Loss of heterozygosity on the long arm of human ... - Nature

Oncogene (1997) 15, 2727 ± 2733  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Loss of heterozygosity on the long arm of human chromosome 7 in sporadic renal cell carcinomas Viji Shridhar1, Qi C Sun2, Orlando J Miller3, Gregory P Kalemkerian1, John Petros2 and David I Smith4 1

Karmanos Cancer Institute, Division of Hematology/Oncology, Department of Internal Medicine; Wayne State University School of Medicine, Detroit, Michigan 48201; 2Department of Urology and Winship Cancer Center, Emory University School of Medicine, Atlanta, Georgia 30322; 3Center for Molecular Medicine and Genetics, Wayne State University School of Medicine, Detroit, Michigan 48201; and 4Division of Experimental Pathology, Department of Laboratory Medicine and Pathology, Mayo Foundation, Rochester, Minnesota 55905, USA

Cytogenetic and molecular analysis of DNA sequences with highly polymorphic microsatellite markers have implicated allele loss in several chromosomal regions including 3p, 6p, 6q, 8p, 9p, 9q, 11p and 14q in the pathogenesis of sporadic renal cell carcinomas (RCCs). Deletions involving the long arm of chromosome 7 have not been described in RCCs although they have been seen in several other tumor types. However, there have been no detailed analysis of loss of heterozygosity (LOH) of 7q sequences in sporadic RCCs. We therefore studied LOH for DNA sequences on 7q with 10 highly polymorphic markers in 92 matched normal/tumor samples representing sporadic RCCs including papillary, nonpapillary, and oncocytomas in order to determine whether allelic loss could be detected in a tumor type with no visible 7q rearrangements at the cytogenetic level. We found chromosome 7q allele loss in 59 of 92 cases (64%) involving one, two, or more microsatellite markers. The most common allele loss included loci D7S522 (24%) and D7S649 (30%) at 7q31.1-31.2, a region that contains one of the common fragile sites, FRA7G. By comparative multiplex PCR analysis, we detected a homozygous deletion of one marker in the 7q 31.1-31.2 region in one tumor, RC21. These results support the idea that a tumor suppressor gene in 7q31 is involved in the pathogenesis of sporadic renal cell carcinomas. Keywords: 7q31.2 loss; renal cell carcinoma; fragile sites

Introduction Renal cell carcinoma (RCC) is the most common tumor of the adult kidney. Loss of DNA sequences from chromosome 3 may be one of the leading events in the development of RCCs (Zbar et al., 1987; Anglard et al., 1994; Ogawa et al., 1991), but allele loss from other chromosomal regions points to the involvement of additional tumor suppressor genes in the pathogenesis of RCCs, as it is in several other solid tumors including those of colon, breast, lung, and bladder (Fearon et al., 1990a; Callahan et al., 1989; Tsai et al., 1990; Weston et al., 1989). In RCCs, a cytogenetically detectable deletion of chromosome 7 sequences has not been reported. Correspondence: DI Smith Received 2 June 1997; revised 23 July 1997; accepted 24 July 1997

