Novel Dual-Reporter Preclinical Screen for Antiastrocytoma
Agents Identifies Cytostatic and Cytotoxic Compounds
JESSICA J. HAWES, JOHN D. NERVA, and KARLYNE M. REILLY
Astrocytoma/glioblastoma is the most common malignant form of brain cancer and is often unresponsive to current phar-
macological therapies and surgical interventions. Despite several potential therapeutic agents against astrocytoma and
glioblastoma, there are currently no effective therapies for astrocytoma, creating a great need for the identification of effec-
tive antitumor agents. The authors have developed a novel dual-reporter system in Trp53/Nf1-null astrocytoma cells to simul-
taneously and rapidly assay cell viability and cell cycle progression as evidenced by activity of the human E2F1 promoter in
vitro. The dual-reporter high-throughput assay was used to screen experimental therapeutics for activity in Trp53/Nf1-null
astrocytoma. Several compounds were identified demonstrating selectivity for astrocytoma over primary astrocytes. The
dual-reporter system described here may be a valuable tool for identifying potential antitumor treatments that specifically tar-
get astrocytoma. (Journal of Biomolecular Screening XXXX:xx-xx)
astrocytoma, Nf1, p53, E2F1, luciferase
© 2008 Society for Biomolecular Scienceswww.sbsonline.org1
mon malignant form of brain cancer and are often unresponsive
to surgical intervention and current pharmacological therapy.
The best current treatment option is surgical removal; however,
astrocytomas/GBM diffusely infiltrate the central nervous system,
rendering resection difficult or impossible, leading to poor patient
prognosis.1,2Consequently, the 5-year survival rate for GBM is
less than 5%.3Although a finite list of pharmacological agents
have been reported as potential therapeutic agents against astro-
cytoma and GBM,4the cure rate is still very low, demonstrating
the tremendous need for identifying more effective antitumor
One of the earliest and most common genetic alterations in
astrocytoma is loss of heterozygosity on chromosome 17p.5-7This
chromosomal region encompasses the p53 gene, which plays a
primary role in the progression of multiple types of tumors,
including astrocytoma. Nearly 50% of astrocytomas include loss
of heterozygosity at 17p and/or mutation of p53, with up to
90% of GBMs displaying alterations in the p53 pathway.8The
STROCYTIC GLIOMAS, INCLUDING ASTROCYTOMAS AND
GLIOBLASTOMA MULTIFORME (GBM), are the most com-
p53 protein serves as a key component of the cell cycle check-
point by halting proliferation in response to DNA damage and as
a transcription factor that induces genes responsible for growth
arrest and apoptosis.9Furthermore, tumors known to contain
mutations in p53 are more resistant to radiation, and many antitu-
mor agents are inactive in p53-null tumors, making these tumors
also resistant to chemotherapy.10
The familial cancer syndrome neurofibromatosis type 1 (NF1)
is an autosomal dominant syndrome that predisposes individuals
to developing multiple tumors, including astrocytoma and
GBM.11NF1 patients carry a mutation in the NF1 gene (the Nf1
gene in mice) that encodes for the protein neurofibromin.
Neurofibromin is a tumor suppressor rasGAP protein that down-
regulates the ras signaling pathway linking growth factor signals
to cellular proliferation.12Consequently, loss of neurofibromin
plays a key role in the induction of tumorigenesis,leading to over-
activation of the oncogenic ras pathway.
Because many human astrocytomas contain mutations in Tp53
(Trp53 in mice), encoding the p53 protein, as well as upregulation
of ras signaling that is critical for astrocytoma tumorigenesis and
maintenance, preclinical models that reflect these alterations may
be ideal for characterizing and identifying potential astrocytoma
therapeutics. Mice carrying mutations in Nf1 and Trp53 on the
same chromosome (Nf1–/+;Trp53–/+cis: NPcis) have been char-
acterized as a mouse model of NF113,14and astrocytoma.14,15NPcis
mice undergo spontaneous loss of heterozygosity at the wild-type
copies of Nf1 and Trp53, resulting in the development of brain
tumors with high penetrance and close similarity to human astro-
cytomas.15NPcis brain tumors range from low-grade astro-
cytomas to high-grade GBMs, forming diffusely infiltrative
Mouse Cancer Genetics Program, National Cancer Institute, Frederick, Maryland.
Received Jan 12, 2008, and in revised form May 30, 2008. Accepted for publi-
cation Jun 1, 2008.
Journal of Biomolecular Screening XX(X); XXXX
J Biomol Screen OnlineFirst, published on July 29, 2008 as doi:10.1177/1087057108321085
Copyright 2008 by Society for Biomolecular Sciences.
tumors.13,15Primary tumor cells isolated from NPcis astrocytomas
show loss of the wild-type copies of Nf1 and Trp53 and maintain
tumor cell characteristics similar to human astrocytoma in vitro.15
Thus, NPcis astrocytoma cells can be used to build an in vitro
assay for identifying novel antiastrocytoma therapeutic candidates.
KR158 tumor cells from a grade III NPcis aggressive
anaplastic astrocytoma15were used to generate a green and red
luciferase (G/R-luc) dual-reporter system that simultaneously
assesses activity of the human E2F1 promoter and cellular
cytotoxicity in a high-throughput assay. The G/R-luc dual-
reporter system was used to screen chemically diverse com-
pounds to identify agents with antiproliferative activity in
astrocytoma cells. This system distinguishes cytostatic com-
pounds from cytotoxic agents during the initial screening, dis-
criminating cytotoxic agents from inhibitors of proliferation.
