Ultraviolet B (UVB) light is the major environmental carcinogen contributing to non-melanoma skin cancer (NMSC) development. There are over 3.5 million NMSC diagnoses in two million patients annually, with men having a 3-fold greater incidence of squamous cell carcinoma (SCC) compared with women. Chronic inflammation has been linked to tumorigenesis, with a key role for the cyclooxygenase-2 (COX-2) enzyme. Diclofenac, a COX-2 inhibitor and non-steroidal anti-inflammatory drug, currently is prescribed to patients as a short-term therapeutic agent to induce SCC precursor lesion regression. However, its efficacy as a preventative agent in patients without evidence of precursor lesions but with significant UVB-induced cutaneous damage has not been explored. We previously demonstrated in a murine model of UVB-induced skin carcinogenesis that when exposed to equivalent UVB doses, male mice had lower levels of inflammation but developed increased tumor multiplicity, burden and grade compared with female mice. Because of the discrepancy in the degree of inflammation between male and female skin, we sought to determine if topical treatment of previously damaged skin with an anti-inflammatory COX-2 inhibitor would decrease tumor burden and if it would be equally effective in the sexes. Our results demonstrated that despite observed sex differences in the inflammatory response, prolonged topical diclofenac treatment of chronically UVB-damaged skin effectively reduced tumor multiplicity in both sexes. Unexpectedly, tumor burden was significantly decreased only in male mice. Our data suggest a new therapeutic use for currently available topical diclofenac as a preventative intervention for patients predisposed to cutaneous SCC development before lesions appear.
According to recent SEER Cancer Statistics, between 2005 and 2009, the overall cancer incidence and mortality rate was higher in men than in women ( 1 ). This sex difference is also observed in non-melanoma skin cancer (NMSC), which includes both basal cell carcinoma and squamous cell carcinoma (SCC). Epidemiological studies demonstrate that men have twice the incidence of basal cell carcinoma and three times the incidence of SCC as women ( 2 ). There are over 3.5 million NMSC diagnoses in approximately two million patients each year ( 3 ) with treatment costs reaching over $1.4 billion ( 4 ), making it the fifth most costly cancer in the Medicare population ( 5 ). Although SCC only makes up 16% of all NMSC, it can metastasize more frequently, killing approximately 2500 people each year, of which men are the majority ( 6 ).
A component of solar radiation, ultraviolet B (UVB) is the major environmental carcinogen contributing to the development of both basal cell carcinoma and SCC ( 7–9 ). Though the increased SCC incidence observed in men has been attributed to lifestyle differences with higher levels of sun exposure and less protection from the sun compared with women ( 10–12 ), we have shown previously using a murine model that when exposed to equivalent UVB doses, this sex bias still holds true ( 13 ). With the increasing incidence of NMSC and the knowledge that men and women respond differently to UVB, it is imperative to examine the efficacy of potential treatments in both sexes.
It is well documented that UVB exposure is associated with an inflammatory response ( 14–16 ), which following both acute and chronic exposure is known to play a role in the formation of NMSC ( 17–19 ). Chronic UVB exposure induces constitutive expression of cyclooxygenase-2 (COX-2), which is the primary source of elevated prostaglandin E 2 (PGE 2 ) in the skin ( 20 , 21 ). This prostaglandin synthesis increase plays a key role in carcinogenesis by contributing to uncontrolled proliferation of damaged cells that ultimately form tumors ( 22–26 ). The inflammatory response that coincides with elevated PGE 2 levels is linked to increased DNA damage, which following exposure to stimuli such as UVB, induces the expression of wild-type p53, a tumor suppressor gene ( 27 ). The loss of function of p53 by mutation or allele loss is a common finding in many human malignancies, including cutaneous SCC ( 28 , 29 ).
UVB exposure induces the production of reactive oxygen species (ROS) such as hydrogen peroxide, which if not maintained at a homeostatic level by antioxidant networks, can cause oxidative stress and cellular damage. The main cutaneous antioxidant, catalase, has the role of detoxifying hydrogen peroxide. Skin antioxidant activity decreases with UVB exposure, and specifically, catalase activity level decreases have been associated with both skin carcinogenesis and progression ( 30–36 ). Previously, we reported decreased antioxidant activity ( 13 ), and specifically, catalase activity ( 37 ), in the skin of male mice compared with female mice that may contribute to increased tumorigenesis in males.
