Full text of DOI:10.1023/A:1017371013487

Breast Cancer Research and Treatment 63: 199–212, 2000.
© 2000 Kluwer Academic Publishers. Printed in the Netherlands.
Effects of Bcl-2 modulation with G3139 antisense oligonucleotide on
human breast cancer cells are independent of inherent Bcl-2 protein
expression
Kim N. Chi1, Anne E. Wallis1, Chow Hwee Lee1, Daniel Lopez de Menezes1, Jason Sartor1,
Wieslawa H. Dragowska1, and Lawrence D. Mayer1,2
1Department of Advanced Therapeutics, BC Cancer Agency; 2Faculty of Pharmaceutical Sciences, University of
British Columbia, Vancouver, British Columbia, Canada
Key words: antisense oligodeoxynucleotides, antineoplastic agents, apoptosis, Bcl-2, breast cancer, chemosensiti-
zation
Summary
We have investigated the effects of transient Bcl-2 down-regulation induced by the Bcl-2 antisense oligodeoxy-
nucleotide (ODN) G3139 (Genta Incorporated) in high Bcl-2 protein expressing, estrogen receptor (ER) positive
MCF-7 and low Bcl-2 expressing, ER negative MDA435/LCC6 human breast cancer cells. Treatment with Bcl-2
antisense ODN in vitro caused > 80% reduction of Bcl-2 protein levels in a sequence specific manner for both
cell lines. Maximum mRNA reduction was achieved within 24 h of the first antisense ODN exposure whereas full
protein down-regulation required antisense exposure over 48 h. This Bcl-2 reduction was associated with 80–95%
loss of viable cells compared to untreated cells. Similar cytotoxic effects were observed in both cell lines despite a
nine-fold intrinsic difference in Bcl-2 protein expression suggesting that the relative degree of down-regulation of
Bcl-2 is more important than the absolute reduction. Cell death associated with G3139 exposure exhibited prop-
erties indicative of apoptosis such as mitochondrial membrane depolarization and caspase activation. Combined
treatment with G3139 and cytotoxic agents resulted in additive cytotoxicity in both cell lines. However, under
most conditions studied, the direct cytotoxic activity of G3139 antisense was not synergistic with the cytotoxic
agents. These results suggest that while Bcl-2 clearly constitutes an attractive therapeutic target due to its role in
regulating apoptosis in breast cancer cells, additional mechanisms are important in the control of apoptosis arising
from exposure to anticancer agents in vitro.
Introduction
gene transfection can lead to increased chemosensit-
ivity [1–3]. Increased levels of Bcl-2 have been cor-
related with inhibition of mitochondrial membrane
depolarization and other indicators of apoptosis activ-
ation [4, 5].
While the studies described above suggest thera-
peutic implications of Bcl-2 manipulation in cancer
treatment, it is unlikely that a gene therapy approach
would be workable for widespread chemosensitization
of disseminated tumor cells. However, temporary re-
duction of Bcl-2 protein levels in tumor cells using an-
tisense oligonucleotides (ODN) may represent a more
feasible option to manipulate Bcl-2 expression due to
The role of Bcl-2 as an important pro-survival factor
has been associated with aggressive tumor phenotypes
as well as resistance to various treatment modalities
including chemotherapy and radiation [1]. Given that
the majority of anticancer drugs exert their cytotoxic
effects through the induction of apoptosis, Bcl-2 has
been implicated as a potential mediator of the mul-
tidrug resistance phenotype. This has been supported
by reports that increasing Bcl-2 expression through
gene transfection can protect cells from some cyto-
toxic agents whereas decreasing Bcl-2 by antisense

200
KN Chi et al.
the fact that ODNs are better suited for systemic ad-
ministration compared to many types of gene therapy.
In this context, studies have demonstrated that Bcl-2
antisense ODN treatment alone can induce apoptotic
tumor cell death in cell culture and preclinical tumor
models [6–11]. Furthermore, there is evidence sug-
gesting that phosphorothioate antisense ODN-induced
Bcl-2 down-regulation can chemosensitize tumors in
vivo [6, 9, 10]. However, the fact that Bcl-2 an-
tisense ODNs exhibit inherent cytotoxic activity as
well as possible chemosensitizing properties on these
high Bcl-2 expressing tumor lines complicates resolv-
ing the degree to which Bcl-2 down-regulation and
chemotherapy act synergistically.
In contrast to ODN treatment, Bcl-2 antisense
gene transfected tumor cell systems commonly used
for in vitro analysis of Bcl-2 manipulation effects
are selected on the basis of stable surviving clones
which may not exhibit the cytotoxic properties of
Bcl-2 down-regulation. Compensatory mechanisms
responding to changes in Bcl-2 levels could poten-
tially conceal important processes relevant to Bcl-2
functions in apoptosis control that would be normally
present. Therefore, these lines may not allow the direct
consequences of transient alterations in Bcl-2 levels
to be examined, thereby confounding in vitro/in vivo
correlations.
Materials and methods
Cell lines
MCF-7 (ER positive) and MDA-435/LCC6 (ER negat-
ive) human breast cancer cell lines were obtained from
the NCI (Bethesda, MD) and Dr. R. Clarke (Geor-
getown Univ.), respectively. Cells were maintained in
RPMI or DMEM, respectively, with 10% fetal bovine
serum at 37◦C and 5% CO2.
Antisense oligodeoxynucleotides (ODNs)
Antisense ODNs directed against the open reading
frame of bcl-2 mRNA (50-TCTCCCAGCGTGCGCC-
AT-30 , G3139, Genta Incorporated, San Diego, CA)
were introduced into the breast cancer cells using
liposomes composed of 50%/50% (by wt.) di-
oleoyldimethylammonium chloride (DODAC)/dio-
leoylphosphatidyl-ethanolamine (DOPE) purchased
from Northern Lipids (Vancouver, BC). To ensure
sequence specificity, both a sham (buffer) treat-
ment (control) and a 2-base mismatch ODN (50-
TCTCCCAGCATGTGCCAT-30, G4126, Genta Incor-
porated, San Diego, CA) were used as controls.