However, standard cytologenetic techniques do not detect the entire spectrum of chromosomal changes including microdeletions. LOH analysis with highly polymorphic microsatellite markers is a sensitive molecular method to screen for changes involving allele loss in tumors. LOH for microsatellite markers in 7q31.1-32 may be a common event in the etiology of breast cancer (Zenklusen et al., 1994a; Bieche et al., 1992), ovarian cancer (Zenklusen et al., 1995a), pancreatic cancer (Achille et al., 1996), prostatic cancer (Zenklusen et al., 1994b; Takahashi et al., 1995), and squamous cell carcinomas of the head and neck (Zenklusen et al., 1995b). The smallest region of frequent allele loss contains the two markers D7S522 and D7S649 in 7q31.1-31.2, a region that also contains FRA7G (Berger et al., 1985; Barbi et al., 1984), an aphidicolin-inducible fragile site. The biological signi®cance of at least some fragile sites is that they might predispose chromosomes to breakage which would then result in translocations, deletions, or even ampli®cations that could play a role in tumor development. For example, we have demonstrated that a high percentage of both pancreatic adenocarcinomas (Shridhar et al., 1996) and sporadic RCCs (Shridhar et al., 1997) have breakpoints in FRA3B, the most common fragile site in the human genome. The inhibition of tumorigenicity of a murine squamous cell carcinoma (SCC) cell line by microcell mediated chromosome transfer containing human chromosome 7 provides additional evidence for the presence of a tumor suppressor gene on 7q (Zenklusen et al., 1994c). The inserted chromosome delayed the onset of tumors by 2 ± 3-fold and in some cases competely repressed the tumorigenic potential of the SCC cell line. Insertion of an intact chromosome 7 into an immortalized human ®broblast cell line with 7q3132 LOH suppressed the immortality of these cell lines and restored their ability to senesce (Ogata et al., 1993). Allele loss may not be the only change in chromosome 7 in RCCs. Trisomy and tetrasomy of chromosome 7 was observed in 15 of 19 patients with cytogenetically abnormal RCCs (Miles et al., 1988). We have used an extensive set of highly polymorphic microsatellite markers on 7q21-qter to look for LOH and increase in copy number. Our results indicated there is chromosome 7q allele loss in 59 of 92 cases of RCCs; one of these 59 tumors had a homozygous deletion of a single marker in the 7q 31.1-31.2 region.

7q LOH in renal cell carcinoma V Shridhar et al

2728

Results We examined loss of heterozygosity (LOH) using 10 polymorphic microsatellite markers on the long arm of chromosome 7 in 92 pairs of renal cell carcinoma samples (RCCs). Six of the 10 markers chosen for this analysis map to 7q 31.1-31.2, a region that has been shown to have a high frequency of allele loss in several other tumor types (Zenklusen et al., 1994a, 1994b, 1995a, 1995b; Bieche et al., 1992; Achille et al., 1996; Takahashi et al., 1995; Kuniyasu et al., 1994). Table 1 summarizes the results of the samples that had LOH. Overall, 64% (59/92) of tumor specimens showed changes involving one or both alleles of at least one marker mapped to 7q. Allele loss of the markers tested ranged from 6.2% with D7S796 to 30% with D7S649. The greatest loss was detected with markers in 7q31.2. The percentage of tumors with LOH at each locus is shown in Table 2. All of the markers analysed were highly informative (Table 2). The number of informative cases ranged from 54 (D7S522) to 81 (D7S796). Each adjacent pair of microsatellite markers was scored for the number of times that both markers were informative with a tested sample and for the number of times that a breakpoint could be detected between the two informative markers. Breakpoints between adjacent markers were scored when one marker was informative but retained heterozygosity, whereas the adjacent informative marker showed LOH. This kind of analysis showed that most of the RCC samples that had LOH for one or more microsatellites tested actually had multiple breakpoints on 7q. We found that 42 tumors had two breakpoints, 11 tumors had four breakpoints and tumors RC10 and R37 each had ®ve breakpoints. Thus in most of these samples, the deletions of chromosome 7 DNA sequences were interstitial, including three of the ®ve tumors with LOH of the most telomeric marker D7S636. The other two tumors may have lost only the distal segment of 7q. Figure 1 shows some representative microsatellite analyses of matched normal-tumor pairs with some of the markers. Figure 2 is a schematic representation showing the frequency of allele loss in the tumor samples. This ®gure clearly illustrates that the smallest common deleted region (SCDR) showing heterozygous loss is in 7q31.1-31.2, with markers D7S522 and D7S649 showing more frequent losses than the other markers. In addition, a second smaller SCDR is seen with the marker D7S471 in 7q31.1. Most of the samples analysed were nonpapillary clear cell RCCs and 64% of these (43/67) showed LOH of one or two markers on 7q. RC11 showed a contiguous loss of almost all the markers in 7q31.1-34. Three of the markers (D7S471, D7S490 and D7S495) were not informative in this tumor. R37 and R45 showed contiguous loss of three markers involving D7S522, D7S490, and D7S649. Five nonpapillary RCC samples (RC10, 13, 27, 59 and 68) showed contiguous loss of two markers involving either D7S490 and D7S649 or D7S649 and D7S495 (Table 1). RCC samples R29 and R30 had LOH of D7S522 and D7S490. We also analysed nine papillary and six oncocytomas for LOH of 7q markers. Three of nine papillary RCCs exhibited LOH of one locus (RC1, R24 and R40) while RC62 and R30 showed LOH of two

