These authors contributed equally to this work.
Clinically Relevant Immunosuppressants Influence UVB-Induced Tumor Size Through Effects on Inflammation and Angiogenesis
Article first published online: 30 OCT 2007
American Journal of Transplantation
Volume 7, Issue 12, pages 2693–2703, December 2007
How to Cite
Duncan, F. J., Wulff, B. C., Tober, K. L., Ferketich, A. K., Martin, J., Thomas-Ahner, J. M., Allen, S. D., Kusewitt, D. F., Oberyszyn, T. M. and VanBuskirk, A. M. (2007), Clinically Relevant Immunosuppressants Influence UVB-Induced Tumor Size Through Effects on Inflammation and Angiogenesis. American Journal of Transplantation, 7: 2693–2703. doi: 10.1111/j.1600-6143.2007.02004.x
For the Trans-Atlantic Research Group for Experimental Therapy of Skin Cancers in Immunosuppressed INdividuals (TARGET-SCIN)
- Issue published online: 30 OCT 2007
- Article first published online: 30 OCT 2007
- Received 12 July 2007, revised 15 August 2007 and accepted for publication 31 August 2007
- posttransplant malignancies;
- skin cancer
Immunosuppressive therapies allow long-term patient and transplant survival, but are associated with increased development of UV-induced skin cancers, particularly squamous cell carcinomas. The mechanisms by which CsA, MMF, tacrolimus (TAC) or sirolimus (SRL), alone or in dual combinations, influence tumor development and progression are not completely understood. In the current study, chronically UV-exposed mice treated with SRL alone or in combination with CsA or TAC developed more tumors than mice treated with vehicle or other immunosuppressants, but the tumors were significantly smaller and less advanced. Mice treated with CsA or TAC developed significantly larger tumors than vehicle-treated mice, and a larger percentage in the CsA group were malignant. The addition of MMF to CsA, but not to TAC, significantly reduced tumor size. Immunosuppressant effects on UVB-induced inflammation and tumor angiogenesis may explain these findings. CsA enhanced both UVB-induced inflammation and tumor blood vessel density, while MMF reduced inflammation. Addition of MMF to CsA reduced tumor size and vascularity. SRL did not affect inflammation, but significantly reduced tumor vascularity. Thus the choice of immunosuppressants has important implications for tumor number, size and progression, likely due to the influence of immunosuppressants on UVB-induced inflammation and angiogenesis.
Transplant recipients have a 60–250 fold increased risk for developing nonmelanoma skin cancer (1–4), and tend to develop multiple aggressive skin tumors (5–9) that can be life threatening (9,10). Indeed, one Australian study estimated that 27% of heart transplant recipients who had survived greater than 4 years posttransplantation died of metastatic squamous cell carcinoma (SCC) (11). Sun exposure is the main risk factor for nonmelanoma skin cancer development in immunocompetent individuals (12–14). In addition to sun exposure, the level and duration of immunosuppression contribute to increased skin cancers in transplant patients (4,15). The mechanisms underlying this increase are not completely understood, nor is it known whether all immunosuppressive regimens increase skin cancer to the same extent (16).
Although less than 1–2% of the UV light from the sun is UVB radiation (290–320 nm), UVB is primarily responsible for the skin damage from acute and chronic sun exposure (17–19). One immediate physiologic consequence of UVB exposure is skin inflammation (20–24), characterized by edema (25) and dermal neutrophil infiltration (26,27). Neutrophils contribute to skin inflammation by producing myeloperoxidase (MPO), reactive oxygen intermediates and pro-inflammatory cytokines (reviewed in (24,28–31), which can contribute to tumor growth by enhancing angiogenesis (32–34). Reducing inflammation in these early stages reduces both angiogenesis (reviewed in (35)) and tumor growth (27,36). However, once tumors are established, increasing the inflammatory response through the use of immune modulators has been shown to be an effective skin cancer treatment (37–40).
