IL-12 and IL-23 Affect Photocarcinogenesis Differently

IL-12 and IL-23 Affect Photocarcinogenesis Differently

ORIGINAL ARTICLE IL-12 and IL-23 Affect Photocarcinogenesis Differently Christian Jantschitsch1,2, Michael Weichenthal1, Ehrhardt Proksch1, Thomas Sc...

436KB Sizes 1 Downloads 11 Views

ORIGINAL ARTICLE

IL-12 and IL-23 Affect Photocarcinogenesis Differently Christian Jantschitsch1,2, Michael Weichenthal1, Ehrhardt Proksch1, Thomas Schwarz1 and Agatha Schwarz1 Induction of DNA damage by UVR is the key event in photocarcinogenesis. IL-12 and IL-23 are related heterodimeric cytokines consisting of a common p40 unit and a p35/IL-12 and a p19/IL-23 chain, respectively. Both exert immunomodulatory activities but are also found to reduce UVR-induced DNA damage presumably via induction of DNA repair. As both cytokines are also produced in the skin, they may mitigate the risk to develop UVR-induced skin cancer. This appears to be the case as mice lacking p40 were previously shown to be at higher risk for skin tumors upon chronic UVR exposure. As these mice express neither IL-12 nor IL-23, the individual effects of IL-12 or IL-23 could not be evaluated. Thus, mice lacking p35 (IL-12p35/) or p19 (IL-23p19/) were subjected to chronic UVR exposure. The Kaplan–Meier analysis indicated a significantly increased probability of tumor development in IL-23p19/ but not in IL-12p35/ mice. Taken together, in our model, loss of IL-23, but not of IL-12, enhances development of UVR-induced skin tumors, indicating that IL-23 but not IL-12 may counteract photocarcinogenesis. This may have impact on the development of future strategies utilizing antibodies against IL-12 and IL-23, respectively, for the treatment of inflammatory dermatoses. Journal of Investigative Dermatology (2012) 132, 1479–1486; doi:10.1038/jid.2011.469; published online 2 February 2012

INTRODUCTION Solar UVR, in particular the mid-wave range (UVB, 290–320 nm), is the predominant causal factor for skin cancer (de Grujil, 1999). The key event in photocarcinogenesis is the induction of DNA damage. Even low, suberythemogenic UVB doses induce DNA lesions, mostly cyclobutane pyrimidine dimers (CPDs) and (6-4) photoproducts (Patrick, 1977). The vast majority of photolesions is efficiently removed by the nucleotide excision repair (NER) (de Laat et al., 1999). Hence, defects in the NER, as impressively demonstrated by the disease xeroderma pigmentosum, dramatically increase the risk for skin cancer (Kraemer et al., 1994). In turn, accelerated or enhanced reduction of DNA damage, as principally proven by the use of exogenous DNA repair enzymes, reduces the photocarcinogenic risk (Yarosh et al., 2001). Hence, lowering the load of DNA damage either by modulating the NER or by other mechanisms is an essential component of recent sun protection strategies (Verschooten et al., 2006).

1

Department of Dermatology, Christian-Albrechts-University Kiel, Kiel, Germany and 2Division of General Dermatology, Department of Dermatology, Medical University Vienna, Vienna, Austria Correspondence: Thomas Schwarz, Department of Dermatology, University of Kiel, Schittenhelmstrasse 7, D-24105 Kiel, Germany. E-mail: [email protected]kiel.de Abbreviations: CPD, cyclobutane pyrimidine dimer; EGCG, epigallocatechin-3-gallate; MTT, 3-(4,5-dimthylthiazol-2-yl)-2,5diphenyltetrazolium bromide; NER, nucleotide excision repair; SCC, squamous cell carcinoma; WT, wild type

Received 26 July 2011; revised 16 November 2011; accepted 27 November 2011; published online 2 February 2012