markers. RC62 showed contiguous loss of markers D7S490 and D7S649. Also lost in this tumor was D7S796 in 7q22. Three of the six oncocytomas tested showed LOH of one or two markers (Table 2: RC61, R29 and R34). The oncocytomas that showed LOH of 7q markers did not display microsatellite instabilities (MIN) characteristic of oncocytomas (Thrash-Bingham et al., 1995a). Three oncocytomas including RC33, RC46, and R32 showed MIN at most of their tested loci. The MIN in these samples were of type II RER+ (replication error) phenotype involving contractions or Table 1

Result of microsatellite analysis of the 65 RCC samples with 10 polymorphic 7q markers. Black ®lled squares, LOH; white un®lled squares, retained; UI, uninformative; A, ampli®cation; MIN, microsatellite instability; HD, homozygous deletion; ND, not determined. Only those RCC samples that showed LOH for one or more of the tested markers are shown on this Table


expansions of two to four base pairs. The oncocytomas with MIN did not show LOH of the tested markers. In addition, two clear cell RCC samples RC63 and RC73 had MIN of one or two loci (Table 1). All 92 pairs of RCCs were examined for homozygous deletions by comparative multiplex PCR analysis. Primer sets D7S471 and D7S490, D7S486 and D7S1805, D7S796 and D7S522, D7S649 and D7S495 were multiplexed in the same reactions. In sample RC21, primer pair sets D7S471 and D7S490 ampli®ed a product from the normal DNA, whereas only D7S490 ampli®ed a product from the corresponding tumor DNA. The inability to amplify a product using the D7S471 marker within the same reaction that ampli®ed a product using the D7S490 marker leads us to believe that both copies of the D7S471 locus were deleted in the tumor sample. All the comparative multiplex PCRs for the tumors with homozygous deletions were reproduced at least twice. In addition, we con®rmed the presence of homozygous deletions by repeating the same reactions without radiolabeling. The products of the multiplex PCR reactions were run on a 2.2% agarose gel and photographed after ethidium bromide staining. As a control, we also ampli®ed samples that we knew did not have a homozygous deletion for that locus. Degradation of the tumor DNA in RC21 was ruled out as a possibility in its inability to amplify D7S471 because other markers larger than D7S471 were ampli®ed in this tumor (for example D7S522 and D7S649 which amplify 220 bp and 280 bp products, respectively). Figure 3a shows the autoradio-

Table 2 Chromosome 7q LOH in 92 renal cell carcinomas

Markers D7S796 D7S471 D7S523 D7S486 D7S522 D7S490 D7S649 D7S495 D7S1805 D7S636

Location

# of informative cases (% of total)

7q22 7q31.1 7q31.1 7q31.1 7q31.1 7q31.1 7q31.1 ± 31.2 7q31.3 ± 32 7q34 7q35

81 79 75 61 54 80 64 62 66 79

# of cases with LOH

(88) (86) (82) (66) (61) (87) (70) (67) (73) (86)

% LOH

5 10 7 4 13 17 19 10 8 5

graphic exposure of RC21 ampli®ed with primer pair sets that revealed the homozygously deleted marker run on an agarose gel. Figure 3b shows the same reactions on an agarose gel. Trisomy or tetrasomy of chromosome 7 in RCCs has been reported (Miles et al., 1988). We, therefore, looked for allelic imbalance indicative of increased copy number of one allele by comparative multiplex PCR without radiolabeling. An example of this analysis in RC22 is seen in Figure 4. In sample pairs RC22 and RC49, primer pair sets D7S486 and D7S1805 ampli®ed the normal DNA to the same extent. But in the corresponding tumor DNA, the product of ampli®cation of D7S1805 was twice as much compared to the product of ampli®cation by the marker D7S486 in the same reaction. However in RC48, the level of ampli®cation of the two tested microsatellites was identical. This is re¯ective of an increase in copy number, indicating the presence of extra copies of the locus. We believe this may be a result of mitotic recombination rather than the presence of extra copies of the chromosome as other loci in the same region did not show ampli®cation. Similar analysis with other primer pairs in some of the samples are indicated as `A' for ampli®cation in Table 1. There was no correlation between tumor grade (not listed) or cell type and 7q LOH, MIN, or ampli®cation in renal cell carcinomas. Discussion Renal cell carcinoma (RCC) is the most common tumor of the kidney in both men and women. Cytogenetic studies with both sporadic and hereditary cases of RCC have focused attention on the short arm of human chromosome 3. Structural rearrangements including deletions and translocations have been observed in which breakpoints have been assigned to

