Gene amplification in the murine SEWA system

Mutation Research, 276 (1992) 285-290

285

© 1992 Elsevier Science Publishers B.V. All rights reserved 0165-1110/92/$05.00

MUTREV 0336

Gene amplification in the murine SEWA system G6ran Levan, Fredrik St~hl and Yvonne Wettergren Department of Genetics, Gothenburg University, S-400 33 Gothenburg (Sweden) (Accepted 8 January 1992)

Keywords: DNA amplification; SEWA murine cells; Double minutes (DM); Homogeneously staining regions (HSR); C-Bandless chromosomes (CM); Chromatin bodies (CB)

Summary Considerable work with DNA amplification has been carried out in the murine SEWA ascites tumor cell system. In SEWA cells there is 'spontaneous' amplification of the c-myc oncogene, and transitions between different cytogenetic expressions of gene amplification such as DM (double minutes), CM (C-bandless chromosomes) and HSR (homogeneously staining regions) of the amplified DNA have been recorded during serial in vivo transplantations. In SEWA cells it has also been shown that the c-myc-containing DM will he lost under in vitro conditions, but are rapidly recovered if the cells are reinjected into animals. Additional gene amplification has been superimposed on the c-myc amplification in SEWA cells by stepwise selection in vitro, leading to resistance to different drugs, such as methotrexate, actinomycin D, colcemid and vincristine. Cytogenetically, DNA amplification is multifaceted and, in addition to the structures mentioned, it may also take the form of CB (chromatin bodies), which have been shown to be the carriers of resistance genes in hybrids between multidrug-resistant SEWA cells and Chinese hamster CHO cells. In most instances, DM are noncentromeric and distributed by a 'hitch-hiking' mechanism at mitosis; in one colcemid-resistant SEWA line, however, we have shown that the DM carry active centromeres. The molecular mechanism behind DNA amplification appears to be complex. We have shown that in four independently derived multidrug-resistant SEWA sublines the amplicons resided on circular molecules which were about 2500 kb long and carried at least five genes, including the three mouse mdr genes. Within the circles the DNA was unrearranged compared to the organization of the DNA in sensitive cells.

The cell material in which the most work on various cytogenetic aspects of gene amplification has been done is probably the murine SEWA system. Here detailed analysis of chromosome behavior has been carried out both on 'sponta-

Correspondence: Dr. G6ran Levan, Department of Genetics, Box 33031, S-400 33 Gothenburg (Sweden). Tel. (31) 853290; Fax (31) 826286.

neous' gene amplification of an oncogene and on amplification induced by drugs, such as methotrexate, actinomycin D, colcemid and vincristine.

Spontaneous gene amplification in SEWA cells The SEWA cells were originally derived from a polyoma-induced osteosarcoma (Sj6gren et al., 1961). Material from the transplantable SEWA

286

tumor was later transformed into ascites tumor cells, which were carried in serial transplantation in mice of the A.SW strain (Nordenskj61d, 1964). Stationary suspension cultures in vitro can readily be established from samples of ascitic fluid. Even in the early photographs of the p a p e r by Hellstr6m et al. (1962) it is possible to detect double minutes (DM) in S E W A cells. However, it was not until later that a systematic study of the DM in S E W A cells was initiated. It was found that the S E W A D M were C-band-negative (Levan et al., 1976) and that they lacked centromeres (Levan and Levan, 1978). During both metaphase and anaphase the D M were unaffected by the mitotic spindle. At anaphase the DM were more or less randomly distributed to the daughter cells by a surprisingly efficient 'hitch-hiking' mechanism involving relic nucleolar material. The noncentromeric nature of DM provided the explanation for two characteristics of DM-containing cell populations - the great variation in the number of DM per cell and the high frequency of interphase cells displaying small micronuclei. It was found that S E W A cells maintained a high level of DM when kept in serial in vivo transplantation. In contrast, there was a rapid loss of DM when S E W A cells were explanted in vitro. The frequency of cells with DM dropped from about 90% to less than 5% over a 100-day period of in vitro growth (Levan et ,al., 1977). However, when the in vitro population was reinjected intraperitoneally into mice, cells with D M gradually increased in importance, and after some 20 transplantation generations over a period of about 200 days the frequency of cells with DM, as well as the average number of D M per cell, had returned to the original level (Levan et al., 1977). These findings indicated that the D M carried genes that were important for the in vivo growth of S E W A cells, and they led us to suggest that the S E W A D M might represent amplification of 'genes for malignancy' (Levan et al., 1980). In fact, it was shown a few years later that the D M contained the c - m y c oncogene, which was present in about 30-fold amplification (Schwab et al., 1985). We have also shown that the level of c - m y c amplification in S E W A cells is correlated with the degree of tumorigenicity of these cells (Martinsson et al., 1988).

