Identification of Tumor Suppressor Genes via Cell Fusion and Chromosomal Transfer

Loss of DNA or chromosomal deletion was frequently reported in different sporadic tumors, suggesting that those lost regions may contain putative tumor suppressor genes (TSGs). To map and isolate candidate genes from vast randomly lost areas, functional and complementary evidence is usually needed to define these areas to critical regions (CR). This is particularly important when one is dealing with sporadic cancers where clearly defined familial predisposition is present but high cancer risk families are not available for position cloning. A numerous studies have revealed extensive DNA deletions in nasopharyngeal carcinoma (NPC) and esophageal squamous cell carcinoma (ESCC). To identify candidate genes from these cancers, we used cell fusion and microcell-mediated chromosome transfer (MMCT) to introduce the whole chromosome or a chromosome fragment, into NPC and ESCC cell lines. Combined with other molecular approaches, we have successfully identified a number of novel TSGs on various human chromosomes. The nasopharynx is located behind the nasal cavity in the upper part of the pharynx (Fig. 1). NPC is a type of malignancy which arises from the epithelial cells in the nasopharynx. NPC is an unique cancer, which is commonly found in, Southeast Asia, North Africa, Middle East, and Arctic regions, but rare in most parts of the world (Jeyakumar et al 2006, Wei et al 2005). The esophagus is a tube, which is about 25 cm long, for the food passage from mouth to stomach (Fig. 2). Esophageal cancer (EC) is classified into two major histologic subtypes: squamous cell carcinoma and adenocarcinoma (Daly et al 2000). Esophageal adenocarcinoma (EAC) arises from the cells of glands responsible for producing mucous in the esophageal wall. The majority of the cases (>80%) are ESCC in Hong Kong, while EAC shows a climbing incidence in Western countries. EC varies greatly in geographical distribution; high-risk areas include north-central China in Henan and Shanxi (Holmes et al 2007, Qi et al 2005, Wu et al 2006).


Introduction
Loss of DNA or chromosomal deletion was frequently reported in different sporadic tumors, suggesting that those lost regions may contain putative tumor suppressor genes (TSGs).To map and isolate candidate genes from vast randomly lost areas, functional and complementary evidence is usually needed to define these areas to critical regions (CR).This is particularly important when one is dealing with sporadic cancers where clearly defined familial predisposition is present but high cancer risk families are not available for position cloning.A numerous studies have revealed extensive DNA deletions in nasopharyngeal carcinoma (NPC) and esophageal squamous cell carcinoma (ESCC).To identify candidate genes from these cancers, we used cell fusion and microcell-mediated chromosome transfer (MMCT) to introduce the whole chromosome or a chromosome fragment, into NPC and ESCC cell lines.Combined with other molecular approaches, we have successfully identified a number of novel TSGs on various human chromosomes.The nasopharynx is located behind the nasal cavity in the upper part of the pharynx (Fig. 1).NPC is a type of malignancy which arises from the epithelial cells in the nasopharynx.NPC is an unique cancer, which is commonly found in, Southeast Asia, North Africa, Middle East, and Arctic regions, but rare in most parts of the world (Jeyakumar et al 2006, Wei et al 2005).The esophagus is a tube, which is about 25 cm long, for the food passage from mouth to stomach (Fig. 2).Esophageal cancer (EC) is classified into two major histologic subtypes: squamous cell carcinoma and adenocarcinoma (Daly et al 2000).Esophageal adenocarcinoma (EAC) arises from the cells of glands responsible for producing mucous in the esophageal wall.The majority of the cases (>80%) are ESCC in Hong Kong, while EAC shows a climbing incidence in Western countries.EC varies greatly in geographical distribution; high-risk areas include north-central China in Henan and Shanxi (Holmes et al 2007, Qi et al 2005, Wu et al 2006).

Somatic cell fusion and tumor suppression
The theory for specific chromosomes contributing to tumor suppression was originally proven by somatic cell genetics (Harris et al 1969).In their study, Harris and colleagues fused the mouse tumorigenic cells with non-tumorigenic cells to form a hybrid cell; some hybrid cells maintained the phenotypes of the parental cell lines and did not form tumors in the tumorigenicity assay.This suggests that non-tumorigenic cells contain genes, which are dominant and capable of suppressing the tumorigenic cell growth.The tumor suppression was further observed in human cell fusion experiments (Stanbridge, 1976 and1992).The unstable tumorigenic hybrids showed loss of the selected chromosomes, resulting in the reemergence of its tumorigenicity properties.These findings suggested that some chromosomes may contain special regions, which can inhibit tumor formation, and thus provided the basis for further development of monochromosome transfer approaches to study the role of a particular chromosome in tumor development.These earlier experiments led to the hypothesis that the human genome might contain a group of genes suppressing tumor growth.
Adapted from Cancer Research UK and American Society of Clinical Oncology phenotype of cancer cells.The nude mouse tumorigenicity assay allows us to identify genes that fit the ultimate definition of a TSG, which requires functional evidence for the suppression of the malignant phenotype of cancer cells.The functional identification of TSGs in nude mouse assay can exclude the non-specific cytotoxic or cytostatic effects introduced by artificial gene expression.It also identifies a new category of molecules that are not cytotoxic in vitro, but are involved in the signaling cascades in different biological processes that hallmark cancer development such as the cell adhesion and sustained angiogenesis, inflammatory responses, evading immune destruction, and the tumor microenvironment.3. How to identify a TSG from a pool of genes?