We previously demonstrated that treating female mice topically with the COX-2 inhibitor, celecoxib, decreased PGE 2 levels as well as the percentage of p53-positive epidermal cells in the acute UVB-induced cutaneous response ( 38 ). Chronic UVB exposure and subsequent topical celecoxib treatment decreased both tumor number and grade in female mice; however, its efficacy in males has not been described. Because male mice display lower levels of inflammation and antioxidant activity as compared with female mice after UVB exposure ( 13 , 37 ), we sought to investigate the extent to which a topical anti-inflammatory drug would affect tumor burden as well as PGE 2 or catalase levels in the skin of male versus female mice after chronic UVB exposure.
Because celecoxib is not available to patients in a topical formulation, we conducted studies using topical diclofenac, a non-steroidal anti-inflammatory drug (NSAID) and COX-2 inhibitor, which is currently prescribed to patients with actinic keratotic lesions—precursors of SCC. Topical formulations are better tolerated and cause fewer adverse effects compared with oral COX inhibitors as the systemic levels are greatly decreased with topical treatments ( 39 ). Open-label studies have demonstrated that diclofenac is indeed well tolerated and effective short term for treating actinic keratotic lesions ( 40 , 41 ). However, its efficacy as a preventative agent in patients without evidence of precursor lesions but with significant UVB-induced cutaneous damage has not been examined. Additionally, potential differences between men and women in the efficacy of diclofenac treatment have not been reported.
A recent case-control study suggested a dose-dependent, cumulative, protective effect for the reduction of skin cancer risk, specifically SCC, with the use of a number of NSAIDs ( 42 ). In this study, we have modeled both male and female compliant patients who eliminated sun exposure and began using topical diclofenac regularly to reverse existing damage resulting from chronic UVB exposure and prevent tumor development. Our results demonstrate that despite the observed sex differences in the inflammatory response, prolonged topical diclofenac treatment of chronically UVB-damaged skin effectively reduced tumor multiplicity in male and female Skh-1 hairless mice. Interestingly, with topical diclofenac treatment, tumor burden was significantly decreased, and tumors were of lower histologic grade only in male mice. Although PGE 2 levels were decreased in both sexes, levels of the antioxidant, catalase, were increased with diclofenac treatment in male mice only. These data support inflammation as a key factor contributing to UVB-induced tumor development in both male and female mice. Taken together, our study emphasizes differences in the response of male and female murine skin and demonstrates a potential new therapeutic use for this currently available topical treatment as a preventative intervention for patients predisposed to cutaneous SCC development.
Materials and methods
Outbred, male and female Skh-1 mice (6–8 weeks old; Charles River Laboratories, Wilmington, MA) were housed in the vivarium at The Ohio State University according to the requirements established by the American Association for Accreditation of Laboratory Animal Care. All procedures were approved by the institutional animal care and use committee before the initiation of any studies. Mice were dorsally exposed to 2240 J/m 2 UVB, previously determined to be one minimal erythema dose, 3× weekly on non-consecutive days for 10 weeks. UVB dose was calculated using UVX radiometer and UVB sensor (UVP, Upland, CA) and emitted by Phillips FS40 UV bulbs (American Ultraviolet Company, Lebanon, IN). Mice were treated topically with vehicle (Surgilube®; Savage Laboratories, Melville, NY) or 500 µg diclofenac (Solaraze®, 3% diclofenac sodium; Doak Dermatologics, Fairfield, NJ) in vehicle for 15 weeks with no additional UVB exposure. This dose of diclofenac was chosen based on previous studies in our laboratory. Surgilube is an inert vehicle used clinically as a surgical lubricant. Tumors >1mm in diameter were measured weekly with calipers. After killing, 0.5cm 2 section of dorsal skin and all tumors were fixed as described previously ( 37 ) while remaining dorsal skin was snap frozen in liquid nitrogen.
Hematoxylin and eosin-stained tissue sections of tumors isolated from male and female mice were graded in a blinded manner by a board-certified veterinary pathologist (D.F.K.) as described previously ( 13 ).