Thus, the G/R-luc dual-reporter system could significantly
decrease the time and cost required to screen compound
libraries. This system was also used to examine the pharmacol-
ogy of identified antitumor agents. The G/R-luc dual-reporter
system is a valuable tool in the identification and characteriza-
tion of potential antitumor treatments specifically targeting
G/R-luc cell line
For construction of the pEf-CBGluc plasmid expressing the
green luciferase gene under control of the human E2F1 promoter,
green click beetle luciferase from pCBG681uc (Promega,
Madison, WI) was subcloned in place of the firefly luciferase
gene into pEf-luc (gift from Dr. Eric Holland).16The hygromycin
resistance gene was PCR cloned into pEf-CBGlu upstream of the
E2F1 promoter for clonal selection. pHygro-Ef-CBGluc was sta-
bly transfected into grade III KR158 astrocytoma cells15using
Fugene (Roche Applied Science, Indianapolis, IN) to generate G-
luc astrocytoma cells. For construction of pCMV-CBRluc, the
cytomegalovirus (CMV) promoter was cloned in place of the
SV40 promoter into pCBR-Basic (Promega), which contains
the modified click beetle red-emitting luciferase. The PKG pro-
moter and puromycin resistance pac gene were cloned upstream
of the CMV promoter for clonal selection. pPuro-CMV-CBRluc
was stably transfected into G-luc astrocytoma cells to generate the
G/R-luc astrocytoma dual-reporter cell line. All cell lines were
maintained as described previously.15
At the time of assay, growth media were replaced with 50 µL
fresh media immediately followed by 50 µL Chroma-Glo
(Promega) lysis and luciferase reagent and incubated at room tem-
perature. At 15 and 30 min after lysis,green (537 nm) and red (613
nm) luminescence was detected with a Fluorostar (BMG
Technologies, Durham, NC) microplate reader by quantitating
photon emissions passing through 540-nm and 615-nm filters.
Luciferase induction assay
G/R-luc cells were plated in 96-well black optical bottom
plates at a density of 15,000 cells per well. Six hours after plating,
the media were changed to starving media (SV) lacking serum or
to fresh growth media (GM), incubated for 24 h at 37 °C, and
changed again to SV or GM for an additional 24 h at 37 °C. For
time induction experiments,SV was replaced with GM at 4,8,18,
24, and 30-h time points prior to the luciferase assay.
G/R-luc dual-reporter validation
Serial dilutions of either U0126 (Calbiochem, San Diego,
CA) or nocodazole (Sigma-Aldrich, St. Louis, MO) were
added to G/R-luc cells 6 h after plating. Approximately 40 h
after compound addition, green and red luciferase expression
was determined using a dual-luciferase assay. Cells treated with
growth media containing DMSO vehicle alone (V) were used
as positive controls for cell proliferation, and cells treated with
SV were used as negative controls. Green luminescence values
for the compound of interest (GλC) and vehicle positive controls
(GλV) were used to determine growth inhibition (GI) values at
each compound concentration (equation (1)). Lethal concentra-
tion (LC) values were determined using red luminescence for
the compound of interest (RλC) values and SV negative controls
(RλSV) (equation (2)).
GI = [(GλV –GλC)/GλV] × 100.
LC = [(RλC–RλN)/RλN] × 10.
The concentration of compound at the GI50value (GI = 50),
representing a 50% reduction in green luminescence, and the
LC50value (LC = –50), representing a 50% reduction in red
luminescence, was determined by plotting the sigmoidal dose-
response curves using GraphPad Prism (GraphPad Software,
San Diego, CA). The half-maximal inhibitory concentration
(IC50) was also determined for nocodazole using the sigmoidal
dose-response curve and GraphPad Prism.
G/R-luc cells were plated in 96-well black optical bottom
plates with each well receiving identical numbers of cells.
Depending on the experimental run, cells were plated at a den-
sity of 3000 to 5000 cells per well. The plating density was
chosen to be higher than the lower limit of detection of 2500
cells and low enough to allow 2 cell doublings over the course
of the experiment in uninhibited controls. Six hours after plat-
ing, the media were changed to fresh media containing controls
or 10 µM of compounds from the National Cancer Institute
Hawes et al.
2www.sbsonline.orgJournal of Biomolecular Screening XX(X); XXXX
(NCI) diversity set library (Drug Synthesis and Chemistry
Branch, Developmental Therapeutics Program [DTP], Division
of Cancer Treatment and Diagnosis, NCI, Frederick, MD). Two
sets of negative controls, SV containing DMSO (vehicle) or
100 ng/mL nocodazole, were included in duplicate. GM con-
taining DMSO was included in duplicate as a positive control
for cell growth and a negative control for growth inhibition.
Then, 10 µM of the phosphatidylinositol 3-kinase (PI3K)
inhibitor LY294002 (LY; Calbiochem) was included as a posi-
tive control for inhibition of growth signaling and was also
used to assess plate-to-plate variability. Two wells of each con-
trol were included in the first and last columns of each plate for
a total of 4 wells.
