Celecoxib Reduces the Effects of Acute and Chronic UVB Exposure in Mice Treated with Therapeutically Relevant Immunosuppressive Drugs

Abstract

Solid organ transplant recipients have a greatly increased risk for the development of non-melanoma skin cancers. We have previously shown in our mouse model that sirolimus given in combination with cyclosporine A resulted in fewer and smaller tumors than cyclosporine A alone. In the current study we tested the hypothesis that an anti-inflammatory agent celecoxib applied topically after UVB exposure would further reduce UVB induced skin cancer in mice treated with cyclosporine A and sirolimus. The effect of celecoxib treatment on acute inflammation, initiation/promotion, and tumor development was examined through a set of four experiments. Delayed tumor onset was observed in both tumor development experiments. Reduced tumor size and number compared to vehicle was observed when CX was administered concurrently with UVB and when CX was administered after cessation of UVB treatments, respectively. Prostaglandin E2 was confirmed to be significantly reduced in the dorsal skin of mice concurrently treated with immunosuppressants, CX and UVB for 13 weeks, suggesting a reduction in the inflammatory response could be the mechanism by which CX reduced tumorigenesis. Furthermore, topical celecoxib treatment following acute UVB exposure reduced dermal neutrophil number and activity compared to vehicle. In all of these experiments unirradiated and vehicle treated mice were utilized as controls. In conclusion, these data suggest that even in the presence of cyclosporine A and sirolimus, topical celecoxib treatment can result in reduced inflammation, tumor number, and size; properties which may be beneficial in the therapeutic reduction of skin cancer development in solid organ transplant recipients.

INTRODUCTION

It is becoming apparent that with more effective immunosuppressive therapies improving post-transplant survival, serious complications such as post-transplant malignancies are an increasingly important issue for transplant patients. The most common solid malignancy in transplant recipients is nonmelanoma skin cancer (NMSC). NMSC is not classically thought of as a mortal disease; however, in the transplant population NMSC are more common and more aggressive than in the general population ^1–4. Often these patients present with multiple lesions in sun exposed areas known as field cancerization, which can be difficult to treat by excision, currently the standard of care. More research is needed to develop less invasive and more effective standards of care for these patients ⁵. While the majority of research to date has focused on chemotherapeutic agents, studies examining effective chemopreventative modalities need to be explored as well.

Therapeutic immunosuppressants used to prevent graft rejection directly and indirectly affect carcinogenesis in these patients. Cyclosporine A (CsA), is a previously favored transplant immunosuppressant that modulates the adaptive immune system through inhibition of calcineurin thus preventing activation of nuclear factor of activated T cells (NFAT). Though CsA has done much in the past to improve transplant survival, its usage is now being reevaluated due to its' multitude of side effects. Focusing on the carcinogenesis process, CsA has been shown to promote cancer development by inhibiting DNA damage repair and altering the tumor microenvironment ^{6, 7}. Sirolimus (SRL), also known as rapamycin, is a macrolide immunosuppressant that has been used as a therapeutic immunosuppressant and more recently as a cancer therapeutic ⁸. SRL inhibits adaptive immunity by arresting cells in the G1 phase of the cell cycle through the inhibition of mammalian target of rapamycin (mTOR) ^{9, 10}. This same action blocks angiogenesis which along with other pro-apoptotic and anti-proliferative effects make this family of drugs potential chemotherapeutic agents ¹¹.

Previous studies have shown that the combination of CsA and SRL synergistically protect against graft rejection ¹². This synergy allows for a reduction in the dose of both drugs thereby reducing the negative side effects ^{13, 14}. In fact, Kahan et al. have shown a decrease in malignancy in patients receiving dual (CsA and SRL) therapy compared to CsA alone ¹⁵. However, malignancies are still reported. We have previously shown using the Skh-1 murine model, that combined treatment with CsA and SRL resulted in the fewest and smallest skin tumors ^{7, 16}. Similarly to Kahan et. al. tumor development was not completely prevented and therefore, more effective intervention strategies need to be developed to prevent these cancers from forming. Introducing a potential cancer chemotherapeutic agent into the immunosuppressive regime may reduce the problem of post-transplant malignancies.

