Investigative Ophthalmology & Visual Science, Vol. 31, No. 12, December 1990
Copyright © Association for Research in Vision and Ophthalmology
Evaluation of Antiproliferative Agents
Using a Cell-Culture Model
Richard I. Senderoff, Paul A. Weber, D. Ronald Smith, and Theodore D. Sokoloski
Chinese hamster ovarian (CHO) cells in culture were used to evaluate the relative antiproliferative
potential of drugs. These agents have been used to improve the clinical response after glaucoma
filtering surgery. The following drugs were evaluated: S-fluorouracil (5-FU) as the benchmark, 5fluorouridine (FUR), 5-fluorodeoxyuridine (FUDR), 5'-deoxy-5-fluorouridine (DFUR), bleomycin,
and cytarabine (ARA-C). In addition to cell counting, a colorimetric assay based on the tetrazolium
salt, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was used to follow growth
response. The MTT assay was found to be extremely convenient and an indirect measure of cell
activity, offering an alternate or addition to a measure of cell number. All agents tested were shown to
inhibit cellular proliferation. Dose-response curves for each agent indicate the following absolute
potency: FUDR > FUR > ARA-C > 5-FU = bleomycin > DFUR. Besides absolute potency, an
evaluation of the effects of equivalent inhibitory concentrations of each drug on growth rate was
assessed. Several agents affected the proliferation rate patterns differently. Based on these studies, it
is suggested that the in vitro model can identify potential agents through an assessment of their overall
activity profile in CHO cells, which includes not only their potency based on dose response, but their
onset of activity, duration of effect, and potential for toxicity. Invest Ophthalmol Vis Sci 31:25722578,1990
Several different approaches have been used to improve the success rates in glaucoma filtering surgery
(GFS) in patients with a poor prognosis. Perhaps the
most promising area of research involves pharmacologic alteration of the cellular response in scar formation. Although investigation of the effectiveness of
corticosteroids,1'2 D-penicillamine,23 and 0-aminopropionitrile3 as adjuncts to GFS has been reported,
most research at present is focused on the evaluation
of antiproliferative agents. The agent most closely
studied is 5-fluorouracil (5-FU), which acts primarily
as an S-phase specific agent.4 Recent studies show
that subconjunctival injections of 5-FU promote bleb
longevity in animals5 and improve the success of filtering operations in patients with poor surgical prognoses.6"9 The proposed mechanism of action responsible for the increased success rates is inhibition of
fibroblast proliferation and consequent attenuation
of postoperative scar formation. Despite encouraging
results, subconjunctival injections of 5-FU are associated with numerous disadvantages. Under current
treatment recommendations, patients receive once or
twice daily subconjunctival injections over 2 weeks.
Not only is this frequency of administration inconve-
From the Department of Ophthalmology and College of Pharmacy, Ohio State University, Columbus, Ohio.
Supported by the Ohio Lions Eye Research Foundation.
Submitted for publication: November 9, 1989; accepted May 15,
1990.
Reprint requests: P. A. Weber, Department of Ophthalmology,
The Ohio State University Hospitals, 5116 Hospital Clinics, 456
West 10th Avenue, Columbus, OH 43210.
nient, but the patient also is uncomfortable with each
injection. In addition, the injections are associated
with corneal epithelial defects in approximately 45%
of cases and wound or needle tract leaks in approximately 41% of cases.610
Approaches to improve the clinical response to
treatment with antiproliferative agents include more
effective agents and better ways to deliver them in
terms of dosing schedules and novel drug-delivery
systems. In exploring these possibilities for application to GFS, an in vitro model was sought to screen
drug candidates with respect to physiologic properties
to provide important information in identifying
more efficient compounds and appropriate protocols
to be used in their clinical evaluation. Also, since the
design of a drug-delivery system is dependent on its
biologic profile, knowledge of a drug's activity in
terms of onset of activity, duration of action, and
potential for toxicity is critical.
