UBC Faculty Research and Publications

Candidate metastasis suppressor genes uncovered by array comparative genomic hybridization in a mouse… Yi, Yajun; Nandana, Srinivas; Case, Thomas; Nelson, Colleen; Radmilovic, Tatjana; Matusik, Robert J; Tsuchiya, Karen D Sep 26, 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13039_2009_Article_46.pdf [ 604.82kB ]
JSON: 52383-1.0223792.json
JSON-LD: 52383-1.0223792-ld.json
RDF/XML (Pretty): 52383-1.0223792-rdf.xml
RDF/JSON: 52383-1.0223792-rdf.json
Turtle: 52383-1.0223792-turtle.txt
N-Triples: 52383-1.0223792-rdf-ntriples.txt
Original Record: 52383-1.0223792-source.json
Full Text

Full Text

ralssBioMed CentMolecular CytogeneticsOpen AcceResearchCandidate metastasis suppressor genes uncovered by array comparative genomic hybridization in a mouse allograft model of prostate cancerYajun Yi1, Srinivas Nandana2, Thomas Case2, Colleen Nelson3, Tatjana Radmilovic1, Robert J Matusik2 and Karen D Tsuchiya*4,5Address: 1Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA, 2Department of Urologic Surgery and Vanderbilt Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN, USA, 3Department of Urologic Sciences, University of British Columbia, The Prostate Centre at Vancouver General Hospital, Vancouver, British Columbia, Canada, 4Clinical Research Division, Fred Hutchinson Cancer Research Center and Department of Laboratories, Seattle Children's Hospital, WA, USA and 5Department of Laboratory Medicine, University of Washington School of Medicine, Seattle, WA, USAEmail: Yajun Yi - andrew.yi@vanderbilt.edu; Srinivas Nandana - srinivas.r.nandana@vanderbilt.edu; Thomas Case - tom.case@vanderbilt.edu; Colleen Nelson - colleen.nelson@ubc.ca; Tatjana Radmilovic - tatjana_radmilovic07@yahoo.com; Robert J Matusik - robert.matusik@vanderbilt.edu; Karen D Tsuchiya* - karen.tsuchiya@seattlechildrens.org* Corresponding author    AbstractBackground: The purpose of this study was to identify candidate metastasis suppressor genesfrom a mouse allograft model of prostate cancer (NE-10). This allograft model originally developedmetastases by twelve weeks after implantation in male athymic nude mice, but lost the ability tometastasize after a number of in vivo passages. We performed high resolution array comparativegenomic hybridization on the metastasizing and non-metastasizing allografts to identifychromosome imbalances that differed between the two groups of tumors.Results: This analysis uncovered a deletion on chromosome 2 that differed between themetastasizing and non-metastasizing tumors. Bioinformatics filters were employed to mine thisregion of the genome for candidate metastasis suppressor genes. Of the 146 known genes thatreside within the region of interest on mouse chromosome 2, four candidate metastasis suppressorgenes (Slc27a2, Mall, Snrpb, and Rassf2) were identified. Quantitative expression analysis confirmeddecreased expression of these genes in the metastasizing compared to non-metastasizing tumors.Conclusion: This study presents combined genomics and bioinformatics approaches foridentifying potential metastasis suppressor genes. The genes identified here are candidates forfurther studies to determine their functional role in inhibiting metastases in the NE-10 allograftmodel and human prostate cancer.BackgroundProstate cancer (PCa) is a heterogeneous disease and theexpression have been reported in PCa, yet identificationof many of the specific genes that drive the progression ofPublished: 26 September 2009Molecular Cytogenetics 2009, 2:18 doi:10.1186/1755-8166-2-18Received: 6 April 2009Accepted: 26 September 2009This article is available from: http://www.molecularcytogenetics.org/content/2/1/18© 2009 Yi et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Page 1 of 9(page number not for citation purposes)ability to predict its clinical outcome is limited. Numer-ous chromosomal abnormalities and alterations in genethese tumors is still lacking. The finding of the TMPRSS2/ETS fusion and the overexpression of ETS transcriptionMolecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/18family members in the majority of PCa illustrates the suc-cess of utilizing a bioinformatics approach to gene discov-ery [1], but the consequences of many other recurrentacquired genomic alterations remain to be elucidated.Studies of human PCa are hindered by the biologic andgenetic heterogeneity of this disease not only betweenindividuals, but also within a given individual. Geneti-cally engineered mouse models of PCa provide an in vivoexperimental system in which tumors with the sameunderlying etiology can be sampled during the course ofprogression at defined time points. The LPB-Tag mousemodel of prostate cancer is one such model that has beenwell characterized [2-4].We have previously described the establishment of anallograft (NE-10) from a primary prostate tumor from theLPB-Tag mouse model that consistently metastasized by12 weeks after transplantation in nude mice in early pas-sages [4]. Conventional cytogenetic analysis of the NE-10allograft revealed numeric and structural chromosomeabnormalities, including a deletion of distal chromosome2 that was consistently present over multiple in vivo pas-sages of the allograft in nude mice [4]. A similar chromo-some 2 deletion has also been described in a mousemodel of acute promyelocytic leukemia [5]. After