By separating the controls into the first and last columns,
the controls are the first and last samples read on each plate,
and the consistency between both groups controls for the sta-
bility of the reaction over the period of reading the plate.
Approximately 40 h after addition of compounds, green and
red luminescence was determined using a dual-luciferase assay.
Vehicle-alone internal controls were used to determine the raw
activity threshold (RAT) for normalization of individual 96-
well assays based on the inhibition of cell proliferation. The
RAT is the level of green luminescence corresponding to
50% growth inhibition compared with the total growth of the
vehicle-alone control. Green luminescence of the compound of
interest (GλC) and of the vehicle-alone controls (GλV) was used
to calculate the GI for screened compounds (see equation (1)).
Compounds resulting in less than 50% inhibition of green lumi-
nescence, GI < 50, such that their green luminescence mea-
surement exceeded the RAT, were excluded as inactive
compounds. Red luminescence of nocodazole-arrested controls
(RλN) was used to identify cytotoxic compounds with negative
LC values (see equation (2)), where RλCis the red luminescence
of the compound of interest. Positive hits from the initial screen
were assayed again to confirm reproducibility of antiprolifera-
tive activity and to eliminate false positives.
Dual-luciferase assays were also used to generate drug-
response curves and calculate GI50values for active compounds
identified from the screen. G/R-luc dual-reporter cells were
treated with 4-fold logarithmic dilutions of compound for
approximately 40 h, followed by a dual-luciferase assay. IC50
values were determined for cytostatic compounds using the sig-
moidal dose-response curve and GraphPad Prism.
G/R-luc astrocytoma cells were grown in 96-well plates as
described earlier and treated with 100 ng/mL nocodazole, 1 µM
NSC#676693, 1 µM deoxybouvardin, or DMSO 6 h after plat-
ing. After 24 h, the media were replaced in all wells; half of the
cells were allowed to recover in media without inhibitors, and
half of the cells were maintained in the presence of inhibitors.
Cells were allowed to recover for 32 h, and green luminescence
was determined with Chroma-Glo luciferase assays described
earlier. Green luminescence was calculated as percent control
of vehicle DMSO controls. Significance was determined using
repeated-measures analysis of variance (ANOVA) and Tukey’s
honestly significant differences (HSD) and Bonferroni post hoc
tests with p < 0.05 considered significant.
Alamar blue assay
G/R-luc, KR158, primary astrocytes, or SF295 (human
grade IV astrocytoma/GBM) cells were plated as described for
the high-throughput screen in 96-well tissue culture plates.
Compounds were added to the cells at logarithmic dilutions in
triplicate and incubated at 37 ºC. After approximately 40 h, GM
containing 30% Alamar blue (Invitrogen, Carlsbad, CA) was
added to the cells to a final concentration of 10% Alamar blue
and incubated at 37 °C for 4 h. Fluorescence was measured
with 560-nm excitation and 590-nm emission filters using a
Novostar (BMG Technologies) microplate reader.
Primary astrocytes were harvested as previously described17
with minor modifications. Neocortex was dissected from the
brains of wild-type C57BL/6J pups at P1. Neocortical cells
were dissociated by mechanical dissociation in Dulbecco’s
modified Eagle’s medium (DMEM)/20% fetal bovine serum
(FBS), pelleted by centrifugation, and plated at a density of 2 ×
106cells per 10-cm plate in DMEM/20% FBS and incubated at
37 °C for 14 days. Surviving astrocytes were plated at 5000
cells per well in 96-well plates for Alamar blue assays.
Astrocytoma cells were plated on glass coverslips in 12-well
plates at a density of 100,000 cells per well. Six hours after
plating, 1 µM of NSC#676693 or DMSO (vehicle control) was
added to the cells and incubated at 37 °C for 48 h. Cells were fixed
in 4% paraformaldehyde, stained with Texas Red-X phalloidin-
conjugated antibody (Molecular Probes, Eugene, OR), and
mounted onto slides with Prolong gold antifade reagent with
4′,6-diamidino-2-phenylindole (DAPI; Invitrogen).
RESULTS AND DISCUSSION
The G/R-luc dual-reporter system was generated by stably
expressing green and red click beetle luciferase reporters in
previously characterized grade III mouse astrocytoma cells,15
where expression of green luciferase was under control of the
human E2F1 promoter and expression of red luciferase was
driven by the constitutively active human CMV immediate-
early promoter. Therefore, green luciferase expression is con-
trolled by active cellular proliferation, and red luciferase
Dual-Reporter Assay Targeting Astrocytoma
Journal of Biomolecular Screening XX(X); XXXXwww.sbsonline.org3
expression is actively expressed in all healthy cells. Because
the peak luminescence for the green (537 nm) and red (613 nm)
click beetle luciferases is well separated, it is possible to simul-
taneously assay both using filtered luminescence with little
spectra overlap.18,19Although filtered luminescence decreases
the total luminescence signal intensity by up to 90%, back-
ground signal is extremely low. The G/R-luc reporter system is
sensitive enough to detect significant green-filtered lumines-
cence over background in as few as 2500 cells (Fig. 1A).
