Carcinogenesis vol.19 no.5 pp.
723–729, 1998
ACCELERATED PAPER
COX-2 expression is induced by UVB exposure in human skin:
Implications for the development of skin cancer
ShaAvhrée Y.Buckman1, Alane Gresham2, Pamela Hale1,
George Hruza1, Jason Anast1, Jaime Masferrer3 and
Alice P.Pentland4,5
1Washington University School of Medicine, St Louis, MO, 2Monsanto
Corporate Research, Monsanto Company, Chesterfield, MO, 3Searle
Inflammatory Disease Research, Monsanto Company, St Louis, MO and
4Rochester University School of Medicine, Rochester, NY, USA
5Towhom correspondence should be addressed at: Department of
Dermatology, University of Rochester, 601 Elmwood, Box 697, Rochester,
NY 14642, USA
Email: Alice_Pentland@urmc.rochester.edu
Extensive documentation has validated the role of UV
irradiation as a tumor initiator and promoter, inducing both
squamous and basal cell carcinomas. Human epidermis is
a tissue which undergoes active metabolism of arachidonic
acid to prostaglandins which is regulated by the action of
prostaglandin H synthase (also known as cyclooxygenase).
One mechanism for the promotional activity of UV light
may involve its ability to induce prostaglandin formation.
Work in our laboratory has demonstrated that acute
exposure of human keratinocytes to UVB irradiation results
in increased production of prostaglandin E2 (PGE2). When
cultured human keratinocytes were examined after irradi-
ation with 30 mJ/cm2 UVB in vitro, Western blot analysis
showed a 6-fold increase in COX-2 protein which was
evident at 6 h and peaked 24 h after irradiation. Further-
more, when human subjects were irradiated on sun-pro-
tected skin with up to four times their minimal erythema
dosage (MED) and biopsied 24 h later, upregulation of
COX-2 protein expression was observed via immunofluo-
rescence microscopy. RNAase protection assays supported
this observation, showing induction of COX-2 message
which peaked at ~12 h following irradiation in vitro.
Furthermore, human squamous cell carcinoma biopsies
exhibited strongly enhanced staining for COX-2 protein
via immunohistochemistry and Western analysis when
compared to normal non-sun-exposed control skin.
Together, these data demonstrate acute upregulation of
COX-2 via UVB irradiation and suggest the need for further
studies of COX-2 expression as a potential pharmacological
target mediating human skin tumor development.
Introduction
Chronic sun exposure is a well-recognized etiologic agent for
skin cancer. Approximately 90% of the estimated 900 000–
1 200 000 new cases of squamous and basal cell carcinomas
diagnosed in the US each year are attributable to UV light
exposure. The majority of these tumors are basal cell carcin-
*Abbreviations: PGE2, prostaglandin E2; MED, minimal erythema dosage;
UVB, ultraviolet B irradiation; COX, cyclooxygenase; COX-1, cyclo-
oxygenase-1; COX-2, cyclooxygenase-2; NSAIDS, non-steroidal anti-
inflammatory drugs; OCT, optimal cutting temperature; SCC, squamous
cell carcinoma.
© Oxford University Press
omas which are relatively benign, however squamous cell
carcinoma (the second most common cutaneous malignancy)
accounts for ~2000 deaths annually. Extensive documentation
has validated the role of ultraviolet B irradiation (UVB*)
(290–320 nm) as both a tumor initiator and promoter (1–4).
One possible mechanism for the promotional activity of UV
light involves its ability to induce prostaglandin formation.
Increased production of free arachidonic acid and prosta-
glandins are characteristic responses of human keratinocytes
following acute exposure to UVB irradiation (5–8). These
products may then function as tumor promoters in UV-initiated
tissue or they may enhance initiation because of their ability
to function as oxidants (9,10).
Prostaglandins are formed by the dual action of a phospholip-
ase which releases arachidonic acid from cellular membrane
phospholipids and cyclooxygenase (COX) which converts
arachidonic acid to prostaglandins (11). Presently, there are
two cyclooxygenase enzymes which have been identified in
humans (60% homology): cyclooxygenase-1 (COX-1) and
cyclooxygenase-2 (COX-2). Previous research suggests that
COX-1 is a housekeeping gene which modulates physiologic
responses such as regulation of renal and vascular homeostasis,
and gastroprotection of the stomach. By comparison, COX-2
is primarily an immediate early gene whose synthesis can be
stimulated rapidly and transiently in response to tumor pro-
moters (12), IL-1β (13), endotoxins (14), and serum (15).
Recent observations in knockout mice, however, suggest that
long-standing dogma regarding the functions of COX-1 and
COX-2 may be oversimplified (16).
Human epidermis actively synthesizes prostaglandins and
previous studies have demonstrated that prostaglandin E2
(PGE2) generation can regulate epidermal cell proliferation
in vitro (17). Prostaglandins generated by the arachidonic
acid cascade have been implicated in various models for
tumorigenesis. Elevated levels of PGE2 have been observed
in squamous and basal cell carcinomas of the skin and may
correlate with an increased propensity for metastatic and
invasive behavior (9,18).
