Chromosomal instability and cytoskeletal defects in
oral cancer cells
William S. Saunders*†, Michele Shuster‡, Xin Huang‡, Burhan Gharaibeh‡, Akon H. Enyenihi*, Iver Petersen§,
and Susanne M. Gollin‡
Departments of *Biological Sciences and‡Human Genetics, University of Pittsburgh and the University of Pittsburgh Cancer Institute, Pittsburgh, PA 15260;
and§Institute of Pathology, University-Hospital Charite, D-10098 Berlin, Germany
Edited by Janet D. Rowley, The University of Chicago Medical Center, Chicago, IL, and approved October 25, 1999 (received for review July 1, 1999)
Oral squamous cell carcinomas are characterized by complex, often
near-triploid karyotypes with structural and numerical variations
superimposed on the initial clonal chromosomal alterations. We
used immunohistochemistry combined with classical cytogenetic
analysis and spectral karyotyping to investigate the chromosomal
segregation defects in cultured oral squamous cell carcinoma cells.
During division, these cells frequently exhibit lagging chromo-
somes at both metaphase and anaphase, suggesting defects in the
mitotic apparatus or kinetochore. Dicentric anaphase chromatin
bridges and structurally altered chromosomes with consistent long
arms and variable short arms, as well as the presence of gene
amplification, suggested the occurrence of breakage–fusion–
bridge cycles. Some anaphase bridges were observed to persist
into telophase, resulting in chromosomal exclusion from the re-
forming nucleus and micronucleus formation. Multipolar spindles
were found to various degrees in the oral squamous cell carcinoma
lines. In the multipolar spindles, the poles demonstrated different
levels of chromosomal capture and alignment, indicating func-
tional differences between the poles. Some spindle poles showed
premature splitting of centrosomal material, a precursor to full
separation of the microtubule organizing centers. These results
indicate that some of the chromosomal instability observed within
these cancer cells might be the result of cytoskeletal defects and
quence of genetic instability can be an alteration in copy number
of one or more genes, a change in gene expression, or a change
in gene structure such that the protein sequence is altered (1).
These genetic changes can lead to either increased or diminished
protein activity or can create a new gain-of-function activity for
the altered protein. Genetic instability can result from changes
in chromosome structure, through errors in DNA metabolism,
repair, recombination, or other rearrangements of the DNA
sequence (2), or from misregulation of the cell cycle, for
example, uncoupling DNA replication from cell division (3) or
centrosomal duplication from division (4). Abnormalities in the
chromosomal segregational apparatus are also likely to play an
important role in genetic instability. These include centrosomal
defects, defects in kinetochore–microtubule attachment, and
movement of chromosomes relative to the poles. While it is likely
that all of these changes contribute to carcinogenesis, the extent
to which each plays a part is still largely uncharacterized.
One of the most striking manifestations of genetic instability
in cancer cells is the variation observed in the karyotypes of
different cells, even within the same tumor. A well-characterized
example is squamous cell carcinomas of the oral cavity. The
karyotypes of oral squamous cell carcinomas (OSCC) are com-
plex, often near triploid, and are composed of multiple numer-
ical and structural abnormalities, including deletions, balanced
and unbalanced translocations, isochromosomes, dicentric chro-
mosomes, and homogeneously staining regions (5). Despite the
numerous and diverse abnormalities, classical cytogenetic anal-
yses of OSCC cells have revealed several consistent chromo-
enetic instability or changes in chromosome structure and
numbers is an important facet of oncogenesis. The conse-
somal changes, such as breakpoints at bands 1p11–1p13, 3p14,
and 11q13 (6) and amplification of band 11q13 (7, 8). Chromo-
somal rearrangements may be driven and?or fixed by activation
11q13 amplicon includes CCND1, the cyclin D1 gene, which
plays a role in promoting cell division (9).