Furthermore, green luciferase expression in rapidly dividing
cells is linear with respect to cell number up to 40,000 cells,
when cell proliferation begins to be inhibited by increasing cell
densities and cell contact inhibition. The threshold for red lumi-
nescence sensitivity is slightly higher than green luminescence
(Fig. 1B). Red luminescence is linear with cell number up to a
very high cell density (160,000 cells per well), suggesting that
red luciferase expression correlates with cell number regardless
of the proliferation rate or E2F1 promoter activity.
Green luciferase expression reflects active cellular prolifer-
ation and is significantly lower when cellular proliferation is
slowed in the absence of serum (Fig. 1C). When growth media
containing serum are added back to serum-starved cells, green
luminescence increases. Green luminescence is significantly
increased 6 h after the addition of serum and rises rapidly after
20 h (Fig. 1D). The increase in green luciferase expression at 6
h coincides with activation of transcription and protein transla-
tion downstream of the E2F1 promoter. The rapid rise in green
luciferase expression after 20 h is consistent with the timing of
the cell cycle, requiring approximately 24 h to complete. These
data suggest that the human E2F1 promoter drives the green
luciferase expression in the G/R-luc astrocytoma cell line in an
efficient and proliferation-dependent manner.
Well-characterized growth inhibitors were used to validate
the G/R-luc dual-reporter system for determining LC50, GI50, or
IC50values in a 96-well dual-luciferase assay. Because green
luciferase expression correlates with activity of the E2F1 pro-
moter and active cell cycle, green luminescence was used to
calculate GI50values, the concentration of compound at which
cell proliferation is inhibited by 50% as compared with con-
trols. Because red luciferase expression correlates with total
cell number, red luminescence values were used to calculate
LC50values, the lethal concentration of compound at which
the total cell number is decreased by 50% as compared with noco-
dazole-arrested controls. The MEK inhibitor U0126 restricts
cell growth in multiple cell lines, including astrocytoma.20In
G/R-luc astrocytoma cells, the GI50 value for U0126 was found
to be 5 µM (Fig. 2A), and the LC50value for U0126 was deter-
mined to be 96 µM using dual-luciferase assays (Fig. 2B), and
these values are consistent with previously reported data.21
Nocodazole is a cell cycle inhibitor that depolymerizes micro-
tubules and arrests cells during mitosis or in G2. Nocodazole treat-
ment reaches Emaxat about 55% growth inhibition with a threshold
dose of 0.33 µM (100 ng/mL; Fig. 2C), which is consistent with
previous reports in which nocodazole was used at this dose to
induce complete growth arrest.22,23Because nocodazole is cytosta-
tic and Emaxdoes not reach 100% growth inhibition, the IC50value
is more informative to compound activity than the GI50value.
Hawes et al.
4www.sbsonline.orgJournal of Biomolecular Screening XX(X); XXXX
y = 0.044x - 37.741
R2 = 0.999
y = 0.01x + 750.0
R2 = 0.95
Green-filtered luminescence is sensitive at low cell densities and is lin-
ear in rapidly dividing populations. (B) Red-filtered luminescence is
linear with respect to cell number even at high cell densities. Green
luminescence is significantly reduced when G/R-luc cells are arrested
in the absence of serum (SV) compared with growth media (GM) and
increases when growth media (SV + GM) are added back to the cells
(C). Green luciferase is induced 6 h after the addition of GM, corre-
sponding to the time of new transcription in response to growth factor
signaling, and rapidly increases after 20 h, corresponding to an
increase in cell numbers through cell division (D).
Characterization of the G/R-luc dual-reporter system. (A)
Therefore, the G/R-luc dual-reporter assay was used to determine
the IC50value for nocodazole as being 0.06 µM in astrocytoma
cells (Fig. 2C). The U0126 and nocodazole pharmacology ana-
lyzed with the G/R-luc dual-reporter system is congruent with
known pharmacology of these compounds and validates this system
for determining LC50, GI50, and IC50values.
To verify the potential of the G/R-luc dual-reporter assay for
high-throughput screens (HTS), we determined the screening
window coefficient (Z′ factor) for both the green- and red-filtered
luminescence (see equation (3)). The Z′ factor takes into account
the experimental standard deviation and the signal-to-noise ratio
to determine the quality and potential of HTS assays.24The Z′
factor was determined using raw luminescence data values from
duplicate positive and baseline controls on each plate used in the
HTS as follows. Cells grown in 10% FBS growth media in the
presence of DMSO vehicle (GM) were used as positive growth
controls, and cells cultured in SV in the presence of 100 ng/mL
nocodazole to arrest cell growth were used as baseline controls.
The 2 positive controls and 2 baseline controls were used to deter-
mine the Z′ factor for each plate, and Z′ factors from 6 indepen-
dent runs were averaged to determine the Z′ factor for the HTS.
The Z′ factor equation was modified to fit the dual-luciferase
model (C), such that σV= standard deviation in GM controls,
σN= standard deviation in Noc controls, µV= average of the raw
luminescence data for GM controls, and µN= average of the raw
luminescence data for Noc controls. The Z′ factor for the green
luminescence is consistently within the range for an excellent
HTS assay, 0.5 ≤ Z′ < 1, in both 96-well and 384-well formats
(Fig. 2E). Although red luminescence exhibits a higher standard
deviation between replicates, the Z′ factor for red luminescence is
consistently within the range of a good HTS assay, 0.2 ≤ Z′ < 0.5,
in both 96-well and 384-well formats. Thus, the G/R-luc dual-
reporter assay has a high potential to discriminate changes in red
and green luminescence compared with positive and negative
controls in HTS assays.