repeatedin vivo passages, the NE-10 allograft eventually lost theability to metastasize. In this study, we have taken advan-tage of the differential metastasizing behavior of the NE-10 allograft, arising from the same original tumor, toscreen the genome for candidate genes that play a role inmetastasis. High resolution genomic technology, com-bined with novel bioinformatics approaches, enabled usto identify different regions of chromosome imbalancesbetween the two allograft lines, and to propose candidatemetastasis suppressor genes within a region of chromo-some 2 that was found to harbor a larger deletion in met-astatic compared to non-metastatic tumors.MethodsNE-10 allograft modelThe 12T-10 line of the LPB-Tag mouse model of prostatecancer was generated using a transgene that consists of therat probasin promoter driving the SV40 large T antigenwith deletion of the small T antigen [3]. These micedevelop low-grade and high-grade prostatic intraepithe-lial neoplasia at 2-5 month of age, with progression toinvasive and metastatic, high-grade, androgen-independ-ent carcinoma demonstrating neuroendocrine differentia-tion at 6-14 months of age. A primary prostate tumorfrom the ventral prostate of a 12T-10 transgenic mousewas used to establish an allograft model by implantationsubcutaneously in male athymic nude mice [4]. After 18weeks, the allograft was passaged to another male nudetwelve weeks after implantation. All metastases from theallografts were histologically similar to the metastasis seenin the 12T-10 mice. Later allograft passages showed histo-logically identical features to the early passages; however,fewer allografts developed metastases. By passage 15, met-astatic potential was completely lost, at least up to thepoint where it was no longer feasible to maintain the micedue to the size of the subcutaneous tumors. Tumors con-sisting of non-metastasizing subcutaneous allografts(SQnon-met), metastasizing subcutaneous allografts(SQmet), and liver metastases (LiverMet) were collectedat 12 weeks post-implantation and snap-frozen for arrayCGH. All procedures involving mice were approved by theVanderbilt University Medical Center and Fred Hutchin-son Cancer Research Center Institutional Animal Care andUse Committees.Array comparative genomic hybridization (CGH)Array CGH was initially carried out using mouse bacterialartificial chromosome (BAC) arrays produced by theGenomics Shared Resource, Fred Hutchinson CancerResearch Center as described [6]. The mouse BAC cloneset was obtained from A. Bradley, Wellcome Trust SangerInstitute [7] and provides an average resolution of 1 Mb.DNA was isolated using the Gentra Puregene genomicDNA purification kit (Qiagen, Valencia, CA). For the BACarrays, three SQnon-met tumors (passage 19), threeSQmet (passages 4, 9, and 12), and four LiverMet (twoeach from passages 9 and 12) were analyzed. Onematched SQmet and LiverMet from the same nude mouseat passage 9 was included in the analysis. Pooled DNAobtained from normal kidney from four different CD-1males was used as a reference. DNA labeling, hybridiza-tion, scanning, and data analysis was performed asdescribed previously [6].Array CGH was repeated on a subset of two tumors eachfrom the SQnon-met and LiverMet groups using the Agi-lent mouse 105K oligonucleotide CGH arrays. Thesearrays were designed using UCSC mm7 (NCBI build 35,August, 2005), and have an average probe spacing of 15Kb. One ug of RsaI/AluI digested DNA was labeled witheither Cy3 or Cy5 using the BioPrime Array CGH genomiclabeling system (Invitrogen Corp., Carlsbad, CA). Approx-imately 4 ug each of Cy5 labeled test (tumor) DNA andCy3 labeled reference (normal female C57Bl/6 liver) DNAwas combined with 25 ug of mouse Cot-1 (InvitrogenCorp.), Agilent 10× blocking agent and 2× hybridizationbuffer (Agilent Technologies, Santa Clara, CA) to a finalvolume of 260 ul. Hybridization and washes were carriedout according to the Agilent oligo array-based CGH proto-col v. 4.0. Scanning was performed on an Agilent scannerand data extraction was carried out using Agilent featurePage 2 of 9(page number not for citation purposes)mouse and the process was repeated to establish the NE-10 line [4]. Initial passages in all mice developed grosslyvisible metastases to liver and micrometastases to lung byextraction software v.9.1, employing linear and Lowessnormalization. Results were analyzed and chromosomeplots generated using CGHanalytics software v. 6.0. The Z-Molecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/18score algorithm with the threshold set at 2.0 and a 2 Mbwindow was used for determining gains and losses. TheLPB-Tag tumors were generated on an outbred CD-1 back-ground, and we did not have non-neoplastic DNA availa-ble from the mouse whose tumor was used to establishthe allograft. As C57Bl/6 reference DNA was used for theoligonucleotide array CGH experiments, gains and lossessmaller than 2 Mb were ignored in order to avoid mistak-ing benign copy number variants between mouse strainsfrom acquired copy number alterations in the tumors.To compare array CGH profiles between the differentgroups of tumors, the Differential Gene locus MAPpingsoftware (DIGMAP version 2.0) was used to analyze arrayCGH data sets [8]. Using the default parameters specifiedin the program, DIGMAP displayed array CGH data as aheat map based on log2 ratio