6.2 12.7 9.3 6.6 24 21.2 29.7 16.1 12.6 6.3

Cumulative results of 7q LOH in 92 sporadic RCC samples

D7S471 NT

R37

D7S486 D7S796 D7S649 D7S486 NT NT NT NT

R1

RC59

RC68

R8

D7S522 NT

R29

Figure 1 Representative examples of RCC samples showing LOH. Autoradiographs represent normal (N) and matched tumor (T) DNA samples from patients whose case numbers are below each panel. Arrows point to the loss of an allele (LOH) in tumors compared to the corresponding normal DNA. The speci®c markers showing LOH are indicated above the respective panels

Figure 2 Graphic representation of 7q21.3-qter and approximate positions of the microsatellite markers. Histogram shows the percentage LOH for each of these markers in the informative RCC samples. Y-axis, markers used; X-axis, percentage of tumors showing LOH

2729


2730

a

a NT NT

D7S1805 NT

D7S796 NT

D7S1805 NT

RC49

RC20

RC22

D7S471

b N

T

N

T

N

T

D7S1805 D7S486

RC49

RC21RC22 b T

N

T

200 bp

D7S471

100 bp

D7S490 RC21

RC48

Figure 4 (a) Representative examples of RCC samples showing ampli®cation. Autoradiographs represent normal (N) and matched tumor (T) DNA samples from patients whose case numbers are below each panel. Arrows point to the ampli®ed band in tumors compared to the corresponding normal DNA. The speci®c markers showing ampli®cation are indicated above the respective panels. (b) Products of multiplex PCR on agarose gel. Lanes 1 and 2: Normal and tumor DNA from RC22 ampli®ed with primer pair sets D7S1805 (top band) and D7S486 (Note the increase in intensity (ampli®cation) of D7S1805 product in tumor DNA, lane 2). Lanes 3 and 4: Normal and tumor DNA from RC49 ampli®ed with primer pair sets D7S1805 and D7S486 (Note the increase in intensity (ampli®cation) of D7S1805 product in the tumor DNA in lane 4).

D7S490

H2O Lym N

RC22

RC22

Figure 3 RC21 showing homozygous deletion. (a) Autoradiography of products of multiplex PCR. Primer pair sets D7S471 and D7S490 ampli®ed a product from the normal DNA whereas only D7S490 ampli®ed a product from the corresponding tumor DNA. (b) Agarose gel electrophoresis of products of multiplex PCR. Lane 1: 100 bp ladder; Lane 2: Lymphoblastoid DNA; Lanes 3 and 4: Normal and tumor DNA from RC21 ampli®ed with primer pair sets D7S471 (top band) and D7S490 (Note absence of D7S471 product in tumor DNA, lane 3); Lanes 5 and 6: Normal and tumor DNA from RC22 ampli®ed with primer pair sets D7S471 (top band) and D7S490

three distinct regions of 3p (Shridhar et al., 1997; Cohen et al., 1979; Wang et al., 1984). It is widely accepted that alterations on this chromosome are probably the initiating events in the development of RCC. However, several other studies have demonstrated additional deletions in RCC which included portions of other chromosomal regions including 5q, 6p, 6q, 11p, 13q, 14q and 17p (Morita et al., 1991; Trash-Bingham et al., 1995b; Presti et al., 1993; Kovacs et al., 1989). Recent studies in several different types of cancers have shown that multiple genetic alterations are required for the transformation of a normal cell into a malignant cell. A total accumulation of events including activation of oncogenes and/or inactivation of tumor suppressor genes leads to colon carcinoma (Fearon et al., 1990a) or breast cancer