& 2012 The Society for Investigative Dermatology

IL-12 is a heterodimeric cytokine consisting of a p35 and a p40 chain. Because of its immunomodulatory activities it is a major player in orchestrating both innate and acquired immune responses (Trinchieri, 1998). One of its major features among many others is its crucial role in the development of T helper 1 responses. In addition, IL-12 exerts the capacity to reduce the amounts of UVR-induced DNA damage (Schwarz et al., 2002). This activity appears to be mediated via the NER as the effect is not observed in NER-deficient mice and cells. This was a very surprising observation as it was the first discovery that the NER can be modulated by cytokines. For quite a long time IL-12 is known to prevent UVR-induced immunosuppression (Mu¨ller et al., 1995; Schmitt et al., 1995; Schwarz et al., 1996). This effect appears to be related to its capacity to reduce UVR-induced DNA damage (Schwarz et al., 2005), which is the major molecular mediator in photoimmunosuppression (Kripke et al., 1992). IL-23 was described several years after the discovery of IL-12. It is closely related to IL-12 as it shares the same p40 unit that associates with a p19 chain (Hunter, 2005). Accordingly, IL-23 shares some biological activities with IL-12, but it also exerts different biological effects (Kastelein et al., 2007). These have been mainly linked to a T helper 17 cell response (Mills, 2008). Recently, we observed that IL-23 exerts similar effects on UVR-induced damage as IL-12 (Majewski et al., 2010). IL-23 reduces the amounts of CPDs both in vitro and in vivo. Studies utilizing Xpa knockout mice indicated that this effect involves the NER. Accordingly, IL-23 prevents the suppression of the induction of contact hypersensitivity by UVR (Majewski et al., 2010). Taken together, IL-23 appears to exert similar effects on UVRinduced DNA damage as IL-12. www.jidonline.org 1479

C Jantschitsch et al. IL-23, IL-12, and Photocarcinogenesis

Since the discovery that IL-12 can reduce UVR-induced DNA damage, speculations exist whether IL-12 prevents photocarcinogenesis through this capacity (Schwarz et al., 2002). This assumption was supported by the fact that IL-12 can be produced in the skin (Aragane et al., 1994; Mu¨ller et al., 1994). Consequently, photocarcinogenesis studies utilizing p40 knockout mice were performed and revealed in fact a higher risk of developing skin cancer when compared with chronically UVR exposed wild-type (WT) mice (Maeda et al., 2006). Hence, it was concluded that IL-12 exerts a mitigating effect on photocarcinogenesis. However, this conclusion has to be re-examined because, as we know today, p40 knockout mice are deficient in not only IL-12 but also IL-23. To examine whether IL-12 and IL-23 exert similar effects on the induction of UVR-mediated skin cancer or differ in this respect, photocarcinogenesis studies were performed with p35 and p19 knockout mice. Only p19-deficient mice revealed a higher probability of developing skin tumors but not p35-deficient mice. The risk of p19-deficient mice was lower when compared with the previous results with p40 knockout mice. This implies that loss of IL-12 seems to be compensated by IL-23, but not vice versa, explaining the slightly enhanced carcinogenesis risk in the absence of IL-23. Loss of both, however, further enhances the risk of UVR-induced skin cancer.

control mice, development of a spontaneous skin tumor was observed. Higher rate of non-epithelial malignant tumors in mice lacking IL-23 or IL-12

Tumors were excised when their volume exceeded 1 cm3 or when they became necrotic. Paraffin sections were stained with hematoxylin and eosin and sections analyzed in a blinded fashion (Supplementary Figure S1 online). As expected, the majority of UVR-induced malignancies in WT mice were of epithelial origin (64%), either squamous cell carcinomas (SCCs, 52%) or basal cell carcinomas (12%; Figure 2 and Table 1a). In contrast, malignant tumors from IL-23p19/ and IL-12p35/ mice were predominantly sarcomas. This phenomenon was more pronounced in IL-23p19/ (77% sarcomas) than in IL-12p35/ animals (64% sarcomas; Figure 2 and Table 1a). As demonstrated in Table 1a, the total number of tumors induced in IL-23p19/ animals was remarkably lower than in WT or IL-12p35/ mice. On the first glance, this may be contradictory to the data shown in Figure 1, but because of the fact that the tumors grew substantially faster than those in WT or IL-12p35/ mice, IL-23p19/ mice had to be killed much earlier before additional tumors could develop. 100%

RESULTS Enhanced photocarcinogenesis in mice lacking IL-23

To address the influence of IL-12 and IL-23 on photocarcinogenesis, mice lacking p35 (IL-12p35/) or p19 (IL-23p19/) and WT mice (C57BL/6) were exposed to a chronic UVB regimen known to induce skin tumors after B200 days (Beissert et al., 2001; Maeda et al., 2006). Unirradiated mice of each group served as controls. Beginning with the first exposures to UVR the entire observation period covered 80 weeks. The Kaplan–Meier analysis revealed that IL-12p35/ as well as IL-23p19/ mice developed UVR-induced skin tumors faster than WT mice (Figure 1a and b). However, this effect was statistically significant only in IL-23p19/ animals. In none of the

wt (n =22) / IL23–/– (n =22)

P = 0.0123

0

100

200

300 400 Days

500

600

36% 23%

0% IL-12p35 –/–

IL-23p19–/–

WT

Figure 2. Distribution of UVR-induced epithelial and non-epithelial malignant tumors. Percent of epithelial (black; squamous cell carcinomas (SCCs) and basal cell carcinomas (BCCs)) and non-epithelial (white; sarcomas) malignant tumors in wild-type (WT), IL-23p19/, and IL-12p35/ mice.

