Aberrant Splicing of the PTPRD Gene MimicsMicrodeletions Identified at this Locusin Neuroblastomas

Prakash Nair,1 Katleen DePreter,2 Jo Vandesompele,2 Frank Speleman,2 and Raymond L. Stallings1*

1Departmentof Pediatrics,Greehey Children’s Cancer Research Institute,The Universityof TexasHealth Science Center at San Antonio,San Antonio,TX 782292Center for Medical Genetics,Ghent University Hospital,Ghent,Belgium

Neuroblastoma (NBL), a pediatric tumor arising from precursor cells of the sympathetic nervous system, is characterized by

numerous recurrent large-scale chromosomal imbalances. High resolution oligonucleotide array CGH analysis of NBL has pre-

viously identified microdeletions that are confined to the 50 UTR of the protein tyrosine phosphatase receptor D (PTPRD)

gene, implicating this gene in the pathogenesis of these tumors. Here, we demonstrate that the 50 UTR of this gene, consisting

of 11 noncoding exons, is also aberrantly spliced in >50% of NBL primary tumors and cell lines. The loss of exons from the 50

UTR region through aberrant splicing results in aberrant mRNA isoforms that are similar to those generated through microde-

letions. The aberrant splicing or microdeletion of 50 UTR exons in such a high proportion of tumors indicates that loss of these

exons dys-regulates the mRNA sequence. To further validate the role of PTPRD in NBL, we have examined the expression of

this gene in normal fetal adrenal neuroblasts (the cell of origin of NBL) and in tumors from patients with either low stage or

high stage disease. This gene is expressed at lower levels in high stage NBL tumors, particularly those with amplification of

MYCN, relative to low stage tumors or normal fetal adrenal neuroblasts, consistent with the possibility that loss of the 50 UTRexons have destabilized the mRNA. This article contains Supplementary Material available at http://www.interscience.wiley.

com/jpages/1045-2257/suppmat. VVC 2007 Wiley-Liss, Inc.

INTRODUCTION

Neuroblastoma (NBL), a pediatric cancer

derived from primitive cells of the sympathetic

nervous system (Brodeur, 2003), can be subdivided

into at least three genetic subtypes with dramati-

cally different clinical behaviors, those with MYCNamplification, 11q loss, or near triploid (Vandesom-

pele et al., 1998). The near triploid tumors gener-

ally have favorable clinical outcomes, while tumors

with either MYCN amplification or 11q loss often

become refractory to treatment (Brodeur et al.,

2001).

Hemi- and homozygous microdeletions affecting

either the 50 UTR or coding sequence regions of

the protein tyrosine phosphatase receptor D

(PTPRD) gene have been reported in NBL (Stal-

lings et al., 2006). In addition, expression of

PTPRD is higher in neuroblasts that were microdis-

sected from normal fetal adrenal gland than in tu-

mor material derived from patients with Stage 4

disease (De Preter et al., 2006). These normal neu-

roblasts have been shown to be the cell of origin of

NBL. Interestingly, PTPRD appears to play a wide-

spread role in cancer, as microdeletions (Sato et al.,

2005; Zhao et al., 2005; Purdie et al., 2007; Stark

and Hayward, 2007) or mutations (Sjoblom et al.,

2006) affecting this gene have been detected in

several other forms of cancer. PTPRD protein

interacts with a putative metastasis suppressor,

MIM, which is involved with cytoskeletal remodel-

ing (Woodings et al., 2003; Gonzalez-Quevedo

et al., 2005).