(Callahan et al., 1989). This is probably also true for renal cell carcinoma. Our study clearly indicates that allele loss of markers in 7q31.1-31.2 occurs in more than 60% of renal cell carcinomas. Morita et al. (1991) reported LOH of marker D7S396 (7q35-36) in 26% (6/23) of clear cell RCCs. Thrash-Bingham et al. (1995b) saw no allele loss with D7S550 (7q36) in 23 cases (18 nonpapillary and ®ve papillary RCCs). In contrast to the above two reports where only one marker was analysed, we have used a much higher density of markers on 7q in the present study. This enabled us to score for LOH more precisely. The deletion map generated from this study resembles similar SCDR reported by other groups for cancers in breast (Zenklusen et al., 1994a), colon (Zenklusen et al., 1995b), ovarian (Zenklusen et al., 1995a), prostate (Zenklusen et al., 1994b; Takahashi et al., 1995), pancreas (Achille et al., 1996), squamous cell carcinoma of the head and neck (Zenklusen et al., 1995b) and gastric carcinomas (Kuniyasu et al., 1994) for this region of the chromosome. It is of considerable interest that the region showing the greatest frequency of loss in the RCC samples coincides with the location of an aphidicolin-inducible fragile site in band 7q31.1-31.2 (FRA7G) (Berger et al., 1984; Barbi et al., 1984). Detailed analysis of the most common fragile site in the human genome, FRA3B in 3p14.2, has revealed frequent breakpoints in this region in cancers, even those with no visible cytogenetic alterations. The FRA7G fragile site, however, has not been characterized in as great detail as FRA3B (Paradee et al., 1995; Wang et al., 1993; Smeets et


al., 1986; Kovacs et al., 1988). We have recently shown a clustering of breakpoints in the FRA3B region in pancreatic adenocarcinomas (Shridhar et al., 1996) similar to the observations reported here. We also observed a dramatic clustering of breakpoints in 3p14.2 with frequent multiple interstitial deletions in RCCs (Shridhar et al., 1997). It will be interesting to see if a similar phenomena is also true for FRA7G in various types of cancers. This awaits a detailed molecular characterization of FRA7G. The same frequency of 7q allele loss is seen in both nonpapillary (64%) and papillary (66%) RCCs tested. Of the six tumors analysed by Morita et al. (1991) with D7S396, ®ve were clear cell RCCs and one was of the granular type. No papillary tumors were analysed. Based on the very small number of papillary tumors analysed in this study, it is dicult to draw any conclusions regarding LOH on 7q and this subtype of RCC tumors. The MIN in oncocytomas observed here is similar to the type II RER+ phenotype. The oncocytoma tumors that showed MIN did not have LOH of the 7q markers. Both ®ndings are similar to what was reported in RCCs by Thrash-Bingham et al. (1995a). Homozygous deletions are rare genetic events. These events have played a critical role in the identi®cation of tumor suppressor genes such as RB (Friend et al., 1986), WT1 (Call et al., 1990) and DCC (Fearon et al., 1990b). Recently, several chromosomal regions have been shown to be homozygously deleted in dierent types of cancers (Diaz et al., 1988; Brown et al., 1993; Wieland et al., 1992; Rabbitts et al., 1990). So far, there have been no reports of homozygous deletions involving 7q. In addition to the in vitro studies of Zenklusen (1994c) and Ogata (1993) which showed suppression of tumorigenicity in cell lines (with 7q LOH) transfected with whole chromosome 7, the identi®cation of homozygous deletions in RC21 lends additional evidence for the existence of a tumor suppressor gene in region 7q31.1-31.2. We are currently attempting to characterize this region of homozygous deletion in RC21 by Southern hybridizations and FISH with probes generated from bacterial arti®cial chromosomes (BACs). It has been suggested that the 7q31.1 region may be the site of a putative tumor suppressor gene involved in tumor progression and metastasis of several other cancers (Collard et al., 1987). This model may very well apply to renal cell carcinomas also. Allelotyping with microsatellites cannot formally distinguish whether the change in the ratio between alleles is due to loss of one or gain of the other. Therefore it is possible to overestimate the frequency of allele loss. Comparative multiplex PCR aided in limiting this misinterpretation. For example in RC22 and RC49 where allelic gain was suspected, this was con®rmed by repeating the reactions without radio labeling and visualizing the products of the same multiplex reactions by agarose gel electrophoresis. Skewing of microsatellite markers in favor of the smaller allele was controlled for as we always compared allele density relative to that for normal DNA from the same person. Thus, the use of comparative multiplex PCR analysis not only enabled us to delineate regions of homozygous deletions but

