Mutation Research, 276 (1992) 329-337 © 1992 Elsevier Science Publishers B.V. All rights reserved 0165-1110/92/$05.00

329

MUT 0329

D N A sequences amplified in cancer cells: an interface between tumor biology and human genome analysis Yosef Shiloh 1, Orna Mor 1, Ari Manor 1, Irit Bar-Am 1 Galit Rotman 1, James Eubanks 2, Mordechai Gutman 3, Guglielmina N. Ranzani 4, Jane Houldsworth 5, Glen Evans 2 and Lydia Avivi 1 1 Department of Human Genetics, Sackler School of Medicine, TelAviv University, RamatAviv 69978 (Israel), 2 Molecular Genetics Laboratory, The Salk Institute, San Diego, CA 92138-9211 (U.S.A.), 3 Department of Surgery, Rokach Hospital, Tel Aviv (Israel), 4 Dipartimento di Genetica e Microbiologia, Universit~ di Pavia, 27100 Pavia (Italy) and 5 Laboratory of Cancer Genetics, Memorial Sloan-Ketering Cancer Center, New York, N Y 10021 (U.S.A.) (Accepted 26 November 1991)

Keywords: DNA amplification; Cancer; Oncogenes; Human genome; Molecular oncology

Summary

There is growing evidence that amplification of specific genes is associated with tumor progression. While several proto-oncogenes are known to be activated by amplification, it is clear that not all the genes involved in DNA amplification in human tumors have been discovered. Our approach to the identification of such genes is based on the 'reverse genetics' methodology. Anonymous amplified DNA fragments are cloned by virtue of their amplification in a given tumor. These sequences are mapped in the normal genome and hence define a new genetic locus. The amplified domain is isolated by long-range cloning and analyzed along three lines of investigation: new genes are sought that can explain the biological significance of the amplification; the structure of the domain is studied in normal cells and in the amplification unit in the cancer cell; attempts are made to identify molecular probes of diagnostic value within the amplified domain. This application of genome technology to cancer biology is demonstrated in our study of a new genomic domain at chromosome 10q26 which is amplified specifically in human gastric carcinomas.

DNA amplification: a common genomic alteration in cancer cells

Our knowledge of the roIe of specific genes in the initiation and development of cancer has ex-

Correspondence: Yosef Shiloh, Ph.D., Department of Human Genetics, Sackler School of Medicine, Tel Aviv University, Ramat Aviv 69978 (Israel). Tel.: 972-3-6409760; Fax: 972-3-6409900.

panded rapidly in recent years. It is now established that inactivation of tumor suppressor genes and activation of proto-oncogenes are intimately connected with the acquisition and progression of the neoplastic phenotype (Bishop, 1991; Stanbridge, 1990). The structural alterations involved in these genetic events are either point mutations or gross aberrations, such as chromosomal translocations, insertions, deletions and DNA amplification. While the initial events in this cascade are probably subtle genomic aberrations,

330

progression of the neoplastic process brings considerable genomic instability. A simple model would describe a ripple effect, where additional genomic alterations affect additional genes, some of which further advance malignancy and increase genomic instability (Fig. 1). Evidence of the association between DNA amplification and tumor progression has come from several sources in recent years. Cytogenetic analysis has repeatedly demonstrated the frequent occurrence in tumors of the two classical hallmarks of DNA amplification: homogeneously staining regions (HSRs) and double minute bodies (DMs) (Brieux et al., 1984; Bruderlein et al., 1990; Gebhart et al., 1986; Limon et al., 1989; Uehara et al., 1987). Molecular studies have shown that certain proto-oncogenes are amplified in subsets of various malignancies, most notably genes of the myc, erb and F G F families (see Schwab and Amler, 1990, for review). In vitro and clinical studies have tied proto-oncogene amplification to enhanced malignancy and advanced tumor stage (Brooks et al., 1987; Heerdt et al., 1991; Martinsson et al., 1988; Schwab and Amler, 1990). It is possible, however, that D N A amplification occurs also at earlier stages of tumor development (Schimke, 1990), but at those stages it is random, encompasses very large genomic segments and yields a low copy number of the products. But when such random amplification happens to include specific genes in certain tissues, it endows the cells with a selective advantage that