We focused on the development of an in vitro
model, based on Chinese hamster ovarian (CHO)
cells grown in culture to assess the cellular response to
antiproliferative agents. These cells were chosen because they are easy to grow in a monolayer and represent a homogeneous, rapidly proliferating, continuous cell line which is commercially available. Their
use thus could lead to a highly reproducible protocol,
which is essential in comparative drug screening. Our
purpose was not to simulate in vivo fibroblast overproliferation, but to develop an in vitro model which
identifies compounds as having antiproliferative
properties and provides additional information with
which to compare several agents. The antiprolifera-
2572
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933380/ on 07/28/2017
No. 12
EVALUATION OF AGENTS FOR GLAUCOMA FILTRATION SURGERY / Senderoff er ol
tive compounds examined included: 5-FU as the
benchmark for comparison, 5-fluorouridine (FUR),
5-fluorodeoxyuridine (FUDR), 5'-deoxy-5-fluorouridine (DFUR), bleomycin sulfate, and cytarabine
(ARA-C). The compounds are believed to have different modes of action which offered the possibility
for differentiation using the method described: 5-FU
and its metabolites inhibit DNA synthesis through
inhibition of thymidylate synthetase, bleomycin sulfate causes scission of single- and double-stranded
DNA, and ARA-C is converted to ARA-C triphosphate, a potent inhibitor of DNA polymerase.
Materials and Methods
The Ohio State University Cell Culture Service
provided CHO, Eagle's minimal essential medium,
bovine calf serum, penicillin, streptomycin, amphotericin B, trypsin, 3-(4,5-dimethylthiazol-2-yl)2,5-diphenyl^trazolium bromide (MTT), Isotone
(Coulter Electronics Inc., Hialeah, FL) and spectrophotometric-grade dimethylsulfoxide (DMSO). The
5-FU was purchased from Sigma (St. Louis, MO),
ARA-C from Quad Pharmaceuticals (Indianapolis,
IN), and bleomycin sulfate from Mead Johnson Oncology Products (Bethel Park, PA). The 5-FU,
FUDR, and DUFR were provided by Hoffman-LaRoche (Nutley, NJ). The analytic equipment consisted of a Coulter counter (Model ZBI, Coulter) and
a multiwell scanning spectrophotometer (Model
MR700; Dynatech Laboratories, Chantilly, VA).
The CHO cells were maintained by incubation at
37°C in 90% humidity and 5% CO2. The cells were
grown in a modified Eagle's minimal essential medium that was supplemented with 10% bovine calf
serum. Penicillin, streptomycin, and amphotericin B
were added to protect against microbial contamination.
Correlation of Assay Techniques
Cell response was determined with a Coulter
counter and by colorimetric assay using MTT.
The following procedures were used. The CHO
cells were removed enzymatically from T25 flasks
using a 15-min incubation with 0.01% trypsin. Next
they were diluted to appropriate concentrations using
a Coulter counter and then plated into 24-well tissueculture plates. Growth curves were generated to establish appropriate inoculation densities to achieve
cell counts ranging from 10,000-300,000 cells/well
after a 24-hr incubation. The experimental design
consisted of using eight wells per inoculation concentration: four wells were analyzed with the Coulter
counter and four wells by the MTT assay. The
Coulter counter assay involved: (1) rinsing the cell
wells with phosphate-buffered saline (PBS), (2) disso-
2573
ciating the cells by adding 200 yX of 0.01% trypsin to
each well and incubating for 15 min, and (3) adding
0.8 ml of PBS to each well, making appropriate dilutions with Isotone and analyzing the samples on the
Coulter counter. The MTT assay procedure involved:
(1) adding 100 /A of a MTT stock solution (5 mg
MTT/ml PBS) to each well, (2) removing all the
media from the wells after a 4-hr incubation at 37°C,
(3) adding 0.5 ml of DMSO to each well, and (4)
transferring 100 fi\ of the solution from each well to a
96-well plate and reading on a multiwell scanning
spectrophotometer (X = 570 nm). It was noted that a
minimum of 10-min exposure to DMSO was required to dissolve the MTT formazin crystals after
which the absorbance remained constant for 20 min.