also allelic imbalance due to increased copy number of one allele which is often seen on chromosome 7q in RCCs (11). Five of 92 RCCs showed ampli®cation of the markers tested as compared to 59 of 92 cases with LOH of at least one marker on 7q; thus, LOH of 7q microsatellites was more common than ampli®cation in RCCs. Chromosomal fragile sites may predispose chromosomes to breakage and recombination, and thus could be very important in the loss of DNA sequences containing tumor suppressor genes. Recently, Coquelle et al. (1997) demonstrated that the expression of fragile sites triggers intrachromosomal mammalian gene ampli®cation and sets boundaries to the early amplicons. Hence, the presence of FRA7G at 7q31.2 may play an important role in both chromosomal loss and ampli®cation in this region. We have recently isolated and mapped a YAC (positive for markers D7S486 and D7S522) that crosses this fragile site by FISH analysis in an aphidicolin-induced chromosome 7-only somatic cell hybrid (unpublished observation). The observation of a homozygous deletion in RC21 with the marker D7S471 further support that this chromosomal region may harbor an important tumor suppressor gene involved in the generation of RCC and possibly of other solid tumors.

Materials and methods Tumor samples Renal cell samples were collected between 1989 and 1995. Thirty of the samples were collected from Harper Hospital at Detroit Medical Center and 62 from Emory University School of Medicine; the samples from Harper Hospital are labeled R and the samples from Emory University School of Medicine are labeled RC. Tumor samples were obtained from 68 patients with clear cell carcinoma, nine with papillary renal cell carcinomas, four with mixed/granular cell type, and four with no such information available. Also included in this study were six oncocytomas and one squamous cell carcinoma. One of the Emory samples was classi®ed as chromophobe RCC; this was a clear cell RCC, chromophobe variety, grade 1. Due to the fact that the RCC tumor samples were obtained from many dierent hospitals, there was no one classi®cation of RCC that was universally applied. The prevailing community standard at the time that most of these tissues were obtained was that of Thoenes et al. (1986). DNA extraction High molecular weight DNA was isolated from tumor and normal tissue by standard phenol chloroform methods. Tumor DNA was obtained from snap-frozen tissue whereas paired noncancerous DNA was obtained from normal kidney (histopathologically con®rmed) obtained from each patient. Microsatellite analysis We used 10 pairs of markers on chromosome 7q which were all obtained from Research Genetics (Huntsville, AL). The markers used in this study were D7S796, D7S471, D7S523, D7S486, D7S522, D7S490, D7S649, D7S495, D7S1805 and D7S636. The PCR reaction mixed contained: 50 ng of genomic DNA, 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2 200 mM concentration of each primer, and

2731


2732

0.5 units of Taq polymerase (Promega) in a 12.5 ml reaction volume. One primer from each microsatellite was endlabeled using T4 kinase and [32P]-ATP as previously described (Shridhar et al., 1994). The conditions for ampli®cation were: 948C for 2 min, then 30 cycles of 948C for 1 min, 52 ± 578C for 1 min, and 728C for 1 min in a Perkin Elmer-Cetus 4800 Gene-Amp PCR system. To detect a homozygous deletion and/or an increase in copy number, we multiplexed two primer sets in one reaction. Primer sets D7S471 and D7S490, D7S486 and D7S1805, D7S649 and D7S495 were multiplexed. These sets of primers had identical annealing temperatures and their products diered in size by about 50 ± 100 bp. If in one sample both sets of primers ampli®ed a product from the normal DNA and only one primer set ampli®ed a product from the corresponding tumor DNA, it was scored as a homozygous deletion. This reaction was repeated twice to con®rm the occurrence of such an event. The PCR products were denatured and run on 6% polyacrylamide sequencing gels containing 8 M urea. The