We have begun to investigate the source of chromosomal
instability in cultured cells derived from human OSCC. We have
found high levels of both numerical and structural chromosome
variations, including marker chromosomes and dicentrics. Dur-
ing mitotic division, we observed multiple segregational defects,
including multipolar spindles, lagging chromosomes, and an-
aphase bridges, leading to chromosomal breakage at telophase
into centric and acentric micronuclei. In some cases, the an-
aphase bridges contained DNA from chromosome 11. An ad-
ditional observed defect was the untimely and unnatural division
of the nuclear mitotic apparatus protein NuMA at the centro-
somes, apparently an early step in the splitting of the spindle
poles. These observations support the conclusion that many of
the cell-to-cell variations in chromosome structure and numbers
in cancer cells result from (i) dicentric chromosomes forming
anaphase bridges, which break and reform similar to that
observed originally in maize by Barbara McClintock (10) and (ii)
specific structural defects in the spindle and chromosome seg-
Materials and Methods
Cell Culture. In this study, we examined three of our OSCC cell
lines, UPCI:SCC003, UPCI:SCC131, and UPCI:SCC172, un-
cloned outgrowths developed from primary tumors from con-
senting patients who had not been treated previously with
chemotherapy or radiation therapy (S.M.G., J. K. Reddy, S.
Comsa, K. M. Rossie, C. M. Lese, M. S., B. N. Appel, R. Wagner,
E. N. Myers, and J. T. Johnson, unpublished work). The cells
were cultured in MEM with Earle’s salts supplemented with
nonessential amino acids (Life Technologies, Grand Island,
NY), L-glutamine, gentamicin, and 10% FBS (Irvine Scientific).
Classical Cytogenetic Analysis. The cell cultures were harvested by
using standard cytogenetic techniques (11). Metaphase cells
from all three cell lines were digitally imaged and karyotyped
after trypsin-Giemsa banding (7) with a CytoVision Ultra Sys-
tem (Applied Imaging, Santa Clara, CA). Karyotypes have been
or will be published elsewhere (ref. 7, and C. M. Lese, T. Ried,
This paper was submitted directly (Track II) to the PNAS office.
SKY, spectral karyotyping; MTOC, microtubule organizing center; BFB, breakage–fusion–
bridge; FISH, fluorescence in situ hybridization; NuMA, nuclear mitotic apparatus protein;
CREST, calcinosis, Raynaud’s phenomenon, esophageal dysmotility, sclerodactyly, and
†To whom reprint requests should be addressed. E-mail: firstname.lastname@example.org.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
January 4, 2000 ?
vol. 97 ?
no. 1 ?
W. Gottberg, J. W. Wilson, S. C. Reshmi, J. K. Reddy, J. T.
Johnson, E. N. Myers, and S.M.G., unpublished work). To
examine the number of centromeres per chromosome, the
metaphase cells were stained with C-banding using barium
hydroxide and Giemsa stain (12).
Chromosome 11 Painting. For chromosome 11 painting of cells
growing in situ on Lab-Tek chamber slides (Nalge Nunc, Na-
perville, IL), the cell culture medium was removed by aspiration,
and the cells were fixed in 3:1 methanol:acetic acid for 100 min,
rinsed several times with 70% acetic acid solution, and air dried.
The slides were aged for 30 min in 2? SCC at 37°C, pretreated
with RNase for 1 hr at 37°C, then with pepsin for 10 min at 37°C,
followed by postfixation in 1% formaldehyde for 7 min. Slides
were denatured for 2 min at 70°C in 70% formamide?2? SSC,
then dehydrated and air-dried prior to hybridization. The Spec-
trum Green directly labeled WCP11 probe (Vysis, Downers
Grove, IL) was hybridized according to the manufacturer’s
directions. Slides were counterstained with the DNA-specific
fluorescent dye 4?,6-diamidino-2-phenylindole (DAPI; Sigma)
in antifade solution (Oncor), stored at ?20°C prior to viewing,
analyzed with an Olympus (New Hyde Park, NY) BHS fluores-
cence microscope, and images were captured with a CytoVision
Spectral Karyotyping (SKY). Prior to SKY, the metaphase cells
were G-banded as above. Within 24 hr, the slides were destained
and hybridized using the SkyPaint 24 color fluorescence in situ
hybridization (FISH) painting kit (Applied Spectral Imaging,
Carlsbad, CA) as described in (13). Briefly, the G-banded slides
were washed twice in xylene for 5 min each, rinsed in methanol
for 5 min, and dehydrated through a graded ethanol series. Slides
were postfixed immediately (10 min in 1% formaldehyde?1?
PBS?50 mM MgCl2), washed for 5 min in PBS, dehydrated
through a graded ethanol series, and air-dried. Slides were
denatured, hybridized, and washed according to the manufac-
turer’s instructions. Hybridized metaphase cells were captured
and analyzed on the SkyVision I System with SKYVIEW software
(Applied Spectral Imaging).
