Journal of Cell Science 104, 31-36 (1993)
Printed in Great Britain © The Company of Biologists Limited 1993
31
Proliferative behaviour of high-ploidy cells in two murine tumour lines
Carmen de la Hoz* and Antonio Baroja
Dpto. de Biología Celular y Ciencias Morfológicas, Facultad de Medicina y Odontología, Universidad del País Vasco,
Leioa, 48940 Vizcaya, Spain
*Author for correspondence
SUMMARY
The presence of high-ploidy cells in malignant tumours
has long been documented. However, the biological significance of these cells is not known and there is a great
deal of controversy over their proliferative potential. We
have analysed the behaviour of these cells in two murine
tumour lines, B16F10 melanoma and 3T3A31M
angiosarcoma, determining their DNA content by
microspectrophotometry and using time-lapse film
studies. We have found a discrepancy between the presence of high-ploidy cells in metaphase and the absence
of hyperploid telophases. High-ploidy metaphases may
be aborted (mitotic polyploidization), prolonged in time
or evolve in the form of multipolar, generally tripolar,
mitoses. Our results suggest that high-ploidy cells are
capable of proliferating, despite certain peculiarities in
their cell cycle, and constitute a tumour subpopulation
whose role in neoplasia merits further study.
INTRODUCTION
This controversy has led us to study cell proliferation in
high-ploidy cells. For our research, we selected two murine
tumour lines, B16F10 melanoma and 3T3A31M angiosarcoma. According to our results, high-ploidy tumour cells
have proliferative capacity, although it is inferior to that of
modal-ploidy cells in usual conditions.
In malignant tumours, and especially the well-differentiated
ones, there is often a predominant cell population whose
DNA content is above the normal diploid value. This population constitutes the so-called tumour stem line (Böhm
and Sandritter, 1975). In many cases, moreover, there are
a number of cells whose ploidy is higher than that of the
predominant or modal population. High-ploidy cells are
present even in incipient tumours or tumour lines (which
are fairly homogeneous in DNA content), and have long
been singled out by pathologists, who regard them as a clear
sign of malignancy. Today, however, the biological significance of high-ploidy tumour cells is still not known (Liautaud-Roger et al., 1990).
Specifically, one controversial aspect of the behaviour of
these cells is their proliferative capacity. While some
authors suggest that the proliferative potential of these cells
might be reduced (Gustavino et al., 1991) or even nil
(Holzner and Golob, 1968; Frankfurt et al., 1985), other
authors claim that high-ploidy cells proliferate in a manner
similar to that of modal ploidy cells (Dooley et al., 1991).
It is important to learn what their proliferative potential is,
since this will help us determine the role of high-ploidy
cells in the tumour population. If high-ploidy cells turned
out to be incapable of dividing, they would simply represent a marginal tumour subpopulation reflecting an aberrant
cell proliferation, and the stem cell function (Pierce, 1978)
would be reserved for modal-ploidy cells. If, on the contrary, high-ploidy cells were capable of undergoing productive mitosis, they would be of potential use to the
tumour.
Key words: tumour proliferation, high-ploidy cells, malignancy
MATERIALS AND METHODS
Cells and culture conditions
This study concentrated on two murine tumour lines, B16F10
melanoma and 3T3A31M angiosarcoma. The B16F10 line is a cell
variant of B16 melanoma selected for its tendency to generate a
high number of pulmonary tumour foci when it is inoculated in
the tail vein (Fidler, 1973). The 3T3A31M line was developed
from the 3T3A31 line through a process of in vivo and in vitro
selection (Zvibel and Raz, 1985). These tumour lines were cultured in DMEM medium to which 10% FCS was added. The cultures were incubated at 37˚C in a humidified atmosphere of 10%
CO2 and were used in experiments when they were in a phase of
exponential growth. Every two days in the case of B16F10
melanoma, and every three days in the case of 3T3A31M angiosarcoma, the cells were subcultured by enzymic procedures (0.1%
trypsin, 2 mM EDTA solution). In order to avoid possible genetic
drifts in the line, cells cultured for more than one month were discarded and new ones were drawn from the reserve kept in liquid
nitrogen for further experiments.