UVB wt (n =22) UVB IL23–/– (n =22)

64%

36%

b

All tumors IL-23 p19 –/– 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

50%

Probability of tumor development

Probability of tumor development

a

77% 64%

All tumors IL-12 p35 –/– 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

UVB wt (n =22) UVB IL12–/– (n =22) wt (n =22) / IL12–/– (n =22)

P = 0.6636

0

100

200

300 400 Days

500

600

Figure 1. Kaplan–Meier analysis of tumors induced by UVR. Mice were exposed thrice weekly for the first 4 weeks with 2.5 kJ m–2, for the next 4 weeks with 5 kJ m–2, and finally for 16 weeks with 10 kJ m–2 of UVB (UVR wild type (WT), UVR IL-23p19/, and UVR IL-12p35/). Controls were untreated mice of each mouse strain used (WT, IL-23p19/, and IL-12p35/). The probability of tumor development was estimated according to Kaplan–Meier analysis. The graph depicts probabilities of tumor development for each of the treatment groups, with ticks indicating censored observations, i.e., mice dying without tumor development before the end of the observation time. Each group consisted of 22 mice. (a) Probability of tumor development in IL-23p19/ versus WT mice. (b) Probability of tumor development in IL-12p35/ versus WT mice. The data for the WT mice are identical in a and b.

1480 Journal of Investigative Dermatology (2012), Volume 132

C Jantschitsch et al. IL-23, IL-12, and Photocarcinogenesis

In the IL-23p19/ group, 71% of tumor-bearing animals had to be killed because of the tumor volume above 1 cm3 or excessive necrosis/ulceration. In contrast, in the IL-12p35/ group the rate was 52%, and in the WT group it was only 18%. In addition, some of the IL-23p19/ mice died prematurely without tumor development, thus contributing to the lower total number of tumors in this group.

Non-epithelial tumors develop significantly earlier in IL-23p19/ mice

Anatomic distribution of tumors

The majority of tumors developed on the back (Table 1b). When stratified by tumor type, it turned out that SCCs and

Table 1a. Histological diagnoses of tumors induced by UVR Tumor types

IL-12p35/ IL-23p19/ Wild type

8 (36.4%)

1 (7.8%)

13 (52%)

0

2 (15.4%)

3 (12%)

Basal cell carcinomas Sarcomas

14 (63.6%)

10 (76.9%)

9 (36%)

Total

22 (100%)

13 (100%)

25 (100%)

0

1 (100%)

4 (100%)

4 (80%)

0

0

Benign tumors Papillomas Dermatofibromas Trichoepitheliomas

1 (20%)

0

0

Total

5 (100%)

1 (100%)

4 (100%)

Table 1b. Anatomic location of tumors induced by UVR Anatomic location Ear Back Eye

IL-12p35/

IL-23p19/

Wild type

7 (25.9%)

3 (21.4%)

12 (41.4%)

15 (55.6%)

10 (71.4%)

15 (51.7%)

5 (18.5%)

1 (7.1%)

1 (3.4%)

Face

0

0

1 (3.4%)

Total

27 (100%)

14 (100%)

29 (100%)

Separate Kaplan–Meier analysis for UVR-induced epithelial (SCCs and basal cell carcinomas) and non-epithelial malignant tumors revealed that occurrence of epithelial malignant tumors was accelerated neither in IL-12p35/ nor in IL-23p19/ mice (Figure 3a and b). In contrast, non-epithelial malignant skin tumors (sarcomas) developed significantly earlier in IL-23p19/ than in WT animals (Figure 3c and d). Higher proliferative capacity of malignant tumors in mice lacking IL-12 or IL-23

Malignant tumors Squamous cell carcinomas

basal cell carcinomas preferentially developed on the ears, whereas sarcomas were observed mainly on the back (Table 1c). On the ears, where 77.3% of SCCs but only 3% of sarcomas developed (Table 1c), the number of tumors was substantially lower in both IL-12p35/ and IL-23p19/ mice in comparison with WT mice (Table 1a and b).