between test and referenceDNA for each tumor. The chromosome regions with sig-nificantly different log2 ratios were marked as differentialflagged regions (DFRs) by direct visualization and compu-tational screening using a T-test based sliding window(TTSW) analysis. For this analysis, a window size of 40genes was used and DFRs represent regions greater than 3standard deviations from the whole genome average Tscore.Expression arraysExpression arrays were performed on two SQnon-mettumors (passages 11b, 18c) and four SQmet tumors (pas-sages 2, 5, 14a, 14c). Allograft passages were labeled a, b,c, etc. when the same graft passage was implanted intomore than one mouse. As a control, mouse reference RNAfrom two postnatal day 1 mice was pooled. RNA was iso-lated using the RNeasy kit (Qiagen Valencia CA), includ-ing treatment with DNase. RNA was quantified using aNanoDrop (ThermoScientific). All samples analyzed hada A260/230 ratio greater than 1.8. RNA quality was ana-lyzed on an Agilent 2100 BioAnalyzer. All samples ana-lyzed had a 28S peak greater than 18S peak, or an RINnumber greater than 7. The 16 k mouse arrays used for thestudy were printed at the Vancouver Microarray Facilityusing Operon's 16 K 70 mer oligomers printed on ami-nosaline slides. Ten ug of total RNA was labeled usingGenisphere's Dendrimer 350 expression array detectionkit for microarrays according to the manufacturer's proto-col. Samples were co-hybridized with ten ug of the abovedescribed mouse reference RNA. Arrays were pre-hybrid-ized in 5 × SSC, 0.1% SDS and 0.2% BSA for 45 minutes.Pre-hybridization buffer was washed off with 3 × 30 sec-ond water washes and a 2 minute wash in isopropanol.Slides were spun dry at 2000 rpm for 4 minutes. The sam-ples were applied to arrays containing 60 × 22 mm lifterslips (Erie Scientific). Slides were subsequently treated assities were extracted using ImaGene V8.0 software (Bio-Discovery).To identify chromosome regions that demonstrate differ-ential gene expression between metastatic and non-meta-static tumors, expression array datasets were analyzedbetween SQnon-met and SQmet mice using DIGMAP asdescribed above for array CGH data. A TTSW genome scanwas also performed as described above.Identification of candidate metastasis suppressor genes in chromosome 2 DFRA total of 146 known genes in the chromosome 2 differ-ential region of deletion (DFR) between metastatic andnon-metastatic tumors (nucleotides 122,316,740-139,585,560) were retrieved from the UCSC mm7 data-base http://genome.ucsc.edu. Two bioinformatics filterswere designed to determine which of these 146 genesmight function as candidate metastasis suppressor genes.A functional filter utilized the Gene Ontology (GO) data-base http://www.geneontology.org, PubMed http://www.ncbi.nlm.nih.gov/PubMed/, and Ingenuity PathwayAnalysis (IPA, http://www.ingenuity.com) to evaluate the146 genes for potential metastasis suppressor function. Aparallel filter employing our recently developed humancancer expression signature database (EXALT) was used toindependently validate the genes as candidate metastasissuppressors in silico [9]. A weighted score was assigned toeach gene for both filters (see Additional file 1).Quantitative reverse transcription (RT)-PCRTotal RNA was isolated using the RNeasy Mini kit (Qia-gen, Valencia, CA) according to the manufacturer's proto-col, including the recommended DNase treatment step.Three ug of RNA was reverse transcribed using the Reac-tionReady First Strand cDNA Synthesis Kit (SuperArrayBioscience Corp., Frederick, MD). Primer sets for mouseHprt1, Slc27a2, Mal, Snrpb, and Rassf2 were purchasedfrom SuperArray (proprietary primers, sequence not dis-closed).PCR reactions were carried out in triplicate in a 25 μL vol-ume using SYBR Green Master Mix (SuperArray). Thestandard two-step amplification with an annealing tem-perature of 60°C was performed in an ABI 7900 PCRmachine. Hprt1 was chosen as an endogenous controlgene because Hprt1 expression values were stable amongtest samples in the microarray expression data set.To allow for a comparison between samples and the twogroups, quantities of all target genes in the test samplesand a common reference RNA (Mouse XpressRef Univer-sal Total RNA, SuperArray) were normalized to the corre-Page 3 of 9(page number not for citation purposes)described in the Genisphere Array 350 expression arraydetection kit for microarrays. Arrays were scanned on aAxon 4200AL scanner (Molecular Devices). Image inten-sponding Hprt1 levels. Relative expression levels (foldchanges) were calculated using the relative standard curvemethod as outlined in the manufacturer's technical userMolecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/18manual (SuperArray). A standard curve was generatedusing the fluorescent data from 10-fold serial dilutions ofthe common reference RNA sample. Four tumors eachfrom the metastasizing and non-metastasizing allograftswere analyzed. Statistically significant differences inexpression between the two groups were determinedusing a Student's T-Test.ResultsDifferences in copy number alterations between metastatic and non-metastatic tumorsCopy number alterations in tumors were identified byarray CGH performed using BAC arrays for all tumors,and oligonucleotide arrays for a subset of tumors. ArrayCGH using both BAC and oligonucleotide platformsdetected