Genetic alteration:

enhances the expansion of a new cell clone which gradually becomes dominant (Fig. 1). An important outcome of the correlation between amplification of certain genes and advanced tumor stage is the clinical value of protooncogene amplification. The two best examples of the use of this phenomenon as a prognostic indicator are amplification of N-myc in neuroblastomas and of c-erbB-2 in breast and ovarian cancers (Schwab, 1990; Slamon et al., 1989). It should be noted that while only a relatively small fraction of the known proto-oncogenes are amplified in human tumors, the frequency of HSRs a n d / o r DMs in various human tumors is much greater than the amplification of recognized genes. In several instances these cytogenetic manifestations of DNA amplification were found in tumor samples that did not show amplification of known proto-oncogenes (Bruderlein et al., 1990). The unavoidable conclusion from these observations is that yet unrecognized DNA sequences are amplified in various human tumors. Their identification represents an interesting and important challenge. The search for novel DNA sequences amplified in h u m a n tumors

Identification of new amplified sequences is expected to result in isolation of new genes involved in tumor progression and new probes for use at the patient's bedside. Investigation of the

Inactivation of tumor-suppressor genes; proto-oncogene acdvatJon; genomic rearrangements; random, low-copy number DNA amplification

Inactivation and~r loss of a tumor-suppressor gene

ExtensiveONA

rearrangements; specific, high-copy number DNA amplification

(--\ I I I

j, /

Phenotypic

change:

Tumor initiation

/

/ j ..~

f

J

,~ [FurZher development of thel ~ " neoplastic phenotype -~ ~

--.~[

]Tumor progresslonJ

- 1

J

Fig. 1. A model for the association of genomic alterations with the initiation and advancement of the neoplastic cellular phenotype. Genomic instability plays a more major role at the later stages of tumor progression. At earlier stages large r a n d o m segments are amplified at low copy number. Occasionally, the amplification of a specific genomic domain would endow the cell with a certain selective advantage. Cell clones with high copy-number of that region will eventually dominate the tumor.

331

structure of the amplified sequences in normal and cancer cells should contribute to our understanding of the mechanism of this intricate process. This aspect is particularly interesting for researchers of genome structure. DNA amplification is a prime example of the potential plasticity of the cellular genome: it involves the overreplication and reorganization of megabase stretches of DNA, with extensive formation of extrachromosomal intermediates which seem to play a central role in the process (Amler and Schwab, 1989; Dolf et al., 1991; Hunt et al., 1990; VanDevanter et al., 1990; Windle et al., 1991). The functional, structural and clinical aspects of DNA amplification are well demonstrated in the study of this phenomenon in the childhood solid tumor neuroblastoma (Schwab, 1990). The "neuroblastoma model" (Shiloh et al., 1987b) has therefore served as our guideline in the search for new DNA sequences amplified in human solid tumors.

Fig. 2 depicts our strategy for isolating new amplified sequences from human tumors. Cytogenetic screening of tumor cell lines is accompanied by molecular screening using the in-gel renaturation method of Roninson (1983). This technique enables the detection of any DNA sequences amplified at least 10-fold, without using molecular probes. The copy number of known genes in the tumor DNA is tested by an extensive battery of oncogene probes. Samples showing amplification of DNA sequences that contain no previously recognized genes become candidates for cloning experiments. Isolation of random fragments from the putative amplified domain is then carried out using the PERT (phenol-enhanced reassociation technique) method. This experiment is based on competitive reassociation of restriction enzyme-digested tumor DNA with an excess of sheared normal DNA, followed by shotgun cloning of the reassociation products. The

Expected result:

Experiment: Cytogeneticanalysis Screeiing~

In-gelrenaturation

Identification of anonymous amplified sequences

Hybridization with oncogeneprobes

PERTcloning

Randomfragmentsfrom the amplifieddomain

Mapping

A genetic locus

Long-rangecloning

A genomiccontig

Functional analysis

New genes

Structural analysis

Ampliconstructure

Clinical analysis

Molecular probes for the clinician

Fig. 2. An experimental approach to the identification of novel genomic loci amplified in human tumors, based on the 'reverse genetic' methodology.