Measurement of DNA content of cells in different
cell cycle phases
The DNA content of the tumour lines was quantified by absorption microspectrophotometry of cells cultured on slides and
32
C. de la Hoz and A. Baroja
stained using the Feulgen procedure. In contrast to flow cytometry, this method makes it possible to visually select the cells to
be measured so that cells in interphase, metaphase and telophase
can be adequately differentiated. The acid hydrolysis of the Feulgen method was carried out with 1 M HCl at 60˚C for 11 minutes. Later the cells were incubated at room temperature in Schiff
reactant for 1 hour. For DNA quantification, a microspectrophotometer equipped with a fast-scanning stage (UMSP-05, Zeiss,
Germany) connected to a computer (M-20, Olivetti, Italy) was
used. Measuring conditions included a 40× Plan Apo objective
(1.0 numerical aperture), 0.8 µm diaphragm and a 40× Plan condenser (0.60 numerical aperture). A 560 nm monochrome light
was used for stimulation. For quantification, the integration
method was chosen (Caspersson, 1979). Using software developed
in our department (Esteban, 1986), we obtained DNA content histograms of the lines. In order to express the DNA content in C
units, murine lymphocytes were stained at the same time as the
tumour cultures and used as our 2C control.
Three subpopulations at the interphase, metaphase and
telophase stages of the cell cycle were selected. The DNA content of at least 600 cells in each subpopulation was quantified.
From the DNA-content histogram, we then determined the proportion of cells in the different phases of the cell cycle and the
proportion of high-ploidy cells, using the peak reflection method
(Dean, 1986).
This method basically consists of plotting the data from the
DNA-content histogram onto two Gaussian curves corresponding
to the cells in G0/G1 and G 2 phases, in the case of the interphase
cells. The cells located in the interval between the two curves are
the cells in S phase. In the case of cells in metaphase and
telophase, the nuclear DNA content is plotted on a single Gaussian curve. In all cases, high-ploidy cells were considered to be
those situated to the right of the Gaussian curves.
Cell proliferation analysed by time-lapse film
studies
The cell cultures were filmed using a system designed in our
department that makes it possible to monitor the evolution of the
cells over time under programmed conditions of temperature,
humidity and CO2 atmosphere. The culture flask was placed in a
black perspex box fitted to the stage of an inverted microscope
(IM 35, Zeiss). Temperature was maintained at 37˚C by a heating coil connected to a fan that keeps air circulating inside the
box. The film was made using a Beaulieu 4008 Z-II camera and
super 8 film. The camera was fitted with an optoelectronic sensor
causing the film to move forward frame by frame. A computer
was used to programme filming conditions, including the frame
change rate, which for these experiments was set at one per
minute. Black and white photographs were obtained from super 8
photograms using Agfapan-X 100 film.
RESULTS
DNA content and presence of high-ploidy cells in
tumour populations in different phases of the cell
cycle
Fig. 1 corresponds to a microphotograph taken of a B16F10
culture stained using the Feulgen method and observed
through confocal microscopy. In this figure, next to a group
of modal-ploidy cells, a high-ploidy cell can be seen.
The DNA content of the B16F10 cells in the interphase,
metaphase and telophase of the cell cycle is shown in Fig.
2A. The DNA-content histogram of the cells in interphase
Fig. 1. Microphotograph of a Feulgen-stained B16F10 culture
taken using confocal microscopy. Note the presence of a highploidy nucleus.
has a classical bimodal pattern corresponding to that of a
proliferative cell population. A small percentage of cells
located to the right of the Gaussian curve corresponds to
the cells in the G2 phase of the cell cycle and is made up
of cells whose ploidy is higher than that of the modal population. These are what we have termed high-ploidy cells.
