[CANCER RESEARCH
45, 5050-5057,
October 1985]
Increased Gene Amplification in L5178Y Mouse Lymphoma Cells with
Hydroxyurea-induced Chromosomal Aberrations1
Anna B. Hill2 and Robert T. Schimke
Department of Biological Sciences, Stanford University, Stanford, California 94305
ABSTRACT
MATERIALS
Chromosomal aberrations and dihydrofolate reducÃ-ase gene
amplification are observed in L5178Y mouse lymphoma cells
after treatment with hydroxyurea. The types of aberrations in
clude polyploidy, endoreduplication, chromosome fragmentation,
and the presence of extrachromosomal
DMA. Hydroxyureatreated cells analyzed by cell sorting showed a subpopulation of
cells with increased DMA and increased dihydrofolate reductase.
This subpopulation shows a high incidence of chromosome
aberrations and an increased frequency of dihydrofolate reduc
tase gene amplification. Hydroxyurea-treated cells with the nor
mal amount of DNA and dihydrofolate reductase have few ab
errations and a low frequency of dihydrofolate reductase gene
amplification. We propose that hydroxyurea treatment causes
overreplication of DNA and that recombination of overreplicated
DNA can lead to chromosome aberrations and gene amplifica
tion.
INTRODUCTION
Previous studies in our laboratory have examined the process
of gene amplification in cultured mammalian cells. The frequency
of cells resistant to MTX3 can be increased by 10-fold or more
by pretreatment of cells with agents such as HU, UV light, and
different carcinogens (1, 2). These agents which enhance DHFR
amplification share the feature that they inhibit DNA synthesis.
Mariani and Schimke (3) showed that, if DNA synthesis was
inhibited during the second hour of S phase in synchronized
CHO cells, DNA synthesized before the HU block was rereplicated once the HU was removed. This overreplication resulted
in an increase in the number of DHFR genes and led to enhanced
resistance to MTX.
HU is a specific inhibitor of ribonucleotide reductase, inhibits
DNA synthesis (4), and is known to induce chromosome aber
rations in vivo and in vitro (4-6). The present studies examine
the relationship between HU-induced chromosome aberrations
and the frequency of DHFR gene amplification. Our results show
that the highest frequency of MTX resistance and gene amplifi
cation occurs in the cell population with many types of chromo
some aberrations including polyploidy, chromosome fragmenta
tion, and the presence of extrachromosomal DNA. We propose
that the generation of chromosomal aberrations is a conse
quence of recombination events subsequent to overreplication
of DNA.
1Supported by a research grant from the NIH (GM 14931) and a fellowship to
A. B. H. from the NIH (IF32CA07197).
'Present address: Department of Radiation Oncology, School of Medicine,
University of Arizona, Tucson, AZ 85724.
3 The abbreviations used are: MTX, methotrexate; HU, hydroxyurea; BrdUrd, 5bromodeoxyuridine; SCE, sister chromatid exchange; F-MTX, fluorescein-conjugated methotrexate; DHFR, dihydrofolate reductase; FACS, fluorescence-activated
cell sorter; HBSS, Hanks' buffered saline solution; CHO, Chinese hamster ovary;
cDNA, complementary DNA.
Received 4/1/85; revised 6/25/85; accepted 6/28/85.
CANCER
RESEARCH
AND
METHODS
Cell Culture and HU Treatment. The Jsens cells used in these
experiments are L5178Y mouse lymphoblastoid cells; the C3 subclone
of Jsens has been previously isolated by Dolnick et al. (7). C3 cells are
highly resistant to MTX and contain 800 to 1000 copies of the mouse
DHFR gene localized in a homogeneously staining region on chromo
some 2. Both these cell lines have stable karyotypes with 40 to 42
chromosomes. Jsens and C3 cells were maintained in RPMI medium
supplemented with 10% horse serum, 100 units of penicillin, and 100 ^g
of streptomycin per ml. The medium used for C3 cells also contained
200 fiM MTX. The cells were grown in suspension at 37°C in an
atmosphere of 5% CO2 and passaged by a 1:50 dilution every 3 days.
Exponentially growing cultures of Jsens and C3 cells at concentrations
of 5 x 10* cells/ml were treated with a freshly made HU stock (Sigma).
One M or 100 mM stocks of HU were dissolved in medium and added to
the cultures in 100- to 500-/J aliquots. Control and treated cultures were
incubated for 6 h at 37°C. HU was removed by centrifuging the cultures
at 225 x g
and the cell
centrifuged
placed into
for 5 min to pellet the cells. The old medium was removed,
pellet was resuspended in prewarmed HBSS. The cells were
and washed one more time in HBSS. The cells were then
prewarmed medium and incubated at 37°C.
In some experiments, Jsens cells were grown in soft agar so that
surviving cells could form individual colonies. When cells were grown
without MTX, 200 to 500 cells were suspended in 5 ml of medium
containing 0.4% agar (Seaplaque). The cells were then placed over a 2ml underiayer of medium containing 0.9% agar in a 60 mw Retri dish.
Individual colonies were visible in 4 to 6 days. For MTX selections, MTX
stocks were prepared as described by Brown ef al. (1). One x 10s to 5
x 106 cells were suspended in medium containing 0.4% agar and 10 nM
MTX. These cells were placed over a 5-ml underiayer of medium con
taining 0.9% agar and 10 nw MTX in a 100 mM Retri dish. Individual
colonies were visible in 2.5 to 3 wk. At 1-wk intervals after cells were
plated out, the plates were briefly cooled until the agar became solid,
and 10 ml of fresh medium containing 0.4% agar and 10 nivi MTX were
carefully pipetted over the cells. The number of surviving colonies was
counted using a dissecting microscope. Individual clones were removed
by suction from a pipet, and clones were expanded by continuous growth
in suspension.
Determination of DNA Synthesis. Either 6 h after the addition of HU
or 6 h after the removal of HU, the number of Jsens and C3 cells in each
culture was determined by using a hemocytometer. The cultures were
then labeled with [mef/)y/-3H]thymidine (1 ¿iCi/ml;New England Nuclear;
6.7 Ci/mmol) for 15 min. 3H incorporation was stopped by the addition
of ice-cold HBSS. The cells were collected by centrifugation as above
and washed 2 more times with cold HBSS. The cell pellet was lysed in
10 mM Tris, pH 8.0, containing 0.5% sodium dodecyl sulfate, 5 mM
EDTA, and proteinase K (50 ^g/ml) and incubated at 37°C overnight.