Vector
3.1 MMCT An intact single chromosome can be introduced into the recipient cell line using MMCT techniques.The establishment of mouse donor cells with a singl e c o p y o f h u m a n chromosome is described previously (Fournier and Ruddle 1977a, Saxon et al 1985, Anderson and Stanbridge 1993).Figure 3 briefly outlines the procedures of MMCT.Donor cells were seeded on day 1 and incubated with 0.04 to 0.1 µg/ml colcemid for 48 to 56 hours to arrest cells in metaphase.Recipient cells were seeded on day 2 with the cell confluence reaching 80-90% the next day.Enucleation of the donor cells was achieved by centrifugation at 13,000 rpm for 65 to 70 min at 34C with 10 g/ml cytochalasin B (Sigma, MO, USA).The heterogeneous microcell mixture was then filtered through successive polycarbonate filters of 8/5/3 m pore sizes (Whatman, Middlesex, UK) to remove large microcells and other cell debris.Microcells were pelleted by centrifugation and resuspended thoroughly in 1 ml DMEM containing 50 g/ml phytohemagglutinin-p (PHA-P).The recipients were rinsed in PBS once and the microcell/PHA-P suspension was added to the recipient cells and incubated for 20-40 min.Microcells were fused to the recipient cell with 1-2 ml ice-cold PEG-1000.The treated cells were rinsed with PBS four times and fed with medium.The selection medium [hypoxanthine, aminopterin, and thymidine (HAT) + G418] was added on day five.
After about three-four weeks selection, only the recipient cells harboring the transferred chromosome containing the neomycin gene on the exogenously transferred chromosome survive; potential hybrids were picked and expanded for further analysis.

DNA slot blot hybridization
The contamination of mouse DNA in MCHs was excluded by slot blot hybridization (Fig. 4).
Five micrograms of genomic DNA of recipient cell lines, HONE1, microcell donor, MCHs, and A9 in 6X SSC (0.9 M NaCl and 90 mM Na Citrate, pH 7.4) were transferred onto a HybondTM N nylon membrane (Amersham Biosciences, Uppsala, Sweden) with a BIO-DOT® SF slot blot apparatus (BioRad, CA, USA).The samples on the membrane were denatured for 10 min with a denaturing solution (1.5 M NaCl and 0.5 M NaOH) and then neutralized in 1 M NaCl and 0.5 M Tris-Cl, pH 7 for 5 min.The membrane was dried and cross-linked by UV on an UV transilluminator (Stratagene, CA, USA).A Gene ImagesTM AlkPhos DirectTM Labelling and Detection System (Amersham Biosciences, Uppsala, Sweden) was used according to manufacturer's protocol.In brief, a total of 100 ng of A9 DNA was denatured by heat for 5 min.The probe was labeled by covalently linking to alkaline phosphatase by incubating at 37C for 30 min with crosslinker solution and reaction buffer.The blot was prehybridized with 0.125 ml/cm2 hybridization buffer containing 12% w/v urea, 0.5 M NaCl and 4% blocking reagent at 55C for at least 15 min.The labeled probe was added into hybridization buffer and hybridized to the membrane containing samples at 55C overnight.The blot was then washed twice with 55C primary washing buffer (2 M Urea, 0.1% w/v SDS, 50 mM Na phosphate pH 7, 150 mM NaCl, 1 mM MgCl 2 , 0.2% w/v blocking reagent) for 10 min with gentle agitation.The blot was then washed twice with secondary wash buffer (50 mM Tris-Cl, 100 mM NaCl, 2 mM MgCl 2 , pH 10) at room temperature for 5 min.A chemiluminescent signal was generated by adding 40 μl/cm 2 CDP-Star detection reagent (Amersham Biosciences, Uppsala, Sweden) and detected on X-ray film.

Microsatellite typing
Microsatellite typing was used in genomic analysis in this study (Fig. 5).The polymorphic markers, showing a high heterozygosity frequency, were used.PCR was performed to amplify genomic regions of particular markers.PCR reactions were carried out with 100 ng of DNA template, 1X PCR buffer, 2.5 mM MgCl 2 , 200 μM dNTP, 0.1 unit of AmpliTaq Gold DNA polymerase (Applied Biosystems, CA, USA), 0.3 μM of each fluorescence-labeled primer.PCR products were injected into semi-automated ABI PRISMTM 3100 Genetic Analyzer (Applied Biosystems, CA, USA) for capillary electrophoresis and analysis by Genotyper software.All markers were labeled with 6-FAM/VIC/NED fluorescence dye and have annealing temperatures of 55C.

Whole chromosome painting FISH
The confirmation of the presence of an extra copy of intact human chromosome in the hybrids was performed by whole chromosome painting FISH analysis.To obtain metaphase cells, logarithmic growth phase cells were treated with 0.05 g/ml colcemid for 4-6 hours.
After harvesting the cells by centrifugation at 4000 rpm for 5 min, they were subjected to hypotonic treatment (0.075 M KCl) for no more than 25 min at 37C.Cells were fixed with methanol/acetic acid (3:1) overnight and then dropped onto glass slides.The whole chromosome paint (WCP) probes were used and they are a mixture of DNA sequences specific to a particular chromosome that are directly labeled with one of the Vysis™ fluorophores.The slides were denatured at 73C for 5 min [70% (v/v) formamide, 2 X SSC (pH7), 0.1 mM EDTA].Slides were dehydrated for 1 min in 70%, 85%, and 100% ethanol successively.Probes were hybridized to the slides in LSI/WCP® hybridization buffer at 37C in dark for overnight.The slide was washed at 73C in washing solution I (0.4 X SSC, 0.3% NP40) for no more than 2 min, followed by washing with washing solution II ( 2 X SSC, 0.1% NP40), and counterstained with DAPI.Fluorescence signals were captured on an Olympus BX51 fluorescence microscope with the Spot software.

In vivo nude mouse tumorigenicity assay and establishment of tumor segregants (TSs)
After successful chromosome transfer, MCHs containing an exogenous chromosome were obtained and the tumorigenicity assay was then performed.The tumorigenicity of the each NPC/ESCC cell line was investigated by subcutaneous injection of 1 X 10 7 cells into 6-8 week old female athymic BALB/c nu/nu mice.This assay has been repeatedly and successfully used for assessing the degree of tumor suppression in MCH populations (Goyette et al 1992, Saxon et al 1985, Weissman et al 1987).For each tested cell line, at least six sites were injected on three mice (two sites per animal).The animals were monitored regularly for tumor formation and palpable nodules were measured weekly using calipers.
After a long period of incubation in the nude mice (>12 weeks), if notable tumors formed, they were surgically removed aseptically and then washed with PBS, followed by mincing the tissues as small as possible in a petri dish.The minced tissue was covered with 5 ml medium and left without any disturbance for at least three days.Medium was withdrawn and 10 ml of fresh medium was added.The cells were then harvested and were the TSs/revertants used in further analysis with their matched MCHs.