Immunofluorescence. Cells at passage 4–35 were grown on glass
coverslips (Corning) for 24–72 hr, fixed in methanol for 30 min
at ?20°C, and air-dried. Samples were either processed within 1
hr or stored at ?20°C with desiccant. The coverslips were
hydrated in PBS containing 1% BSA and subject to indirect
immunofluorescence with the following primary Abs: rabbit Abs
a patient with the (CREST) variant of scleroderma (calcinosis,
Raynaud’s phenomenon, esophageal dysmotility, sclerodactyly,
and telangiectasias) at 1:80 or undiluted mouse monoclonal
anti-tubulin. The coverslip with primary Ab was placed in a
humid chamber at 37°C for 1 hr and then washed several times
with PBS?1% BSA. The secondary Abs were Cy3-conjugated
sheep anti-mouse IgG (Sigma) used at 1:1,000, FITC-conjugated
goat anti-mouse (Sigma) at 1:100, or FITC-conjugated goat
anti-human IgG at 1:100 (Roche Molecular Biochemicals). In all
cases, the coverslips were counterstained with DAPI at 1–3
For these analyses, we selected three of our OSCC cell lines,
designated UPCI:SCC003, UPCI:SCC131, and UPCI:SCC172.
The TP53 gene (exons 5–8) in all three tumors and derived cell
lines was sequenced and was found to be wild type. The tumor
tissue was negative for human papillomavirus by the PCR. To
examine the variation in chromosome number and structure
between different cells within a single culture, G-banding and
SKY were performed. The latter technique utilizes a DNA-
hybridization probe mixture in which individual chromosomes
are labeled with one to five fluorophores. Chromatin from each
chromosome can be discriminated by spectral imaging as having
a distinct dye spectrum, enabling characterization of the origin
of segments of rearranged chromosomes (Fig. 1A). The karyo-
type of the UPCI:SCC172 cell line was observed to be highly
aneuploid, containing gains, losses, and structural abnormalities
of numerous chromosomes, including 1, 3, 7, 9, and 11. No two
karyotypes were identical, varying in chromosome copy number
and?or structure. This is illustrated in the supplemental material
(Table 1) on the PNAS web site (www.pnas.org) for five well-
characterized UPCI:SCC172 cells examined by G-banding and
SKY. Karyotypic variability is observed for most of the more
than 30 OSCC cell lines we have karyotyped by G-banding (C.
M. Lese, T. Ried, W. Gottberg, J. W. Wilson, S. C. Reshmi, J.
K. Reddy, J. T. Johnson, E. N. Myers, and S.M.G., unpublished
Of particular note was the observation that many cells within
UPCI:SCC172 displayed what appeared to be related chromo-
somal structural variations, examples of which are shown in Fig.
1B. The SKY analysis shows these exceptional chromosomes to
include segments of chromosomes 3, 4, 8, and 11. The marker
chromosomes, labeled a–c, appear to have much of their DNA
in common, but have different short-arm sequences. All three
marker chromosomes share portions of chromosomes 4 and 11,
but have different short-arm terminal segments derived from
chromosomes 3 and 8. Unfortunately, cytogenetic nomenclature
ignores this sort of variation, focusing only on ‘‘clonal’’ events in
chromosome ‘‘c’’ shown in B. (B) Normal chromosomes 3, 4, 8, and 11 from this culture are shown. Marker chromosomes, also from UPCI:SCC172, are indicated
as a, b, and c. (C) A representative dicentric chromosome is shown from a Colcemid-arrested UPCI:SCC131 cell C-banded to highlight the centromeres.
(A) Display image of metaphase chromosomes from a Colcemid-arrested UPCI:SCC172 cell after SKY analysis. Arrowhead identifies the marker
www.pnas.orgSaunders et al.
common between cells. Marker chromosomes have also been
observed in the UPCI:SCC003 cell line, but the UPCI:SCC131
line has not produced sufficiently high-quality metaphase
spreads to hybridize optimally with the SKY technique. These
results demonstrate the remarkable evolutionary variation in
chromosome structure between different OSCC cells within the
same culture. Because the cell lines were uncloned outgrowths
from tumor biopsies, it is possible that the variant marker
chromosomes may have evolved in the primary tumors prior to
culture. To determine whether the chromosomal structural
variation in the cultured cells included changes in the number of
centromeres per chromosome, we also performed C-banding of
metaphase chromosome spreads, which revealed the presence of
varying numbers of dicentric chromosomes (Fig. 1C).