Thus, in all experiments spectrophotometric readings
were taken 15 min after the addition of DMSO.
Dose-Response Studies
The CHO cells were removed enzymatically from
T25 flasks using 0.01% trypsin. They were diluted to
3750 cell/ml media using a Coulter counter and then
plated in 24-well tissue-culture plates. The inoculation density was 3750 cells/well. After allowing 24 hr
for attachment, the agents under study were added to
the wells at various concentrations. Each plate was
then incubated for an additional 120 hr. Parallel experiments were done using both the Coulter counter
and MTT assays. Cells in plain media were used as
controls. All experiments were done in triplicate. The
dose-response data for each individual run (n = 3)
was used to determine a 50% inhibitory concentration (IC50) as the concentration at the intersection of
50% of control with the slope of steepest descent determined by linear regression. Differences in relative
potencies were evaluated using Tukey's studentized
range test after transforming the data by taking the
natural logarithms to equalize the variances.
Drug Effect over Time
The CHO cells were removed enzymatically from
T25 flasks using 0.01% trypsin and then diluted to
7500 cells/ml using a Coulter counter. One ml of the
cell suspension was added to each well of a 24-well
tissue-culture plate. After incubating for 24 hr at
37°C, the medium was replaced in designated test
wells with an antiproliferative agent at its 90% inhibitory concentration (IC90). In drug treatment studies,
two experimental conditions were used. In one subset, controls were compared with cells run in the continuous presence of drug at its IC90. In another, wells
initially treated with drug were flushed at 72 hr using
two rinses of PBS; the PBS was subsequently replaced
with plain media. The media in all wells was replaced
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933380/ on 07/28/2017
2574
Vol. 31
INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / December 1990
every 24 hr. Four control wells and four test wells
were assayed by both the MTT assay and the Coulter
counter assay every 12-24 hr depending on the
growth rate. Total incubation periods ranged from 84
(control wells)-288 hr (FUR wells) depending on the
antiproliferative agent. The six agents listed previously were used in both experimental designs.
Onset of activity was assessed by estimating the time
at which the curve for the drug-treated system deviated from the control curve. The duration of effect
was estimated by noting the time at which the slope
of the proliferation curve increased after removal of
the drug. A potential for toxicity was judged through
a determination of the point at which the slope of the
proliferation curve decreased after continuous drug
treatment.
Results
Correlation of Assay Techniques
Figure 1 summarizes the results that compare the
Coulter counter (x-axis) and MTT (y-axis) methods
of assessing cell population. Each point represents
one of 12 different initial CHO cell inoculations that
resulted in cell counts ranging from 10,000-300,000
cells/well after 24 hr. At each population density or
data point represented, eight wells were analyzed:
four by Coulter counting and four by the MTT procedure. The length of the bars around the average in
the x- and y-directions represents the error in the
respective methods.
Dose-Response Studies
Cell count or MTT absorbance in the presence of
varying amounts of the six drugs studied relative to
controls without drug was determined and plotted as
a function of the logarithm of the drug concentration
added. Figure 2 gives the results. Solid lines are the
results from the Coulter counter and broken lines
those from the MTT assay. The error bars represent
the error about the mean percent of control for four
replicates. Table 1 lists the average IC50 values for the
three individual runs and the standard error found.
Drug Effect Over Time
Our results found in the studies assessing drug effects over time are shown in Figure 3 in which MTT
optical densities of two experimental groups (continuous drug and drug flushed at 72 hr) is compared
with the control group (no drug). In each panel (representing a different drug), the upper curve represents
the control system where no drug was added, the
middle curve represents the system where drug was
added to the culture at its IC90 concentration but then
o
50
100
150
200
CELL NUMBER X 10E-3
250
300
Fig. 1. The relationship between absorbance as measured with
the MTT formazan method to determine cell number and the cell
number determined using a Coulter counter.
flushed and replaced with drug-free medium, and the
lower curve represents systems where the cells are
exposed to drug at its IC90 concentration for the entire time. These plots were compared in a semiquantitative way to identify trends in the time-action profiles and used to differentiate the agents. Significant
differences were not observed in onset of activity between the agents, and all agents were active within 24
hr of drug treatment. The duration of effect for FUR,
bleomycin sulfate, ARA-C, 5-FU, and FUDR was
192, 144, 96, 72, and 72 hr, respectively, as reflected
by the time when cell response began to increase after
drug removal. The estimations of potential for toxicity were 120, 120, 144, 144, and 168 hr for FUR,
ARA-C, bleomycin sulfate, FUDR, and 5-FU, respectively, determined as the point at which the slope
of the curve decreased with continuous treatment.