The genomic architecture of PTPRD is complex,

with 36 coding sequence exons and a 50 UTR

spliced together from 11 noncoding exons. Multi-

ple isoforms are generated by either alternate splic-

ing (Pulido et al., 1995) or by alternate transcrip-

tional start sites (Sato et al., 2005) in a tissue spe-

cific manner. The predominant isoform in brain

has an extended 711 base pair 50 UTR (L isoform),

while the isoform (S) expressed in kidney lacks the

extended 50 UTR (Sato et al., 2005). In addition,

the brain isoform is characterized by the absence

*Correspondence to: Raymond L. Stallings, Royal College of Sur-geons in Ireland, 123 St Stephen’s Green, York House, Dublin 2,Ireland or Children’s Medical and Research Foundation, Our Lady’sHospital for Sick Children, Dublin 12, Ireland.E-mail: rstallings@rcsi.ie

Supported by: Association of International Cancer Research.

Received 28 September 2007; Accepted 26 October 2007

DOI 10.1002/gcc.20521

Published online 29 November 2007 inWiley InterScience (www.interscience.wiley.com).

VVC 2007 Wiley-Liss, Inc.

GENES, CHROMOSOMES & CANCER 47:197–202 (2008)

of exons 14 to 18 corresponding to amino acid resi-

dues 568 to 978 of the 4th through 7th fibronectin

III-like domain and by the insertion of a 12 base

pair mini-exon sequence between exons 23 and 24

(Pulido et al., 1995).

In spite of the increasing evidence for the role of

PTPRD in tumor pathogenesis, relatively little is

understood about the expression and splicing of

this gene in tumors. It was demonstrated that lung

carcinoma cell lines with PTPRD deletions have

lower levels of mRNA transcripts relative to nor-

mal lung tissue (Sato et al., 2005). In addition, Pur-

die et al. (2007) have shown that an intragenic de-

letion of PTPRD identified in a squamous cell car-

cinoma created a fusion allele consisting of exon

B8 of the 50 UTR fused to coding sequence exon

5, which is expressed in the tumor cells, perhaps

conveying a dominant negative effect. Here, we

examine both the expression and splicing of

PTPRD in primary NBL tumors and cell lines.

MATERIAL ANDMETHODS

Cell Lines and Tissues

Protein tyrosine phosphatase receptor D

(PTPRD) studies were performed on primary NBL

tumor tissues obtained from a tumor bank at Our

Lady’s Hospital for Sick Children (Dublin, Ire-

land) and on cell lines obtained from either the

American Type Culture Collection or the Child-

ren’s Hospital of Los Angeles. The International

Neuroblastoma Staging System was used to desig-

nate disease stage (Shimada et al., 1999). Basically,

Stage 1 or 2 disease involves local-regional tumors,

where metastasis (if any) is limited to nearby

lymph nodes, while Stage 4 involves metastasis to

distant organ sites.

Affymetrix HU-133a Expression Microarray

Analysis of Microdissected Fetal Neuroblasts

and Primary Tumors

Details of the characterization of microdissected

normal fetal adrenal neuroblasts, including micro-

array expression analysis, has been provided in

detail by (De Preter et al., 2006).

Quantitative Real Time PCR and PCR Analysis

of mRNA Splicing

All primers used for quantitative real time PCR

analysis of either cDNA or genomic DNA are

listed in Supplementary Table 1. Expression of

PTPRD was quantified using two-step reverse tran-

scriptase quantitative PCR analysis. The primers

for real time PCR analysis were derived from exon

B1 (50 UTR)(forward primer) and exon 1 (coding

sequence)(reverse primer). Both of these exons

were found to be consistently expressed in all

tumors and cell lines that were analyzed for

mRNA splicing.

Primers used in the analysis of mRNA splicing

were designed to specifically amplify exon regions

B2, B4, B5, B7, B8, B10, B11, E2, or E3 (forward

primers) to exon E5 (reverse primer) from cDNA.