332 result is a library enriched for sequences amplified in the tumor D N A (Shiloh et al., 1987a). Greater portions of the amplified domain are isolated by further cloning of sequences flanking the ' P E R T fragments', which are then studied according to the various aspects mentioned above.

Identification of a genomic domain specifically amplified in human gastric cancers In the course of our search for tumor cell lines demonstrating amplification of anonymous D N A sequences, the gastric carcinoma cell line K A T O III (Sekiguchi et al., 1978) was the first to fulfill the requirements of a good candidate. Fig. 3 shows metaphases of K A T O III and another gastric carcinoma cell line, SNU16 (Park et al., 1990), which was made available to us at a later stage. The prominent H S R in chromosome 11 of K A T O

III and numerous DMs of SNU16 point to extensive D N A amplification. In-gel renaturation confirmed the presence of amplified D N A in K A T O III, while all the proto-oncogenes we tested were present at single copy in this cell line (Mor et al., 1991). P E R T cloning resulted in 21 fragments highly amplified in K A T O III, and some 200 kb of D N A flanking 6 of them were isolated from large-insert genomic libraries. Southern blotscreening of 71 cell lines derived from a variety of tumors showed amplification of these sequences in one other gastric carcinoma, SNU16 (Fig. 4). This result indicated that the amplification was not a fortuitous in vitro event peculiar to K A T O III, but a genomic alteration characterizing certain gastric carcinomas. Nakatani et al. (1990) also detected the amplified sequences in K A T O III using in-gel renaturation. These investigators used this technique to isolate random fragments

O

g

i

•

i

J

Fig. 3. Metaphase spreads of the gastric carcinomas KATO III (left) and SNUI6 (right), showing a large HSR, and numerous DMs, respectively. The arrow points to the HSR.

333 from the amplified region, independently of our work. An initial step in the analysis of new amplified sequences is their localization in the genome of the cancer cell. We expected the sites of amplification to be the H S R of K A T O III and the DMs of SNU16. Fluorescent in situ hybridization of several amplified fragments to metaphase chromosomes of both cell lines confirmed this expectation (Fig. 5). This result demonstrates the amplification of the same D N A sequences in two different chromosomal structures: the permanent HSR and the extrachromosomal unstable DMs. For the geneticist, identification of new amplified sequences marks the embarkation on an unexplored domain of the human genome. The location of the new domain in the normal genome becomes a key question. Interestingly, the sequences amplified in the two gastric carcinomas were localized by somatic cell hybrid analysis and in situ hybridization to chromosome 10, region q26 (Mor et al., 1991), which indeed had not been extensively studied before and was devoid of

K.

123

9,4--

proto-oncogenes. One gene previously mapped to this chromosomal band, ornithine aminotransferase, was not amplified in the two cell lines. Assuming that the driving force behind extensive amplification of a specific genomic segment is associated with the function of a gene or genes contained in that region, we set out to search for expressed sequences within the first 200 kb isolated from the 10q26 domain. Using Northern blotting analysis we identified a unique fragment highly expressed in KATO III and SNU16. Sequence analysis revealed an exon bounded by intronic sequences. A Genebank computer search identified this exon as part of the K-sam gene identified independently by Hattori et al. (1991) among the amplified sequences previously isolated from K A T O III by Nakatani et al. (1990). This gene codes for a heparin-binding growth factor receptor. It is of interest that two groups identified the same gene among randomly isolated fragments of the amplified domain. The nature of the protein product of the gene, a tyrosine kinase receptor resembling several

123

123

16.0 --

6.6 5.5 4',0 m

2.5--

! i ¸¸ i ! i 0.5

A

m

B

C

Fig. 4. Southern blots of normal DNA (1), KATO III (2) and SNU16 (3), probed by 2 different PERT clones (A and C) and a phage clone containing 15 kb of DNA flanking a third PERT fragment. (Reproduced with permission of Oxford University Press from Mor et al., 1991.)