Monolayers of cells in exponential growth were incub-
ated with the liposome-ODN complex in serum-free
media for 4 h every 24 h on two consecutive days.
In the studies presented here, we have investigated
the effects of transient Bcl-2 down-regulation using
the phosphorothioate antisense ODN G3139 in estro-
gen receptor (ER) positive MCF-7 and the lower Bcl-2
expressing ER negative MDA435/LCC6 human breast
cancer cells. These cell lines were selected to reflect
the clinical observation that ER positive tumors tend
to have high expression of Bcl-2 while ER negative
tumors have lower expression [19, 20].
G3139 is directed against the first six codons fol-
lowing the translation initiation site of Bcl-2 mRNA
and has been shown to effectively reduce both Bcl-2
mRNA and protein expression in a sequence spe-
cific manner [12]. Various molecular and cell survival
analyses were performed on tumor cells exposed to
G3139 alone and in combination with a variety of ap-
optosis inducing agents in order to resolve the direct
cytotoxicity and chemosensitizing potential of Bcl-
2 antisense ODN treatment in high and low Bcl-2
expressing breast cancer cell lines. The results of
these studies provide useful information regarding
the laboratory investigation of Bcl-2 antisense ODNs
for therapeutic purposes in breast and other types of
cancer.
Western analysis
Bcl-2, Bcl-XL, Bax, and Mcl-1 protein expression
was evaluated by western analysis using commer-
cially available antibodies. Cell lysates were pre-
pared by washing cells once with ice cold phos-
phate buffered saline followed by the addition of
lysis buffer (50mM Tris, 150mM NaCl, 0.1% SDS,
0.5% sodium deoxycholate, 1% NP-40, 0.02% so-
dium azide) containing CompleteTM, mini protease
inhibitor cocktail (Boehringer Mannheim/Roche Dia-
gnostics, Indianapolis, IN). Samples were centrifuged
and the supernatant stored at −70◦C. Lysates equal-
ized for protein content were separated in 0.1% SDS,
12.5% polyacrylamide gels and transferred to Im-
mobilon polyvinylidene difluoride or nitrocellulose
membranes. Membranes were blocked (1% BSA,
0.02% Tween-20, 0.05% sodium azide, and 1X PBS
for Immobilon membranes; 5% milk, 0.05% sodium
azide, and 1X TBS for nitrocellulose membranes) then
probed with antibodies for Bcl-2 (DAKO, Carpinteria,
CA), Bcl-X (Santa Cruz Biotechnology Inc., Santa
Cruz, CA), Bax (Santa Cruz Biotechnology Inc., Santa

Bcl-2 antisense ODN effects on breast cancer cells
201
Cruz, CA) and Mcl-1 (Santa Cruz Biotechnology
Inc., Santa Cruz, CA). Blots were washed, incubated
with horseradish peroxidase-conjugatedsecondary an-
tibodies (DAKO), then visualized using an enhanced
chemiluminescence western blotting analysis system
(Amersham Pharmacia Biotech).
Mitochondrial membrane potential measurement and
cell cycle analyses
Alterations in mitochondrial membrane potential
following ODN treatment was monitored using a
cationic lipophilic dye 3,30-dihexiloxa dicarbocyanine
(DiOC6) [13, 14]. At selected times after ODN treat-
ment, 5×105 cells were harvested and incubated with
DiOC6 (40nM) for 30min at 37◦C. In some experi-
ments, simultaneous cell cycle analysis was evaluated
using Hoechst staining (20µg/ml) for 30 min at 37◦C.
After incubation with appropriate dyes, cells were
washed twice with cold PBS containing 0.1% BSA
(PBS-B), re-suspended in PBS containing 0.5 µg/ml
PI, and were immediately analyzed by flow cytometry
(Coulter Epics, Miami, FL). Changes in the membrane
potential were normalized for the cellular mitochon-
drial mass by staining with 100nM nonylacridine
orange (NAO).
RNase protection assays
Total RNA was isolated using Trizol (Life Technolo-
gies, Rockville, MD). Assays were carried out using
the RiboQuant multiprobe (Pharmingen, San Diego,
CA) protection assay system with the template set
hApo-2c as described in the manufacturer’s protocol.
Quantitation was determined using phosphoimaging
analysis (Phosphoimager and ImageQuant v5.0 soft-
ware, Molecular Dynamics, Sunnyvale, CA).
For cell cycle analyses, 1×106 cells were washed
with ice-cold PBS and then slowly permeabilized
with 70% ethanol, dropwise at 4◦C for 30 min. Cells
were then washed with PBS-B and RNAase-treated
(1 mg/ml, DNAase-free) for 30 min. at 37◦C. Finally,
cells were stained with 50 µg/ml PI in PBS and the
PI was allowed to equilibrate in cells on ice for 30 min
prior to analysis on flow cytometer (Coulter). Files of a
minimum of 10,000 ungated events were collected and
later analyzed using Expo analysis software (Coulter).
Viability assays
Equal numbers of cells (2–5×103) were seeded in 96-
well microtiter plates and allowed to adhere overnight.
After ODN treatments, 48–72h exposures to varying
concentrations of cytotoxics (doxorubicin, paclitaxel,
cisplatin) and direct apoptosis inducing agents (C6
ceramide) were performed. Direct counts were per-
formed utilizing hemocytometer and trypan blue dye
exclusion. Cell viability after ODN and/or cytotoxic
treatment was also assessed for total number of viable
cells using a conventional 3-(4,5-Dimethylthiazol-
2-yl)-2,5-diphenyl tetrazolium bromide (MTT) dye
reduction assay and optical density determined at
570 nm. Clonogenicity was assessed using plating ef-
ficiencies for serial dilutions of single cell suspensions
as detected by methylene blue staining of single adher-
ent colonies. All assays were performed in triplicate.