Transition of DM into other chromosomal structures

Some S E W A sublines could be maintained by in vivo transplantation for several years without any apparent change in the frequency of cells with DM. However, under the same conditions the DM were replaced by other cytogenetic structures in other sublines. The unexpected finding that DM could be replaced by a new chromosome type, so-called C-bandless chromosomes (CM), was reported by Levan et al. (1978). The original CM were medium-sized chromosomes that were easily distinguishable from the ordinary mouse chromosomes because they were metacentric. They clearly contained a centromere but lacked the C-heterochromatic region near the centromere which is characteristic of mouse chromosomes. Like SEWA DM, the CM were early replicating (Levan et al., 1978) and they were shown to contain amplified c-myc by in situ hybridization (Schwab et al., 1985). Several additional cases of the genesis of CM have been encountered later, one of which is reported in Levan et al. (1981). Less unexpectedly, there were other SEWA sublines in which DM were replaced by homogeneously staining regions (HSR). There seemed to be an 'antagonism' between D M and these other forms of gene amplification, meaning that individual cells normally will exhibit either DM or CM or HSR, provided that they represent amplification of the same gene. However, in most cell populations that contained a mixture of cells with D M and CM, it was possible to find occasional cells exhibiting both structures (Levan et al., 1981). Possibly these cells were representatives of a transient stage, during which the D M were integrated into the CM. Corresponding results were obtained with populations containing a mixture of cells with D M and HSR. Normally, after some passages the DM cells would completely disappear leaving ceil populations with only CM or HSR. Methotrexate (MTX) resistance in SEWA cells

Martinsson et al. (1982) looked into the question of whether it was possible to obtain amplification of a second gene in S E W A cells, on top of the amplification of c - m y c present before. Three

287 SEWA sublines, one with CM and two with (different) HSR, were subjected to selection in three different concentrations of MTX ranging from 0.1 to 10 /zg/ml. It was found that there was a distinct difference in the primary response to MTX in the three lines, and it was possible to get growing resistant cells only in one of the sublines at the lowest concentration of MTX. This primarily resistant line contained 80% cells with DM. By stepwise selection it could be made resistant with relative ease to much stronger MTX concentrations. It was shown that the degree of resistance was closely correlated with the number of DM per cell. Furthermore, it was shown by Southern hybridizations that the Dhfr gene was amplified (Jakobsson, 1989) and by protein electrophoresis that there was overproduction of D H F R protein (Jakobsson et al., 1983) in these cells. A SEWA line carried in 10 /xg/ml of MTX for 3 months had DM in 100% of the cells. Nine months later nearly 50% of the cells were without DM due to the presence of a subpopulation with an HSR. This mixed population was transferred to a medium with 50 /xg/ml of MTX. After a brief period of sluggish growth the population began to grow normally. Cytogenetic analysis revealed that all cells contained large numbers of DM and that no cells with HSR could be detected. We concluded that HSR may be favored when conditions remain unchanged, but that cells with DM have superior flexibility and could be recruited to meet a sudden and drastic shift in the environment (Martinsson and Levan, 1982). The same mixed cell population was used as material in a cell fusion experiment aiming at the transfer to V79 Chinese hamster cells of the MTX resistance (Jakobsson et al., 1983). The hybrids were selected in MTX and were found to contain approximately the sum of the chromosomes from both parental species. Only a few hybrids were studied in detail, but all of them contained the Dhfr HSR from the SEWA parent and no DM, perhaps indicating that it is easier to transfer MTX resistance to a cell hybrid by means of an HSR than with DM. Multidrug resistance (MDR) in SEWA cells The same drug-sensitive SEWA subline that had been used for the MTX experiments was also