Direct UVB damage accompanied by inflammation is thought to be the main cause of NMSC. A strong body of literature has linked the repeated inflammatory response associated with UV exposure to skin carcinogenesis ^17–20. Many studies, including those from our lab, have shown that repeated inhibition of UV induced inflammation can dramatically reduce skin tumor development.

A key signaling molecule in UVB induced inflammation is prostaglandin E2 (PGE₂). This prostaglandin and its receptors have been shown to be important in the development of UVB induced squamous cell carcinoma of the skin ^{21, 22}. Celecoxib, a non-steroidal anti-inflammatory drug, is a specific inhibitor of COX-2, an inducible rate-limiting enzyme in the PGE₂ synthesis pathway. Celecoxib has been shown to be effective in the inhibition of carcinogenesis in many organs including the skin ^{17, 21, 23}.

In the current studies we hypothesized that CX could be used in animals treated with immunosuppressive drugs to reduce skin carcinogenesis. Our studies found that while CX did affect tumor development, the specific effects were dependent on the treatment schedule in which it was given.

MATERIALS AND METHODS

Animals

Female Skh-1 mice, aged 6–8 weeks, were purchased from Charles River Laboratories (Wilmington, MA). Animals were housed in an Ohio State University housing facility and handled according to the requirements of the American Association for Accreditation of Laboratory Animal Care.

Immunosuppressants

An injectable form of cyclosporine A, Sandimmune (Novartis; Basel, Switzerland), and 99% pure sirolimus (Rapamycin; LC Laboratories Inc., Woburn, MA) were used. Sirolimus was dissolved in DMSO combined with Sandimmune and then diluted with PBS to a final concentration equaling 20mg CsA/2mg SRL per kg body weight when given in a 100 μl ip injection.

Celecoxib

CX in a dry crystallized powder was donated by Pfizer (New York, NY). For topical treatment, CX was dissolved in acetone at a concentration of 2.5 mg/ml. Each time CX was administered, dorsal murine skin (approximately 6 cm² area) was treated topically with 2, 100 μl applications resulting in a combined does of 500 μg.

Treatments

Skin cancer has been described as a three step process involving initiation promotion and progression. UV is considered a complete carcinogen, influencing all of these steps. Because of this we designed four treatment models to examine different stages of tumor development. Model 1: The effects of combined immunosuppression and CX treatment on UVB mediated tumor initiation and promotion was examined by immunosuppressing the mice by daily injection (ip) with CsA+SRL or vehicle and concurrently exposing mice to UVB three times weekly followed immediately by topical administration of CX. All treatments were given for 12 weeks. Then all treatments were stopped and tumors were allowed to develop for an additional 18 weeks without any further treatment (n=10 per group; Figure 1A). Model 2: To determine the effects of concurrent anti-inflammatory treatment and immunosuppression on tumor development in previously initiated and promoted skin, mice were exposed to UVB three times weekly for 15 weeks to induce tumor development. At week 15 UVB exposure was stopped and mice were injected daily (ip) with CsA+SRL and treated topically with CX or vehicle three times weekly for nine additional weeks (n=10 per group; Figure 1B). Model 3: Examination of the cutaneous microenvironment during the initiation/promotion phase was investigated by concurrent administration of UVB three times weekly, daily ip injections of CsA+SRL, and topical application of CX or vehicle three times weekly for 13 weeks, and subsequent euthanization (n=4 per group; Figure 1C). Model 4: To examine effects of immunosuppression and CX on acute UVB exposure, as an indication of the effects of these treatments on cancer initiation mice were injected daily (ip) for 10 days with CsA+SRL followed by a single exposure to UVB and a single topical treatment with CX or vehicle on day eight (n=10 per group; Figure 1D). These mice were euthanized 48 hours later at the peak of the inflammatory phase. Control groups not exposed to UVB but receiving the same drug treatments were included in each of the models.