GAPDH was used to normalize the expression lev-

els in each sample. cDNA was first synthesized

from total RNA using oligo dT primers and the

TaqMan Reverse Transcriptase (RT) kit from

Applied Biosystem (Foster City, CA). Hundred

and fifty nanogram of total RNA was used for

cDNA synthesis in a total reaction volume of 30 ll,which included the 1X RT buffer, 5.5 mM MgCl2,

500 lM dNTP, 2.5 lM oligo dT, 0.4 U/ll of RNase

inhibitor, and 1.25 U/ll Multiscribe Reverse Tran-

scriptase. The cycling parameters were 258C for

10min, 488C for 30min, and 958C for 5min. The syn-

thesized cDNA was used for the real-time Q-PCR

reaction on an ABI 7500 using triplicate runs per

sample-primer pair. Each Q-PCR reaction (20 ll)contained 2 ll of cDNA, 0.2 lM each of forward

and reverse primers, and 1X SYBR green master

mix. The cycling condition was 1 cycle each of

2 min at 508C, and denaturation at 958C for 10 min,

followed by 40 cycles of 958C for 15 sec and 608Cfor 60 sec. The relative quantitation assay type was

used for experimental setup and the data was ana-

TABLE 1. PTPRD 50 UTR Exon Expression

Sample Stage 50 UTR exons not expresseda

T-39 4 B2, B4, B5, B7, B10b, B11c

T-53 DUB 2T-31 4T27 2aT1 2IMR32Kelly B7, B10CHP212SK-N-FI B2, B4, B7b

SK-N-DZ B2, B4, B5, B7, B8, B11, E3b

SMS-LHNSHSY5Y B2, B4, B5, B7, B8, B10NGP B2, B4, B7, B11SK-N-BE B2, B4, B5, B7, B8, B11NB69LA-N-6 B2c, B4, B5, B7, B8, B10SK-N-AS B2, B4, B5b, B7, B8c, B11c

aExons B2, B4, B5, B7, B8, B10, B11 from the 50 UTR and E2 and E3

from the 50 region of the coding sequence were assayed.bFaint band.cAbnormal size.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

198 NAIR ET AL.

lyzed using the comparative Ct (DDCt) method.

The relative expression levels (compared with

total RNA from brain tissues) were calculated from

the respective Ct value as 22DDCt, where 2DDCtrepresents the ratio of the normalized expression

in sample and the control brain tissue. The relative

expression 6 standard error is plotted for each

sample. Real timer PCR analysis of genomic DNA

was carried out as previously described (Selzer

et al., 2005).

RESULTS

PTPRD Expression in Neuroblastoma

De Preter et al. (2006) have recently determined

that fetal adrenal neuroblasts are the cell of origin

of neuroblastomas (NBLs) and have examined the

expression profiles of microdissected normal fetal

adrenal neuroblasts and different genetic subtypes

of primary NBL using Affymetrix HU-133A arrays

(De Preter et al., 2006). A review of the primary

data from this study shows that PTPRD is

expressed at higher levels in both normal fetal ad-

renal neuroblasts and loco-regional tumors relative

to MYCN amplified or 11q- tumors derived from

patients with high stage disease (i.e., metastatic

disease) (Fig. 1). No information regarding the

expression of the L isoform (containing the 50

UTR) in NBL was obtainable from this study since

only the extreme 30 region of the PTPRD gene is

represented on these arrays. Therefore, we have

carried out expression analysis of PTPRD on 29 pri-

mary NBL tumors representing low stage near trip-

loid tumors with favorable histopathology (n 5 9),

high stage 11q- tumors with unfavorable histopa-

thology (n 5 10) and high stage MYCN amplified

tumors with unfavorable histopathology (n 5 10)

using quantitative real time PCR (Fig. 2). Previous

analysis of 15 of these tumors by array CGH indi-

cated that none of these tumors had detectable de-

letion of the PTPRD gene (Stallings et al., 2006).

PTPRD deletion status of the other 14 tumors is

unknown. Primers specific for the L isoform, as

originally described by Sato et al. (2005), came

from exon B1 (forward primer)(exons designated B

are from the 50 UTR) and from exon 1 (reverse pri-

mer)(see supplementary information on PCR pri-

mers). The L isoform was expressed in NBL, simi-

lar to brain tissue. Consistent with the microarray

study (De Preter et al., 2006), the mean expression

in low stage hyperdiploid tumors (mean 5 0.7) was

greater than either the 11q- (mean 5 0.3) or MYCNamplified (mean 5 0.07) tumors. The difference in

mean expression levels between the three tumor

groups was statistically significant using the non-

parametric Kruskal-Wallis test (P < 0.001).