Fifty ng of salmon sperm DNA were added as a carrier to each sample.
Trichloroacetic acid precipitation and quantitation of [3H]thymidine incor
porated were as previously described (3). Rates of [3H]thymidine incor
poration were expressed as total precipitatile radioactivity incorporated
per 103 cells. Each determination was done in triplicate.
Chromosome Preparations. To observe SCEs, cells were grown for
24 h (2 cell cycles in normal cells) in the presence of 10 UM BrdUrd. Two
h before cell fixation, Colcemid (2 x 10~7 M final concentration) was
added. Chromosome fixations and fluorescence plus Giemsa staining
were as described by Hill and Wolff (8), except that the cells were swollen
in 0.075 M KCI for 10 min. SCEs were analyzed in 50 metaphases for
each point, and the data were expressed as the mean number of SCEs
per cell. These same slides were studied to quantitate other types of
chromosome abnormalities.
VOL. 45 OCTOBER
1985
5050
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
GENE AMPLIFICATION
AND CHROMOSOMAL
CHROMOSOME
FIXATION
Fluorescein MTX, DNA Fluorochrome Staining, and Use of the
FACS. For F-MTX staining, Jsens control and HU-treated cells were
placed in medium with 5 n»F-MTX supplemented with 30 pM each of
glycine, hypoxanthine, and thymidine 24 h before FACS analysis. Before
sorting, the medium containing F-MTX was removed, and the cells were
resuspended in medium lacking F-MTX. F-MTX purification and cell
sorting on the FACS II were as described by Johnston ef al. (9).
Distribution of fluorescence for a cell population was determined from at
least 10,000 cells.
Jsens cells were sorted into 2 populations; the bright population in
the control cells was the cells with F-MTX fluorescence of 90 units or
higher. The bright population of the HU-treated cells was all cells that
had F-MTX fluorescence greater than or equal to the.control bright
population. The control dull population was those cells with F-MTX
fluorescence less than 90 units; the HU dull population also had F-MTX
fluorescence of less than 90 units. F-MTX dull and F-MTX bright popu
lations were sorted from each sample and placed into medium. For
chromosome analysis, sorted cells were incubated for 8 h in the presence
of 2 x 10~7 M Colcemid, and chromosome preparations were made as
BrdU
EARLY BrdU LABEL
LATE BrdU LABEL
BrdU
CHROMOSOME
FIXATION
* 3H INCORPORATION
20
10
30
40
50
60
TIME (hours)
fetoprotein cDNA (12). Nick translations and hybridization conditions
were as described in Brown er al. (13). Filters were autoradiographed,
and the degree of hybridization was estimated by scanning exposed Xray film with a Quick-Scan densitometer (Helena Laboratories).
RESULTS
Transient Inhibition of DNA Synthesis in L5178Y Cells. The
experimental protocol in which HU was used to block DNA
synthesis in actively growing Jsens and C3 cells is shown in
Chart 1. After 6 h of incubation with 1 and 10 mw HU, [3H]thymidine incorporation is reduced to less than 5% of control
(control, 51 x 1(T3 counts/cell; HU, 1.8 x 1fr3 counts/cell). Six
h after removal of HU, [3H]thymidine incorporation returned to
40 to 80% of the control (control, 41 x 10~3 counts/cell; HU, 16
to 34 x 10~3 counts/cell). These results show that HU at these
concentrations is a strong inhibitor of DNA synthesis, but DNA
synthesis resumes rapidly once HU is removed. These data are
consistent with the transient HU inhibition studies of Brown ef
a/. (1) and Mariani and Schimke (3).
HU Enhancement of Sister Chromatid Exchange Occurring
during S-Phase. The presence of SCEs is considered a mani
festation of S-phase-specific DNA damage. Detection of SCEs
requires that BrdUrd be incorporated into DNA for 2 cell cycles,
and we have used 2 protocols to study SCEs. In the first protocol,
BrdUrd was added to the medium either directly after the HU
RESEARCH
*
U
5 x IO5
above. Cells from sorted populations were also placed in soft agar with
and without MTX to determine the number of viable cells and the
frequency of MTX resistance.
Determination of the amount of DNA in viable cells was previously
described by Arnt-Jovin and Jovin (10). The amount of DNA per cell was
measured by staining Jsens cells with 5 ^M Hoechst 33342 for 1 h at
37°C prior to FACS analysis. A UV laser (355 nm) was used to analyze
the Hoechst-stained cells. Jsens cells were sorted into 2 populations;
the control bright population was those cells with Hoechst fluorescence
of 110 units or higher. The bright population in the HU-treated cells
included all cells with fluorescence greater than or equal to the control
bright population. The control dull population was those cells with less
than 110 units of Hoechst fluorescence; the HU dull population was also
those cells with Hoechst fluorescence of 110 units or less. Control and
HU-treated cells were sorted into Hoechst dull and bright populations,
and these cells were analyzed for chromosome abnormalities and growth
in soft agar with and without MTX as described above.
Assessment of DHFR Gene Amplification. Individual clones isolated
after growth in soft agar were expanded by growing cells in suspension
in the presence of 10 nw MTX. DNA from these clones was isolated and
purified as previously described (1). One ng of DNA from each clone was
bound to a nitrocellulose filter using the slot hybridization technique (1).
As a standard, increasing amounts of Jsens DNA were also applied to
each filter. Filters were prepared in duplicate so samples could be
hybridized to both ^P-labeled DHFR cDNA (11) or "P-labeled mouse «-
CANCER
ABERRATIONS
Chart 1. Protocol of hydroxyurea treatment of Jsens and C3 cells. Jsens or C3
cells (5x10*) were seeded into flasks, and 18 h later, exponentially growing cells
were treated with 1 or 10 mw HU for 6 h. Ten mw BrdUrd (BrdU} was added directly
to cultures after HU removal (early BrdUrd label) or 6 h after HU was removed (late
BrdUrd label). The cells were placed in Colcemid, and metaphase spreads were
examined at various times thereafter (generally 24 h after addition of BrdUrd).