multiple gains and losses in both non-metastatictumors (SQnon-met) and tumors with metastatic poten-tial (SQmet and LiverMet). Both array platforms uncov-ered similar chromosome imbalances, but theoligonucleotide arrays allowed us to more precisely mapregions of gain and loss. Many copy number alterationswere common to both groups of tumors, including dele-tions involving regions of chromosomes 1, 2, 3, 4, 8, 13,16, 17 18, 19, whole chromosome loss of 14, and gains ofchromosome 8 (Figure 1 and Table 1); however, differ-ences in copy number changes between the tumors withand without metastatic potential were also uncovered(Figure 1 and Table 1). One notable difference betweenthe two groups of tumors was a differential region of dele-tion in distal chromosome 2. Both the metastatic andnon-metastatic tumors shared a common 27.4 Mb inter-stitial deletion of distal chromosome 2, from bands F3 toH3 (139,585,560-167,088,181 bp); however, the meta-static tumors (LiverMet) showed a larger 45 Mb deletion(122,316,740-167,088,181 bp) that extended proximallyto band E5 (Figure 2). The non-metastatic tumors showedclonal heterogeneity for both the larger and smaller dele-tions, with low-level mosaicism for the larger deletion(Figure 2 top), but the clone or clones with the larger dele-tion were enriched in the metastatic tumors (Figure 2 bot-tom).Differences in gene expression profiles between metastatic and non-metastatic tumorsTo obtain differential gene expression profiles betweennon-metastatic and metastatic NE-10 tumors, the differ-ent groups of tumors (SQnon-met, SQmet, and LiverMet)were profiled using gene expression microarrays and ana-lyzed by DIGMAP. A T-test sliding window whole genomescan was performed and plotted by chromosome (seeAdditional file 2). Based on this analysis, chromosome 2had the greatest number of regions showing the largestdifference in gene expression between metastatic (SQmetfrom 179 to 181 Mb (see Additional file 2). This region isdistal to the common region of deletion found in both themetastatic and non-metastatic tumors. Three othersmaller peaks were also located slightly more centromeric,from 155 to 168 Mb, within the common region of dele-tion. Thus, regions of differential gene expression betweenmetastatic and non-metastatic tumors were located bothwithin and distal to the common chromosome 2 deletion;however, we were unable to detect any clusters of genesthat demonstrated down-regulation in the metastaticcompared to the non-metastatic tumors. There were alsono significant differences in gene expression detected inthe differential region of deletion on chromosome 2between the metastatic and non-metastatic tumors.Identification of candidate metastasis suppressor genes in Heat map of oligonucleotide array CGH resultsFigure 1Heat map of oligonucleotide array CGH results. Array CGH results are partitioned by chromosome number and clustered by probe location in the chromosomes. The length of each chromosome bar is based on the total number of probes for that chromosome present on the array. Within each chromosome bar, each row represents a separate NE-10 tumor, and the non-metastatic tumors (SQnon-Met) are separated from the metastatic tumors (LiverMet) by the white line (SQnon-Met above and LiverMet below the line). Black represents normal copy number in tumor compared to the normal reference; red represents increased copy number; green represents decreased copy number. A larger interstitial deletion in chromosome 2 can be visualized in the metastatic compared to the non-metastatic tumors (blue box). Relative loss of the X chromosome in the metastatic tumors, and most of the X chromosome in the non-meta-static tumors, is seen because array CGH for these tumor samples was performed using a sex-mismatched (female) ref-erence DNA. The non-metastatic tumors also show a region of copy number gain on the X chromosome.Page 4 of 9(page number not for citation purposes)and LiverMet) and non-metastatic (SQnon-met) tumors.A detailed T-test sliding window plot of chromosome 2revealed the highest DFR expression peak in band H4,chromosome 2 differentially deleted regionMetastasis suppressor genes are expected to show inactiva-tion or decreased activity in metastatic tumors because ofMolecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/18genetic or epigenetic changes that result in loss or down-regulation of expression. The finding of a differentiallydeleted region on chromosome 2, potentially resulting inloss of a metastasis suppressor gene in the metastatictumors, led us to focus on this region for more detailedanalysis. Given that only a subset of genes in the differen-tially deleted region was represented on the expressionarrays, we employed an informatics approach to identifyall potential candidate metastasis suppressor genes in theregion of interest (chromosome 2 E5-F3; nucleotides122,316,740-139,585,560). Two parallel bioinformaticsfilters were designed to carry out this analysis. Out of atotal of 146 genes in the region of interest, 11 candidatemetastasis suppressor genes were identified based on thefunction-based filters, while 22 candidates were identifiedfilter using the EXALT database (see Additional file 3). Afinal list of four candidate genes (Slc27a2, Mall, Snrpb, andRassf2) was chosen based on having the highest totalChromosome 2 oligonucleotide