Effects of p53 or p27 overexpression on cyclooxygenase-2 gene expression in head and neck squamous cell carcinoma cell lines.

Abstract

Although cyclooxygenase-2 (COX-2) has been suggested to play an important role in carcinogenesis, the effects of tumor suppressors on COX-2 gene expression and the combined antitumor effects of tumor suppressors and COX-2 inhibitors have rarely been investigated.
The effects of p53 or p27 gene transfer on COX-2 expression by adenoviral vector and the combined effects of p53 or p27 gene transfer and COX-2 inhibitor exposure on the proliferation of cancer cells were investigated in head and neck squamous cell carcinoma (HNSCC) cell lines.
Overexpression of p53 markedly downregulated the transcription of COX-2, but the overexpression of p27 did not affect COX-2 levels in HNSCC cell lines. The combined antitumor effects of p53 or p27 gene transfer and of a COX-2 inhibitor (NS 398) were mainly at least additive in terms of the inhibition of cell proliferation and cell cycle arrest and additive in terms of apoptotic induction.
These results suggest that the overexpression of p53 could exert antitumor effects, at least in part, through the suppression of COX-2 gene expression, whereas growth suppression by the overexpression of p27 probably occurs by mechanisms other than the downregulation of COX-2 expression. In addition, the administration of COX-2 inhibitors, as an adjunct to p53 or p27 gene therapy, could offer a new strategy of cancer treatment and prevention.

Full text

Page 1

EFFECTS OF p53 OR p27 OVEREXPRESSION ON
CYCLOOXYGENASE-2 GENE EXPRESSION IN HEAD AND NECK
SQUAMOUS CELL CARCINOMA CELL LINES
Dong Wook Lee, MD, PhD,1Seok-Woo Park,2Soo-Yeun Park, MD, PhD,3Dae-Seog Heo, MD, PhD,2,4
Kwang Hyun Kim, MD, PhD,5,6Myung-Whun Sung, MD, PhD2,5,6
1Department of Otorhinolaryngology–Head and Neck Surgery and Medical Research Institute, College of Medicine,
Chungbuk National University, Cheongju, Korea
2Cancer Research Institute, College of Medicine, Seoul National University, Seoul, Korea
3Department of Anatomy, College of Medicine, Ewha Womans University, Seoul, Korea
4Department of Internal Medicine, College of Medicine, Seoul National University, Seoul, Korea
5Department of Otorhinolaryngology–Head and Neck Surgery, College of Medicine, Seoul National University,
28 Yongon-Dong, Chongno-Gu, Seoul, 110-744, Korea. E-mail: mwsung@snu.ac.kr
6Clinical Research Institute, College of Medicine, Seoul National University, Seoul, Korea
Accepted 27 February 2004
Published online 9 July 2004 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hed.20025
Abstract:
2) has been suggested to play an important role in carcino-
genesis, the effects of tumor suppressors on COX-2 gene
expression and the combined antitumor effects of tumor
suppressors and COX-2 inhibitors have rarely been investigated.
Methods. The effects of p53 or p27 gene transfer on COX-2
expression by adenoviral vector and the combined effects of
p53 or p27 gene transfer and COX-2 inhibitor exposure on the
proliferation of cancer cells were investigated in head and neck
squamous cell carcinoma (HNSCC) cell lines.
Results. Overexpression of p53 markedly downregulated
the transcription of COX-2, but the overexpression of p27 did not
affect COX-2 levels in HNSCC cell lines. The combined antitumor
effects of p53 or p27 gene transfer and of a COX-2 inhibitor
(NS 398) were mainly at least additive in terms of the inhibition
Background. Although cyclooxygenase-2 (COX-
of cell proliferation and cell cycle arrest and additive in terms of
apoptotic induction.
Conclusions.These results suggest that the overexpression
of p53 could exert antitumor effects, at least in part, through the
suppression of COX-2 gene expression, whereas growth
suppression by the overexpression of p27 probably occurs by
mechanisms other than the downregulation of COX-2 expres-
sion. In addition, the administration of COX-2 inhibitors, as an
adjunct to p53 or p27 gene therapy, could offer a new strategy of
cancer treatment and prevention. A 2004 Wiley Periodicals, Inc.
Head Neck 26: 706–715, 2004
Keywords: cyclooxygenase-2; p53; p27; COX-2 inhibitor; head
and neck squamous cell carcinoma
Cyclooxygenase (COX) enzymes catalyze the
rate-limiting step in arachidonic acid metabolism,
resulting in the production of prostaglandin,
prostacyclin, and thromboxanes. Two COX iso-
forms have been identified: COX-1 and COX-2.1
COX-1 is expressed constitutively in most cells,
where it is involved in the maintenance of tissue
Correspondence to: M.-W. Sung
Contract grant sponsor: This work was partly supported by Korea
Research Foundation Grant (KRF-2002-041-E00134), by 2002 BK21
project for Medicine, Dentistry and Pharmacy, Korea, and by Seoul
National University Hospital Research Grant (04-2002-035-0).
B 2004 Wiley Periodicals, Inc.
HEAD & NECK August 2004
706
Effects of p53 or p27 on Cyclooxygenase-2 Gene Expression