Statistical analysis:
Data were analyzed by the Student’s t-test or one way
analysis of variance (ANOVA). The levels of statist-
ical significance were set at p < 0.05 (two-sided). All
statistical calculations were performed using SPSS for
Windows, version 7.5.1 software.
Caspase-3 activity assay
Results
Cells were harvested at various times after ODN
treatment and caspase-3 levels in cells determined us-
ing a caspase-3 apoptosis detection kit (Santa Cruz
Biotechnology Inc., Santa Cruz, CA). Caspase-3
activity in cell lysates was determined fluorometric-
ally utilizing a synthetic substrate, DEVD-AFC (tet-
rapeptide DEVD-7-amino-4-trifluoromethylcoumarin
AFC) which is cleaved to a fluorescent AFC analog.
The measured fluorescence of free AFC (Excitation
400 nm / Emission 505nm) was used to determine the
amount of caspase-3 activated in ODN-treated cells.
Effect of transient Bcl-2 manipulation on growth
properties of ER negative and positive breast cancer
cells
The effects of the Bcl-2 antisense ODN G3139 and
its two base mismatch control (G4126) on Bcl-2 pro-
tein levels in MCF-7 and MDA435/LCC6 cells were
examined over the time course utilized for ODN
treatment and chemosensitivity assays. Preliminary
experiments revealed that ODN-DODAC/DOPE lipo-
somal complexes were necessary for significant Bcl-2

202
KN Chi et al.
Figure 1. Bcl-2 antisense ODN (G3139) decreases Bcl-2 protein expression in MCF-7 and MDA-435/LCC6 cell lines. (A) Western analysis of
intrinsic Bcl-2 protein expression in MDA-435/LCC6 (lane 1) and MCF-7 (lane 2) cells. (B) Western analysis of Bcl-2 protein in (A) MCF-7
and (B) MDA-435/LCC6 cells 48 h after exposure to sham control (C), G3139 antisense (AS) or G4126 mismatch (MM) ODN treatments.
(C) Timeline of relative Bcl-2 protein expression in (A) MCF-7 and (B) MDA-435/LCC6 cells after C, AS, or MM treatment.
protein reduction and that optimal down-regulation
was achieved using a lipid to DNA charge ratio of
1.3:1 (lipid/ODN wt. ratio of 5). Maximum Bcl-2
reduction was obtained at a G3139 concentration of
800nM in both cell lines; reduced ODN concentra-
tions led to decreased (< 50%) down-regulation and
concentrations beyond 1 µM resulted in non-sequence
specific increased cell death, consistent with toxicity
arising from cationic lipid/DNA complexes. Experi-
mental results presented herein were performed us-
ing oligonucleotides at a concentration 800nM in
both cell lines. As shown in Figure 1A, direct com-
parison between ER positive MCF-7 and ER neg-
ative MDA435/LCC6 cells revealed that the former
breast cancer cells exhibit 9-fold higher intrinsic (un-
treated control) Bcl-2 protein levels compared to the
MDA435/LCC6 cell line. Extensive reductions in Bcl-
2 protein levels are obtained for both MCF-7 and

Bcl-2 antisense ODN effects on breast cancer cells
203
MDA435/LCC6 breast cancer cells after two 4 h ex-
posure to G3139 compared to control, untreated cells
whereas minimal effects are observed with the 2-
base mismatch control ODN (Figure 1B). Densito-
metry analysis revealed that Bcl-2 levels were sup-
pressed 80–95% by G3139 compared to ≤20% down-
regulation in the presence of the mismatch control.
The sequence specific Bcl-2 effect was maintained
between 48 and 72 h after ODN treatment and protein
levels increased to within 50% of control after 96h
(Figure 1C). Western analysis for Bcl-XL, Bax, and
Mcl-1 protein revealed no significant changes in pro-
tein expression levels 24–72h after G3139 treatment
relative to mismatch ODN or untreated controls (data
not shown).
served for an additional 48 h, after which time Bcl-
2 mRNA increased back to control levels. In the
MDA435/LCC6 cells the level of bcl-2 mRNA was de-
creased by approximately 60–80% in antisense treated
cells compared with a less than 25% decrease in mis-
match treated cells. The bcl-2 mRNA reduction was
evident even after 96 h.
In MCF-7 cells, the decrease in Bcl-2 mRNA
expression was accompanied by marginal decreases
in Bcl-x and Bad expression, however, no signific-
ant changes in Bcl-w, Bak or Bax expression were
observed. There was also a trend towards increased
expression of Mcl-1 RNA in antisense treated MCF-7
cells. In MDA435/LCC6 cells, treatment with G3139
resulted in a trend towards decreased expression levels
of Bad, but the level of Mcl-1, Bcl-w, Bcl-x Bak
and Bax were unchanged. However, in all cases these
trends were not statistically significant.
Ribonuclease protection assay analysis was carried
out using the hAPO-2c probe kit (Pharmingen) which
allowed cell line-specific differences to be noted. Ex-
pression of the survival antagonist Bik was observed
in MCF-7 cells while no expression of the apop-
tosis inhibitor Bfl-1 mRNA was detected. In contrast,
for MDA435/LCC6 cells the situation was reversed
with Bfl-1 expression being clearly indicated with
no detectable expression of Bik observed. This cell
line-specific gene expression was unaffected by the
treatment with either antisense ODN or doxorubicin
(data not shown).
The levels of Bcl-2 mRNA were assessed by
ribonuclease protection assay (RPA) analysis after
ODN treatment. Phosphoimaging analysis indicated
that Bcl-2 mRNA expression was decreased by the
antisense treatment. The reduction in Bcl-2 mRNA
observed in MCF-7 cells at 24 h (Figure 2) was ob-
The direct cytostatic/cytotoxic effects of G3139 on
MCF-7 and MDA435/LCC6 breast cancer cells were
determined by monitoring the number of live cells and
percent viability before and at various times after treat-
ment with G3139, mismatch ODN or control buffers.