array CGH plotsFigure 2Chromosome 2 oligonucleotide array CGH plots. Chromosome 2 plots are shown for one representative non-metastatic (SQnon-Met) and one representative metastatic (LiverMet) tumor. The normalized log2 ratio is on the Y axis, and the chromosome bands are designated on the X axis. Copy number gains and losses are designated by the bars above or below a log2 ratio of 0, respectively. The shaded area reflects the magnitude of the gain or loss. The meta-static tumors show only the larger chromosome 2 deletion (bottom panel), whereas the non-metastatic tumors demon-strate clonal heterogeneity, with the both the smaller dele-tion, as well as low-level mosaicism for the larger deletion (top panel).Table 1: Summary of acquired copy number changes identified by oligonucleotide array CGH in liver metastases (LiverMet) from NE-10 allografts, and in non-metastasizing (SQnon-met) NE-10 allograftsChromosome SQnon-met LiverMet1 - loss *84525268-103082423 84525268-1030824232 - loss 139585560-167088181 122316740-1670881813 - loss 11809017-qter 11809017-qter4 - loss 3010281-131339250 3010281-1313392504 - gain N.A. 132282308-1492334005 - loss 3003879-114937488 N.A.7 - loss N.A. whole chromosome8 - gain 8397754-17345126 8397754-173451268 - gain 20825979-35687899 20825979-356878998 - loss 37627333-qter 37627333-qter10 - gain 79500370-89346058 N.A.10 - loss 89748171-qter N.A.12 - gain 3021012-12590005 N.A.13 - loss whole chromosome 3015154-6062483414 - loss whole chromosome whole chromosome15 - gain 99348507-qter N.A.16 - loss 3026317-36861999 whole chromosome17 - loss 3023355-10795310 whole chromosome17 - loss 29138819-qter whole chromosome18 - loss 35576728-qter whole chromosome19 - loss 46971550-qter 46971550-qterX - gain 145268084-149114686 N.A.N.A. = not applicable*Nucleotide sequences are from NCBI build 35Page 5 of 9(page number not for citation purposes)using a filter that employed a cancer expression signature score.Molecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/18Quantitative RT-PCR of 4 candidate metastasis suppressor genes in chromosome 2 differentially deleted regionQuantitative RT-PCR for the four candidate metastasissuppressor genes was performed in tumors with meta-static potential (SQMet) and tumors without metastaticpotential (SQnon-Met). Metastatic liver tumors were notanalyzed due to the presence of contaminating normalliver tissue in some of the tumors. All four candidatemetastasis suppressor genes demonstrated decreasedexpression in the metastasizing tumors compared to thenon-metastasizing tumors, with Slc27a2 and Snrpb show-ing statistically significant decreased expression in themetastasizing tumors (Figure 3).DiscussionUsing high-resolution array CGH, we uncovered multiplecopy number changes common to both the metastaticand non-metastatic tumors, consistent with a phenotypeof genomic instability in this model. Many of the regionsof loss overlap with frequent losses that have beenobserved by conventional CGH and array CGH per-formed on localized and metastatic human PCa, includ-ing human 4q22.3-q31.1, 5q21.1-q21.3, 6q15-q16.2,8p21.1-p23.1, 10q23.1, 10q25.1-q25.3, 13q14.2-q22.1,16q12.1-q24.3, 18q12.3-q21.1, 18q21.31-q21.32,18q22.1-q23 [10-12]. Loss of specific tumor suppressorgenes implicated in human PCa, such as Nkx3-1 and Rb1,was also present in the mouse NE-10 allograft. Otheralterations that are frequently seen in human PCa, such asloss of PTEN, loss of CDKN1B, and gain of MYC, were notobserved in the NE-10 model.In addition to copy number alterations common to boththe metastatic and non-metastatic tumors, we alsoobserved differences in copy number alterations betweenthese two groups of tumors that could be responsible fortheir divergent behavior. These differences included lossof chromosome 2 (122.3-139.6 Mb), gain of chromo-some 4 (132.2-149.2 Mb), loss of chromosome 7, loss ofchromosome 16 (36.9 Mb -qter), loss of chromosome 17(10.8- 29.1 Mb), and loss of chromosome 18 (3.3-35.6Mb) in the metastatic, but not the non-metastatic tumors.Our attention was drawn to a differentially deleted regionof chromosome 2 for a number of reasons. First, haploin-sufficiency or inactivation of metastasis suppressor genescan occur through deletion. We have also detected distalchromosome 2 deletions in approximately 35% of meta-static tumors from multiple independent mice from theoriginal 12T-10 transgenic line (unpublished data), con-firming that this is a recurring chromosome abnormalityin tumors from this model. In addition, a similar deletionof chromosome 2 has previously been reported in amouse model of acute promyelocytic leukemia [5]. Also,the larger region of deletion that is enriched in the meta-static tumors demonstrates conserved synteny with theshort arm of human chromosome 20 that has been impli-cated in metastatic PCa [13]. In the mouse model of acutepromyelocytic leukemia, loss of one copy of the Sfpi1(Pu.1) gene due to the chromosome 2 deletion, andreduced expression of this gene, have been implicated inthe progression of leukemia in these mice [14]. The Sfpi1gene is not an obvious candidate metastasis suppressorgene in the NE-10 allograft model, as the gene mapsapproximately 31 Mb centromeric to the proximal chro-mosome 2 deletion breakpoint in the metastatic tumors.Cd82 (Kai1) and Cd44, known metastasis suppressorgenes for prostate [15,16], are also located many mega-bases (~29 Mb and 20 Mb, respectively) proximal to thechromosome 2 deletion. Expression array analyses that weperformed also did not uncover candidate metastasis sup-pressor genes within this differentially deleted region;therefore, we employed a bioinformatics approach toidentify potential candidate genes that may be involved inthe metastatic behavior of these tumors.The candidate metastasis suppressor genes in the differen-Relative expression levels between metastatic and non-meta-static NE-10 tum rs by RT-PCRFigure 3Relative expression levels between metastatic and non-metastatic NE-10 tumors by RT-PCR. The X-axis shows four candidate metastasis suppressor genes and the Y-axis shows relative fold change (log2 based) between SQMet and SQnonMet NE-10 tumors (top panel). The relative expression levels are designated by the height of bars, and standard deviation error bars are in the direction of SQMet samples. An asterisk indicates significantly decreased expres-sion in SQMet samples (P < 0.05). The corresponding Page 6 of 9(page number not for citation purposes)tial region of deletion of chromosome 2 that we identi-fied, based on prior evidence of metastasis suppressorfunction and down-regulation in cancer, consisted ofgenomic locations of the four candidate genes on mouse chromosome 2, and the regions of conserved synteny in human, are shown below.Molecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/18Slc27a2, Mall, Snrpb, and Rassf2. All of these genes demon-strated decreased expression in the metastasizing com-pared to the non-metastasizing NE-10 tumors byquantitative RT-PCR. MALL (MAL, T-cell differentiationprotein) is a raft-associated integral membrane proteinthat is involved in membrane trafficking processes.Expression of the MAL protein has been demonstrated inspecific types of normal epithelial cells throughout therespiratory system, the gastrointestinal tract, and the gen-itourinary tract, including strong expression in the ductaland acinar cells of the prostate [17]. Loss of or decreasedMALL expression, sometimes as the result of DNA meth-ylation of the promoter region, has been found in a vari-ety of benign and malignant epithelial tumors comparedto their normal epithelial counterparts, consistent with arole for this protein in tumor suppression [17-21]. Fur-thermore, MAL has been shown to enhance apoptosisthrough the Fas pathway, and suppress tumorigenicity,invasion, and motility [21]. Although loss or decreasedMALL expression has not been evaluated in human pros-tate cancer, the expression of this protein in normal pros-tate epithelium and its ability to suppress invasion andmotility render it a biologically plausible metastasis sup-pressor gene for prostate cancer.SLC27a2 encodes a protein that is an isozyme of the long-chain fatty-acid-coenzyme A ligase family, and as suchplays a role in lipid biosynthesis and fatty acid degrada-tion. It may also be involved in translocation of long-chain fatty acids across membranes [22]. Long-chain fattyacids participate in many cellular functions and have beenimplicated as modulators of carcinogenesis, partlythrough their ability to activate peroxisome proliferator-activated receptors (PPAR). There is evidence that PPARgamma regulates prostatic epithelial differentiation andmay restrict epithelial proliferation; therefore, it is possi-ble that decreased expression of Slc27a2 in the NE-10 allo-graft model could alter the tumor suppressor activity ofPPAR gamma and contribute to metastatic behavior.With regard to RASSF2 and SNRPB, Goodzari and co-workers found evidence for a prostate cancer metastasissuppressor gene on the short arm of human chromosome20 [13]. The differential region of chromosome 2 deletionin metastatic compared to non-metastatic tumors in theNE-10 model shows conserved synteny with a region ofhuman 20p. RASSF2 and SNRPB also map to the shortarm of human chromosome 20, although they are telom-eric to the most likely metastasis suppressor region of 20p(20p11.23-p12) that was proposed [13]. SNRPB encodesa protein that is one of several nuclear proteins that arefound in common among U1, U2, U4/U6, and U5 smallribonucleoprotein particles (snRNPs). These snRNPs areof other forms of epithelial cancer, such as colon, gastric,breast, and lung carcinoma [23-25]. The mechanism ofinactivation of RASSF2 described in these tumors is aber-rant promoter methylation, primarily in early tumors. Inaddition, inactivation of the A isoform of RASSF2 by pro-moter methylation has been found to correlate with ahigher frequency of lymph node metastases in patientswith nasopharyngeal carcinoma [26].In summary, using an approach based on bioinformaticfilters applied to array CGH data on a divergent metasta-sizing and non-metastasizing mouse model of PCa, wehave identified genes that are biologically plausiblemetastasis suppressor genes. Our studies have identifiedan association between deletion with decreased expres-sion of these genes and the metastasizing NE-10 model,suggesting a biological role for these genes in the meta-static process. We believe the divergence in behavior of theallograft can be attributed to clonal heterogeneity, withboth metastasizing and non-metastasizing clones presentin early allograft passages, but with selection for a non-metastasizing clone or clones after