Although most of the metaphases have a DNA content
corresponding to that of the modal population following the
duplication of their DNA, here the metaphasic population
is heteroploid and includes a small percentage of highploidy cells (2.4%).
Interestingly, the DNA content of the B16F10 telophases
is very homogeneous and no high-ploidy cells in telophase
were detected in the sample studied.
Fig. 2B shows the DNA-content histograms of 3T3A31M
cells in interphase, metaphase and telophase. As with the
B16F10 line, here the interphase population includes a percentage of high-ploidy cells, but in this case the percentage is higher (9%). The cells in metaphase also include a
subpopulation of high-ploidy cells whose DNA content is
practically double that of the modal cells. This subpopulation is approximately 7% of the total. Again, the telophasic daughter cells constitute a very homogeneous subpopulation: in a sample of 600 nuclei, we were unable to find
any high-ploidy telophasic cells.
Evolution of high-ploidy tumour cells recorded in
time-lapse film studies
In analysing the filmed records of the B16F10 and
3T3A31M cultures, and, specifically, in observing the evolution of the high-ploidy cells in metaphase, as well as occasional standard cell divisions giving rise to two daughter
cells (which lasted longer than usual), we detected two
types of sequence, which are represented schematically in
Fig. 3. (a) Mitosis, which had progressed until metaphase,
is interrupted, and a single interphasic cell is generated.
This phenomenon, called mitotic polyploidization, generates, from a single high-ploidy cell, another cell of higher
ploidy (Fig. 3A). This could account for the existence of
High-ploidy tumour cells
A
B
B16F10 MELANOMA
33
3T3A31M ANGIOSARCOMA
100
100
Interphase
Interphase
6c
4c
8c
12c
DNA CONTENT
DNA CONTENT
150
150
Metaphase
8c
Metaphase
12c
DNA CONTENT
DNA CONTENT
300
300
Telophase
Telophase
4c
DNA CONTENT
6c
DNA CONTENT
Fig. 2. DNA-content histograms of the B16F10 (A) and 3T3A31M (B) lines in three different phases of the cell cycle. At least 600 nuclei
were measured in each of the subpopulations considered. The filled histogram bars represent high-ploidy cells. The DNA content of
interphasic cells has the classical bimodal profile delimited by the two Gaussian curves corresponding to cells in G0/G1 and G2 phases of
the cell cycle, respectively. This histogram includes 2.5% (B16F10) and 9% (3T3A31M) of high-ploidy cells. The DNA content of the
metaphasic subpopulation has been plotted on a single Gaussian curve. In this case there is also a small percentage of high-ploidy
metaphases (2.4% in the B16F10 line and 2.5% in the 3T3A31M line). However, the telophasic subpopulation is highly homogeneous
and the presence of high-ploidy telophasic daughter cells is indetectable in the sample analysed.
high-ploidy metaphasic cells in the absence of the corresponding high-ploidy telophases. (b) High-ploidy metaphasic cells divide productively in a tripolar fashion, giving
rise to three daughter cells (Fig. 3B). Careful culture film
monitoring of these daughter cells revealed that they reassume an interphasic morphology with evident signs of
dynamic cell activity, and provided evidence that the daughter cells from tripolar divisions were, in turn, capable of
dividing (Fig. 4). This type of evolution is compatible with
the absence or scarcity of high-ploidy telophase cells, since
one high-ploidy metaphasic cell can generate three modalploidy daughter cells.
Tripolar mitosis in tumour cultures
On Feulgen-stained slides, a count was made of the number
of tripolar mitoses and their proportion out of the total
mitoses in the two tumour lines under study. Fig. 5 shows
a tripolar mitosis taken from a B16F10 culture. The results
of our quantification of this type of mitosis were as fol-
lows: 0.3% tripolar mitoses in the B16F10 line and 2.4%
tripolar mitoses in the 3T3A31M line. The higher proportion in the latter line is consistent with the higher incidence
of high-ploidy cells in this line.