Fig, 5. Localization

of amplified

DNA sequences

using fluorescent KATO

in situ hybridization. The probe, a fragment III (left) and to the DMs of SNU16 (right).

derived

from the lOq26 region,

hybridizes

to the HSR of

335 1

2

3 4

5

6

7

8

g

A genomic walk

10 11 1 2 13 14

In a sense, the study of D N A sequences amplified in cancer cells (Fig. 2) bears several features of the 'reverse genetics' approach developed for isolating disease genes that code for unrecognized proteins (Orkin, 1986). In both strategies a genetic locus of some biological or clinical significance is defined and mapped, and novel genes of, yet unknown functions are sought. This requires the cloning and physical mapping of a fairly large genomic domain. Powerful methods were devised for this purpose, and have been refined since the initiation of the Human Genome Project. Longrange cloning using yeast artificial chromosome (YAC) vectors now enables the isolation of megabase contigs from any defined region (Burke, 1990). In adapting the genomic approach to the problem, we initiated long-range cloning experiments aimed at isolating the entire amplified domain on a YAC contig. This should facilitate the search for additional genes included in the amplified domain, as well as its structural analysis and the derivation of probes for the clinician. We used the K-sam gene as a starting point in this chromosome walk. Fig. 7 shows the restriction maps of two genomic segments corresponding to this gene, which were isolated from the Washington University genomic YAC library (Burke et al., 1987). Clone C18F7 of 240 kb turned out to represent a non-contiguous fragment, a common artifact in YAC libraries. Clone B169E10 of 200 kb appears to be a contiguous genomic fragment and was chosen for further study. It is of interest that this YAC clone con-

Fig. 6. Amplificationof the K-sam gene in one of 14 primary gastric carcinomas.

proto-oncogenes, implicates it as a possible driver of the amplification of the 10q26 domain. This result validates our approach for the isolation from amplified D N A sequences of new genes potentially involved in the malignant process. An important component of our working scheme is the testing of the copy number of every D N A fragment isolated in the course of these experiments in fresh tumor tissues. Fig. 6 shows a 10-fold amplification of the K-sam fragment in one of 14 primary gastric carcinomas. Thus far, a total of 143 such tumors have been tested and 5 showed amplification of this gene. Nakatani et al. (1990) also detected amplification of this gene in a fraction of gastric tumors. The clinical significance of this finding is now being tested on other types of tumors. However, this preliminary finding completes the successful application of the 'neuroblastoma model' to gastric carcinomas.

Mtul

EagI Mtul C18F7

,

SatI t

SatI B169 El0 I I

I

l

Sfi I',ik.--EagI - -

MtuI

Sfi I-...~EE~gslHIrSa[I I

10 kb

Fig. 7. Two clones isolated from the Washington UniversityYAC library (Burke et al., 1987) using PCR. A sequence-taggedsite (STS) within the K-sam gene was used for the screening. Clone C18F7 is non-contiguous, while clone B169E10 represents contiguous 200 kb of genomicDNA. The K-sam STS is located in the BssHII-SalI fragment.

336

1 23

123

transferase probe and Adam Sartiel for art work. This study was supported by a joint research grant from the Deutsches Krebsforschungszentrum (DKFZ) and the Israel National Council for Research and Development.

References

URA

C E N

Fig. 8. Amplification of the right ('URA') and left ('CEN') end fragments of clone B169E10 (Fig. 7) in KATO III (lane 2) and SNU16 (lane 3).

tains a CpG island, flagged by a cluster of CpGrich restriction sites (Fig. 7). CpG islands are considered to mark the 5' regions of active genes (Bird, 1987). Fig. 8 shows hybridization of end fragments of the B169E10 YAC clone to Southern blots of KATO III and SNU16. It is evident that the entire fragment lies within the amplicons of both cell lines. These end fragments are now being used for further chromosome walking aimed at extending our grip on the amplified domain. At present these experiments are entirely in the realm of the genome project. But, as we characterize the genes in the amplified region at chromosome 10q26, and draw the clinical-pathological correlates of their amplification, we will move into the domain of cancer biology where we seek to understand how the amplification of these particular genes is associated with the development of a specific cellular phenotype. Walking on this interface between tumor biology and the exploration of the human genome holds promise for both sides of this interface.