To determine the source of this chromosomal variation, we
examined mitotic segregation in the OSCC cell lines by indirect
immunofluorescence with Abs to NuMA to mark the centro-
somes (provided by Duane Compton, Dartmouth Medical
to mark the kinetochores (provided by Carol Feghali and
Thomas Medsger, University of Pittsburgh, Pittsburgh, PA), and
anti-tubulin Abs to mark the spindle (provided by Charles
Walsh, University of Pittsburgh). During the metaphase stage of
mitosis, chromosomes attach to both spindle poles by means of
kinetochore microtubules and align equidistantly between the
two poles to form the chromosomal metaphase plate (14). In the
OSCC cell lines, most of the metaphase spindles appeared
structurally and functionally normal (Fig. 2A, and see Fig. 3).
The most striking visible difference from the norm was the
appearance of multipolar spindles. In these aberrant structures,
the chromosomes appeared to attempt to achieve alignment
between the multiple poles, forming bizarre Y or cruciform
shapes (Fig. 2B).
To identify the position of the spindle poles unambiguously,
we immunolabeled them with Abs to NuMA (15). NuMA is a
nuclear matrix protein that relocates to the spindle poles during
mitosis. NuMA staining was seen at all spindle poles, but, in
some, the NuMA labeling was seen to have split into two parts,
even though only a single microtubule organizing center
(MTOC) was present (Fig. 2C, Inset shows the marked pole
without NuMA staining). This division of the NuMA immuno-
labeling is likely to be a precursor to full spindle pole separation
found in multipolar spindles. When we stained the cells with
anti-kinetochore Abs, we observed that the multiple spindle
poles did not necessarily have an equal influence on the chro-
mosomes migrating to the metaphase plate. In some cases, the
chromosomes were evenly divided between the poles (Fig. 2B),
but, in others, there was a marked asymmetry in metaphase
chromosomal alignment. In Fig. 2D, nearly normal metaphase
alignment is achieved, with the chromosomes equidistant be-
tween one major pole and two minor poles (arrows). In addition,
a small subset of the chromosomes is aligned between the two
minor poles. In Fig. 2E, two minor sites of microtubule organi-
pole draws one or two chromosomes away from the metaphase
alignment between the major poles. Thus, the different poles,
presumably depending on their microtubule-nucleation ability,
exerted different influences on the final alignment of the chro-
shows an example of a tripolar spindle from a tissue sample
obtained from a laryngeal carcinoma biopsy.
We also observed segregational errors associated with bipolar
spindles in the OSCC cells, including lagging metaphase and
anaphase chromosomes and anaphase bridges. The segrega-
tional defects were counted and are compared between cell lines
in Fig. 3. Anaphase chromatin forming continuous bridges con-
necting the two sets of separating chromosomes is shown in Fig. 4.
with the DNA dye, DAPI (blue). A normal metaphase from this culture is shown in A, and multipolar metaphase spindles in B–E. (C, Inset) The anti-tubulin image
alone of the pole marked with an arrowhead. (D and E) Arrows mark minor spindle poles. (Bars, 5 ?m.) (F) A tissue section from a formalin-fixed,
paraffin-embedded, laryngeal carcinoma was stained with hematoxylin and eosin. An atypical tripolar mitosis (arrow) and a normal bipolar mitotic figure
(arrowhead) are shown.
UPCI:SCC131 cells were immunolabeled with Abs to tubulin (yellow), and kinetochores (red in A, D, and E) or NuMA (purple in C) and counterstained
Saunders et al.
January 4, 2000 ?
vol. 97 ?
no. 1 ?
Anaphase bridges can be formed by dicentric chromosomes being
pulled simultaneously toward both spindle poles and are consistent
To demonstrate that the bridges contained dicentric chromosomes,
we immunostained with anti-kinetochore Abs. Anaphase bridges
usually contained centromeres, typically seen as two or more
well-separated fluorescent signals (Fig. 4A), although centromer-
eless bridges were also observed. Thus, both the presence of
multicentric chromosomes and the kinetochore staining of bridges
suggested that at least some of the bridges were caused by chro-
mosomes being pulled simultaneously toward both poles in an-
Anaphase bridging has been proposed to be a mechanism for
gene amplification (16), and the UPCI:SCC003, -131, and -172
cells lines express amplification of chromosomal band 11q13 in
the form of a homogeneously staining region (ref. 7, and X.H.,
W.S.S., M.S., C. M. Lese, and S.M.G., unpublished work). To
determine whether chromosome 11 could be found in the
anaphase bridges, we used chromosomal painting with chromo-
some 11 sequences as a FISH probe. Chromosome 11 was found
in 16% of anaphase bridges, suggesting that anaphase bridges
may be involved in amplification of chromosome 11 in these
OSCC cells (Fig. 4C).