The importance of prostaglandins in tumorigenesis is evid-
enced by data demonstrating the ability of various inhibitors
of prostaglandin production to inhibit growth and metastasis
of tumors in vivo (19–21). Most notably, recent epidemiological
studies have shown that non-steroidal anti-inflammatory drugs
(NSAIDS) which directly inhibit cyclooxygenase reduce the
risk of colorectal cancer (22). The putatitive role of eicosanoids
as viable tumor promoters is further suggested by the recent
work of Oshima et al. (23) in APC knockout mice (a model
for familial adenomatous polyposis) bred with mice expressing
wild type or disrupted COX-2 genes. Their work has clearly
demonstrated the role of COX-2 as an early rate-limiting step
in the pathway from colon polyps to cancer (23,24). Additional
studies have illustrated that pharmacological intervention by
NSAIDS can prolong the lag phase of chemically-induced
colon cancer and decrease the polyp number in Min mice (25).
723
S.Y.Buckman et al.
Moreover, Tsujii and DuBois (26) demonstrated that cells
overexpressing COX-2 were resistant to apoptosis, therefore
providing a putative mechanism for the ability of NSAIDS to
induce apoptosis and subsequently inhibit tumorigenesis (26).
In the context of skin tumorigenesis, chemical carcinogenesis
models have implicated the role of prostaglandins in mediating
tumor promotion (27). Unfortunately however, these studies
do not employ agents with clear physiological implications
for humans. Studies on skin cancer in humans are often limited
to epidemiological data for self-evident reasons. This study
explores the effects of acute UVB irradiation in humans and
combines these data with in vitro studies of human ker-
atinocytes to demonstrate the role of cyclooxygenase in mediat-
ing prostaglandin upregulation in response to irradiation.
Furthermore, the long term effects of repetitive UV exposure
are elucidated by evaluation of cyclooxygenase expression in
squamous cell skin cancer as well as in actinic keratoses, the
precursor dysplastic skin lesions which precede the develop-
ment of squamous cell carcinoma. Collectively, in the context
of data currently available, our observations may not only
suggest a role for COX-2 expression in responsivity to acute
UV irradiation but also further implicate its role as a potential
pharmacological target for intervention in skin tumorigenesis.
Materials and methods
Cell culture and preparation of primary human keratinocytes
Primary adult human keratinocyte cultures were obtained from human epi-
dermis removed during reductive mammoplasty and panniculectomies as
described previously (28). The tissue was defatted and the epidermis was
separated from the dermis following overnight incubation in 0.25% trypsin.
Epidermal cells were then scraped in DMEM (Gibco BRL, Gaithersburg,
MD) containing 5% fetal bovine serum, penicillin (5 units/ml), streptomycin
(5 µg/ml) and 2.5 mM N-2-hydroxyethypiperazine-N9-2-ethanesulfonic acid
(HEPES; pH 7.4). Isolated cells are then plated onto type I collagen-coated
petri dishes at a density of 7.53105 cells/cm2 and used at 2–3 days post-
confluence.
Irradiation of human keratinocyte cultures
Cultured keratinocytes were exposed to a bank of eight Westinghouse FS-20
bulbs which emit light predominantly in the UVB range, 290–320 nm.
Irradiance was measured at 1 mW/cm2 at a 30 cm distance measured by a
UVX-digital radiometer with a UVX 31 filter. Human keratinocyte cultures
were irradiated with 30 mJ/cm2 UV light
Irradiation of human subjects and skin biopsy
Following institutional approval and informed consent, human volunteers who
sunburn easily were recruited for irradiation and biopsy studies. After
initial screening to eliminate those who took aspirin-like drugs, and/or
photosensitizing medications, these subjects were exposed to varying doses
of UV light which would approximate (0.53, 13, 23, and 43) their individual
minimal erythema dose. After a 24 h time period, 6-mm punch biopsies were
taken from irradiated and non-irradiated sites. Relative erythema (redness)
was surveyed by a Minolta chromameter and biopsies were frozen in optimal
cutting temperature compound (OCT) (Miles Inc., Elkhart, IN) immediately
after collection and analyzed for COX-2 expression by immunofluorescence
microscopy.
