Content uploaded by Laurent Meijer
Author content
All content in this area was uploaded by Laurent Meijer
Content may be subject to copyright.
Autophagy: A Novel Mechanism of Synergistic Cytotoxicity between
Doxorubicin and Roscovitine in a Sarcoma Model
Laura A. Lambert,1Na Qiao,2Kelly K. Hunt,1Donald H. Lambert,1Gordon B. Mills,3
Laurent Meijer,4and Khandan Keyomarsi1,2
Departments of 1Surgical Oncology, 2Experimental Radiation Oncology, and 3Systems Biology, University of Texas - M. D. Anderson Cancer
Center, Houston, Texas; and 4Protein Phosphorylation and Disease, Centre National de la Recherche Scientifique,
Station Biologique, Roscoff, France
Abstract
Doxorubicin is a genotoxic chemotherapy agent used in treat-
ment of a wide variety of cancers. Significant clinical side
effects, including cardiac toxicity and myelosuppression,
severely limit the therapeutic index of this commonly used
agent and methods which improve doxorubicin efficacy could
benefit many patients. Because doxorubicin cytotoxicity is
cell cycle specific, the cell cycle is a rational target to enhance
its efficacy. We examined the direct, cyclin-dependent kinase
inhibitor roscovitine as a means of enhancing doxorubi-
cin cytotoxicity. This study showed synergistic cytotoxicity
between doxorubicin and roscovitine in three sarcoma cell
lines: SW-982 (synovial sarcoma), U2OS-LC3-GFP (osteosarco-
ma), and SK-LMS-1 (uterine leiomyosarcoma), but not the
fibroblast cell line WI38. The combined treatment of doxoru-
bicin and roscovitine was associated with a prolonged G
2
-M
cell cycle arrest in the three sarcoma cell lines. Using three
different methods for detecting apoptosis, our results revealed
that apoptotic cell death did not account for the synergistic
cytotoxicity between doxorubicin and roscovitine. However,
morphologic changes observed by light microscopy and
increased cytoplasmic LC3-GFP puncta in U20S-LC3-GFP cells
after the combined treatment suggested the induction of
autophagy. Induction of autophagy was also shown in SW-982
and SK-LMS-1 cells treated with both doxorubicin and
roscovitine by acridine orange staining. These results sug-
gest a novel role of autophagy in the enhanced cytotoxicity by
cell cycle inhibition after genotoxic injury in tumor cells. Fur-
ther investigation of this enhanced cytotoxicity as a treat-
ment strategy for sarcomas is warranted. [Cancer Res 2008;
68(19):7966–74]
Introduction
Sarcomas are a broad group of mesenchymal tumors that are
notoriously chemoresistant. More than 50 histological types of
sarcoma are described (1) with an overall 5-year survival for all
stages of 50% to 60% (2–4). Only 20% of sarcomas respond to
doxorubicin, which is the current standard of systemic therapy care
for these tumors (5). Unfortunately, the clinical utility of
doxorubicin, which is also used to treat a wide range of solid
and nonsolid tumors, is limited by significant side effects,
particularly irreversible cardiac toxicity (6). For these reasons,
efforts to improve the efficacy of doxorubicin are essential and
could benefit many patients suffering from a variety of cancers.
The mechanism of doxorubicin cytotoxicity involves ‘‘poisoning’’
the topoisomerase II enzyme, thereby interfering with the
separation of daughter DNA strands and chromatin remodeling.
In addition, doxorubicin intercalates into double-stranded DNA
producing structural changes that interfere with DNA and RNA
synthesis (6). Because it primarily damages double-stranded DNA,
cells in the S phase of the cell cycle are more susceptible. Because
of this cell cycle specificity, the cell cycle is a rational target for
enhancing doxorubicin efficacy.
Cyclin-dependent kinases (Cdk) are essential cell cycle regula-
tory proteins, which ensure accurate and appropriate transition
between cell cycle phases and ultimately cell division (7). In
addition, Cdk function is an important determinant of the cellular
response to DNA damage, including that caused by genotoxic
chemotherapies, such as doxorubicin (8). To date, 13 different Cdks
have been identified (Cdk1–13; ref. 9). The activity of individual Cdk
types is restricted to specific phases of the cell cycle and dependent
on binding with activating cyclin proteins (10). Because Cdks are
involved in regulating many important aspects of the cell cycle,
they provide an appealing target for cell cycle-directed anticancer
therapy.
Currently, >100 direct and indirect Cdk inhibitors are available.
One of the most studied, and apparently selective, direct Cdk
inhibitors is the purine analogue, roscovitine (11). Roscovitine
directly inhibits Cdks by occupying the ATP binding site of the
catalytic subunit of the kinase protein and is relatively more
specific for Cdk1 and Cdk2 than other Cdk inhibitors (12). It is a
well-tolerated, oral agent, which has shown ability to arrest tumor
growth as a single agent in phase I clinical trials (13) and is
currently being tested in phase II clinical trials against non–small
cell lung cancer and nasopharyngeal cancer. Other preclinical
studies have shown enhanced cytotoxicity with a variety of chemo-
therapeutic agents when combined with roscovitine (14, 15).
Because roscovitine can inhibit both Cdk1 and Cdk2, it has the
potential to enhance the cytotoxicity of genotoxic chemotherapeu-
tic agents by the combined effect of delayed cell cycle progression
as well as inhibition of the DNA damage response pathway.
In the present study, we investigated the effect of the
combination of doxorubicin with roscovitine in three different
sarcoma cell lines and one immortalized fibroblast cell line. Our
results show a nonapoptotic synergistic cytotoxicity in all three
sarcoma cell lines, but not the fibroblasts, treated with both
doxorubicin and roscovitine. This synergistic cell death was
accompanied by both a prolonged G
2
-M arrest and significant
induction of autophagy.