To conduct validation tests as outlined by the Assay Guidance
Manual,25we used 10 µM of the PI3K inhibitor LY294002 (LY;
Calbiochem). In KR158 astrocytoma cells, 10 µM LY acts as a
cytostatic compound near the RAT by inhibiting E2F activity by
40% to 50%. During a potency analysis, the minimum significant
ratio (MSR) between multiple HTS runs was 1.26 (MSR < 3),
demonstrating that there was good individual agreement between
the multiple runs, and the assay passed the reproducibility test.
The upper and lower limits of agreement (LsA) fell between 0.94
and 1.49, demonstrating that the assay also passed the LsA crite-
rion (0.33 < LsA < 3.0) and equivalence test. Then, 10 µM LY was
included in all HTS assays as the primary internal control and to
monitor assay drift between runs. The overall MSR across the
diversity set HTS screen (described below) was 7.13, demonstrat-
ing that the assay reproducibility was stable over time (MSR <7.5).
The G/R-luc dual-reporter assay was used to screen the NCI
diversity set for compounds with antiastrocytoma activity (Fig.
3A,B). The NCI diversity set is composed of 1982 chemically
diverse compounds chosen to represent the various chemical
structure groups found in the larger set of almost 140,000 com-
pounds. G/R-luc cells were treated with 10 µM of each com-
pound for 40 h, followed by a dual-luciferase HTS assay.
Vehicle-alone controls were used to normalize the individual
assays (see the Methods section for details) (Fig. 3A) and cal-
culate the RAT as the level of luminescence corresponding to a
50% reduction in the E2F1 promoter activity relative to vehicle
alone. Thus, raw data can be quickly screened to find active
compounds that show luminescence measurements below RAT
for individual plates. Compounds resulting in a decrease in red
luminescence, LC < 0, as compared with nocodazole-arrested
controls were identified as cytotoxic compounds (Fig. 3B).
Compounds that did not result in a decrease in red luminescence
as compared with arrested controls, such that LC ≥ 0, represented
noncytotoxic compounds. Thus, positive 50% growth inhibition,
GI > 50, and positive LC values, LC ≥ 0, represent cytostatic
Z′ = 1 – . (3)
Dual-Reporter Assay Targeting Astrocytoma
Journal of Biomolecular Screening XX(X); XXXX www.sbsonline.org5
Log (µM)Log (µM)
curves for (A) U0126 and (C) nocodazole were used to calculate GI50
values. Concentration response curves were used to calculate LC50values
for (B) U0126 and (D) nocodazole. Screening window coefficients (Z′
factor) were determined for green-filtered (Z′GREEN) and red-filtered
(Z′RED) luminescence (E) from 6 independent runs in high-throughput
screening assays. GI, growth inhibition; LC, lethal concentration.
G/R-luc dual-reporter validation. Sigmoidal dose-response
compounds that inhibit activity of the E2F1 promoter but do not
generally inhibit transcription factors, as evidenced by the activity
of the CMV promoter, or kill the cells.
Compounds with GI > 50 in the initial screen were rescreened
to confirm reproducibility and eliminate false positives. During
rescreening, cells were also examined briefly under a light micro-
scope to determine the frequency of false positives in which com-
pound treatment inhibited both green and red luminescence but
did not kill the cells. These compounds are likely to inhibit activ-
ity of the E2F1 and CMV promoters indiscriminately or activity
of the luciferase enzymes but are not specific for the E2F1 pro-
moter. The false-positive rate was found to be less than 0.2%. The
remaining compounds were further divided into 3 groups: cyto-
toxic, cytostatic, and cytotoxic/cytostatic. Ninety-seven cytotoxic
compounds were identified (4.9% of tested compounds). The
cytostatic group consisted of 68 compounds (3.4% of tested
compounds) that reproducibly inhibited cellular proliferation
without cytotoxicity. The third group, cytotoxic/cytostatic, con-
sisted of 9 compounds that were cytotoxic in 1 screen and cyto-
static in another. These compounds are likely to be cytotoxic,such
that the concentration used at screening, 10 µM, is at a threshold
concentration level near the LC50.
To further refine the list of compounds with activity against
brain tumors, we compared the compounds identified from the
compound screen in the G/R-luc astrocytoma cell line with the
NCI-60 data that were available through the NCI DTP program
(http://www.dtp.nci.nih.gov/index.html). Thirty-seven of the
identified compounds have activity in at least 1 human CNS
cell line, and 14 showed significant activity in at least 3 or more
CNS cell lines. The G/R-luc astrocytoma cells were then used
to determine the GI50value for the top candidates in a 96-well
Camptothecin and several of its derivatives, including topote-
can, were identified as cytotoxic agents in the G/R-luc astrocy-
toma cell line, with GI50values in the nanomolar range (Table 1).