Protein determination and Western blot analysis
Confluent cultures of human keratinocytes were scraped and sonicated in
buffer containing: 50 mM Tris–HCl, 10% glycerol, and 2% SDS in the
presence of protease inhibitors leupeptin (5 µg/ml), pepstatin A (1 µg/ml),
soybean trypsin inhibitor (50 µg/ml). Lysates were normalized using the Bio-
Rad Dc Protein Assay Kit (Bio-Rad, Richmond, CA) and diluted in buffer
containing: 50 mM Tris–HCl, 10% glycerol, 2% SDS, 100 mM DTT, and
0.1% bromophenol blue. Lysates were electrophoresed on 7.5% SDS–PAGE
gels and transferred to nitrocellulose membranes overnight at 30 mA constant
current at 4°C. Membranes were blocked for 1 h at room temperature in 5%
nonfat dry milk in Tris–buffered saline (TBS–20 mM Tris-HCl, 137 mM NaCl
pH 7.6) with 0.1% Tween 20. After blocking, the membrane was incubated
overnight with an anti-pGHS-1 monoclonal antibody (1:1000) (Cayman
Chemical, Ann Arbor, MI) or a pGHS-2 polyclonal antibody (Oxford Biomed-
ical Research, Oxford, UK) (1:1000) diluted in TBS-Tween containing 1%
724
non-fat dry milk. Membranes were subsequently washed 3310 min in 0.1%
TBS-Tween, incubated in Protein A-HRP (Amersham, Arlington Heights, IL)
(1:2000) for 1 h, washed, and visualized by enhanced chemiluminescence
(Amersham).
Assessment of cyclooxygenase activity and PGE2
The levels of PGE2 in the supernatants from cultured cells were determined
by ELISA using a kit obtained from Cayman Chemical (Ann Arbor, MI).
Cyclooxygenase activity was determined by incubating intact keratinocytes
for 15 min at 37°C with 30 µM arachidonic acid substrate prior to removal
of supernatant for PGE2 ELISA assay. Preliminary studies revealed that this
concentration of arachidonic acid was sufficient to saturate the COX enzyme
in irradiated primary keratinocyte cultures. Inhibitor studies were performed
by incubating keratinocytes with SC-58125 {1-[(4-methylsulfonyl)phenyl]-3-
trifluoromethyl-5-[(4-fluoro)phenyl]pyrazole} (IC50 COX-1 .100 µM; IC50
COX-2 5 0.09 µM) provided by Dr Jaime Masferrer (G.D.Searle Corporation,
St Louis, MO). Indomethacin was obtained from Sigma (St Louis, MO).
Immunohistochemistry
Biopsy samples were embedded in OCT (Miles Inc., Elkhart, IN) and serially
sectioned at a thickness of 6 µM onto positively charged slides. After warming
to room temperature, the slides were immersed in acetone for 10 min at
–20°C and allowed to air dry followed by 235 min washes in PBS. Nonspecific
binding sites were blocked for 20 min at room temperature in buffer containing
1% BSA, 0.2% non-fat milk, 0.3% Triton X-100, and 20% donkey serum
diluted in PBS. Sections were incubated overnight in a humidified chamber
at 4°C in primary antibody against COX-2 at a dilution of 1:1000. The slides
were then washed with PBS for 235 min and incubated in CY-3 conjugated
donkey anti-rabbit secondary antibody (Jackson Immunochemicals, West
Grove, PA) for 30 min at 37°C in a humidified chamber. Finally, slides were
washed 335 min in PBS followed by a final rinse in distilled water prior to
mounting in FluorSave (Calbiochem, San Diego, CA) and visualization by
fluorescence microscopy.
Frozen tumor biopsies were sectioned, fixed and washed as above, however
due to autofluorescence of tumor sections, a light microscopic alkaline
phosphatase method was utilized. Specifically, endogenous avidin/biotin sites
were blocked using an avidin/biotin blocking kit as described by the manufac-
turer (Zymed Immunochemicals, San Francisco, CA). Sections were permeabil-
ized by incubation for 30 min in PBS containing 1% BSA, 0.2% non-fat milk,
0.3% Triton X-100, and 0.2% saponin. Sections were blocked in PBS
containing 10% normal goat serum (Sigma, St Louis, MO) for 30 min, then
incubated in anti-COX-2 at a dilution of 1:1000 overnight at 4°C. The
following day, slides were washed 335 min in PBS and a biotinylated goat
anti-rabbit secondary antibody (Zymed Immunochemicals, San Francisco, CA)
was applied for 1 h at room temperature. Slides were washed 335 min in
PBS followed by incubation with a streptavidin-alkaline phosphatase conjugate
(Shandon-Lipshaw, Pittsburgh, PA) for 45 min. Immunoreactive complexes
were detected by a Vector Red substrate (Vector Laboratories, Burlingame,
CA) and counterstained using a 50:50 mixture of methyl green and alcian
blue in 1% sodium acetate buffer, mounted in GVA (Zymed Immunochemicals),
and visualized by light microscopy.
RNAase protection assay
Confluent keratinocyte monolayers were washed three times with PBS, lysed
in 4 M guanidine containing 0.72% beta-mercaptoethanol, and extracted in
phenol/chloroform. Nucleic acids were precipitated with isopropanol and the
concentration was determined by absorbance at 260 nm. Integrity of the RNA
was checked by agarose gel electrophoresis and ethidium bromide staining.
RNAase protection assays were performed using a kit (Ambion, Austin, TX).
