Báo cáo khoa học: The in vitro effects of CRE-decoy oligonucleotides in combination with conventional chemotherapy in colorectal cancer cell lines potx

The
in vitro
effects of CRE-decoy oligonucleotides in combination
with conventional chemotherapy in colorectal cancer cell lines
Wai M. Liu*, Katherine A. Scott*, Sipra Shahin and David J. Propper
New Drug Study Group, Barry Reed Oncology Laboratory, St. Bartholomew’s Hospital, London, UK
The cAMP response element consensus sequence directs the
transcription of a wide range of genes. A 24-mer single-
stranded cAMP response element decoy oligonucleotide
(CDO) has been shown to compete with these sequences for
binding transcription factors and therefore interferes with
cAMP-induced gene transcription. We have examined the
effect of this CDO alone and in combination with a range of
common chemotherapeutic agents in colorectal cancer cell
lines. CDO had a potent anti-proliferative effect in colorectal
cell lines, yet, a similar enhancement of cell death was
not observed. Simple drug–drug interaction studies showed
that combining CDO with chemotherapy resulted in an
enhancement of the antiproliferative effects. Furthermore,
this cytostatic effect was protracted and associated with an
increase in senescence-associated b-galactosidase activity at
pH 6. There is a possible role for p21
waf1
in mediating this
effect, as the enhancement of cell growth inhibition was not
observed in cells lacking the ability to correctly upregulate
this protein. Additionally, significant decreases in cyclin-
dependent kinase (CDK) 1 and CDK 4 function were seen
in the responsive cells. These data provide a possible model
of drug interaction in colorectal cell lines, which involves the
complex interplay of the molecules regulating the cell cycle.

Clinically, the cytostatic ability of CDO could improve and
enhance the antiproliferative effects of conventional cyto-
toxic agents.
Keywords: cAMP response element; colorectal cancer; oligo-
nucleotide decoy factors; synergy.
The regulation of transcription by using short sequence
oligonucleotides has been a focus for developing new drug
therapies[1–5].Thisisbasedupontheprinciplethatrepression
of key genes associated with malignancy might provide novel
therapeutic targets. Also, dysregulation of response elements
within promoter regions of genes has been implicated in
neoplastic transformation, thus restoring correct and appro-
priate function may reverse the aberrant phenotype [1].
There are two approaches using oligonucleotides to
achieve this. First, the development of dominant mutants
with dysfunctional activation domains, which compete with
wild-type counterparts in binding to target genes [2]. This
antisense oligonucleotide approach results in agents that,
through their ability to bind specific RNA and DNA
sequences are highly selective. However, this genomic
approach has only met with a limited degree of success, as
there have been conflicting reports to suggest that the
efficacy of these antisense oligonucleotides may not exclu-
sively be a result of sequence-binding, but to some other yet
unknown mechanism predominant in cells that sensitizes
them to cell killing [6,7]. In addition, it is also possible that
the reagents used to maximize delivery of these oligonucle-
otides to the target cell may actually directly interfere
with cellular processes, resulting in nonspecific effects [8,9].
Another consideration for the use of generic antisense

oligonucleotides is the diversity and number of possible
fusion sequences in cancer, which can actually prevent a
particular disease from being treated successfully with just
a single agent. For example, the bcr-abl translocation in
chronic myeloid leukaemia can have as many as seven
distinct junctional sequences that would require their own
antisense oligonucleotide [10,11]. Consequently, treatment
would have to be adapted for each individual patient,
making the concept of using oligonucleotides less attractive.
The second approach involves the use of short strands of a
nucleotide sequence as a decoy factor, which competes with
the response elements within the promoter regions of genes
that bind transcription factors [1,12]. In a similar manner to
the first approach, specificity is achieved through sequence
binding. However, this is enhanced further, as relevant
transcription factors are specifically sequestered by the decoy
oligonucleotides, resulting in an effect that is both sustain-
able and nongenomic in nature. Additionally, as protein–
protein interactions would be distal from native enhancer
sites, nonspecific interference of these sites would be reduced.
The cAMP response element (CRE) consensus sequence
is intimately involved in the transcription of a wide range of
Correspondence to W. M. Liu, Drug Resistance Team, Section of
Medicine, Institute of Cancer Research, Haddow Laboratories,
15 Cotswold Road, Sutton, Surrey SM2 5NG, UK.
Fax: + 44 208 661 3541, Tel.: + 44 208 722 4429,
E-mail:
Abbreviations: CRE, cAMP response element; CREB, CRE-binding
protein; CDK, cyclin-dependent kinase; CDO, CRE-decoy
oligo-nucleotide; 5-FU, 5-fluorouracil; SO, scrambled mismatch