Although camptothecin failed clinical trials in the 1970s, topote-
can was approved by the Food and Drug Administration (FDA) in
1996 as a secondary treatment for ovarian and small-cell lung
cancers and is currently being investigated for the treatment of
astrocytoma.26Although these compounds are not novel, their
identification from the compound library serves as positive con-
trols for hit identification in the HTS assay and further validates
the use of the G/R-luc dual-reporter assay as a therapeutic screen-
ing tool to identify active compounds. Bouvardin and deoxybou-
vardin were found to have cytotoxic activity in the 28- to 41-nM
range. The bouvardins also have been reported as potential anti-
tumor agents.27,28Eight other compounds were identified from the
NCI diversity set as having cytotoxic activity in the p53/NF1-null
G/R-luc dual-reporter astrocytoma cells (Table 1). Four of these
compounds have potent cytotoxic activity in the nanomolar range.
Washout experiments were used to confirm that the G/R-luc
dual-reporter system differentiates between cytostatic and cyto-
toxic activities. Green luminescence is inhibited in the pres-
ence of cytostatic compounds nocodazole and NSC#676693
Hawes et al.
6 www.sbsonline.orgJournal of Biomolecular Screening XX(X); XXXX
GM SVNocLY B6B8E4 E9 H10 H11 A11
E4 E9 H10 H11 A11
high-throughput screening (HTS) assay. (A) Growth inhibition (GI)
values from green luciferase data were normalized to cells treated with
growth media + vehicle (GM), and (B) lethal concentration (LC) val-
ues from red luciferase data were normalized to nocodazole-treated
cells (Noc). White bars indicate the internal controls run in quadrupli-
cate on each plate: GM (growth media + DMSO vehicle), SV (starv-
ing media + DMSO vehicle), Noc (starving media + nocodazole), and
LY (growth media + LY294002). Black bars indicate wells containing
active compounds that were identified in the screen, asterisks indicate
cytotoxic compounds, and arrows indicate cytostatic compounds.
Shaded bars indicate nonactive compounds chosen at random from the
plate. Compounds were run in single wells in the primary screen
shown here. RAT, raw activity threshold.
(A, B) Representative data plots from the dual-luciferase
GI50Values of Cytotoxic Compounds
(Tables 2, 3; Fig. 4) and the cytotoxic compound deoxybou-
vardin (Tables 1,3; Fig. 4). When inhibitors are removed from the
media, green luminescence significantly increases in cells pre-
treated with nocodazole and NSC#676693 but not in those previ-
ously treated with deoxybouvardin (Fig. 4). Therefore, G/R-luc
dual-reporter astrocytoma cells were able to recover from cytosta-
tic treatment with nocodazole and NSC#676693 but not from the
cytotoxicity of deoxybouvardin. These data further suggest that
green luminescence in the G/R-luc dual-reporter system can be
used to distinguish between cytostatic and cytotoxic inhibition.
At least 5 of the cytostatic compounds identified in the HTS
screen were found to alter the cellular morphology of the astrocy-
toma cells (Table 2). Treatment with NSC#676693,NSC#128687,
NSC#158383, or NSC#131734 results in an apparent increase in
cellular cytoplasm and change in cell shape. On the other hand,
treatment with NSC#131053 results in a decrease in total cell size.
Furthermore, treatment of anaplastic astrocytoma cells with
NSC#676693 results in a clear morphological change that lacks
the astrocytic projections and cell-spreading characteristics of
KR158 cells and is consistent with inhibition of cytoskeletal regu-
lation (Fig. 5B). Because these compounds are cytostatic and Emax
does not approach 100% growth inhibition, the concentration
resulting in 50% growth inhibition (GI50) is not equal to the con-
centration resulting in half of the maximal inhibitory effect (IC50
values). Therefore,IC50values are more reflective of the activity of
cytostatic compounds than GI50values. The IC50values for all but
1 of the morphology-altering compounds are in the low micromo-
lar to nanomolar range. Although these compounds are cytostatic
rather than cytotoxic up to 10 µM, all but 1 of these compounds
(NSC#131053) reaches 50% growth inhibition. This is consistent
with nocodazole treatment, which arrests the cells in mitosis and
reaches Emaxnear 55% growth inhibition (Fig. 2C). NSC#676693
also reaches Emaxat 55% inhibition of the E2F1 promoter over a 2-
fold log concentration from 1 to 10 µM (Fig. 5A), and washout
experiments confirm that cells recover after treatment with
NSC#676693 similarly to nocodazole (Fig. 4),thereby suggesting a
specific cytostatic rather than cytotoxic function for NSC#676693.
Because the dual-reporter assay has been established in a
mouse grade III astrocytoma line, it is important to validate the
results of this assay in human tumor lines and in tumors of dif-
ferent grades. GI50or IC50 values for potent compounds were
assessed in other astrocytoma cell lines to determine if the
inhibitory effects were common to different astrocytoma cells
from both mouse and human (Table 3). Alamar blue assays,
which assess innate metabolic activity and cell viability, were
used to generate therapeutic response curves and calculate GI50
or IC50values for cell lines not containing the red and green
luciferase reporters. The Alamar blue assay in grade III KR158
mouse astrocytoma cells and the dual-reporter assay in grade
III G/R-luc mouse astrocytoma cells yielded similar results.
Thus, GI50and IC50values obtained from the dual-reporter assay
compare with the conventional Alamar blue assay. To assess
whether cytotoxic and cytostatic effects were specific to tumor
cells, we determined GI50or IC50 values for each compound in
primary astrocyte cultures. Only deoxybouvardin exhibited
nonspecific inhibition (Table 3), which suggests that deoxy-
bouvardin cytotoxicity does not discriminate between cancer-
ous and noncancerous cells.