To determine how clinically relevant immunosuppressant treatments influence UVB-induced tumor development and progression, we assessed the effects of CsA, MMF, tacrolimus (TAC) and sirolimus (SRL) alone or in combination on tumor number, size, progression, angiogenesis and on UVB-induced inflammation. We observed that SRL-based regimens resulted in more tumors than vehicle (VEH) controls, but these tumors were significantly smaller. In contrast, CsA did not increase the number of tumors compared to VEH, but the tumors were larger and a greater percentage was malignant. Furthermore, CsA exacerbated UVB-induced inflammation and resulted in more vascular tumors compared to VEH, while the addition of MMF to CsA reduced both inflammation and tumor vascularity. Thus, our data indicate that specific immunosuppressant regimens influence tumor size and progression through effects on inflammation and tumor angiogenesis.
Materials and Methods
Adult female SKH-1 hairless mice (Charles River Laboratories, Wilmington, MA), 8–12/group, were maintained in an accredited Ohio State University vivarium. All procedures were approved by the Institutional Laboratory Animal Use and Care Committee.
UV Irradiation: Mice were exposed to 2240 J/m2 of UVB (27), generated by FS40UVB lamps (American Ultraviolet Company, Lebanon, IN) covered by Kodacel filters (Eastman Kodak, Rochester, NY). UVB levels were measured using a UVX radiometer (UVP Inc, Upland, CA).
Immunosuppressants: CsA (Sandimmune®, Novartis, Basel, Switzerland), MMF (CellCept®, Roche Pharmaceuticals, Nutley, NJ) and tacrolimus (TAC–Prograf®, Astellas, Deerfield, IL) were all purchased through The Ohio State University pharmacy; Sirolimus (rapamycin) (SRL) was obtained from LC Laboratories (Woburn, MA). CsA, MMF or vehicle (VEH, PBS) was injected i.p. once daily at a dose of 20 mg/kg/day, while TAC was given at a dose of 2 mg/kg/day. SRL was administered at a dose of 2 mg/kg/day. The immunosuppressants were either given alone or in the following combinations: CsA + MMF, TAC + MMF, CsA + SRL, TAC + SRL. Drug doses were based on published effective doses in rodents (41). Drug levels were not changed in combination therapies.
Study design (Figure 1): For acute studies, mice received immunosuppressants or VEH i.p. daily for 1 week prior to receiving 0 or 2240 J/m2 UVB. Mice were euthanized, skin fold thickness determined and tissues recovered for analysis 48 h after UVB exposure.
For carcinogenesis studies, mice received 2240 J/m2 UVB thrice weekly for 23 weeks. Beginning in week 10, mice were randomly assigned to receive VEH, CsA, MMF, TAC, SRL, CsA + MMF, CsA + SRL, TAC + MMF or TAC + SRL i.p. once a day for the remaining 13 weeks. Beginning at week 11, tumors ≥1 mm diameter were counted in a blinded fashion. Measurements were performed using digital calipers, and tumor area calculated multiplying length × width. Mice were euthanized and tissues collected for analysis at the end of week 23.
Tumor staging: Tumor stage was determined histologically in a blinded fashion by a board-certified veterinary pathologist (DFK). Grading of tumor progression is based on invasion (42). Exophytic papillomas stages 1–3 are considered benign. Micro-invasive SCC (miSCC) and SCC, which invades the panniculus carnosus are considered malignant.
Myeloperoxidase (MPO) assay: MPO activity was assessed as previously described (43). Briefly, MPO activity in a 10-mm2 dorsal skin punch was measured over a 5-min period at 450 nm with a programmable microplate reader. Results are reported as the fold-increase over activity in matched, non-UVB-exposed controls.