Rates of DNA synthesis were measured by [3H]thymidine incorporation at the end
of the HU treatment and 6 h after HU removal.
Table 1
Hydroxyurea effect on sister chromatid exchange
Jsens and C3 cells were treated with HU for 6 h; after HU was removed, cells
were either placed into medium containing BrdUrd (early BrdUrd label) or allowed
to recover for 6 h before BrdUrd was added (late BrdUrd label). Cells were fixed
24 h after the addition of BrdUrd; metaphase spreads were made, and slides were
stained by the fluorescence plus Giemsa technique.
BrdUrd
label9.9
±0.5*
C3
cellsJsens
mu HU
lOmMHUControl
20.3 ±1.7
15.1
±1.39.2
BrdUrd
label9.4
±0.5
10.2 ±0.4
11.3
±0.59.9
±0.5
±0.6
41.7 ±2.6
9.5 ±0.4
23.4 + 1.7Late
9.7 ±0.5
* Mean ±SE basedon examinationof 50 metaphasespreadsfor each treatment.
cellsTreatmentControl1
1 mMHU
10mwHUEarly
treatment (early BrdUrd label) or BrdUrd was added after the
cells were allowed to recover from the treatment for 6 h (late
BrdUrd label) (Chart 1). The cells were fixed 24 h later, and the
number of SCEs per cell was determined.
An increase in the number of SCEs produced by HU is detected
when the BrdUrd was added directly after the HU treatment
(Table 1). A 1 HIM HU treatment gave a 4-fold increase in SCEs
in Jsens cells, while the increase was 2-fold in C3 cells. Ten mw
HU induced fewer SCEs than 1 mw HU in both the C3 and Jsens
cell lines. In each case, the increase in number of SCEs induced
by 10 mw HU was half that induced by a 1 mw HU treatment.
Fig. 1, A and D, shows examples of metaphases with many
SCEs induced by the HU treatment.
If BrdUrd was added 6 h after removing HU from the culture,
the number of SCEs was not significantly different from the
control. The Jsens cells showed no effect with either a 1 mw or
a 10 ITÃŒM
treatment, while the C3 cells increased by only 2 SCEs
per cell at 10 mw HU; neither value is significantly different from
the control using Student's f test (P > 0.05). These results show
that HU induces a large increase in the number of SCEs at the
time of treatment, but that once the HU is removed, the number
of SCEs detected rapidly diminishes. The events that generate
SCEs are completed within 6 h after HU removal, a time when S
phase is still continuing in many cells.
Many metaphases present 24 h after the HU treatment was
fragmented and had other chromosome aberrations. These cells
did not complete an S phase or underwent only a single S phase
in the presence of BrdUrd, as indicated by the lack of differentially
VOL. 45 OCTOBER
1985
5051
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
GENE AMPLIFICATION
AND CHROMOSOMAL
labelled chromatids. Thus, it was impossible to detect SCEs in
such cells. To determine whether highly fragmented chromo
somes had a high frequency of SCEs, Jsens cells were treated
with BrdUrd for 12 h before 1 mw HU was added to the cultures,
and following removal of HU, the cells were cultured for 24 h in
medium containing BrdUrd. Under these conditions, fragmented
metaphases had differentially labeled chromatids when stained
by the fluorescence plus Giemsa technique; i.e., such cells had
undergone an S phase prior to HU block and had completed the
S phase in which HU had transiently inhibited DMA synthesis. In
most of the metaphases with fragmented chromosomes, a very
high number of SCEs was evident; an example is shown in Fig.
1F. We interpret this finding to indicate that fragmentation of
chromosomes as induced by HU is a lethal event; i.e., cells can
complete the S phase in which HU-induced fragmentation has
occurred, but they cannot progress through an S phase subse
quent to chromosome fragmentation. The decrease in number
of SCEs observed with 10 HIM HU compared with 1 mw HU as
detected with BrdUrd labeling in the first 6 h after removal of HU
(Table 1) is probably explained by the fact that 10 rriM HU results
in a greater number of cells with fragmented chromosomes;
hence, they would not progress through a second S phase to
allow detection of SCEs.
Chromosomal Aberrations Produced by Transient Inhibition
of DMA Synthesis. A feature of both the Jsens and C3 cell lines
is that the karyotype is stable under normal growth conditions.
Forty to 42 chromosomes are seen in both cell types; aneuploid
and polyploid cells are rare (Table 2). Yet after cells were treated
with HU for 6 h and fixed 24 h after HU removal and after
metaphases were stained with the fluorescence plus Giemsa
technique, a wide range of chromosomal aberrations was ob
served. Features such as polyploidy (Fig. 1A), endoreduplication
(Fig. 1D), chromatid breaks, translocations, and acentric frag
ments (Fig. 1C), and dicentrics (Fig. 1£)were observed com
monly. Metaphases with intact chromosomes and varying
amounts of extrachromosomal DMA were also seen frequently
(Fig. ^B). Staining with Hoechst 33258 verified that the extrachromosomal material was DMA. The same spectrum of chro
mosomal aberrations was seen when cells were treated only
with HU, i.e., no BrdUrd. Thus the chromosomal aberrations
observed are the consequence of HU treatment.
A summary of chromosome aberrations seen in Jsens and C3
cells 24 h after the HU was removed is shown in Table 2.
Chromosome aberrations of all types were increased in both C3
and Jsens cells. Twenty to 30% of all metaphases were abnormal
after the 1 and 10 mw HU treatments. There were some differ
ences in the frequency of aberrations induced in C3 and Jsens
ABERRATIONS
cells. The Jsens cells had more fragmented metaphases than
did the C3 cells, while C3 cells had a larger number of meta
phases with extrachromosomal DNA than did the Jsens cells.
An important observation is that the majority of chromosome
aberrations appeared in metaphases seen 24 h after the HU
treatment. If metaphases were examined 12 or 72 h after a 10
HIM HU treatment, 92 to 95% of the metaphases were normal.
Thus the time for observation is critical. Cells with aberrant
chromosomes either fail to divide multiple times (hence not
detected at mitosis), or the aberrations are resolved, since few
chromosome aberrations are seen 72 h after HU treatment.