Acknowledgements We wish to thank Dr. Adi F. Gazdar of the National Cancer Institute (U.S.A.) and Dr. J.-G. Park of Seoul National University (Korea) for the cell line SNU16, Dr. Rima Rozen of McGill University (Canada) for the ornithine amino-

Amler, L., and M. Schwab (1989) Amplified N-myc in human neuroblastoma cells is often arranged as clustered tandem repeats of differently recombined DNA, Mol. Cell. Biol., 9, 4903-4913. Bird, A.P (1987) CpG islands as gene markers in the vertebrate nucleus, Trends Genet., 3, 342-347. Bishop, J.M. (1991) Molecular themes in oncogenesis, Cell, 64, 235-248. Brieux de Salum, S., I. Slavutsky, S. Besuchio and A. Pavlovsky (1984) Homogeneously staining regions (HSR) in a human malignant melanoma, Cancer Genet. Cytogenet., 11, 5360. Brooks, B., J. Baney, M.M. Nau, A. Gazdar and J. Minna (1987) Amplification and expression of the myc gene in small cell lung cancer, Adv. Viral Oncol., 7, 155-172. Bruderlein, S., K. Van der Bosch, P. Schlag and M. Schwab (1990) Cytogenetics and DNA amplification in colorectal cancers, Genes Chromosomes Cancer, 2, 63-70. Burke, D.T. (1990) YAC cloning: options, Genet. Anal., 7, 94-99. Burke, D.T., G.F. Carle and M.V. Olson (1987) Cloning of large segments of exogenous DNA into yeast by means of artificial chromosome vectors, Science, 236, 806-812. Dolf, G., R.E. Meyn, D. Curley, N. Prather, M.D. Story, B.M. Boman, M.J. Siciliano and R.R. Hewitt (1991) Extrachromosomal amplification of the epidermal growth factor receptor gene in a human colon carcinoma cell line, Genes Chromosomes Cancer, 3, 48-54. Gebhart, E., S. Bruderlein, M. Augustus, E. Sieber, J. Feldner and W. Schmidt (1986) Cytogenetic studies on human breast carcinomas, Breast Cancer Res. Treat., 8, 125-138. Hattori, Y., H. Odagiri, H. Nakatani, K. Miyagawa, K. Naito, H. Sakamoto, O. Katoh, T. Yoshida, T. Sugimura and M. Terada (1991) K-sam, an amplified gene in stomach cancer, is a member of the heparin-binding growth factor receptor genes, Proc. Natl. Acad. Sci. (U.S.A.), 87, 59835987. Heerdt, B., S. Molinas, D. Deitch and L.H. Augenlicht (1991) Aggressive subtypes of human colorectal tumors frequently exhibit amplification of the c-myc gene, Oncogene, 6, 125-129. Hunt, J.D., M. Valentine and A. Tereba (1990) Excision of N-myc from chromosome 2 in human neuroblastoma cells containing amplified N-myc sequences, Mol. Cell. Biol., 10, 823-829. Limon, J., R. Lundgren, P. Elfving, S. Heim, U. Kristoffersson, N. Mandahl and F. Mitelman (1989) Double minutes in two primary adenocarcinomas of the prostate, Cancer Genet. Cytogenet., 39, 191-194.