Requests for reprints: Khandan Keyomarsi, Department of Experimental
Radiation Oncology, Unit 0066, University of Texas, M. D. Anderson Cancer Center,
1515 Holcombe Boulevard, Houston, TX 77030. Phone: 713-792-4845; Fax: 713-794-
5369; E-mail: kkeyomar@mdanderson.org.
I2008 American Association for Cancer Research.
doi:10.1158/0008-5472.CAN-08-1333
Cancer Res 2008; 68: (19). October 1, 2008 7966 www.aacrjournals.org
Research Article
Materials and Methods
Cell culture. Fetal bovine serum (FBS) and cell culture medium were
purchased from Hyclone Laboratories. All other chemicals were reagent
grade. SW-982, SK-LMS-1, and WI38 (immortalized fibroblasts) were
purchased from American Type Culture Collection and cultured in DMEM
(SW-982, SK-LMS-1) or a-MEM (WI38) with 10% FBS. U20S cells stably
transfected with LC3-GFP protein (U2OS-LC3-GFP) were kindly donated by
Dr. Gordon B. Mills (M. D. Anderson Cancer Center, Houston, TX) and
maintained in DMEM with 10% FBS. All cells were cultured and treated at
37jC in a humidified incubator containing 6.5% CO2. A 5-mmol/L stock
solution of doxorubicin (Sigma) was made in sterile water and maintained at
20jC. A 10-mmol/L stock solution of roscovitine (synthesized and kindly
provided by Drs. Herve Galons, Universite´ Rene´ Descartes, Paris, France and
Laurent Meijer, Protein Phosphorylation and Disease, Centre National de la
Recherche Scientifique, Station Biologique, Roscoff, France) was made in
DMSO and maintained at 20jC. Fresh drug was prepared for each
experiment.
Cell cycle analysis. For DNA content analysis, harvested cells were
centrifuged at 1,000 gfor 5 min, washed with PBS, and fixed in 70% ethanol.
Cells were then treated with RNase (10 Ag/mL) for 30 min at 37jC, washed
with PBS, resuspended, and stained in 1 mL of 69 Amol/L propidium iodide
(PI) in 38 mmol/L sodium citrate for 30 min. The cell cycle phase
distribution was determined by analytic DNA flow cytometry as described
by Keyomarsi and colleagues (16). The percentage of cells in each phase of
the cell cycle was analyzed using Modfit software (Verity Software House).
High-throughput clonogenic assay and combination index. 3-(4,5-
Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays were
modified for use as high-throughput clonogenic assays (HTCA) to
determine cell viability. Cells were plated in 96-well plates at a density of
1.5 to 3 10
3
cells per well. After 24 h, cells were treated with drug-free
medium, 0.001 to 0.02 Amol/L doxorubicin for 24 h, 5 to 25 Amol/L
roscovitine for 48 h, or doxorubicin for 24 h then roscovitine for 48 h. Drug-
free medium was replaced every 48 h after treatment. Six replicates were
plated for each treatment group. Ten days after plating, 50 AL of 2.5 mg/mL
MTT solution was added to each well and incubated for 4 h. The resultant
blue formazan crystals were solubilized in 100 AL of buffer (0.04 N HCl and
1% SDS in isopropyl alcohol) for 1 h. Absorbance was read at 590 nm using a
Wallac-1420 plate reader. Drug interactions were assessed using CalcuSyn
software version 2.1 (Biosoft, Inc.) to determine the combination index of
the combined treatment of doxorubicin and roscovitine.
Clonogenic assays. Cells (1 10
3
) were plated onto 100-mm
2
dishes.
After 24 h, cells were treated with fresh drug-free medium, 0.01 Amol/L
doxorubicin (SK-LMS-1), 0.005 Amol/L doxorubicin (U2OS-LC3-GFP and
SW-982) for 24 h, 20 Amol/L (SK-LMS-1) or 10 Amol/L (U2OS-LC3-GFP and
SW-982) roscovitine for 48 h, or doxorubicin for 24 h then roscovitine for
48 h. Drug-free medium was applied at the end of each treatment. After
14 d, the cells were fixed and stained with crystal violet in 100% ethanol
suspension. Plates were scored for the number of visible colonies of z2 mm.
Apoptosis assays. Both APC-conjugated Annexin V and terminal
deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays were
used to determine the presence of apoptosis. Cells (2 10
5
) were plated on
100-mm
2
plates. After 24 h, cells were treated with fresh drug-free medium,
doxorubicin for 24 h, roscovitine for 48 h, or doxorubicin for 24 h then
roscovitine for 48 h as described above. Cells were harvested and stained
with Annexin V-APC and PI (Trevigen, Inc.) or by TUNEL using the APO-
DIRECT kit (BD Biosciences PharMingen) according to the manufacturer’s
instructions (Trevigen, Inc.). Apoptosis was detected by flow cytometry.
Western blot analysis for apoptosis. For Western blot analysis, cells
were homogenized by sonication and high-speed centrifugation. Cell lysate
supernatant was assayed for total protein content and subjected to Western
blot analysis as described by Rao and colleagues (17). Briefly, 25 Ag of protein
from each condition was electropheresed on a 10% SDS-polyacrylamide gel
and transferred to Immunobilon P for 2 h. Membranes were washed at 4jCin
Blotto [5% nonfat dry milk in 20 mmol/L Tris, 137 mmol/L NaCl, 0.25% Tween
(pH 7.6)] overnight. After 6 washes in TBST [20 mmol/L Tris, 137 mmol/L
NaCl, and 0.05% Tween (pH 7.6)], membranes were incubated in primary
antibody (Parp; Cell Signaling) for 2 h (1 Ag/mL in Blotto). Membranes were
washed and incubated with goat anti-mouse horseradish peroxidase
conjugate at a dilution of 1:5,000 for 1 h, washed, and developed with
Renaissance chemiluminescence system as directed by the manufacturers
(NEN Life Sciences Products).