Fluorescent staining for CD31: Frozen sections of tumors were fixed in acetone and blocked, then incubated with anti-CD31 antibody (BD Biosciences, San Diego, CA). Staining was visualized using goat anti-rat IgG Alexa Fluor 594 (Invitrogen, Carlsbad, CA). Vascularity was determined similarly to the method of Bolontrade et al. (44). Images were captured at 200× (20× objective, 10× eyepiece) on a digital fluorescent microscope and loaded into Image J v1.38b (http://rsb.ingo.nih.gov/ij/), which was used to adjust the threshold so that only CD31+ staining was measured. A 640 × 640 pixel box was drawn and the percent of the pixels above threshold were measured. This was repeated four times for a total of five measurements/slide, with 8–12 slides/group. Images of a micrometer were captured under identical conditions and used to determine pixels/μm. The area in pixels was then converted to area in mm2. Data are shown as the average area in mm2× 105± SE.
Statistical analyses: To determine if there were significant differences in the numbers of tumors per group, MPO activity or CD31 staining, normally distributed data were analyzed using Student's t-test, while nonnormally distributed data were analyzed using the Wilcoxon rank-sum test. Individual p-values <0.025 were considered significant in the pair-wise comparisons.
To determine if there were significant differences between groups with respect to the size of the tumors, a repeated measures analysis of variance model was fit to the data using the MIXED procedure in 'SAS Institute, Cary, NC', version 9.1. The model assumptions of normality and equal variances of the residuals were checked and the data transformed using a natural log transformation. Outliers in the data still remained after transformation. The residuals from the model were still not normally distributed and this could have an effect on the p-values presented. The exact Wilcoxon rank-sum test was performed to compare diagnosis level between the groups and between the sizes of tumors. For each of these analyses, a Bonferroni adjustment was made to each set of comparisons to assure that the overall significance level remained at 0.05. Thus, to be significant, a p-value from a specific test had to be 0.025 or lower for the two-pair comparisons. For the comparisons with VEH, the p-value had to be 0.0063 or lower.
Effects of immunosuppression on UVB-induced skin cancer
Tumor number and size: To assess the effects of different immunosuppressive treatments on tumor development in a well-established animal model that mimicked a common clinical situation, SKH hairless mice were exposed to UVB for 23 weeks, and immunosuppressed for the final 13 weeks. VEH-treated animals developed 12.7 ± 3.5 tumors per mouse (Figure 2). Neither CsA, TAC nor MMF significantly altered the number of tumors per animal, yielding 10.4 ± 3.8, 14.3 ± 2.5 and 14.9 ± 5 tumors per mouse, respectively. In contrast, SRL treatment, either alone or in combination with CsA or TAC, resulted in increased tumor numbers compared to VEH (20.6 ± 5.1, p = 0.002; 21.0 ± 7.6, p = 0.02; 21.0 ± 5.1, p = 0.002, respectively). We also assessed tumor size in these treatment groups (Table 1). Tumors from CsA or TAC-treated animals were significantly larger than those from VEH-treated mice (p < 0.001 and p = 0.0057, respectively). In contrast, tumors from MMF-treated animals were not significantly different in size from those of VEH-treated animals. The addition of MMF to CsA significantly reduced the tumor size compared to CsA treatment (p = 0.007), such that these tumors were not significantly different in size compared to VEH-treated mice. Addition of MMF to TAC did not significantly decrease tumor size compared to TAC alone. In contrast, all of the SRL treatment groups (SRL, CsA + SRL and TAC + SRL) had significantly smaller tumors than VEH (p < 0.001, p = 0.0028, p < 0.001, respectively).
|Treatment||Number of tumors||Median size, mm2||Size range, mm2||p-Value, compared to vehiclea||p-Value of combination to calcineurin inhibitor alone|
|CsA + MMF||103||5.05||1.1–77.6||0.1216, N.S.||0.007|
|CsA + SRL||168||3.06||1.0–35.91||0.0028||<0.001|
|TAC + MMF||141||5.26||1.1–252.17||0.0022||0.7252, N.S.|
|TAC + SRL||210||2.21||1.0–91.4||<0.001||<0.001|
Tumor progression: To determine if immunosuppression influenced tumor progression, three randomly selected tumors (one each small, medium, large) from each animal were staged on the basis of invasion (42). Figure 3A shows the stages of the benign and malignant tumors in each treatment group. Tumors from VEH-treated animals showed a range of benign and malignant phenotypes, with benign tumors being most prevalent and malignant tumors comprising 26.9% of the tumors examined. With the exception of CsA, similar patterns were observed with other immunosuppressants. Malignant tumors were the most prevalent tumors in CSA-treated animals, comprising 46.4% of the tumors. Addition of MMF or SRL to CSA reduced the percentage of malignant tumors from 46.4% to 26.9% and 22.2%, respectively. The percentage of malignant tumors in animals treated with SRL alone was only slightly decreased, but the percentage of stage 1 papillomas increased from 11.5% to 32%. Thus, it appears that CsA treatment may enhance tumor progression, but the addition of MMF or SRL reduces that enhancement. Further, SRL may slightly slow tumor progression, as stage 1 papillomas were increased in this group.