PTPRD mRNA Splicing in Neuroblastoma

To determine if NBL have the PTPRD splice

variant that is predominant in brain (i.e., lacking

exons 14 to 18), we sequenced PCR products

obtained from cDNA from 12 tumors using a for-

ward primer from E11 and a reverse primer from

E28. All 12 tumors, which included five 11q-

tumors, two MNA tumors and five low stage hyper-

diploid tumors, lacked exons 14 through 18, similar

to the brain isoform. All of these tumors also con-

tained a mini-exon sequence (TAAACCCGACAG)

between exons 23 and 24, similar to the brain

isoform.

Figure 1. PTPRD expression in normal fetal neuroblasts, low stagehyperdiploid neuroblastomas with favorable prognosis, high stage neu-roblastomas with loss of 11q and unfavorable prognosis, and high stageneuroblastomas with MYCN amplification and unfavorable prognosis.PTPRD expression data from Affymetrix HU-133a microarrays wasobtained from De Preter et al. (2006).

Figure 2. PTPRD expression in high stage neuroblastomas withMYCN amplification, high stage tumors with loss of 11q, and low stagenear triploid neuroblastomas. PTPRD expression was ascertained usingquantitative real time PCR. The primers for real time PCR analysiswere derived from exon B1 (50 UTR) and exon 1 (coding sequence).Both of these exons were found to be consistently expressed in alltumors and cell lines that were subsequently analyzed for mRNA splic-ing (Table 1 and Fig. 4).

Genes, Chromosomes & Cancer DOI 10.1002/gcc

199ABERRANT SPLICING OF PTPRD IN NEUROBLASTOMA

In order to further examine the splicing of the 50

region of PTPRD in NBL tumors and cell lines, we

have carried out PCR analysis on cDNAs using for-

ward primers from exons B2, B4, B5, B7, B8, B10,

B11, exon 2 and exon 3 individually combined

with a reverse primer from exon 5 (see supplemen-

tal section for primer sequences). Figure 3a illus-

trates the expected pattern of the PCR products

from brain cDNA (expected to have product from

all primers). As expected, the PCR products

decrease in size going from exons B2 to E3 (Fig.

3a). PCR products using primers from the 50 UTR

exons B2 to B11 in combination with a reverse

primer from exon 5 were absent from kidney

cDNA, as expected (Fig. 3b). Only forward primers

from exons 2 and 3 produced PCR product from

cDNA derived from kidney total mRNA (Fig. 3b),

consistent with a published study indicating that

kidney expresses only the S isoform (Sato et al.,

2005). Analysis of cDNA from five primary tumors

and 12 cell lines indicated that exons B2 through

exon 3 were present in eight samples (Fig. 4a).

However, in nine samples, including both primary

tumors and cell lines, various exons from the 50

UTR were not present in cDNA, as illustrated in

Figures 4b–4d and summarized in Table 1. Exons

B2 and B4 were the most frequently missing 50

UTR exons. In some instances, PCR products

were not of the expected size (Table 1), which

might also be indicative of splicing aberrations.