Autophagy assays. U2OS-LC3-GFP cells (1 10
4
) were plated on
coverslips in 6-well plates. After 24 h, cells were treated with fresh drug-free
medium, doxorubicin for 24 h, roscovitine for 48 h, or doxorubicin for
24 h then roscovitine for 48 h. The cells were fixed with 4% para-
formaldehyde, stained with 4’,6-diamidino-2-phenylindole, dilactate (Invi-
trogen), and observed with fluorescence microscopy (Leica DM 4000 B).
To detect autophagy in SK-LMS-1 and SW-982 cells, 1 10
4
cells were
plated on coverslips in 6-well plates. After 24 h, cells were treated with fresh
drug-free medium, doxorubicin for 24 h, roscovitine for 48 h, or doxorubicin
for 24 h then roscovitine for 48 h. The cells were fixed with 4%
paraformaldehyde for 15 min at room temperature then stained with
acridine orange (AO; Polysciences, Inc.), 1 Ag/mL in PBS at 37jC, in the
dark, and observed immediately with fluorescence microscopy. To quantify
the number of cell with acidic vesicles, cells were seeded into 6-well plates
at a density of 4 10
4
cells per well and cultured overnight. The cells were
stained with 1 Ag/mL AO in DMEM at 37jC for 15 min. After incubation,
the cells were washed with PBS and removed with trypsin-EDTA,
resuspended, and analyzed by flow cytometry.
Statistical analysis. Each experiment was repeated at least thrice. The
isobologram analysis and graphs were carried out using Calcusyn software
(version 2.1; Biosoft, Inc.), which performs multiple drug dose-effect
calculations using the median effects method described by T-C Chou and
P. Talalay (18). Combination index (CI) values of <0.9 indicates synergy, CI of
>0.9 and <1.2 indicate additivity, and CI of >1.2 indicate antagonism.
Comparisons among groups were analyzed by two-sided ttest. A difference
of Pvalue of V0.05 was considered to be statistically significant. All analyses
were done with SPSS software, Version 12.0. The data represent the means
of three or more samples with SE.
Results
Synergistic cytotoxicity between doxorubicin and roscovi-
tine. High-throughput clonogenic assay (HTCA) was used to
compare the cytotoxic effects of doxorubicin alone, roscovitine
alone, and doxorubicin followed by roscovitine in SW-982, U2OS-
LC3-GFP, SK-LMS-1 sarcoma, and WI38-immortalized fibroblast
cell lines. A sequential drug treatment strategy (see Materials and
Methods) was chosen based on previous reports in the literature
demonstrating sequence–specific synergistic effects with adminis-
tration of combination chemotherapy, and Cdk inhibitors (19, 20).
As shown in a three-dimensional graph in Fig. 1A, dose-dependent
increases in cell death were seen in all four cell lines when treated
with doxorubicin alone (Xaxis). For SW-982, the percentage of cell
death increased from 5% to 60% as cells were treated with
increasing doxorubicin doses from 0.001 to 0.015 Amol/L (IC
50
,
f0.0125 Amol/L). For U2OS-LC3-GFP, the percentage of cell death
increased from 5% to 85% as cell were treated with increasing
doxorubicin doses from 0.001 to 0.01 Amol/L (IC
50
, between 0.005
and 0.0075 Amol/L). For SK-LMS-1, the percentage of cell death
increased from 7% to 25% as cell were treated with increasing
doxorubicin doses from 0.001 to 0.02 Amol/L (IC
50
, not reached).
For WI38, the percentage of cell death increased from 13% to 50% as
cell were treated with increasing doxorubicin doses from 0.001 to
0.01 Amol/L (IC
50
, 0.01 Amol/L). The percentage of cell death for all
three sarcoma cell lines treated with roscovitine alone ranged
between 5% and 30% as cells were treated with increasing rosco-
vitine doses from 5 to 20 Amol/L (Yaxis). For WI38, the percentage
of cell death increased from 25% to 95% as cell were treated with
increasing roscovitine doses from 5 to 25 Amol/L (IC
50
,10Amol/L).
Synergistic Cytoxicity through Autophagy
www.aacrjournals.org 7967 Cancer Res 2008; 68: (19). October 1, 2008
When the two drugs were combined, there was a significant
increase in cell death (Zaxis) in all three sarcoma cell lines
compared with either doxorubicin or roscovitine alone in isobolo-
gram determinations. However, in the WI38 cells, the combination
of doxorubicin and roscovitine did not result in such a synergistic
cell death. CIs, as determined using the CalcuSyn software, showed
synergistic cytotoxicity between doxorubicin and roscovitine in the
three sarcoma cell lines but not the WI38 cells (Fig. 1B). For
Figure 1. Synergistic effect of doxorubicin
and roscovitine in sarcoma cells. A, HTCA
was used to compare the cytotoxic effects
of doxorubicin alone, roscovitine alone,
and doxorubicin followed by roscovitine in
SW-982, U2OS-LC3-GFP, SK-LMS-1,
and WI38 cell lines [X-axis: doxorubicin
(Dox)Amol/L; Y-axis: roscovitine (Rosc)
Amol/L; Z-axis: percentage cell death].
B, isobologram analysis showed
synergistic interactions in all three sarcoma
cell lines. Isobologram analysis and
graphs were obtained using CalcuSyn
software, which performs drug dose-effect
calculation using the median effect method
described by Chou and Talalay (18).
Cand D, clonogenic assays also showed
a synergistic effect between doxorubicin
and roscovitine in the three cell lines
(Cont, untreated; D, doxorubicin; R,
roscovitine; D+R, doxorubicin plus
roscovitine). Representative of three
experiments.