In some cells, the anaphase bridge chromatin was excluded
as demonstrated by formation of the anti-tubulin-positive mid-
body region between the dividing daughter cells. At this stage,
the nuclear envelope should be reforming, but the cells have not
yet resolved their anaphase bridge, with centromere staining
persisting in the midbody region (arrows in Fig. 4B). At later
stages of division, it appears that these lagging chromosomes are
excluded from the reforming nucleus and incorporated into
micronuclei separate from the main genome and from each
other (Fig. 4D). This exclusion of bridged chromatin from the
cell nucleus most likely explains the high frequency of micro-
nuclei observed previously in oral cancer cells (17). When the
micronuclei were stained with Abs to kinetochores, both centric
(n ? 120) of the micronuclei contained kinetochores as judged
by immunostaining with CREST autoimmune serum. Appar-
ently, chromosome fragmentation occurred as the chromosomes
tubulin immunolabeling and DAPI staining and examined for five segrega-
tional abnormalities. The percentages of metaphase cells with multipolar
spindles, interphase cells with micronuclei, metaphase, and anaphase cells
with lagging chromosomes, and anaphase cells with chromatin bridges are
given, as well as the percentage of metaphase cells with normal spindles. The
mean and SEM from three coverslips are shown. A total for each cell line of
25–81 anaphase cells was counted for anaphase bridges and lagging chro-
mosomes, 86–164 metaphase cells for multipolar spindles and lagging chro-
mosomes, 1264–1593 interphase cells for micronuclei, and 110–370 meta-
phase cells for normal spindles.
UPCI lines SCC131, SCC172, and SCC003 were treated with anti-
and FISH with a chromosome 11 paint probe (green). (B) Arrows point to centromeres trapped in the forming midbody as these late telophase cells divide. (D)
Arrow points to the trapped lagging chromosome excluded from the reforming nucleus of the cell on the right. (E) Some micronuclei are immunonegative for
anti-centromere Abs. Arrow points to negative micronucleus, and arrowhead points to positive. Examples are from UPCI:SCC131.
Anaphase bridges containing centromeres and chromosome 11. Immunolabeling with Abs to tubulin (yellow), centromeres (red), and with DAPI (blue)
www.pnas.orgSaunders et al.
were pulled simultaneously in both directions, producing both
acentric fragments and centric chromosomes.
Chromosomal loss and rearrangement are known to be impor-
tant signs of genetic instability in cancer cells, but the mecha-
nisms behind these changes are unclear. We have begun an
investigation of the contribution of segregational errors to
chromosomal instability using oral carcinoma cells as our model
system. In these cultures, we found frequent variations in
chromosome numbers and structure between different cells
from the same tumor cell culture. We believe that many of these
abnormalities can be explained by chromosomal segregational
The OSCC cells clearly had difficulty achieving normal meta-
phase alignment and anaphase chromosome separation. One
simple model is that the kinetochores are defective for move-
ment, and that this single defect causes chromosomes to lag at
both metaphase and anaphase. Consistent with this model,
approximately equal numbers of lagging chromosomes were
observed in both metaphase and anaphase cells. One way to test
this directly would be a real-time analysis of chromosome
movement to determine whether the same chromosomes express
a lagging phenotype in both metaphase and anaphase. Alterna-
tively, lagging anaphase chromosomes, like bridged chromo-
somes, could also be the result of dicentric chromosomes, if one
centromere is only partially active, as originally suggested by
Fraccaro et al. (18).
One specific source of delay in anaphase progression was the
formation of anaphase chromatin bridges. Anaphase bridges are
rare in normal mammalian cells, but are common in virally
infected and tumor cells (19) or in patients with chromosome
breakage syndromes, such as Bloom syndrome (20). Anaphase
bridges can be formed as a result of defects in telomere structure
or length (21), defects in DNA replication (22), recombination
(23), or translocations that introduce a second centromere in the
Anaphase bridges were originally observed by Barbara Mc-
Clintock (10), who found that chromosomes in maize could
break, exposing a novel end, which after replication fused with
the sister chromatid to form a dicentric bridge in anaphase.