Increasing tumor size is thought to correlate with increased malignancy. However, only pathological analysis can adequately determine tumor stage. Figure 3B demonstrates the association of tumor size with increasing malignancy. Overall, malignant tumors were significantly increased in large tumors compared to small tumors (p < 0.001) or medium tumors (p < 0.001, not shown). As expected, there was a significant increase in malignancy going from small to large tumors in the control group VEH (p = 0.0012). Further, both CsA (p < 0.001) and MMF (p = 0.0029) groups also showed a similar association of size with malignancy. Interestingly, there was no significant increase in malignancy comparing small versus large tumors in the TAC or SRL groups, or the CsA + MMF group. These data therefore suggest that TAC, SRL or CsA + MMF may slow malignancy compared to VEH.
Effects of immunosuppression on the acute inflammatory response to UVB
MPO activity: UVB-induced inflammation, characterized by an increase in skin MPO activity, is linked to the development and progression of skin tumors. Since CsA-treated mice had larger tumors, we hypothesized that CsA might exacerbate UVB-induced inflammation (Figure 4). Compared to VEH, CsA significantly increased (p = 0.013), while MMF significantly reduced skin MPO activity (p < 0.003). Neither TAC nor SRL altered MPO activity relative to VEH. However, the addition of either MMF or SRL to CsA reduced MPO activity to VEH levels. This reduction was statistically significant compared to CsA alone (p < 0.02).
Blood vessel density: Chronic inflammation is associated with increased angiogenesis (32–34) that is essential for tumor growth and progression (45–47). We hypothesized that the increased inflammation in CsA-treated animals would result in more vascular tumors than VEH. We assessed CD31 immunofluorescence in tumors from VEH, SRL, CsA, MMF and CsA + MMF groups and calculated the percentage of each section staining positively for CD31 (Figure 5A). SRL treatment significantly reduced (p = 0.009) CD31 staining relative to VEH treatment. MMF treatment alone did not significantly alter CD31 staining, although there was a trend toward less staining (p = 0.032) compared to VEH. The addition of MMF to CsA resulted in significantly less CD31 staining compared to CsA and to VEH (p = 0.002, p = 0.017, respectively). Tumors from SRL-treated mice had less CD31 staining compared to VEH-treated mice (Figure 5B, panels a, b), indicating reduced angiogenesis in these tumors. In contrast, tumors from CSA-treated mice had increased CD31 staining compared to tumors from VEH-treated animals (Figure 5B, panel c). MMF did not significantly alter CD31 staining compared to VEH (panel d). However, the addition of MMF to CsA greatly reduced CD31 staining (panel e).
Thus, our data indicate that CsA allows increased tumor growth and progression, likely due in part to an exacerbation of UVB-induced inflammation and angiogenesis. MMF does not appear to directly inhibit angiogenesis significantly, but can reduce the enhanced levels of inflammation and angiogenesis caused by CsA, and may thereby slow tumor growth. In contrast, SRL reduces tumor size in this de novo tumor model through an inhibition of angiogenesis likely separate from effects on inflammation.