oligonucleotides; BrdU, 5-bromo-2¢-deoxyuridine; SA-b-gal,
senescence-associated b-galactosidase.
*Present address: Section of Medicine, Institute of Cancer Research,
Haddow Laboratories, 15 Cotswold Road, Sutton, Surrey SM2 5NG,
UK.
(Received 18 February 2004, revised 22 April 2004,
accepted 7 May 2004)
Eur. J. Biochem. 271, 2773–2781 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04208.x
genes [13]. The promoter region of several of these genes has
been studied, and a common CRE sequence has been noted
upstream of the transcriptional start site [14]. All of the
cAMP responsive gene promoter regions have the same
eight-base enhancer sequence, the CRE, which is the
palindromic sequence 5¢-TGACGTCA-3¢ [13]. Proteins that
bind to these CREs have been identified that are 43 kDa in
size, and contain a basic leucine zipper DNA-binding motif
[15]. Functional studies have shown that this transcription
factor, termed the CRE-binding protein (CREB), couples
gene activation to a wide variety of cellular signals [14], and
thus coordinates a multitude of genes that regulate numer-
ous cellular processes, including cell growth and differenti-
ation [16].
The ubiquitous nature of the CRE consensus site makes it
a good target for chemotherapy. Indeed, it has been shown
that a palindromic trioctamer of this sequence can interfere
with CREB binding, and specifically inhibit PKA subunit
expression, interfering with the CRE-PKA pathway [17].
This causes specific inhibition of growth in cancer cells, and
although the CRE-regulated genes are common in all cell
types, surprisingly, CRE-decoy oligonucleotides (CDOs)

has no significant effect in normal cells. Furthermore, in
animal studies, CDOs induced tumour shrinkage without
obvious toxicity [17]. The mechanism by which CDOs
inhibit cell growth has not been elucidated, although it has
been shown that CRE-decoy treatment reduces cyclin D1
and cyclin-dependent kinase (CDK) 4 levels and retino-
blastoma protein (Rb) phosphorylation. CDO-induced
growth inhibition was independent of p53 status [18,19],
and accompanied by the hallmarks of apoptosis [20], which
together suggests a more profound interaction.
The aims of the present study were threefold: first, to
explore the in vitro effects of CDO alone in a panel of three
colorectal cancer cell lines; second, to investigate the effects
of combining CDO with etoposide (VP16), 5-fluorouracil
(5-FU) or SN38 on cell growth and viability; and third, to
elucidate the cellular mechanisms underlying any synergistic
effects seen in the drug combinations.
Materials and methods
Cell culture
HCT116 and SW620 colorectal cell lines were obtained
from the Cancer Research UK laboratories, and were
maintained in Dulbecco’s modified Eagle’s medium supple-
mented with 10% (v/v) fetal bovine serum. GEO colorectal
cancer cell lines were a gift from G. Tortora (Dipartimento
di Endocrinologia e Oncologia Molecolare e Clinica,
Universita
`
di Napoli, Italy), and were maintained in
McCoy’s 5A with 10% (v/v) fetal bovine serum. HCT116
and GEO cell lines were both wild-type p53, and SW620

lines were mutant p53. No antibiotics were used in our
experiments, and all cell lines were incubated in a humidified
atmosphere with 5% (v/v) CO
2
in air at 37 °C.
Transfection with CDO
CDO and scrambled mismatch oligonucleotides (SO) were
gifts from Y. S. Cho-Chung (National Cancer Institute,
Bethesda, MD, USA), and were phosphothiorated for
stability [21,22]. They were trioctamers of the CRE consen-
sus site, and their complete sequences were: CDO,
5¢-TGACGTCATGACGTCATGACGTCA-3¢;SO,5¢-TGT
GGTCATGTGGTCATGTGGTCA-3¢.
Cells (1 · 10
5
mL
)1
) were plated into 6-well plates, and
allowed to adhere for 24 h. Cells were rinsed in Hank’s
buffered salt solution (Sigma) and refreshed with serum-free
medium before the addition of CDO with Oligofectamine
reagent according to the manufacturer’s protocol (Invitro-
gen Ltd). CDO was added at a 50–200 n
M
final concentra-
tion. After 4 h of incubation, culture medium supplemented
with 20% (v/v) fetal bovine serum was added to make the
volume up to 5 mL. At this stage, SN38, 5-FU or VP16 (all
from Sigma) could be added. Aliquots were removed daily
for assessment of cell number and viability by staining with