Page 2

homeostasis. In contrast, COX-2 is normally
absent in intact cells but is induced in response
to inflammation.1
Recent evidence indicates that COX-2 may
represent a novel target for the prevention and
treatment of various cancers, including head and
neck squamous cell carcinoma (HNSCC). The up-
regulation of COX-2 has been detected in several
different types of human cancer,2–5and epide-
miologic, clinical, animal, and in vitro studies
have indicated that the inhibition of COX-2 could
be useful for the prophylaxis and treatment
of cancers.5–8
COX-2 is up-regulated in response to growth
factors, tumor promoters, and cytokines,9and sev-
eral oncogenes, including v-src, v-Ha-ras, HER-2/
neu, and Wnt genes, have been reported to
enhance the expression of COX-2.10,11In contrast
to oncogenes, the effects of tumor suppressor
genes (eg, p53 or p27) on COX-2 have rarely been
studied, especially in intact cell systems. P53 is
known to be important in the suppression of
cellular growth and transformation.12The induc-
tion of wild-type p53 caused cells to undergo
apoptosis,13and the inactivation of p53 led to the
deregulation of the cell cycle and DNA replica-
tion, selective growth advantage, and tumor
formation.14Significantly, p53 is known to either
increase or suppress the expressions of a number
of target genes.15
p27 is one of the cyclin-dependent kinase
inhibitors (CKI) and exerts its inhibitory effects
on multiple steps of the cell cycle. The main
inhibitory action of p27 arises from its binding
with cyclin E–cyclin-dependent kinase (CDK)2.16
In addition to its role as a CKI, p27 is a putative
tumor suppressor gene. The expression of p27
protein decreases during tumor development,
even though p27 gene mutations are rarely found
in cancer. In addition, decreased p27 expression
is associated with poor prognosis in several tu-
mors, including those of breast and lung cancer,
and the overexpression of p27 may have an an-
titumor effect in HNSCC cell lines.17–19
Some reports have been published on the
effects of p53 on COX-2 gene expression.20,21
But those findings were from genetically engi-
neered mouse embryo fibroblasts or resected tu-
mor specimens. Few reports have been published
on the effects of tumor suppressor genes, such as
p53 or p27, on COX-2 gene expression in intact
cancer cells, especially in HNSCC cells. In this
work, we constructed a recombinant adenovirus-
p53 (ad-p53) and an adenovirus-p27 (ad-p27) and
investigated their effects on COX-2 gene expres-
sion in HNSCC cell lines. We then tested their
antitumor effects on the inhibition of cell prolif-
eration, the arrest of the cell cycle, and the
induction of apoptosis in HNSCC cells. In addi-
tion, we investigated the combined effects of p53
or p27 gene transfer and of a specific COX-2 inhib-
itor on the malignant behaviors of HNSCC cells.
MATERIALS AND METHODS
Cell Lines and Reagents.
cell lines (SNU-1041, SNU-1066, and SNU-1076)
were purchased from the Korean Cell Line Bank
(Seoul National University, Seoul, Korea). All cell
lines were cultured in Roswell Park Memorial
Institute (RPMI)-1640 medium (Gibco-BRL,
Grand Island, NY) supplemented with 10 mg/mL
gentamicin and 10% fetal bovine serum (FBS)
and maintained at 37jC under 5% CO2 in humid-
ified air. All cell culture consumables were ob-
tained from Gibco BRL, Grand Island, NY, unless
otherwise stated. A selective COX-2 inhibitor, N-
(2-cyclohexyloxy-4-nitrophenyl)methane sulfona-
mide (NS-398) (Cayman Chemical Co., Ann Arbor,
MI), was prepared as a 10 mM stock solution.
The required final concentrations of NS-398
were prepared by diluting aliquots of this stock so-
lution in medium, together with an appropriate
volume of dimethyl sulfoxide (DMSO).
Three human HNSCC
Construction of Recombinant ad-p53 and ad-
p27.
Wild-type ad-p53 and ad-p27 were con-
structedinourlaboratoryasreportedpreviously.19
Briefly, the cDNAs of human p53 and p27 (kindly
provided by J. Massague, Memorial Sloan-Ketter-
ing Cancer Center, New York, NY) were subcloned
into the KpnI and BamHI sites of the polylinker of
adenoviral shuttle vector, pAC CMV pLpA (kindly
provided by Robert Gerard, The University of
Texas Southwestern Medical Center, Dallas, TX),
respectively. The resulting pAC CMV-27 and pAC
CMV-53 were cotransfected, respectively, with
pJM17 (also kindly provided by R. Gerard) into
HEK293 cells (human renal embryonal cells im-
mortalizedbystabletransfectionwithE1ofadeno-
virus; purchased from the Korean Cell Line Bank,
Seoul, Korea), using the standard calcium phos-
phate precipitation method. HEK293 cells were
maintained in Dulbecco’s modified Eagle’ medium
(DMEM) (Gibco-BRL, Grand Island, NY) contain-
ing 2% FBS until the onset of a cytopathic effect.
The generating adenoviruses were purified three
times by plaque assay and confirmed by DNA se-
Effects of p53 or p27 on Cyclooxygenase-2 Gene Expression HEAD & NECKAugust 2004
707