These bridges then can go through a cycle of breakage, fusion,
and bridging (BFB). An alternative mechanism involving mitotic
recombination and quadriradial formation has also been pro-
posed (20). BFB cycles have been proposed to be an important
mechanism of gene amplification in the acquisition of drug
resistance (25–27). Recent work has shown a role for fragile sites
in BFB-induced gene amplification (16). Telomeric fusions and
BFB cycles have also been reported in other cancer cells and are
thought to lead to specific types of chromosomal rearrange-
ments, including deletions and ring chromosomes (28–30). We
believe that many of the chromosomal differences we are seeing
between OSCC cells in culture are because of BFB cycles. BFB
cycles could explain the high variation between cells within a
culture, because dicentric chromosomes are inherently unstable
and may change in virtually every generation (31). As an
example, BFB cycles most likely produce the amplification of
chromosome 11q13 with concomitant deletion of distal se-
quences (8). This is in contrast to recombinational models for
gene amplification (23), which would not produce the chromatin
bridges we have observed in OSCC cells unless the recombina-
tional intermediates persisted into anaphase.
Micronuclei are commonly observed in oral cancer cells and
are an important biomarker for this disease (17, 32). Micronuclei
could be formed either from lagging anaphase chromosomes or
anaphase bridges, both of which occur in our OSCC cultures. In
human lymphocytes from healthy volunteers, micronuclei are
rare and nearly all contain intact chromosomes, and were
interpreted to result from lagging anaphase chromosomes (33).
In contrast, more than two-thirds of the micronuclei from the
OSCC cells studied here were negative for centromere immu-
nostaining, suggesting that many of the micronuclei in oral
cancer cells result from chromosome-fragmentation events, such
as could occur with anaphase bridging. The appearance of
acentric micronuclei is consistent with the expected DNA dam-
age resulting from BFB cycles. Pharmacological agents that lead
to DNA damage usually result in acentric micronuclei, whereas
agents that produce aneuploidy, by interference with microtu-
bule dynamics, usually yield centromere-positive micronuclei
(33, 34). These observations, combined with our data showing
dicentric chromosomes being excluded from the reforming
result from anaphase bridges.
The frequency of 20–30% micronuclei in our OSCC samples
is much higher than the 0.21–0.7% of exfoliated oral epithelial
cells from healthy volunteers (35). However, the frequency of
30% centric micronuclei in this study is roughly similar to the
50% centromere-positive micronuclei observed in these cells
(35). This comparison suggests that the defect in OSCC cells that
results in micronucleus formation, compared with normal oral
rather the frequency with which they form. In patients with
Bloom syndrome, one centromere of dicentric chromosomes is
usually inactivated (20). We propose that this mechanism for
centromere inactivation may be lost in our OSCC cells, contrib-
uting to the high levels of BFB events and the resulting genomic
instability of these cultures.
Multipolar spindles are frequently observed in cancer cells
(19), but the origin is unclear. One model proposes that multi-
polar spindles are the result of excess centrioles, for example,
from centriolar replication in the absence of division. Neoplastic
progression in transgenic mice with an elastase-SV40 large T
antigen gene was associated with increased numbers of centri-
oles, up to five at a single pole (36). However, whereas pericen-
triolar material is required for microtubule nucleation at spindle
poles (37), centrioles are not. When multipolar spindles were
induced by Colcemid treatment in Chinese hamster ovary cells,
60% of the MTOCs of tripolar cells lacked any centriole (38). It
seems clear that multipolar spindles can be the result of sepa-
ration of the pericentriolar material without centriolar duplica-
tion. Our results indicate that the NuMA protein is able to form
duplicated foci prior to spindle pole body separation. At the
spindle pole, NuMA binds to the microtubule motor dynein, and
the dynein-activating complex dynactin (39, 40) and can poly-
merize and has been proposed to tether and stabilize microtu-
bules at the spindle poles (39, 41, 42). We suggest that accumu-
lation of otherwise normally functioning NuMA at the spindle
poles of these cancer cells may recruit enough microtubule
crosslinking activity, for example, the dynein motor, to induce
in the absence of centrosomes if dynein is added (43). The
splitting of the poles can produce major or minor MTOCs, as
defined by the number of microtubules and the extent of
influence on chromosome alignment. This variation in activity
could simply be the result of the total amount of pericentriolar
material that each pole receives when they separate or whether
specific unidentified factors or modifications are present. We
favor the interpretation that the amount of pericentriolar ma-
terial that each nascent pole receives from the original deter-
mines its relative activity, but additional studies are required to
address this issue.