Briefly, 5 µg of total RNA from each condition was allowed to hybridize with
an antisense RNA probe for COX-1 or COX-2 overnight at 45°C. Subsequently,
each condition was treated with RNAase A, ethanol precipitated, and loaded
on to an 8 M urea polyarcrylamide gel which was subsequently fixed and
dried with overnight exposure to X-ray film at –80°C.
Results
UVB irradiation enhances cyclooxygenase activity
Our laboratory has previously characterized the early time
course for the effect of UVB irradiation on prostaglandin
synthesis in human keratinocytes. These data illustrated that
synthesis and activation of a high-molecular weight calcium-
dependent cytosolic phospholipase A2 (cPLA2) mediated PGE2
production during the first 6 h following irradiation (29). As
shown in Figure 1A, PGE2 production begins to upregulate at
early time points following irradiation and synthesis plateaus

Fig. 1. (A) Timecourse of PGE2 production in cultured human keratinocytes
following UVB irradiation. Keratinocytes were either sham-irradiated at
time 0 or irradiated with 30 mJ/cm2 UVB. Cumulative PGE2 from the
supernatants was quantitated at the indicated timepoints by ELISA. T 5 0
represents synthesis over 24 h in unirradiated cultures. Data represent mean
6 SE for three experiments. (B) Cyclooxygenase activity was determined
by incubating previously irradiated keratinocytes with 20 µM arachidonic
acid for 10 min at 37°C followed by ELISA for PGE2. Data represent mean
6 SE (n 5 3).
between 10 and 24 h post-injury. Furthermore, cyclooxygenase
activity increases by ~3-fold by 24-h post injury and therefore
contributes to enhanced prostaglandin synthesis (Figure 1B).
These data suggest that the increase in cyclooxygenase activity
may temporally follow the increase in cPLA2 activity previ-
ously described to occur at early time points after UV irradi-
ation. This increase in cyclooxygenase activity coupled with
an enhancement in prostaglandin production has been observed
consistently in a variety of keratinocyte preparations derived
from multiple individuals with varying skin types.
Expression of COX-2 protein in UVB irradiated human
keratinocytes
Human keratinocyte lysates were collected from cells harvested
at control, 6- and 24-h following UVB irradiation in the
presence and absence of serum. Similar experiments were
performed in the absence or presence of 20 µM dexamethasone.
COX-2 expression is induced in human skin cancer
Fig. 2. Western blot analysis for protein following UVB-irradiation of
cultured keratinocytes using (A) anti-COX-1 and (B) anti-COX-2
antibodies. Post-confluent keratinocytes were harvested at the indicated
number of hours following irradiation. Cells in panel (C) were incubated
with 20 µM dexamethasone for the indicated time periods following
irradiation then immunoblotted with anti-COX-2. Fifty µg of protein were
loaded in each lane. The positive control in (A) is purified sheep seminal
vesicle cyclooxygenase.
This dose of dexamethasone was determined to completely
inhibit IL-1 stimulated COX-2 induction in fibroblasts (data
not shown). The blots were probed with a COX-1 monoclonal
or a COX-2 polyclonal antibody. COX-1 expression was
unaffected by UVB irradiation. Densitometry of the resulting
bands indicated an approximate 6-fold increase in COX-2
protein at the 24 h time point which appeared to correlate with
the increased PGE2 production discussed above. Dexametha-
sone treatment only modestly inhibited (,50%) COX-2 induc-
tion (Figure 2). These findings may reflect the inability of
dexamethasone to effectively inhibit the cyclooxygenase-2
enzyme found in human keratinocytes. In order to further
verify these results, studies were undertaken to assess inhibition
of cyclooxygenase activity utilizing a specific COX-2 inhibitor.
Inhibition of UVB induced cyclooxygenase activity
Further evidence for the role of COX-2 upregulation mediating
prostaglandin production was provided by the use of a COX-
2 selective inhibitor (SC-58125). The pharmacological profile
for SC-58125 inhibition of PGE2 synthesis indicates that
prostaglandin synthesis is COX-2 dependent at 24 h post UVB
irradiation. Furthermore, SC-58125 (10 µM) was able to
completely inhibit cyclooxygenase activity at a dose selective
for COX-2 (IC50 5 0.09 µM). Similarly, indomethacin
(0.1 µM), a non-selective cyclooxygenase inhibitor, was able
to completely inhibit activity (Figure 3). These data further
confirm the specificity for COX-2 upregulation following UVB
irradiation of human keratinocytes in culture.
Expression of COX-2 mRNA in irradiated cultured human
keratinocytes
Increased COX-2 protein in irradiated keratinocytes suggested
that UV irradiation induced synthesis of COX-2. The expres-
sion of COX-2 mRNA was therefore evaluated by RNAase
protection. Enhanced COX-2 message was observed as early
as 6 h post-irradiation. Densitometric analysis indicates an
725
S.Y.Buckman et al.
approximate 12-fold induction in COX-2 mRNA at the 12 h
time point following irradiation. In contrast, COX-1 mRNA
was unaffected by UVB-irradiation (Figure 4).