TS analysis
The TSs show a restoration of tumorigenicity in the tumorigenicity assay by elimination of critical regions (CRs) associated with tumor suppression (Cheng et al 2002, Cheung et al 2009).Loss of donor genetic materials detected by microsatellite typing (see Section 3.2.2) was considered significant, when the intensity of the donor to recipient allele ratio in a TS versus its corresponding MCH parental cell line, was equal to or less than 0.8.The significance of these losses was simultaneously verified by bacterial artificial chromosome (BAC) FISH.Furthermore, the differential gene expression profiling between the tumorsuppressive MCH cell lines versus their derived TSs and the highly tumorigenic recipient cell line was revealed by oligo microarray analysis.The BAC FISH and oligo microarray analyses will be discussed in detail.

Bacterial artificial chromosome (BAC) FISH
The BAC DNA probes were prepared by labeling the isolated BAC DNAs by nick translation.
In brief, BAC clones were incubated in 5 ml LB medium containing 12.5 μg/ml chloramphenicol at 37C overnight.Bacterial cells were pelleted down and resuspended in 200 μl of ice-cold alkaline lysis solution I. Four hundred microliters of freshly prepared alkaline lysis solution II were added, mixed by gentle inversion several times, and placed on ice for no more than 5 minutes.A total of 300 μl of ice-cold solution III was added, and the mixture was put on ice for 5 minutes, followed by centrifugation at maximum speed for five minutes at 4C.Supernatant was transferred into a fresh 1.5 ml tube containing 900 μl of isopropanol at room temperature and mixed by gentle inversion.The precipitated DNA was collected by centrifugation at maximum speed for five minutes at 4C; supernatant was discarded and the pellet was rinsed with 1 ml of 75% ethanol.The tube was centrifuged for 5 min at room temperature and ethanol was removed.The DNA pellet was air-dried and dissolved in 50 μl TE.A total of 500 ng of BAC DNA was labeled with 10 μM SpectrumGreen/Orange TM-dUTP (Vysis, IL, USA), nick translation buffer (Vysis, IL, USA), 10 μl nick translation enzyme (Vysis, IL, USA) at 14C for 16 hours and stopped at 70C for 10 minutes.The probe was precipitated with 1 μg of COT-1 DNA (Invitrogen, CA, USA) and 2 μg salmon sperm DNA (Invitrogen, CA, USA), 4 μl water, 0.1 volume of 2 M sodium acetate, 2.5 volumes 100% ethanol and centrifuged at maximum speed for 30 min at 4C.The supernatant was removed.The pellet was air-dried and resuspended in 3 μl water and 7 μl LSI/WCP hybridization buffer.Procedures for the hybridization and detection were the same as that used for the whole chromosome painting FISH.Representative BAC FISH results are shown in Figure 6.

Oligonucleotide microarray
In order to investigate the differentially expressed genes of the tumor-suppressive MCHs versus their matched tumorigenic TSs, oligonucleotide microarray analysis was performed; a TSG is presumably up-regulated in MCHs and down-regulated in TSs (Robertson et al 1997).Therefore, genes with that expression profile are potential TSGs in NPC.This method can also help to identify candidate genes that are downstream of the functional pathways.The 19K or 28K oligonucleotides were spotted on glass slides using a custom-built microarray spotter at the Genome Institute of Singapore.The incorporation of Cy3 and Cy5 to cDNAs, competitive hybridizations, and processing of array images were performed as described (Lin et al 2004).
In brief, a total of 20 μg of total RNA was mixed with 1 μg oligo dT and denatured at 65C for 5 min.Reverse transcription was done by using 1X first strand buffer, 4 μl of 0.1 M DTT, 2 μl 20X low dT/aa-dUTP mix (1 M dATP, 1 M dGTP, 1 M dCTP, 200 mM dTTP, and 800 mM amminoallyl dUTP), 100 unit of RNase OUT inhibitor (Invitrogen, CA, USA), and 400 units of SuperScript II reverse transcriptase (Invitrogen, CA, USA) and incubated at 42C for 60 min.Another 200 units of SuperScript II reverse transcriptase was added to the mixture and then it was incubated at 42C for 60 min.Five microliter of 500 mM EDTA (pH 8) and 10 μl of 1M NaOH were added to the reaction mixture and incubated at 65C for 45 min, followed by adding 25 μl of 1 M Tris-Cl (pH 7.5) and 400 μl ddH 2 O and purified on a YM-30 column (Millipore, MA, USA).One microliter of 500 mM NaHCO 3 (pH 9) and fluorescent dye, Cy3 or Cy5 (GE Health Care Life Sciences, NJ, USA), were added to 9 μl of the cDNA for dye incorporation at room temperature for 1 hour.The chips, which contain synthesized 60-mers of 19K or 28K oligonucleotides, were prehybridized by using DIG Easy Hyb (Roche Diagnostics, Basel, Switzerland) at 42C for one hour.Four microliter of 4 M hydroxylamine (Sigma, MO, USA) was added to the labeled cDNA at room temperature for 15 min.Thirty-five microliter of 100 mM NaOAc (pH 5.2) was added and mixed with the paired sample labeled with Cy3 and Cy5.The reaction mixtures were purified on YM-30 columns, containing 40 μg of herring sperm DNA (Invitrogen, CA, USA) and 22 μl DIG Easy Hyb were added, followed by denaturing at 65C for 5 min.The cDNA probe was hybridized onto slides and incubated in the MAUI hybridization chamber (BioMicro Systems, UT, USA) at 42C for 16 hours.After hybridization, slides were washed with 2X SSC + 0.1% SDS for 30 s, 1X SSC for 30 s and 0.2X SSC for 30 s and 0.05X SSC for 5 s.The slides were then scanned with GenePix 3100 microarray scanner and analyzed by GenePix Pro 4.0 and 5.0 software (Axon Instruments, CA, USA).The results were uploaded to the database from Genome Institute of Singapore (GIS MAdb, http://gismadb.gis.a-star.edu.sg).The hierarchical clustering was performed by CLUSTER program and the results were viewed by the TREE VIEW program (M.Eisen; http://rana.lbl.gov/EisenSoftware.htm).Stringency was set so that only data sets with expression ratios higher than 1.4 in hybrids/recipient cells and ratios less than 0.6 in hybrids/TSs for both duplicates were selected for further studies.