The Kelly and SK-N-AS cell lines had homozy-

gous deletion of the PTPRD 50 UTR region, as

indicated by array CGH and real time PCR analy-

sis of genomic DNA (Stallings et al., 2006), so that

absence of some of these exons in cDNA would be

expected. The presence of exon B8 and the ab-

sence of exons B7 and B10 in Kelly (Fig. 4b), how-

ever, indicates that either the deletions are more

complex than what was originally predicted by the

array CGH data (Stallings et al., 2006) or that a

combination of deletion and aberrant splicing has

led to the noncontiguous expression of these

exons. Apart from Kelly and SK-N-AS, there was

no evidence for homozygous deletion of PTPRD in

the other cell lines or tumors using either array

CGH or real time PCR analyses. One cell line, SK-

N-DZ had a hemizygous deletion of a large seg-

ment of the 9p which includes PTPRD, as ascer-

tained by CGH (Schleiermacher et al., 2003). This

cell line is missing exons B2, B4, B5, B7, B8, and

B11 of the 50 UTR in PTPRD cDNA (Fig. 4d). To

Figure 3. Analysis of exons from the 50 region of PTPRD in cDNAderived from brain (A) and kidney (B). Forward primers from exons B2,B4, B5, B7, B8, B10, and B11 from the 50 UTR or coding sequencesexons E2 and E3 were used in conjunction with a reverse primer fromexon 5. Primers are listed in the supplemental information section. Asexpected, each forward primer from a 50 UTR exon produced a PCRproduct, decreasing in size from exons B2 through B11 from braincDNA. Forward primers from coding exons E2 and E3 also producedPCR products of decreasing size. No PCR products were observedwith 50 UTR forward primers from kidney, which has been reported topossess only the S isoform of PTPRD (lacking the 711 bp 50 UTR). PCRproducts were observed, however, using forward primers from codingexons E2 and E3.

Figure 4. Analysis of PTPRD 5’ exons in cDNA derived from theneuroblastoma cell lines CHP-212 (A), Kelly (B), SK-N-BE (C), and SK-N-DZ (D) using forward primers from exons B2, B4, B5, B7, B8, B10,and B11 from the 50 UTR or coding sequences exons E2 and E3 in con-junction with a reverse primer from exon E5, as described in Figure 2.CHP-212 (A) had a splicing pattern similar to that of normal brain tis-sue, with all primer sets producing PCR product. Kelly cells lackedexons B7 and B10, while retaining exon B8. The absence of B7 and B10is due to complex homozygous deletion of corresponding regions ofgenomic DNA. Both SK-N-BE and SK-N-DZ lacked exons B2, B4, B5,B7, B8, and B11, presumably through aberrant splicing of mRNA. Inaddition, SK-N-DZ had an exon 3 fragment that was larger then theexpected size, perhaps representing either additional mRNA splicingaberrations or a PCR generated artifact. Neither cell line had homozy-gous deletion of any part of the PTPRD gene (Fig. 4).

Genes, Chromosomes & Cancer DOI 10.1002/gcc

200 NAIR ET AL.

determine whether small-scale cryptic deletions of

the PTPRD region might exist on the ‘‘non-

deleted’’ chromosome 9 homologue, which might

have escaped detection by CGH analysis, we per-

formed real time PCR analysis on genomic DNA

using 13 primers, including primers from exons B2,

B3, B4, B5, B6, B7, and B8 (Fig. 5). Although the

chromosome 9 hemizygous deletion was apparent

from the PCR analysis, there was no evidence for

homozygous deletion, indicating that some form of

aberrant mRNA splicing led to the loss of the these

50 UTR exons in cDNA. The SK-N-BE cell line,

which had a similar expression profile for the 50

UTR exons as SK-N-DZ, was also analyzed for

possible deletion of small regions of the PTPRDgene (Fig. 5). No evidence for either hemi- or

homozygous deletion was obtained, confirming

that aberrant splicing was responsible for the miss-

ing exons in the cDNA.

DISCUSSION

The loss of 50 UTR exons in PTPRD mRNA

occurs at high frequency (>50%) in NBL tumors

and cell lines as a consequence of either homozy-

gous genomic deletion or aberrant mRNA splicing.

The high frequency of exon loss strongly impli-

cates this gene, and particularly the 50 region of

this gene, in NBL pathogenesis. Both microarray

expression and real time PCR analyses of exon

regions that are not deleted or aberrantly spliced

indicate that the expression of this gene has been

decreased in NBL from patients with high stage

disease. The possibility that loss of these 50 UTR

exons decreases mRNA stability is consistent with

these results and warrants further study. It is possi-

ble that multiple mechanisms might exist for down

regulating PTPRD activity in high stage NBLs,

given that tumors and cell lines that express the

entire set of 50 UTR exons also have relatively low

levels of PTPRD mRNA. Aberrant hypermethyl-

ation of the promoter region of this gene might be

another mechanism for down-regulation.