Catalase activity assay
Frozen dorsal skin was crushed, and 15mg of dorsal skin was used for analysis of catalase activity using the Catalase Assay Kit (Cayman Chemical, Ann Arbor, MI) according to manufacturer’s instructions.
PGE 2 EIA
Frozen dorsal skin was crushed, and 15mg of dorsal skin was used for analysis of PGE 2 levels using the PGE 2 EIA Kit-Monoclonal (Cayman Chemical) according to manufacturer’s instructions.
p53. Paraformaldehyde-fixed/OCT-embedded dorsal skin sections were cut (10 µm) onto Superfrost Plus® microscope slides (Fisher Scientific) and stored at −80°C for future analysis. Slides were thawed overnight at room temperature, baked at 60°C for 30min and then rehydrated in Clear Rite 3 (Richard-Allen Scientific) and a graded series of ethanol. Sections were circled with Immedge pen (Vector Laboratories), and endogenous peroxidase activity was blocked with 3% H 2 O 2 in water for 10min at room temperature. Slides were incubated in Antigen Unmasking Solution (Vector Laboratories) for 25min in a vegetable steamer reaching 95°C. Slides were cooled for 20min at room temperature and blocked in 1× casein (Vector Laboratories) for 10min followed by incubation with primary p53 antibody [clone CM5p; Novocastra (Leica Microsystems), Buffalo Grove, IL] at a 1:500 dilution in 1× casein at room temperature for 1h. Slides were incubated in Rabbit Link and Label (Biogenex, Fremont, CA), each for 30min at room temperature. Slides were washed in phosphate-buffered saline containing 0.05% Tween between the various incubation steps. Following a final phosphate-buffered saline containing 0.05% Tween wash, slides were incubated in diaminobenzidene solution (Vector Laboratories) for 10min at room temperature. Slides were washed in deionized water, counterstained with hematoxylin 2 and dehydrated in an increasing series of ethanol followed by Clear Rite 3 and then cover slipped with VectaMount mounting medium (Vector Laboratories). p53 foci were counted as three or more adjacent p53-positive cells and examined in five fields of view at ×20 magnification.
Ki67. Protocol is the same as for p53 through incubation with 1× casein. Slides were incubated with primary Ki67 antibody (Dako, Carpinteria, CA) at a 1:200 dilution in 1× casein overnight at 4°C in a humid chamber. Slides were incubated in biotinylated IgG (Vector Laboratories) at a 1:200 dilution in 1× casein for 30min at room temperature. Slides were incubated in ABC Elite (Vector Laboratories) for 30min at room temperature. Slides were washed in phosphate-buffered saline containing 0.05% Tween between the various incubation steps. Slides were incubated in diaminobenzidene solution (Vector Laboratories) for 10min at room temperature. Slides were washed in deionized water, counterstained and dehydrated as described. Ki67-positive cells were examined in five fields of view at ×60 magnification.
The results presented in this paper were part of a larger experiment involving four treatment groups (of which diclofenac was one) and a single control group. Dunnett’s adjustment ( 43 , 44 ) for multiplicity was used for comparing the primary outcome of tumor burden at 24 weeks between the treatment groups and control in order to restrict the probability of a type I error to 5%. The number of control mice was inflated compared with the treatment groups to increase the power of the comparison ( 43 ). Residual plots verified the model assumptions of normality and homoscedasticity, and a logarithmic transformation was utilized if necessary. Continuous outcome data were analyzed using an analysis of variance approach with linear contrasts for testing the comparisons of interest. A mixed-effects regression model with a random slope and intercept by subject was used to model tumor growth from the time of tumor origination. For count data, Poisson regression was used. All analyses were conducted in SAS version 9.2 (SAS Institute, Cary, NC). P -values ≤ 0.05 were considered statistically significant.
Diclofenac topical treatment decreased tumor number and burden
To examine the effects of topical diclofenac treatment as a preventative agent against tumor development, we exposed male and female Skh-1 hairless mice to 2240 J/m 2 UVB (previously determined to be one minimal erythema dose in our laboratory) for 10 weeks to model chronic sun exposure. Mice were then topically treated with either surgilube (vehicle) or diclofenac for 15 weeks without further UVB exposure to model a lifestyle change. Non-irradiated male and female mice treated with either vehicle or diclofenac did not develop tumors. Additionally, mice that received no vehicle treatment did not exhibit a significantly different tumor burden compared with mice treated with vehicle, indicating that the vehicle had no significant effect on tumorigenesis in this study (data not shown).