The growing adherent cells were exposed to ODNs
for 4 h upon initiation of the experiment and a re-
peated exposure 24 h later. As shown in Figure 3, in
MDA435/LCC6 cells, down-regulation of Bcl-2 us-
ing 800nM G3139 resulted in significant (p < 0.05)
reduction in the number of viable cells over 72 h after
initiation of treatment. Viable cell numbers for cells
treated with mismatch ODN were comparable to those
for control cells at earlier time points (p > 0.05) and
reducing the G3139 concentration below 600 nM res-
ulted in growth curves similar to mismatch ODN. At
24 h after the first exposure to G3139, cell numbers
were similar to those initially plated. However, at later
times, cell numbers were reduced by approximately
50% relative to the initial cell count and 90% relative
to cell counts for the control group (Figure 3). Sim-
Figure 2. G3139 reduces Bcl-2 mRNA expression in the MCF-7
cell line. Ribonuclease protection assay analysis of MCF-7 mRNA
24 h after treatment with sham control (C), G3139 antisense ODN
(AS) or mismatch ODN (MM).

204
KN Chi et al.
Figure 3. G3139 treatment reduces viable cell number. Direct
counts of viable MDA435/LCC6 cells utilizing trypan blue dye
exclusion after treatment with sham control (C), G3139 antisense
ODN (AS), or mismatch ODN (MM).
Figure 4. G3139 induces cytotoxic effects on MCF-7 and
MDA435/LCC6 cells. Percentage of nonviable cells utilizing direct
cell counts and trypan blue dye exclusion 48 h after sham control,
G3139 antisense or mismatch ODN treatment.
ilar results were obtained with MCF-7 cells where the
number of viable cells 48h after initiation of ODN
treatment were 14% and 76% relative to untreated
controls for G3139 and mismatch groups, respectively.
The results from cell growth determination sug-
gest that Bcl-2 down-regulation with G3139 results
in cytotoxicity rather than simply cytostasis with the
concentrations used in these experiments. This inter-
pretation was corroborated by examining the viability
of adherent cells using trypan blue dye exclusion.
Control and mismatch ODN treated cells exhibited
high cell viability where less than 12% of the adherent
cells were unable to exclude trypan blue. In contrast,
the number of non-viable MCF-7 and MDA435/LCC6
cells 48 h after G3139 treatment were 47% and 38%
of the total adherent cell population, respectively (Fig-
ure 4). In addition, replating equal numbers of viable
control, mismatch and G3139 treated cells after ODN
treatment resulted in differential growth rates for the
various groups. At 72h after replating, the number of
tumor cells in the G3139 treatment group were 42% of
that for control cells whereas the number of mismatch
treated cells were 72% of control (Figure 5).
Analysis of mechanisms associated with G3139
induced cytotoxicity
Experiments were performed to reveal the nature of
the direct cytotoxicity caused by exposing MCF-7 and
MDA435/LCC6 breast cancer cells to G3139. Partic-
ular attention was given to monitoring morphological
and molecular endpoints typically associated with ap-
optosis. A hallmark feature of apoptotic cell death
is the development of a cell morphology where the
nuclei condense and often fragment or bleb into nu-
merous vacuoles. We examined this by performing
fluorescence microscopy of DAPI stained tumor cells
exposed to G3139, mismatch ODN and untreated con-
trols in both cell lines. As demonstrated in Figure
6, DAPI fluorescence associated with control or mis-
match treated cells was strong and evenly distributed
within the nuclear compartment for the vast majority
of cells. In contrast, cells treated with Bcl-2 anti-
sense ODN exhibited more punctated fluorescence

Bcl-2 antisense ODN effects on breast cancer cells
205
9). In control cells, approximately 65% of cells ex-
hibited a characteristic live-cell DNA profile (Figure
9A, quadrant A1) with very few apoptotic cells (Fig-
ure 9A, quadrantsA2 and A3). A minor effect was
observed during or after the exposure of the cells to
mismatch ODN (Figure 9B). For G3139 treated cells,
the relative number of live cells was reduced to 17%
(Figure 9C, quadrant A1) with a significant increase
in the hypodiploid cell population (Figure 9C, quad-
rants A2 and A3) compared to untreated control or
mismatch-treated cells.
Flow cytometric cell cycle analyses using PI stain-
ing showed similar cell cycle distributions for control
and mismatch-treated MDA435/LCC6 cells (Figure
10A-a vs. b). G3139 treatment resulted in signi-
ficantly increased numbers of apoptotic cells (61%,
sub-Go/G1) (Figure 10B-c) compared to controls.
Similar results were observed for MCF-7 cells;
however, the increased fragility of these cells to pro-
cessing steps resulted in elevated non-viable cells in
control and ODN treatment groups, which precluded
quantitative comparisons. It is also important to note
that G3139 did not have any effects on mitochondrial
membrane potential, PI profile or Hoechst/PI stain-
ing in K562 cells which display negligible amounts of
Bcl-2 protein.
Figure 5. G3139 reduces growth rate in MCF-7 cells. Total num-
ber of viable cells determined by MTT dye reduction assay after
treatment with sham control (C), G3139 antisense (AS) or mismatch
(MM) ODN.
with condensed nuclei, many of which displayed fea-
tures consistent with nuclear blebbing. Cells with very
faint DAPI staining in what appeared to be nuclear
remnants could also be detected in this group. An-
other feature characteristic of apoptotic cell death is
the activation of caspase-3. MDA435/LCC6 cells ex-
posed to G3139 had 2.3 fold higher caspase-3 activity
than an equal number of mismatch ODN treated cells
and 3.9 fold higher activity than untreated controls
(p < 0.0001). Mismatch ODN treated cells showed
only marginal activation of caspase-3 relative to con-
trols (Figure 7).
Figure 8 presents flow cytometry plots of
MDA435/LCC6 cells for DiOC6 detection of the
mitochondrial membrane potential (normalized for
mitochondrial content using nonylacridine orange).