Immunohistochemical analysis of COX-2 in UVB irradiated
human skin
In order to demonstrate evidence for similar regulation of
COX-2 expression in an in vivo model, human subjects were
irradiated with doses of UVB ranging between 0.5–4 times
the minimal erythema dose (MED) exposure. Staining for
COX-2 was minimal in unirradiated tissue but was enhanced
in a dose dependent manner in experimental biopsies. As
Fig. 3. SC-58125 (COX-2 specific inhibitor) completely blocks UVB-
irradiation induced enhancement of cyclooxygenase activity. Post-confluent
human keratinocytes were irradiated and supernatants were assayed for
PGE2 production by ELISA following incubation with the specified dosages
of either indomethacin or SC-58125.
Fig. 4. RNAase protection assay of cyclooxygenase expression in UVB-
irradiated human keratinocytes using either a COX-1 (A) or COX-2 (B)
specific probe. Total RNA was prepared from post-confluent cells and
harvested at the times indicated following irradiation. Ten µg of RNA was
loaded in each lane.
Fig. 6. Immunostaining for COX-1 and COX-2 in normal skin and squamous cell carcinoma tissues. Negative control (A) is normal epidermis stained with
secondary antibody alone. Normal skin: (B) anti-COX-1, (D) anti-COX-2. Overlying epidermis from squamous cell carcinoma (C) anti-COX-1, (E) anti-
COX-2. Immunoreactive COX-2 expression was identified within tumor nests of SCC (F) and surrounding blood vessels (H). By comparison, blood vessels
from normal skin did not express significant COX-2 staining (G). Actinic keratoses, precursor lesions to squamous cell carcinoma, also exhibit enhanced
staining for COX-2 (I) when compared to normal skin (D).
726
shown in Figure 5, no staining is present in the control biopsy.
However, at four times the MED, COX-2 localization is
specific to the epidermis and could be competed fully in the
presence of 100-fold excess of COX-2 recombinant protein
(data not shown).
Immunohistochemical analysis of COX-2 in actinic keratoses
and human squamous cell carcinomas
Epithelial tissues from squamous and basal cell carcinomas,
actinic keratoses, as well as normal tissues were evaluated for
immunoreactive COX-2 (as denoted by the red staining shown
in Figure 6). Consistent with the human explant data, immuno-
staining for COX-2 was usually minimal in normal skin, with
subtle expression within regions of differentiated epidermis.
By comparison, squamous cell carcinomas exhibited strongly
enhanced positive staining in the overlying epidermis (adjacent
to chronically sun-exposed skin) and within regions of distinct
tumor involvement. Staining was also evident in inflammatory
mononuclear cells as well as vascular endothelial cells and
hair follicles. These findings are consistent with observations
made by Leong et al. (28). Immunoreactive COX-2 was
localized to the overlying epidermis of the squamous cell
carcinomas as well as the tumor nests, macrophages, blood
vessel endothelia, hair follicles, sebaceous glands, and sweat
glands. Basal cell carcinomas biopsies (data not shown)
exhibited less intense staining of the overlying epidermis as
well as attenuated staining of the tumor islands. Normal human
skin demonstrated minimal COX-2 staining within regions of
more differentiated epidermis and distinctly positive staining
for COX-2 within the sebaceous glands. COX-1 expression
was uniformly present within both normal human skin and
squamous cell carcinoma biopsies. Most interestingly, actinic
keratoses (sun-exposure induced precursor lesions to squamous
cell carcinoma) also exhibited prominent COX-2 staining
which permeated the full thickness of the epidermis. Increased
COX-2 expression observed in squamous cell tumors was
further confirmed by Western analysis of frozen tumor speci-
mens (Figure 7).
Discussion
Two elements are necessary for PGE2 synthesis: i) phospholip-
ase A2 which controls the availability of free arachidonic acid
substrate and ii) cyclooxygenase activity which metabolizes
this substrate to available prostaglandins. We have previously
demonstrated the role of cytosolic phospholipase A2 in mediat-
ing erythema and PGE2 formation at early time points following
UVB injury (29). The work described in this study investigates
the role of COX-2 as an important enzyme mediating PGE2
synthesis in response to acute UVB irradiation in human
epidermal keratinocytes. COX-2 protein induction occurs ~24
h following irradiation in tissue culture and in human in vivo
models. This upregulation in total protein is matched with a
concurrent enhancement in total prostaglandin production,
cyclooxygenase activity, and COX-2 message. Furthermore,
the upregulation of cyclooxygenase activity in UVB stimulated
keratinocytes is efffectively inhibited by the COX-2 selective
 COX-2 expression is induced in human skin cancer

Fig. 5. UVB irradiation enhances COX-2 expression in irradiated human skin.

Volunteers were irradiated with dosages of UVB approximating 0–43 their
MED and were subsequently biopsied after 24 h for documentation of COX-2
expression by immunohistochemistry. Frozen blocks of full-thicknes skin were
cut into 6-µm sections and stained for COX-2 expression with a specific
polyclonal antibody and analyzed by immunofluorescence microscopy.