DISCUSSION
Despite the fact that the presence of high-DNA-content
tumour cells has long been documented, the role of these
cells in the biology of the tumour has not been elucidated.
The number of these cells is generally higher in undifferentiated tumours and in advanced neoplastic stages, and
their presence is therefore normally associated with a worse
clinical prognosis (Auer et al., 1984; Matsuura et al., 1986).
However, this is not true of all tumours, since there are
cases in which a low DNA content is associated with a poor
prognosis (Pui et al., 1991).
One reason why the behaviour of high-DNA-content
34
C. de la Hoz and A. Baroja
Evolution of high ploidy mitoses
Fig. 3. Graphic representation of two types of evolution of the
high-ploidy metaphases observed in filmed images of cultures of
the tumour lines studied. (A) Shows how a high-ploidy metaphase
divides in a tripolar manner, giving rise to three daughter cells.
(B) Shows a case of mitotic polyploidization, i.e. the interruption
of a mitosis in metaphase to regenerate an interphasic cell of
higher ploidy than the original.
Mitotic polyploidization
particular tumour and its degree of malignancy. Very little
research has been aimed at determining the biological properties of the tumour cells with a high DNA content (Brodsky and Uryvaeva, 1985).
Among the controversial aspects of the biological performance of high-DNA-content tumour cells is their proliferative potential. Cells whose DNA content is above the
normal diploid value seem clearly able to proliferate, since
the predominant or stem line tumour population frequently
has a higher than diploid DNA content (Büchner et al.,
1985). What is not clear is the proliferative potential of
tumour cells whose DNA content is higher than the modal
(that of the stem line), whatever that content may be.
Although some authors claim that these cells have a proliferative capacity similar to that of the rest of the population (Dooley et al., 1991), others suggest that the polyploid
A
B
Tripolar mitosis
tumour cells remains obscure is that most of the current literature on the subject consists of clinical studies aimed at
establishing a correlation between the DNA content of a
Fig. 4. Evolution of a tripolar mitosis in a
culture of the 3T3A31M cell line (timelapse microcinematography sequence).
(A) (0 min) Cell that will undergo tripolar
mitosis (1) has become rounded and
refringent. Next to cell 1 is another,
clearly smaller, cell in mitosis. (B) (30
min) Cell 1 has taken on the tripolar
shape. (C) (1 h) The three medium-sized
daughter cells (2, 3 and 4) have adopted
an interphasic morphology. (D) (2 h 30
min) and (E) (4 h 30 min) Cell 2 and the
other two daughter cells have moved in
opposite directions, thus lengthening the
distance between them. (F) (14 h 30 min)
Cell 2 has begun mitosis. (G) (16 h 10
min) Daughter cell 4, resulting from the
tripolar division of cell 1, begins mitosis.
Note the two daughter cells from the
division of cell 2 (5 and 6). (H) (18 h 10
min) The third daughter cell (3) derived
from the tripolar division begins mitosis.
Next to it are the two daughter cells from
the division of cell 4 (cells 7 and 8).
High-ploidy tumour cells
Fig. 5. Microphotograph of a culture of B16F10 cells stained by
the Feulgen method, showing a tripolar mitosis.
cells in general have a longer cell cycle and a smaller probability of undergoing further mitoses (Gustavino et al.,
1991), while still others suggest that the mitotic capacity is
lost altogether (Holzner and Golob, 1968).
From our results, we deduce that the tumour cells whose
ploidy is higher than that of the stem line are capable of
dividing, although their cell cycle has certain peculiarities.
