN2A locus has been described in CMM and

lung carcinoma tumors (Puig et al., 1995; Kim et al.,

1997; Ruiz et al., 1998). Functional chromosome

transfer studies identified distinct regions at 9p21

implicated in the development of CMM (Parris

et al., 1999). Finally, a locus modifying CMM risk

located proximal to CDKN2A was described in mela-

noma families harboring CDKN2A mutations (van

der Velden et al., 1999).

FISH analysis and Southern blotting excluded

the disruption of TEK and C9orf11 genes by the

translocation breakpoint (not shown). The break-

point was then finely mapped by designing South-

ern probes at regular intervals between TEK and

C9orf11 using information taken from the human

genome sequence (PCR primers for probes are

detailed in Supplementary Table 1; Supplementary

material for this article can be found at http://

www.interscience.wiley.com/jpages/1045-2257/supp

mat). The breakpoint was limited to a 1.5 kb

region, at *33 kb proximal and*20 kb distal from

30-TEK and 50-C9orf11 genes, respectively (Fig. 3).

The level of sequence conservation across genomes

of different species is relatively high around the

breakpoint, which suggests the presence of con-

served regulatory or expressed sequences between

TEK and C9orf11 genes.To determine the existence of an unidentified

gene between TEK and C9orf11, three comple-

mentary molecular approaches were used: (i) com-

plete sequencing of public domain ESTclones; (ii)

normal human melanocyte cDNA library screen-

ing; and (iii) RT-PCR reactions using cDNA prepa-

rations from different tissues (including normal

skin tissue), cell types (including normal human

epidermal melanocyte cell cultures) and primers

designed in evolutionarily conserved sequence

regions (Supplementary Table 1). First, complete

sequence of all EST clones in the region was

obtained, with an exon–intron structure found in

only one of them (IMAGE 2275835). This clone

and the overlapping clone IMAGE 1677562 have

polyA+ sequences that define a full transcript of

*1.5 kb in length. A possible ORF within this tran-

script encodes for a protein of 99 amino acids (AA).

Subsequently, melanocyte cDNA library screen-

ing using probes belonging to all ESTs identified a

single polyA+ clone that corresponded exactly to

the 1677562 sequence. Northern analysis of differ-

ent tissues using this sequence as a probe revealed

a complex expression pattern consisting of at least

six different transcripts of up to *6 kb, and main

expressed transcripts of *0.5 and *1.5 kb (Sup-

plementary Fig. 2; Supplementary material for this

article can be found at http://www.interscience.

wiley.com/jpages/1045-2257/suppmat), the latter in

agreement with the expected mRNA size based on

sequencing of ESTs.

Finally, RT-PCR reactions extended the cDNA

sequence distally from clones 2275835 and

1677562. Thus, 13 additional exons were identi-

fied. On the other hand, 50-RACE reactions

extended the cDNA proximally, 208 bp to the 50-end of 2275835 sequences. It was therefore deter-

mined that the complete C9orf14 gene sequence

contains 17 exons and that the translocation break-

point is located in the 10th intron. Searching for

additional functional elements at the C9orf14locus, we identified a small noncoding RNA com-

plementary to the mRNA sequence of C9orf14 rep-

resented by the first two exons, suggesting com-

plex regulation of C9orf14 expression (Supplemen-

tary Fig. 2).

It was then analyzed whether C9orf14 is differ-

entially expressed between cancer cell lines and

between melanoma tumors relative to nevi and

normal melanocytes. Using a publicly available

expression data set containing 60 cancer cell lines

of different tissue origin (NCI60) (Scherf et al.,

2000), differential expression of the AI078569

probe corresponding to the 1677562 sequence was

observed (Fig. 2B). Thus, three (SK-MEL-5 and

28, and MALME-3M) out of seven melanoma cell

lines showed downregulation of C9orf14. No altera-

Genes, Chromosomes & Cancer DOI 10.1002/gcc

158 PUJANA ETAL.

Figure 2. DNA copy number alteration and gene expression analysisof 9p21 and 12q13 loci. A: Clustering of melanoma tumors based onDNA copy number analysis (log2 tumor/normal tissue ratios) of completehuman chromosomes 9 (HSA9) and 12 (HSA12). Sample and clone identi-fications are not shown because of resolution constraints [available inGEO record GSE2631 (Curtin et al., 2005)]. Common loss of 9p21including CDKN2A and TEK-C9orf11 loci is observed. No alterations ofHSA12 are observed except for low-frequency amplification that includesACVRL1 and ACVR1B loci. B: NCI60 cancer cell lines gene expression analy-sis at 9p21 and 12q13 loci. Melanoma cell lines are shown in red. Differen-tial expression (white squares) between melanoma cell lines is shown forAI078569 (IMAGE 1677562 sequence; C9orf14). Two probes match the