used in experiments with MDR. In the first experiments SEWA cells were subjected to stepwise selection in actinomycin D or in vincristine. It was possible to obtain resistant lines with DM, very much in the same fashion as with MTX (Dahll6f et al., 1984). Again there was a clear correlation between the level of DM and the degree of resistance. The experiments were extended to include M D R lines induced by stepwise selection in colcemid (St~hl et al., 1988). When SEWA lines exhibiting M D R were maintained for long periods of time in the same concentration of drug, the DM were usually replaced by HSR. In this material we obtained clear evidence for the actual integration of the DM sequences during the genesis of HSR. Three subclones were derived from a DM-containing M D R subline. After several months HSR replaced the DM in all three subclones. The HSR were in different locations in each of the three sublines (Martinsson, 1987). Molecular analysis showed that each of the HSR sublines contained an amplicon with a 'novel joint' which was characteristic of the amplicon present in the DM (St~ihl et al., 1988). Gene amplification in the SEWA M D R lines seemed superficially very similar to the MTX-resistant lines. However, there were some clear differences. For one thing, both the DM and (especially) the HSR in the M D R lines were frequently more pycnotic when stained in Gbanding. In contrast to the HSR containing amplification of c-myc or Dhfr, the HSR in the M D R lines were often C-band-positive. This was studied further by in situ hybridization. The two HSR chromosomes in an actinomycin D-resistant line contained both mdr gene-specific sequences and mouse satellite DNA (Jakobsson et al., 1987). The latter is normally restricted to the centromeric regions of mouse chromosomes, but the HSR studied in this line were located terminally in two mouse chromosomes. This finding pointed to the possibility that repetitive DNA might be involved in facilitating the amplification process. On the other hand, we had mapped the mouse mdr gene to a position quite close to the centromere in mouse chromosome 5 (Martinsson and Levan, 1987), and it could be feasible that centromeric heterochromatic DNA sequences had

288 merely b e e n coamplified in these HSR. However, Bostock and Clark (1980) have described large H S R that contained satellite D N A in M T X - r e sistant mouse cells, and the Dhfr gene is located distally in mouse c h r o m o s o m e 13. To add to the confusion, a n o t h e r H S R from a S E W A M D R line was studied with kinetochore-specific antibodies (Jakobsson et al., 1989), and it was shown that the H S R exhibited n u m e r o u s fluorescent spots of (probably inactive) kinetochore material (Fig. 1A). In this material a fourth cytogenetic structure related to gene amplification was encountered. Cells from an M D R S E W A line with two Cband-positive H S R were fused to Chinese hamster C H O cells (Jakobsson et al., 1987). In contrast to the findings with the a b o v e - m e n t i o n e d MTX-resistant hybrids, 21 out of 22 M D R hybrid clones contained no recognizable mouse c h r o m o somes. Instead, they exhibited, in addition to normal-looking hamster c h r o m o s o m e s , n u m e r o u s heteropycnotic chromatin bodies (CB) which were clearly distinct from ordinary D M (Fig. 1B). T h e

CB were shown to be derived from the S E W A H S R , since they contained the same elements: amplified mouse mdr genes and mouse satellite D N A . The CB were sometimes chromosome-like and sometimes DM-like, and they were relatively stable even u n d e r n o n s e l e c t i v e c o n d i t i o n s (Jakobsson, 1988). By means of kinetochorespecific antibodies each CB was shown to contain one active c e n t r o m e r e (Jakobsson et al., 1989).

Molecular structure of amplicons As is evident from the results described above, D N A amplification in S E W A cells is a multifaceted p h e n o m e n o n . T o the four different cytogenetic structures of amplified D N A a fifth can be added, in that we have recently come across a colcemid-resistant S E W A line which exhibits clearly centromeric DM. This is not the rule for such lines, however. Colcemid-resistant S E W A lines exist which display nonpycnotic, noncentromeric DM, as well as those displaying nonpycnotic or heteropycnotic HSR. This multitude of

--,-O

Fig. 1. (A) Fluorescent staining with antikinetochore antibodies showing the presence of kinetochore material in the HSR of an actinomycin D-resistant SEWA subline (arrow). (B) C-banding of MDR hybrid cells derived from fusion of Chinese hamster CHO and actinomycin D-resistant SEWA cells. All the chromosomes have been derived from the hamster parent except the strongly C-band-positive CB (arrows), which contain amplification of murine mdr genes and murine satellite DNA.

