Chromosome segregation from cell hybrids IV. Movement and position of segregant set chromosomes in early-phase interspecific cell hybrids

Interspecific cell hybrids retain a virtually intact set of chromosomes of one parent (the ‘retained set’) while preferentially losing chromosomes of the other parent (the ‘segregant set’). The mitotic spindle has long been thought to be involved in this directional chromosome segregation (Migeon, 1968; Handmaker, 1973), and there are several observations consistent with the idea that loss occurs at mitosis (Schall & Rechsteiner, 1978; Dawson & Graves, 1984). Loss of groups of chromo-somes could result from aberrant interactions between segregant set chromosomes and the spindle fibres of the hybrid cells (Handmaker, 1971, 1973). We have recently shown that segregation is independent of hybrid spindle constitution (Zelesco & Graves, 1987). There remains the possibility that aberrant interactions between spindle and centromere result from intrinsic or induced differences in efficiencies with which retained and segregant set chromosomes engage the spindle.

Two distinct movements of chromosomes characterize normal mitotic behaviour; congression at the meta-phase plate and anaphase migration to opposite poles of the spindle. Both movements are effected by correct interaction of centromeres with spindle fibres. If segre-gant set chromosomes were to fail to attach to spindle fibres, attached less efficiently, or became prematurely disengaged, aberrant positioning or movement of segregant chromosomes should be observed at meta-phase and anaphase in hybrid cells. The exclusion of segregant chromosomes from the reconstituted daughter nuclei might also be apparent as micronuclei at telophase.

We used C-banding and G-11-banding to determine the parental origin of chromosomes in hamster–mouse and hamster–human cell hybrids. We screened populations of early-stage hybrids for aberrant metaphase congression and anaphase and telophase segregation and for non-random positioning of segregant chromo-somes.

Rodent lines were: Bio, a TK-deficient Chinese hamster line; Cmd4, an HPRT-deficient Chinese hamster line; and LTA, a TK-deficient mouse line (described in full by Hope & Graves, 1978; Zelesco & Graves, 1987). Human lines were: EUE, an HPRT-deficient HeLa derivative; and Daudi (Klein et al. 1967), a human suspension line obtained from Dr N. Hoogenraad, La Trobe University, Victoria.

Cells were routinely cultured in Dulbecco’s modified Eagle’s medium (DME, GIBCO) supplemented with 10% foetal calf serum (Flow, Australia), 60 μgml^-1 penicillin (Sigma) and 50 μgml^-1 streptomycin (Glaxo, Australia).

Disaggregated parental cells were mixed in a 1:1 ratio and fused in suspension with polyethylene glycol, M_r 1000 (BDH or Ajax, Australia) diluted to a 50% (w/v) solution in serum-free DME. Fused cells were cultured in 100mm diameter Petri dishes (Corning) in HAT medium (Szybalski et al. 1962) containing 10^-4 M-hypoxanthine, 4× 10^-7 M-amin-opterin and 1·6× 10^-5 M-thymidine (Sigma), in a humidified CO₂ incubator. Populations of hamster–mouse fused cells were harvested for cytological study, but for hamster–human fusions, because of the low frequencies of hybrids, macro-scopic colonics were isolated by suction with a micropipette, transferred to individual microscope-slide growth chambers (Lab-Tek, Miles Laboratories, Australia) and propagated until sufficient cells were present for cytological study.

Petri dishes containing 20–40 hamster–mouse hybrid colonies were sampled at random between 10 and 20 days after fusion. Cells were trypsinized and gently fixed without colchicine or hypotonic treatment, by slowly adding a few drops of fixative (3:1 (v/v), methanol:acetic acid) to the suspension in a centrifuge tube. Cells were centrifuged gently and resuspended in fresh fixative twice, then slides were prepared by spreading the suspension gently onto dry micro-scope slides and air drying. Slide chambers containing individual hamster–human hybrids were sampled 22–27 days after fusion. Cells were gently fixed in situ by removal of half the medium, and addition of fixative, which was sub-sequently replaced twice and aspirated; slides were then air dried.

Mouse and hamster chromosomes were differentiated by C-banding (Leversha et al. 1980). Hamster and human chromo-somes were differentially stained (magenta and pale blue, respectively) by a modification of the G-11 banding technique (Bobrow & Cross, 1974; A. Dobrovic, Flinders University, Australia). Slides were placed in 2×SSC (SSC is 0·15M-NaCI, 0·015 M-sodium citrate) at 58–62°C for 5 min, rinsed in double-distilled water, incubated in 4% Giemsa (Harlcco) in 10^-2M-NaOH at 37°C for 9 min, rinsed and air dried. Cells were examined using a Leitz Dialux 2 microscope and photographed using Kodak technical pan, Agfa Copex or Ilford FP4.