727
S.Y.Buckman et al.
Fig. 7. COX-2 expression is enhanced in SCC biopsy samples analyzed by
Western immunoblot. Extracted protein from frozen biopsy sections of
normal skin and SCC were normalized to 50 µg protein per lane and probed
with an anti-COX-2 polyclonal antibody.
inhibitor, SC-58125. The data described in this paper clearly
demonstrate cyclooxygenase induction in human epidermal
UVB injury. Furthermore, this report demonstrates that squam-
ous cell skin cancer, much like colon cancer development,
may be linked with chronic activation of the prostaglandin
biosynthetic pathway resulting from recurrent UVB exposure.
In the series of tumor biopsies evaluated, COX-2 was
prominently expressed in squamous cell carcinomas within the
overlying sun-exposed epidermis as well as within the tumor
nests. Positive staining was also observed within the endothe-
lium and smooth muscle layers of the blood vessels and
infiltrative macrophages of SCC biopsies. Clearly the main
source of prostaglandin generation in these tumors appeared to
be resulting from the keratinocytes themselves. By comparison,
basal cell carcinomas demonstrated very minimal immunoreac-
tive staining for COX-2 within the tumor nests. These data
were also confirmed by Western analysis of the tumor biopsies.
As has been reported in the case of colon carcinomas stained
for COX-2 expression, COX-2 induction is not present in all
squamous cell biopsies (present in ~70% of biopsies studied),
therefore it may not be the only mediator of tumorigenesis. In
fact, COX-2 expression may more directly be correlated with
tumor aggressivity (30). The results of the data presented in
this report correlate well with observations described by Leong
et al. (28). However, by comparison, the data reported here
attempt to link upregulated cyclooxygenase expression with
chronic UVB irradiation as an important factor involved in
the development of skin carcinoma.
Actinic keratoses are considered to be precursor lesions to
squamous cell carcinomas which result from chronic exposure
to sunlight. When these biopsies were stained for immunoreac-
tive COX-2, they too exhibited increased staining for COX-2
when compared to normal epidermis (although with markedly
less intensity than the SCC biopsies). By inclusion of immu-
nohistochemial staining of biopsies from normal skin, actinic
keratosis, and squamous cell carcinoma (SCC), it is evident
that increased COX-2 expression is an early event that is
present even in precursor dysplastic lesions which precede the
development of SCC. Furthermore, these findings provide
evidence that chronic UVB irradiation may exacerbate COX-
2 expression in vivo. These data may further suggest that
COX-2 activation is increased prior to the development of
overt malignancy and may therefore play a role in tumor
promotion.
Recent clinical studies have provided clear evidence for the
role of COX-2 in the mediation of colorectal neoplasia.
Epidemiological studies have suggested the capacity of
NSAIDS to decrease the incidence of colorectal carcinoma
(31,32). Similarly, elevated prostaglandin levels have been
correlated with higher metastatic potential in human breast
cancer (33–35). Studies have indicated that skin tumors also
have enhanced production of arachidonic acid metabolites (9)
suggesting that prostaglandins may also play a role in skin
728
carcinogenesis. Evidence indicates the plausibility of PGE2 as
well as PGF2α as possible cocarcinogens in chemically-treated
mice which develop squamous cell carcinoma. Co-administra-
tion of these prostaglandins appears to shorten the latency
period to skin tumor formation (36). Similarly, studies in
chemically-initiated mouse models suggest the ability of prosta-
glandin inhibitors to slow the course of skin tumor progression
(27,37). Indomethacin treatment in mouse photocarcinogenesis
models significantly decreases tumor number (38). With this
evidence in mind, this report directly demonstrates the preval-
ence of COX-2 in UV-induced human squamous cell carcinoma
and its precursor lesion, actinic keratosis.
Current data suggest a model by which UV light injury can
function as an initiative stress followed by upregulation of
COX-2 which may function in a variety of ways: (i) it may
serve to enhance prostaglandin E2 production which may
function as a mitogen in an initiated cell population; (ii) it
may inhibit apoptosis (26) and thus promote retention of UV-
induced ‘sunburn cells’ (normally discarded by the epidermis
as apoptotic cells); (iii) it may alter the ability of the cells to
attach to substrate and thus enhance their propensity to exhibit
tumorigenic growth (26) and (iv) it may enhance the formation
of DNA adducts or decrease their repair (10,39).
The results reported here illustrate the need for evaluation
of the potential efficacy of COX-2 selective inhibitors as
chemotherapeutic agents to inhibit the course of UV-induced
skin cancer. Future studies should combine the use of currently
available COX-2 knockout mouse models or overexpression
studies to more definitively demonstrate the role of COX-2 in
the mediation of photocarcinogenesis in animal models.
Acknowledgements
We thank Dorothy Edwards and Monika Siedemann for their assistance in the
preparation of frozen human skin sections. We also thank Dr Aubrey Morrison
for his assistance in the preparation of this manuscript. This work was
supported by a grant from the Monsanto Corporation and from the NIH DK-
38111 and NIH 5F31CA63925–05.
