Download Chromosomal Changes without DNA

[CANCER RESEARCH 46, 4607-4612, September 1986]
Chromosomal Changes without DNA Overproduction in Hydroxyurea-treated
Mammalian Cells: Implications for Gene Amplification1
Peter Hahn,2 Leon N. Kapp, William F. Morgan, and Robert B. Painter
Laboratory of Radiohiology and Environmental Health [P. H., L. N. K., W. F. M., R. B. P.] and the Department of Microbiology [R. B. P.], University of California,
San Francisco, California 94143
ABSTRACT
It has been reported that a 6-h incubation of early S-phase Chinese
hamster cells with hydroxyurea promotes DNA overproduction, i.e.,
replication of DNA a second time within a single cell cycle, and that this
could be the basis for gene amplification in drug-treated mammalian
cells. When we incubated methotrexate-resistant Chinese hamster cells
that were approximately 2 h into the S phase with hydroxyurea for 6 h,
DNA that had been replicated before the incubation with hydroxyurea
(early S-phase DNA) was replicated again within 11 h after the hydrox
yurea treatment. However, incubation with colchicine or Colcemid after
hydroxyurea treatment virtually abolished this overreplication, as well
as that of the amplified dihydrofolate reducÃ-asegenes in these cells,
indicating that the second replication had occurred in a second cell cycle.
Cells collected in the first mitosis after incubation with hydroxyurea
never contained overreplicated DNA but did contain abundant chromo
some aberrations. Early S-phase DNA replicated again on schedule
during the first few hours after mitosis. Asymmetric segregation of
chromosome fragments or unequal sister chromatid exchange may be the
actual basis for gene amplification in drug-treated mammalian cells.
INTRODUCTION
Gene amplification is a process whereby the number of copies
per cell of a given gene increases, and consequently the amount
of that gene's product also increases (1). In mammalian cells,
the most frequently studied system is the development of re
sistance to methotrexate by amplification of the DHFR1 gene
(2, 3). This process has been studied primarily in cultured cells;
amplification events in mammalian cells have been observed in
nature (4), but thus far only in aneuploid cells. Gene amplifi
cation in Drosophila (5) involves developmentally controlled
processes that cause overreplication of small portions of the
genome by iterative initiation at single DNA replication origins.
Because intermediates are envisaged in these processes, in
which multiple copies of both DNA strands lie within the same
structure, the term "onion-skin model" has been coined to
describe this method of gene amplification (6, 7). In applying
this model to gene amplification in drug-treated mammalian
cells, a seminal idea is that the two DNA growing points of a
replicón diverging from its origin are stalled, i.e., completely
blocked, by an inhibitor of DNA synthesis. The inhibitor may
be either the selective agent itself, e.g., methotrexate, or a
second DNA synthesis inhibitor, e.g., hydroxyurea. This stalling
is thought to lead to derangement of a DNA replication control
system, so that abnormal reinitiation at the origin(s) occurs,
leading to a 2-fold or more amplification of that region of the
genome.
Received 11/20/85: revised 3/17/86, 6/10/86; accepted 6/10/86.
The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1This work was supported by the United States Department of Energy (Con
tract DE-AC03-76-SF01012) and by NIH National Research Service Award 5
T32 ES07106 from the National Institute of Environmental Health Sciences.
2To whom requests for reprints should be addressed, at Laboratory of Radiobiology and Environmental Health. LR 102. University of California, San Fran
cisco. CA 94143.
'The abbreviations used are: DHFR. dihydrofolate reducÃ-ase:CHO. Chinese
hamster ovary; BrdUrd. bromodeoxyuridine; SCE, sister chromatid exchange;
SSC, 0.15 M sodium chloride-0.015 M sodium citrate.
Experimental approaches to elucidate the mechanism of gene
amplification in mammalian cells have generally fallen into two
categories: (a) structural analysis of the DNA or chromosomes
of cells that have been selected (8-11) or (b) search for agents
that will increase the spontaneous levels of gene amplification
and analysis of the mode of action of these agents to arrive at
a conclusion about the underlying process (12-16). Mariani
and Schimke (16) presented evidence that treatment of Chinese
hamster cells during early S phase with hydroxyurea, a potent
inhibitor of ribonucleotide reducÃ-ase, induces all DNA repli
cated before the addition of hydroxyurea to be replicated again
within that same cell cycle. This would result in a 2-fold
amplification of that DNA replicated before the block.
A problem with the concept of stalled growing points is that
inhibitors of DNA synthesis such as hydroxyurea do not block
DNA growing points; they only slow them in a dose-dependent
manner (17, 18). Therefore, we sought another explanation for
gene amplification in mammalian cells. We repeated and ex
tended the experiments of Mariani and Schimke (16), and our
results led not only to an alternative explanation for their data,
but also to alternative explanations for the mechanism of gene
amplification in drug-treated mammalian cells.
