Analysis of sister-chromatid exchanges and tumorigenicity in cell hybrids

It has been suggested that sister-chromatid exchange (SCE) might reflect a pathway by which cells can repair certain forms of chromosome damage(Latt & Juergens, 1977). SCE has also turned out to be of practical use to the investigator as an efficient indicator of chromosome damage, which potentially could be used to disclose carcinogens in our environment. A great many noxious agents capable of inducing SCE in tissue culture cells are carcinogenic (see, for example, Wolff, 1977). Apparently, the interaction of these agents, or their metabolites, with DNA is a necessary pre-requisite for both SCE and tumorigenesis.

The problem we have been concerned with is to find an in vitro correlate with tumorigenicity in a system of cell hybrids in the mouse. In the course of our studies we thought it worthwhile to explore the relationship between SCE and tumorigenicity in these cells, which were originally obtained from crosses between diploid fibroblasts or lymphocytes and cells of a malignant mouse melanoma (Jonasson, Povey & Harris, 1977; Jonasson & Harris, 1977; Straus, Jonasson & Harris, 1977). The results of these experiments are to be presented below.

These are independent clonal derivatives of a spontaneous malignant melanoma tumour in the mouse. Both clones were selected for resistance to 6-thioguanine and have been fully described by Jonasson et al. (1977).

This is an autotetraploid derivative of the melanoma (Salazar, Sidebottom & Harris, 1975).

These were isolated from 12-to 14-day mouse embryos of the appropriate genetic background.

Suspensions of lymphocytes were obtained by teasing subcutaneous and mesenteric lymph nodes from adult mice.

Samples of peripheral blood were collected from 7 different individuals and allowed to sediment. Unseparated leukocytes in the buffy coat layers were used for the SCE experiments. Lymphocytes used for cell fusion were separated by the method described by Böyum (1968).

All cells, except those in lymphocyte cultures, were maintained in plastic flasks in Eagle’s Minimum Essential Medium plus 10% foetal calf serum. They were subcultivated by trypsinization. Details of culture techniques, cell fusion, isolation of hybrids and assay for tumorigenicity were given by Jonasson et al. (1977), Jonasson & Harris (1977), and Straus et al. (1977). No further tumorigenicity tests have been done for this work.

Experiments with cells grown in monolayers were performed in 240-ml plastic flasks (Nunclon-Delta) filled with 20 ml MEM with 10% foetal calf serum and incubated in the dark at 37 °C. BrdU was added 24 h after seeding the culture at approximately (5–10) × 10³ cells per cm² and incubation continued for the appropriate period of time for each individual cell type to reach mitosis after 2 cycles of DNA replication in the presence of BrdU.

Lymphocyte cultures were grown for 48 h (mouse) or 72 h (human) in glass bottles with 5 ml of Ham’s Fio medium with 15% foetal calf serum and PHA. BrdU was present during the whole of the incubation period.

The procedures were described by Alves & Jonasson (1978), and Jonasson, Alves & Ström (1979). Sister-chromatid exchanges within the centromeric region of the chromosomes were disregarded when calculating SCE frequencies, since the majority of them might represent inversions of the chromatid rather than true SCEs (see the references above).

Fig. 1 A shows the ‘spontaneous ‘frequency of SCE plotted versus the BrdU concentration in the growth medium for 3 different cell types used in the present work: parental mouse melanoma cells, diploid mouse fibroblasts, and a hybrid cell clone obtained from a cross between diploid fibroblasts and melanoma cells. There was no indication of any plateau level of SCEs in any of the 3 experiments. A point of interest is that the hybrid (Clone 100A) showed significantly lower frequencies of SCEs per chromosome than both parental types of cells. This phenomenon turned out to be of general occurrence and will be commented further upon in the following sections.