289

shapes that amplified D N A may take on makes it difficult to envisage the molecular mechanisms behind the amplification process. Several different hypotheses have been put forward and it is not easy to select the correct one. In fact, it seems likely that many different processes operate. We have recently performed experiments which may have a bearing on this. It has been an enigma why there is commonly coamplification of several genes during the development of MDR, in spite of the fact that it is known from transfection experiments by several authors that a single gene can confer the resistance (Gros et al., 1986; Ueda et al., 1987; Devault and Gros, 1990). In five out of six independently derived M D R SEWA sublines we found that in the amplicons a 2500-kb 5orcln NruI

class NruI

5

segment containing at least 5 different genes was completely unrearranged compared to the organization of the native DNA in drug-sensitive cells (St~ihl et al., 1992). In four of them the amplicons were shown to be situated on circular DNA molecules, each probably corresponding to one DM (Fig. 2). These findings imply that on each side of a D N A segment, encompassing about 2500 kb and containing at least 5 genes, there are relatively small DNA regions, in which each of these independently generated circles have recombined. Subsequent amplification must have occurred without further recombination. The findings are definitely in conflict with some of the proposed models for gene amplification. Probably, the mechanism by which these M D R lines

mdrla

MluI

MluI

mdrlb/mdr~ Notl

Nrul

NatI

MluI

Miu!

MhJT

Mh,T

k.-

l~luI

clmmm

MluI ~

mdrla

5 Notl

NruI

mdrlb/mdr2

8orc|r'l NotI/Nrul 100

kb

NcuI t.,

•-•

l'4[iuI

MluI

Fig. 2. Circularization of the amplicon in a S E W A colcemid-resistant subline. At the top a partial restriction m a p of genomic D N A in a drug-sensitive S E W A line is shown, with the approximate position of five genes (sorcin, class 5, m d r l a , m d r l b and mdr2) indicated. Below, a circular D N A element is shown, which is the amplicon in a colcemid-resistant derivative line. This element probably represents the units of the D M present in this line. Similar circular elements were detected in three other S E W A sublines, which had been independently derived by stepwise selection in colcemid or vincristine (St~h] eta]., 1992).

290

have arisen is not a general mechanism for gene amplification. Rather, our results emphasize that many different processes may be involved in D N A amplification in mammalian cells.

Acknowledgements This work was supported by grants from the Swedish Cancer Society, the Erik Philip-S6rensen Foundation, the IngaBritt and Arne Lundberg Research Foundation, CANCIRCO and BioV~ist. References Bostock, C.J., and E.M. Clark (1980) Satellite DNA in large marker chromosomes of MTX-resistant mouse cells, Cell, 19, 709-715. Dahll6f, B., T. Martinsson and G. Levan (1984) Resistance to actinomycin D and to vincristine induced in a SEWA mouse tumor cell line with concomitant appearance of double minutes and a low molecular weight protein, Exp. Cell Res., 152, 415-426. Devault, A., and P. Gros (1990) Two members of the mdr family confer multidrug resistance with overlapping but distinct drug specificities, Mol. Cell. Biol., 8, 1652-1663. Gros, P., Y.B. Neriah, J.M. Croop and D. Housman (1986) Isolation and expression of complementary D N A that confers multidrug resistance, Nature (London), 323, 728-731. Hellstr6m, I., K.E. Hellstr6m and H.O. Sj6gren (1962) Further studies on superinfection of polyoma-induced mouse tumors with polyoma virus in vitro, Virology, 16, 282-300. Jakobsson, A.-H. (1988) The semistability of centric chromatin bodies during long-term passage of multidrug-resistant mouse-Chinese hamster cell hybrids, Anticancer Res., 8, 307-312. Jakobsson, A.-H. (1989) Cytogenetic and molecular analysis of drug resistance in interspecific cell hybrids, PhD Thesis, Gothenburg University. Jakobsson, A.-H., B. Dahll6f, T. Martinsson and G. Levan (1983) Transfer of methotrexate resistance by somatic cell hybridization, Hereditas, 99, 293-302. Jakobsson, A.-H., U. Arnason, A. Levan, T. Martinsson, C. Hanson and G. Levan (1987) Novel cytogenetic expression of gene amplification in actinomycin D-resistant somatic cell hybrids: transfer of resistance by centric chromatin bodies, Chromosoma, 95, 408-418. Jakobsson, A.-H., Y. Wettergren and T. Haaf (1989) Chromatin bodies in multidrug resistant hybrid cells have centromeres and originate from homogeneously staining regions, Anticancer Res., 9, 267-272. Levan, A., and G. Levan (1978) Have double minutes functioning centromeres?, Hereditas, 88, 81-92. Levan, G., N. Mandahl, U. Bregula, G. Klein and A. Levan (1976) Double minute chromosomes are not centromeric regions of the host chromosomes, Hereditas, 83, 83-90.