References
1. Devary,Y., Gottlieb,R.A., Smeal,T. and Karin,M. (1992) The mammalian
ultraviolet response is triggered by activation of src tyrosine kinases. Cell,
71, 1081–1091.
2. Ley,R., Applegate,L.A., Padilla,R.S. and Stuart,T.D. (1989) Ultraviolet
radiation-induced malignant melanoma in Monodelophis domestica.
Photochem. Photobiol., 50, 1–5.
3. Hall,E.J., Astor,M., Bedford,J., Borek,C., Curtis,S.B., Fry,M., Geard,C.,
Hei,T., Mitchell,J. and Oleinick,N. (1988) Basic radiobiology. Am. J. Clin.
Onc., 11, 220–252.
4. Marks,R. (1995) An overview of skin cancers. Cancer Suppl., 75, 607–612.
5. Hawk,J.L.M. and Parrish,J.A. (1993) Responses of Normal Skin to
Ultraviolet Radiation. Plenum Medical Book Publishers, New York, pp.
219–260.
6. Hruza,L.L. and Pentland,A.P. (1993) Mechanisms of UV-induced
inflammation. J. Invest. Dermatol., 100, 35S–41S.
7. Kang-Rotondo,C.H., Miller,C.C., Morrison,A.R. and Pentland,A.P. (1993)
Enhanced keratinocyte prostaglandin synthesis after UV injury is due to
increased phospholipase activity. Am. J. Physiol., 264, C396–C401.
8. Grewe,M., Trefzer,U., Balhorn,A., Gyufko,K., Henninger,H. and
Krutmann,J. (1993) Analysis of the mechanism of ultraviolet (UV)
B radiation-induced prostaglandin E2 synthesis by human epidermoid
carcinoma cells. J. Invest. Derm., 101, 528–531.
9. Vanderveen,E.E., Grekin,R.C., Swanson,N.A. and Kragballe,K. (1986)
Arachidonic acid metabolites in cutaneous carcinomas. Arch. Dermatol.,
122, 407–412.
10. Cerutti,P.A. and Trump,B.F. (1991) Inflammation and oxidative stress in
carcinogenesis. [Review]. Cancer Cells, 3, 1–7.
11. DeWitt,D. and Smith,W.L. (1995) Yes, but do they still get headaches?
Cell, 83, 345–348.
 COX-2 expression is induced in human skin cancer

12. Kujubu,D.A., Fletcher,B.S., Varnum,B.C., Lim,R.W. and Herschman,H.R. (1980) Prostaglandin in human breast cancer: evidence suggesting that
(1991) TIS10, a phorbol ester tumor promoter-inducible mRNA from elevated prostaglandin production is a marker of high metastatic potential
Swiss 3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase for neoplastic cells. J. Natl Cancer Inst., 64, 1061–1070.
homologue. J. Biol. Chem., 266, 12866–12872. 35. Fulton,A.M. and Heppner,G.H. (1985) Relationship of prostaglandin E and
13. Maier,J.A., Hla,T. and Maciag,T. (1990) Cyclooxygenase is an immediate- natural killer sensitivity to metastatic potential in murine mammary
early gene induced by interleukin-1 in human endothelial cells. J. Biol. adenocarcinoma. Cancer Res., 45, 4779–4784.
Chem., 265, 10805–10808. 36. Lupulescu,A. (1977) Effect of prostaglandins on skin tumorigenesis.
14. Xie,W., Robertson,D.L. and Simmons,D.L. (1992) Mitogen-inducible Nature, 272, 634–636.
prostaglandin G/H synthase: A new target for nonsteroidal anti- 37. Muller-Decker,K., Scholz,K., Marks,F. and Furstenberger,G. (1995)
inflammatory drugs. Drug Devel. Res., 25, 249–265. Differential expression of prostaglandin H synthase isozymes during
15. O’Banion,M.K., Winn,V. and Young,D.A. (1992) cDNA cloning and multistage carcinogenesis in mouse epidermis. Mol. Carcinogen., 12,
functional activity of a glucocorticoid-regulated inflammatory cyclo- 31–41.
oxygenase. Proc. Natl Acad. Sci., 89, 4888–4892. 38. Reeve,V.E., Matheson,M.J., Bosnic,M. and Boehm-Wilcox,C. (1995) The
16. Langenbach,R., Morham,S.G., Tiano,H.F., Loftin,C.D., Ghanayem,B.I., protective effect of indomethacin on photocarcinogenesis in hairless mice.
Chulada,P.C., Mahler,J.F., Lee,C.A., Goulding,E.H., Kluckman,K.D., Cancer Lett., 95, 213–9.
Kim,H.S. and Smithies,O. (1995) Prostaglandin synthase 1 gene disruption 39. Liehr,J. and Wang,M. (1995) Lipid hydroperoxide-induced endogenous
in mice reduces arachidonic acid-induced inflammation and indomethacin- DNA adducts in hamsters: Possible mechanisms of lipid hydroperoxide-
induced gastric ulceration. Cell, 83, 483–492. mediated carcinogenesis. Arch. Biochem. Biophys., 316, 38–46.