No decrease in slope was observed for DFUR after
continuous drug treatment.
Discussion
Correlation of Assay Techniques
Until reduced by dehydrogenase enzymes in the
mitochondria of living cells, MTT is a yellow substrate. Once reduced, a purple compound (MTT formazan) is formed which is insoluble in media but can
be dissolved in DMSO and quantified spectrophotometrically. This method for assaying viable cells has
been shown to be much more rapid and equally as
accurate as other methods such as the hemocytometer, the Coulter counter, incorporation of radioactive nucleotides, and 51Cr labeling of cell proteins.
The amount of MTT formazan produced (as the reduced form) is directly proportional to the number of
living cells, with many cell lines.''
In this study, a linear relationship between cell
number (as determined by Coulter counter) and cell
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933380/ on 07/28/2017
EVALUATION OF AGENTS FOR GLAUCOMA FILTRATION SURGERY / Senderoff er ol
No. 12
101010"'
10°
10'
5-FU CONCENTRATION. UG/ML
10
1
100 -
c
1
1 1
1II
FT,
=f
I-T-T-TTTT
\
o
u
ft? 25 t-
M
'
10
10-
10101010"
FUR CONCENTRATION. UG/ML
10°
10
1010"'
10"
10'
10*
DFUR CONCENTRATION. UG/ML
10 s
10"4
101010"
10°
ARA-C CONCENTRATION. UG/ML
10'
;
T
50 -
T"
2575
-
\
I
V
V
I
\
\
\
\
X
\
1
1
1
111
10101010"
FUDR CONCENTRATION. UG/ML
10-
10°
1010"'
10°
10'
10*
BLEOMYCIN CONCENTRATION. U/L
Fig. 2. The dose-response relationships found for the several antiproliferative agents. The solid line denotes the relative number of cells
(response) as determined through the Coulter counter. The stippled line is the result found where relative cell number is determined via the
MTT formazan method. The 50% inhibitory concentration is the midpoint of the sigmoid curve.
activity (as determined by the MTT method) was
found for CHO cells in the range studied (Fig. 1).
With cell numbers greater than 300,000 (resulting in
optical densities greater than 1.2), the results were
erratic, and the absorbance was no longer proportional to the number of cells. The error associated
with each method was similar as evidenced through
the error bars (Fig. 1) and standard deviation (Table
1). The linearly correlated counting and activity
measures may have been fortuitous. The MTT
method quantitates cell number by the measured
amount of enzyme contained in the cell spectrophotometrically, which may give an indirect measure of
metabolism and cell viability and may be a more
significant assessment of potency than cell number
for a group of compounds like the antiproliferatives.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933380/ on 07/28/2017
2576
INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / December 1990
Table 1. Fifty percent inhibitory concentrations of
several antiproliferative drug candidates
Drug*
Coulter counter]
MTT%
DFUR
Bleomycin
5-FU
ARA-C
FUR
FUDR
34.5 (9.66)
7.17 (1.98)
4.71 (0.510)
0.560 (0.0404)
0.0614(0.0139)
0.0205(0.00316)
83.8 (4.00)
10.1 (0.888)
11.7 (1.44)
0.722 (0.0337)
0.130 (0.0371)
0.0577 (0.00359)
* See text for abbreviations used.
t Cell response was estimated using the Coulter counter method. Values
represent cone. X 107 M (±SD X 107).
% Cell response was estimated using the MTT method. Values represent
cone. X 107 M (±SD X I0 7 ).
This possibility can only be proved in clinical studies.