CDKN2A gene in this U95 platform: 73130_s_at (U26727) and 73132_r_at(AA909181). As expected, 73130_s_at probe indicates downregulation ofCDKN2A in most melanoma cell lines. However, 73132_r_at probe doesnot correlate with the expected results. This apparent inconsistency isprobably due to the Affymetrix heuristics rules being relaxed (r-suffix) forthis particular probe, which suggests poorer reliability. No comparableexpression differences between melanoma cell lines are observed for12q13 probes. C: C9orf14 expression levels (log2 transformed) in normalskin, melanocytes, keratinocytes, nevi, and melanomas. Results are shownfor Affymetrix U133 Plus 2.0 probe 1563728_at. Downregulation is par-ticularly observed in late stage melanomas.

tions were observed in probes located on 12q13.

These observations could suggest genetic or epige-

netic alterations of the C9orf14 locus associated

with CMM development. Analysis of an expression

data set (GEO record GSE4587) (Smith et al.,

2005) containing data from normal skin, short-term

cultures of epidermal melanocytes and keratino-

cytes, nevi and melanomas, revealed downregula-

tion of C9orf14 in melanomas (P < 0.05) (Fig. 2C).

Specifically, downregulation was observed in late

stage melanomas (vertical growth phase, metastatic

growth phase, and lymph node metastasis).

Downregulation of C9orf14 expression in ovarian

cancer cell lines was also observed in the NCI60

set (OVCAR-5 and 8, and SK-OV-3). Analysis of a

publicly available ovarian expression data set (Lu

et al., 2004) revealed downregulation of C9orf14 in

serous carcinomas relative to normal ovarian sur-

face epithelium (P < 0.05). This observation, to-

gether with data describing the genetic loss of

9p21 in ovarian clear cell carcinomas (Dent et al.,

2003), might also suggest that C9orf14 is involved

in ovarian tumorigenesis.

Next, the role of the C9orf14 encoded protein of

99 AAwas investigated through an analysis of its cel-

lular localization in normal human mammary epithe-

lial cells (HME/TERT) (DiRenzo et al., 2002). Cen-

trosome localization was then revealed by colocaliza-

tion of EmGFP-C9orf14 and g-tubulin (Fig. 4). The

centrosome localization could suggest a role in micro-

Figure 3. Genomic location of the 9p21 translocation breakpoint. Direction of transcription is indicatedby gray arrows inside each gene structure. The ORF (99 AA) is indicated for the first two exons. Level ofsequence conservation across genomes of different species is shown at the bottom.

Figure 4. Cellular localization of the C9orf14 gene product. Results for EmGFP-C9orf14 expression,g-tubulin (TUBG1) immunofluorescence, and DAPI counterstaining are shown in HME/TERT cells. Whiteand yellow arrows indicate the centrosome of a cell expressing EmGFP-C9orf14 and of a cell not express-ing EmGFP-C9orf14, respectively.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

160 PUJANA ETAL.

tubule organization that influences metastatic poten-

tial and chromosome segregation fidelity (Doxsey

et al., 2005). A critical role for the CDKN2A gene

product preventing centrosome dysfunction in pri-

mary cells has recently been reported (McDermott

et al., 2006) and may point to a functional association

with the C9orf14 gene product.Taken together, the molecular characterization

of a chromosome translocation in related individu-

als with LNN and CMM, and integrative genomics

analysis suggests that C9orf14 is a tumor-suppressor

for nevus development, late stage melanoma, and,

maybe, for other types of human cancers. The

strategy developed here can be applied to other

genes in which an underlying molecular mecha-

nism is suspected. Thus, the integration of \omic"data highlights likely candidates with increased

confidence (Mootha et al., 2003; Dahia et al.,

2005). Genetic analysis of melanoma-prone fami-

lies with linkage to 9p21 but without alterations

in CDKN2A, CDKN2B, or p14ARF, together with

population-based association studies, should define

the role of C9orf14 in melanoma susceptibility.

REFERENCES

Bahuau M, Vidaud D, Kujas M, Palangie A, Assouline B,Chaignaud-Lebreton M, Prieur M, Vidaud M, Harpey JP, Lafour-cade J, Caille B. 1997. Familial aggregation of malignant mela-noma/dysplastic naevi and tumours of the nervous system: Anoriginal syndrome of tumour proneness. Ann Genet 40:78–91.

Bataille V, Bishop JA, Sasieni P, Swerdlow AJ, Pinney E, Griffiths K,Cuzick J. 1996. Risk of cutaneous melanoma in relation tothe numbers, types and sites of naevi: A case-control study. Br JCancer 73:1605–1611.