Trypan blue, and cell cycle distribution by flow cytometry.
DNA analysis
The distinct phases of the cell cycle were distinguished by
flow cytometry, according to methods described previously
[23]. The acquisition of data was performed within 1 h using
a FACSCalibur (BD Biosciences). Five thousand cells were
analysed for each data point, and the percentages of cells in
sub-G
1
(apoptotic fraction, cells with a reduced propidium
iodide stain but similar morphology), G
1
,SandG
2
/M
phases were determined using the cell cycle analysis program
WINMDI
v2.4.
Flow cytometric analysis of BrdU incorporation
The degree of incorporation of the thymidine analogue
5-bromo-2¢-deoxyuridine (BrdU; Sigma) in HCT116 and
GEO cells was measured by flow cytometry. Following
culture in CDO and drugs, cells (5 · 10
5
mL
)1
)were
transferred into fresh culture medium containing 10 l
M
BrdU for 30 mins before fixing with ice-cold 70% (v/v)

ethanol and permeabilization in 2
M
HCl with 0.5% (v/v)
Triton X-100. Samples were washed and incubated with
fluorescein isothiocyanate-conjugated mouse anti-BrdU
according to the manufacturer’s instructions (PharMingen).
The cell cycle distribution was resolved by staining with
propidium iodide, and BrdU fluorescence specifically within
the S-phase was measured by using the FACSCalibur.
Ten-day clonogenic assays
Cells were harvested from initial cell cultures and
resuspended in DMEM. Cells (1 · 10
5
mL
)1
)wereplated
in semisolid cultures containing 0.9% (w/v) methylcellulose
and 30% (v/v) fetal bovine serum (Stem Cell Tech). Culture
dishes were incubated at 37 °C in a humidified atmosphere
with 5% (v/v) CO
2
. The number of colonies containing more
than 50 cells was assessed on day 10.
Immunoblotting and immunoprecipitation analysis
For immunoblot analysis, total cellular protein was solubi-
lized and resolved by SDS/PAGE using 15% acrylamide
with a 5% stacking gel as described previously [23]. Pri-
mary antibody probing was performed with anti-p21
waf1
(0.2 lgÆmL

)1
), anticyclin D (2 lgÆmL
)1
), anti-CDK 4
2774 W. M. Liu et al. (Eur. J. Biochem. 271) Ó FEBS 2004
(2 lgÆmL
)1
), anticyclin B (2 lgÆmL
)1
), or anti-CDK 1
(2 lgÆmL
)1
) (all from PharMingen). Anti-(b-actin) Ig was
used to confirm equal sample loading (1 : 2000; Oncogene
Research Products). Following a washing step in 0.1% (v/v)
Tween in Tris-buffered saline (Sigma; 100 m
M
Tris pH 7.6,
150 m
M
NaCl), horseradish peroxidase-conjugated anti-
species IgG1 was used as the secondary antibody (1 : 1000;
DAKO Ltd). Bands were visualized by the ECL plus
detection system (Amersham Biosciences Ltd).
For the analysis of cyclin–CDK interaction, cells were
lysed in a modified RIPA buffer (50 m
M
Tris, 250 m
M
NaCl,

5m
M
EDTA, 50 m
M
NaF, 10 lgÆmL
)1
phenylmethane-
sulfonyl fluoride, 0.5 lgÆmL
)1
leupeptin, 2 lgÆmL
)1
soybean
trypsin inhibitor, 0.5 lgÆmL
)1
aprotinin, 2 lgÆmL
)1
N-tosyl-
L
-phenylalanine chloromethyl ketone, 0.1% (v/v) Triton X-
100; all Sigma), and clarified by centrifugation. Protein was
used for immunoprecipitation with either anticyclin D or
cyclin B and protein A-sepharose (Amersham). Resultant
immune complexes were washed twice with RIPA buffer,
denatured in Laemmli buffer, and resolved by SDS/PAGE
(15% acrylamide).
Analysis of SA-b-gal activity
Cellular senescence-associated b-galactosidase (SA-b-gal)
activity was assessed as previously described by this group
[24]. Briefly, cells were washed twice in ice-cold NaCl/P
i

,
before fixing in 2% (v/v) formaldehyde and 0.2% (v/v)
glutaraldehyde. Cells were then washed twice in ice-cold
NaCl/P
i
, before overnight incubation at 37 °CinX-Gal
staining solution (1 mgÆmL
)1
5-bromo-4-chloro-3-indolyl
b-
D
-galactoside in 40 m
M
citric acid/sodium phosphate
pH 6, 5 m
M
potassium ferricyanide, 5 m
M
potassium
ferrocyanide, 150 m
M
sodium chloride, 2 m
M
magnesium
chloride). Samples were then washed twice in ice-cold NaCl/
P
i
prior to assessing the percentage of cells staining positive
for SA-b-gal activity by light microscopy.
Statistical analysis

All statistical analyses were performed using
MINITAB
version 10 (State College, PA, USA). Data was normally
distributed as established by Shapiro–Wilk testing, and
parametric analyses were used throughout. Differences
between variables and control cultures, as determined by
analysis of variance, were further characterized by paired
Student’s t-tests.
Results
Exposure to single-agent CDO
A concentration-dependent reduction in cell number and
cell viability was observed in HCT116 and GEO cell lines
cultured with CDO. However, no changes to cell number or
viability was observed in the cells treated with the SO
control (Fig. 1). Conversely, in SW620 cells that are
intrinsically more resistant to cytotoxic agents in general,
CDO had no effect on cell proliferation at equi-molar
concentrations (cells per mL and percentage viability with
1.6 l
M
CDO: 1.1 · 10
6
and 89.7% vs. 1.7 · 10
6
and 91.3%
in SO-cultured control cells). Flow cytometric analysis
revealed concomitant increases in the sub-G
1
population of
cells, indicative of apoptosis (Fig. 1).