Hamster–mouse hybrids were recovered from Cmd4 × LTA fusions at frequencies of 2×10^-5 to 4× 10^-5. Several hamster–mouse hybrids were isolated and propagated individually in order to determine the direction and extent of chromosome segregation in this hybrid combination. Table 1 shows that some hybrids segregated about half the hamster complements and retained one set (or sometimes almost two sets) of mouse chromosomes. However, the reverse pattern was also observed in several hybrids. Thus the mass populations of early hybrids would have contained mixtures of hybrids segregating hamster, and hybrids segregating mouse, chromosomes.

Hybrids were recovered from Bio × EUE and Bio × Daudi crosses at much lower frequencies (approx. 10^-6). All retained hamster chromosomes (1, or sometimes 2 sets) and segregated human chromosomes. Established Bio–EUE hybrids retained two to eight human chromosomes, but early-phase hybrids appeared to have more. An exact count of human chromosomes in G-11-banded early hybrids was difficult because, in the absence of colchicine and hypotonic treatment, chromosomes often overlapped. However, a careful analysis of 30 and 21 (respectively) favourable metaphase cells showed that BIo–EUE early hybrids retained an average of 17, and Bio–Daudi early hybrids retained 12 (Table 1).

If chromosomes are lost from hybrid cells as the result of delayed congression or anaphase separation, we would expect to detect, at high frequencies, segregant chromosomes preferentially displaced from the meta-phase plate, lagging at anaphase, and forming micro-nuclei at telophase. If loss results from formation of tri-or tetrapolar spindles, multipolar mitoses should be frequent. We studied mitotic cells from hamster-mouse and hamster-human hybrids that had been fixed in situ or in suspension without colchicine or hypotonic pretreatment, in order to preserve the spatial relationships of chromosomes.

Since a newly formed hybrid cell might be expected to possess four rather than two mitotic centres, it is possible that tripolar and tetra-polar mitoses are frequent, and result in aberrant segregation and chromosome loss. We observed the frequencies of multipolar mitoses in the early-phase hamster–mouse and hamster-human hybrids. Table 2 shows that the frequency of multipolar mitoses in hamster–mouse hybrids was extremely low, and of the order of the frequency in parental cells. The frequency among early hamster-human hybrids was higher, but hardly more than the frequency among the EUE parents.

Metaphase spreads were selected for analysis only if they were unbroken, were oriented to give either a polar or lateral view, and were differentially stained. These rigorous scoring criteria were felt to be essential for a valid analysis, although they left rather small sample sizes for hamster-human early hybrids; only two of 12 samples (BIo–EUE fixed at 26 days, and BIo–Daudi at 24 days) contained cells deemed to be scorable. Chromosomes were scored as delayed in congression if they either lay within or were outside the ring in radial spreads, and had no attachment with the equatorial plate in a lateral view. It was reasoned that, when a three-dimensional cell is viewed in two dimensions, lagging chromosomes within the space occupied by the spindle will usually appear within the ring in a polar view, and detached from the band of chromosomes in an equatorial view. Lagging chromosomes outside this space will also be detached from the band in an equatorial view, but will appear outside the ring in a polar view. In hamster–mouse hybrids, chromosomes could be identified unequivocally as to parental origin by their strongly C-banding centromeric regions (Fig. 1A). In G-11-banded hamster–human hybrids, hamster chromosomes stained magenta and human chromosomes pale blue (Fig. 1B).

Table 3 shows the proportion of metaphase parental and hybrid cells with one or more lagging chromo-somes. A high proportion of hamster–mouse hybrid cells possessed laggards; however, this proportion was no greater than the proportion of parental cells with laggards (Table 3); also, both mouse and hamster chromosomes were represented among laggards, frequently within the same cell (Fig. 1A). Among the scorable hamster–human hybrid cells, too, it was clear that hamster chromosomes lagged as frequently as human (Table 3, Fig. 1B).

There were few anaphase figures in either hybrid type; all were scored. Table 4 shows that both mouse and hamster chromosomes may lag at anaphase in hamster–mouse hybrids, often in th, same cell (Fig. 2); hovever, the frequency of laggard is no greater in hybrids than in parental cells. Then were too few scorable hamster–human anaphase fig tires to permit any meaningful analysis.

If segregant chromo somes were delayed in their anaphase movement am are lost because they are excluded from daughte nuclei, we would expect to observe a higher frequency of micronuclei among early telophase figures in hybrit cells than in parental cells. We scored all available earl-telophase figures for hamster–mouse and hamster-human hybrids. Hamster–mouse hybrids did not shov an increase in micronucleus frequency (Table 5). No did the Blo–Daudi hybrids; however, BIo–EUE dit contain 6% cells with micronuclei (compared to 0 am 2% in parent cells); this difference is non-significant a shown by a X² text.