MATERIALS AND METHODS
Cell Culture. CHO-B11 cells (from A. Hill, Stanford University) are
the same cells used by Mariani and Schimke (16) and are resistant to
high concentrations of methotrexate because they contain an amplified
array of approximately 50 copies of stably integrated DHFR genes.
These cells were maintained as monolayers in Ham's F-12 medium
lacking glycine, hypoxanthine, and thymidine and supplemented with
10% dialyzed fetal bovine serum, penicillin (50 units/ml), streptomycin
(50 Mg/ml), and 0.5 MMmethotrexate.
Cell Synchrony. Mitotic cell populations were selected by means of
an automated synchrony apparatus (Talandic Research Corp., Pasa
dena, CA) (19). CHO-B11 ceils were plated into 750-cm2 roller bottles
and maintained at 1 rpm. For the experiment described in Fig. 1, the
synchronized cells were selected in Ham's medium lacking glycine,
hypoxanthine, and thymidine. For the experiments described in Fig. 3,
to ensure uniform, nonconfluent late log phase monolayers that would
yield a maximum number of mitotic cells, we subcultured the cells in
each roller bottle 2:1 into new roller bottles containing complete Ham's
F-12 supplemented with whole serum 24 h before the beginning of
synchrony.
Before beginning the experiments with mitotically selected cells, we
rotated the monolayers in the roller bottles at 75 rpm for 1 min each
hour to remove loosely attached cells and debris and then refed the
cells. This was repeated hourly for 24 h. Under these conditions, it was
possible to obtain 4 to 5 x 10* mitotic cells in a single selection with a
mitotic index of >95%. Therefore, a single selection was sufficient to
perform each of the experiments described in Figs. 3 and 4. The cell
cycle was determined in each experiment with each population of BI1
cells. The length of the cycle, d and S phase, was observed each time
that the cells were used for hydroxyurea-BrdUrd experiments. To
measure S phase, we pulse labeled synchronized cells for 30 min with
['H]thymidine (2 ^Ci/ini; specific activity, 60 Ci/mmol). Cells were
then washed with SSC and trypsinized. One-half of the cells were
counted in a Coulter Counter, and the remainder of the sample was
4607
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1986 American Association for Cancer Research.
HYDROXYUREA
AND GENE AMPLIFICATION
used to determine 'II cpm incorporation in the acid-precipitable cellular
fraction (19).
For experiments, cells were treated essentially as described by Mariani and Schimke (16). Starting immediately after selection, the cells
were incubated in medium containing BrdUrd (10~5 M) and [3H]thy02
x
midine (3 jiCi/ml; 50 Ci/mmol) for 8 h. At the end of this time the
medium was removed, and medium containing BrdUrd (10~5 M) and
Q.
O
hydroxyurea (0.3 HIM)was added. After 6 h this medium was removed,
the cells were washed with fresh medium, and medium containing
BrdUrd (10 M) was added. At this point, variations in the protocol
were introduced. The details of each protocol are given in "Results."
Cesium Chloride Equilibrium Density Centrifugation. Cellular DNA
was isolated as previously described (20). Cell monolayers were washed
with SSC and lysed with 0.1 % sodium dodecyl sulfate. The lysate was
incubated for 1 h with RNase (100 ¿ig/ml,37°C)and for l h with
Pronase (100 ng/m\, 37°C).The preparation was extracted with chloroform:isoamyl alcohol (24:1 ) 2 or 3 times. The aqueous layer was then
removed and dialyzed against SSC. After dialysis, the DNA samples
were sheared with a Virtis homogenizer. Exactly 4.5 ml of sample were
added to 5.9 g of CsCl and centrifuged at 37,000 rpm for 48 h or more
in a Beckman TI-50 rotor. After centrifugation, a hole was made in the
bottom of each tube, samples were collected, and the DNA in the
solution was precipitated with 4% perchloric acid and passed through
glass fiber filters. The 3H on the filters was measured by liquid scintil
1
10
15
20
Fraction Number
Fig. 1. CsCl density gradient profiles showing the effect of colchicine on
overreplication of early S-phase DNA in CHO-BI1 cells. Two h after cells had
entered S phase, cells were treated with 0.3 mM hydroxyurea for 6 h and allowed
to progress through the cell cycle in the presence (A) or absence (•)of colchicine
for 11 h. I, position of heavy-heavy DNA.
lation spectrometry.
Dot Blot Analysis. For detection of DHFR genes, fractions from
CsCl gradients containing the heavy-light or heavy-heavy DNA were
pooled, dialyzed, ethanol precipitated, and subjected to the dot blot
procedure of Kafatos et al. (21). The probe was a "P-Iabeled mouse
complementary DNA (22).
Cytogenetic Analysis. Synchronized CHO-B11 cells (mitotic index,
>95%) were treated as described for the CsCl centrifugation experi
ment, except that no radioisotope was added to the cultures. After the
6-h incubation with hydroxyurea (0.3 or 1.0 mM), fresh medium con
taining BrdUrd (10 ' M) was added to the cultures, and 2 h later
Colcemid (2 x 10~7 M) was added. Separate cultures were maintained
in medium containing BrdUrd for 12 h before addition of Colcemid.