Figure 1. Treatment schedules used.

Four treatment schedules were used to determine the effects of CX treatment on carcinogenesis. Model 1: tumor development with treatment (A), model 2: development in the absence of continued treatment (B), model 3: initiation/promotion (C) and model 4: acute UVB induced inflammation (D). Black bars indicate time periods in which daily injections of CsA+SRL (20 and 2 mg/kg respectively) were administered with thrice weekly application of topical CX (500μg), except in the acute schedule (A) were CX was only administered once after the single UVB exposure. White bars indicate treatment with UVB, either a single exposure or thrice weekly exposures. Time periods are indicated below each treatment schedule in either days or weeks. All mice were euthanized at the end of each treatment schedule.

Individual exposures to UVB light (290–320nm) were 2240 J/m², equivalent to 1 minimal erythemic dose, using Phillips FS40 UVB bulbs (American Ultraviolet Company, Murray Hill, NJ), fitted with Kodacel filters (Eastman Kodak, Rochester, NY). During chronic exposure (Figure 1A, 1B and 1C), mice were exposed dorsally to UVB light thrice weekly on non-consecutive days. Immunosuppressants were administered daily by ip injection. When CX was administered with UVB (Figure 1A, 1C, 1D) it was administered at 500 μg/200 μl immediately after individual UVB treatments. When CX was administered without UVB treatment (Figure 1B) it was applied topically thrice weekly on non-consecutive days, the identical schedule as if UVB was being administered. At the end of each experiment mice were euthanized by lethal inhalation of CO₂.

Edema

Skin fold thickness was used as an estimate of edema and was measured using digital calipers 48 hours after UVB exposure, which we have determined to be the peak of the UVB-induced inflammatory response.

De Novo Tumors

Beginning at week 12 palpable neoplastic lesions located on the dorsal skin measuring greater than 1mm in any direction were counted and measured (L × W) on a weekly basis. Measurements were made using digital calipers and the investigator was blinded to the treatment groups being measured.

Myeloperoxidase (MPO) Activity Assay

Ten mm dorsal skin punches were homogenized in 1.25 ml of 0.5% hexadecyl-phosphate buffer (pH 6.0). The homogenate was sonicated, frozen in liquid nitrogen, and thawed, three times. Samples were centrifuged for 10 minutes at 14,000 r.p.m. and 4°C. O-dianisidine dihydrochloride (0.167 mg/ml o-dianisidine dihydrochloride, and 0.0005% H2O2, in 50 mM potassium phosphate buffer pH 6.0) was combined with protein isolate in a 96-well plate. Activity was measure spectrophotometrically at 450 nm for 5 minutes with a programmable microplate reader (Molecular Devices, Menlo Park, CA). The data was expressed as mean units of myeloperoxidase, one unit being equal to the amount of myeloperoxidase necessary to degrade 1 μmol of peroxidase/minute at 25°C.

Western Immunoblot for Cleaved Caspase-3

Whole dorsal skin was crushed with a mortar and pestle in liquid nitrogen. Protein was isolated by dissolving crushed tissue in NP40 buffer and centrifuging at 13,000 rpm. Total protein was measured by bicinchoninic acid assay. Supernatant was used in western immunoblot analysis. Briefly, 40 μg of total protein was electrophoresed under reducing conditions with Laemmli buffer (BioRad, Hercules, CA) in a 12.5% criterion Polyacrylamide gel (BioRad, Hercules, CA). The protein was then transferred, via electrophoresis, to a PVDF membrane. Non-specific binding was blocked with 5% BSA in tris buffered saline with 0.1% Tween 20 (TBST) and incubated with anti-cleaved caspase-3 antibody (Becton Dickinson, Mountainview, CA) over night at 4°C. After rinsing in TBST the membrane was incubated with an HRP conjugated goat anti-rabbit antibody (Cell Signaling, Danvers, MA) for 1 hour. The immunocomplex was detected with LumiGLO chemiluminescent substrate (Cell Signaling, Danvers, MA) and Biomax light film (Kodak, Rochester, NY) was exposed. Exposed film was digitized by scanning at a resolution of 600 dpi. ImageJ was used to determine the integrated density of each band. The same membrane was stripped and re-blotted in a similar fashion with an antibody against β-actin (Cell Signaling, Danvers, MA), as a loading control.