17. Pentland,A.P. and Needleman,P. (1986) Modulation of keratinocyte
proliferation in vitro by endogenous prostaglandin synthesis. J. Clin. Received on October 29, 1997; revised on January 30, 1998; accepted on
Invest., 77, 246–251. February 5, 1998
18. Klapan,I. (1986) Prognostic significance of plasma prostaglandin E
concentration in patients with head and neck cancer. J. Cancer Res. Clin.
Oncol., 118, 308–313.
19. Snyderman,C.H., Abbas,M.M., Wagner,R. and D’Amico,F. (1995) Inhibiton
of growth of murine squamous cell carcinoma by a cyclooxygenase
inhbitor increases leukotriene B4 production. Arch. Otolaryngol. Head
Neck Surg., 121, 1017–1020.
20. Hial,V., Horakova,Z., Shaff,R.E. and Beaven,M.A. (1976) Alteration of
tumor growth by aspirin and indomethacin: studies with two transplantable
tumors in mouse. Eur. J. Pharmacol., 37, 367–376.
21. Lynch,N.R., Castes,M., Astoin,M. and Salomon,J.C. (1978) Mechanism of
inhibition of tumour growth by aspirin and indomethacin. Brit. J. Cancer,
38, 503–512.
22. Rao,C.V., Tokumo,K., Rigotty,J., Zang,E., Kelloff,G. and Reddy,B.S.
(1991) Chemoprevention of colon carcinogenesis by dietary administration
of piroxicam, alpha-difluoromethylornithine, 16 alpha-fluoro-5-androsten-
17-one, and ellagic acid individually and in combination. Cancer Res.,
51, 4528–4534.
23. Oshima,M., Dinchuk,J.E., Kargman,S.L., Oshima,H., Hancock,B.,
Evans,J.F. and Taketo,M.M. (1996) Suppression of intestinal polyposis in
Apc716 knockout mice by inhibition of cyclooxygenase 2 (Cox-2). Cell,
87, 803–809.
24. Prescott,S.M. and White,R.L. (1996) Self-promotion? Intimate connections
between APC and Prostaglandin H Synthase-2. Cell, 87, 783–786.
25. Jacoby,R.F., Marshall,D.J., Newton,M.A., Novakovic,K., Tutsch,K.,
Cole,C.E., Lubet,R.A., Kelloff,G.J., Verma,A., Moser,A.R. and Dove,W.F.
(1996) Chemoprevention of spontaneous intestinal adenomas in the Apc
Min mouse model by the nonsteroidal anti-inflammatory drug piroxicam.
Cancer Res., 56, 710–714.
26. Tsujii,M. and DuBois,R.N. (1995) Alteration in cellular adhesion and
apoptosis in epithelial cells overexpressing prostaglandin endoperoxide
synthase 2. Cell, 83, 493–501.
27. Furstenberger,G., Gross,M. and Marks,F. (1989) Eicosanoids and multistage
carcinogenesis in NMRI mouse skin: role of prostaglandins E and F in
conversion (first stage of tumor promotion) and promotion (second stage
of tumor promotion). Carcinogenesis, 10, 91–96.
28. Leong,J., Hughes-Fulford,M., Rakhlin,N., Habib,A., Maclouf,J. and
Goldyne,M.E. (1996) Cyclooxygenases in human and mouse skin and
cultured human keratinocytes: association of COX-2 expression with
human keratinocyte differentiation. Exp. Cell Res., 224, 79–87.
29. Gresham,A., Masferrer,J., Chen,X., Lealkhi,S. and Pentland,A.P. (1996)
Increased synthesis of high-molecular-weight cPLA2 mediates early UV-
induced PGE2 in human skin. Am. J. Physiol. Cell Physiol., 39, C1037–
C1050.
30. Tsujii,M., Kawano,S. and DuBois,R.N. (1997) Cyclooxygenase-2
expression in human colon cancer cells increases metastatic potential.
Proc. Natl Acad. Sci. USA, 94, 3336–3340.
31. Rao,C.V., Rivenson,A., Simi,B., Zang,E., Kelloff,G., Steele,V. and
Reddy,B.S. (1995) Chemoprevention of colon carcinogenesis by sulindac,
a nonsteroidal anti-inflammatory agent. Cancer Res., 55, 1464–1472.
32. Marnett,L.J. (1992) Aspirin and the potential role of prostaglandins in
colon cancer. [Review]. Cancer Res., 52, 5575–89.
33. Liu,X. and Rose,D.P. (1996) Differential expression and regulation of
cyclooxygenase-1 and -2 in human breast cancer cell lines. Cancer Res.,
56, 5125–5127.
34. Rolland,P.H., Martin,P.M., Jacquemier,J., Rolland,A.M. and Toga,M.

729