Page 3

quencingoftheviralDNAandbyWesternblotting.
A large-scale stock of adenoviruses was prepared
by the standard CsCl method. Recombinant ade-
novirus without any therapeutic gene (ad-null)
was used as the control virus inall experiments.
Transfection of Cell Lines with ad-p27 and ad-
p53.
Exponential growing HNSCC cells were
transfected with ad-p27 (100 MOI) and ad-p53
(100 MOI) for 90 minutes with gentle frequent
shaking and then incubated with complete media
for the experiment.
Western Blot Analysis.
fected with 100 MOI of ad-p53 or ad-p27 (or ad-
null for the control virus) for 90 minutes and then
grown for 48 hours in complete media. Western
blotting for p53, p27, COX-2, and COX-1 was
performed with the electrochemiluminescence
(ECL) Western blotting system (Amersham, Eng-
land). Monoclonal mouse anti-human p53 anti-
body (1:1000, Santa Cruz Biotech, Santa Cruz,
CA), polyclonal rabbit anti-human p27 antibody
(1:1000, Santa Cruz Biotech), and monoclonal
mouse anti-human COX-2 and COX-1 antibodies
(1:1000, Transduction Laboratories, Lexington,
HNSCC cells were trans-
KY) were used as the primary antibodies for p53,
p27, COX-2, and COX-1, respectively.
Cell Proliferation Assay.
HNSCC cell lines was plated in 96-well plates
(5 ? 103/well). After 12 hours of incubation, cells
were transfected with 100 MOI of ad-null, ad-p53,
or ad-p27 for 90 minutes and then incubated with
complete media. Cells of the control group were
incubated with serum-free media for 90 minutes
and then regrown in complete media. Cell pro-
liferation was determined by 3-(4,5-dimethylthi-
azol-2-yl)-2,5-diphenyltetrazonium bromide
(MTT) assay every other day.
To evaluate the combined effects of p53 or
p27 gene transfer and NS-398 exposure on the
proliferation of the targeted cell lines, at day 1
after transfection of ad-p53 or ad-p27 into the
three HNSCC cells, old medium was replaced
with fresh medium containing NS-398 at concen-
trations of 20 or 40 AM, or an equivalent volume
of DMSO. Then at days 3 and 5, an MTT assay
was performed.
Each of the three
Cell Cycle Analysis.
transfected with 100 MOI of ad-null, ad-p53, or
ad-p27 for 90 minutes and grown for 48 hours in
complete media. Cell cycle alterations were
measured by flow cytometry with use of the Cycle
HNSCC cell lines were
FIGURE 2. Quantitative image analysis of the Western blotting
bands. When p27 overexpressed, compared with ad-null, a
certain settled pattern of change in total optical densities of the
bands for cyclooxygenase (COX)-1 or COX-2 was not observed in
three head and neck squamous cell carcinoma (HNSCC) cell lines
tested. But in case of p53, compared with ad-null, total optical
densities of the bands for COX-2 were consistently decreased by
approximately 90% or more in all three HNSCC cell lines tested.
FIGURE 1. Western blotting for p27, p53, and cyclooxygenases
in head and neck squamous cell carcinoma (HNSCC) cell lines.
Western blot analyses for p27, p53, and cyclooxygenases were
carried out 48 hours after transfection with 100 MOI of ad-p27 or
ad-p53 (Ad-null, untransduced). The results demonstrated that
p27 did not affect the levels of cyclooxygenase (COX)-2 and
COX-1 despite its overexpression after transfection with ad-p27
compared with ad-null in all HNSCC cell lines tested. In the case
ofp53,itwasoverexpressedaftertransfectingad-p23intoallthree
HNSCC cell lines, and this was accompanied by the nearly com-
plete suppression of COX-2 expression compared with ad-null in
all HNSCC cell lines tested. However, the overexpression of p53
did not affect the levels of COX-1 versus ad-null.
HEAD & NECK August2004
708
Effects of p53 or p27 on Cyclooxygenase-2 Gene Expression