Combination with other chemotherapeutic agents
Preliminary experiments indicated that SW620 cells were
resistant to both CDO and chemotherapy at the concen-
trations studied, and so were excluded from the combina-
tion studies. These simple combination studies involved
culturing cells simultaneously with each agent at the
concentration that reduced cell numbers by 25% (IC
25
).
Culturing HCT116 and GEO cells with these equi-toxic
drug concentrations resulted in different responses
in these cell lines. Specifically, combining CDO with a
chemotherapeutic drug had no significant effect in HCT116
cells, but significantly reduced cell numbers in GEO
cultures. Also, this effect was greater than expected
(hyper-additive) (Fig. 2A). This can be illustrated most
clearly with the results for GEO cells cultured with CDO
and 5-FU; by simply comparing the total reduction in cell
number in cultures treated with CDO and 5-FU together
(relative to the SO control) with the expected reduction in
Fig. 1. The effect of CDO in HCT116 and GEO cell lines on day 3. The activity of CDO in the sensitive cell lines was fitted a standard E
max
model.
Representative DNA histograms following culture with CDO in HCT116 cells are also shown. Each point represents the means and SD of at least
three separate experiments. A, Apoptosis; SO, scrambled mismatch oligonucleotide control.
Ó FEBS 2004 CRE-decoys in colorectal cancer cells (Eur. J. Biochem. 271) 2775
cell number (calculated as the numerical sum of the
reductions in cell number seen in the cultures with the two
agents separately [25] (· 10
5

cellsÆmL
)1
: )22.3 ± 1.8 vs.
)13.8 ± 2.6; P < 0.001). These results were confirmed by
the flow cytometric data, which showed no enhancement of
the apoptotic fraction (sub-G
1
population) of HCT116 cells
Fig. 2. The effect of combining CDO with
cytotoxic agents in HCT116 and GEO cells.
Cells were cultured with CDO (C) alone or in
combination with 5-FU, SN38 or VP16. There
were significant reductions in cell number in
GEO cells that was not seen in HCT116 cells.
(A) Representative DNA histograms of GEO
cells cultured with VP16, 5-FU and SN38 in
the presence or absence of CDO. (B) Individ-
ual fractions of events within the sub-G
1
population (apoptosis). Each data point
represents the mean and SDs of six separate
experiments; P-values were calculated from
paired Student’s t-tests.
2776 W. M. Liu et al. (Eur. J. Biochem. 271) Ó FEBS 2004
cocultured with CDO and any cytotoxic agent (Fig. 2B).
This was similarly established by comparing the size of the
sub-G
1
population in the combination sample with the
numerical sum of the separate sub-G

1
populations seen in
cells treated with the individual agents. Additionally, flow
cytometry revealed no apparent blockades in the G
1
,Sor
G
2
phases of the cell cycle, suggesting that the reduction in
cell number may have been the result of a general and
simultaneous blockade of all three phases of the cell cycle.
Cell proliferation is reduced
The reduction in cell number may have been a result of an
inhibition of cellular proliferation. Therefore, at the end of
each of the culture schedules, cells were pulsed with BrdU
for 30 mins. The extent of BrdU incorporation was then
measured by flow cytometry. In HCT116 cells, there were
no significant differences in the measured level of BrdU
incorporation and the expected level (Fig. 3). In contrast,
there was a significant reduction in BrdU fluorescence in
GEO cells cocultured with CDO and cytotoxic drugs
compared to those treated with drugs separately (Fig. 3).
This was most apparent for BrdU incorporation in cells
cocultured with CDO and 5-FU (% BrdU incorporation
normalized to control cells with SO: 72.6 ± 4.2% in cells
treated with both drugs vs. 90.2 ± 4.6% and 98.9 ± 0.8%
in cells treated with the two individually).
Cell growth arrest is protracted
The extent of treatment-induced growth-arrest was investi-
gated in GEO cells only, as inhibition of cell proliferation