Following 10 weeks of only UV exposure and 15 weeks of preventative topical treatment with diclofenac, male mice exhibited decreased tumor multiplicity with 57.8% fewer tumors compared with the mice treated with vehicle ( P < 0.0001, Figure 1A ). At week 24, male mice treated topically with diclofenac exhibited significantly decreased tumor burden by 82.5% ( P < 0.0001) compared with the mice treated with vehicle ( Figure 1B ). Female mice treated topically with diclofenac also demonstrated decreased tumor multiplicity with 63.1% times fewer tumors compared with mice treated with vehicle ( P < 0.0001, Figure 1A ). Female mice treated topically with diclofenac displayed a 51.7% decrease in average tumor burden that was not statistically different ( P = 0.1098) compared with mice treated with vehicle ( Figure 1B ). However, female mice treated with vehicle exhibited approximately half the tumor burden as vehicle-treated males, which was significantly different ( P = 0.0033).
Examining the change in tumor burden over time, we found that male mice treated topically with diclofenac had a significantly decreased tumor growth rate of 6.9%, while mice treated with vehicle exhibited a growth rate of 15.9% ( P = 0.0057, Figure 2A ). Though female mice treated topically with diclofenac also exhibited a slight decrease in tumor growth rate compared with mice treated with vehicle, 8.6% versus 11.1%, respectively, this decrease was not significant ( P = 0.3949, Figure 2B ).
Diclofenac topical treatment decreased the severity of tumors in male mice
Tumors were isolated from mice after 10 weeks of UVB exposure followed by 15 weeks of topical treatment and scored by a board-certified veterinary pathologist (D.F.K.). Tumors classified as papilloma were considered benign, and tumours classified as microinvasive or fully invasive SCC were considered malignant. Male mice treated with vehicle developed tumors of each grade, whereas male mice preventatively treated with diclofenac developed tumors graded as papilloma (grades 1–3) or microinvasive SCC (grade 1 only). The estimated per tumor malignancy rate for male mice treated with vehicle was 0.20, indicating that one out of five tumors could be expected to be malignant, as compared with the malignancy rate of 0.06 for mice treated with diclofenac. There was a trend toward a lower rate, but it was not statistically significant ( P = 0.2413). Importantly, male mice treated with diclofenac developed no fully invasive SCC lesions, whereas mice treated with vehicle developed 5.3% of graded tumors that were fully invasive SCC lesions.
Female mice treated with vehicle developed all grades of papilloma, microinvasive SCC grade 1 and fully invasive SCC. Mice preventatively treated with diclofenac developed all grades of papilloma, microinvasive SCC grades 1 and 2, and fully invasive SCC. Interestingly, although female mice treated topically with diclofenac exhibited decreased tumor burden compared with vehicle-treated mice, the estimated per tumor malignancy rate was 0.25 compared with 0.11 for mice treated with vehicle, which trended toward an increased malignancy rate but was not statistically significant ( P = 0.2574).
Diclofenac topical treatment increased catalase activity levels in male mice
Protein was extracted from tumor-free, dorsal skin in order to examine catalase activity levels. Even after 15 weeks without UVB exposure, catalase activity levels were significantly decreased in the vehicle-treated male skin as compared with the non-irradiated controls ( P = 0.0004, Figure 3A ), but not in female skin ( P = 0.2121, Figure 3B ). Topical diclofenac treatment applied to male skin significantly increased catalase activity from levels observed in the vehicle-treated skin ( P = 0.0035, Figure 3A ), and in fact, catalase activity is almost restored to levels observed in non-irradiated skin. In contrast, catalase activity levels in female skin treated topically with diclofenac were not significantly altered compared with those observed in vehicle-treated skin ( P = 0.1590, Figure 3B ).