Immunohistochemical Staining for p53 and Ly6G

Formalin fixed paraffin embedded tissue was cut 5 μm thick and mounted on glass slides. Samples were deparaffinized and rehydrated in a graded series of ethanol washes. 1N hydrochloric acid was used to uncover antigens. Non-specific binding was blocked by incubation in 10% goat serum in PBS. Rabbit anti-p53 antibody (Vector Laboratories, Burlingame, CA) was incubated at room temperature for 1 hour. Vector Link and Label were used to tag the primary antibody with HRP and diaminobenzidine was used to visualize the antigen. The mean number of foci (two or more adjacent cells with positive nuclei in the epidermis) was calculated for 5, ×100 magnification fields. Detection of Ly6G⁺ Cells, as a marker for neutrophils, was performed as described previously by Wilgus et al ¹⁷. The mean number of cells per field was determined by counting 5 fields (×600 magnification) within the dermis.

Prostaglandin E2 (PGE₂) ELISA

Whole dorsal skin was crushed with a mortar and pestle in liquid nitrogen. Eighty μg of crushed tissue was weighed out and PGE₂ was extracted by methanol for 30 minutes at room temperature. Tissue particulate was removed by centrifugation and 20 μl of extract was dried by vacuum centrifugation. Dried samples were reconstituted with ELISA diluent buffer and assayed according to Cayman Chemicals PGE₂ ELISA protocol (Cayman Chemicals, Ann Arbor, MI). Data is expressed as PGE₂: Tissue (pg/mg).

Statistical Analysis

Linear regression modeling was used to compare the shift and slope in the tumor number, total tumor area and average area per tumor over time. Data for these tests was transformed with square root. Students two tailed T-test with α=0.05 was used for all other comparisons. P<0.05 was considered to be statistically significant.

RESULTS

CX delays tumor onset and reduces tumor size but not number when given concurrently with CsA+SRL and UVB

In model 1 mice were treated for 12 weeks with UVB, CsA+SRL (ip), and topical CX or vehicle concurrently at which point all treatments were stopped and tumor development was monitored for an additional 18 weeks (Figure 1A). It was observed that CX treatment significantly delayed tumor onset. Using Kaplan Meyer survival analysis the difference in median survival time was estimated to be delayed by 3.3 weeks from week 22.7 in vehicle to week 26 in CX-treated mice (p<0.001; Figure 2A). Tumor multiplicity was not significantly affected by treatment with CX (Figure 2C). Conversely, a reduction in the mean tumor size was observed with CX-treatment. Using linear regression modeling, a significant shift downward of −0.5 mm² in tumor size over time was observed in CX-treated mice compared to vehicle (p<0.05; Figure 2B).

Figure 2. Tumor development without further treatment.

Mice were treated with UVB followed immediately by topical CX or vehicle concurrently with CsA+SRL injections for 12 weeks. Mice were then allowed to develop tumors without any further treatments for 18 weeks (Figure 1B). Mice treated with CX showed significantly delayed time to first tumor (A, p=0.006). There were no differences in tumor multiplicity (B) but a significant reduction in the rate of tumor growth was seen in CX-treated mice compared to vehicle (C; p<0.01).