Mitotic cells were collected at 3-h intervals after Colcemid treatment.
Metaphase preparations were processed for cytogenetic analysis by
routine methods and stained by the fluorescence-plus-Giemsa technique
(23). This procedure permitted us to determine cytogenetically if any
parts of the chromosomes had replicated more than once before arriving
at metaphase. The detailed results of this analysis have been reported
by Morgan et al. (24). In some experiments, the cells were allowed to
go through another cell cycle in the presence of BrdUrd before collec
tion of mitotic cells, so that we could measure the frequency of SCE.
RESULTS
When mitotically synchronized CHO-B11 cells, the cell line
used by Mariani and Schimke, were incubated with BrdUrd and
[3H]thymidine for 8 h, BrdUrd and hydroxyurea for 6 h, and
BrdUrd for 11 h, the results were almost identical to those
reported by Mariani and Schimke (Ref. 16; Fig. 1). More than
half of the DNA that had replicated in the presence of [3H]BrdUrd during the first 8 h after mitotic collection (early Sphase DNA) was found in the "heavy-heavy" region of the CsCl
gradient, corresponding to the density of DNA labeled in both
strands with BrdUrd. Therefore, the early S-phase DNA had
entered another round of replication. However, when parallel
cultured cells were simultaneously treated in exactly the same
way except for the addition of 2 x IO"5 M colchicine to the
medium for 11 h after removal of hydroxyurea, only a very
small amount of the early S-phase DNA had completed a
second round of replication (Fig. 1). It should be noted that,
although colchicine and Colcemid inhibit Chinese hamster cell
division, they delay entry into a second cell cycle by only 2 to
3 h (25). For this reason, these spindle inhibitors are often used
a
8
12
16
Hours
Fig. 2. Cell cycle kinetics of CHO-BI 1 cells grown in complete medium. Cells
were harvested as described in "Materials and Methods."
to induce tetraploidy (26). The small amount of heavy-heavy
DNA visible in the colchicine-treated culture is very likely due
to reentry into S phase of the second cell cycle after a failed
attempt to undergo cell division. The experiment described in
Fig. 3 was designed to address this point further.
The timing of the appearance of the heavy-heavy DNA varied
from one experiment to another; occasionally there was no
heavy-heavy DNA in cells incubated with or without colchicine
at the 11-h posthydroxyurea time point. We attribute this
variability to differences in cell cycle time brought about by the
restrictive nature of the selective medium (Ham's F-12 lacking
glycine, thymidine, and hypoxanthine and supplemented with
dialyzed fetal calf serum plus 0.5 /J.Mmethotrexate). The growth
rate of cells in this medium is dependent on feeding schedules
and cell density. Therefore, for the next experiment we used
synchronized cells grown in complete F-12 medium supple
mented with whole fetal calf serum, and we checked the cell
cycle (Fig. 2) immediately before the experiment (16-h cycle
time, with S phase commencing at 4 to 6 h). The synchronized
cells collected at the zero time point of the cell cycle determi
nation in Fig. 2 were used for the following experiment.
The cells were plated into one large flask and incubated with
BrdUrd and [3H]thymidine for 8 h and then with BrdUrd and
hydroxyurea for 6 h. After the hydroxyurea treatment, the cells
were incubated with medium containing both BrdUrd and Col
cemid (a more reversible microtubule inhibitor than colchicine),
so that every 2 h we could shake off the cells that had accu
mulated in mitosis. Again, Colcemid inhibited cell division and
allowed the cells to accumulate in mitosis for a few hours, but
it only postponed, rather than blocked, entry into the following
4608
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1986 American Association for Cancer Research.
HYDROXYUREA
AND GENE AMPLIFICATION
cell cycle, even in the continued presence of Colcemid. Half of
the cells collected at each shake-off were lysed immediately; the
other half were transferred to another flask to continue incu
bation for a total of 20 h after hydroxyurea treatment before
being lysed. In this way, we were able to distinguish between
events occurring before and after the first mitosis after hydrox
yurea treatment. A diagram and explanation of the protocol for
this experiment are given in Fig. 3.
The results of CsCl equilibrium density gradient analysis of
the various cell samples showed the following, (a) None of the
mitotic cells that accumulated with Colcemid contained meas
urable quantities of overreplicated (heavy-heavy) DNA at the
time that they were collected (4, 6, 8, or 10 h after removal of
hydroxyurea) (Fig. 4). (b) All cell populations that had been
collected in mitosis and reincubated with Colcemid in a second
flask (for a total of 20 h after removal of hydroxyurea) contained
replicated DNA, the percentage of which increased with time
of incubation after collection (Table 1). The apparent linear
increase in heavy-heavy DNA with time probably reflects partial
decay of synchrony induced by the hydroxyurea treatment.