Diclofenac topical treatment decreased cutaneous PGE 2 levels
To confirm that topical diclofenac treatment preventatively applied effectively inhibited COX-2-mediated PGE 2 production, cutaneous PGE 2 levels were determined by enzyme immunoassay. In chronically irradiated male ( Figure 4A ) and female ( Figure 4B ) vehicle-treated skin, even after 15 weeks without UVB exposure, PGE 2 levels were significantly increased from levels observed in non-irradiated skin ( P < 0.0001). Skin of male ( Figure 4A ) and female ( Figure 4B ) mice exposed to UVB for 10 weeks and treated topically with diclofenac for 15 weeks exhibited significantly decreased PGE 2 levels as compared with chronically irradiated, vehicle-treated mice ( P = 0.0001).
Male mice have increased numbers of p53-positive foci compared with female mice
Tumor-free, dorsal skin sections were examined for p53-positive foci via immunohistochemistry with an antibody detecting both wild-type and mutant p53 ( Figure 5A and 5B ). The number of p53-positive foci measured over five fields was not significantly altered with preventative topical diclofenac treatment in male skin (2.5 for vehicle-treated versus 1.8 for diclofenac-treated, P = 0.2323) or female skin (0.89 for vehicle-treated versus 0.63 for diclofenac-treated, P = 0.4918). Corresponding with the increased tumor burden, the mean number of foci was significantly higher in vehicle-treated male skin compared with vehicle-treated female skin (2.5 versus 0.89, P = 0.0004, Figure 5C ).
Diclofenac does not affect proliferation 15 weeks after cessation of UVB exposure
Tumor-free, dorsal skin sections were stained for Ki67. The percentage of Ki67-positive cells was not different ( P = 0.4708) between the vehicle-treated and diclofenac-treated male skin after 10 weeks of UVB exposure followed by 15 weeks of topical treatment with no additional UVB exposure. In female skin, the percentage of Ki67-positive cells exhibited a decreasing trend from 13.4% to 9.4% with preventative topical diclofenac treatment compared with vehicle-treated mice ( P = 0.0596). Interestingly, the percentage of Ki67-positive cells was significantly higher in male skin compared with female skin with both vehicle and diclofenac topical treatment ( P < 0.006 and P = 0.0004, respectively, Figure 6 ).
Our previous studies have demonstrated that when exposed to equivalent, chronic UVB light, male mice exhibit a higher tumor burden compared with female mice ( 13 ), which we confirmed in this study. We also have shown previously that female mice preventatively treated topically with the anti-inflammatory drug, celecoxib, exhibited decreased tumor number, grade, PGE 2 levels and the number of p53-positive cells ( 20 , 38 ). In this study, we show that the readily available and currently used NSAID, diclofenac, elicits similar results in male mice treated topically with diclofenac after sustaining significant UVB damage. We demonstrate that prolonged, topical diclofenac treatment of chronically UVB-damaged skin effectively reduced tumor multiplicity in both sexes. Unexpectedly, tumor burden was only significantly decreased in male mice, where we also observed a slower rate of growth and a trend toward a lower malignancy rate, with no fully invasive SCCs developing in diclofenac-treated male mice. We also observed a sex difference in catalase activity in that male mice treated with diclofenac exhibited increased catalase activity levels compared with vehicle-treated male mice, whereas catalase activity levels in female mice remained unchanged.
The observed increase in catalase activity levels in male skin may be an important factor because we have observed endogenously lower antioxidant activity levels ( 13 ) and specifically, lower catalase activity levels in male skin ( 37 ). Additionally, both decreased antioxidant enzyme levels and the resulting oxidative protein damage in chronically UV-exposed skin have been demonstrated ( 33 ). Recently, we reported a link between decreased skin catalase activity, increased Gr-1 + CD11b + myeloid cell infiltration and increased tumorigenesis in male mice but not in female mice ( 37 ). Elevated levels of PGE 2 have been shown to induce higher levels and more suppressive Gr-1 + CD11b + myeloid cells that produce high levels of ROS ( 45 , 46 ). By decreasing PGE 2 levels in male skin with topical diclofenac treatment, Gr-1 + CD11b + myeloid cell infiltration is decreased (unpublished data), therefore, decreasing myeloid-associated ROS levels. As a result, catalase activity levels may be increased or maintained due to the overall lower levels of cutaneous ROS production. The increased skin catalase activity in male mice treated with diclofenac compared with male mice treated with vehicle may indicate that restoration of this antioxidant is important for decreasing tumorigenesis in male but not in female mice.