With conventional HTS assays, it is necessary to run a
secondary assay to eliminate false positives that are cytotoxic or
Dual-Reporter Assay Targeting Astrocytoma
Journal of Biomolecular Screening XX(X); XXXX www.sbsonline.org7
GI50Values of Cytostatic Compounds
ND, not determined.
a. IC50values are reported for cytostatic activities reaching Emaxbelow 100% growth
GI50Values from Alamar Blue Metabolic
NA, no activity.
a. IC50values are reported for the cytostatic compound NSC#676693.
static and cytotoxic activities in washout experiments. Cytostatic com-
pounds nocodazole and NSC#676693 and the cytotoxic compound
deoxybouvardin (white bars) inhibit green luminescence. When com-
pounds are removed (shaded bars), cells treated with nocodazole or
NSC#676693 recover, but cells treated with deoxybouvardin do not.
Green luminescence was calculated as % vehicle (DMSO) control.
*p < 0.05.
The G/R-luc dual-reporter system distinguishes between cyto-
nonspecific to the target of interest, such as the E2F1 promoter.
Because the dual-luciferase assay design discriminates between
green and red luminescence, it is possible to distinguish luciferase
expression under the control of different promoters in a single assay.
This system distinguishes between specific inhibition of the E2F1
promoter (green luminescence) and nonspecific transcription factor
inhibition and general cytotoxicity (red luminescence) in the initial
screen. Thus,the dual-luciferase reporter assay system can cut costs
and time significantly by identifying potential therapeutic candi-
dates from a single HTS assay. This system further can be used to
determine GI50, LC50, and IC50values as well as study the additive
or synergistic effects of combinatorial treatments in an HTS fashion.
The G/R-luc astrocytoma dual-reporter assay was validated as
a stable and reproducible HTS tool to identify novel compounds
with antiproliferative activity in astrocytoma cells. Out of 1982
chemically diverse compounds in the NCI diversity set compound
library, 14 were identified that also have significant activity in
other CNS tumor cell lines. Several compounds with known
antiproliferative activity in CNS tumors were also identified, such
as camptothecin and several of its derivatives, and served as posi-
tive controls for successful identification of therapeutic agents
using the G/R-luc dual-reporter assay. Three novel compounds—
NSC#207895, NSC#268665, and NSC#606985—were also iden-
tified as having potent cytotoxic activity in mouse and human
astrocytoma cells with specificity for tumor cells over primary
astrocytes. Future experiments will be required to determine the in
vivo bioavailability and use of these compounds as antiastrocy-
toma therapeutic agents.
The G/R-luc dual-reportor HTS also identified 6 compounds
in the NCI diversity set that are cytostatic and specifically inhibit
cellular proliferation accompanied by alteration of astrocytoma
cell morphology. For example, NSC#676693 is a highly potent
Hawes et al.
8www.sbsonline.orgJournal of Biomolecular Screening XX(X); XXXX
(A). NSC#676693 reaches Emaxat 50% growth inhibition over 2 log concentrations. Texas Red–conjugated phalloidin was used to visualize actin
filaments and morphological changes in cells treated with (B) DMSO vehicle or (C) 1 µM NSC#676693. Scale bar = 100 µm. DAPI, 4′,6-
Representative drug-response curves for the cytostatic compound NSC#676693 compared with the cytotoxic compound camptothecin
cytostatic inhibitor, specific for astrocytoma cells compared with Download full-text
primary astrocytes, and induces morphological changes in astro-
cytoma cells. These compounds may be useful tools to under-
standing astrocytoma tumorgenicity. Future in vivo experiments
will determine the utility of these cytostatic compounds as antias-
We described here a novel dual-reporter assay that uses filtered
luminescence to simultaneously assess and distinguish between
activity of the E2F1 promoter and nonspecific transcription factor
inhibition and cytotoxicity. The G/R-luc dual-reporter system is an
efficient and promising tool for the identification and study of anti-
astrocytoma therapeutic agents in vitro.
This research was supported by the Intramural Research
Program of the National Institutes of Health (NIH), National
Cancer Institute. This research was performed while JJH held a
National Research Council Research Associateship Award at
the National Cancer Institute. All experiments were conducted
in compliance with the current laws of the United States.
1.Central Brain Tumor Registry of the United States: Primary Brain
Tumors in the United States—Statistical Report. Chicago: Central Brain
Tumor Registry of the United States, 2002.
Ohgaki H: Genetic pathways to glioblastomas. Neuropathology 2005;
McLendon RE, Halperin EC: Is the long-term survival of patients with
intracranial glioblastoma multiforme overstated? Cancer 2003;98:1745-1748.
Gilbert MR, Loghin M: The treatment of malignant gliomas. Curr Treat
Options Neurol 2005;7:293-303.
Ichimura K, Bolin MB, Goike HM, Schmidt EE, Moshref A, Collins VP:
Deregulation of the p14ARF/MDM2/p53 pathway is a prerequisite for
human astrocytic gliomas with G1-S transition control gene abnormali-
ties. Cancer Res 2000;60:417-424.