was not seen in the HCT116 cells. At the end of the
treatment schedules, the surviving fractions of GEO cells
were plated in short-term semisolid cultures in the absence
of drug, and colony formation was assessed on day 10. The
total number of colonies seen in control plates containing
untreated cells was 254.2 ± 18.9, which was not signifi-
cantly different from the number seen in plates pretreated
with SO alone (251.9 ± 32.1; P ¼ 0.809). However, the
number of colonies in the CDO-treated samples was
significantly less than that observed in the control plates
(180.3 ± 33.1; P < 0.001; a reduction of around 71
colonies). Colony numbers were also reduced in the plates
Fig. 3. Effect of combining CDO with cyto-
toxic agents in HCT116 and GEO cells. The
numerical sum of BrdU incorporation into
cultures containing CDO or any of the cyto-
toxic agents alone (expected) was compared to
the observed level of incorporation in cultures
with the two agents used simultaneously
(observed). For example, the extent of BrdU
incorporation into cells treated with CDO
alone and into cells treated with VP16 alone
was summed, and compared to the extent of
BrdU incorporation into cells treated with
both CDO and VP16 together. Each column
represents the mean and SDs of at least three
separate experiments; P-values were calcula-
tedfrompairedStudent’st-tests.
Ó FEBS 2004 CRE-decoys in colorectal cancer cells (Eur. J. Biochem. 271) 2777
pretreated with cytotoxic drugs (e.g. 163.7 ± 23.4 in SN38-

treated cells; P < 0.001; a reduction of around 88 colonies)
(Fig. 4A). Therefore the expected reduction in colony
number caused by coculturing with the two agents was
159. However, the observed number of colonies was actually
49 ± 13.4—a reduction of around 202 that was signifi-
cantly greater than the calculated expected reduction,
consistent with an enhanced suppression in growth
(Fig. 4B).
Combining CDO with cytotoxic drugs increases
SA-b-gal activity
As combination treatment induced a protracted reduction
in cell number, and did not induce significantly more
apoptosis, we sought evidence for senescence by assessing
SA-b-gal activity. In control and SO-treated cells, % SA-
b-gal positive cells after a 3-day culture were 13.3 ± 5.2%
and 11.7 ± 4.1%, respectively, and increased slightly
following culture with CDO alone (22.5 ± 5.2%; P ¼
0.027 vs. SO-treated cells). Similarly, coculturing cells with
cytotoxic drugs and SO alone also increased SA-b-gal
staining slightly compared to the SO-control (Fig. 5).
Concurrent culture of CDO with any cytotoxic drug
resulted in further increases in SA-b-gal staining that were
significantly greater than those seen in cells cultured with SO
and drug (all P < 0.001), indicating a synergistic effect of
CDO and cytotoxic drug in inducing senescence (Fig. 5).
Cyclin-associated CDK protein levels are reduced
Whole cell lysates from GEO and HCT116 cells treated with
drugs were separated by electrophoresis and immuno-
probed for p21
waf1

, cyclin B, cyclin D, CDK 1 and CDK 4.
Combining CDO with any chemotherapeutic drug resulted
in changes in protein levels that appeared to be similar,
irrespective of the chemotherapy used. Consequently, the
effects of 5-FU on the cell lines are presented, which were
most representative of the effects seen with any of the three
drugs studied.
p21
waf1
levels did not change after culturing with any
combination of drug in HCT116 cells, but were significantly
increased in GEO cells treated with the combination of
CDO and 5-FU (Fig. 6A). In both HCT116 and GEO cells,
there were no significant changes in the levels of CDK 4,
cyclin D or CDK 1 in response to drug combinations. Only
cyclin B appeared to be affected, its level being reduced in
combination cultures (Fig. 6B). This observation was con-
fusing, did not correlate with the flow cytometry data (no
G2-block was observed; Fig. 2B), and was not associated
with a respective change in its partner CDK 1. As the
association of CDKs with their partner cyclins is crucial to
function, we sought to resolve this by measuring the levels of
CDK 1 and CDK 4 coimmunoprecipitating with the
respective anticyclin antibody (Fig. 6C). GEO cells (blots:
i–ii) and HCT116 cells (blots: iii–iv) were cultured with
5-FU in the presence or absence of CDO, and results
showed that the amount of each CDK precipitating with
their respective cyclin was significantly reduced only in the
GEO cell line.
Discussion

This study was undertaken to determine whether combining
CDO with conventional chemotherapeutic drugs might
have synergistic anticancer effects in colorectal cancer cell
lines. We confirmed that CDO was cytotoxic in two of the
cell lines studied at nanomolar concentrations. Additionally,
we showed that combining CDO with chemotherapeutic
drugs resulted in enhanced inhibition of cell proliferation,
which was associated with an increase in p21
waf1
expression,
loss of CDK function, and the generation of cells with
senescence characteristics.
In the first part of the investigation, we determined the
effect of continuous exposure to CDO on cell viability and
growth. IC
50
values showed that CDO was an effective
cytotoxic drug in HCT116 and GEO cells (300 n
M
and
Fig. 4. Effect on colony formation of combining CDO with cytotoxic
agents. GEO cells were cultured for 4 days with CDO in the presence
or absence of SN38 (S), 5-FU or etoposide. Equal number of cells were
removedfromeachoftheseculturesandplatedontomethylcellulose
for assessment of colony numbers on day 10. (A) Typical magnitude of
colony numbers seen in plates, using SN38 as an example. (B) The
differences in colony number (respective to controls) following treat-
ment were calculated (expected), and compared with the actual
(observed) numbers. Each column represents the mean and SDs of
at least four separate experiments. *P<0.05, between the expected