Several lines of evidence indicate that PTPRDmight be acting as a metastasis suppressor gene, as

opposed to an initial suppressor of tumorigenesis

in NBL. First, we have demonstrated that in a

matched primary and metastatic tumor sample

only the metastatic sample had microdeletion of

this gene (Stallings et al., 2006). Secondly, the fact

that primary tumors derived from patients with

metastatic disease have lower expression of

PTPRD relative to loco-regional tumors further

suggests that inactivation of this gene is important

for metastasis. It is also noteworthy that PTPRD

interacts with a putative metastasis suppressor,

MIM, which is involved with cytoskeletal remodel-

ing (Woodings et al., 2003; Gonzalez-Quevedo

et al., 2005).

The involvement of PTPRD in several forms of

cancer, as indicated by both microdeletions and

mutations (Sato et al., 2005; Zhao et al., 2005; Sjo-

blom et al., 2006; Stallings et al., 2006; Purdie

et al., 2007; Stark and Hayward, 2007) should initiate

Figure 5. Real time quantitative PCR analysis of genomic DNA using13 primer sets (see supplemental section on primers) spanning thePTPRD gene. MJ90 is a diploid human fibroblast cell line used as a con-trol, while SK-BN-BE and SK-N-DZ are neuroblastoma cell lines. Ahemizygous deletion was evident for all primer sets in SK-N-DZ (whichhas been reported to be hemizygously deleted for a large segment of

chromosome 9p by CGH (Schleiermacher et al., 2003), while no dele-tion for PTPRD was detected in SK-N-BE. The absence of a homozygousdeletion in these cell lines indicates that the missing 50 UTR exons fromthe PTPRD cDNA (Fig. 4) resulted from mRNA splicing. GAPDH wasused for normalization. [Color figure can be viewed in the online issue,which is available at www.interscience.wiley.com.]

Genes, Chromosomes & Cancer DOI 10.1002/gcc

201ABERRANT SPLICING OF PTPRD IN NEUROBLASTOMA

a greater interest for this gene in future studies.

The role of the 50 UTR should be a particular pri-

ority given that some tumors have microdeletions

affecting only this region or have splice variants

that basically mimic these deletions.

REFERENCES

Brodeur GM. 2003. Neuroblastoma: Biological insights into a clini-cal enigma. Nat Rev Cancer 3:203–216.

Brodeur GM, Look AT, Shimada H, Hamilton VM, Maris JM, HannHW, Leclerc JM, Bernstein M, Brisson LC, Brossard J, LemieuxB, Tuchman M, Woods WG. 2001. Biological aspects of neuroblas-tomas identified by mass screening in Quebec. Med PediatrOncol 36:157–159.

De Preter K, Vandesompele J, Heimann P, Yigit N, Beckman S,Schramm A, Eggert A, Stallings RL, Benoit Y, Renard M, DePaepe A, Laureys G, Pahlman S, Speleman F. 2006. Human fetalneuroblast and neuroblastoma transcriptome analysis confirmsneuroblast origin and highlights neuroblastoma candidate genes.Genome Biol 7:R84.

Gonzalez-Quevedo R, Shoffer M, Horng L, Oro AE. 2005. Receptortyrosine phosphatase-dependent cytoskeletal remodeling bythe hedgehog-responsive gene MIM/BEG4. J Cell Biol 168:453–463.

Pulido R, Serra-Pages C, Tang M, Streuli M. 1995. The LAR/PTPd/PTP sigma subfamily of transmembrane protein-tyrosine-phos-phatases: Multiple human LAR, PTP d, and PTP sigma isoformsare expressed in a tissue-specific manner and associate with theLAR-interacting protein LIP.1. Proc Natl Acad Sci USA 92:11686–11690.