gels were dried and autoradiographed for 16 ± 24 h and scored for LOH. Multiple exposures were used before scoring for LOH. Allelic imbalance indicative of LOH were scored when there was a loss of intensity of one allele in the tumor sample with respect to the matched allele from normal tissue. The evaluation of the intensity of the signal between the dierent alleles was determined by visual examination by two independent viewers (VS and GK). For the tumor which showed a homozygous deletion with one of the primer pair sets, the same reaction was repeated without end-labeling, the ampli®ed product was run on a regular 2% agarose gel, and the ethidium bromide stained gel was photographed. Some of the tumors which were suspected to contain an increased copy number of a particular allele were also analysed in a similar fashion.

Acknowledgements This work was supported by NIH grant CA48031 (to DIS).

References Achille A, Biasi MO, Zamboni G, Bogina G, Magalini AR, Pederzoli P, Perucho M and Scarpa A. (1996). Cancer Res., 56, 3808 ± 3813. Anglard P, Tory K, Brauch H, Weiss GH, Latif F, Mreino MJ, Lerman MI, Zbar B and Linehan WM. (1994). Cancer Res., 51, 1073 ± 1077. Barbi G, Steinbach P and Vogel W. (1984). Hum. Genet., 68, 290 ± 294. Berger R, Bloom®eld CD and Sutherland GR. (1985). Cytogenet. Cell Genet., 40, 490 ± 535. Bieche I, Champeme MH, Matifas F, Hacene K, Callahan R and Lidereau, R. (1992). Lancet , 339, 139 ± 143. Brown AG, Ross FM, Dunne EM, Steel M and WeirThompson EM. (1993). Nature Genet., 3, 67 ± 72. Call KM, Glaster TM, Ito CY, Buckler AJ, Pelletier J, Haber DA, Rose EA, Kral A, Yager H, Lewis WH, Jones CA and Houseman DE. (1990). Cell, 60, 509 ± 520. Callahan R and Campbell G. (1989). J. Natl. Cancer. Inst., 81, 1780 ± 1786. Cohen AJ, Li FP, Berg S, Marchetto DJ, Tsai S, Jacobs SC and Brown RS. (1979). New Engl. J. Med., 301, 592 ± 595. Collard JG, van de Poll M, Scheer A, Roos E, Hopman AHM, Geurts van Kessel AHM and van Dongen JJM. (1987). Cancer Res., 47, 6666 ± 6670. Coquelle A, Pipiras E, Toledo F, Buttin G and Debatisse M. (1997). Cell, 89, 219 ± 225. Diaz MO, Zielmin S, Le Beau MM, Pitha P, Smith SD, Chilcote RR and Rowley JD. (1988). Proc. Natl. Acad. Sci. USA, 85, 5259 ± 5263. Fearon ER and Vogelstein B. (1990). Cell, 61, 759 ± 767. Fearon ER, Cho KR, Nigro JM, Kern SE, Simons JW, Ruppert JM, Hamilton SR, Preisenger AC, Thomas G, Kinzler KW and Vogelstein B. (1990). Science, 247, 49 ± 56. Friend SH, Bernards R, Rogelj S, Weinberg R, Rapport JM, Albert D and Dryja TP. (1986). Nature, 323, 643 ± 646. Kovacs G and Frish S. (1989). Cancer Res., 49, 651 ± 659. Kovacs G, Erlandsson R, Boldog F, Ingvarsson S, MullerBrechlin R, Klein G and Sumegi J. (1988). Proc. Natl. Acad. Sci. USA, 85, 1571 ± 1575. Kuniyasu H, Yasui W, Yokozaki H, Akagi M, Akama Y, Kitahara K, Fujii K and Tahara E. (1994). Int. J. Cancer, 59, 597 ± 600. Miles J, Michalski K, Kouba M and Weaver DJ. (1988). Cancer. Genet. Cytogenet., 34, 135 ± 142. Morita R, Ishikawa J, Tsutsumi M, Hikiji K, Tsukada Y, Kamidono S, Maeda S and Nakmura Y. (1991). Cancer Res., 51, 820 ± 823.