Methylation study of candidate tumor suppressor genes
Promoter hypermethylation is a major mechanism for silencing TSG expression.Promoter hypermethylation only occurs at cytosine (C), which is located on the 5' end of guanine (G) of promoter CpG islands.Addition of a methyl group to cytosine by methyltransferase blocks the binding of the activator or inhibitor to the promoter sequence and consequently there is transcriptional silencing of the gene.By using 5-aza-2'-deoxycytidine, which can demethylate the methyl groups in the CpG islands, gene expression can be restored.Using this approach, methylation status of genes of interest may be elucidated.Besides using the 5aza-2"-deoxycytidine, bisulfite genomic sequencing (BGS) and methylation-specific PCR (MSP) are other important methods used to analyze promoter hypermethylation.In the bisulfite treatment, the genomic DNAs are modified.All of the unmethylated cytosines are converted to uracils, while the methylated cytosines remain unchanged.By analyzing the promoter sequences and using sequence-specific primers, the promoter hypermethylation status can be determined.

Bisulfite treatment
Genomic DNA was treated with sodium metabisulfite to analyze the promoter hypermethylation status.Two micrograms of DNA were dissolved in 10 μl of water, followed by adding 1.1 μl of 3N NaOH and incubation at 37C for 15 min.Then 104 μl of urea/metabisulfite solution (6.24 M urea and 2 M sodium metabisulfite; Sigma, MO, USA), 6 μl of 10 mM hydroquinone (Sigma, MO, USA) and 100 μl of mineral oil were added into the mixture and incubated at 55C for no more than 15 hours.The mixture was then purified by QIAquick PCR purification Kit (Qiagen, CA, USA) according to the manufacturer's protocol and eluted with 200 μl elution buffer.The purified DNA was denatured in 23 μl of 3N NaOH at 37C for 15 min.The DNA was precipitated by adding 2 μl of tRNA (10 μg/μl; USB, OH, USA), 50 μl of 10 M ammonium acetate, and 500 μl of 100% ethanol, and stored at -20C for 30 min.The mixture was centrifuged at 12K rpm at 4C for 20 min; the supernatant was removed.The DNA pellet was washed with 500 μl of 70% ethanol and centrifuged at full speed at 4C for 20 min.The DNA pellet was air-dried and dissolved in 100 μl of 10 mM Tris-Cl (pH 8.5) solution.

Bisulfite genomic sequencing (BGS) and methylation-specific PCR (MSP)
BGS and MSP PCR reactions were carried out with 5 μl of bisulfite-treated DNA template, 1X PCR buffer, 2.5 mM MgCl 2 , 200 μM dNTP, 0.1 unit of AmpliTaq Gold DNA polymerase (Applied Biosystems, CA, USA), 0.3 μM of primers.For the BGS, PCR products were then cloned into the TA cloning pMD18T Simple vector (TaKaRa Biotechnology, Dalian, China) and transformed into competent cells.Mini-preparation was performed to extract the plasmid DNA.For each examined cell line, five individual clones were checked to confirm the promoter hypermethylation status.The plasmids were then sequenced by using pMD18T7 forward and reverse sequencing primers and BigDye 3.1 terminator.Signals were detected using an ABI PRISMTM 3100 Genetic Analyzer (Applied Biosystems, CA, USA).

5-Aza-2'-deoxycytidine treatment
The re-expression of the candidate genes silenced by promoter hypermethylation was investigated using a demethylation reagent.Cells were treated with 5 μM 5-aza-2'deoxycytidine (Sigma, MO, USA) for five days; freshly diluted drug was changed daily.
After 5 days treatment, the cells were harvested for RNA extraction.

Other functional assays we use to study the tumor suppressive phenotypes
In order to examine the tumor suppressive phenotypes of NPC and EC cells expressing a candidate TSG, various functional assays were performed according to the major hallmarks of cancer (Hanahan et al 2011).Simple proliferation, cell cycle, colony formation, and soft agar assays are routinely performed to assess the inhibitory effects of the transgenes on cell growth ability in normal culture conditions, the change of cell cycle status, the ability of colony formation from a single cells, and the cell growth potential in an the anchorageindependent environment.In addition, the migration, invasion, and angiogenic properties of the tumor suppressive transfectants were also examined.Some of those functional assays will be discussed in detail.

Three-dimensional matrigel culture
Matrigel contains a gelatinous protein mixture of extracellular matrix proteins such as laminin and collagen, which recapitulates the natural environment of tumor cells for better investigation of the functional impact of a candidate TSG.We aim to see whether there is any change in the numbers, sizes, and morphologies of colonies of transfectants with the transgenes versus the vector-alone, when grown in matrigel.In brief, matrigel basement membrane matrix (BD Biosciences, San Jose, CA, USA) was thawed on ice and 100 μl was coated as a bottom layer onto each well of a 24-well cluster plate.Subsequently, a total of 5000 cells resuspended in 0.5 ml growth medium was seeded on top of this bottom matrigel layer.After 2 weeks, images were captured using an inverted light microscope (Nikon TMS, Ontario, Canada) at 20X magnification.

Cell migration and invasion assays
The cell migration and invasion abilities of cells were determined using a micropore chamber assay (BD Biosciences, San Jose, CA, USA).The chamber used for migration study consists of an 8 m pore size PET membrane at the bottom of the insert, while the chamber used for invasion study is a migration chamber coated with a thin layer of matrigel basement matrix.The migration and invasion abilities of each clone were determined by the number of cells passing through the membrane and the matrigel basement matrix, respectively.