HU Effects on Cell Survival and MTX Resistance. Since
many chromosomal abnormalities were present after HU treat
ment, it was important to determine the long-term viability of
treated cells. Although Jsens and C3 cells normally grow in
suspension, they can form colonies from single cells when grown
in 0.4% soft agar. Eighteen h after HU treatment, Jsens cells
were counted and placed into soft agar to determine relative
plating efficiencies. At the same time, cells were placed in soft
agar containing 10 nw MTX to determine the number of MTXresistant colonies that arose after the HU pretreatment.
Treatment with 1 or 10 mw HU reduced the plating efficiency
of the Jsens cells to 28% of the control. When HU-treated cells
were tested for resistance to 10 nw MTX, a 4.8- to 6.5-fold
increase in resistant colonies was observed (Table 3). The in
crease in MTX-resistant colonies was dependent on when cells
were plated into MTX. If the cells were allowed to recover from
the HU treatment 72 h before they were placed into soft agar,
the relative plating efficiency without MTX increased so that 73%
of the cells formed colonies in soft agar. When these same cells
are selected in 10 nw MTX, the number of resistant colonies
from a 10 mw HU treatment was only 1.7-fold higher than from
control cells. These results are consistent with the studies of
Brown ef al. (1), which showed that the frequency of MTX
resistance was most enhanced if cells were plated into MTX 15
h after the removal of HU.
MTX resistance can occur by several mechanisms including
altered transport of MTX into cells (14), altered affinity of DHFR
for MTX (15, 16), and amplification of the DHFR gene (17). Soft
agar clones resistant to 10 nw MTX from control, 1 mw HU, and
10 mw HU-treated cells were expanded by growing them in
suspension, and DNA from each clone was analyzed to deter
mine if the DHFR gene had been amplified. The analysis was
done by a modified dot hybridization technique in which 1 ng of
DNA from each clone was hybridized to 32Pnick-translated cDNA
to DHFR. At the same time, 1 ng of DNA from each clone was
hybridized to 32Pnick-translated a-fetoprotein. Previous studies
Table 2
Chromosome aberrations induced by hydroxyurea
Jsens and C3 cells were treated with HU for 6 h; after HU was removed, cells were placed in medium containing BrdUrd. Colcemid-arrested
h later. Values are mean of 200 metaphases for each treatment.
cells were examined 24
Chromosome aberrations as % of total metaphases (24 h after treatment)
Treatment
Normal
Polyploid
Endoreduplicated
96.0
72.0
68.0
3.5
5.5
7.0
0.0
1.5
1.0
Fragmented
Extrachromosomal
DMA
C3 cells
Control
1 mwHU
lOmwHU
0.0
7.0
1.0
0.0
15.0
22.0
Jsens cells
Control1
mMHU10rriMHU10
HU10mMHU72hafterHU100.074.083.092.095.00.04.02.02.04.00.00.51.00.00.00.010.011.04.01.00.04.03.01.00.0
mM HU 12 h after
CANCER
RESEARCH
VOL. 45 OCTOBER
1985
5052
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
GENE AMPLIFICATION
AND CHROMOSOMAL
ABERRATIONS
Table3
Hydroxyurea-induced
increase in methotrexate-resistant
colonies
DHFR
Unsorted Jsens cells were treated with HU for 6 h; after HU was removed, cells
were allowed to recover for 18 h and then placed in medium containing soft agar
to determine plating efficiencies. At the same time, cells were placed in soft agar
containing 10 nw MTX. Colonies were counted 1 to 3 wk later. Control and HUtreated cells were sorted into F-MTX dull and bright as well as Hoechst dull and
bright subpopulations as described in "Results." These subpopulations were placed
ALPHA
FETOPROTEIN
DHFR
GENE
COPY
in soft agar as were the unsorted populations. The frequency of MTX-resistant
cells includes a correction factor for the effects of treatments (HU and sorting) on
plating efficiencies (in all cases, the plating efficiency in normal medium is set as
100%).
1.1
Frequency
of survivors in
Plating efficiency
in soft agar
Unsorted Jsens cells
Control
1 niMHU
10 mw HU
10mMHU(72hafter
HU)
71.0 ±8.9*
20.5 ±4.5
19.5 ±2.8
52.0 ±5.3
1.1
10 â„¢
methotrexate
(x 10-*)
CONTROL
Fold increase
over control
1.2
1.3
0.92 ±0.1
5.93 ±0.9
4.44 ±0.5
1.57 ±0.3
1.0
6.5
4.8
1.7
1.5
1.8
1 MM HU
1.7
Cell-sorted
populationsControl,
dullControl,
F-MTX
F-MTXbrightlOmMHU,
2.0
2.5
F-MTXdull10mwHU,
2.4
10 MM HU
F-MTXbrightControl,
3.1
dullControl,
Hoechst
HoechstbrightlOmwHU,
4.3
Hoechstdull10
Fig. 2. Assessment of gene amplification in methotrexate-resistant clones. DNA
Hoechstbright13.73.210.50.833.231.226.610.40.663.302.865.690.480.100.321.541.05.04.38.61.00.20.73.2'
row HU,
was isolated from clones resistant to 10 nu MTX, and 1 v<3of DNA was applied to
nitrocellulose filters by the slot hybridization technique (see 'Materials and Meth
ods"). One filter was hybridized with MP-DHFR cDNA; an identical filter was
Mean ±SE.
hybridized to a-Å“P-fetoprotein cDNA. Filters were placed on X-ray film, and films
by Brown ef al. (1) showed that a-fetoprotein does not amplify
under MTX selection; therefore it acts as an internal standard to
measure the amount of DMA present.
The comparison between DHFR cDNA hybridization and afetoprotein cDNA hybridization in the 12 clones studied is shown
in Fig. 2. The amount of hybridization was quantitated by densitometry, and the degree of DHFR amplification was determined
by comparing the ratio of DHFR to a-fetoprotein hybridization;
clones showing a ratio of 1.5 or greater were considered to have
amplified DHFR genes (Fig. 2). With this technique, all 8 of the
clones pretreated with HU showed increased DHFR gene copy
number relative to the control. None of the 4 MTX-resistant
clones from control cells had amplified the DHFR gene. These
results in lymphoma cells show that HU-pretreated cells have a
higher frequency of DHFR gene amplification to account for the
increased frequency of MTX resistance. These data also suggest
that the effect of HU on gene amplification is dose dependent,
since a 10 HIM pretreatment induced a higher number of DHFR
gene copies as compared to the number induced by a 1 HIM HU
pretreatment.