Untreated control cells are characterized by strong
DiOC6 staining indicative of a large mitochondrial
membrane potential (Figure 8A). Minimal effects
were observed during or after the exposure of the cells
to mismatch ODN (Figure 8B). In contrast, G3139
induced significant depolarization (low DiOC6 stain-
ing) in cells 8 h (data not shown) and 24 h after Bcl-2
antisense ODN treatment (Figure 8C).
Combined antisense ODN and chemotherapy
treatment
Down-regulation of Bcl-2 with the antisense ODN
leads to direct cytotoxic activity in both ER positive
and ER negative human breast cancer cells without
the need to provide additional apoptotic stimuli. How-
ever, a great deal of interest exists in the ability
of decreasing Bcl-2 levels to sensitize tumor cells
to conventional chemotherapy regimens. We there-
fore evaluated the effects of combining Bcl-2 down-
regulation through G3139 with exposure to various
anticancer drugs and the apoptosis initiating agent C6-
ceramide. In Figure 11, the number of viable cells
relative to untreated controls are presented for MCF-
7 and MDA435/LCC6 tumor cells exposed 48 h to
various concentrations of doxorubicin. Pretreatment
of cells with mismatch ODN prior to doxorubicin
exposure had a modest effect on the cytotoxicity of
the anticancer agent in MCF-7 cells (Panel A) and
minimal effects in MDA435/LCC6 cells (Panel B).
However, cells pre-treated with G3139 exhibited sig-
nificantly increased cytotoxicity when incubated with
doxorubicin. This effect was most notable at low drug
Simultaneous Hoechst 33342/PI determination for
cell cycle versus viability was also performed (Figure

206
KN Chi et al.
Figure 6. MDA435/LCC6 cells treated with G3139 display typical apoptotic morphology. Fluorescent micrographs of MDA435/LCC6 cells
0
0
stained with DNA dye 4 -6 diamino-2-phenyl indole (DAPI) and cytocentrifuged onto glass slides (Leica, 400×). Cells treated with Bcl-2
antisense G3139 (panel B) demonstrate striking apoptotic morphology (punctated fluorescence with condensed nuclei (arrow a) and faint DAPI
staining with nuclear remnants (arrow b)) compared to relatively normal MM-treated cells (panel A).
concentrations (below 0.1µM doxorubicin) where the
number of viable cells in the G3139 treated group were
40–50% that observed in control cells.
Although the results above demonstrate that com-
bining Bcl-2 antisense ODNs with anticancer drugs
such as doxorubicin can provide significant increases
in cytotoxicity over the individual agents, this in itself
does not necessarily establish the presence of synergy
between the two treatments. A series of studies were
performed where the cytotoxicity of various agents
known to induce tumor cell death through apoptosis
were combined with mismatch ODN or G3139 treat-
ment and the degree of cell kill was normalized to that
obtained with ODN exposure alone (relative viable
Figure 7. Increased casapase-3 activity in MDA435/LCC6 cells
cells = # cells
÷ # cells-drug+ODN). In this
+drug+ODN
after G3139 exposure. MDA435/LCC6 cells were treated with
either sham control (C), G3139 antisense (AS) or mismatch (MM)
ODN and the cellular activity of caspase-3 was analyzed 48 h after
using a fluorometric apoptosis detection kit.
case, exposure of MCF-7 and MDA435/LCC6 cells to
the anticancer drugs doxorubicin, paclitaxel and cis-
platin as well as the apoptosis inducing C6-ceramide

Bcl-2 antisense ODN effects on breast cancer cells
207
Figure 8. G3139 induces depolarization of the mitochondrial membrane potential. Flow cytometric histograms of MDA435/LCC6 cells for
DiOC6 detection of mitochondrial membrane potential after sham control (panel A), mismatch (panel B) or G3139 antisense (panel C) ODN
treatment.
Figure 9. G3139 treatment results in direct cell apoptosis measured by flow cytometry analyses. (column A) MDA435/LCC6 cells stained
with Hoechst (Param B) and PI (PMT4). Based on the Hoechst/PI staining, quadrants A1: shows normal cells with typical DNA profile, A2:
apoptotic and necrotic cells, and A3: late apoptotic or debris. G3139 treated cells (panel C) have reduced numbers of normal/live cells and
higher percentages of apoptotic cells as compared to sham control (panel A) or Bcl-2 mismatch ODN treatment (panel B).
did not reveal any differences in the chemosensitiv-
ity of cells that would not have died from G3139
alone (Figure 12). IC50 values for control (no ODN)
MDA435/LCC6 cells were virtually identical to those
In order to ensure that the observations above were
not a result of the culturing, drug exposure or cytotox-
icity assay conditions utilized, experiments examining
synergy between anticancer agents and G3139 in-
duced down-regulation of Bcl-2 were conducted using
a variety of experimental designs. Cytotoxicity curves
were generated for doxorubicin and paclitaxel with
ODN pretreatment using a wide variety of tumor cell
plating densities (10–80% confluent), drug exposure
obtained for tumor cells pretreated with G3139 or
∼
the mismatch ODN control (IC50 values 0.3 µM,
=
1.0 nM, 10 µM, and 8 µM for doxorubicin, paclitaxel,
cisplatin and C6-ceramide, respectively). Very similar
results were obtained with MCF-7 cells.

208
KN Chi et al.
Figure 10. Quantitative cell-cycle analysis of G3139 and doxorubicin treated cells using propidium iodide staining and flow cytometry.
MDA435/LCC6 cells were treated with either sham control (row a), mismatch (row b) or G3139 antisense ODN (row c) followed by exposure to
doxorubicin. (column A) 24 h after ODN/sham treatment and no doxorubicin exposure; (column B) after a 2-h exposure to 0.5 µM doxorubicin
following ODN/sham treatment; (column C) after a 24-h exposure to 0.5 µM docorubicin following ODN/sham treatment.
Figure 12. Pre-treatment with Bcl-2 antisense ODN does not res-
ult in a synergistic increase in the chemosensitivity of MCF-7 and
MDA435/LCC6 cells. Cells were pre-treated with sham control,
Bcl-2 antisense or mismatch ODN. After replating at equivalent cell
number and density, MCF-7 (panel A) and MDA435/LCC6 (panel
B) cells were exposed for 48-h to doxorubicin. Relative viable cells
were determined by MTT dye reduction assay.