Cancer Research
Cancer Res 2008; 68: (19). October 1, 2008 7968 www.aacrjournals.org
example, in SW-982 cells, treatment with either 0.005 Amol/L
doxorubicin alone or 10 Amol/L roscovitine alone resulted in a
cell death rate of only 15%. However, when treated first with
0.005 Amol/L doxorubicin then 10 Amol/L roscovitine, the
percentage of cell death increased to 55% (CI, <0.9), suggesting
synergism. Similarly, in U2OS-LC3-GFP, treatment with either 0.005
Amol/L doxorubicin alone or 10 Amol/L roscovitine alone resulted
in a cell death of 30% and 7%, respectively. However, when treated
first with 0.005 Amol/L doxorubicin then 10 Amol/L roscovitine, the
percentage of cell death increased to 75% (CI<0.9). For SK-LMS-1,
treatment with either 0.02 Amol/L doxorubicin alone or 20 Amol/L
roscovitine alone resulted in a cell death of 25%. However, when
treated first with 0.02 Amol/L doxorubicin then 20 Amol/L
roscovitine, the percentage of cell death increased to 64% (CI,
<0.9). Hence, in all three sarcoma cell lines, the combination of
doxorubicin and roscovitine resulted in synergistic cell killings. On
the other hand, for the WI38-immortalized fibroblast cells, the CIs
were predominantly antagonistic (CI, >1.2) and not synergistic. This
finding suggests that the increased cytotoxicity of the combined
treatment is tumor cell specific.
The HTCA was complemented with conventional clonogenic
assays in the sarcoma cell lines, and also showed a synergistic
effect between doxorubicin and roscovitine in all three sarcoma
cell lines (Fig. 1Cand D). For SW-982 cells, there was a 10%
decrease in the number of colonies formed after treatment with
10 Amol/L roscovitine alone, a 30% decrease after treatment with
0.005 Amol/L doxorubicin alone, and a 64% decrease in the number
of colonies formed after treatment with doxorubicin (0.005 Amol/L)
then roscovitine (10 Amol/L; PV0.006). For U2OS-LC3-GFP cells,
there was a 6% decrease in the number of colonies formed after
treatment with 10 Amol/L roscovitine alone, a 64% decrease after
treatment with 0.005 Amol/L doxorubicin alone, and a 82%
decrease in the number of colonies formed after treatment with
doxorubicin (0.005 Amol/L) then roscovitine (10 Amol/L; PV0.001).
For SK-LMS-1 cells, there was no significant change in the number
of colonies formed after treatment with 20 Amol/L roscovitine
alone. There was a 47% decrease after treatment with 0.01 Amol/L
doxorubicin alone and a 85% decrease in the number of colonies
formed after treatment with doxorubicin (0.01 Amol/L) then
roscovitine (20 Amol/L; PV0.008). Fig. 1Cshows representative
clonogenic plates from each experiment, and Fig. 1Dshows
quantization of the clonogenic assays under each condition.
Treatment with doxorubicin followed by roscovitine induces
prolonged G
2
arrest. To determine the cell cycle effects of
doxorubicin and roscovitine, both alone and in combination, DNA
content was evaluated using PI staining followed by flow
cytometry. Cell cycle distributions for the three sarcoma cell lines,
both untreated and treated, are shown in Fig. 2A. After 24 hours of
Figure 2. Effect of doxorubicin and roscovitine on cell cycle. A, SW-982 cells, U2OS-LC3-GFP cells, and SK-LMS-1 cells were treated with either 0.005 Amol/L
(SW-982 and U20S-LC3-GFP) or 0.01 Amol/L (SK-LMS-1) doxorubicin for 24 h, 10 Amol/L (SW-982 and U20S-LC3-GFP) or 20 Amol/L (SK-LMS-1) roscovitine for 48 h,
the combination of doxorubicin and roscovitine, or were untreated. At the completion of drug treatment (T1) and 72 to 96 h after the completion of drug treatment (T2),
cells were harvested and the cell cycle distribution was determined using fluorescence-activated cell sorting (B). Cell cycle distribution histograms of SK-LMS-1.
Representative of three experiments.
Synergistic Cytoxicity through Autophagy
www.aacrjournals.org 7969 Cancer Res 2008; 68: (19). October 1, 2008
treatment with doxorubicin alone, all 3 sarcoma cell lines showed
an increase in the percentage of cells in G
2
-M phase: 5% in SW-982,
6% in U20S-LC3-GFP, and 9% in SK-LMS-1 (Fig. 2B) compared with
untreated cells. After 48 hours of treatment with roscovitine alone,
there was an increase in the percentage of cells in G
1
by 13% for
SW-982. U2OS-LC3-GFP and SK-LMS-1 showed a 7% and 25%
increase in the percentage of cells in G
2
-M phase compared with
untreated cells, respectively. Combined treatment with doxorubicin
for 24 hours followed by roscovitine for 48 hours resulted in an
increase in the percentage of cells in G
2
-M phase compared with
untreated cells by 5% for SW-982, 24% for U2OS-LC3-GFP, and 45%
for SK-LMS-1. Collectively, comparison of the changes in cell cycle
distribution in response to each drug alone versus in combination
suggests significant modulation of the cell cycle pathway by the
combination of the two drugs. To assess the involvement of the cell
cycle as the modulator of the observed synergistic response, we
next performed cell cycle analysis at various time points: at the
completion of drug treatment (T1) and 72 to 96 hours after the
completion of drug treatment (T2).