Thus, in the DNA-content histograms of tumour subpopulations in different phases of the cell cycle, the absence of
heterogeneity in the DNA content of the cells in telophase
is especially noteworthy, in contrast to the presence of
metaphasic cells whose DNA content is above the modal
value. This discrepancy suggests that, while, in the tumour
lines analysed, the high-DNA-content cells in metaphase
are capable of initiating the mitotic process, the division is
not completed in the same way as in the rest of the cell
population.
The evolution of the high-DNA-content tumour
metaphases is a problem whose solution will determine the
role attributed to high-ploidy cells (cells whose ploidy is
over that of the predominant population) within the context
of the heterogeneous tumour population, since this role will
depend, among other factors, on their proliferative capacity. Several decades ago, Bader (1959) deduced, on the
basis of a much narrower distribution of DNA content in
anaphases than in metaphases, that most of the metaphases
with DNA content higher than the modal never completed
mitosis. This interpretation supported the existence of
tumour stem cells with modal ploidy. However, other
authors have recently questioned this model, suggesting that
all cells complete their division regardless of their chromosome number or their DNA content, and claim that any
tumour cell is a potential stem cell (Dooley et al., 1991).
These authors find support for this claim in their finding of
heterogeneous DNA content in the telophases of certain
tumour lines.
Although our results show a discrepancy between the
presence of high-ploidy cells over the modal value in the
metaphases and the homogeneity of DNA content of the
telophases, we do not consider that this entails the inability of high-DNA-content metaphases to complete cellular
division, as was suggested previously (Bader, 1959; Pfitzer
and Pape, 1973; Frankfurt et al., 1985). It is true that some
35
of these divisions are abortive, as we have verified in filmed
records of tumour line cultures, as follows: the cell rounds
up and detaches itself from the substratum, initiating a mitosis that is suddenly interrupted, after which the cell reattaches itself to the substratum and resumes an interphase
morphology. This phenomenon, known as mitotic polyploidization, has been described in detail by Brodsky and
Uryvaeva (1985). Moreover, it is compatible with the
ploidy discrepancies between the metaphases and
telophases that we have detected.
However, there are other possible evolutions of highDNA-content metaphases compatible with the absence of
corresponding high-DNA-content telophases. For example,
chromosome migration to the cell poles could be hindered
by the presence of anomalous or especially numerous chromosomes, a situation which is habitual in high-DNA-content metaphases. Indeed, in our filmed records of tumour
cultures, we have verified that metaphase lasts longer in
hyperploid metaphases, a phenomenon that has also been
reported by other authors (Sisken et al., 1985). This increase
in the duration of the metaphase may also contribute to the
discrepancy in the proportion of high-DNA-content
metaphases and telophases.
Moreover, a high-ploidy metaphase is able to evolve
through multipolar, especially tripolar, mitosis, as verified
in our time-lapse film studies of cell cultures. This phenomenon, which is not exceptional in tumour biopsies
(Robbins and Kumar, 1990), is found, in the tumour lines
we studied, in proportions ranging between 0.5% and 2.5%
of total mitoses. Spontaneous depolyploidization through
multipolar mitosis is frequent in other biological models
(Afon’kin, 1989), and is compatible with the absence of
high-DNA-content telophases in a tumour population with
hyperploid metaphases.
In conclusion, our results suggest that high-ploidy tumour
cells are capable of proliferating, although this potential is
quantitatively inferior to that of the modal cells. The role
of these cells in tumour development merits further study.
REFERENCES
Afon’kin, S. Yu. (1989). Induced and spontaneous polyploidization in large
amebas. Int. Rev. Cytol. 115, 231-265.
Auer, G.U., Eriksson, E., Azavedo, E., Caspersson, T.O., and Wallgren,
A.S. (1984). Prognostic significance of nuclear DNA content in
mammary adenocarcinomas in humans. Cancer Res. 44, 394-396.
Bader, S.A. (1959). A cytochemical study of the stem cell concept in
specimens of a human ovarian tumor. J. Biophys. Biochem. Cytol. 51,
217-229.