By adding 0·5 μm 5-fluorodeoxyuridine (FdU), which is known to enhance the incorporation of BrdU into DNA by blocking endogenous synthesis of thymidylic acid, it seemed possible to modify the response to BrdU (Fig. 1 B). In the low to moderate BrdU concentration range a plateau was then observed, corresponding for each cell type approximately to the SCE level observed at 50 μ M BrdU without FdU. Furthermore, the cells seemed BrdU dependent. Too low a BrdU concentration resulted in chromosome fragmentation (Fig. 2) and there was also a diminished number of mitoses (data not given). Therefore, no SCE data could be obtained at very low BrdU concentrations in the presence of FdU (Fig. 1B). If one may assume that exogenous thymidine supply is required when FdU is present, it is conceivable that the low level plateau observed under these circumstances should correspond to a nearly full substitution of newly synthesized DNA with BrdU for thymidine. Needless to say, since we have not made any measurements to confirm this interpretation, other explanations are also possible. At higher BrdU concentrations, beyond 50 μ M, there was an increase of SCE frequencies even in the presence of FdU. which thus would give a strong indication of a biphasic character of the concentration–dependence curve, i.e. it is likely to be the result of at least 2 different superimposed phenomena. From such considerations it seemed justified for practical purposes to conclude that very little or nothing was to be gained from a thorough investigation of the BrdU dose-dependence relationship of each individual clone of cells to be tested for SCE and tumorigenicity. Throughout the experiments to be reported we have therefore used as a standard dose 50 μ M BrdU without the addition of FdU.

Sister-chromatid exchange frequencies in parental cells of the different crosses are shown in Table 1. There was no indication of any significant difference between mouse diploid fibroblasts and the melanoma cells. Autotetraploid melanoma cells (Clone mmc38) did not differ significantly from the nearly diploid Clone PG19 either. The slightly lower frequencies obtained with mouse and human lymphocytes are not necessarily comparable to those of the 2 other cell types since lymphocytes cannot be grown in monolayer cultures.

As shown in Table 2, SCE frequencies in the C57BL melanoma × mouse diploid fibroblasts or lymphocyte hybrid clones were consistently lower than those observed in the parental cells (Table 1). Four melanoma × fibroblast hybrid clones were examined. All of them were chromosomally fairly complete biparental hybrids which had retained more than three-quarters of the diploid fibroblast set. Clone 100A was originally obtained from a cross between Clone mel 2C and CBA T6 fibroblasts and showed a mean number of 0·21 SCE per chromosome. A subclone (Clone 100A Subclone 1) showed the same frequency. On the other hand, 2 other similar hybrid clones derived from a cross between PG 19 and T13H fibroblasts, Clones 7 and 8 (which thus represent a cross with another inbred strain of mice) showed 0·33 and 0·34 SCE per chromosome, respectively. Clone 4A, finally, from a cross between PG19 and T50H fibroblasts showed 0·26 SCE per chromosome.

These hybrids between malignant and diploid mouse cells have a severely reduced ability to generate progressive tumours when injected into immunosuppressed syngeneic animals, as compared to the malignant parental melanoma cells. They are therefore, arbitrarily, referred to here as being ‘non-malignant’; which thus should be taken in the relative sense rather than absolute.

Malignant hybrids (which had been explanted from the occasional tumours formed upon the injection of large numbers of hybrid cells into syngeneic animals) did not differ significantly from the corresponding ‘non-malignant’ hybrid, even if, in some cases, there was a tendency for slightly higher counts in the malignant cells (Tables 2,3). For example, the malignant cells Clone 100A Tumour 1 and Clone 100A Subclone 1 Tumour 2B showed a mean number of 0·23 SCE per chromosome, as compared to 0·20 SCE per chromosome for their ‘non-malignant’ counterparts (Clone 100A and Clone 100A Subclone 1).

It also appears, from the data presented in Table 2, that hybrid clones derived from crosses where the same inbred strain of mice was used as diploid parent showed more or less identical frequencies of SCE per chromosome. However, the observation might well be fortuitous.

Apparently, the lower SCE frequencies observed in the hybrids were not due to the presence of 2 diploid sets of chromosomes since autotetraploid melanoma cells (Clone mmc38) showed a high level of SCE per chromosome, indistinguishable from that of the nearly diploid melanoma (Table 1). This conclusion was further supported by the observation that PG 19 × human diploid lymphocyte hybrids containing either several sets of mouse chromosomes plus a variable number of human chromosomes (Clone 1 and 2) or the PG 19 chromosomes only (Clone 19) showed rather similar frequencies of SCE (Table 3); all of which were significantly lower than the frequency observed in the melanoma itself (Table 1).

Furthermore, as seen from Table 3 the phenomenon of a relatively low SCE frequency in the hybrids could not be related to the presence or absence of a normal allele for HGPRT. Selection against it in 0·5 μ g/ml 6-thioguanine did not alter the low frequency characteristics of either of the 2 hybrid clones (Clone 1-6TG and Clone 2-6TG) that were tested (Table 3).