HUVEC tube formation assay
The tube formation assay is an in vitro angiogenesis assay to assess the tube formation ability of vascular endothelial cells.Briefly, the Human Umbilical Vein Endothelial Cell (HUVEC) was cultured in collagen pre-coated TC dishes.The conditioned medium of each cell line was collected by incubating the cells with DMEM/RPMI serum-free medium for 24 hours.The 96-well plate pre-coated with 50 μl growth factor-reduced Matrigel (BD Bioscience, San Jose, CA, USA) per well was used.A total of 4 x 10 4 HUVEC cells resuspended in 100 μl conditioned media supplemented with 1% serum were seeded in each well.The HUVEC cells were then incubated at 37C for 6 hours to allow the formation of tube-like structures.The tubes formed were captured under microscopy and the total tube length formed by each sample was measured by Spot software (Diagnostic Instruments Inc, Sterling Heights, MI, USA).

Matrigel plug angiogenesis study
In vivo angiogenesis study was performed by matrigel plug angiogenesis analysis.A total of 5 x 10 6 cells mixed with 50 μl DMEM and 250 μl ice-cold matrigel was injected into nude mice subcutaneously to allow gel plug formation.Each sample was injected into one flank of five female athymic nude mice.The matrigel mixed with the cells polymerizes to form a solid gel plug, which allows cell growth and blood vessel formation.After inoculation for 7 days, the matrigel plug was excised, fixed with formalin, embedded in paraffin, sectioned, and mounted onto slides.The slides were stained with hematoxylin and eosin (H&E) for histological observation.The blood vessels formed were stained with monoclonal antibody.
The slide was scanned and the signal was quantified by ImageScope v10 software (Aperio, Vista, CA, USA).

Human angiogenesis antibody array
The angiogenesis-related proteins excreted from the TSGs versus the vector-alone transfectants were detected by the human angiogenesis antibody array (RayBiotech, Norcross, GA, USA).The array included two membranes pre-coated with 43 angiogenesis cytokine antibodies.The conditioned media were obtained as described in Section 3.6.3.After blocking and complete washing, the membranes were hybridized with biotinconjugated antibody at room temperature for 2 hours, followed by hybridizing with HRPconjugated streptavidin at room temperature for 2 hours.The membranes were then washed again and hybridized with the detection buffer.The signals were exposed to a X-ray film and quantified by Quantity One ® software (Biorad, Hercules, CA, USA).

Examples of candidate TSGs studied via MMCT and other functional approaches
The technique of microcell fusion to transfer single normal chromosomes was used as a functional assay for the TSG activity (Goyette et al 1992, Saxon et al 1986, Weissman et al 1987).This has been particularly useful in confirming TSG functions associated with specific chromosomes, where the mapping location is suspected, but no candidate TSG has been cloned.In addition, by transferring chromosomes possessing interstitial deletions, it has been possible to map the location of TSGs more precisely (Dowdy et al 1991).The analysis of tumor biopsy specimens, xenografts, and derived cell lines have identified an extensive chromosomal alterations in human genome, including those derived from chromosomes 1, 3, 9, 11, 12, 13, 14, 16, and17 (Huang et al 1989, Mitelman et al 1983), Consistent with Knudson's "two-hit" theory, studies of chromosomal loss or deletion may provide useful clues for identification of critical genes involved in inherited cancers.However, it should be appreciated that the limitations of deletion or loss studies in sporadic cases may miss critical regions for several reasons.First, the minimally deleted regions may be obscured when all of the genome shows some degree of LOH, which is common in many sporadic tumors.Second, vast evidence has shown that some TSGs, notably p16 are silenced or imprinted by epigenetic events such as methylation, irrespective of their LOH status in tumors.Third, LOH was studied in primary tumors with karyotypic complexity, gene dosage changes, and contamination by normal cells.Finally, although LOH is commonly detected in practically all types of human cancers, neither LOH nor CGH (comparative genomic hybridization) studies provide functional evidence.To map novel TSGs, functional and complementary approaches are needed for distinguishing a critical region from extensive randomly lost areas.Chromosome 3 was the first chromosome used for the functional studies of tumor suppression in NPC.In this study we used a series of intact and deleted copies of human chromosome 3 derived from normal cells, with discrete interstitial deletions in the p arm, for transfer into the tumorigenic NPC HONE1 cell line.By using the MMCT approach, we successfully transferred these chromosome 3 fragments into HONE1 cells and localized a tumor suppressive region on this chromosome.Comparison of the tumorigenic potential of the MCHs containing these exogenous chromosome 3 fragments identified chromosome 3p21.3as the first tumor suppressive region in NPC (Cheng et al 1998).This area was subsequently confirmed to harbor several TSGs associated with the development of NPC and other common cancers (Lerman and Minna 2000, Lo et al 2001, Yau et al 2006, Hesson et al 2007).Using these functional approaches, we also investigated known and candidate TSGs in other chromosomes, including chromosomes 9, 11, 13, 14 and 17.These chromosomes contain important TSGs, such as p16, RB and p53 and a number of newly-identified TSGs.However, we only introduce a few of examples of these studies in this chapter due to the space limitation.(Cheng et al, 2000, 2002, 2003and 2004, Ko et al 2005and 2008, Lung et al 2008aand 2008b, Cheung et al 2009, Lo et al 2007and 2010)

Chromosome 3: ADAMTS9
In a subsequent study, an intact and two truncated human chromosomes 3 obtained from the same panel of chromosome 3 donor cells, were transferred into the highly tumorigenic ESCC SLMT-1 cell line.Similarly, the ability of these transferred chromosomes to functionally complement defects in the ESCC cell line was assessed by examining the impact of this transfer on tumorigenic potential in nude mice.PCR-microsatellite and BAC FISH analyses were used to narrow down and identify the CR associated specifically with tumor suppression.A 1.61 Mb CR located between markers D3S1600 and D3S1285 was found to be necessary for the tumorigenic suppression of ESCC.These findings further suggest that the CR present in the exogenous chromosomes contains functional tumor suppressive elements.
In the study, we identified a candidate TSG, ADAMTS9, and one non-coding RNA, ENST351926 mapping to 3p14.2, which are located in the chromosome 3 CR of ESCC.The expression of ADAMTS9 in tumor suppressive MCHs was confirmed by reverse transcription (RT)-PCR.The positive expression of the gene was observed in all tumor suppressive MCHs, but was not found in tumorigenic MCHs and TSs, strongly suggesting that ADAMTS9 plays an important role in tumor suppression.The pseudogene is located upstream of the ADAMTS9 promoter region and whether it can serve as a riboregulator or gene expression regulator remains to be determined (Lo et al 2007).