Thus, our studies of metaphase, anaphase am telophase have provided no evidence that chromosom movement is impaired in any way in early-phas hamster–mouse or hamster–human cell hybrids.

C-banded populations of hamster–mouse early hybrid were examined for any’ evidence of non-randor chromosome positioning at metaphase. There were n cells, out of 23 radial and 37 side-on views, in which th mouse and hamster chromosome sets were spatiall separated (side-by-side or concentric) ; mouse an hamster chromosomes were always intermixed within metaphase spreads.

A completely different result was obtained when the positions of hamster or human sets in 40 G-11-banded early hamster–human hybrids were examined. In all 40 scorable hybrid cells (representing both crosses), the pale-blue human chromosomes all lay nearer the centre of radial mitoses, usually completely within a ring of magenta-stained hamster chromosomes (Fig. 3). The positions of hamster and human chromosomes in metaphase figures viewed side-on were consistent with this interpretation, in that a blue core was seen to be encased by superimposed magenta chromosomes. The outermost regions of the metaphase plate contained only magenta chromosomes. This clearly non-random arrangement of chromosomes was seen not to be related to chromosome size, since some group A human chromosomes were central while some group IV hamster chromosomes (which are less than half the size) were peripheral. Nor did the staining difference reflect merely the positions of chromosomes, since occasional hamster chromosomes within the ring were conspicuously magenta stained, while occasionally a blue-stained human chromosome was observed among the peripheral chromosome ring.

We made light-microscope studies of the movement and position of chromosomes in the early phase of two types of cell hybrids. Both sets of hamster–human hybrids that we studied showed rapid preferential segregation, invariably of human chromosomes, while the hamster–mouse hybrids showed a more gradual loss of either hamster or mouse chromosomes in different clones. Hamster–mouse hybrids derived from different combinations of cell lines have been described that segregate mouse (Scaletta et al. 1967) or hamster chromosomes (Graves, 1975); however, it is unusual to find both directions of segregation among hybrids from a single cross. Evidently the (unknown) factors that determine direction of segregation must be rather evenly balanced in these hybrids (and in similar hybrids described by Marin & Pugliatti-Crippa, 1972), so that direction may be established by variations between fused cells; perhaps the dosage of some factor, shown to be important by fusions between polyploid lines (Graves & McMillan, 1984), depends on the cycle phase of the cell at the time of fusion in these cell combinations.

We studied dividing cells fixed with the minimum of disruption early in the generation of hamster–mouse and hamster–human hybrids, in order to test the hypothesis that chromosome segregation results from multipolar mitosis, delayed metaphase congression, anaphase lagging or exclusion from the telophase nucleus. From the numbers of chromosomes lost, and the numbers of cell generations that elapsed between fusion and fixation, we could roughly estimate the level of abnormal mitotic events required to account for segregation. Populations of hamster–mouse hybrids fixed 10–20 days post-fusion could have undergone up to 20 cell generations. Individual hamster–human hybrids isolated as colonies of a few hundred cells, then propagated for a further 7 days could have undergone up to 15. Both these figures are likely to be large overestimates, since we have assumed cell generation times equivalent to those of the parent lines; however, the cycle times of interspecific hybrids are initially very much longer than those of either parent (Graves & Koschei, 1980). Even so, the loss of up to 12 hamster or 24 mouse chromosomes over 20 cell generations would be expected to give rise to an observable abnormality in more than half the mitotic mouse–hamster hybrid cells, while the loss of about 35 human chromosomes over 15 cell generations should be detectable as an average of 2·3 abnormalities per cell. We observed nothing like this level of mitotic abnormality. We found no evidence that multipolar mitosis, delayed metaphase congression, anaphase segregation or micronucleus formation were any more frequent than in parental cells. Nor did segregant chromosomes appear to be involved more frequently in such irregularities than retained chromosomes. It is possible that irregularities, such as multipolar mitoses (Oftebro, 1968) and asyn-chrony (Johnson & Rao, 1970), may have already had their effect in the very earliest divisions, and so were not detected in the hybrids studied here. From our findings, however, we must conclude that aberrant mitosis or delayed movement of segregant chromosomes cannot account for chromosome segregation.

This finding was unexpected, for two reasons. First, interspecific barley hybrids (sexual, not somatic), which segregate chromosomes from one parental set during early embryogenesis, have a high frequency of chromosome lagging (Bennett et al. 1976), suggesting that segregation in this system does result from delayed mitotic movement. Thus, it appears that there is only limited analogy between chromosome segregation in mammalian somatic hybrids and interspecific plant hybrids. Second, our failure to observe abnormal chromosome movement leaves in question the fate of segregated chromosomes. If they are not lost during early mitoses, they may be fragmented and degraded during interphase. We did observe some signs of fragmentation and degradation in early hybrids and have frequently noted cells of established hybrids with multiple rearrangements, or fuzzy puddles of staining material, which may represent the discarded chromo-somes; however, these are hard to quantify. Another possibility is that segregant chromosomes may fail to undergo replication during the S phase.