Analysis of Hydroxyurea-treated Cells Using the FACS.
Studies by Mariani and Schimke (3) showed that, if HU was
added during S phase to synchronously growing CHO cells,
overreplication of DNA occurred when HU was removed from
the culture. We were interested in studying whether Jsens cells
had an increased cellular DNA content following treatment with
HU. The amount of DNA was measured by staining cells with
the fluorochrome Hoechst 33342 and sorting them with the
FACS 24 h after HU was removed. The data from these experi
ments are displayed as a 2-dimensional contour plot of cell
distribution in which cell size is displayed along the Y axis and
DNA content on the X axis. A series of such distributions for
CANCER
RESEARCH
were scanned by a densitometer to determine the degree of hybridization. The
DHFR gene copy was calculated by comparing DHFR hybridization to a-fetoprotein
hybridization in each done. Clones with a DHFR gene copy of greater than 1.5
were considered amplified. The first 4 slots are increasing amounts of nonresistant
Jsens DNA; these amounts are 0.5 <*gDNA, 1 ,ig DNA, 2 (Jg DNA, and 3 ng DNA.
Clones from the HU-pretreated cells are amplified for DHFR; clones from control
cells resistant to MTX are not amplified.
control cells and cells treated with 10 HIM HU is shown in Chart
2. In control cells, integration analysis indicated that 15.8% of
the cells showed a fluorescence intensity of 110 units or greater
(Chart 2A). Twenty-four h after the removal of cells from 10 HIM
HU, 35.2% of the cells show fluorescence greater than 110 units
(Chart 2B). The cells that show higher Hoechst fluorescence also
have a larger amount of forward scatter during the sorting, which
indicates increased cell size. If cells are fixed 24 h after the
removal of HU and examined under the microscope, many of
these cells have a diameter of more the 2 times control cells.
The increased amount of DNA per cell and the increased cell
size are transient properties of the cell population. If cells are
allowed to recover for 72 h prior to FACS analysis, the population
is very similar to the control with respect to cell size and DNA
content (Chart 2C).
Jsens cells treated with 10 HIM HU were also analyzed 24 h
after the removal of HU to determine if individual cells in the
population had an increased amount of DHFR enzyme. This was
measured by staining cells with a fluorescein conjugate of MTX
(F-MTX) and measuring the number of highly fluorescent cells
on the FACS. A number of studies from this laboratory (9, 18)
have shown that the fluorescence intensity of cells stained with
F-MTX is a function of DHFR content.
Contour plots for cells stained with F-MTX are shown in Chart
2, D to F. The brightest cells in the control population, those with
VOL. 45 OCTOBER
1985
5053
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
GENE AMPLIFICATION
AND CHROMOSOMAL
ABERRATIONS
Table 4
Chromosome aberrations in sorted cells
Jsens cells were treated with 10 HIM HU for 6 h; after HU was removed, cells
either were placed in medium containing F-MTX and analyzed on the cell sorter 24
h later or were placed into fresh medium and stained with Hoechst 33342 prior to
FACS analysis 24 h later. Control and HU-treated cells were sorted for bright and
dull F-MTX fluorescence or dull and bright Hoechst fluorescence (see "Materials
and Methods"). The sorted populations were placed into medium containing Col
cemid, and cells were fixed 8 h later. Metaphase spreads were examined from
each cell population. Each value is the mean of analysis of 50 metaphases.
Metaphases (%)
PotyFrag-ploid
chromosomal
rescenceDullBrightDullBrightDullBrightDullBrightNormal100989672100959
DyeF-MTXHoechstTreat
mentControlControl10niM
merited02214000180021000010000405830
DNA
HUlOmM
HUControlControl10rriM
40
M)
F MIX
120
160
200
FLUORESCENCE
40
80
FMTX
160
120
200
FLUORESCENCE
IO
80
FMTX
120
160
200
HUlOmMHUFluo
FLUORESCENCE
Chart 2. Contour plots of Jsens cell populations analyzed by the fluorescenceactivated cell sorter. Chart 2, A to C, shows cells stained for 1 h with Hoechst
33342 to determine DNA content per cell. A, control cells; 15.8% of the population
has a fluorescence intensity of 110 units or greater. In B, cells are pretreated with
10 mM HU and analyzed 24 h after HU removal; 35.5% of the population has a
fluorescence intensity of 110 units or greater. In C, cells are pretreated with 10 mM
HU and analyzed 72 h after HU removal; only 11% of the population has a
fluorescence intensity of 110 units or greater. Chart 2, D to F, shows cells stained
for 24 h with F-MTX to determine DHFR enzyme content. D, control cells; 8.3% of
the population has a fluorescence intensity of 90 units or greater. In £,cells are
pretreated with 10 mw HU and analyzed 24 h after the removal of HU; 51.6% of
the population has a fluorescence intensity of 90 units or greater. In F, cells are
pretreated with 10 mw HU and analyzed 72 h after HU removal; only 4% of the
cells have a fluorescence intensity of 90 units or greater. Contour plots are based
on 10,000 cells; each line represents 20% of the population.
F-MTX fluorescence greater than 90 units, make up only 8.3%
of the population (Chart 2D). Twenty-four n after the removal of
HU, 51.6% of the cells have a fluorescence intensity of greater
than 90 units (Chart 2E). In addition to having more DHFR, these
cells are also larger in size. This increase of DHFR per cell and
the larger cell size are transient; thus if the cells are allowed to
recover from HU for 72 h before FACS analysis, the pattern is
similar to untreated cells (Chart 2F).
In order to show that the HU-treated cells with bright F-MTX
fluorescence were also bright for Hoechst fluorescence, control
and HU-treated cells were stained with both F-MTX and Hoechst.
Greater than 50% of the cells bright for F-MTX fluorescence
were also bright for Hoechst fluorescence (data not shown).
From these data, we conclude that, after cells are treated with
HU, a subpopulation of cells is generated in which the cells are
larger than normal cells, have increased DHFR enzyme content,
and have an increased amount of DNA per cell.