As shown in Fig. 2A, untreated cells in all three sarcoma cell lines
showed persistent cell division over the time course with increased
percentage of cells residing in G
1
phase at the time of final analysis
(T2), consistent with a contact-induced G
1
cell cycle arrest
(compare control cells in T1 with T2). In addition, the percentage
of cells in G
1
phase also increased in all three sarcoma cell lines
treated with either doxorubicin or roscovitine alone, suggesting the
persistence of cell division after drug treatment (compare
doxorubicin alone cells in T1 with T2). In comparison, after
combined treatment with doxorubicin and roscovitine, SW982 cells
showed a smaller percentage of cells in G
1
compared with the
initial time point (19%) and a persistence of cells in G
2
-M phase
compared with untreated cells at the final time point (T2, 13%
versus 3% of untreated cells). For U2OS-LC3-GFP cells, there was no
significant change in the percentage of cells in G
1
compared with
the initial time point (41% versus 46%, respectively) but a
persistence of cells in G
2
-M phase compared with untreated cells
at the final time point (14% versus 2%). SK-LMS-1 cells showed an
increase in the percentage of cells in G
1
compared with the
combined treatment at the initial time point (19%) as well as a
persistence of cells in the G
2
-M phase (15% versus 3%) compared
with untreated cells at the final time point (Fig. 2B). These findings
translated into a 4-, 7-, and 5-fold increase in the percentage of
SW982, U2OS-LC3-GFP, and SK-LMS-1 cells remaining in G
2
-M
phase, respectively, compared with untreated cells at the final time
point with the combination treatment. These results suggest that
treatment by roscovitine after doxorubicin not only inhibits the cell
cycle but also the ability of the cell to recover from cytotoxic injury.
All of these experiments were performed at least thrice in triplicate
with similar results. Figure 2Bis representative of a single
experiment in the SK-LMS-1 cell line.
Synergistic cytotoxicity by doxorubicin and roscovitine is
not due to increased apoptosis. We detected unique morpho-
logical changes by light microscopy of untreated and treated cells
particularly in U2OS-LC3-GFP and SK-LMS-1 cells (Fig. 3A–B ). As
compared with untreated cells, cells treated with both doxorubicin
and roscovitine became larger, displayed prominent fibrils, and
were often multinucleated (see arrows). Based on the synergistic
cytotoxicity observed with the combined drug treatment, it was
clear that cell death was present but not through apoptosis. Three
different methods were used to determine the degree of cell death
due to apoptosis after treatment with either the single agents or
the combination: Annexin V-APC staining, TUNEL staining, and
expression of Parp (Fig. 4).
As shown in Fig. 4A, there was a significant increase in Annexin
V-APC–positive SW-982 cells after treatment with either doxoru-
bicin or roscovitine alone or in combination compared with
untreated cells (PV0.004). Roscovitine treatment alone lead to
the greatest fold increase in Annexin V-APC–positive SW-982 cells
(4-fold), and this was significant when compared with either
doxorubicin treatment alone or in combination with roscovitine
Figure 3. Morphologic changes observed by light microscopy. A, U2OS-LC3-GFP cells and (B) SK-LMS-1 cells were treated with either 0.005 Amol/L
(U20S-LC3-GFP) or 0.01 Amol/L (SK-LMS-1) doxorubicin for 24 h, 10 Amol/L (U20S-LC3-GFP) or 20 Amol/L (SK-LMS-1) roscovitine for 48 h, the combination of
doxorubicin and roscovitine, or left untreated. At 6 d, the cells were observed by light microscopy. Arrows, enlarged cells with prominent micofibrils and multinucleated
cells seen in combination treatment.
Cancer Research
Cancer Res 2008; 68: (19). October 1, 2008 7970 www.aacrjournals.org
(PV0.021). However, as shown previously, the combined treatment
resulted in greater cell death than either doxorubicin or roscovitine
alone, and because apoptosis was not significantly increased after
the combination treatment, this finding supports the hypothesis of
a nonapoptotic mechanism of synergistic cell death.
In comparison, there was no significant difference in Annexin
V-APC–positive U2OS-LC3-GFP cells after treatment with either
doxorubicin or roscovitine alone or in combination compared with
untreated cells. For SK-LMS-1 cells, there was a significant increase
in the Annexin V-APC–positive cells treated with the combination
of doxorubicin and roscovitine compared with untreated cells and
cells treated with doxorubicin alone (PV0.05). This increase was
not significantly different when compared with roscovitine alone.
However, because the percentage of Annexin V-APC–positive cells
was only 12% in cells treated with the combination of doxorubicin
and roscovitine, these findings also support an alternative
mechanism of synergistic cell death due to the combination
treatment. Similar results were found using the TUNEL assay. SK-
LMS-1 cells showed a slight increase in the TUNEL-positive cells
treated with the combination of doxorubicin and roscovitine. This
Figure 4. A, fluorescence-activated cell
sorting–based apoptosis analyses. SW-982
(top left), U20S-LC3-GFP (bottom left ), and
SK-LMS-1 cells (right) were treated
with either 0.005 Amol/L (SW-982 and
U20S-LC3-GFP) or 0.01 Amol/L (SK-LMS-1)
doxorubicin for 24 h, 10 Amol/L (SW-982
and U20S-LC3-GFP) or 20 Amol/L
(SK-LMS-1) roscovitine for 48 h, the
combination of doxorubicin and roscovitine,
or were untreated. At 6 d, the cells were
harvested and stained with either Annexin
V-APC or TUNEL and subjected to
fluorescence-activated cell sorting. *,
significant difference compared with
untreated cells (P< 0.05); c, significant
difference compared with D + R (P< 0.05).
Representative of three individual experiments.
B, effect of combined treatment with
doxorubicin and roscovitine on apoptosis.
SW-982, U20S-LC3-GFP, and
SK-LMS-1 cells were treated with either
0.005 Amol/L (SW-982 and U20S-LC3-GFP)
or 0.01 Amol/L (SK-LMS-1) doxorubicin for
24 h, 10 Amol/L (SW-982 and U20S-LC3-GFP)
or 20 Amol/L (SK-LMS-1) roscovitine for 48 h,
the combination of doxorubicin and roscovitine,
or left untreated. Western blot analysis using
antibody specific for Parp (full-length and
cleaved) was performed at 0 and 72 h after
treatment. Positive cont, SW-982 cells
treated with 4 Amol/L camptothecin for 12 h.
Representative of three individual experiments.