Our findings provide strong support for the model that a
substantial proportion of the genetic instability in cancer cells is
because of specific structural defects in chromosome segrega-
tion. These include BFB cycles leading to gene amplification,
Saunders et al.
January 4, 2000 ?
vol. 97 ?
no. 1 ?
chromosome fragmentation and loss, and segregational errors Download full-text
from separation of microtubule-organizing material. In fact, it
may be explained by the observed segregational defects. The
formation of supernumerary spindle poles appears to be pre-
ceded by the separation of the centrosomal protein NuMA into
distinct foci. The extra poles could have only a minor influence
on the metaphase alignment, pulling one or a few chromosomes
from the main body of chromosomes or major effects, splitting
the chromosomes equally in four or more directions simulta-
neously. It will be of interest to determine whether the relative
amounts of NuMA that partition to these minor poles determine
their ability to capture chromosomes and perturb the bipolar
The cell lines were initiated and cared for initially by Dr. Christa M.
Lese, Ms. Jaya Reddy, or Ms. Seia Comsa, the latter two under the
guidance of Ms. Robin Wagner. We thank Dr. Chandramohan S.
Ishwad and Ms. Diana Kerestan for their high quality DNA sequence
analyses and human papilloma virus testing; Drs. Charles Walsh,
Duane Compton, Carol Feghali, and Thomas Medsger for Abs; and
Dr. Thomas Ried of the National Cancer Institute and Mr. Dirk
Soenksen, Ms. Cathy Jaenish, and Ms. Elizabeth Weeks of Applied
Spectral Imaging for training and assistance with the spectral karyo-
typing. These studies were supported by an Oral Cancer Prevention,
Research, and Treatment Center Planning Grant (P20DE12378) (to
Dr. Eugene Myers), Grants CB-171 and JFRA-624 from the American
Cancer Society (to W.S.), National Institutes of Health Grants
F32DE05707 (to M.S.), R01DE10513, R01DE12008, P30CA47904-
11S39019, and S10RR11879 (to S.M.G.), Smokeless Tobacco Re-
search Council Grant 0660 (to S.M.G.), and the Project Grant
1P60DE13059-01 of the Oral Cancer Center at the University of
Pittsburgh (to W.S. and S.M.G.). Cytogenetic analyses were carried
out in the University of Pittsburgh Cancer Institute Cytogenetics
1. Lengauer, C., Kinzler, K. W. & Vogelstein, B. (1998) Nature (London) 396,
2. Miyagawa, K. (1998) Int. J. Hematol. 67, 3–14.
3. Elledge, S. J. (1996) Science 274, 1664–1672.
4. Winey, M. (1996) Curr. Biol. 6, 962–964.
5. Cowan, J. M. (1992) Otolaryngol. Clin. North Am. 25, 1073–1087.
6. Jin, Y., Jin, C., Wennerberg, J., Mertens, F. & Hoglund, M. (1998) Cancer Res.
7. Lese, C. M., Rossie, K. M., Appel, B. N., Reddy, J. K., Johnson, J. T., Myers,
E. N. & Gollin, S. M. (1995) Genes Chromosomes Cancer 12, 288–295.
8. Jin, Y., Hoglund, M., Jin, C., Martins, C., Wennerberg, J., Akervall, J.,
Mandahl, N., Mitelman, F. & Mertens, F. (1998) Genes Chromosomes Cancer
9. Akervall, J. A., Michalides, R. J., Mineta, H., Balm, A., Borg, A., Dictor, M. R.,
Jin, Y., Loftus, B., Mertens, F. & Wennerberg, J. P. (1997) Cancer 79, 380–389.
10. McClintock, B. (1941) Cold Spring Harbor Symp. Quant. Biol. 9, 72–81.
11. Heo, D. S., Snyderman, C., Gollin, S. M., Pan, S., Walker, E., Deka, R., Barnes,
E. L., Johnson, J. T., Herberman, R. B. & Whiteside, T. L. (1989) Cancer Res.
12. Sumner, A. T. (1972) Exp. Cell Res. 75, 304–306.
13. Veldman, T., Vignon, C., Schrock, E., Rowley, J. D. & Ried, T. (1997) Nat.
Genet. 15, 406–410.