In addition to increasing catalase activity, diclofenac topical treatment also reduced the number of p53-positive foci observed in male mice. Though not statistically significant, there is a clear trend toward fewer foci observed. As p53 is indicative of overall DNA damage, a decrease in the number of p53-positive cells observed demonstrates that diclofenac treatment may be promoting repair or preventing damage altogether. Previous studies demonstrate that ingenol mebutate, a well-tolerated topical antineoplastic drug used for treatment of both NMSC and actinic keratotic lesions ( 47–49 ), effectively reduced UVB-induced lesions and p53 patches in Skh-1 mice ( 50 ). However, the side effects of this treatment included scar formation and skin tightening, which were not observed in this study. Because the mice in this study were exposed to UVB for 10 weeks and then treated for 15 weeks without further UVB exposures prior to examining tissue, further studies examining various time points in addition to examining levels of mutated p53 would be necessary to determine the role diclofenac is playing in the formation of p53 foci.
Importantly, preventative diclofenac treatment led to decreased tumor burden in male mice and significantly lower PGE 2 levels in both sexes compared with vehicle-treated mice, underscoring the importance of COX-2-mediated inflammation in the UVB-induced carcinogenesis process. PGE 2 also contributes to angiogenesis ( 51 ) and invasion ( 52–55 ), in addition to inflammation. Notably, angiogenesis has been reportedly increased in actinic keratoses compared with adjacent skin ( 56 ). Further, studies have indicated that an ‘angiogenic switch’ is turned on during this stage, which contributes to the invasive nature and progression of these lesions toward becoming SCCs ( 57 , 58 ). Although not examined in this study, blocking PGE 2 production with diclofenac treatment may be decreasing angiogenesis and invasive properties that contribute to the lack of development of fully invasive SCC lesions in male mice that are present in the vehicle-treated mice.
We have shown previously that males have lower levels of acute UVB-induced inflammation compared with female mice and suggested that the inflammatory response was not as important to tumorigenesis in male mice ( 13 ). This study offers support to the fact that the inflammatory response in male mice, regardless of magnitude, is indeed important for tumorigenesis because decreasing levels of inflammatory mediators and increasing the antioxidant capacity within male skin correlated with decreased tumor number and burden in male mice topically treated with diclofenac.
Recent studies have refocused on the preventative effects of NSAIDs against skin cancer ( 42 , 59 ), highlighting the relevance of our study. Although many studies focus on oral NSAIDs that may result in harmful side effects over time, the use of topical NSAIDs results in lower systemic levels and therefore, a lower risk of these side effects including gastrointestinal hemorrhage and peptic ulcer disease ( 39 , 60 ). Our data demonstrate a potential NSAID alternative for patients with cardiac or other health issues that may be heightened with oral NSAID use.
In summary, we have shown in a model of UVB-induced SCC that topically treating male and female Skh-1 hairless mice prophylactically with the NSAID and COX-2 inhibitor, diclofenac, decreased tumor multiplicity and PGE 2 production in both sexes. In addition, topically applied diclofenac also interestingly increased skin catalase levels and decreased tumor burden as well as the percentage of malignant tumors in male mice. These findings are especially relevant because in humans, men have a greater risk and incidence of cancer overall ( 1 ), with a 3-fold greater incidence of SCC ( 2 ). To the best of our knowledge, this is the first report of a sex-associated difference in the efficacy of topical diclofenac treatment, implying that it may be beneficial to retrospectively dissect out potential sex-based differences in previously reported clinical outcomes in order to more effectively treat each sex. As we have previously observed that male mice have lower antioxidant activity as compared with female mice, further studies are needed to determine if the increased catalase levels are contributing to the therapeutic effects of the diclofenac treatment, or if inhibiting inflammation is indeed the primary factor in both sexes. Overall, our data suggest that sustained diclofenac treatment applied topically to chronically UVB-damaged skin before the appearance of precursor lesions, in combination with the elimination of UVB exposure, may be beneficial to patients who are predisposed to developing SCC.
National Institutes of Health (CA133629).
Conflict of Interest Statement: None declared.
non-melanoma skin cancer
non-steroidal anti-inflammatory drug
- PGE 2
prostaglandin E 2
reactive oxygen species
squamous cell carcinoma
We would like to acknowledge Kristen Duckro for excellent technical support.