Nonmelanoma skin cancers, especially SCC, are dramatically increased in immunosuppressed populations, such as transplant recipients (1,4,9,48), HIV/AIDS patients (6,49,50), cancer patients (5,51–54) and patients with autoimmune diseases (55–58). Organ transplant recipients have the highest risk for developing nonmelanoma skin cancers, and skin cancers in these patients demonstrate more aggressive characteristics than skin cancers in nontransplant recipients, including early dermal invasion, an infiltrative growth pattern and greater depth of invasion at diagnosis (59). This is thought to be due to a combination of sun exposure and the continued administration of immunosuppressive drugs to prevent or treat organ rejection (4,15).
Using data from human transplant patients, it has been difficult to determine which specific immunosuppressive drugs are responsible for the increased skin cancer risk (16) or to determine the general mechanisms underlying the increased risk (16,60,61). Most transplant recipients are adults when they receive their new organs, and have accumulated years of sunlight-related skin damage. Although some patients modify their behavior to use sunscreen and stay out of the sun, most do not (62–64). We therefore used the hairless mouse model of posttransplantation skin cancer to study the effects of specific immunosuppression regimes on UVB-induced inflammation and to mimic both previous and ongoing sun exposure of adult transplant patients. It is important to note that this is a de novo model of tumor development, and the animals, other than being treated with immunosuppressants, are immunocompetent.
CsA is reported to promote tumorigenesis through increased TGF-β production (65), decreased DNA repair (66) and increased angiogenesis (67). We expected that CsA-treated animals would develop more tumors than VEH-treated mice in response to chronic UVB exposure. Although CsA and VEH-treated mice developed similar numbers of tumors, the tumors were significantly larger in the CsA group (median sizes of 5.55 mm2 compared to 3.97 mm2) and overall, more of the tumors had progressed to malignancy (46.4% in CsA). TAC treatment also resulted in larger tumors compared to VEH treatment. MMF treatment alone had no effect on tumor number or size. Combining MMF with CsA significantly reduced the median tumor size from 5.55 mm2 to 5.05 mm2, and MMF blocked the development of extremely large tumors. In contrast, combining MMF with TAC did not reduce tumor size or the variability in sizes. Some of the size differences, although statistically significant, were subtle, and may not be biologically important. CsA + MMF reduced the median size to 5.05 mm2, which was not different from VEH. However, 5.09 mm2 for TAC was significantly larger. There was a wide range of tumor sizes, with some tumors in the CsA and TAC + MMF groups becoming extremely large. These very large tumors were invariably malignant. Thus, the ability of MMF to prevent the growth of very large tumors may be important. The outbred nature of the hairless mice may partially explain the wide range of tumor sizes. Definitive comparisons of the different treatment groups for their effects on tumor progression are not possible, as we only had 26–30 tumors per group available for analysis and the p-values to achieve significance needed to be less than 0.0063. Nonetheless, there was a trend toward increased malignancy with CsA and reduced malignancy with SRL, compared to VEH. Likewise, the addition of MMF to CsA appeared to ameliorate the increased malignancy observed with CsA alone, but this did not achieve statistical significance. We did observe a statistically significant association of increasing tumor size with increasing malignancy in the VEH, CsA and MMF groups, and among the different tumor sizes overall. Based on these data, we suggest that CsA specifically promotes tumor growth and progression, but that this effect can be ameliorated by the addition of MMF to the treatment regimen. Although we used established dosing regimens for all drug combinations, a general limitation of our studies is that we were unable to obtain systemic levels for all drug combinations. Therefore, we cannot rule out that blood levels of immunosuppressants may have influenced our results.