337 Martinsson, T., F. Stahl, P. Pollwein, A. Wenzel, A. Levan, M. Schwab and G. Levan (1988) Tumorigenicity of SEWA murine cells correlates with degree of c-myc amplification, Oncogene, 3, 437-441. Mor, O., Y. Messinger, G. Rotman, I. Bar-Am, Y. Ravia, R.L. Eddy, T.B. Shows, J.-G. Park, A.F. Gazdar and Y. Shiloh (1991) Novel DNA sequences at 10q26 are amplified in human gastric carcinoma cell lines: molecular cloning by competitive DNA reassociation, Nucleic Acids Res., 19, 117-123. Nakatani, H., H. Sakamoto, T. Yoshida, J. Yokota, E. Tahara, T. Sugimura and M. Terada (1990) Isolation of an amplified DNA sequence in stomach cancer, Jpn. J. Cancer Res., 81, 707-710. Orkin, S.H. (1986) Reverse genetics and human disease, Cell, 47, 845-850. Park, J.-G., H. Frucht, R.V. LaRocca, D.P. Bliss, Y. Kurita, T.-R. Chen, J.G. Henslee, J.B. Trepel, R.T. Jensen, B.E. Johnson, Y.-J. Bang, J.-P. Kim and A.F. Gazdar (1990) Characteristics of cell lines established from human gastric carcinoma, Cancer Res., 50, 2773-2780. Roninson, I.B. (1983) Detection and mapping of homologous, repeated and amplified DNA sequences by DNA renaturation in agarose gels, Nucleic Acids Res., 11, 5413-5431. Schimke, R.T. (1990) The search for early genetic events in tumorigenesis: an amplification paradigm, Cancer Cells, 2, 149-151. Schwab, M. (1990) Amplification of the MYCN oncogene and deletion of putative tumor suppressor gene in human neuroblastomas, Brain Pathol., 1, 41-46. Schwab, M., and L. Amler (1990) Amplification of cellular oncogenes: a predictor of clinical outcome in human cancer, Genes Chromosomes Cancer, 1, 81-93. Sekiguchi, M., K. Sakakibura and G. Fuji (1978) Establishment of cultured cell lines derived from a human gastric carcinoma, Jpn. J. Exp. Med., 48, 61-68. Shiloh, Y., L.M. Kunkel, E. Rose, B. Korf and S.A. Latt

(1987a) Rapid cloning of multiple amplified DNA sequences from human neuroblastoma cell lines by competitive DNA reassociation, Gene, 51, 53-59. Shiloh, Y., E. Rose, B. Korf, J. Shipley, K. Sakai, G. Brodeur, R.C. Seeger and S.A~. Latt (1987b) Analysis of DNA amplification in tumor cells: the neuroblastoma model, in: H. zur Hausen and J. Schlehofer (Eds.), Accomplishments in Oncology: The Role of DNA Amplification in Carcinogenesis, Lippincott, Philadelphia, PA, pp. 186-197. Slamon, D.J., W. Godolphin, L.A. Jones, J.A. Holt, S.G. Wong, D.E. Keith, W.J. Levin, S.G. Stuart, J. Udove, A. Ullrich and M.F. Press (1989) Studies of the H E R - 2 / n e u proto-oncogene in human breast and ovarian cancer, Science, 244, 707-712. Smith, K.A., P.A. Gorman, M.B. Stark, R.P. Groves and G.R. Stark (1990) Distinctive chromosomal structures are formed very early in the amplification of CAD genes in Syrian hamster cells, Cell, 63, 1219-1227. Stanbridge, E.J. (1990) Human tumor suppressor genes, Annu. Rev. Genet., 24, 615-657. Stark, G.R., M. Debatisse, E. Giulotto and G.M. Wahl (1989) Recent progress in understanding mechanisms of mammalian DNA amplification, Cell, 57, 901-908. Uehara, M., M. Kida and M. Kamakura (1987) Rings and double minutes in a case with blastic phase of chronic myelocytic leukemia, Cancer Genet. Cytogenet., 25, 253258. VanDevanter, D.R., V.D. Piaskowski, J.T. Casper, E.G. Douglass and D.D. Von Hoff (1990) Ability of circular extrachromosomal DNA molecules to carry amplified MYCN proto-oncogene in human neuroblastomas in vivo, J. Natl. Cancer Inst., 82, 1815-1821. Windle, B., B.W. Draper, Y. Yin, S. O'Gorman and G.M. Wahl (1991) A central role for chromosome breakage in gene amplification, deletion formation, and amplicon integration, Genes Develop., 5, 160-174.