CX reduces tumor number but not size when administered after cessation of UVB

To examine the effect of topical CX treatment on tumor development in an immunosuppression model without continued UVB exposure, mice were exposed to UVB for 15 weeks followed by treatment with only CsA+SRL (ip) and topical CX or vehicle for an additional 9 weeks (model 2; Figure 1B). The time in which it took to develop 10 tumors was significantly delayed in the CX-treated mice compared to vehicle control, as measured by Kaplan Meyer Survival analysis (p=0.001; Figure 3A). Accordingly, mice treated topically with CX had fewer tumors per mouse. Using linear regression modeling the rate of tumor onset over time was significantly decreased from 1.09 tumors per week in the vehicle-treated group to 0.71 tumors per week in the CX-treated group (Figure 3B; p=0.006). Tumor size however, was not significantly altered by treatment with CX in this treatment schedule (Figure 3C).

Figure 3. Tumor development with CsA+SRL.

Mice were treated for 15 weeks with UVB only. After week 15 mice were treated daily with CsA+SRL and thrice weekly with topical CX or vehicle for 9 weeks without continued UVB exposure (Figure 1A). Kaplan Meyer survival analysis shows that treatment with CX significantly delayed the time it took for individual mice to reach a threshold of 10 tumors (A; p<0.001). Similarly, in the presence of CX tumor multiplicity was significantly reduced compared to vehicle-treated mice (B; p=0.006), however, mean tumor size was not significantly affected (C).

CX reduces dermal infiltrating neutrophils and MPO activity

Immunohistochemical detection of Ly6G was used to determine the number of dermal infiltrating neutrophils in the dorsal skin of mice treated acutely with UVB according to the protocol depicted in model 4 (Figure 1D). Treatment with CX reduced the mean number of neutrophils in the dermis of these mice by over three fold compared to vehicle (p=0.02; Figure 4A). MPO activity in the same skin was significantly decreased in the CX-treated mice compared to vehicle from 0.004 units to 0.002 units respectively (p=0.048). Furthermore when compared to activity levels in unirradiated control mice only UVB irradiated vehicle-treated and not CX-treated dorsal skin had a significant increase in MPO activity (p=0.02; Figure 4B).

Figure 4. Reduced neutrophil infiltration and MPO activity with CX treatment.

Mice were treated for 10 days with CsA+SRL and either topical vehicle or CX. A single exposure to UVB was given on day 8 (Figure 1D). Dorsal skin harvested 48 hours after UVB exposure was paraffin embedded. Mice treated topically with CX had significantly fewer dermal infiltrating neutrophils as determined by immunohistochemical staining for Ly6G (A). Bars represent the mean number of cells per X600 field ±SD (* p=0.02). MPO activity was measured spectrophotometrically from protein isolates of whole dorsal skin (B) treated as previously described (black bars), or from control mice not exposed to UVB (white bars). Bars represent mean MPO activity in arbitrary units ±SD (* p=0.04, † p=0.02).

CX reduced levels of activated caspase-3 and the number of p53⁺ foci

Dorsal skin from mice in model 3 treated concurrently with UVB, CsA+SRL (ip) and topical CX or vehicle was examined for p53⁺ foci by immunohistochemistry and cleaved caspase-3 protein levels by western immunoblot. P53⁺ foci were reduced with topical CX treatment from 1.4 in the vehicle to 0.6 per ×100 magnification field. This difference, however, was not statistically significant (p=0.21; Figure 5A). Activated caspase-3 levels were reduced 2.8 fold from vehicle-treated mice (p=0.025; Figure 5B).

Figure 5. Topical CX reduces cleaved caspase-3 but not p53 foci.

Dorsal skin exposed to UVB, treated ip with CsA+SRL and topically with CX or vehicle (Figure 1C) was evaluated for p53 foci. P53 foci were detected by immunohistochemistry with no significant differences detected. Representative photomicrographs of acetone-treated and CX-treated epidermis are shown with a CsA+SRL+acetone unirradiated control shown as an insert, black bars are 50 μm (A). Individual groups of p53 positive cells were counted in 5 X600 fields. Bars represent the mean ±SD (B; p=0.569). A significant decrease in cleaved caspase-3 was detected in CX-treated dorsal skin when measured by western immunoblot. Representative bands are shown with bands for β-actin loading control (C). A graph depicting densitometric analysis of the 17kDa band is shown in D. Bars represent mean value ±SD (* p=0.025).