The MTT assay offers other distinct advantages over
the Coulter counter method: (1) only viable cells are
measured, (2) no washing steps which increase sample variation and processing time are required, (3) the
cells do not need to be dissociated from the culture
plate, and (4) a multiwell scanning spectrophotometer can be used which measures a large number of
wells with a high degree of precision at virtually identical time points. Compared with Coulter counter
measures, the MTT assay did tend to give larger estimates of apparent cell population after drug treatment, but the small differences did not affect the interpretation of either dose response or proliferation
rate data.
Dose-Response Studies
Tukey's studentized range test was used to verify
the differences in relative potency based on the IC5o
determinations. Bleomycin sulfate and 5-FU were
found to be equipotent; all other agents tested had
different potencies at a level of significance of 0.05.
The rank order found was: FUDR > FUR > ARA-C
> 5-FU = bleomycin > DFUR. The statistical results
were identical for IC5o determinations using both the
MTT and Coulter counter assay techniques.
Drug Effect Over Time
Although dose-response curves for numerous antiproliferative agents have been generated in our laboratories and those of others,12"14 there are no published reports to our knowledge that use cell-culture
techniques to examine the effects of antiproliferative
agents over time. This type of analysis can reveal
relative differences in onset of activity, duration of
effect (reversibility), and potential for toxicity. Each
of the six agents studied had its own unique set of cell
response versus time curves after treatment with
Vol. 31
equivalent doses (IC90). Although differences in onset
of action were not found, analysis of the results revealed differences in activity in terms of potential
reversibility of cell growth inhibition after drug removal and the propensity to apparent toxicity after
continuous cell treatment. Based on the estimates of
the duration of effect, FUR and bleomycin sulfate
produce a cellular inhibition that is twice as long as
5-FU and FUDR, whereas 5-FU and FUDR produce
the same degree of inhibition. Although decreasing
cell numbers with time after continuous drug treatment may have been due to secondary factors such as
nutritional deficiency in the media or conversion of
the drug to a more potent metabolite, it seemed reasonable to assume that this represented toxicity and
that the relative toxicity may be assessed by noting
how soon the growth curve slope became negative.
Judging toxicity potential in this way, DFUR was the
only agent tested that might be relatively free of toxicity. The other agents showed a potential for toxicity
in the order listed (low to high): 5-FU > FUDR
> bleomycin sulfate > ARA-C > FUR. These findings are semiquantitative and subjective, but they indicate that equivalent doses of antiproliferative
agents do not produce equivalent responses in terms
of their activity over time.
Summary
Our procedures represent a simple, reproducible
method to evaluate drug activity with rapidly proliferating cells. The reproducibility of the model
allows experiment-to-experiment and interlaboratory
comparisons of activity. This protocol can be used to
screen a series of compounds with respect to differences in their relative activity. The method shows
that the overall activity profile of an antiproliferative
agent should include not only its potency based on
dose response, but its onset of activity, duration of
effect, and potential for toxicity determined by timecourse studies. The differences noted can be applied
to the design of clinical studies by providing equivalent doses and dosing schedules. The studies emphasize that antiproliferative potential based on equivalent doses derived from absolute potency does not
necessarily confer therapeutic equivalence. The regimen for each drug also should be optimized before
comparisons of clinical response can be made. When
considering a controlled drug-delivery system, the
significance of these differing properties is even
greater, since ultimately it is the absolute potency,
duration of effect, and potential for toxicity of the
active agent which determines the required delivery
rate.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933380/ on 07/28/2017
No. 12
0
EVALUATION OF AGENTS FOR GLAUCOMA FILTRATION SURGERY / Senderoff er ol
1 2 3 4
0 1 2
0
5 6 7 8 9 10 11 12 13 14 15
TIME (DAYS)
3 4 5 6 7 8 9 10 11 12 13 14 15
TIME (DAYS)
1 2 3 4
5 6 7 8 9 10 11 12 13 14 15
TIME (DAYS)
2577
0.0
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
TIME (DAYS)
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
TIME (DAYS)
0
1 2 3 4
5 6 7 8 9 10 11 12 13 14 15
TIME (DAYS)
Fig. 3. The relationship between absorbance (cell response) and time under different experimental conditions. In each panel the upper curve
represents the control where no drug is added, the middle curve represents the system where drug is added to the culture at its IC90
concentration but thenflushedand replaced with drug-free medium, and the lower curve refers to systems where the cells are exposed to drug
at its IC90 concentration over the entire time course. Each panel is a different drug: A, 5-FU; B, FUR; C, FUDR: D, bleomycin; E ARA-C F
DFUR.