multiple in vivo pas-sages. This selection could be due to preferential subcuta-neous growth of the non-metastasizing compared to themetastasizing component of the allograft, such that overtime, a larger and larger percentage of the allograft con-sists of the non-metastasizing population. It is possiblethat the clone or clones with the larger chromosome 2deletion are enriched in the metastatic tumors because thedifferentially deleted region itself is responsible for themetastatic behavior, either alone or in combination withother genetic alterations, or that the larger deletion is sim-ply a marker for a clone that is selected for because of thepresence of some other alteration that confers metastaticpotential. Additional studies are needed to prove that agene or genes within the differentially deleted region ofchromosome 2 are responsible for suppression of meta-static behavior.ConclusionWe have taken advantage of a mouse allograft model ofPCa (NE-10) with divergent metastasizing and non-metastasizing behavior to identify regions of the genomethat potentially harbor metastasis suppressor genes. Usinga combination of genomics and bioinformaticsapproaches, we identified candidate genes from a differ-entially deleted region on mouse chromosome 2 betweenthe metastasizing and non-metastasizing allograft lines.The genes presented here are candidates for further studiesto determine their functional role in inhibiting metastasesin the NE-10 allograft model and human PCa.List of abbreviationsPage 7 of 9(page number not for citation purposes)involved in pre-mRNA splicing. RASSF2 (RAS associationdomain family 2) is a negative effector of Ras and hasbeen implicated as a tumor suppressor gene in a numberNE-10: mouse allograft established from a primary tumorfrom the LPB-Tag mouse model of prostate cancer; PCa:prostate cancer; SQnon-met: non-metastasizing subcuta-Molecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/18neous allograft; SQmet: metastasizing subcutaneous allo-graft; LiverMet: liver metastases; CGH: comparativegenomic hybridization; BAC: bacterial artificial chromo-some; DFRs: differential flagged regions; TTSW: T-test slid-ing window.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsYY interpreted the expression array results, conducted thebioinformatics, directed and interpreted the RT-PCRexperiments, and helped draft the manuscript. SN partici-pated in the expression array experiments and helpedrevise the manuscript. TC maintained the allografts andharvested tissues. CN provided oversight of the expressionarray experiments. TR performed the quantitative RT-PCRexperiments. RJM provided the allograft model, designedthe expression array experiments, and helped draft themanuscript. KDT conducted and interpreted the arrayCGH experiments, and drafted the manuscript. Allauthors read and approved the final manuscript.Additional materialAcknowledgementsWe thank the Genomics Resource Core at the Fred Hutchinson Cancer Research Center for assistance with array CGH. This work was supported by NIH grants NCI R01-CA76142 (R.J.M.), DOD grant DAMD17-03-1-Frances Williams Preston Laboratories of the T.J. Martell Foundation (R.J.M.).References1. Rubin MA, Chinnaiyan AM: Bioinformatics approach leads to thediscovery of the TMPRSS2:ETS gene fusion in prostate can-cer.  Lab Invest 2006, 86:1099-1102.2. Kasper S, Sheppard PC, Yan Y, Pettigrew N, Borowsky AD, Prins GS,Dodd JG, Duckworth ML, Matusik RJ: Development, progression,and androgen-dependence of prostate tumors in probasin-large T antigen transgenic mice: a model for prostate can-cer.  Lab Invest 1998, 78:319-333.3. Masumori N, Thomas TZ, Chaurand P, Case T, Paul M, Kasper S,Caprioli RM, Tsukamoto T, Shappell SB, Matusik RJ: A probasin-large T antigen transgenic mouse line develops prostate ade-nocarcinoma and neuroendocrine carcinoma with meta-static potential.  Cancer Res 2001, 61:2239-2249.4. Masumori N, Tsuchiya K, Tu WH, Lee C, Kasper S, Tsukamoto T,Shappell SB, Matusik RJ: An allograft model of androgen inde-pendent prostatic neuroendocrine carcinoma derived froma large probasin promoter-T antigen transgenic mouse line.J Urol 2004, 171:439-442.5. Zimonjic DB, Pollock JL, Westervelt P, Popescu NC, Ley TJ:Acquired, nonrandom chromosomal abnormalities associ-ated with the development of acute promyelocytic leukemiain transgenic mice.  Proc Natl Acad Sci USA 2000, 97:13306-13311.6. Loo LW, Grove DI, Williams EM, Neal CL, Cousens LA, Schubert EL,Holcomb IN, Massa HF, Glogovac J, Li CI, et al.: Array comparativegenomic hybridization analysis of genomic alterations inbreast cancer subtypes.  Cancer Res 2004, 64:8541-8549.7. Chung YJ, Jonkers J, Kitson H, Fiegler H, Humphray S, Scott C, HuntS, Yu Y, Nishijima I, Velds A, et al.: A whole-genome mouse BACmicroarray with 1-Mb resolution for analysis of DNA copynumber changes by array comparative genomic hybridiza-tion.  Genome Res 2004, 14:188-196.8. Yi Y, Mirosevich J, Shyr Y, Matusik R, George AL Jr: Coupled anal-ysis of gene expression and chromosomal location.  Genomics2005, 85:401-412.9. Yi Y, Li C, Miller C, George AL Jr: Strategy for encoding andcomparison of gene expression signatures.  Genome Biol 2007,8:R133.10. Sun J, Liu W, Adams TS, Li X, Turner AR, Chang B, Kim JW, ZhengSL, Isaacs WB, Xu J: DNA copy number alterations in prostatecancers: a combined analysis of published CGH studies.  