Böhm, N. and Sandritter, W. (1975). DNA in human tumors: a
cytophotometric study. Curr. Top. Pathol. 60, 151-219.
Brodsky, V.Ya. and Uryvaeva, I.V. (1985). Genome Multiplication in
Growth and Development. Cambridge University Press.
Büchner, Th., Hiddemann, W., Schumann, J. et al. (1985). DNA
aneuploidy - A common cell marker in human malignancies and its
correlation to grade, stage and prognosis. In Tumor Aneuploidy (ed. Th.
Büchner, C.D. Bloomfield, W. Hiddemann, D.K. Hossfeld and J.
Schumann), pp. 41-52. Berlin, Heidelberg: Springer-Verlag.
Caspersson, T. (1979). Quantitative tumor cytochemistry. Cancer Res. 39,
2341-2355.
Dean, P.H. (1986). Data analysis in cell kinetics research. In Techniques in
Cell Cycle Analysis (ed. J.W. Gray and Z. Darzynkiewicz), pp. 207-253.
Clifton, New Jersey: Humana Press.
Dooley, W.C., Allison, D.C. and Robertson, J. (1991). Discrepancies
36
C. de la Hoz and A. Baroja
among the metaphase, telophase, and the G0/G1 and G2 DNA peaks of
heteroploid cell lines. Cytometry 12, 99-106.
Esteban, J.L. (1986). Desarrollo del equipamiento para el estudio de la
diferenciación celular. Tesis de licenciatura, Universidad del País Vasco.
Fidler, I.J. (1973). Selection of successive tumor lines for metastasis.
Nature New. Biol. 242, 148-149.
Frankfurt, O.S., Chin, J.L., Englander, L.S., Greco, W.R., Pontes, J.E.
and Rustum, Y.M.(1985). Relationship between DNA ploidy, glandular
differentiation and tumor spread in human prostate cancer. Cancer Res.
45, 1418-1423.
Gustavino, B., Bassani, B. and Pacchierotti, F. (1991). Vinblastineinduced numerical chromosome changes and selection processes in
mouse bone marrow cells. Mutat. Res. 248, 45-50.
Holzner, J.H. and Golob, E. (1968). Changes of the DNA content in cells
from cervical cancer under cytostatic and radiation therapy. Acta Cytol.
12, 473-477.
Liautaud-Roger, F., Dufer J. and Coninx, P. (1990). Tetraploid cycle in
ageing solid tumours. Cell Tiss. Kinet. 23, 261-269.
Matsuura, H., Sugimachi, K., Ueo, H., Kuwano, H., Koga, Y. and
Okamura, T. (1986). Malignant potentiality of squamous cell carci-
noma of the esophagus predictable by DNA analysis. Cancer 57, 18101814.
Pfitzer, P. and Pape, H.D. (1973). Characterization of tumor cell
populations by DNA-measurements. Acta Cytol. 17, 19-26.
Pierce, G.B. (1978). Cancer: a Problem of Development Biology, pp. 68-87.
Englewood Cliffs, New Jersey: Prentice-Hall Inc.
Pui, C-H., Dodge, R.K., Look, A.T., et al. (1991). Risk of adverse events in
children completing treatment for acute lymphoblastic leukemia: St. Jude
total therapy studies. VIII, IX and X. J. Clin. Oncol. 9, 1341-1347.
Robbins, S.L. and Kumar, V. (1990). Patología Humana, pp. 217-238.
Mexico: Interamericana. McGraw-Hill.
Sisken, J.E., Bonner, S.V., Grasch, S.D., Powell, D.E. and Donaldson,
E.S. (1985). Alterations in metaphase durations in cells derived from
human tumors. Cell Tiss. Kinet. 18, 137-146.
Zvibel, I. and Raz, A. (1985). The establishment and characterization of
a new BALB/c angiosarcoma tumor system. Int. J. Cancer 36, 261272.
(Received 26 May 1992 - Accepted 12 September 1992)