'
'
' '
Key words: antiproliferatives, in vitro model, cell culture,
CHO cells, potency, growth rate
References
1. Giangiacomo J, Dueker DK, and Adelstein E: The effect of
preoperative subconjunctival triamcinolone administration on
glaucoma filtration. Arch Ophthalmol 104:838, 1986.
2. McGuigan LJB, Cook DJ, and Yablonski ME: Dexametha-
sone, d-penicillamine, and glaucoma filter surgery in rabbits.
Invest Ophthalmol Vis Sci 27:1755, 1986.
3. McGuigan LJB, Mason RP, Sanchez R, and Quigley HA: Dpenicillamine and jS-aminopropionitrile effects on experimentalfilteringsurgery. Invest Ophthalmol Vis Sci 28:1625, 1987.
4. Madoc-Jones H and Bruce WR: On the mechanism of the
lethal action of 5-fluorouracil on mouse L cells. Cancer Res
28:1976, 1968.
5. Gressel MG, Parrish RK, and Folberg R: 5-Fluorouracil and
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933380/ on 07/28/2017
2578
6.
7.
8.
9.
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / December 1990
glaucoma filtering surgery: I. An animal model. Ophthalmology 91:378, 1984.
Hever DK, Parrish RK, Gressel MG, Hever DK, Parrish RK,
Gressel MG, Hodapp E, Palmberg PF, and Anderson DR:
5-Fluorouracil and glaucomafilteringsurgery: II. A pilot study.
Ophthalmology 91:384, 1984.
Hever DK, Parrish RK, Gressel MG, Hever DK, Parrish PK,
Hever DK, Skuta GL, Hodapp E, Palmberg PF, Gressel MG,
and Fever W: 5-Fluorouracil and glaucoma filtering surgery:
III. Intermediate follow-up of a pilot study. Ophthalmology
93:1537, 1986.
Rockwood EJ, Parrish RK, Hever RK, Rockwood EJ, Parrish
RK, Hever DK, Skuta GL, Hodapp E, Palmberg PF, Gressel
MG, and Fever W: Glaucoma filtering surgery with 5-fluorouracil. Ophthalmology 94:1071, 1987.
Ruderman JM, Welch DB, Smith MF, and Shouh DE: A randomized study of 5-fluorouracil and filtration surgery. Am J
Ophthalmol 104:218, 1987.
Vol. 31
10. Lee DA, Hersh P, Kersten D, and Melamed S: Complications
of subconjunctival 5-fluorouracil following glaucoma filtering
surgery. Ophthalmic Surg 18:187, 1987.
11. Mosmann T: Rapid colorimetric assay for cellular growth and
survival: Application to proliferation and cytotoxicity assays. J
Immunol Methods 65:55, 1983.
12. Blumenkranz MS, Claflin A, and Hajek AS: Selection of therapeutic agents for intraocular proliferative disease: Cell culture
evaluation. Arch Ophthalmol 102:598, 1984.
13. Mallick KS, Hajek AS, and Parrish RK: Fluorouracil (5-FU)
and cytarabine (ARA-C) inhibition of corneal epithelial cell
and conjunctival fibroblast proliferation. Arch Ophthalmol
103:1398, 1985.
14. Blumenkranz MS, Hartzer MK, and Hajek AS: Selection of
therapeutic agents for intraocular proliferative disease: II. Differing antiproliferative activity of thefluoropyrimidines.Arch
Ophthalmol 105:396, 1987.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933380/ on 07/28/2017