Because none of the cells contained -'H in heavy-heavy DNA
at the time of mitotic harvest, no early S-phase DNA had
Incubated
¡nSecond
Colcemid
Lysed
Flask
BrdUrd
immediately
16
20
24
28
32
12
Total Hours in Culture
Fig. 3. Protocol for experiment to determine the time of a second replication
of DNA in cells that were about 2 h into the S phase before a 6-h incubation with
hydroxyurea (//{/). At 0 h, one T-150 flask was inoculated with synchronized
cells, which were incubated for 8 h in medium containing 3 n<'i of [3H]thymidine
(3HdThd) per ml (50 Ci/mmol) and 1 x 10"' M BrdUrd and were then transferred
to medium containing hydroxyurea and BrdUrd. After 6-h incubation with
hydroxyurea, the medium was changed to one containing 2 x IO M Colcemid
and BrdUrd. At 4.6,8. and 10 h after removal from hydroxyurea, the accumulated
mitotic cells were shaken off the monolayer. At each time point, one-half of the
medium and cells were transferred into a T-75 flask, in which the cells were
incubated until the end of the experiment. The cells in the other half of the
medium were resuspended in saline, lysed by adding sodium dodecyl sulfate (final
concentration, 0.1%), and frozen until the other incubations were completed.
After each shake-off, fresh medium containing Colcemid and BrdUrd was added
to the T-1 SO flask, and incubation continued until the next collection time. All
samples were subsequently analyzed simultaneously by equilibrium density gra
dient centrifugation.
Table I Replication of early S-phase DNA after mitosis
Time in second
flask (h)
DNA replicated
0
0
46
10
12
61
77
14
91
16
" Percentage of 3H-labeled early S-phase DNA found in the heavy-heavy
fraction of equilibrium density gradients after mitotic shake-off, transfer to a
separate flask, and continued incubation (sec Fig. 3).
replicated a second time in the first cell cycle. The highest yield
of mitotic cells was obtained at the collection made 6 h after
removal of hydroxyurea, but there were already many cells in
the mitotic harvest at 4 h. This shows that the most frequent
time for cells to progress from the end of hydroxyurea treatment
to mitosis was only 4 to 6 h, and many cells took less than 4 h.
To show that the presence of Colcemid did not interfere with
the ability of the cells to replicate their DNA a second time
within a single cell cycle, we performed a similar experiment
without Colcemid. Cells from four roller bottles were used for
a single mitotic selection, and all the detached mitotic cells
were plated into a single roller bottle. As in the other experi
ments, [-"Hjthymidine and BrdUrd were present for the first 8
h, followed by 0.3 mM hydroxyurea and BrdUrd for 6 h,
followed by BrdUrd only (no Colcemid). After hydroxyurea was
removed from the medium, the roller bottle was rotated at 75
rpm for 1 min every h, and the detached mitotic cells were
aspirated from the bottle. Medium was replaced with fresh
medium, and the process was repeated at hourly intervals. At
alternate hours, the aspirated mitotic cells were either lysed
immediately and prepared for CsCl gradient centrifugation or
plated into separate flasks and incubated for a total of 12 h
after hydroxyurea removal, then lysed, and prepared for CsCl
gradient analysis. Fewer cells were collected at each time point
than in experiments in which Colcemid accumulated cells in
mitosis. Nevertheless, heavy-heavy DNA was never observed in
the mitotic cells (Fig. 5A) but was observed in the cells that
were allowed to continue past mitosis (Fig. SB), indicating that,
even in the absence of Colcemid, the second replication of DNA
occurred only after mitosis in a second cell cycle.
In a similar experiment, cells were detached by mitotic shakeoff in the absence of Colcemid at 4, 6, 8, 10, 12, 14, and 16 h
after the removal of hydroxyurea and lysed immediately. The
samples were pooled (to increase the sensitivity of the experi
ment) and analyzed by CsCl density gradient analysis (Fig. 6/4),
as was the DNA from the cells remaining on the monolayer
(Fig. 6Ä).(The cells on the monolayer are cells that had escaped
mitotic harvest, some of which had entered S phase of the next
cell cycle.) Heavy-heavy DNA was not observed in the cells
harvested at mitosis, but was observed in the cells remaining
on the monolayer.
We also measured the effects of the 6-h incubation with
hydroxyurea on DNA synthesis; the incorporation of ['H]thymidine into DNA was reduced to 20% of control rates by this
treatment. Therefore, the cells had progressed for about an
hour's equivalent of S phase during the 6-h incubation with
8
10
12
14
Fraction Number
20
Fig. 4. CsCl density gradient profiles of |'H]DNA from CHO-BI1 cells
accumulated in mitosis by Colcemid, harvested 10 h after hydroxyurea treatment,
and lysed immediately. (See legend to Fig. 3.) Profiles from other time points
were the same. HH. heavy-heavy: HL, heavy-light.
hydroxyurea and were not completely blocked, as reported by
Mariani and Schimke (16). Our results are consistent with
other reports on inhibition of DNA synthesis by hydroxyurea
(17, 18).
Even though the major part of early S-phase DNA was not
replicated twice in the same S phase after the hydroxyurea
block, it was possible that the DHFR gene itself had been
overreplicated. To determine if this was so, the heavy-light
4609
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1986 American Association for Cancer Research.