Figure 11. Increased cytotoxic effect of doxorubicin after
pre-treatment with G3139. Equal numbers of MCF-7 (panel A)
and MDA435/LCC6 (panel B) were pre-treated with sham control,
Bcl-2 antisense or mismatch ODN. Relative viable cells were de-
termined by MTT dye reduction assays after a 48-h exposure to
doxorubicin.

Bcl-2 antisense ODN effects on breast cancer cells
209
times (4–120h, continuous and pulsed), concomit-
ant drug and ODN exposure as well as clonogenicity,
MTT and direct cell counting (using trypan blue) as-
says for determining the number of viable cells. Cyto-
toxicity curves with negligible differences in drug IC50
values between G3139, mismatch ODN and ODN-free
tumor cell groups were obtained under all culturing
conditions and drug exposure times as well as for
clonogenic and total viable cell detection assays. Also,
no differences were observed for either breast cancer
cell line when lower doses of ODN were employed
which led to attenuated Bcl-2 down-regulation (data
not shown).
In the case of concomitant exposure of the breast
cancer cells to drugs during ODN treatment (drugs
removed from media during the 4-h ODN pulses
and subsequently returned), a trend toward decreased
drug IC50 values (increased chemosensitivity) was ob-
served for G3139 treated cells (normalized to G3139
alone) compared to mismatch ODN or control cells.
However, the approximately 2-fold sensitization con-
sistently obtained under these conditions could not be
shown to be statistically significant (data not shown).
In addition, MDA435/LCC6 cells pretreated with
G3139 appeared to die faster when exposed to dox-
orubicin compared to mismatch ODN treated cells.
Specifically, the percent reduction of viable cells 8
and 24 h after addition of 5 µM doxorubicin was 23%
and 84%, respectively, for the G3139 group com-
pared to 0% and 75% for mismatch ODN treated cells.
At ≥ 48 h post drug treatment as well as in clono-
genic assays, no differences in the drug cytotoxicity
curves were observed between the two ODN treatment
groups.
Assays for endpoints indicative of apoptosis were
also completed on MDA435/LCC6 cells to allow com-
parisons of the process of drug induced cell death
with tumor cells exposed to G3139 and mismatch
ODN. In the presence of 0.5 µM doxorubicin, flow
cytometric cell cycle analysis using quantitative PI
staining showed that the relative numbers of apop-
totic cells following either sham control or mismatch
ODN treatments increased in a similar manner from
2 to 24 h (Figure 10B vs. C). At 24h, an arrest in
G2/M phase of the cell cycle was observed (Fig-
ure 10C-a,b). Increased numbers of apoptotic cells
were observed after G3139 treatment and this ef-
fect was even more pronounced in the presence of
doxorubicin (Figure 10B,C-c). Flow cytometric ana-
lysis demonstrated increasing mitochondrial mem-
brane potential depolarization in tumor cells after
addition of doxorubicin in a manner qualitatively
similar to that achieved with G3139 treatment (data
not shown).
Discussion
The studies described here were undertaken in or-
der to characterize the in vitro direct cytotoxic and
chemosensitizing effects of Bcl-2 antisense ODN
treatment in a pharmacological context and to cor-
relate such observations with molecular endpoints for
specific cellular processes associated with apoptosis.
Although transfecting tumor cells with the Bcl-2 gene
has provided mechanistic information concerning its
role in controlling apoptosis, demonstrating protection
from apoptosis with gain of Bcl-2 does not a priori
establish that down-regulation in a treatment setting
will provide therapeutic benefits. Similarly, Bcl-2 an-
tisense gene expressing transfectants may have limited
predictive value pharmacologically due to the selec-
tion of surviving clones in the face of a pro-apoptotic
manipulation. In view of these considerations, we
investigated the effects of Bcl-2 antisense ODN expos-
ure using a wide variety of culturing conditions, drug
concentrations and exposure times as well as several
methods for assessing cytotoxicity. Pharmacological
correlations focused on quantitating the absolute num-
ber of viable cells remaining after treatment since this
is ultimately the parameter that will reflect therapeutic
response.
The results obtained here demonstrate that the Bcl-
2 antisense ODN G3139 is very effective in reducing
Bcl-2 mRNA and protein levels for both ER negative
and positive human breast cancer cell lines. Protein
levels remained depressed for over 72 h after ODN ex-
posure in the MCF-7 and MDA/435/LCC6 cells and
this down-regulation was sequence specific, similar
to observations in studies employing this antisense
molecule with other cell lines [6–10]. It appears that
decreasing Bcl-2 protein levels by ≥ 50% results in
significant cell death that is mediated through ap-
optosis and also leads to decreased growth of the
surviving cell population. Very similar results were
obtained for MCF-7 and MDA435/LCC6 cell lines,
which is of interest since they display large differ-
ences in inherent Bcl-2 expression. This suggests that
the relative degree of Bcl-2 down-regulation, possibly
coupled with the baseline ratio of Bcl-2 to other pro-
or anti-apoptotic proteins, may be more important
than absolute decreases in dictating Bcl-2 antisense-

210
KN Chi et al.
induced cytotoxicity. Therefore, the therapeutic bene-
fits of Bcl-2 antisense ODNs may not be restricted to
tumor populations that are selected due to overexpres-
sion of this protein, although it is anticipated that some
finite level of protein expression would be required for
activity.
duration exposure of tumor cells to high concentra-
tions of cytotoxic agents in order to follow endpoints
of apoptosis and assays used to detect cytotoxicity
were often more qualitative in nature [2]. In other
cases, attempts to reveal chemosensitization effects
of Bcl-2 down-regulation utilized simultaneous titra-
tions of antisense ODN and chemotherapeutic agents
[16]. Although increased cell kill was shown with such
combination treatments, normalization of that data for
the survival probability of the individual agents (com-
pare Figures 10 and 11) generates cytotoxicity curves
very comparable to those obtained here.