Rasheed BK, McLendon RE, Herndon JE, Friedman HS, Friedman AH,
Bigner DD, et al:Alterations of the TP53 gene in human gliomas. Cancer
Watanabe K, Sato K, Biernat W, Tachibana O, von Ammon K, Ogata N,
et al: Incidence and timing of p53 mutations during astrocytoma progres-
sion in patients with multiple biopsies. Clin Cancer Res 1997;3:523-530.
von Deimling A, Eibl RH, Ohgaki H, Louis DN, von Ammon K, Petersen I,
et al: p53 mutations are associated with 17p allelic loss in grade II and grade
III astrocytoma. Cancer Res 1992;52:2987-2990.
Harris CC: p53 tumor suppressor gene: from the basic research laboratory
to the clinic—an abridged historical perspective. Carcinogenesis 1996;17:
Weinstein JN, Myers TG, O’Connor PM, Friend SH, Fornace AJ, Kohn
KW, et al: An information-intensive approach to the molecular pharma-
cology of cancer. Science 1997;275:343-349.
Huson S, Hughes R: The Neurofibromatoses:A Pathogenetic and Clinical
Overview. London: Chapman & Hall Medical, 1994.
Martin GA, Viskochil D, Bollag G, McCabe PC, Crosier WJ, Haubruck H,
et al: The GAP-related domain of the neurofibromatosis type 1 gene product
interacts with ras p21. Cell 1990;63:843-849.
13. Cichowski K, Shih TS, Schmitt E, Santiago S, Reilly K, McLaughlin ME,
et al: Mouse models of tumor development in neurofibromatosis type 1.
Vogel KS, Klesse LJ, Velasco-Miguel S, Meyers K, Rushing EJ, Parada LF:
Mouse tumor model for neurofibromatosis type 1. Science 1999;286:
Reilly KM, Loisel DA, Bronson RT, McLaughlin ME, Jacks T. Nf1;Trp53
mutant mice develop glioblastoma with evidence of strain-specific
effects. Nat Genet 2000;26:109-113.
Uhrbom L, Nerio E, Holland EC: Dissecting tumor maintenance require-
ments using bioluminescence imaging of cell proliferation in a mouse
glioma model. Nat Med 2004;10:1257-1260.
Sanchez JF, Sniderhan LF, Williamson AL, Fan S, Chakraborty-Sett S,
Maggirwar SB: Glycogen synthase kinase 3beta-mediated apoptosis of
primary cortical astrocytes involves inhibition of nuclear factor kappaB
signaling. Mol Cell Biol 2003;23:4649-4662.
Almond B, Hawkins E, Stecha P, Garvin D, Paguio A, Butler B, et al: A
New Luminescence: Not Your Average Click Beetle. 2003. Retrieved from
Gammon ST, Leevy WM, Gross S, Gokel GW, Piwnica-Worms D:
Spectral unmixing of multicolored bioluminescence emitted from hetero-
geneous biological sources. Anal Chem 2006;78:1520-1527.
Uht RM, Amos S, Martin PM, Riggan AE, Hussaini IM: The protein
kinase C-eta isoform induces proliferation in glioblastoma cell lines
through an ERK/Elk-1 pathway. Oncogene 2007;26:2885-2893.
Ahn N, Tolwinski NS, Hsiao K, Goueli SA: U0126: An Inhibitor of
MKK/ERK Signal Transduction in Mammalian Cells. Madison, WI:
Promega Corporation, 1999.
Barnouin K, Dubuisson ML, Child ES, Fernandez de Mattos S, Glassford
J, Medema RH, et al: H202 induces a transient multi-phase cell cycle
arrest in mouse fibroblasts through modulating cyclin D and p21Cip1
expression. J Biol Chem 2002;277:13761-13770.
Yamada HY, Gorbsky GJ: Cell-based expression cloning for identifica-
tion of polypeptides that hypersensitize mammalian cells to mitotic
arrest. Biol Proced Online 2006;8:36-43.
Zhang JH, Chung TD, Oldenburg KR: A simple statistical parameter for
use in evaluation and validation of high throughput screening assays. J
Biomol Screen 1999;4:67-73.
Assay Guidance Manual Version 4.1. 2005. Retrieved from http://www
Klautke G, Schutze M, Bombor I, Benecke R, Piek J, Fietkau R:
Concurrent chemoradiotherapy and adjuvant chemotherapy with topotecan for
patients with glioblastoma multiforme. J Neurooncol 2006;77:199-205.
Adwankar MK, Khandalekar DD, Chitnis MP: Combination chemother-
apy of early and advanced murine P388 leukaemia with bouvardin, cis-
diamminedichloroplatinum and vincristine. Oncology 1984;41:370-373.
Jolad SD, Hoffmann JJ, Torrance SJ, Wiedhopf RM, Cole JR, Arora SK,
et al: Bouvardin and deoxybouvardin, antitumor cyclic hexapeptides from
Bouvardia ternifolia (Rubiaceae). J Am Chem Soc 1977;99:8040-8044.
Address correspondence to:
Karlyne M. Reilly, Ph.D.
NCI-Frederick, Mouse Cancer Genetics Program
West 7th Street at Fort Detrick
PO Box B, Building 560, Rm 32-20
Frederick, MD 21702
Dual-Reporter Assay Targeting Astrocytoma
Journal of Biomolecular Screening XX(X); XXXX www.sbsonline.org9