Page 4

Test plus kit protocol (Becton Dickinson, San
Jose, CA).
To evaluate the combined effects of p53 or p27
genetransferandNS-398exposureonthecellcycle
parameters of the targeted cell lines 1 day after
transfection with ad-p53 or ad-p27 into the three
HNSCC cells, old medium was replaced with fresh
medium containing NS-398 at concentrations of
20 or 40 AM or an equivalent volume of DMSO.
Then, at day 3 after the transfection, cell cycle
analysis was performed, as described previously.
Quantification of Apoptosis.
analysis of annexin V-fluorescein isothiocyanate
(FITC) and propidium iodide (PI)–stained cells
was performed with an Annexin V apoptosis kit
(PharMingen, San Diego, CA), as recommended
by the manufacturer. Briefly, HNSCC cells were
transfected with 100 MOI of ad-null, ad-p53, or
ad-p27 for 90 minutes and grown for 48 hours
Flow cytometric
in complete media. Cells were then detached by
a brief trypsin treatment, stained with annexin V
and PI, and sorted by FACSstar flow cytometry
(BectonDickinson).Theproportionsshowingearly
apoptosis (positive for annexin V and negative for
PI) were determined.
To evaluate the combined effects of p53 or p27
gene transfer and NS-398 exposure on the apop-
tosis of targeted cell lines 1 day after transfec-
tion with ad-p53 or ad-p27 into the three HNSCC
cells, old medium was replaced with fresh me-
dium containing NS-398 at concentrations of 20
or 40 AM or an equivalent volume of DMSO. Then,
3 days after the transfection, apoptosis was
determined by FACS as described previously.
Transfection of COX-2 Promoter and Luciferase
Assay.
SNU-1041 cells were cotransfected with
ad-p53 at various MOIs and with luciferase ex-
pression vector containing COX-2 promoter
FIGURE 3. Growth curves of head and neck squamous cell carcinoma (HNSCC) cells after transfecting with ad-p27 or ad-p53. After
transfecting cells with 100 MOI of ad-p27 or ad-p53 (Ad-null, untransduced), cell proliferation was determined by MTT assay every other
day. A statiscally significant inhibition of proliferation was observed in all cell lines transfected with ad-p27 or ad-p53 versus cells
transfected with ad-null (*, one-way analysis of variance [ANOVA], p < 0.05). OD, optical density (OD540).
Effects of p53 or p27 on Cyclooxygenase-2 Gene Expression HEAD & NECK August2004
709