Bevona C, Goggins W, Quinn T, Fullerton J, Tsao H. 2003. Cutane-ous melanomas associated with nevi. Arch Dermatol 139:1620–1624. Discussion 1624.

Briollais L, Chompret A, Guilloud-Bataille M, Bressac-de PailleretsB, Avril MF, Demenais F. 2000. Patterns of familial aggregation ofthree melanoma risk factors: Great number of naevi, light photo-type and high degree of sun exposure. Int J Epidemiol 29:408–415.

Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, KutznerH, Cho KH, Aiba S, Brocker EB, LeBoit PE, Pinkel D, BastianBC. 2005. Distinct sets of genetic alterations in melanoma. NEngl J Med 353:2135–2147.

Dahia PL, Hao K, Rogus J, Colin C, Pujana MA, Ross K, MagoffinD, Aronin N, Cascon A, Hayashida CY, Li C, Toledo SP, StilesCD. 2005. Novel pheochromocytoma susceptibility loci identifiedby integrative genomics. Cancer Res 65:9651–9658.

Dent J, Hall GD, Wilkinson N, Perren TJ, Richmond I, MarkhamAF, Murphy H, Bell SM. 2003. Cytogenetic alterations in ovarianclear cell carcinoma detected by comparative genomic hybridisa-tion. Br J Cancer 88:1578–1583.

DiRenzo J, Signoretti S, Nakamura N, Rivera-Gonzalez R, SellersW, Loda M, Brown M. 2002. Growth factor requirements andbasal phenotype of an immortalized mammary epithelial cell line.Cancer Res 62:89–98.

Doxsey S, McCollum D, Theurkauf W. 2005. Centrosomes in cellu-lar regulation. Annu Rev Cell Dev Biol 21:411–434.

Duffy DL, Macdonald AM, Easton DF, Ponder BA, Martin NG.1992. Is the genetics of moliness simply the genetics of sun expo-sure? A path analysis of nevus counts and risk factors in Britishtwins. Cytogenet Cell Genet 59:194–196.

Goldgar DE, Cannon-Albright LA, Meyer LJ, Piepkorn MW, ZoneJJ, Skolnick MH. 1991. Inheritance of nevus number and sizein melanoma and dysplastic nevus syndrome kindreds. J NatlCancer Inst 83:1726–1733.

Grob JJ, Gouvernet J, Aymar D, Mostaque A, Romano MH, ColletAM, Noe MC, Diconstanzo MP, Bonerandi JJ. 1990. Count ofbenign melanocytic nevi as a major indicator of risk for nonfami-lial nodular and superficial spreading melanoma. Cancer 66:387–395.

Hayward N. 2000. New developments in melanoma genetics. CurrOncol Rep 2:300–306.

Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, Speed TP.2003. Summaries of Affymetrix GeneChip probe level data.Nucleic Acids Res 31:e15.

James MR, Roth RB, Shi MM, Kammerer S, Nelson MR, Stark MS,Dumenil T, Montgomery GW, Hayward NK, Martin NG, BraunA, Duffy DL. 2005. BRAF polymorphisms and risk of melano-cytic neoplasia. J Investig Dermatol 125:1252–1258.

Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. 2003.Cancer statistics, 2003. CA Cancer J Clin 53:5–26.

Kim SK, Ro JY, Kemp BL, Lee JS, Kwon TJ, Fong KM, Sekido Y,Minna JD, Hong WK, Mao L. 1997. Identification of three dis-tinct tumor suppressor loci on the short arm of chromosome 9 insmall cell lung cancer. Cancer Res 57:400–403.

Kumar R, Angelini S, Snellman E, Hemminki K. 2004. BRAF muta-tions are common somatic events in melanocytic nevi. J InvestigDermatol 122:342–348.

Laud K, Marian C, Avril MF, Barrois M, Chompret A, GoldsteinAM, Tucker MA, Clark PA, Peters G, Chaudru V, Demenais F,Spatz A, Smith MW, Lenoir GM, Bressac-de Paillerets B. 2006.Comprehensive analysis of CDKN2A (p16INK4A/p14ARF) andCDKN2B genes in 53 melanoma index cases considered to be atheightened risk of melanoma. J Med Genet 43:39–47.

Liu L, Goldstein AM, Tucker MA, Brill H, Gruis NA, Hogg D,Lassam NJ. 1997. Affected members of melanoma-prone familieswith linkage to 9p21 but lacking mutations in CDKN2A do notharbor mutations in the coding regions of either CDKN2B orp19ARF. Genes Chromosomes Cancer 19:52–54.