ADAMTS9 (A disintegrin-like and metalloprotease with thrombospondin type 1 motif 9)
As described, using a functional genomic mapping approach, we identified a CR for tumor suppression at 3p14.2 and discovered the important role of A Disintegrin-like And Metalloprotease with ThromboSpondin type 1 motif 9 (ADAMTS9), a gene previously mapped to this region (Clark et al 2000) in ESCC (Holmes and Vaughan 2007).ADAMTS9 encodes a member of a large family of 19 metalloproteases involved in maturation of precursor proteins, extracellular matrix remodeling, cell migration, and inhibition of angiogenesis (Apte 2004, Porter et al 2005).Although the related matrix metalloproteases and ADAM proteases have been clearly implicated in tumor progression and angiogenesis, the role of ADAMTS proteases in cancer is less clearly defined.ADAMTS1 was first identified as an anti-angiogenic molecule (Vazquez et al 1999), and shown to have antitumor effects.Recently, methylation studies identified another family member ADAMTS18 as crucial in several human cancers including NPC and ESCC (Jin et al 2007).
As mentioned in the previous section, ADAMTS9 was identified as one of the differentially expressed genes in these non-tumorigenic MCHs and their matched TS cell lines in our previous study (Lo et al 2007).It was suggested that ADAMTS9 is associated with tumor suppression in human esophageal cancer.Promoter hypermethylation contributes to ADAMTS9 gene silencing in ESCC (Lo et al 2007, Lung et al 2008b).However, the functional impact of ADAMTS9 on cancer development had not been explored.In the follow-up study, we evaluated the hypothesized anti-angiogenic and tumor suppressive functions of ADAMTS9 in ESCC, by stringent tumorigenicity and matrigel plug angiogenesis assays (Lo et al 2010).ADAMTS9 activation suppressed tumor formation in nude mice.In vivo angiogenesis assays revealed a reduction in microvessel numbers in gel plugs injected with tumor-suppressive cell transfectants.Similarly, conditioned media from cell transfectants dramatically reduced the tube-forming capacity of HUVECs.By using the angiogenesis antibody array, we found that these activities were associated with a reduction in expression levels of the pro-angiogenic factors, MMP9 and VEGFA, which were consistently reduced in ADAMTS9 transfectants.Based on the deletion patterns of the ADAMTS9 transcripts in tumors and a TS derived from the tumorigenic transfectants, we speculate that the tumorsuppressive activity of ADAMTS9 in ESCC was associated with the thrombospondin (TSP) domains in the C-terminal region of the gene.Taken together, our data strongly suggest that ADAMTS9 plays a critical role in the "angiogenic switch" and transforms in the ESCC cell lines from a pro-angiogenic to a non-angiogenic phenotype.

Chromosome 9: ENG, DEC1
The transfer of chromosome 9 containing an interstitial deletion at 9p21 (where the wellknown TSG, p16 is located) to NPC HONE1 cell line did not result in tumor suppression in the nude mouse assay, but p16 cDNA suppressed growth of HONE1 cells in vitro assay.It suggests that p16 gene plays an important role in this cancer (Cheng et al., 2000).The similar transfer of chromosome 9 into SLMT1 provided the first functional tumor suppressive evidence in ESCC (Yang et al 2005).The result suggested that gene(s) other than p16 on chromosome 9 is (are) important for ESCC tumorigenesis.The ESCC chromosome 9 MCHs exhibited a delayed latency period in tumor formation compared with that of the parental SLMT1 cells.The delay in tumor growth kinetics was hypothesized to be associated with the loss or inactivation of wild type alleles from the exogenous transferred donor chromosome 9. Detailed microsatellite marker-PCR deletion mapping analysis of the tumor suppressive chromosome 9 MCHs and their corresponding derived TSs delineated that the critical regions that may harbor candidate TSGs to a 2.4 Mb region at 9q33-q34 around D9S112.

ENG (Endoglin)
The MMCT-identified CR at 9q32-34 is a gene-rich region.ENG (Endoglin), mapping to 9q33-q34.1, is a component of the transforming growth factor beta (TGF-) receptor complex and is involved in tumor angiogenesis by modulating the biological effect of TGF-.Significant down-regulation of ENG was detected at frequencies of 87  (Craft et al 2007).In endothelial cells, high endoglin expression stimulates the type I activin receptor-like kinases (ALK1) pathway and indirectly inhibits ALK5 signaling for endothelial cell proliferation and migration, thus promoting the state of angiogenesis.ENG may act differently in cancer cell versus endothelial cells.ENG may suppress cancer cell proliferation in the pre-malignant stage; meanwhile, its expression could promote angiogenesis facilitating the cancer progression in the late malignant stage.The mechanism of ENG in suppressing ECSC tumor requires further study (Wong et al 2008).