Synergistic Cytoxicity through Autophagy
www.aacrjournals.org 7971 Cancer Res 2008; 68: (19). October 1, 2008
difference was statistically significant when compared with
doxorubicin alone (P= 0.05) but not roscovitine alone. However,
similar to the Annexin-APC analysis, the percentage of TUNEL-
positive cells was low (4%) in cells treated with the combination of
doxorubicin and roscovitine again supporting an alternative
mechanism of synergistic cell death due to the combination
treatment.
Western blot analysis was also performed to determine whether
the enhanced cell death was occurring through apoptosis. As
shown in Fig. 4B, in all three sarcoma cell lines, the expression of
the cleaved form of the apoptosis marker, Parp, decreased after
treatment with the combination of doxorubicin and roscovitine
compared with either doxorubicin or roscovitine alone. This
finding further suggests that apoptosis is not the only mechanism
of cell death due to the combined treatment of doxorubicin and
roscovitine. Because the morphologic changes we observed with
the drug combination were reminiscent of the homeostatic
condition of autophagy, we explored whether an autophagy-
mediated mechanism of cell death was associated with the
synergistic cytotoxicity.
Synergistic cytotoxicity by doxorubicin and roscovitine is
associated with autophagy. Microtubule-associated protein-1
light chain-3 (LC3), the homologue of the yeast Apg8/Aut7p gene,
localizes on the autophagosomal membrane during autophagy (21).
The LC3-GFP fused protein is used frequently to detect autophagy
through the increased presence of GFP puncta within the
cytoplasm (21). Figure 5Ashows the difference in the presence of
LC3-GFP punctuate structures in U2OS-LC3-GFP cells treated with
doxorubicin alone, roscovitine alone, or with doxorubicin followed
by roscovitine compared with untreated cells. The pictures clearly
Figure 5. Induction of autophagy by combined treatment with doxorubicin and roscovitine. A, U20S-LC3-GFP cells were treated with either 0.005 Amol/L doxorubicin
for 24 h, roscovitine for 48 h, the combination of doxorubicin and roscovitine, or left untreated. Cells were observed under fluorescence microscopy to detect the
presence of LC3-GFP puncta, which are indicative of autophagy. Representative of three separate experiments. B, SK-LMS-1 cells were treated with either 0.01 Amol/L
doxorubicin for 24 h, 20 Amol/L roscovitine for 48 h, the combination of doxorubicin and roscovitine, or left untreated. Seventy-two hours after treatment, cells were stained
with AO and observed with fluorescence microscopy to detect the presence of AO puncta. C, SW-982 and SK-LMS-1 cells were treated with either 0.005 Amol/L
(SW-982) or 0.01 Amol/L (SK-LMS-1) doxorubicin for 24 h, 10 Amol/L (SW-982) or 20 Amol/L (SK-LMS-1) for roscovitine 48 h, the combination of doxorubicin and
roscovitine, or left untreated. Seventy-two hours after treatment, cells were stained with AO and quantified by fluorescence-activated cell sorting. *, significant difference
compared with untreated cells (P< 0.05); c, significant difference compared with D + R (P< 0.05). Representative of three separate experiments.
Cancer Research
Cancer Res 2008; 68: (19). October 1, 2008 7972 www.aacrjournals.org
show that treatment with either drug alone leads to increased
autophagocytic LC3-GFP puncta, compared with untreated U2OS-
LC3-GFP cells. Furthermore, cells treated with doxorubicin alone
had more LC3-GFP puncta than cells treated with roscovitine
alone. However, U2OS-LC3-GFP cells treated with both doxorubicin
and roscovitine showed the most LC3-GFP puncta compared with
cells treated with either drug alone. These findings suggested that
autophagy may be associated with the mechanism of synergistic
cytotoxicity of doxorubicin and roscovitine.
To determine whether autophagy was also present in SW-982
and SK-LMS-1, the cells were stained with AO. AO is concentrated
in acidic vesicles such as the autpohagolysosome and has been
used as a measure of autophagy (22). SK-LMS-1 and SW-982 cells
treated with either doxorubicin or roscovitine developed signif-
icantly increased cytoplasmic AO puncta compared with untreat-
ed cells (Fig. 5B). Furthermore, SK-LMS-1 and SW-982 cells
treated with both doxorubicin and roscovitine showed more AO
puncta compared with cells treated with either doxorubicin or
roscovitine alone. Quantization of AO staining by flow cytometry
showed a significant increase in both the SW-982 and SK-LMS-1
cells treated with both doxorubicin and roscovitine compared
with untreated cells and cell treated with either doxorubicin or
roscovitine alone (PV0.003 for all conditions; Fig. 5C). Taken
together, these results suggested an autophagic mechanism of
synergistic cytotoxicity due to the combined treatment of
doxorubicin and roscovitine.
Discussion
Roscovitine synergistically increases doxorubicin cytotox-
icity in sarcoma cells but not fibroblasts. Doxorubicin is a
commonly used cytotoxic chemotherapy agent with a significant
and use-limiting side effect profile. The ability to increase
doxorubicin efficacy would positively effect many patients suffering
from a variety of cancers. In this report, we show synergistic
cytotoxicity between the Cdk inhibitor roscovitine and doxorubicin
in three different sarcoma cell lines. However, immortalized human
fibroblast cells were not responsive to such synergistic action,
suggesting that the synergism observed is tumor specific. Previous
studies have shown that Cdk inhibition mediates tumor cell–
specific cell cycle arrest and cell death (23, 24). These effects occur
through Cdk inhibition in both the early and late phases of the cell
cycle. For example, within the G
1
phase, Cdk-mediated phosphor-
ylation of Rb leads to E2F-1 transcription factor activation and up-
regulation of the genes required for the transition into S phase (25).
Cdk inhibition during G
1
can lead to G
1
arrest and in S phase can
lead to inappropriately persistent E2F-1, resulting in both S-phase
delay and apoptosis (26).