HYDROXYUREA
AND GENE AMPLIFICATION
Colchicine
HL
HH
• HH (i)
Fraction
Number
Fig. 5. CsCI density gradient profiles of [3H]DNA from CHO-BI1 cells
incubated with [3H]thymidine and BrdUrd for 8 h, then with hydroxyurea for 6
h. and then with BrdUrd only (no Colcemid) until shake-off in mitosis. A, DNA
from cells lysed immediately after collection in mitosis at 12 h after hydroxyurea;
B. DNA from mitotic cells transferred to a second flask at 7 h after hydroxyurea
and incubated for 5 h in medium with BrdUrd. HH, heavy-heavy; HL, heavylight.
30 - A
Fig. 7. Dot blot analysis of "P-labeled DHFR-hybridizable sequences in
heavy-heavy (HH) and heavy-light (HL) DNA from CHO-B11 cells incubated
with or without colchicine. Cells were treated as described in Fig. 1 but yielded a
CsCI density gradient profile indicating that only about one-fifth of the DNA in
cells not incubated with colchicine had rereplicated. The 3!P in the heavy-heavy
DNA spots was autoradiographed with [////(<)] or without (HH) an intensifying
screen. The spots represent one-fourth of the sample and were excised after
autoradiography and counted in a liquid scintillation spectrometer for the pres
ence of 3H used in the initial labeling. There were 78 cpm in heavy-heavy DNA
from cells treated with colchicine and 319 cpm in heavy-heavy DNA from cells
treated without colchicine. There were 10 cpm in a random, unlabeled nitrocel
lulose fragment.
Table 2 Effect of hydroxyurea, metholrexale, and hydroxyurea plus melholrexate
on survival ofCHO-Bll cells
25
20
TreatmentMTX"
(0.5 MM)*
±3C
HU (0.3 mM) + MTX (0.5 MM)
279 ±5
MTX only (50 UM)
136 ±20
HU (0.3 mM) + MTX (50 ^M)Colonies/plate397
59 ±4Survival
°MTX, methotrexate; HU. hydroxyurea.
4 CHO-B11 cells are resistant to 0.5 \M methotrexate.
' Mean ±range.
15
10
Q.
O
~m
5
5
(%
of control)100
70
34
IS
20
15
10
10
15
Fraction Number
Fig. 6. CsCI density gradient profiles of [3H]DNA from CHO-B11 cells
incubated with pHjthymidine and BrdUrd for 8 h. then with hydroxyurea and
BrdUrd for 6 h, and then with BrdUrd only (no Colcemid). A, profile of DNA
from mitotic shake-offs at 4, 6, 8, 10, 12, 14, and 16 h after removal of
hydroxyurea. Cells were lysed immediately after shake-off, and cell samples were
pooled for analysis. B, profile of DNA from the cells remaining on the monolayer.
HH, heavy-heavy; HL, heavy-light.
(representing DNA labeled in only one strand with BrdUrd)
and heavy-heavy regions of density gradients from cells incu
bated with or without colchicine (as in Fig. 1) were probed with
"P-labeled mouse DHFR complementary DNA (Fig. 7). In this
experiment, only one-fifth of the 3H was found in the heavyheavy region of density gradients from cells incubated without
colchicine. The results of the hybridization showed a very strong
*-P signal for DHFR sequences in the heavy-light regions of
detectable signal for DHFR sequences was found in the heavyheavy DNA of colchicine-treated cells. Had the DHFR gene in
these cells replicated to the same extent as the total DNA of
cells incubated without colchicine, there would have been the
same number of copies of the DHFR gene in the heavy-heavy
region of the gradients whether or not the cells had been
incubated with colchicine.
We were unable to repeat the abrupt gene amplification
results that Mariani and Schimke (16; their Fig. 3) reported for
early S-phase cells that had been incubated with hydroxyurea
for 6 h. Their results indicated that this treatment, followed by
incubation in 100-fold excess methotrexate, virtually eliminated
the cytotoxic effects of both drugs. We incubated synchronized
cells with hydroxyurea for 6 h (beginning 2 h into S phase) and
then transferred them into medium containing 50 //M metho
trexate or, as one control, 0.5 /¿Mmethotrexate. (CHO-B11
cells are resistant to 0.5 pM methotrexate.) For additional
controls, synchronized cells were plated directly into 0.5 //M or
50 fiM methotrexate alone. The results show that both hydrox
yurea plus 0.5 Õ/M
methotrexate and 50 MMmethotrexate alone
killed cells, and that the effects of the two treatments were
additive (Table 2), not mutually protective, as reported by
Mariani and Schimke (16). We cannot explain the difference
between our results and those of Mariani and Schimke (16).
The predominant cytogenetic effects of the 6-h incubation
gradients of DNA from all cells, whether or not they had been
incubated with colchicine, and a less strong signal for DHFR
sequences in the heavy-heavy region of the gradients of DNA
from cells not incubated with colchicine. However, only a barely
4610
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1986 American Association for Cancer Research.