Murata Y, Tamari M, Takahashi T, Horio Y, Hibi K, Yokoyama S, Inazawa J, Yamakawa K, Ogawa A, Takahashi T and Nakamura Y. (1994). Hum. Mol. Genet., 3, 1341 ± 1344. Ogata T, Ayusawa D, Namba M, Takahashi E, Oshimura M and Oishi M. (1993). Mol. Cell. Biol., 13, 6036 ± 6043. Ogawa O, Kakehi Y, Ogawa K, Koshiba M, Sugiyama T and Yoshida O. (1991). Cancer Res., 51, 949 ± 953. Paradee W, Mullins C, He Z, Glover TW, Wilke C, Opalka B, Schutte J and Smith DI. (1995). Genomics, 27, 358 ± 361. Presti JC, Reuter VE, Cordon-Cardo C, Mazumdar M, Fair WR and Jhanwar SC. (1993). Cancer Res., 53, 5780 ± 5783. Rabbitts P, Bergh J, Douglas J, Collins F and Waters J. (1990). Genes Chrom. Cancer, 2, 231 ± 238. Shridhar R, Shridhar V, Wang X, Paradee W, Dugan M, Sarkar F, Wilke C, Glover TW, Vaitkevicius V and Smith DI. (1996). Cancer Res., 56, 4347 ± 4350. Shridhar V, Siegfried J, Hunt J, del mar Alanso M and Smith DI. (1994). Cancer Res., 54, 2084 ± 2087. Shridhar V, Wang L, Rosati R, Paradee W, Shridhar R, Mullins C, Sakr W, Grignon G, Miller OJ, Sun QC, Petros J and Smith DI. (1997). Oncogene, 14, 1269 ± 1277. Smeets DFCM, Scheres JMJC and Husinx TWJ. (1986). Human Genet., 72, 215 ± 220. Takahashi S, Shan Al, Ritland SR, Delacey KA, Bostwick DG, Lieber MM, Thibodeau SN and Jenkins RB. (1995). Cancer Res., 55, 4114 ± 4119. Thones W, Storkel S and Rumpelt HJ. (1986). Pathol. Res. Pract., 181, 125 ± 143. Thrash-Bingham CA, Greenberg RE, Howard S, Bruzel A, Bremer M, Goll A, Salazar H, Freed JJ and Tartof K. (1995b). Proc. Natl. Acad. Sci. USA, 92, 2854 ± 2858. Thrash-Bingham CA, Salazar H, Freed JJ, Greenberg RE and Tartof K. (1995a). Cancer Res., 55, 6189 ± 6195. Tsai YC, Nichols PW, Hiti AL, Williams Z, Skinner DG and Jones PA. (1990). Cancer Res., 50, 44 ± 49. Wang N and Perkins KL (1984). Cancer Genet. Cytogenet., 11, 479 ± 481. Wang N-D, Testa JR and Smith DI. (1993). Genomics, 17, 341 ± 347. Weston A, Willey JC, Modali R, Sugimura H, McDowell EM, Resau J, Light B, Haugen A, Mann DL, Trump BF and Harris CC. (1989). Proc. Natl. Acad. Sci. USA, 86, 5099 ± 5103. Wieland I, Bohm M and Bogatz S. (1992). Proc. Natl. Acad. Sci. USA, 89, 9705 ± 9709. Zbar B, Brauch H, Talmadge C and Linehan M. (1987). Nature, 327, 721 ± 724.


Zenklusen JC, Barrett JC, Oshimura M and Coti CJ. (1994c). Oncogene, 9, 2817 ± 2825. Zenklusen JC, Bieche I, Lidereau R and Conti CJ. (1994a). Proc. Natl. Acad. Sci. USA, 91, 12155 ± 12158. Zenklusen JC, Thompson JC, Klein-Szanto AJP and Conti CJ. (1995b). Cancer Res., 55, 1347 ± 1350.

Zenklusen JC, Thompson JC, Troncoso P, Kagan J and Conti CJ. (1994b). Cancer Res., 54, 6370 ± 6373. Zenklusen JC, Weitzel JN, Ball HG and Conti CJ. (1995a). Oncogene, 11, 359 ± 363.

2733