The G
2
-M transition is also susceptible to Cdk inhibition. Cdk1
activation is essential for progression from the G
2
to M cell cycle
phases, the progression of mitosis through metaphase and cell
survival during the mitotic checkpoint (27, 28). Conditional
knockdown of Cdk1 has been shown to result in extensive DNA
rereplication and apoptosis (29). Inhibition of Cdk1 has also been
shown to increase apoptosis after genotoxic stress, whereas
prolonged Cdk1 inhibition alone can cause significant tumor
cell–specific apoptosis (23, 24). Furthermore, Cdk1 inhibition
during the spindle assembly checkpoint has been shown to cause
tumor cell–specific cell cycle progression without cell division
resulting in mitotic catastrophe (30). Because most conventional
chemotherapy agents cause DNA damage and activate cell cycle
checkpoints, it is logical that disruption of these checkpoints
through a Cdk inhibitor, such as roscovitine, would increase their
cytotoxic effects.
Roscovitine-enhanced doxorubicin cell death is associated
with autophagy. We show here that apoptosis is not the only
mechanism of synergistic cell death caused by the combination of
doxorubicin and roscovitine. Instead, autophagy is significantly
increased in the cells treated with the combination of doxorubicin
and roscovitine. Autophagy is a membrane-trafficking process
which, under normal conditions, degrades cytosolic proteins and
organelles through engulfment into double-membraned vesicles
(autophagosomes). Autophagosomes fuse with lysosomes to form
autolysosomes and the contents are degraded (31). Autophagy is
induced above basal levels by a wide variety of stimuli including
nutrient deprivation and genotoxic stress and has a direct effect on
cell viability. In this study, we have shown that autophagy can also
be induced by the direct Cdk inhibitor, roscovitine, in the setting of
previous cytotoxic treatment.
Cell cycle and autophagy. One of the more intriguing findings
of this study is the induction of autophagy by the combination of
doxorubicin and roscovitine in all three sarcoma cell lines despite
different initial cell cycle effects of the individual drug treatments.
Interestingly, the combined treatment with doxorubicin and
roscovitine lead to a prolonged G
2
-M arrest in all three sarcoma
cell lines. Therefore, we postulate that this prolonged arrest, caused
by the combined activation of the DNA damage checkpoint by
doxorubicin followed by the inhibition of the Cdk1-cyclinB
complex by roscovitine, may be a trigger for the induction of
autophagy and ultimately cell death (Fig. 6).
Figure 6. Proposed mechanism of autophagy-mediated synergistic cell death
by genotoxic chemotherapy and cyclin-dependent kinase inhibition.
Synergistic Cytoxicity through Autophagy
www.aacrjournals.org 7973 Cancer Res 2008; 68: (19). October 1, 2008
The relationship between cell cycle regulation and autophagy is
not yet clearly defined. A variety of agents have shown induction
of autophagy and autophagic cell death in association with a G
2
cell cycle arrest (32, 33). Based on our findings of the combination
treatment of doxorubicin and roscovitine leading to synergistic
cell death, induction of autophagy, and prolonged G
2
cell cycle
arrest, we postulate the following model. DNA damage by
doxorubicin (and potentially other genotoxic chemotherapy agents
or treatments) activates the DNA-PK/ATM/ATR kinases, initiating
two potential Cdk1-cyclin B inactivating cascades: (a) phosphor-
ylation of Cdc25 by CHK1 and CHK2 and (b) (when present)
activation of p53 and up-regulation of direct Cdk1-cyclin B
inhibitors such as 14-3-3j, GADD45, and p21. Injured cells
undergo cell cycle arrest and autophagy to facilitate the repair
of the damaged DNA. Once repaired, passage into mitosis occurs.
However, pharmacological inhibition of the Cdk1-cyclin B complex
by roscovitine also induces cell cycle arrest and autophagy,
thereby preventing the cells from entering mitosis. Furthermore,
due to the pivotal role of Cdk1-cyclin B in the G
2
-M transition, the
G
2
-M DNA damage checkpoint may also be inhibited by
roscovitine. In this manner, autophagy induced by a G
2
cell cycle
arrest in the setting of genotoxic injury and Cdk inhibition, may be
exploitable as a mechanism of cell death. Further investigation
into the mechanism of synergy between cytotoxic agents and Cdk
inhibition as well the role of the cell cycle in the initiation and
outcome of autophagy (e.g., cell survival or cell death) will
potentially provide new approaches to tumor-specific anticancer
therapies.
Disclosure of Potential Conflicts of Interest
L. Meijer: coinventor on patent on roscovitine, licensed to Cyclacel. The other
authors disclosed no potential conflicts of interest.
Acknowledgments
Received 4/9/2008; revised 6/18/2008; accepted 8/4/2008.
Grant support: Grant number CA87458 from the NIH (K. Keyomarsi), National
Cancer Institute P50CA116199 (K. Keyomarsi), and NIH T32 CA009599 (L.A. Lambert).
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Robert Bast for the helpful discussions.
Cancer Research
Cancer Res 2008; 68: (19). October 1, 2008 7974 www.aacrjournals.org
References
1. Zahm SH, Fraumeni JF, Jr. The epidemiology of soft
tissue sarcoma. Semin Oncol 1997;24:504–14.
2. Gaynor JJ, Tan CC, Casper ES, et al. Refinement of
clinicopathologic staging for localized soft tissue
sarcoma of the extremity: a study of 423 adults. J Clin
Oncol 1992;10:1317–29.
3. Jaques DP, Coit DG, Hajdu SI, Brennan MF. Manage-
ment of primary and recurrent soft-tissue sarcoma of
the retroperitoneum. Ann Surg 1990;212:51–9.
4. Singer S, Demetri GD, Baldini EH, Fletcher CD.
Management of soft-tissue sarcomas: an overview and
update. Lancet Oncol 2000;1:75–85.