Chromosome Abnormalities and MTX Resistance in Cell
Populations Sorted Using the FACS. By analysis with the
FACS, we have shown that a subpopulation of cells with high
fluorescence for F-MTX (Chart 2, D to F) and Hoechst 33242
(Chart 2, A to C) appeared 24 h after HU treatments. These
subpopulations were analyzed for the frequency of chromosomal
abnormalities and increased resistance to MTX. Control and HUtreated cells were stained separately with F-MTX and Hoechst
and sorted into bright and dull subpopulations. The cells were
then treated with Colcemid, and metaphase spreads were ex
amined. The types of abnormalities observed are summarized in
Table 4.
When sorting cells for variable staining with F-MTX, metaphases observed in the control dull population were normal.
Among the control bright cells, only 2% of the metaphases were
CANCER RESEARCH
abnormal. In contrast, in cells treated with 10 mw HU, chromo
some aberrations were common. Four % of the treated dull cells
were abnormal, while 28% of the bright cells had chromosome
alterations. Many of the latter were polyploid, had extrachromosomal DNA, and/or were fragmented.
When the same cells were stained with Hoechst 33342 and
sorted into dull and bright subpopulations, the metaphases in
the control dull population were all normal, while 5% of the
metaphases in the control bright population were abnormal. In
the 10 HIM HU-treated cells, the dull population had few aberrant
metaphases; 8% of the metaphases were abnormal. The great
est number of chromosome aberrations was seen in the HUtreated bright population. Fifty-eight % of these metaphases
were abnormal, and many types of chromosome aberrations
were seen (Table 4).
These studies show that the subpopulations of large cells
produced by HU treatment, which has more DHFR and more
DNA per cell, also has a high frequency of cells with chromosome
aberrations. To determine whether there were viable cells in this
subpopulation and if these cells display increased resistance to
MTX selection, experiments were performed on control cells and
on cells sorted 24h after HU treatment (Table 3). Cells sorted by
the FACS had a much lower viability than unsorted cells. In cells
sorted with F-MTX, only 13.7% of the control dull cells formed
colonies in soft agar. Both control bright cells and the 10 HIM HU
bright cells had low plating efficiencies compared to the control
dull cells and the 10 mM HU dull cells. However, all populations
did produce viable colonies. When these same F-MTX dull and
bright populations were selected in soft agar containing 10 DM
MTX, almost 9 times as many colonies resistant to MTX occurred
in the 10 HIM HU bright cells than from the control dull cells. An
intermediate increase in MTX resistance was seen in the control
bright cells and the 10 mM HU dull cells (Table 3). In a separate
experiment, control and 10 HIM HU-treated cells were stained
with Hoechst 33342 and sorted 24 h later. An increased fre
quency of MTX resistance was seen only in the Hoechst bright
subpopulation.
From the series of cells sorted with F-MTX, 6 MTX-resistant
colonies from the control cells (3 dull and 3 bright) and 6 MTXresistant colonies from 10 ITIM HU-treated cells (3 dull and 3
bright) were expanded by growing them in suspension; meta
phases and DHFR gene copy number for each colony were
analyzed. No gross chromosomal aberrations were seen in any
of the clones, although polyploid cells were more common than
VOL. 45 OCTOBER
1985
5054
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
GENE AMPLIFICATION
AND CHROMOSOMAL
in the parent population. Analysis of the DHFR gene copy number
showed that none of the control dull or HU dull colonies had
amplified the DHFR gene. One of the 3 control bright clones and
2 of the 3 HU bright clones had amplified the DHFR gene.
Because of the small number of colonies examined, we can
made no conclusions concerning the relative frequency of gene
amplification among the various cell populations.
DISCUSSION
Previous studies by Mariani and Schimke (3) examined effects
of HU on synchronously growing CHO cells that contained 50
copies of the DHFR gene. When DMA synthesis was blocked by
HU in the second hour of S phase, a large number of survivor
cells could form colonies in medium containing a 100-fold in
crease in MIX concentration. Few control cells or cells treated
with HU in other parts of the cell cycle could survive this MIX
selection. In these same experiments, Mariani and Schimke
found that the 10% of the genome replicated before the addition
of HU in the second hour of S phase was rereplicated after HU
was removed. These studies provided biochemical evidence that
part of the genome is rereplicated after DNA synthesis is tran
siently inhibited.
Our present studies have examined the effects of HU on
lymphoma cells to determine if DNA synthesis inhibition increases
the frequency of gene amplification and results in over-replication
of DNA. Our approach has been to use the FACS to measure
DNA content by measuring the amount of Hoechst dye bound
intracellularly. The data presented here show that Jsens cells
that survive DNA synthesis inhibition by HU have a high rate of
resistance to MTX due to an increased frequency of DHFR gene
amplification. HU-pretreated cells stained with Hoechst show a
significant portion of the population with increased fluorescence,
indicating increased DNA content. When metaphase chromo
somes of HU-treated cells with high Hoechst fluorescence were
examined, many of these cells showed extra chromosomes or
small, extrachromosomal DNA. The data from these studies and
from Mariani and Schimke (3) demonstrate an increased amount
of DNA by 2 independent experimental approaches. Studies by
Johnston ef a/.4 show in Chinese hamster ovary cells that critical
parameters in demonstrating an increase in DNA content as
measured by Hoechst staining are the length of HU exposure as
well as the population density of the exposed cells.
The results from our laboratory are consistent with the idea
that at least some forms of DNA synthesis inhibition result in
uncoordinated DNA synthesis. Mariani and Schimke (3) inter
preted their results to indicate that rereplication occurred in the
time frame of a single cell cycle, defined as the time from M to
M. An alternative (not mutually exclusive) hypothesis is that
treatment with HU results in a partial dissociation between S and
M phases, such that some cells progress into the period of
chromosome condensation, i.e., into M at a time when they are
in the first portion of a second S phase. Either hypothesis results
in cells with additional DNA per cell. After HU treatment, many
of the surviving cells have more DHFR enzyme and more DNA
per cell. If these cells are placed under drug selection, a high
proportion of cells with extra functional copies of the DHFR gene
will survive MTX selection and form a resistant colony. Other
overreplicated DNA presented after HU treatment will be highly
unstable and be lost in the absence of a specific selection.