As soon as UVR-induced tumors had developed, their volume was recorded at least once weekly and the average growth rate in mm3 per week was calculated from the pooled tumors of each group. There was a trend of increased growth of tumors in IL-12p35/ and IL-23p19/ animals compared with WT mice (Figure 4a). This was mostly because of the growth behavior of non-epithelial malignant tumors (sarcomas; Figure 4b), whereas lack of IL-12 or IL-23 did not influence the growth rate of epithelial tumors (Figure 4c). These results were confirmed by evaluating the expression of Ki-67. Immunohistochemical stainings were evaluated in a blinded fashion and tumors classified as low (0–10%), medium (11–30%), and high (430%) Ki-67-positive cells. The fraction of high Ki-67-positive ( ¼ highly proliferating) tumors was higher in IL-12p35/ and IL-23p19/ animals than in WT mice (Figure 4d). This difference was much more pronounced when sarcomas were analyzed separately (data not shown). To study the in vitro growth behavior, cell lines were established from SCCs and sarcomas and cultivated. As in the IL-23p19/ group only a single SCC developed (Table 1a), the in vitro proliferation rate was analyzed for non-epithelial malignant tumors (sarcomas) only. A total of 16 sarcoma cell lines were subjected to 3-(4,5-dimthylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) proliferation assays, which were performed immediately (0 hours) and 72 hours

Table 1c. Correlation of tumor types with anatomic location Anatomic location

SCC

BCC

17 (77.3%)

3 (60%)

Back

4 (18.2%)

1 (20%)

Eye

1 (4.5%)

0

Face

0

1 (20%)

0

0

0

0

Total

22 (100%)

5 (100%)

33 (100%)

5 (100%)

4 (100%)

1 (100%)

Ear

Sarcoma

Papilloma

Dermatofibroma

Trichoepithelioma

0

0

1 (100%)

26 (78.8%)

5 (100%)

4 (100%)

0

6 (18.2%)

0

0

0

1 (3%)

Abbreviations: BBC, basal cell carcinoma; SCC, squamous cell carcinoma.

www.jidonline.org 1481

C Jantschitsch et al. IL-23, IL-12, and Photocarcinogenesis

UVB wt (n =22) UVB IL23 –/– (n=22) wt (n =22) / IL23–/– (n=22)

P = 0.4839

c

100

200

300 Days

400

500

600

UVB wt (n=22) UVB IL23 –/– (n=22) wt (n=22) / IL23–/– (n=22)

P = 0.0265

0

100

200

300

400

500

Epithelial malignant tumors IL-12 p35 –/– 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

UVB wt (n=22) UVB IL12 –/– (n =22) wt (n=22) / IL12 –/– (n=22)

P = 0.2182

0

d

Non-epithelial malignant tumors IL-23 p19 –/– 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

Probability of tumor development

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0

Probability of tumor development

b

Epithelial malignant tumors IL-23 p19 –/–

Probability of tumor development

Probability of tumor development

a

600

Days

100

200

300 Days

400

500

600

Non-epithelial malignant tumors IL-12 p35 –/– 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

UVB wt (n =22) UVB IL12–/– (n=22) wt (n =22) / IL23 –/– (n =22)

P = 0.2741

0

100

200

300

400

500

600

Days

Figure 3. Kaplan–Meier analysis of epithelial tumors and sarcomas induced by UVR. Data displayed in Figure 1 were reanalyzed and the probability of developing epithelial malignant tumors (a, b) and sarcomas calculated separately (c, d). The graph depicts probabilities of tumor development for each of the treatment groups, with ticks indicating censored observations, i.e., mice dying without tumor development before the end of the observation time. Panels a and c show the data for IL-23p19/ versus WT mice, and b and d for IL-12p35/ versus WT mice. The data for the WT mice are identical in a and b and c and d, respectively.