CX reduced cutaneous PGE₂ levels following chronic UVB exposure

To make certain that CX was effectively inhibiting COX-2, whole skin PGE₂ levels were determined by ELISA. Skin from model 3 was used to evaluate PGE₂ levels in the skin because this time point is before the establishment of tumors, which could influence PGE₂ levels, but still shows relative levels in the skin following chronic UVB-exposure. The PGE₂ levels in dorsal skin from mice treated concurrently for 13 weeks with UVB, CsA+SRL and CX (Figure 1C) was significantly reduced compared to that in vehicle-treated controls (Figure 6; p=0.02).

Figure 6. Topical celecoxib reduces whole skin PGE₂.

Mice were treated for 13 weeks with UVB and CsA+SRL followed immediately by topical administration of either vehicle or CX (Figure 1C). Whole dorsal skin was crushed and PGE₂ was extracted by methanol and measured via an ELISA based immunoassay. Treatment with celecoxib significantly reduced PGE₂ levels compared to vehicle. Bars represent the mean ±SD (A; * p=0.05).

DISCUSSION

We have shown previously that combined therapy with SRL and CsA reduces UVB induced cancer incidence in immunocompetent mice ⁷ compared to vehicle alone or CsA alone. The effect however was not as effective as that seen with CX in the absence of immunosuppression ¹⁷. In the current set of studies we have shown that topical CX treatment can delay and reduce UVB induced carcinogenesis in animals exposed to immunosuppressive treatment.

Multiple models were used to examine the effects of CX on different stages of UVB-induced carcinogenesis in mice being treated with therapeutic immunosuppressants. Models 1 and 2 demonstrated the chemotherapeutic efficacy of CX whereas model 3 allowed for determining a potential mechanism by which CX exerted its chemopreventative effects. Models 1 and 3 may also be representative of the efficacy of CX we could potentially expect from pediatric transplant recipients undergoing immunosuppression during exposure to UV, a complete carcinogen. Model 2 may represent the potential efficacy of CX on adult transplant recipients showing the effects of immunosuppression on previously carcinogen exposed skin. Since we had previously shown a link between the effects of CX on the acute UVB-induced inflammatory response and tumor development we included model 4 to evaluate the effects of this drug on acute inflammation in immunosuppressed mice.

We have previously shown that topical CX treatment of immunocompetent SKH-1 mice was effective as a chemopreventative agent but did not induce regression of established tumors ²⁴. The current study showed similar effects of CX in that when it was applied concurrently with UVB exposure (model 1), tumor onset was delayed. Interestingly, CX given during chronic UVB exposure and immunosuppressive therapy (weeks 1–12 as shown in Figure 1A) was able to cause a decreased rate of tumor growth over 17 additional weeks without any further therapy (Figure 2B). Initiation of concurrent systemic immunosuppression and topical CX treatment in chronically UVB exposed tumor bearing mice (model 2 in Figure 1B) slowed new tumor emergence as seen in the reduced tumor number, but did not affect the tumor size (Figure 3). These results demonstrate that CX may be an effective chemopreventative agent in immunosuppressed mice as well as in immunocompetent mice ^{17, 24, 25}.