and observed.
2778 W. M. Liu et al. (Eur. J. Biochem. 271) Ó FEBS 2004
360 n
M
, respectively), but ineffective in the more resistant
SW620 cells (>1600 n
M
). Results showed dose-dependent
decreases in cell number and concomitant decreases in cell
viability in the sensitive cell lines. Flow cytometric analysis
showed that this cell death was not specific to any particular
phase of the cell cycle, and was associated with an increase
in the sub-G
1
(apoptotic) portion of the cell cycle.
We next investigated the effect of combining CDO with
the chemotherapeutic drugs 5-FU, SN38 and etoposide in
the two CDO-sensitive cell lines. There was neither
enhancement of cell death nor a greater reduction in the
number of HCT116 cells when CDO was combined with
any chemotherapeutic drug. However, CDO/chemotherapy
combinations in GEO cells resulted in significant reductions
in cell number. This was further investigated by assessing
BrdU incorporation. Results confirmed the synergistic
reduction in cell number, and were in agreement with the
observation that CDO has widespread effects on genes
controlling cell proliferation [26].
We then investigated the extent of the cell growth
inhibition in longer-term clonogenic assays. As the syner-
gistic arrest in cell growth was observed only in GEO cells,

these studies were performed in this cell line only. Results
confirmed significantly enhanced decreases in the number of
colonies cultured with both CDO and cytotoxic drugs,
compared to the reductions observed in cells cultured with
the drugs individually. This suggested a protracted effect of
CDO in combination with chemotherapy, so we stained
cells for SA-b-gal activity, and showed that CDO alone did
not increase the extent of staining. However, in cells that
had been cocultured with CDO and a cytotoxic drug,
staining was significantly increased, indicating the presence
of senescence. Our data are consistent with a model in which
CDO induces a sustained arrest (senescence). Others have
shown that senescence is mediated in part by p21
waf1
activation [24,27]. Therefore, we assessed p21
waf1
levels in
GEO and HCT116 cell lines, and showed that combination
therapy induced p21
waf1
in the GEO cell line only (the cell
line in which synergistic effects were observed). A possible
explanation for this difference could be the higher basal
p21
waf1
levels in the HCT116 cell line compared to the GEO
cell line, suggesting a possible fault in their pathway. Hence
treatment with CDO would make HCT116 cells both
less likely to respond with an increase in p21
waf1

,orto
mount a functional p21
waf1
response. This requires further
investigation.
The correct binding of cyclins to CDKs is required for
successful cell cycle progression [28,29]; for example, the
correct formation of cyclin D/CDK 4 complex is required
for pRb hyper-phosphorylation, in order for it to release
transcription factors necessary for G
1
/S transition. Cells
arrest if this complex is not formed. It has been suggested
that CDO reduces pRb hyper-phosphorylation through
p53 stabilization [18]. This would cause an increase in
p21
waf1
and induce cell cycle arrest through its CDK-
inhibitory function [30]. Our results showed that absolute
CDK 1 and CDK 4 levels in HCT116 and GEO cells
were unchanged after treatment with CDO and cytotoxic
drug. Small changes were seen in cyclin B levels in some
of the treatments; the significance of which was unclear.
However, as it is generally accepted that the heterodime-
rization of catalytic CDKs with the cyclin subunits is the
major determinant of cell cycle fate, rather than their
absolute numbers in isolation [31,32], we next performed
immunoprecipitation assays in an attempt to clarify the
relationship between CDKs, cyclins and cell cycling follo-
wing combination treatments. Results showed that levels

of CDKs associated with their respective cyclins were
significantly reduced, but only in the GEO cell lines where
p21
waf1
expression was increased by treatment. This
suggests that protracted inhibition of cell growth was
mediated through reduced cyclin/CDK function. There
have only been two studies reporting CDO-induced
inhibition of cyclin/CDK operation and cell proliferation
[18,19], which were in concordance with our results, and
highlighted a modulatory effect of CDO on cell cycle
progression. However, in contrast with our results, these
studies showed a reduction in cyclin D and E expression.
Disappointingly, the specific interactions between CDKs
and cyclins were not assessed, and the effects of CDO on
p21
waf1
were not investigated, making direct comparison
with our data more difficult.
Fig. 5. Effect of CDO and cytotoxic agents on
SA-b-gal staining. GEO cells were cultured
with VP16, 5-FU or SN38 and CDO or the
scrambled mismatch oligonucleotide (SO)
control, before staining for SA-b-gal activity.
Each column represents the mean and SDs of
six separate; P-values were calculated from
paired Student’s t-tests.
Ó FEBS 2004 CRE-decoys in colorectal cancer cells (Eur. J. Biochem. 271) 2779
In summary, these data provide a possible model of drug
interaction in GEO cells, which involves the complex