Pros-tate 2007, 67:692-700.11. Lapointe J, Li C, Giacomini CP, Salari K, Huang S, Wang P, Ferrari M,Hernandez-Boussard T, Brooks JD, Pollack JR: Genomic profilingreveals alternative genetic pathways of prostate tumorigen-esis.  Cancer Res 2007, 67:8504-8510.12. Kim JH, Dhanasekaran SM, Mehra R, Tomlins SA, Gu W, Yu J, Kumar-Sinha C, Cao X, Dash A, Wang L, et al.: Integrative analysis ofgenomic aberrations associated with prostate cancer pro-gression.  Cancer Res 2007, 67:8229-8239.13. Goodarzi G, Mashimo T, Watabe M, Cuthbert AP, Newbold RF, PaiSK, Hirota S, Hosobe S, Miura K, Bandyopadhyay S, et al.: Identifica-tion of tumor metastasis suppressor region on the short armof human chromosome 20.  Genes Chromosomes Cancer 2001,32:33-42.14. Walter MJ, Park JS, Ries RE, Lau SK, McLellan M, Jaeger S, Wilson RK,Mardis ER, Ley TJ: Reduced PU.1 expression causes myeloidprogenitor expansion and increased leukemia penetrance inmice expressing PML-RARalpha.  Proc Natl Acad Sci USA 2005,102:12513-12518.15. Dong JT, Lamb PW, Rinker-Schaeffer CW, Vukanovic J, Ichikawa T,Isaacs JT, Barrett JC: KAI1, a metastasis suppressor gene forprostate cancer on human chromosome 11p11.2.  Science1995, 268:884-886.16. Yoshida BA, Chekmareva MA, Wharam JF, Kadkhodaian M, StadlerWM, Boyer A, Watabe K, Nelson JB, Rinker-Schaeffer CW: Pros-tate cancer metastasis-suppressor genes: a current perspec-tive.  In Vivo 1998, 12:49-58.17. Marazuela M, Acevedo A, Adrados M, Garcia-Lopez MA, Alonso MA:Additional file 1Weighted scoring for the selection of candidate genes in 2E5-2F3. Detailed description of the weighted scoring method that was used to select candidate suppressor genes in the differentially deleted region of chromo-some 2 between metastasizing and non-metastasizing allografts.Click here for file[http://www.biomedcentral.com/content/supplementary/1755-8166-2-18-S1.DOC]Additional file 2T-test sliding window whole genome and chromosome 2 analysis. Comparison of expression array results between metastatic and non-met-astatic allografts using a T-test sliding window analysis.Click here for file[http://www.biomedcentral.com/content/supplementary/1755-8166-2-18-S2.PPT]Additional file 3Complete candidate gene list. Complete list of candidate genes from the 2E5-2F3 identified from the function-based bioinformatics filter and the cancer expression signature filter.Click here for file[http://www.biomedcentral.com/content/supplementary/1755-8166-2-18-S3.DOC]Page 8 of 9(page number not for citation purposes)0035 (K.D.T.), NCI Howard Temin Award CA114033 (Y.Y.), and the Expression of MAL, an integral protein component of themachinery for raft-mediated pical transport, in human epi-thelia.  J Histochem Cytochem 2003, 51:665-674.Publish with BioMed Central   and  every scientist can read your work free of charge"BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime."Sir Paul Nurse, Cancer Research UKYour research papers will be:available free of charge to the entire biomedical communitypeer reviewed and published immediately upon acceptancecited in PubMed and archived on PubMed Central Molecular Cytogenetics 2009, 2:18 http://www.molecularcytogenetics.org/content/2/1/1818. Lind GE, Ahlquist T, Lothe RA: DNA hypermethylation of MAL:a promising diagnostic biomarker for colorectal tumors.Gastroenterology 2007, 132:1631-1632. author reply 163219. Mimori K, Nishida K, Nakamura Y, Ieta K, Yoshikawa Y, Sasaki A, IshiiH, Alonso MA, Mori M: Loss of MAL expression in precancer-ous lesions of the esophagus.  Ann Surg Oncol 2007, 14:1670-1677.20. Hatta M, Nagai H, Okino K, Onda M, Yoneyama K, Ohta Y, NakayamaH, Araki T, Emi M: Down-regulation of members of glycolipid-enriched membrane raft gene family, MAL and BENE, incervical squamous cell cancers.  J Obstet Gynaecol Res 2004,30:53-58.21. Mimori K, Shiraishi T, Mashino K, Sonoda H, Yamashita K, YoshinagaK, Masuda T, Utsunomiya T, Alonso MA, Inoue H, Mori M: MALgene expression in esophageal cancer suppresses motility,invasion and tumorigenicity and enhances apoptosis throughthe Fas pathway.  Oncogene 2003, 22:3463-3471.22. Stahl A: A current review of fatty acid transport proteins(SLC27).  Pflugers Arch 2004, 447:722-727.23. Hesson LB, Wilson R, Morton D, Adams C, Walker M, Maher ER,Latif F: CpG island promoter hypermethylation of a novelRas-effector gene RASSF2A is an early event in colon car-cinogenesis and correlates inversely with K-ras mutations.Oncogene 2005, 24:3987-3994.24. Endoh M, Tamura G, Honda T, Homma N, Terashima M, Nishizuka S,Motoyama T: RASSF2, a potential tumour suppressor, issilenced by CpG island hypermethylation in gastric cancer.Br J Cancer 2005, 93:1395-1399.25. Cooper WN, Dickinson RE, Dallol A, Grigorieva EV, Pavlova TV,Hesson LB, Bieche I, Broggini M, Maher ER, Zabarovsky ER, et al.: Epi-genetic regulation of the ras effector/tumour suppressorRASSF2 in breast and lung cancer.  Oncogene 2008,27:1805-1811.26. Zhang Z, Sun D, Van do N, Tang A, Hu L, Huang G: Inactivation ofRASSF2A by promoter methylation correlates with lymphnode metastasis in nasopharyngeal carcinoma.  Int J Cancer2007, 120:32-38.yours — you keep the copyrightSubmit your manuscript here:http://www.biomedcentral.com/info/publishing_adv.aspBioMedcentralPage 9 of 9(page number not for citation purposes)


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items