DEC1 (Deleted in Esophageal Cancer 1)
DEC1 (Deleted in Esophageal Cancer 1) is in the vicinity of the CR at 9q32-34 and is downregulated frequently in ESCC cell lines and tumor tissues (Leung et al 2008).The DEC1 protein localizes to both the cytoplasm and nucleus.The vesicular pattern of DEC1 in the cytoplasm appears to localize at the Golgi and Golgi-endoplasmic reticulum intermediate compartment.DEC1 is clinically important as the tissue microarray (TMA) study suggested an association of DEC1 expression with lymph node metastasis, early onset ESCC, and familial status (Wong et al 2011b).DEC1 stably transfected clones provided functional evidence for cell growth inhibition in vitro and significant delay in tumor growth in vivo (Yang et al., 2005).DEC1 stable clones showed significantly fewer colony numbers as compared to the vector-alone control.Restoration of DEC1 expression also negatively affected anchorage-independent growth properties of an ESCC cell line (Leung et al 2008).
Using cDNA microarray analysis to reveal the differential expression profiling between tumor suppressive DEC1 clones versus the vector-alone transfectant, DUSP6 (dual-specificity phosphatase 6) was identified as one of the downstream targets of DEC1, as it is up-regulated in DEC1 stable transfectants, C4 and C9, compared to vector-alone stable transfectants.It is expected that in clinical specimens, the higher expression of DEC1 associates with higher expression of DUSP6.This association was observed in tumor tissues in younger aged ESCC patients group.This association was only significant in these two groups, which was limited by the number of available samples (only 26 normal counterpart tissues and 74 tumor tissues).Subsequent functional study of DUSP also indicated that DUSP6 plays a crucial role for ESCC carcinogenesis by inhibiting cell invasion and impairing the epithelialmesenchymal transition (EMT)-associated phenotype (Wong et al 2011a).

Chromosome 13: THSD1
Chromosome 13q deletions are frequent events in several human cancers, including ESCC (Hu et al 2000, Pack et al 1999), nasopharyngeal (Tsang et al 1999) and lung (Tamura et al 1997) cancers.Molecular evidence of gross deletions which implicate the existence of TSGs came from numerous LOH and CGH studies.Our group has pioneered the identification of the TSGs on chromosome 13 by the functional complementation approach in both NPC and ESCC (Cheng et al 2004, Ko et al 2008).The transfer of intact chromosome 13 in HONE1 cells identified a critical region essential for the viability and growth of NPC MCHs at chromosome 13q12, but it was not RB gene (Cheng et al 2004).By the microsatellite deletion mapping, D13S893 at 13q12, a minimally deleted region of 0.7 Mb, was found to be nonrandomly eliminated in the six chromosome 13 MCHs bounded by markers D13S1287 and D13S260.The growth suppressive activity involved at least one novel growth control gene for NPC tumorigenesis.Our subsequent study employed the same MMCT technique in an ESCC model with the ESCC cell line, SLMT1.The transfer of an intact chromosome 13 into this highly tumorigenic recipient cell line conferred tumor suppressive activity, and identified critical regions at 13q12.3, 13q14.11,and 13q14.3.Of interest, a 0.373 Mb at the critical region 2 (CR2), mapped to 13q12.3 and was co-localized to the same CR identified in the NPC model system in our chromosome 13 transfer study (Cheng et al 2004, Ko et al 2008).

THSD1 (Thrombospondin type I domain-containing 1)
The ESCC functional studies implicate the importance of chromosome 13q14 in tumor suppression; TS microsatellite-deletion mapping analysis localized two CRs (CR3 at D13S263 and CR4 at D13S133) at chromosomal region 13q14, which are frequently eliminated (Ko et al 2008).Differential gene expression profiles of a reference immortalized normal esophageal epithelial cell line, three tumor-suppressing MCHs, and their tumorigenic parental SLMT1 cell line were revealed by cDNA oligonucleotide microarray analysis.Nine 13q14 candidates genes, including RB1, were identified to show downregulation in SLMT-1 as compared to NE1, the immortalized normal esophageal epithelial cell line, and the MCHs.RB1 is a well-known TSG mapped to 13q14, but our Western blot analysis indicated that the active form of RB was not increased in the tumor suppressive MCHs (data not shown).

Chromosome 14: LTBP-2
Chromosome 14 loss is commonly found in different cancers, including esophageal (Ihara et al 2002), renal (Yoshimoto et al 2007), lung (Weir et al 2007), and colon cancers (Mourra et al 2007).In NPC extensive chromosome 14 allelic loss has been reported (Dodd et al 2006, Lo et al 2000, Lung et al 2001, Shao et al 2002).This suggests the importance of chromosome 14 in tumor development.In our earlier NPC study, chromosome 14q11.2-13.1 and 14q32.1 regions were found to associate with tumor suppression.In those chromosome 14 MCHs, non-random eliminations of two CRs were consistently observed and associated with tumor growth in tumorigenicity assays (Cheng et al 2003).In a later study, a new panel of chromosome 14 MCHs with an intact exogenous chromosome 14 was established.The potent ability to suppress tumor growth in the in vivo tumorigenicity assay of all intact chromosome 14 MCHs suggests the ability of chromosome 14 to suppress tumor formation in HONE1 cells.This is consistent with chromosome 14 harboring candidate TSGs involved in NPC development (Cheung et al 2009).An intact chromosome 14 was also transferred into the ESCC SLMT1 cell line.The tumorigenic potential of microcell hybrids containing the transferred chromosome 14 provided functional evidence that tumor-suppressive regions of chromosome 14 are essential for esophageal cancer.TSs emerging in nude mice during the tumorigenicity assay was analyzed by detailed PCR-microsatellite typing and dual-colour BAC FISH to identify critical non-randomly eliminated regions.A 680-kb CR mapped to 14q32.13 and an approximately 2.2-Mb CR mapped to 14q32.33 were delineated.(Ko et al 2005).Microarray differential gene expression profiling of tumor-suppressive chromosome 14 MCH cell lines and their tumorigenic TSs identified LTBP-2 (latent transforming growth factor β binding protein 2) mapped to 14q24 as a candidate TSG important for ESCC (Chan et al 2011).