Interestingly, SRL treatment resulted in increased tumor numbers, but the tumors were significantly smaller than tumors in VEH-treated mice and more of these were stage 1 papillomas (32% vs. 11.5% for VEH). In contrast to VEH, the SRL treatment did not show a statistically significant association of increasing size and increasing malignancy, suggesting that SRL may slow or inhibit tumor progression. The increase in tumor number with SRL was unexpected based on previous reports indicating an anti-tumor effect of SRL. Many reports showing anti-tumor effects of mTOR inhibitors have assessed the growth of transformed cell lines injected into mice, including the landmark paper by Guba et al. (67), which concluded that SRL limited tumor growth by inhibiting angiogenesis. Two reports have investigated mTOR inhibitors in de novo tumor formation in genetically engineered mice. Koehl et al. showed an anti-tumor effect of SRL compared to MMF in p53−/− mice; in these studies, mice treated with SRL had reduced tumor burden compared to controls or to animals treated with CsA or MMF (68). Likewise, Mabuchi et al. showed that another mTOR inhibitor, everolimus, reduced tumor formation and progression in a transgenic model of spontaneous ovarian adenocarcinoma (69). These two reports did not directly assess the mechanism underlying this effect, but the findings are consistent with effects of mTOR inhibitors on angiogenesis. There are two main differences between our study and the previous reports. First, our source of mTOR inhibitor and mode of administration is different: Koehl et al. obtained SRL from Wyeth and Mabuchi obtained their mTOR inhibitor from Novartis, while we used a generic version of SRL obtained through LC laboratories. Secondly, Koehl et al. fed the SRL to the mice, while we administered SRL i.p. We cannot rule out the possibility that the fluctuation of SRL levels with i.p. injection influenced tumor development. Perhaps most importantly, both of the previous reports examined the spontaneous development of tumors in genetically engineered animal models, while our data were obtained from a nongenetically engineered model of UVB carcinogenesis closely paralleling the clinical situation. Mice were exposed to UVB for 10 weeks prior to being immunosuppressed, so much of the mutational damage had already occurred, similar to an adult transplant recipient who has already had a lifetime of sun exposure. Our data indicate that SRL does not prevent tumor appearance or, by inference, the mutational damage preceding tumorigenesis. However, the markedly reduced size of the tumors and the slightly delayed tumor progression suggested that SRL inhibition of angiogenesis may account for our observations. In addition, spread of the large tumors in the other treatment groups could mask several smaller tumors, potentially leading to a misleadingly low tumor number. Mutational analysis may provide information regarding this possibility.
To investigate the mechanisms by which the different immunosuppressive treatments affected tumor size, we assessed the effects of immunosuppression on UVB-induced inflammation. The link between chronic inflammation from sun exposure and skin cancer development has been established in both humans (70–72) and animal models (27,73). Chronic inflammation is believed to promote genetic changes, tumor angiogenesis, tumor progression and metastasis (reviewed in (32,33,74)). In animal models, administration of anti-inflammatory compounds, orally or topically, reduces UVB-induced inflammation in the skin and also reduces the number of subsequent skin tumors by 35–50% (36,75).
Inflammatory responses to UVB were measured in immunosuppressed hairless mice 48 h after exposure. We analyzed edema and MPO activity, which are routinely used as indicators of the degree of the cutaneous inflammatory response (42,43). Edema results from enhanced vascular permeability, while MPO is produced by neutrophils to catalyze the conversion of hydrogen peroxide to hypochlorous acid and is generally used as an indicator of neutrophil activation (76). Only CsA increased both skin edema (not shown) and MPO activity. Calcineurin inhibitors like CsA are used to treat atopic dermatitis and other skin inflammatory conditions (77); however, these are applied topically, not used systemically as was done in our study. Moreover, the two FDA-approved topical calcineurin inhibitor formulations are pimecrolimus and TAC, not CsA. We tested systemic TAC in our system and found that it had no effect on UVB-induction of MPO activity. Furthermore, CsA is reported to decrease DNA repair in the skin, and persistent DNA damage may independently lead to increased inflammation (66). Interestingly, while SRL alone had no effect on MPO activity, the addition of SRL to CsA significantly reduced MPO activity compared to CsA alone.
MMF significantly reduced MPO activity, alone or when combined with CsA. Although MMF is used widely to prevent transplant rejection and was designed to inhibit inosine monophosphate dehydrogenase, it has multiple effects on many different immune cells (reviewed in (78)). Farivar et al. reported that MMF treatment reduced MPO activity and leukocytic infiltration in a model of lung reperfusion injury, clearly demonstrating that MMF has anti-inflammatory properties (79). MMF is used in dermatology to reduce inflammation in dermatitis and vasculitis, and its efficacy is thought to be related to effects on endothelial cells that reduce neutrophil infiltration (80–82).