interaction of proteins involved in cell cycle regulation. To
recapitulate, combining CDO with cytotoxic chemotherapy
reduced CDK activity. This ultimately resulted in reduced
cell cycling, which was manifest as a general reduction in cell
proliferation and the appearance of senescence characteris-
tics. The activation of p21
waf1
appeared to play an
important role in mediating this effect, as we also showed
that an inability to upregulate this protein, as seen in
HCT116 cells, resulted in the absence of enhanced cell
growth inhibition. The central role of p21
waf1
in mediating
senescence is currently being investigated in isogenic cell
lines by gene expression profile methodologies, and will
form the basis of a future publication. Nevertheless, it
appears that clinically, the cytostatic ability of CDO could
improve and enhance the conventional cytotoxic effect of
other chemotherapies in some cancers. Any synergistic
effect however, may be independent of pathways controlling
p21
waf1
expression.
Fig. 6. The effect of CDO and 5-FU on cell cycle regulating proteins. The cytotoxic agents appeared to have similar effects on the proteins studied;
therefore, immunoblots from only the CDO and 5-FU combination are presented. The patterns of protein expression following culture with CDO
and 5-FU together were similar in GEO and HCT116 cells. (A) The only protein that was expressed differently in the cell lines was p21
waf1
,which
was increased in GEO but unchanged in HCT116 cells. (B) Standard immunoblots for cyclin and CDK levels in GEO cells. (C) Immuno-

precipitation experiments highlighting cyclin/CDK interactions in the GEO cell line (i–ii) and the HCT116 cell line (iii–iv). The results of
densitometry analyses are given in (A) and (B), and are expressed as a percentage of each individual control.
2780 W. M. Liu et al. (Eur. J. Biochem. 271) Ó FEBS 2004
Acknowledgements
We thank Prof. Yoon Cho-Chung for supplies of the CRE decoy
oligonucleotides and Dr Gianpaolo Tortora for the provision of the
GEO cell line. We also thank Dr Simon Joel for helpful discussions.
This work was supported by the New Drug Study Group discretionary
fund.
References
1. Cho-Chung, Y.S., Park, Y.G. & Lee, Y.N. (1999) Oligonucleo-
tides as transcription factor decoys. Curr. Opin. Mol. Ther. 1,386–
392.
2. Wang, H., Prasad, G., Buolamwini, J.K. & Zhang, R. (2001)
Antisense anticancer oligonucleotide therapeutics. Curr. Cancer
Drug Targets 1, 177–196.
3. Cho-Chung, Y.S. (2002) Antisense DNAs as targeted therapeutics
for cancer: no longer a dream. Curr. Opin. Invest. Drugs 3,934–
939.
4. Cho-Chung, Y.S. (2003) Antisense DNAs as targeted genetic
medicine to treat cancer. Arch. Pharm. Res. 26, 183–191.
5. Stephens, A.C. & Rivers, R.P. (2003) Antisense oligonucleotide
therapy in cancer. Curr. Opin. Mol. Ther. 5, 118–122.
6. Bergan, R., Connell, Y., Fahmy, B., Kyle, E. & Neckers, L. (1994)
Aptameric inhibition of p210bcr-abl tyrosine kinase autopho-
sphorylation by oligodeoxynucleotides of defined sequence and
backbone structure. Nucleic Acids Res. 22, 2150–2154.
7. Smetsers, T.F., Linders, E.H., van de Locht, L.T., de Witte, T.M.
& Mensink, E.J. (1997) An antisense Bcr-Abl phosphodiester-
tailed methylphosphonate oligonucleotide reduces the growth of

chronic myeloid leukaemia patient cells by a non-antisense
mechanism. Br. J. Haematol. 96, 377–381.
8.Juliano,R.L.,Alahari,S.,Yoo,H.,Kole,R.&Cho,M.
(1999) Antisense pharmacodynamics: critical issues in the trans-
port and delivery of antisense oligonucleotides. Pharm. Res. 16,
494–502.
9. Juliano, R.L. & Yoo, H. (2000) Aspects of the transport and
delivery of antisense oligonucleotides. Curr. Opin. Mol. Ther. 2,
297–303.
10. Gewirtz, A.M. (1999) Oligonucleotide therapeutics: clothing the
emperor. Curr. Opin. Mol. Ther. 1, 297–306.
11. Jahagirdar, B.N., Miller, J.S., Shet, A. & Verfaillie, C.M. (2001)
Novel therapies for chronic myelogenous leukemia. Exp. Hematol.
29, 543–556.
12. Tomita, N., Morishita, R., Tomita, T. & Ogihara, T. (2002)
Potential therapeutic applications of decoy oligonucleotides.
Curr. Opin. Mol. Ther. 4, 166–170.
13. Roesler, W.J., Graham, J.G., Kolen, R., Klemm, D.J. & Mcfie,
P.J. (1995) The cAMP response element binding protein synergizes
with other transcription factors to mediate cAMP responsiveness.
J. Biol. Chem. 270, 8225–8232.
14. Shaywitz, A.J. & Greenberg, M.E. (1999) CREB: a stimulus-
induced transcription factor activated by a diverse array of
extracellular signals. Annu. Rev. Biochem. 68, 821–861.
15. Montminy, M.R. & Bilezikjian, L.M. (1987) Binding of a nuclear
protein to the cyclic-AMP response element of the somatostatin
gene. Nature 328, 175–178.
16. De Cesare, D., Fimia, G.M. & Sassone-Corsi, P. (1999) Signaling
routes to CREM and CREB: plasticity in transcriptional activa-
tion. Trends Biochem. Sci. 24, 281–285.