Lu KH, Patterson AP, Wang L, Marquez RT, Atkinson EN,Baggerly KA, Ramoth LR, Rosen DG, Liu J, Hellstrom I, SmithD, Hartmann L, Fishman D, Berchuck A, Schmandt R, WhitakerR, Gershenson DM, Mills GB, Bast RC, Jr. 2004. Selection ofpotential markers for epithelial ovarian cancer with gene expres-sion arrays and recursive descent partition analysis. Clin CancerRes 10:3291–3300.

McDermott KM, Zhang J, Holst CR, Kozakiewicz BK, Singla V,Tlsty TD. 2006. p16(INK4a) prevents centrosome dysfunctionand genomic instability in primary cells. PLoS Biol 4:e51.

Mootha VK, Lepage P, Miller K, Bunkenborg J, Reich M, HjerrildM, Delmonte T, Villeneuve A, Sladek R, Xu F, Mitchell GA,Morin C, Mann M, Hudson TJ, Robinson B, Rioux JD, LanderES. 2003. Identification of a gene causing human cytochrome coxidase deficiency by integrative genomics. Proc Natl Acad SciUSA 100:605–610.

Parris CN, Harris JD, Griffin DK, Cuthbert AP, Silver AJ, NewboldRF. 1999. Functional evidence of novel tumor suppressor genesfor cutaneous malignant melanoma. Cancer Res 59:516–520.

Pho L, Grossman D, Leachman SA. 2006. Melanoma genetics: Areview of genetic factors and clinical phenotypes in familial mela-noma. Curr Opin Oncol 18:173–179.

Pollock PM, Harper UL, Hansen KS, Yudt LM, Stark M, RobbinsCM, Moses TY, Hostetter G, Wagner U, Kakareka J, Salem G,Pohida T, Heenan P, Duray P, Kallioniemi O, Hayward NK, TrentJM, Meltzer PS. 2003. High frequency of BRAF mutations innevi. Nat Genet 33:19–20.

Puig S, Ruiz A, Castel T, Volpini V, Malvehy J, Cardellach F, LynchM, Mascaro JM, Estivill X. 1997. Inherited susceptibility to sev-eral cancers but absence of linkage between dysplastic nevus syn-drome and CDKN2A in a melanoma family with a mutation inthe CDKN2A (p16INK4A) gene. Hum Genet 101:359–364.

Puig S, Ruiz A, Lazaro C, Castel T, Lynch M, Palou J, Vilalta A,Weissenbach J, Mascaro JM, Estivill X. 1995. Chromosome 9pdeletions in cutaneous malignant melanoma tumors: The minimaldeleted region involves markers outside the p16 (CDKN2) gene.Am J Hum Genet 57:395–402.

Rigel DS. 1996. Malignant melanoma: Perspectives on incidenceand its effects on awareness, diagnosis, and treatment. CA CancerJ Clin 46:195–198.

Rivers JK. 2004. Is there more than one road to melanoma? Lancet363:728–730.

Ruiz A, Puig S, Lynch M, Castel T, Estivill X. 1998. Retention ofthe CDKN2A locus and low frequency of point mutations in pri-mary and metastatic cutaneous malignant melanoma. Int J Cancer76:312–316.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

161C9orf14 AS A CANDIDATE TUMOR-SUPPRESSOR

Scherf U, Ross DT, Waltham M, Smith LH, Lee JK, Tanabe L,Kohn KW, Reinhold WC, Myers TG, Andrews DT, Scudiero DA,Eisen MB, Sausville EA, Pommier Y, Botstein D, Brown PO,Weinstein JN. 2000. A gene expression database for the molecularpharmacology of cancer. Nat Genet 24:236–244.

Serrano M, Lee H, Chin L, Cordon-Cardo C, Beach D, DePinhoRA. 1996. Role of the INK4a locus in tumor suppression and cellmortality. Cell 85:27–37.

Seykora J, Elder D. 1996. Dysplastic nevi and other risk markers formelanoma. Semin Oncol 23:682–687.

Smith AP, Hoek K, Becker D. 2005. Whole-genome expressionprofiling of the melanoma progression pathway reveals marked

molecular differences between nevi/melanoma in situ andadvanced-stage melanomas. Cancer Biol Ther 4:1018–1029.

Smyth GK. 2004. Linear models and empirical Bayes methods forassessing differential expression in microarray experiments. StatAppl Genet Mol Biol 3:1.

Swerdlow AJ, Green A. 1987. Melanocytic naevi and melanoma: Anepidemiological perspective. Br J Dermatol 117:137–146.

van der Velden PA, Sandkuijl LA, Bergman W, Hille ET, FrantsRR, Gruis NA. 1999. A locus linked to p16 modifies melanomarisk in Dutch familial atypical multiple mole melanoma(FAMMM) syndrome families. Genome Res 9:575–580.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

162 PUJANA ETAL.