We used CD31 staining to assess tumor blood vessel density as a marker for angiogenesis in tumors from VEH, CsA, MMF, CsA + MMF and SRL groups. Tumors from SRL-treated mice had less, while tumors from CsA-treated mice had more CD31 staining compared to VEH. These data confirm the findings of Guba et al. (67), in a physiologic, de novo tumor model. There was a trend toward reduced CD31 staining with MMF treatment, although it did not reach significance. Interestingly, when combined with CsA, MMF significantly reduced CD31 staining. Recently, Koehl et al. (83) observed that MMF had powerful, consistent anti-angiogenic effects on vascular endothelial cells in vitro, but had less consistent effects in an in vivo model using injected, transformed cell lines. These authors presented data suggesting that the different effects may relate to the bioavailability of MMF in vivo (83). We used significantly lower doses of MMF (20 mg/kg/day vs. 40 mg/kg/day and 80 mg/kg), so we cannot rule out that MMF bioavailability was limited in our studies. However, additional explanations are that MMF may act preferentially to reduce neutrophil activity rather than acting on endothelial cells to directly reduce angiogenesis. Thus, MMF may have variable direct effects on angiogenesis in vivo. However, when CsA is present and neutrophil activity is dramatically increased, MMF could reduce the enhanced neutrophil infiltration and/or activity and therefore indirectly reduce angiogenesis. Alternatively, MMF may act on UVB-exposed and/or neoplastic epithelium, which elaborates several angiogenic factors.
Our experiment mimicked an adult transplant recipient who received regular UVB exposure throughout their life and did not alter their sun behavior posttransplantation. Thus, inflammation and DNA damage were repeatedly induced throughout the study. Although our experimental design may best reflect the current clinical situation, understanding the precise effects of immunosuppression on skin tumor formation and progression requires additional experimental designs. For example, it will be necessary to assess the effects of immunosuppression when UVB exposure precedes immunosuppression, as in transplant recipients with a lifetime of sun exposure who alter their sun behavior posttransplantation. Data indicate that when immunosuppressive treatments are administered in the absence of continued UVB exposure, CsA and SRL treatments both result in fewer tumors than VEH, but CsA tumors are larger, and SRL tumors smaller, than VEH (Wulff et al., in review). Likewise, immunosuppression may have different effects when administered concomitantly with UVB exposure from the beginning of the experiment, mimicking the experience of a pediatric transplant recipient. Preliminary data indicate that in this instance, SRL treatment results in fewer and smaller tumors than VEH (De Gruijl and Geissler, in review).
In conclusion, our data indicate that the choice of immunosuppressant administered after chronic UVB exposure can have significant effects on tumor number and size. It is likely that the influence of immunosuppressants on UVB-induced inflammation and angiogenesis play important roles in mediating these effects.
This work was supported in part by the National Institutes of Health (NIH P30 CA16058 [AMV, AKF, TMO, DFK], R03 CA 110054 [AMV] R01 CA 109204 [TMO]), the American Heart Association (AMV) and Roche Pharmaceuticals CEL 508(AMV). The authors would like to thank Drs Ginny Bumgardner, Gregg Hadley, Frank deGruijl and Edward Geissler for helpful comments and discussions. These data were presented, in part, at the 2006 World Transplant Congress and the 2007 American Transplant Congress.
- 15Skin cancer in heart transplant recipients: Risk factor analysis and relevance of immunosuppressive therapy. Circulation 2000; 102(19 Suppl 3): III222–III227., , et al.
- 24Ultraviolet light induced injury: Immunological and inflammatory effects. Immunol Cell Biol 2001; 79: 547–568., , .Direct Link:
- 46The role of angiogenesis in tumor growth. Semin Cancer Biol 1992; 3: 65–71..