HYDROXYUREA
AND GENE AMPLIFICATION
with hydroxyurea on S-phase cells were chromosome aberra
tions, increased polyploidy, and increased SCEs (Table 3; Fig.
8); no radioisotope was used in these experiments. We followed
the fate of aberrations into the second generation and found
that they were still elevated over controls but were reduced
compared to the first generation.
DISCUSSION
Our equilibrium density gradient results without colchicine
or Colcemid treatment are essentially identical to those of
Mariani and Schimke (16). However, the addition of colchicine
or Colcemid delayed the appearance of a second replication of
early S-phase DNA. The experiments with removal of mitotic
cells show that early S-phase DNA was replicated a second
time only after the cells had completed mitosis and entered the
next S phase, and not, as Mariani and Schimke proposed,
during a single cell cycle as aberrant rereplication. Our deter
mination of the frequency of DHFR sequences in the heavyheavy region of density gradients shows that the DHFR gene
was also not replicated twice within a single cell cycle in
hydroxyurea-treated cells. These observations are consistent
with reports from other studies in which overreplication of
DNA within a single cell cycle was sought but not detected
(27-29).
The possibility that DNA was replicated twice in a single cell
cycle, but that this process was inhibited in some way by
colchicine or Colcemid, is refuted by our experiments. In nonColcemid-treated cells, there was no overreplicated DNA in
cells harvested at mitosis, but early S-phase DNA replicated
again after mitosis (Figs. 5 and 6).
The apparent rapid traverse of the cell cycle after removal of
hydroxyurea was probably due primarily to the fact that 0.3
iriM hydroxyurea inhibits DNA synthesis to only about 20% of
Table 3 Aberrant and polyploitl cells and SCE per chromosome induced by
hydroxyurea
For aberration yields. 500 cells were analyzed; for SCE frequency. 50 cells
from each of two duplicate experiments were scored.
divisionHydroxy First
divisionPolyploid(%)3.0
urea
(mm)0
(%)4.617.6
0.3
1.0Aberrant21.0Poly-
(%)3.612.6
ploid
(%)2.2
chromosome0.54
±0.023°
10.0
0.78 ±0.021
7.68.6Aberrant
14.2Second 9.0SCE/ 1.30 ±0.086
" Mean ±SE.
control. Therefore, at the end of hydroxyurea treatment, the
cells had actually progressed much further through S phase
than would be expected if hydroxyurea formed an absolute
block to DNA synthesis. Mariani and Schimke (16) found that
mitotic cells plated immediately into hydroxyurea and BrdUrd
showed an appreciable amount of DHFR-hybridizing activity
in heavy-light DNA. This indicates that, even in their hands,
hydroxyurea was not an absolute block to replication. Thus, the
presumptive overreplication they observed can be explained by
this incomplete hydroxyurea block. Furthermore, several re
ports have indicated that other events associated with cell cycle
traverse continue in the presence of a strong inhibition of DNA
synthesis (30-32). Therefore, cell cycle progression may be
more rapid than normal after release from an inhibition of
DNA synthesis. Finally, it should be noted that Mariani and
Schimke previously reported the CHO-B11 cell line to have an
average cell cycle time of only 12 h (33), indicating that the 12
h during which these cells were allowed to recover from the
hydroxyurea treatment would be sufficient for them to enter a
second S phase, even if hydroxyurea acted as an absolute block
and no other cell cycle perturbations were induced.
The "onion-skin" model has been applied to integrated SV40
sequences; in this case there seems to be no doubt that DNA
replication is reinitiated within a single cell cycle (7). However,
the SV40 origin of replication is absolutely required (34), and
one can envisage this process as one in which the SV40 origin
acts autonomously after DNA damage, i.e., escapes control by
the host system, as during initiation of the lytic cycle of tem
perate bacteriophage. The experiments reported here show that
overreplication of early S-phase DNA following a 6-h hydrox
yurea block does not occur, at least within the limits of detection
of this technique (~3%) in CHO cells. Although overreplication
of small parts of the genome might occur at levels below this,
there is still no direct evidence that overreplication occurs at
any level in drug-treated mammalian cells. Therefore, alterna
tives to the onion-skin model should be considered.
Our results suggest two alternative mechanisms for gene
amplification, (a) We have observed that hydroxyurea, even at
0.3 mM, induces large amounts and many kinds of chromosome
aberrations (Table 3; Fig. 8). We speculate that the first step in
gene amplification occurs during the first step of selection,
when the drug, e.g., methotrexate, induces chromosome aber
rations. After mitosis, a chromosome fragment containing the
DHFR gene may occasionally segregate with the intact chro
mosome containing the gene, resulting in a cell with three
instead of two DHFR genes. Fragments of various sizes are
Fig. 8. Examples of chromosome aberra
tions seen in CHO-BI1 cells after a 6-h hy
droxyurea treatment. A, an acentric fragment
(arrow) resembling a double minute; H, multi
ple chromosome fragments and rearrange
ments.