An important aspect of these studies is the extent of cell death
induced by HU treatment. If Jsens cells are examined for viability
by trypan blue exclusion directly after HU is removed, few cells
are dead (data not shown). However, only 30% of these cells
4Johnston, ef al., manuscript in preparation.
CANCER
RESEARCH
ABERRATIONS
are able to form colonies in soft agar. When HU-treated cells are
sorted through the FACS, the cell viability as measured by colony
formation diminishes to 10% or less. A similar, but less extensive
decrease in viability after sorting was reported by Mariani and
Schmike in Chinese hamster ovary cells (3). The reason for such
a reduced viability is unknown, but we speculate that a combi
nation of mechanical fragility of cells and reduced viability as a
result of numerous chromosomal aberrations is involved. The
dyes used in FACS sorting are somewhat toxic and may further
lower plating efficiencies.
Chromosome analysis of HU-treated cells shows many types
of chromosome aberrations. The mechanism by which HU in
duces chromosome aberrations is not clear. HU slows down the
rate of DNA chain elongation (19). Even in the presence of high
HU concentrations, low levels of DNA synthesis continue (20).
DNA synthesized during HU treatment is abnormally small and
is ligated into large-molecular-weight DNA at a reduced rate.
Studies by Radford ef a/. (21) showed that the accumulation of
low-molecular-weight DNA was correlated with increased cytotoxicity in HU-treated cells. Radford et al. (21) proposed that
DNA synthesis inhibition is the key event in HU-induced cytotoxicity; DNA synthesis inhibition leads to chromosome lesions,
specifically chromosome breaks which are the ultimate cause of
cell death.
Other DNA synthesis inhibitors have effects similar to HU.
Rainaldi ef al. (22) showed that when aphidicolin, 1-/3-o-arabinofuranosylcytosine, and thymidine were used to inhibit DNA syn
thesis, a large increase in SCEs was seen if synchronized cells
were treated in the early part of S phase. Woodcock and Cooper
(23) showed that when 9-#-D-arabinofuranosyladenine,
1-/3-Darabinofuranosylcytosine, and cyclohexamide inhibited DNA syn
thesis, aberrant double replication of chromosomal segments
resulted. They proposed that double replication of chromosomal
DNA is a general consequence of freezing DNA replication forks
and not specific to the type of inhibitor. Studies by Tlsty ef al.
(2) showed that 3T6 cells pretreated with UV light or W-acetoxyA/-acetylaminofluorene had an increased frequency of MTX re
sistance and gene amplification. Treatment of Chinese hamster
ovary cells with UV light results in chromosomal aberrations
similar to those described here.5 Increases in MTX resistance
were seen when cells were pretreated by aphidicolin,6 and John
ston ef al* have shown that treatment of Chinese hamster cells
with aphidicolin results in increased DNA per cell. Collectively
such studies suggest that DNA synthesis inhibition produces
altered DNA replication that can later lead to specific gene
amplification under selection conditions, chromosome aberra
tions, and cell death.
A striking finding of our studies is that the subset of HU-treated
cells with additional DNA per cell is the same subset of cells with
chromosomal aberrations. Such cells cannot have simply under
gone arrest in M phase, followed by endoreduplication in a
second S phase, inasmuch as the additional DNA per cell is not
a multiple of the 2C content; furthermore the chromosomal
aberrations are of multiple types, i.e., fragmented chromosomes,
dicentric chromosomes, extrachromosomal DNA, and varying
degrees of polyploidy. In addition, we do not believe that the
increase in DNA content per cell results from fusion of cells with
fragmented chromosomes for 2 reasons, (a) Following resump
tion of DNA synthesis after HU treatment, DNA per cell as
determined by the FACS increases progressively over a 6- to
24-h period.7 If cell fusion was occurring, the number of cells
with increased DNA would have increased, not the amount of
8T. D. Tlsty, unpublished results.
6P. C. Brown, unpublishedresults.
7Johnston, ef a/., unpublished observation.
VOL. 45 OCTOBER
1985
5055
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
GENE AMPLIFICATION
AND CHROMOSOMAL
DNA per cell, (b) Coculturing of HU-treated L5178Y C3 cells
(which have more than 800 copies of the DHFR gene) and HUtreated CHO DHFR- cells (24) did not result in any (frequency
< 7 x 1Q-8) CHO cells converting to DHFR+.8 This is an
extremely sensitive method for detecting uptake and function of
DNA (25).
Our current hypothesis is that overreplication of DNA is a prior
event in which various modes of recombination result in observed
chromosomal aberrations. We speculate that DHFR gene ampli
fication and chromosome aberrations result from overreplication
of DNA after HU treatments. Thus, to the extent that various
agents used in cancer chemotherapy kill cells by chromosome
fragmentation, the very same molecular process results in
emergence of drug resistance by gene amplification.
ACKNOWLEDGMENTS
We would like to thank Elise Mosse for expert technical help, Randy Johnston
for help with the FACS sorting, and Randy Johnston, Peter Brown, and Thea Tlsty
for many fruitful discussions. We wish to thank Steven Sherwood and David
Smousefor critically reading the manuscript and also providing helpful discussions.
REFERENCES
1. Brown, P. C., Tlsty, T. D., and Schimke, R. T. Enhancementof methotrexate
resistance and dihydrofolate reducÃ-asegene amplification by treatment of
mouse 3T6 cells with hydroxyurea. Mol. Cell. Biol., 3: 1097-1107,1983.
2. Tlsty, T. D., Brown, P. C., and Schimke, R. T. UV radiation facilitates metho
trexate resistance and amplification of the dihydrofolate reducÃ-asegene in
cultured 3T6 mouse cells. Mol. Cell. Biol., 4: 1050-1056, 1984.
3. Mariani,
D., and
Schimke,
R.Biol.
T. Gene
amplification
in a single cell cycle
in
Chinese B.
hamster
ovary
cells. J.
Chem.,
259:1901-1910,1984.
'
4. Timson, J. Hydroxyurea. Mutât.Res., 32: 115-132, 1977.
5. Kaung, D. T., and Swartzendruber, A. A. Effect of chemotherapeutic agents
on chromosomes of patients with lung cancer. Dis. Chest, 55: 98-100, 1969.