5. Verweij J, Lee SM, Ruka W, et al. Randomized phase II
study of docetaxel versus doxorubicin in first- and
second-line chemotherapy for locally advance d or
metastatic soft tissue sarcomas in adults: a study of
the european organization for research and treatment of
cancer soft tissue and bone sarcoma group. J Clin Oncol
2000;18:2081–6.
6. Rubin E, Hait W. Drugs That Target DNA Top-
oisomerases. In: Kufe D, Bast R, Hait W, et al., editors.
Cancer Medicine 7. Hamilton (Ontario): BC Decker, Inc;
2006. p. 690–3.
7. Sherr CJ. G1 phase progression: cycling on cue. Cell
1994;79:551–5.
8. Melo J, Toczyski D. A unified view of the DNA-damage
checkpoint. Curr Opin Cell Biol 2002;14:237–45.
9. Shi J, Feng Y, Goulet AC, et al. The p34cdc2-related
cyclin-dependent kinase 11 interacts with the p47
subunit of eukaryotic initiation factor 3 during apopto-
sis. J Biol Chem 2003;278:5062–71.
10. Meyerson M, Enders GH, Wu CL, et al. A family of
human cdc2-related protein kinases. EMBO J 1992;11:
2909–17.
11. Meijer L, Borgne A, Mulner O, et al. Biochemical and
cellular effects of roscovitine, a potent and selective
inhibitor of the cyclin-dependent kinases cdc2, cdk2 and
cdk5. Eur J Biochem 1997;243:527–36.
12. Meijer L, Raymond E. Roscovitine and other purines
as kinase inhibitors. From starfish oocytes to clinical
trials. Acc Chem Res 2003;36:417–25.
13. Benson C, White J, De Bono J, et al. A phase I trial of
the selective oral cyclin-dependent kinase inhibitor
seliciclib (CYC202; R-Roscovitine), administered twice
daily for 7 days every 21 days. Br J Cancer 2007;96:29–37.
14. Abal M, Bras-Goncalves R, Judde JG, et al. Enhanced
sensitivity to irinotecan by Cdk1 inhibition in the p53-
deficient HT29 human colon cancer cell line. Oncogene
2004;23:1737–44.
15. Crescenzi E, Palumbo G, Brady HJ. Roscovitine
modulates DNA repair and senescence: implications
for combination chemotherapy. Clin Cancer Res 2005;
11:8158–71.
16. Keyomarsi K, Conte D, Jr., Toyofuku W, Fox MP.
Deregulation of cyclin E in breast cancer. Oncogene
1995;11:941–50.
17. Rao S, Lowe M, Herliczek TW, Keyomarsi K.
Lovastatin mediated G1 arrest in normal and tumor
breast cells is through inhibition of CDK2 activity and
redistribution of p21 and p27, independent of p53.
Oncogene 1998;17:2393–402.
18. Chou T-C, Talalay P. Analysis of combined drug
effects: a new look at a very old problem. Trends Pharm
Sci 1983;4:450–4.
19. Motwani M, Delohery TM, Schwartz GK. Sequential
dependent enhancement of caspase activation and
apoptosis by flavopiridol on paclitaxel-treated human
gastric and breast cancer cells. Clin Cancer Res 1999;5:
1876–83.
20. Shah MA, Kortmansky J, Motwani M, et al. A phase I
clinical trial of the sequential combination of irinotecan
followed by flavopiridol. Clin Cancer Res 2005;11:3836–45.
21. Klionsky DJ, Abeliovich H, Agostinis P, et al. Guide-
lines for the use and interpretation of assays for
monitoring autophagy in higher eukaryotes. Autophag y
2008;4:151–75.
22. Paglin S, Hollister T, Delohery T, et al. A novel
response of cancer cells to radiation involves autophagy
and formation of acidic vesicles. Cancer Res 2001;61:
439–44.
23. Ongkeko W, Ferguson DJ, Harris AL, Norbury C.
Inactivation of Cdc2 increases the level of apoptosis
induced by DNA damage. J Cell Sci 1995;108:2897–904.
24. Vassilev LT, Tovar C, Chen S, et al. Selective small-
molecule inhibitor reveals critical mitotic functions of
human CDK1. Proc Natl Acad Sci U S A 2006;103:
10660–5.
25. Lundberg AS, Weinberg RA. Functional inactivation
of the retinoblastoma protein requires sequential
modification by at least two distinct cyclin-cdk com-
plexes. Mol Cell Biol 1998;18:753–61.
26. Phillips AC, Vousden KH. E2F-1 induced apoptosis.
Apoptosis 2001;6:173–82.
27. Nigg EA. Mitotic kinases as regulators of cell division
and its checkpoints. Nat Rev Mol Cell Biol 2001;2:21–32.
28. O’Connor DS, Wall NR, Porter AC, Altieri DC. A
p34(cdc2) survival checkpoint in cancer. Cancer Cell
2002;2:43–54.
29. Itzhaki JE, Gilbert CS, Porter AC. Construction by
gene targeting in human cells of a "conditional’ CDC2
mutant that rereplicates its DNA. Nat Genet 1997;15:
258–65.
30. Weaver BA, Cleveland DW. Decoding the links
between mitosis, cancer, and chemotherapy: the mitotic
checkpoint, adaptation, and cell death. Cancer Cell 2005;
8:7–12.
31. Baehrecke EH. Autophagy: dual roles in life and
death? Nat Rev Mol Cell Biol 2005;6:505–10.
32. Aoki H, Takada Y, Kondo S, Sawaya R, Aggarwal BB,
Kondo Y. Evidence that curcumin suppresses the growth
of malignant gliomas in vitro and in vivo through
induction of autophagy: role of Akt and extracellular
signal-regulated kinase signaling pathways. Mol Phar-
macol 2007;72:29–39.
33. Kuo PL, Hsu YL, Cho CY. Plumbagin induces G2-M
arrest and autophagy by inhibiting the AKT/mammalian
target of rapamycin pathway in breast cancer cells. Mol
Cancer Ther 2006;5:3209–21.