after seeding. The absorption values at 0 hours were set as 1 and the increase of proliferation after 72 hours was calculated. Cell lines derived from sarcomas occurring either in IL-12p35/ (7 cell lines) or in IL-23p19/ (6 cell lines) revealed a slightly increased proliferation rate. The differences, however, were not significant when compared with the proliferation of tumor cell lines generated from WT mice (3 cell lines; Figure 4e). This correlated with the findings obtained in the colony formation assays. Both the numbers and sizes of colonies were higher in cell lines derived from either IL-12p35/ or IL-23/p19/ mice in comparison with tumor cells obtained from WT mice (Figure 4f). DISCUSSION The discovery that IL-12 exerts the capacity to reduce UVRinduced DNA damage via modulating the NER was quite surprising and unexpected, as it was assumed that the NER as a kind of emergency system is not subjected to any regulation and not susceptible to external influences. Soon thereafter, it was reported that this activity is not confined to IL-12 but applies also to other mediators, including IL-18 and a-melanocyte-stimulating hormone (Bo¨hm et al., 2005; Schwarz et al., 2006). Because of the structural similarities, IL-23 shares some of the biological activities with IL-12. Hence, it was not surprising that IL-23 exerts similar effects on UVR-induced DNA 1482 Journal of Investigative Dermatology (2012), Volume 132

damage (Majewski et al., 2010). As UVR-induced DNA damage is the major molecular mediator of UVR-induced immunosuppression (Kripke et al., 1992), it was predicted that any cytokine that reduces DNA lesions should prevent UVR-induced immunosuppression. This in fact applies for IL-12 and IL-23; injection of either cytokine prevents the UVR-mediated suppression of sensitization (Schwarz et al., 1996; Majewski et al., 2010). Additional speculations about the biological relevance of these unique features of IL-12 and IL-23 extended to carcinogenesis. Both cytokines can be produced in the skin (Aragane et al., 1994; Mu¨ller et al., 1994; Takenaka et al., 2011). Hence, one could surmise that a constant release of IL-12 or IL-23 could reduce the risk of photocarcinogenesis by reducing UVR-induced DNA damage. Accordingly, this question was addressed by utilizing p40 knockout mice. They revealed a significantly enhanced risk of developing skin tumors upon chronic UVR exposure (Maeda et al., 2006). These findings were interpreted that IL-12 can in fact reduce photocarcinogenesis. The option that IL-23 might as well contribute in this system was not considered. Shortly thereafter, p35 knockout mice were found to be more prone to skin cancer upon chronic UVR exposure in comparison with WT mice (Meeran et al., 2006). This study also indicated a link to NER. Basically, this paper studied the effect of

C Jantschitsch et al. IL-23, IL-12, and Photocarcinogenesis

a

b

All tumors

Non-epithelial tumors 600

Increase of tumor volume per week in mm3

Increase of tumor volume per week in mm3

600

NS

400 NS 200

0

*

200

0 IL-12p35 –/– IL-23p19 –/–

IL-12p35 –/– IL-23p19 –/–

WT

c

d Epithelial tumors

600

Percent of malignant tumors

Increase of tumor volume per week in mm3

400

NS

400

200 NS

NS

0 IL-12p35 –/– IL-23p19 –/–

WT

100% 40.9 high

24 46.2 32

50%

27.3 med

30.8

31.8 low

23

44

0% IL-12p35 –/– IL-23p19 –/–

e

WT

WT

f Increase of proliferation

4

NS

3 NS 2

1

0 IL-12p35 –/–

IL-23p19 –/–

WT

IL-12p35 –/– IL-23p19 –/–

WT

Figure 4. Proliferation of UVR-induced malignant tumors in vivo and in vitro. Tumor volumes were recorded at least once weekly and the average growth rate in mm3 per week was calculated. UVR-induced malignant tumors were pooled according to the mouse strain in which they developed (wild type (WT), IL-23p19/, and IL-12p35/). (a) The volumes of all tumors, (b) and of non-epithelial and (c) epithelial tumors are shown. Error bars depict SD. NS, not significant versus WT, *P ¼ 0.032 versus WT. (d) Immunohistochemistry of malignant tumor samples was carried out using an antibody against Ki-67. The fraction of proliferating cells, resembling Ki-67-positive cells in the tumor, was estimated. The frequency of 0–10% Ki-67-positive cells was defined as low (white), 11–30% as medium (med, gray) and over 30% as high (black) proliferative capacity. (e) Cells were obtained from 16 surgically removed sarcomas of WT (3 tumors), IL-23p19/ (7 tumors), and IL-12p35/ (6 tumors) mice. Proliferation was measured using 3-(4,5-dimthylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) proliferation assay immediately (0 hours) and 72 hours after onset of culture. The values at 0 hours were set as 1 and the increase of proliferation after 72 hours was calculated. For each tumor cell line, MTT assays were at least performed twice independently. Data are expressed as mean±SEM. NS versus WT. (f) To determine the colony-forming efficiency, tumor cell lines (sarcomas) were seeded at low concentrations into Petri dishes and the outgrowth of colonies was analyzed after 14 days by visualizing the colonies with crystal violet. One representative of four independently performed experiments is shown.