Our results demonstrated that both G3139 and an-
ticancer drug treatments induced cell death through
classical apoptosis pathways associated with mito-
chondrial membrane depolarization, caspase activa-
tion, and nuclear condensation/blebbing. When these
two treatment modalities were combined, increased
cytotoxicity was observed as reflected by reduced
numbers of viable cells. The decreased percent viab-
ility of tumor cells observed in the combined treat-
ment suggested enhanced sensitivity to chemotherapy
induced by Bcl-2 reduction. However, from a pharma-
cological standpoint, this implied synergy should be
reflected by decreased viable cell numbers compared
to the survival probabilities arising from combining
two active cytotoxic agents. This anticipated result
was not observed in our investigations under a wide
spectrum of conditions using metabolic and clono-
genicity assays for cell survival. In general, we were
unable to demonstrate that MCF-7 and MDA/LCC6
breast cancer cells surviving G3139 treatment were
more sensitive to anticancer drugs or C6-ceramide,
all of which induced apoptosis individually. It could
be argued that the viable cells remaining after G3139
treatment do not display decreased Bcl-2 levels and
therefore would not be expected to be more chemo-
sensitive. However, the cells recovered for Bcl-2 pro-
tein determination are the same cells that are used
for cytotoxicity studies. The fact that Bcl-2 protein
levels are down-regulated by >80% at this point ar-
gues against explanations based on cell heterogeneity
with respect to Bcl-2 expression.
We are unable at this time to clearly elucidate
how G3139 induced Bcl-2 down-regulation may be
uncoupled from apoptotic cell death induced by the
anticancer agents studied here. Apoptosis induced in-
dependent of Bcl-2 has been documented in other
systems [15]. However, this does not account for
the apparent discrepancies between the investigations
here and others suggesting the presence of synergistic
Bcl-2 antisense-chemotherapy effects [7, 16, 17]. It
is important to note that most investigations in this
area have focused more on mechanistic studies utiliz-
ing Bcl-2 gene (and antisense gene) transfected cells
rather than identifying specific therapeutic applica-
tions. Furthermore, these systems often utilized short
Previous studies have provided encouraging indic-
ations that combining Bcl-2 antisense ODN treatment
with chemotherapy may result in enhanced antitu-
mor activity in vivo due to increased chemosensitivity
associated with Bcl-2 down-regulation [6, 9, 10]. Un-
fortunately, obtaining definitive evidence of synergy
in animal models is complicated by difficulties in cor-
relating tumor growth changes with actual tumor cell
kill. These complications can arise from in vivo tumor
model features such as combination treatment-related
changes in drug pharmacodistribution properties, the
presence of nonmalignant cells in the tumor mass and
increased fractions of dead tumor cells associated with
hypoxic regions within solid tumors. This is partic-
ularly important when antitumor activity is observed
for both agents being combined in vivo. For example,
in a previous study, dacarbazine treatment of SCID
mice bearing melanoma solid tumors provided a 72%
reduction in tumor mass 21 days after cell inocula-
tion and Bcl-2 antisense caused 60% tumor growth
inhibition [6]. The fact that the mean tumor weight
of animals treated with antisense and dacarbazine in
combination was decreased by 98% strongly supports
synergistic action since additive effects would predict
an 89% reduction in tumor growth. However, given
the additional environmental factors that may influ-
ence therapeutic activity in vivo, such indications must
be somewhat tempered due to the significant therapy
provided by each treatment component individually.
It is important to note that a recent study combining
a Bcl-2 ODN with mitoxantrone in a murine andro-
gen sensitive breast tumor model provided clear evid-
ence for synergistic antisense-related chemosensitiz-
ation [10]. Such synergy could be demonstrated due
to the lack of appreciable direct cytotoxicity of ODN
against this cell line with the conditions employed.
Taken together with the results reported here, future
investigations focused on chemosensitization by mod-
ulation of Bcl-2 must define experimental conditions

Bcl-2 antisense ODN effects on breast cancer cells
211
and models where reduction of Bcl-2 in itself does not
independently lead to significant apoptosis.
3. Vaux DL, Cory S, Adams JM: Bcl-2 gene promotes hema-
topoietic cell survival and cooperates with c-myc to immortal-
ize pre-B cells. Nature 335: 440–442, 1988
Ultimately, the therapeutic utility of antisense
ODNs against Bcl-2 will be dictated by the ability to
achieve significant tumor Bcl-2 down-regulation and
therapeutic activity in patients. Early indications from
clinical trials suggest that G3139 may directly induce
tumor cell death leading to therapeutic responses [18].
The strong cytotoxic effects of Bcl-2 down-regulation
demonstrated here with G3139 on breast cancer cells
exhibiting both high and low levels of inherent Bcl-
2 expression suggest that such genetic manipulations
may have therapeutic potential in this patient popu-
lation. This lack of antisense ODN dependence on
inherent Bcl-2 levels may also mean that the utility of
Bcl-2 as a therapeutic target is not directly linked to
its prognostic utility in breast cancer where its expres-
sion has been shown in some studies to be a positive
prognostic marker [19, 20]. Whether chemosensitiza-
tion can be achieved clinically is yet to be determined;
however, early results from a clinical trial of G3139
plus dacarbazine in melanoma patients with resistant
disease are encouraging [21].
The results of the studies completed here indicate
that care should be given to avoid oversimplification of
the complex apoptosis pathways that may be targeted
for therapeutic intervention. Clearly, our ability to use
in vitro and in vivo preclinical studies predictively in
the design of clinical trials with agents directed at ap-
optosis factors will require further elucidation of the
pathways governing cell death and how these controls
will compare and contrast moving from the laboratory
to human studies.