Levan, G., N. Mandahl, B.O. Bengtsson and A. Levan (1977) Experimental elimination and recovery of double minute chromosomes in malignant cell populations, Hereditas, 86, 75 -90. Levan, A., G. Levan and N. Mandahl (1978) A new chromosome type replacing the double minutes in a mouse tumor, Cytogenet. Cell Genet., 20, 12-23. Levan, G., I. Hauser-Urfer, R. Pero, F. Mitelman and A. Levan (1980) DM, HSR and CM in the SEWA tumour: amplification of genes for malignancy?, 7th Int. Chromosome Conf., Abstracts, pp. 96-97. Levan, A., G. Levan and N. Mandahl (1981) Double minutes and C-bandless chromosomes in a mouse tumor, in: F.E. Arrighi, P.N. Rao and E. Stubblefield (Eds.), Genes, Chromosomes, and Neoplasia, Raven, New York, pp. 223-251. Martinsson, T. (1987) Gene amplification in SEWA mouse tumor cells, PhD Thesis, Gothenburg University. Martinsson, T., and G. Levan (1982) Chromosomal mechanisms of methotrexate resistance: Experiences from cell lines of the SEWA murine ascites sarcoma, Hereditas, 97, 26-27. Martinsson, T., and G. Levan (1987) Localization of the multidrug resistance-associated 170 kDa P-glycoprotein gene to mouse chromosome 5 and to homogeneously staining regions in multidrug-resistant mouse cells by in situ hybridization, Cytogenet. Cell Genet., 45, 99-101. Martinsson, T., P. Tenning, L. Lundh and G. Levan (1982) Methotrexate resistance and double minutes in a cell line from the SEWA mouse ascites tumor, Hereditas, 97, 123137. Martinsson, T., F. St~hl, 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. Nordenskj61d, B. (1964) Large scale production of polyoma virus in mouse ascites tumor cells in vivo, Virology, 24, 225-227. Schwab, M., G. Ramsay, K. Alitalo, H.E. Varmus, J. Bishop, T. Martinsson, G. Levan and A. Levan (1985) Amplification and enhanced expression of the c-myc oncogene in mouse SEWA tumor cells, Nature (London), 315, 345-347. Sj6gren, H.O., I. Hellstr6m and G. Klein (1961) Transplantation of polyoma virus-induced tumors in mice, Cancer Res., 21,329-337. St~hl, F., T. Martinsson, B. Dahll6f and G. Levan (1988) Amplification and overexpression of the P-glycoprotein genes and differential amplification of three other genes in SEWA murine multidrug-resistant cells, Hereditas, 108, 251-258. St~hl, F., Y. Wettergren and G. Levan (1992) Amplicon structure in multidrug-resistant murine cells: a nonrearranged region of genomic DNA corresponding to large circular DNA, Mol. Cell Biol., in press. Ueda, K., C. Cardarelli, M.M. Gottesman and I. Pastan (1987) Expression of a full-length cDNA for the human 'MDRI' gene confers resistance to colchicine, doxorubicine and vinblastine, Proc. Natl. Acad. Sci. (U.S.A.), 84, 3004-3008.