Page 5

(kindly provided by Dr. Hiroyasu Inoue, National
Cardiovascular Center Research Institute, Osaka,
Japan) for 3 hours with the LipofectAMINE PLUS
Reagent kit (Gibco-BRL, Grand Island, NY).
Luciferase activity was measured 48 hours after
gene transfer with a TR717 Microplateluminom-
eter (Tropix, Bedford, MA), in accordance with the
manufacturer’s instructions. Cells transfected
with COX-2 promoter only were used as controls.
Statistical Analysis.
performed using one-way analysis of variance
(ANOVA), repeated measures ANOVA, and Tu-
key’s Studentized Range test, where appropriate,
with a level of significance at p < .05. Graphically,
data are shown schematically with error bars
and represent plots of means F standard error
of the mean (SEM) (Sigmaplot Scientific Soft-
ware; SPSS Inc., Chicago, IL).
Statistical evaluations were
RESULTS
Western Blot Analysis.
transfecting with 100 MOI of ad-p53 or ad-p27
into HNSCC cell lines, the expression of p53, p27,
COX-2, and COX-1 were measured by Western
blotting. Figure 1 shows that the expression of
p53 increased after ad-p53 transfection and that
this was accompanied by the near complete sup-
pression of COX-2 expression in all HNSCC cell
lines tested, whereas p27 did not affect the levels
of COX-2 despite its overexpression after ad-p27
transfection. As a control, Western blotting of the
cell lysate protein for COX-1 was also carried out,
and overexpressions of p53 or p27 were found not
to affect the levels of COX-1. To confirm the
findings, the procedure was repeated for the same
HNSCC cell lines, and to express the results
numerically, quantitative image analysis was
made for the bands obtained from two different
Western blottings. When p27 overexpressed,
Forty-eight hours after
compared with ad-null, a certain settled pattern
of change in total optical densities of the bands for
COX-2 was not observed in three HNSCC cell
lines tested. But in case of p53, compared with ad-
null, total optical densities of the bands for COX-2
were consistently decreased by approximately
90% or more in all three HNSCC cell lines tested
(Figure 2).
Effects of p53 and p27 on HNSCC Cell Growth.
Transfection with ad-p27 or ad-p53 induced
marked growth suppression compared with cells
untransduced or transduced with ad-null in all
three HNSCC cells (Figure 3).
Cell Cycle Analysis in HNSCC Cells.
gene transfer into HNSCC cells induced signifi-
cant increases in the percentages of cells in the
G0/G1 phases and concomitant reductions in the
number of cells at the S phase versus transfected
with ad-null cells in all three cell lines, suggesting
G0/G1 arrest. Representative percentages of cells
in each phase of the cell cycle 48 hours after
transfection of the respective genes are presented
in Table 1.
P53 or p27
Induction of Apoptosis in HNSCC Cells.
are expressed as the percentage of cells positive
for annexin V and negative for PI, which
represent early apoptotic cell fractions 48 hours
after transfection (Table 2). P53 gene transfer
into HNSCC cells induced significant increases in
the percentages of early apoptotic cells compared
Results
Table 2. Early apoptotic cell fractions after transfection with
ad-p27 or ad-p53 in HNSCC cell lines.
Cell lineEarly apoptotic cell fraction (%)
SNU-1041Control
Ad-null
Ad-p27
Ad-p53
Control
Ad-null
Ad-p27
Ad-p53
Control
Ad-null
Ad-p27
Ad-p53
11.08 F 5.48
10.31 F 4.70
10.30 F 4.06
26.49 F 8.08y
10.08 F 3.34
8.85 F 4.76
18.81 F 4.44y
33.01 F 6.23y
11.68 F 3.12
11.15 F 2.59
17.33 F 4.73y
33.85 F 5.32y
SNU-1066
SNU-1076
Abbreviation: HNSCC, head and neck squamous cell carcinoma.
Note: Data are expressed as mean (F SEM); percentages represent cells
positive for annexin V and negative for PI at 48 hours (n = 6).
yp < .05 (one-way analysis of variance [ANOVA]) compared with cells
transfected with ad-null.
Table 1. Cell cycle analysis after transfection with ad-p27 or
ad-p53 into the SNU 1066 cell line.
Group G0/G1 (%) S (%) G2/M (%)
Control
Ad-null
Ad-p27
Ad-p53
45.77 F 7.48
44.74 F 3.11
54.68 F 8.31y
62.49 F 6.01y
37.57 F 8.69
41.02 F 8.51
29.06 F 4.37y
26.77 F 5.15y
16.66 F 7.30
14.26 F 5.35
16.26 F 5.71
10.74 F 3.40
Note: Data are expressed as the mean (F SEM); the percentage of cells
in given stages of the cell cycle phase is quoted at 48 hours (n = 6).
yp < .05 (one-way analysis of variance [ANOVA]) compared with cells
transfected with ad-null.
HEAD & NECKAugust2004
710
Effects of p53 or p27 on Cyclooxygenase-2 Gene Expression