4. Decaudin D, Geley S, Hirsch T, Castedo M, Marchetti P,
Macho A, Kofler R, Kroemer G: Bcl-2 and Bcl-XL antagon-
ize the mitochondrial dysfunction preceding nuclear apoptosis
induced by chemotherapeutic agents. Cancer Res 57: 62–67,
1997
5. Miyashita T, Reed JC: bcl-2 gene transfer increases relat-
ive resistance of S49.1 and WEHI7.2 lymphoid cells to cell
death and DNA fragmentation induced by glucocorticoid and
multiple chemotherapeutic drugs. Cancer Res 52: 5407
6. Burkhard J, Schlagbauer-Wadl H, Brown BD, Bryan RN, van
Elsas A, Muller M, Wolff K, Eichler HG, Pehamberger H:
bcl-2 antisense therapy chemosensitizes human melanoma in
SCID mice. Nature Med 4: 232–234, 1998
7. Ziegler A, Luedke GH, Fabbro D, Ahmann KH, Stahel RA,
Zangemeister-Wittke U: Induction of apoptosis in small-cell
lung cancer cells by an antisense oligodeoxynucleotide tar-
geting the Bcl-2 coding sequence. J Natl Cancer Inst 89:
1027–1036, 1997
8. Cotter FE, Johnson P, Hall P, Pocock C, al Mahdi N, Cow-
ell JK, Morgan G: Antisense oligonucleotide suppress B-cell
lymphoma growth in a SCID-hu mouse model. Oncogene 9:
3049–3055, 1994
9. Yang D, Ling Y, Almazan M, Guo R, Murray A, Brown B,
Lippman ME: Tumor regression of human breast carcinoma
by combination therapy of anti-bcl-2 antisense oligonucleotide
and chemotherapeutic drugs (Abstract). Proc Amer Assoc
Cancer Res 40: 4814, 1999
10. Tolcher A, Miyake H, Gleave ME: Downregulation of Bcl-2
expression by antisense oligonucleotide (AS-ODN) treatment
enhances mitoxantrone cytotoxicity in the androgen dependent
Shionogi tumor model (Abstract). Proc Amer Assoc Cancer
Res 40: 3198, 1999
11. Reed JC, Stein C, Subasinghe C, Haldan S, Croce CM,
Yum S, Cohen J: Antisense-mediated inhibition of BCL2
protooncogene expression and leukemic cell growth and sur-
vival: comparisons of phosphodiester and phosphorothioate
oligodeoxynucleotides. Cancer Res 50: 6565–6570, 1990
12. Kitada S, Miyashita T, Tanaka S, Reed JC: Investigations
of antisense oligonucleotides targeted against bcl-2 RNAs.
Antisense Res Dev 3: 157–169, 1993
13. Marchetti P, Susin SA, Decaudin D, Gamen S, Castedo
M, Hirsch T, Zamzami N, Naval J, Senik A, Kroemer G:
Apoptosis-associated derangement of mitochondrial function
in cells lacking mitochondrial DNA. Cancer Res 56: 2033–
2038, 1996
14. Bedner B, Li X, Gorczyca W, Melamed MR, Darzynkiewicz
Z: Analysis of apoptosis by laser scanning cytometry. Cyto-
metry 35: 181–195, 1999
15. Memon SA, Moreno MB, Petrak D, Zacharchuk CM: Bcl-2
blocks glucocorticoid- but not Fas- or activation-induced ap-
optosis in a T-cell hybridoma. J Immunol 155: 4644–4652,
1995
Acknowledgements
This work was supported by a grant from the National
Cancer Institute of Canada through funds raised from
the Canadian Cancer Society. K.N. Chi was suppor-
ted by a fellowship from the Canadian Association of
Medical Oncologists/Glaxo Wellcome.
References
16. Zangemeister-Wittke U, Schenker T, Luedke GH, Stahel RA:
Synergistic cytoxicity of bcl-2 antisense oligodeoxynucleotide
and etoposide, doxorubicin and cisplatin in small-cell lung
cancer cell lines. Br J Cancer 78: 1035–1042, 1998
17. Kitada S, Takayama S, De Riel K, Tanaka S, Reed JC:
Reversal of chemoresistance of lymphoma cells by antisense-
mediated reduction of bcl-2 gene expression. Antisense Res
Dev 4: 71–79, 1994
1. Reed JC: Prevention of apoptosis as a mechanism of drug
resistance. Hem Onc Clinics North Am 9: 451–473, 1995
2. Teixera C, Reed JC, Pratt MAC: Estrogen promotes chemo-
therapeutic drug resistance by a mechanism involving Bcl-
2 proto-oncogene expression in human breast cancer cells.
Cancer Res 55: 3902–3907, 1995

212
KN Chi et al.
18. Webb A, Cunningham D, Cotter F, Clarke PA, di Stefano F,
Ross P, Corbo M, Dziewanowska Z: Bcl-2 antisense therapy
in patients with non-Hodgkin lymphoma. Lancet 349: 1137–
1141, 1997
19. Hellemans P, van Dam PA, Weyler J, van Oosterom AT,
Buytaert P, Van Marck E: Prognostic value of bcl-2 expression
in invasive breast cancer. Br J Cancer 72: 354–360, 1995
20. Silvestrini R, Benini E, Veneroni S, Daidone MG, Tomasic G,
Squicciarini P, Salvadori B: p53 and bcl-2 expression correl-
ates with clinical outcome in a series of node-positive breast
cancer patients. J Clin Oncol 14: 1604–1610, 1996
21. Jansen B, Wacheck V, Heere-Ress E, Schlagbauer-Wadl H,
Hollenstein U, Lucas T, Eichler H-GG, Wolff K, Pehamberger
H: A Phase I-II study with dacarbazine and BCL-2 antis-
ense oligonucleotide G3139 (Genta) as a chemosensitizer in
patients with advanced malignant melanoma (Abstract). Proc
Amer Assoc Clin Oncol 18: 2049, 1999
Address for offprints and correspondence: Lawrence D. Mayer,
BC Cancer Agency, 600 West 10th Avenue, Vancouver, B.C., V5Z
4E6, Canada; Tel: (604) 877-6000, ext. 3153; Fax: (604) 877-6011;
E-mail: lmayer@bccancer.bc.ca.