6. Oppenheim, J. J., and Fishbein, W. tf. Introduction of chromosome breaks in
cultured normal human leukocytes by potassium arsenite, hydroxyurea, and
related compounds. Cancer Res., 25: 980-985,1965.
7. Dolnick, B. J., Berenson, R. J., Bertino, J. R., Kaufman, R. J., Nunberg, J. H.,
and Schimke, R. T. Correlation of dihydrofolate reducÃ-ase
elevation with gene
amplification in a homogeneously staining chromosomal region in L5178Y
cells. J. Cell. Biol., S3: 394-402, 1979.
8A. Hill, unpublished observation.
CANCER
RESEARCH
ABERRATIONS
8. Hill, A., and Wolff, S. Increased induction of sister chromatid exchange by
dietnylstilbestrol in lymphocytes from pregnant and premenopausalwomen
Cancer Res., 42. 893-896,1982.
9. Johnston, R. N., Bevertey, S. M., and Schimke, R. T. Rapid spontaneous
dihydrofolate reducÃ-asegene amplificaron shown by fluorescence-aclivaled
cell sorting. Proc. Nail. Acad. Sci. USA, 80: 3711-3715 1983
10 Arndt-Jovin, D. J., and Jovin, T. M. Analysisand sorting of livingcells according
to deoxyribonucleic acid content J. Hislochem. Cytochem., 25: 585-589,
1977.
11 Chaira, A. C. Y., Nunberg, J. H., Kaufman,R. J., Ertich,H. A., Schimke R T
and Cohen, S. N. Phenotypic expression in E. coli of a DNA sequencecoding
for mouse dihydrofolate reducÃ-ase.
Nalure (Lond.),275: 617-624,1978.
12. Law, S., Tamoaki, T., Krevzaler, F., and Dugaiczyk, A. Molecular cloning of
DNA complemenfary lo mouse alphafeloprolein mRNA sequence Gene ro53-61,1980.
13. Brown, P. C., Beverley, S. M., and Schimke, R. T. Relalionship of amplified
dihydrofolate reducÃ-asegenes to double minute chromosomes in unstably
resistant mouse fibroblast cell lines. Mol. Cell. Biol., 7:1077-1083 1981
14 Sirolnak, F. M., Moccio, D. M., Kelleher, L. E., and Goûtas,L. J. Relative
frequency and kinetic properties of transport-defective phenotypes among
methotrexale-resislanl L1210 clonal cell lines derived in vivo. Cancer Res'
15. Flintoff, W. F., Davidson, S. V., and Siminovitch, L. Isolation and partial
characterization of three melholrexale-resislanl phenolypes from Chinese
hamster ovary cells. Somatic Cell Genet., 2: 245-261,1976.
16. Haber, D. A., and Schimke, R. T. Unstableamplificationof an altered dihydro
folate reducÃ-ase
gene associated with double-minutechromosomes Cell 26355-362,1981.
17. Alt, F. W., Kellems, R. E., Bertino, J. R., and Schimke, R. T. Selective
multiplication of dihydrofolate reducÃ-ase
genes in melholrexate-resistanl var
iants of cultured murine cells. J. Biol. Chem., 253:1357-1370,1978.
18. Kaufman,R. J., Bertino, J. R., and Schimke,R. T. Quantitationof dihydrofolate
reducÃ-asein individual parenlal and methotrexate-resistant murine cells- use
of a fluorescence activated cell sorter. J. Biol. Chem.,253:5852-5860 1978
19. Ramseier,H. P., Burkhalter, M., and Gautschi,J. R. Survival of CHO cells thai
replicated DNA in the presenceof hydroxyurea. Exp. Cell Res., 105:445-453,
i y //.
20 Coyle, M. B., and Strauss, B. Cell killing and the accumulata! of breaks in the
DNA of HEP-2 cells incubated in the presence of hydroxyurea. Cancer Res
30:2314-2319,1970.
Radford, I. R., Martin, R. F., Finch, L. R., and Hodgson, G. S. Inhibitionof DNA
synthesis and cell dealh. Biochim. Biophys. Acia, 696: 154-162,1982.
Rainaldi, G., Sessa, M. R., and Mariani, t. Inhibitors of DNA synthesis induce
sister chromatid exchanges at Ihe early S phaseof Ihe cell cycle. Chromosoma
(Beri.), 90:46-49, 1984.
23 Woodcock, D. M., and Cooper, I. A. Evidence for double replication of chro
mosomal DNA segmenls as a general consequenceof DNA replication inhibi
tion. Cancer Res., 41: 2483-2490,1981.
24. Urlaub,G., and Chasin, L. A. Isolationof Chinesehamstercell mutants deficient
in dihydrofolate reducÃ-ase
activity. Proc. Nati. Acad. Sci. USA, 77:4216-4220
1980.
25. McGrogan, M., Simonsen,C. C., Smouse, D. T., Farnham,P. J., and Schimke
R. T. Heterogeneityof Ihe 5'-lermini of mousedihydrofolatereducÃ-ase
mRNAs
J. Biol. Chem., 260: 2307-2314,1985.
VOL. 45 OCTOBER
1985
5056
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
GENE AMPLIFICATION
AND CHROMOSOMAL
ABERRATIONS
1
*
«-
B
-».
<»-
''^
..
V
' •"•
J
, ~
Fig. 1. Chromosomal aberrations induced by hydroxyurea. Metaphases were examined 24 h after the removalof HU, and differentiallystained sister chromatids were
examined for SCEs. A, polyploid cell with many SCEs. B, metaphase with normal chromosomes and a large amount of small extrachromosomal DNA (arrow). C,
metaphase with multiple chromosome gaps and breaks. D, an endoreduplicated metaphase with a large number of SCEs. £,metaphase with a dicentric chromosome
(arrow). F, a highly fragmented metaphase (note break at arrow) showing a very high frequency of SCE.
CANCER
RESEARCH
VOL. 45 OCTOBER
1985
5057
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
Increased Gene Amplification in L5178Y Mouse Lymphoma
Cells with Hydroxyurea-induced Chromosomal Aberrations
Anna B. Hill and Robert T. Schimke
Cancer Res 1985;45:5050-5057.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/45/10/5050
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 [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.