It is conceivable that selection of hybrid cells in special media such as HAT-medium could induce epigenetic changes which are stable. To exclude the possibility that the mere presence of a relatively high concentration of thymidine in HAT-medium would induce a change in the SCE frequency characteristics of the melanoma cells we grew PG19 and mmc38 in H + T medium (i.e. HAT-medium without Aminopterin) for about 10 cell generations and then retested the SCE frequency. The frequencies then obtained were 0·35 and 0·39, respectively (Table 1), which was well within the range previously observed with these clones.

It is to some extent surprising that however conspicious the phenomenon of sister-chromatid exchange may be its genetic implications are not yet clear. There is in fact not one generally accepted theory of the matter, although it has recently been suggested that SCE might reflect a pathway by which cells can repair certain forms of chromosome damage (Latt & Juergens, 1977). It is possible that chromosome damage might be a necessary prerequisite for both SCE and cell transformation leading to tumourous growth. We have therefore found it of interest to investigate the relationship between SCE and tumorigenicity in a system of hybrid cells that previously has been used for the analysis of malignancy.

No method is available that allows us to study SCE without some kind of interference with growth of the living cell. The SCE detection procedures that are now commonly used employ a technique of feeding the cells 5-br0m0de0xyuridine (BrdU) to visualize the SCE events. Not unexpectedly, this substance is mutagenic and known to have an influence on the frequency of SCE. In most cases, it is clear that a majority, if not all, of the so-called spontaneous SCEs are BrdU induced. Had one used a lower concentration of BrdU in the growth medium one would also have scored a lower number of SCEs on the chromosomes. Unfortunately, extrapolation to zero BrdU concentration is not easily done. Another difficulty is that differences in uptake and metabolism of the drug between different cell types will possibly also affect the SCE frequencies. The present data indicate that this might indeed be the case. Higher concentrations of BrdU in the growth medium were needed to precipitate differential staining of sister chromatids in those cell types showing relatively low frequencies of SCE than in those showing the higher frequencies (Fig. 1B). However, staining reactions of this kind cannot advantageously be used for the quantitative evaluation of BrdU uptake and metabolism because they are too variable.

The experiments that we have just described show conclusively that SCE frequencies in hybrids between normal and malignant cells are not correlated in a simple fashion with tumorigenicity. Malignant and ‘non-malignant’ hybrid cells when derived from the same fusion event showed more or less identical SCE frequencies under the experimental conditions that we have used.

We have also shown that SCE frequencies in the C57BL melanoma × diploid fibroblast or lymphocyte hybrid clones were consistently lower than those in the corresponding parental cells under identical growth conditions, even if variation was observed between the different hybrid clones. This apparently epigenetic variation is not surprising in view of the difference that previously has been established in this laboratory between human B and T lymphocytes (Santesson, Lindahl-Kiessling & Mattsson, 1979); T cells showing more than double as much SCE as B cells from the same individual. Therefore, it is conceivable that fusion of 2 different cell types from the same individual would produce a hybrid clone with epigenetic SCE characteristics that were different from both its parents.

The interpretation of the present data is partly obstructed by the fact that ‘fibroblasts’ in culture, like lymphocytes, might not be a homogeneous cell population but rather represent heterogeneous mixtures of cells. Thus, the diploid parent of the hybrids that were isolated might in some cases be a rare cell type which was not necessarily representative of the bulk of the fibroblast culture. Instead it could possibly represent a variety of differentiated cells. This would explain why all fibroblast × melanoma hybrids did not show identical SCE frequency characteristics. On the other hand, as was already mentioned in the Results section, most of the variation that was seen between the different hybrid clones also could very well be explained on the basis of a hypothesis of genetic differences between the different strains of inbred mice from which the diploid parental cells were originally explanted; even if variation of SCE frequencies has not yet been established between different inbred strains of mice. The present material is insufficient to allow a firm conclusion on this matter.

In any case, it is apparent that modulation of the SCE frequency characteristics of the melanoma cells did occur upon cell fusion and this could be related neither to a higher number of chromosomes in the hybrid nor to the special method of selecting hybrid cells in HAT-medium. So far there is no clue to the mechanisms of regulation of SCE frequencies since we have not been able to detect whether the modulation simply is due to differences in the uptake and metabolism of the BrdU, which is used to induce differentiation between sister chromatids, or whether the intrinsic rate of SCE varies for some other reason.

This investigation was supported from the Marcus Borgström Fund and the Swedish Cancer Society (1082-B80-03X).