LTBP-2 (Latent transforming growth factor β binding protein 2)
The extracellular matrix (ECM) protein LTBP-2 (Latent transforming growth factor β binding protein 2) encodes a secretary protein that functions as a component of the ECM microfibrils and belongs to the LTBP/fibrillin family (Chan et al 2011).Unlike other members in LTBP family, LTBP-2 does not form complexes with the small latent TGF-s.Interestingly, in addition to the first reported tumor suppressor role in ESCC, LTBP-2 is related to congenital glaucoma and rheumatoid arthritis which are eye and bone diseases, respectively.LTBP-2 expression at the mRNA and protein levels was down-regulated in both ESCC cell lines and primary tumors.One of the mechanisms responsible for the down-regulation of LTBP-2 is via promoter hypermethylation.Restoration of LTBP-2 in an ESCC cancer cell line resulted in tumor suppression in nude mouse assay, which is partially explained by the significant reduction of colony-forming ability on matrigel 3D culture and anchorage-independent growth in vitro.Further in vitro functional characterization of LTBP-2 demonstrated its inhibitory role in angiogenesis, migration, and invasion of cancer cells.An angiogenesis protein array analysis of conditioned medium from LTBP-2 stable clone revealed the change in expressions of different cytokines, including GM-CSF, RANTES, VEGF, uPAR, I-309, MMP-1, Angiopoietin-1, and MCP-1, which in turn induce a less favorable microenvironment for angiogenesis and tumor growth.ESCC is a deadly disease and patients are usually diagnosed at late stage.In many late stage tumors, the TGF-signaling pathway is involved in the activation of EMT program, which is responsible for cancer cell traits promoting malignancy.In contrast, TGF-is well-known to be anti-proliferative.
Interestingly, an inverse correlation of high LTBP-2 and survival in advanced ESCC stage was detected by IHC staining of primary ESCC tissues.LTBP-2 may indirectly regulate TGFby competing with LTBP-1 for fibrillin-1 binding site (Hirani et al 2007).

Conclusions
In this book chapter, we focus on using MMCT, as a functional approach to identify candidate TSGs.The monochromosome transfers of selected chromosomes into the NPC and ESCC cell lines, HONE1 and SLMT1, were performed to determine whether tumor suppressing activities for NPC and ESCC mapped to chromosomes 3, 9, 11, 13, 14 and 17, as described in our previous reports.While all these experiments have been performed in the two NPC and SLMT1 cell lines HONE1 and SLMT1, it should be appreciated there are a few of well-established and well-characterized NPC and ESCC cell lines available for this kind of study.Not surprisingly, identification of TSGs or regions in these cell lines was subsequently confirmed to be important in other NPC and ESCC cell lines as well as primary tumors.By using the MMCT approach, we successfully identified several CRs associated with tumor suppression in the HONE1/SLMT1 cell line systems.These candidate TSGs mapping to these regions were subsequently studied for their role in tumor suppression assays.We discovered that ADAMTS9 at 3p14.2, ENG and DEC1 at 9q33-q34, are genes mapped into CRs, and important for tumor suppression in ESCC.Gene expression profiling of the 19K and 28K oligonucleotide microarrays, including tumor-suppressive MCH and tumorigenic TS cell lines, was utilized to identify candidate genes (THSD1 and LTBP-2) within the critical tumor suppressive regions.It is clear now that many genes contribute to the development of these two important cancers.

Fig. 5 .
Fig. 5. PCR microsatellite typing of a representative donor, recipient, and hybrid cell line.The recipient HONE1, donor MCH556.15, and MCH HK11.13 cell lines were analyzed with the D11S1394 primers.The combined peak patterns of the hybrid cell line show successful transfer of chromosome 11 in recipient cells.

Table 1 .
Gene transfer methods into mammalian cells.
www.intechopen.comIdentification of Tumor Suppressor Genes via Cell Fusion and Chromosomal Transfer 57 .5% in 16 ESCC cell lines, 39.1% directly in 23 ESCC tumor specimens from Hong Kong, and 33.4% in 18 ESCC tumor specimens from the high-risk ESCC region of Henan, China.Both epigenetic methylation and allelic loss appear to contribute to ENG down-regulation in ESCC tumors.Subsequent functional studies with restoration of ENG in an ESCC cell line demonstrated that ENG plays a critical role in ESCC carcinogenesis.Colony formation efficiency was significantly reduced by over-expression of ENG.In addition, significantly smaller colonies of ENG stable transfectants were formed in Matrigel culture.Significant suppression of invasion efficiency and tumorigenicity were also observed, when comparing the ENG stable transfectants with the vector-alone transfectants.No report had yet verified the functional role of ENG in ESCC tumor cells.This study provides evidence supporting ENG, as a cell invasion and tumorsuppressing gene in ESCC.ENG may be functionally involved in TGF-signaling.Downregulation of ENG in esophageal cancer in this study may provoke cancer progression through blocking the tumor suppression of the TGF-signaling cascade.For the functional impact of ENG in cell migration, ENG may suppress cancer cell motility by a TGF-dependent mechanism involving activation of the type I TGF-receptor and Smad1, as reported in the study of prostate cancer cells (de Fraipont et al 2001)novel candidate TSG(s) other than RB1 should be involved in the observed tumor suppression.RT-PCR was performed for KIAA0853, ESD, CHC1L, PHF11, RFP2, FLJ11712, THSD1, and C13orf9.Real-time PCR results validated the frequent down-regulation of THSD1 and PHF11 in ESCC cell lines.THSD1 is located between FLJ11712 and C13orf9 within 13q14.3, but only specific loss of THSD1 expression in all cancer cell lines was detected.Since THSD1 showed a more prominent loss than that of PHF11, it was the first target chosen for further functional characterization.Epigenetic silencing and LOH were the mechanisms responsible, at least in part, for the loss of THSD1 expression in ESCC tumorigenesis.The wild type THSD1 transfection in SLMT1 resulted in significant reduction of colony formation ability, providing evidence for a growth suppressive role of THSD1 in ESCC tumorigenesis.The function of THSD1 is unknown.It encodes a transmembrane molecule containing a thrombospondin type 1 repeat (TSR), which may be involved in cell adhesion and angiogenesis(de Fraipont et al 2001).Interestingly, analysis of the differential expression levels of this gene in previous microarray studies show that high THSD1 expression positively correlated with a better distant metastasis survival in breast cancer patients.This is consistent with its loss possibly being associated with metastatic tumor spread; studies are needed to evaluate its potential importance as a biomarker for esophageal carcinoma.Further functional studies on THSD1 are now underway to elucidate its tumor suppressive role.