14. Nicklas, R. B. (1997) Science 275, 632–637.
15. Compton, D. A., Szilak, I. & Cleveland, D. W. (1992) J. Cell Biol. 116,
16. Coquelle, A., Pipiras, E., Toledo, F., Buttin, G. & Debatisse, M. (1997) Cell 89,
J. W., Toth, B. B. & Hong, W. K. (1994) Int. J. Cancer 59, 457–459.
18. Fraccaro, M., Lo Curto, F. & Vogel, W. (1978) Prog. Clin. Biol. Res. 26,
19. Sandberg, A. A. (1990) The Chromosomes in Human Cancer and Leukemia
(Elsevier, New York), pp. 96–99.
20. Therman, E., Trunca, C., Kuhn, E. M. & Sarto, G. E. (1986) Hum. Genet. 72,
21. Harley, C. B. & Villeponteau, B. (1995) Curr. Opin. Genet. Dev. 5, 249–255.
22. Poupon, M. F., Smith, K. A., Chernova, O. B., Gilbert, C. & Stark, G. R. (1996)
Mol. Biol. Cell 7, 345–354.
23. Smith, K. A., Gorman, P. A., Stark, M. B., Groves, R. P. & Stark, G. R. (1990)
Cell 63, 1219–1227.
24. Hastie, N. D. & Allshire, R. C. (1989) Trends Genet. 5, 326–331.
25. Smith, K. A., Stark, M. B., Gorman, P. A. & Stark, G. R. (1992) Proc. Natl.
Acad. Sci. USA 89, 5427–5431.
26. Ma, C., Martin, S., Trask, B. & Hamlin, J. L. (1993) Genes Dev. 7, 605–620.
27. Ishizaka, Y., Chernov, M. V., Burns, C. M. & Stark, G. R. (1995) Proc. Natl.
Acad. Sci. USA 92, 3224–3228.
28. Mondello, C., Casati, A., Riboni, R. & Nuzzo, F. (1995) Cancer Genet.
Cytogenet. 79, 41–48.
29. Sawyer, J. R., Sammartino, G., Husain, M., Lewis, J. M., Anderson, B. & Boop,
F. A. (1993) Genes Chromosomes Cancer 8, 69–73.
30. Sawyer, J. R., Thomas, E. L., Roloson, G. J., Chadduck, W. M. & Boop, F. A.
(1992) Cancer Genet. Cytogenet. 60, 152–157.
31. Riboni, R., Casati, A., Nardo, T., Zaccaro, E., Ferretti, L., Nuzzo, F. &
Mondello, C. (1997) Cancer Genet. Cytogenet. 95, 130–136.
32. Stich, H. F., Parida, B. B. & Brunnemann, K. D. (1992) Int. J. Cancer 50,
33. Ford, J. H., Schultz, C. J. & Correll, A. T. (1988) Am. J. Hum. Genet. 43,
34. Degrassi, F. & Tanzarella, C. (1988) Mutat. Res. 203, 339–345.
35. Moore, L. E., Titenko-Holland, N., Quintana, P. J. & Smith, M. T. (1993) J.
Toxicol. Environ. Health 40, 349–357.
36. Levine, D. S., Sanchez, C. A., Rabinovitch, P. S. & Reid, B. J. (1991) Proc. Natl.
Acad. Sci. USA 88, 6427–6431.
37. Gould, R. R. & Borisy, G. G. (1977) J. Cell Biol. 73, 601–615.
38. Keryer, G., Ris, H. & Borisy, G. G. (1984) J. Cell Biol. 98, 2222–2229.
39. Merdes, A., Ramyar, K., Vechio, J. D. & Cleveland, D. W. (1996) Cell 87,
40. Mattagajasingh, S. N., Huang, S. C., Hartenstein, J. S., Snyder, M., Marchesi,
V. T. & Benz, E. J. (1999) J. Cell Biol. 145, 29–43.
41. Yang, C. H. & Snyder, M. (1992) Mol. Biol. Cell 3, 1259–1267.
42. Dionne, M. A., Howard, L. & Compton, D. A. (1999) Cell Motil. Cytoskeleton
43. Verde, F., Berrez, J. M., Antony, C. & Karsenti, E. (1991) J. Cell Biol. 112,
www.pnas.orgSaunders et al.