If the position of a chromosome on the metaphase plate is directly or indirectly related to its chance of elimination, a non-random arrangement of segregant set and retained set chromosomes might be expected in interspecific hybrids. To test this hypothesis, we studied the relative positions of segregant and retained chromosomes in C-banded and G-11-banded hamster–mouse and hamster-human hybrids. Although we detected no non-random positioning of mouse and hamster chromosomes in hamster–mouse hybrids, the positions of hamster and human chromosomes in all hamster–human hybrid metaphase cells were clearly non-random. Segregant (human) chromosomes were observed always to occupy a central position on the metaphase plate, within a ring of hamster chromosomes, and this arrangement was independent of chromosome size.

Other observations of non-random arrangements of genomes in mouse–human cell hybrids have been made (Rechsteiner & Parsons, 1976; Rogers et al. 1983), but these were reported to be transient and considered, therefore, to be unrelated to progressive chromosome segregation. It is likely that in this study the use of disruptive preparative procedures (Colcemid and hypotonic treatment) and the identification of mouse and human chromosomes by morphology alone may have obscured a non-random arrangement. Genome separation has also been reported for interspecific sexual hybrids; for instance, chick–quail hybrids (Bammi et al. 1966) and barley–rye hybrids (Finch et al. 1981; Schwarzacher-Robinson et al. 1987). In interspecific barley hybrids, segregant set chromo-somes were observed to occupy positions outside a ring of retained set chromosomes (Finch, 1983); quite the opposite pattern to that reported here.

There are two different ways in which the aberrant positioning we observed might be related to segregation. Human chromosomes may occupy an aberrant central position on the metaphase plate as a result of aberrant, or at least less efficient interactions between centromeres and spindle fibres. We have recently demonstrated (Zelesco & Graves, unpublished data), that one parental form of β-tubulin is repressed or under-expressed in hamster–mouse hybrids; thus it is possible that there is a limited supply of spindle fibres in interspecific hybrids. If there were an intrinsic difference (due to centromeric DNA sequences) or a difference induced (e.g. by DNA modification, see Drahovsky et al. 1980, 1981; Finch, 1983) in the efficiency with which the hamster and human centromeres bind to spindle fibres, human chromosomes might engage fewer (or more) than hamster, and the resultant of the unequal torsion may force the human set into a central position on the spindle. Aberrant interactions between segregant set chromosomes and the spindle may be directly responsible both for the aberrant positioning of human chromosomes, which we observed, and for aberrant movement resulting in segregation. However, we did not observe this latter effect.

Perhaps, then, aberrant positioning of segregant set chromosomes, such as we and others have observed, has a more direct relationship to segregation. Bennett (1984) has described a strong tendency for parental genomes in interspecific and intergeneric hybrid plant cells to separate concentrically, and has provided evidence from reconstructed nuclei that it is the parental genome that occupies the peripheral domain that is predisposed for mitotic chromosome elimination. The positions of chromosomes in interphase cells is now thought to be highly ordered, and to have some functional significance (Agard & Sedat, 1983). Meta-phase positioning may reflect this interphase organization. Bennett (1984) has demonstrated a correlation in hybrid plant cells between gene activity, including suppression of nucleolar-organizing regions, and intra-nuclear chromosome or genome position. In a somatic cell hybrid, chromosomes of the retained set may assume their normal relative positions, while those of the segregant set cannot be accommodated in a normal spatial arrangement. A breakdown in the normal arrangements of segregant set chromosomes may result in a disruption of the normal synthetic activities of these chromosomes. This could be responsible for the repression of ribosomal RNA (Onishi el al. 1984), histones (Ajiro et al. 1978) and tubulin (Zelesco and Graves, unpublished data) as well as, perhaps, affecting the stability of X chromosome inactivation and the conformation of the inactive X (Gartleret al. 1985). It could also, conceivably, impede normal DNA replication or repair of segregant chromosomes, and lead to their subsequent loss from daughter cells. This hypothesis does not contradict those observations that suggest that the rate of segregation is related to number of cell divisions (Schall & Rechsteiner, 1978), since failure of replication or repair could be limited to the S phase of the cycle. This hypothesis could be tested directly by studying DNA synthesis in rodent and human chromo-somes of early-phase hybrids.

This work was funded by grants from the Australian Research Grants Scheme and the National Health and Medical Research Council. Some of this work was completed while Dr Zelesco held a Commonwealth Postgraduate Research Scholarship.