Purdie KJ, Lambert SR, Teh MT, Chaplin T, Molloy G, RaghavanM, Kelsell DP, Leigh IM, Harwood CA, Proby CM, Young BD.2007. Allelic imbalances and microdeletions affecting thePTPRD gene in cutaneous squamous cell carcinomas detectedusing single nucleotide polymorphism microarray analysis. GenesChromosomes Cancer 46:661–669.

Sato M, Takahashi K, Nagayama K, Arai Y, Ito N, Okada M, MinnaJD, Yokota J, Kohno T. 2005. Identification of chromosome arm9p as the most frequent target of homozygous deletions in lungcancer. Genes Chromosomes Cancer 44:405–414.

Schleiermacher G, Janoueix-Lerosey I, Combaret V, Derre J, Coutu-rier J, Aurias A, Delattre O. 2003. Combined 24-color karyotypingand comparative genomic hybridization analysis indicates pre-dominant rearrangements of early replicating chromosome regionsin neuroblastoma. Cancer Genet Cytogenet 141:32–42.

Selzer RR, Richmond TA, Pofahl NJ, Green RD, Eis PS, Nair P,Brothman AR, Stallings RL. 2005. Analysis of chromosome break-points in neuroblastoma at sub-kilobase resolution using fine-til-ing oligonucleotide array CGH. Genes Chromosomes Cancer 44:305–319.

Shimada H, Ambros IM, Dehner LP, Hata J, Joshi VV, Roald B,Stram DO, Gerbing RB, Lukens JN, Matthay KK, CastleberryRP. 1999. The International Neuroblastoma Pathology Classifica-tion (the Shimada system). Cancer 86:364–372.

Sjoblom T, Jones S, Wood LD, Parsons DW, Lin J, Barber TD,Mandelker D, Leary RJ, Ptak J, Silliman N, Szabo S, BuckhaultsP, Farrell C, Meeh P, Markowitz SD, Willis J, Dawson D, WillsonJK, Gazdar AF, Hartigan J, Wu L, Liu C, Parmigiani G, Park BH,Bachman KE, Papadopoulos N, Vogelstein B, Kinzler KW, Velcu-lescu VE. 2006. The consensus coding sequences of human breastand colorectal cancers. Science 314:268–274.

Stallings RL, Nair P, Maris JM, Catchpoole D, McDermott M,O’Meara A, Breatnach F. 2006. High-resolution analysis of chro-mosomal breakpoints and genomic instability identifies PTPRDas a candidate tumor suppressor gene in neuroblastoma. CancerRes 66:3673–3680.

Stark M, Hayward N. 2007. Genome-wide loss of heterozygosityand copy number analysis in melanoma using high-density single-nucleotide polymorphism arrays. Cancer Res 67:2632–2642.

Vandesompele J, Van Roy N, Van Gele M, Laureys G, Ambros P,Heimann P, Devalck C, Schuuring E, Brock P, Otten J, GselinckJ, De Paepe A, Speleman F. 1998. Genetic heterogeneity of neu-roblastoma studied by comparative genomic hybridization. GenesChromosomes Cancer 23:141–152.

Woodings JA, Sharp SJ, Machesky LM. 2003. MIM-B, a putativemetastasis suppressor protein, binds to actin and to protein tyro-sine phosphatase delta. Biochem J 371(Part 2):463–471.

Zhao X, Weir BA, LaFramboise T, Lin M, Beroukhim R, GarrawayL, Beheshti J, Lee JC, Naoki K, Richards WG, Sugarbaker D,Chen F, Rubin MA, Janne PA, Girard L, Minna J, Christiani D, LiC, Sellers WR, Meyerson M. 2005. Homozygous deletions andchromosome amplifications in human lung carcinomas revealedby single nucleotide polymorphism array analysis. Cancer Res65:5561–5570.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

202 NAIR ET AL.