17.Park,Y.G.,Nesterova,M.,Agrawal,S.&Cho-Chung,Y.S.
(1999) Dual blockade of cyclic AMP response element-(CRE) and
AP-1-directed transcription by CRE-transcription factor decoy
oligonucleotide. gene-specific inhibition of tumor growth. J. Biol.
Chem. 274, 1573–1580.
18. Lee, Y.N., Park, Y.G., Choi, Y.H., Cho, Y.S. & Cho-Chung, Y.S.
(2000) CRE-transcription factor decoy oligonucleotide inhibition
of MCF-7 breast cancer cells: cross-talk with p53 signaling path-
way. Biochemistry 39, 4863–4868.
19. Park, Y.G., Park, S., Lim, S.O., Lee, M.S., Ryu, C.K., Kim, I. &
Cho-Chung, Y.S. (2001) Reduction in Cyclin D1/CDK4/Retino-
blastoma Protein Signalling by CRE-Decoy Oligonucleotide.
Biochem. Biophys. Res. Commun 281, 1213–1219.
20. Alper, O., Bergmann-Leitner, E.S., Abrams, S. & Cho-Chung,
Y.S. (2001) Apoptosis, growth arrest and suppression of inva-
siveness by CRE-decoy oligonucleotide in ovarian cancer cells:
protein kinase A downregulation and cytoplasmic export of CRE-
binding proteins. Mol. Cell. Biochem. 218, 55–63.
21. Henry, S.P., Geary, R.S., Yu, R. & Levin, A.A. (2001) Drug
properties of second-generation antisense oligonucleotides: how
do they measure up to their predecessors? Curr. Opin. Invest.
Drugs 2, 1444–1449.
22. Fisher, A.A., Ye, D., Sergueev, D.S., Fisher, M.H., Shaw, B.R. &
Juliano, R.L. (2002) Evaluating the specificity of antisense oligo-
nucleotide conjugates: a DNA array analysis. J. Biol. Chem. 277,
22980–22984.
23. Liu, W.M., Oakley, P.R. & Joel, S.P. (2002) Exposure to low
concentration of etoposide reduces the apoptotic capability of
leukaemic cell lines. Leukemia 16, 1705–1712.
24. te Poele, R.H., Okorokov, A.L., Jardine, L., Cummings, J. & Joel,

S.P. (2002) DNA damage is able to induce senescence in tumor
cellsinvitroandinvivo.Cancer Res. 62, 1876–1883.
25. Liu, W.M., Stimson, L.A. & Joel, S.P. (2002) The in vitro activity
of the tyrosine kinase inhibitor STI571 in BCR-ABL positive
CML cells: Synergistic interactions with anti-leukaemic agents.
Br.J.Cancer86, 1472–1478.
26. Cho, Y.S., Kim, M.K., Cheadle, C., Neary, C., Park, Y.G.,
Becker, K.G. & Cho-Chung, Y.S. (2002) A genomic-scale view of
the cAMP response element-enhancer decoy: a tumor target-based
genetic tool. Proc. Natl Acad. Sci. USA 99, 15626–15631.
27. Han, Z., Wei, W., Dunaway, S., Darnowski, J.W., Calabresi, P.,
Sedivy,J.,Hendrickson,E.A.,Balan,K.V.,Pantazis,P.&Wyche,
J.H. (2002) Role of p21 in apoptosis and senescence of human
colon cancer cells treated with camptothecin. J. Biol. Chem. 277,
17154–17160.
28. Sherr, C.J. (1993) Mammalian G1 cyclins. Cell 73, 1059–1065.
29. Nurse, P. (1994) Ordering S phase and M phase in the cell cycle.
Cell 79, 583–589.
30. El-Deiry (1998) p21/p53, cellular growth control and genomic
integrity. Curr. Top. Microbiol. Immunol. 227, 121–237.
31. Mcdonald, E.R. & El-Deiry, W.S. (2000) Cell cycle control as a
basis for cancer drug development. Int. J. Oncol. 16, 871–886.
32. Murray, A.W. (2004) Recycling the cell cycle: cyclins revisited.
Cell 116, 221–234.
Ó FEBS 2004 CRE-decoys in colorectal cancer cells (Eur. J. Biochem. 271) 2781