Model 3 is an important time-point to examine because it gives a snapshot of the skin after chronic UVB-exposure and damage but before tumors have become palpable. To evaluate the effects that topical CX treatment had on UVB-induced cutaneous damage we examined the level of caspase-3 as well as the number of p53⁺ foci within the epidermis. Caspase-3 has been identified as being a key mediator of apoptosis in mammalian cells. In response to cellular damage a cell can undergo cell cycle arrest and initiate repair, or if damage is too extensive apoptosis is triggered. This trigger can come from internal or external signals, such as Fas or NF-κβ, which result in a cascade leading to the cleavage of caspase-3, and ultimately cellular death. A primary switch in the decision to undergo cellular repair or apoptosis is the p53 protein. P53 is a tumor-suppressor gene that has been shown to play an important role in multi-step carcinogenesis, particularly in UV-induced tumorigenesis. The p53 gene encodes a 53-kDa phosphoprotein that plays an important role in the control of cell proliferation ²⁶. DNA damage following exposure to stimuli such as UVB induces the expression of wild-type p53 by a post-translational mechanism ²⁷. If the damage can not be repaired this arrest can lead to apoptosis ^{28, 29}. Therefore, p53 plays a key physiological role in limiting mutagenic damage. Thus mutations in p53, which would prevent it from carrying out this important function, would allow for an increased rate of accumulation of genetic damage in the cell. The number of p53⁺ foci can be used to suggest the relative number of precancerous lesions ³⁰. It is important to note that we used an antibody that recognizes both wild type and mutant p53. However, data was collected after the peak of normal p53 expression after UVB. Furthermore, Berg et al have shown that the majority of mutant p53⁺ foci present in the epidermis after chronic UV exposure remain for long periods after cessation of UV ³¹. While there was a trend toward a decrease in p53⁺ foci in the epidermis of CX-treated mice, we did not find statistically significant differences compared to vehicle-treated mice (Figure 4B). This is not surprising considering that in model 1 we did not see significant differences in tumor number with topical CX treatment (Figure 1C). In contrast we detected reduced caspase-3 levels in CX-treated skin (Figure 4A) suggesting CX caused a reduction in cellular damage leading to a reduced stimulus to undergo apoptosis ³² as well as a reduction in the promotion of carcinogenesis.

In model 4 the initial UVB response in mice undergoing therapeutic immunosuppression and topical CX treatment was determined. As we have previously shown there is a correlation between the acute UVB response and long-term outcomes, where reduction in markers of inflammation, including neutrophil infiltration and MPO activity, predicts a reduction in carcinogenesis ¹⁷. Here we found CX treatment of immunosuppressed mice reduced neutrophil infiltration and MPO activity compared to vehicle treated immunosuppressed mice (Figure 4). This data suggests that a potential mechanism by which CX has chemotherapeutic efficacy on UVB mediated skin carcinogenesis in immunosuppressed mice is by inhibiting chronic inflammation.

It is important to consider that recently celecoxib has been shown to inhibit other target enzymes besides COX-2 ³³. Carbonic anhydrase and phosphoinositide 3- kinase are two such enzymes that can affect the carcinogenesis process in a number of organs. Higher levels of carbonic anhydrase have been detected in a chemical murine skin carcinogenesis model and are thought to be associated with a cell proliferative advantage ³⁴. The PI3K/AKT pathway has been implicated in cancer, driving many pro-cancer qualities such as proliferation and escape from apoptosis ³⁵. One of the most important downstream effectors in this pathway is mammalian target of rapamycin (mTOR) ³⁶. In the current studies we have not ruled out the possibility that the mechanism of action of CX was to block either of these pathways. However, the use of sirolimus (an inhibitor of mTOR) in both treatment groups would make the possibility that the inhibitory effects of CX on tumor formation are due to modulation of PI3K activity unlikely. Furthermore, we were able to show that levels of PGE₂, a COX-2 product and key inflammatory component, were reduced, suggesting that the observed effects were a result of the decrease in inflammation. Furthermore, MPO and neutrophil infiltration were reduced after acute UVB exposure (model 4 in Figure 1; Figure 4), which we have previously shown to be associated with reduced tumor formation ¹⁷.

In conclusion, we have shown beneficial effects of topical CX treatment during and after chronic UVB exposure in combination with therapeutic immunosuppression. This suggests that a topical formulation of CX or potentially an alternative COX-2 inhibitor when applied at the appropriate time may be beneficial for the transplant recipient population to reduce and/or delay the onset of NMSC.