B
4611
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1986 American Association for Cancer Research.
HYDROXYUREA
AND GENE AMPLIFICATION
seen in the first metaphase after hydroxyurea treatment (Fig.
8). We further suggest that once these fragments are generated
by methotrexate or hydroxyurea, they behave in a manner
analogous to chromosome-mediated
gene transfer, in which
selected genes are either maintained unstably and frequently
amplified (similar to double minutes) or are later stably inte
grated into a chromosome (35). We also found that the fre
quency of polyploidy increased after incubation with hydrox
yurea (Table 3). We have shown that these extra chromosomes
are not due to overreplication in a single cell cycle (24). It is
probable that polyploid cells were selected for in the mitotic
harvests because they are more resistant to the effects of the
hydroxyurea treatment on cell cycle progression than are neardiploid cells.
(b) A second mechanism is suggested by the observation of
increased SCE in hydroxyurea-treated cells (Table 3). Unequal
SCE has often been suggested as a mechanism for initiation of
gene amplification (for review, see Ref. 36). If only a very small
fraction of hydroxyurea-induced SCE involves unequal ex
change near the DHFR locus, it could explain the observed
hydroxyurea-induced increase in frequency of gene amplifica
tion (13).
The two possible mechanisms discussed here are not mutually
exclusive; together they could accommodate most of the data
that describe gene amplification events in drug-treated mam
malian cells.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
ACKNOWLEDGMENTS
24.
We gratefully acknowledge Barbara Young for expert technical as
sistance. Mary McKenney for outstanding editing, and Leslie Roberts
for patiently retyping the manuscript numerous times.
25.
26.
REFERENCES
1. Alt. F. W.. Kellems, R. E., Berlino, J. R., and Schimke. R. T. Selective
multiplication of dihydrofolate reducÃ-asegenes in methotrexate-resistant
variants of cultured murine cells. J. Biol. Chem., 253: 1357-1370, 1978.
2. Cowell, J. K. Double minutes and homogeneously staining regions: gene
amplification in mammalian cells. Annu. Rev. Genet., 16: 21-59. 1982.
3. Schimke, R. T. Gene amplification in cultured animal cells. Cell, 37: 705713, 1984.
4. Horns, R. C., Jr., Dower, W. J.. and Schimke, R. T. Gene amplification in a
leukemic patient treated with methotrexate. J. Clin. Oncol.. 2: 2-7, 1984.
5. Spradling, A. C. The organization and amplification of two chromosomal
domains containing Drosophila chorion genes. Cell, 27: 193-201, 1981.
6. Botchan, M.. Topp, W., and Sambrook, J. Studies on simian virus 40 excision
from cellular chromosomes. Cold Spring Harbor Symp. Quant. Biol., 43:
709-719. 1978.
7. Lavi, S. Carcinogen-mediated amplification of viral DNA sequences in simian
virus 40-transformed Chinese hamster embryo cells. Proc. Nati. Acad. Sci.
USA. 78: 6144-6148, 1981.
8. Biedler, J. L., and Spengler, B. A. Metaphase chromosome anomaly: asso
ciation with drug resistance and cell-specific products. Science (Wash. DC),
191: 185-187, 1976.
9. Roberts, J. M., Buck, L. B., and Axel, R. A structure for amplified DNA.
Cell, 33: 53-63, 1983.
10. Kaufman. R. J., Sharp, P. A., and Lait, S. A. Evolution of chromosomal
regions containing transfected and amplified dihydrofolate reducÃ-ase se
quences. Mol. Cell. Biol.. 3: 699-711, 1983.
11. Flintoff, W. F., Livingston. E., Duff. C.. and Worton. R. G. Moderate-level
gene amplification in methotrexate-resistant Chinese hamster ovary cells is
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
accompanied by chromosomal translocations at or near the site of the
amplified DHFR gene. Mol. Cell. Biol., 4: 69-76, 1984.
Varshavsky, A. Phorbol ester dramatically increases incidence of methotrex
ate-resistant mouse cells: possible mechanisms and relevance to tumor pro
motion. Cell. 25: 561 -572, 1981.
Tlsty, T., Brown, P. C., Johnston, R., and Schimke, R. T. Enhanced frequency
of generation of methotrexate resistance and gene amplification in cultured
mouse and hamster cell lines. In: R. T. Schimke (ed.), Gene Amplification,
pp. 231-238. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory,
1982.
Brown, P. C., Tlsty, T. D., and Schimke, R. T. Enhancement of methotrexate
resistance and dihydrofolate reducÃ-asegene amplification by treatment of
mouse 3T6 cells with hydroxyurea. Mol. Cell. Biol., 3: 1097-1107, 1983.
Barsoum, J., and Varshavsky, A. Mitogenic hormones and tumor promoters
greatly increase the incidence of colony-forming cells bearing amplified
dihydrofolate reducÃ-asegenes. Proc. Nail. Acad. Sci. USA, SO: 5330-5334,
1983.
Mariani, B. D., and Schimke, R. T. Gene amplification in a single cell cycle
in Chinese hamster ovary cells. J. Biol. Chem.. 259: 1901-1910. 1984.
Gautschi, J. R.. Kern, R. M., and Painter, R. B. Modification of replicón
operation in HeLa cells by 2,4-dinitrophenol. J. Mol. Biol., SO: 393-403,
1973.
Ockey, C. H. Fibre autoradiography of repair and replication in DNA from
single cells: the effect of DNA synthesis inhibitors. Eur. J. Cell Biol., 27:2533. 1982.
Klevecz, R. R., and Kapp, L. N. Intermittent DNA synthesis and periodic
expression of enzyme activity in the cell cycle of WI-38. J. Cell Biol., 58:
564-573, 1973.
Painter, R. B., and Schaefer, A. W. Rate of synthesis along replicons of
different kinds of mammalian cells. J. Mol. Biol.. 45:467-479, 1969.
Kafatos, F. C., Jones, C. W.. and Efstratiadis, A. Determination of nucleic
acid sequence homologies and relative concentrations by a dot hybridization
procedure. Nucleic Acids Res., 7: 1541-1552, 1979.
Chang, A. C. Y., Nunberg, J. H., Kaufman, R. J., Erlich, H. A., Schimke, R.
T., and Cohen, S. N. Phenotypic expression in E. coli of a DNA sequence
coding for mouse dihydrofolate reducÃ-ase.Nature (Lond.), 275: 617-624,
1978.
Wolff, S., Lindsley, D. L., and Peacock, W. J. Cytological evidence for
switches in polarity of chromosomal DNA. Proc. Nati. Acad. Sci. USA, 73:
877-881, 1976.
Morgan, W. F., Bodycote, J., Fero, M. L., Hahn, P. J., Kapp, L. N., Pantelias,
G. E., and Painter, R. B. A cytogenetic investigation of DNA rereplication
after hydroxyurea treatment: implications for gene amplification. Chromosoma (Beri.), 93:191-196, 1986.
Rydberg, B. Repair of double-strand breaks in Colcemid-arrested mitotic
Chinese hamster cells. Int. J. Radiât.Biol., 46: 299-304, 1984.
Wolff, S., and Perry, P. Insights on chromosome structure from sister
chromatid exchange ratios and the lack of both isolabelling and heterolabelling as determined by the FPG technique. Exp. Cell Res., 93: 23-30, 1975.
Ramseier, H. P.. Burkhalter, M., and Gautschi. J. R. Survival of the CHO
cells that had replicated DNA in the presence of hydroxyurea. Exp. Cell
Res.. 105:445-453, 1977.
Meyn, R. E., Hewitt, R. R.. Thomson, L. F.. and Humphrey, R. M. Effects
of ultraviolet irradiation on the rate and sequence of DNA replication in
synchronized Chinese hamster cells. Biophys. J.. 16: 517-525, 1976.
Radford. I. R., Martin, R. F., Finch. L. R.. and Hodgson, G. S. Inhibition of
DNA synthesis and cell death. Biochim. Biophys. Acta, 696: 154-162, 1982.
Salzman, N. P., Pellegrino, M., and Franceschini, P. Biochemical changes
in phytohemagglutinin stimulated human lymphocytes. Exp. Cell Res., 44:
73-83, 1966.
Milner, G. R. Nuclear ultrastructure of the transforming lymphocyte during
inhibition of deoxyribonucleic acid synthesis with hydroxyurea. J. Cell Sci.,
¿•583-591,1969.
Painter. R. B., and Schaefer, A. W. Variation in the rate of DNA chain
growth through the S phase in HeLa cells. J. Mol. Biol., 58: 289-295, 1971.
Mariani, B. D., Slate, D. L., and Schimke, R. T. S phase-specific synthesis
of dihydrofolate reducÃ-asein Chinese hamster ovary cells. Proc. Nati. Acad.
Sci. USA, 78:4985-4989, 1981.
Lavi, S. Carcinogen-mediated activation of SV40 replicons: a model system
for initiation of carcinogenesis. In: R. T. Schimke (ed.), Gene Amplification,
pp. 225-230. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory,
1982.
Klobutcher, L. A., and Ruddle, F. H. Phenotype stabilisation and integration
of transferred material in chromosome-mediated
gene transfer. Nature
(Lond.), 280: 657-660, 1979.
Stark, G. R., and Wahl, G. M. Gene amplification. Annu. Rev. Biochem.,
53:447-491, 1984.
4612
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1986 American Association for Cancer Research.
Chromosomal Changes without DNA Overproduction in
Hydroxyurea-treated Mammalian Cells: Implications for Gene
Amplification
Peter Hahn, Leon N. Kapp, William F. Morgan, et al.
Cancer Res 1986;46:4607-4612.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/46/9/4607
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at pubs@aacr.org.
To request permission to re-use all or part of this article, contact the AACR Publications
Department at permissions@aacr.org.
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1986 American Association for Cancer Research.