epigallocatechin-3-gallate (EGCG) on skin cancer. The authors showed that EGCG reduced the frequency of UVRinduced skin cancer. This effect was not observed in p35 knockout mice, indicating that EGCG acts via IL-12, which was confirmed shortly thereafter by other studies (Schwarz et al., 2008). In addition, reduction of CPDs by EGCG was not only reduced in p35 knockout mice but also in Xpa knockout mice, suggesting that EGCG acts via IL-12 that ultimately modulates the NER, finally resulting in reduction of CPDs (Meeran et al., 2006). As neither of these studies allows to differentiate the effects of IL-12 or IL-23 on photocarcinogenesis independently, this

study was performed comparing IL-23p19/ versus IL-12p35/ mice. The Kaplan–Meier analysis indicated that both knockout strains revealed a higher risk of developing skin tumors when compared with WT mice. However, in IL-12p35/ mice, this trend was minimally pronounced and statistically not significant. All mice that made it to the end of the observation period developed tumors. We selected Kaplan–Meier estimates for calculating the probability of tumor development. This kind of analysis makes maximum use of the information by keeping all animals in the analysis. Our findings obtained with the IL-12p35/ mice are in contrast to the findings by Meeran et al. (2006) who did see www.jidonline.org 1483

C Jantschitsch et al. IL-23, IL-12, and Photocarcinogenesis

an increased risk. The reason for these differences remains to be determined but may be because of the different mouse strains used (C3H/HeN vs. C57BL/6) or because of different irradiation protocols. Retrospective comparison of the data obtained with the IL-23p19/ mice with the previously published findings observed in p40/ mice (Maeda et al., 2006) revealed that p40/ mice apparently are at a higher risk of developing skin cancer. Although a retrospective juxtaposition of data has limitations, it appears possible that the loss of either IL12 or IL-23 can be at least partially compensated by the fellow cytokine. Nevertheless, IL-23 seems to play a more dominant role in protecting from photocarcinogenesis than IL-12. Loss of both cytokines, however, appears to have the maximum impact as shown in the p40/ mice (Maeda et al., 2006). Another striking feature of this study is the shift in the types of tumors induced by UVR in the absence of IL-12 or IL-23. In WT mice, the vast majority were epithelial carcinomas. Surprisingly, this ratio was shifted in favor of sarcomas in the knockout mice. In addition, the more aggressive growth behavior of the tumors in the knockout mice and the increased proliferation rate of cell lines established from these tumors were mostly because of the sarcomas. The increased frequency of sarcomas upon chronic UVR exposure was surprising and similar findings are not reported in the literature. However, one has to be aware that careful histological analysis including histochemical characterization in most of these studies was not performed. Hence, it would be interesting to study retrospectively whether those tumors described as dedifferentiated carcinomas were in fact exclusively carcinomas or whether sarcomas were also among those neoplasms. Accordingly, we tried to reanalyze samples of the previous study (Maeda et al., 2006). Although we did not have access to all samples, we were able to obtain 23 tumors that could be allocated to either p40/ or WT mice. After a careful reanalysis in a blinded fashion by our pathologist, 56% of the tumors obtained from p40/ mice had to be reclassified as sarcomas. In contrast, no sarcoma was found in the WT group. Hence, it appears that the vast majority of tumors described as dedifferentiated carcinomas in our previous paper were in fact sarcomas. Although limited by the fact that we could not reanalyze all tumors of the previous study, the percentage of sarcomas in the p40/ group is in the same range as in this study. Whether the observed sarcomas are truly of dermal origin or the product of epithelial–mesenchymal transition cannot be deducted from our data. It has been described that EGF and transforming growth factor-b may induce epithelial– mesenchymal transition in breast cancer (Ackland et al., 2003; Wendt et al., 2010). The fact that sarcomas developed preferentially on the back and not on the ears gives rise to the speculation that topographic differences in the structure of the skin may be of relevance. This may include the fact that the dermis of the ears is much thinner and consequently much fewer fibroblasts per surface area exposed to UVR. In case of epithelial–mesenchymal transition being the driving force, different environmental conditions on the back and on the ears might influence this process as well. As the rate of 1484 Journal of Investigative Dermatology (2012), Volume 132

sarcomas was substantially increased in the absence of IL-12 or IL-23, these cytokines might directly or indirectly inhibit epithelial–mesenchymal transition. With regard to carcinomas in murine skin, it has been shown that IL-23 promotes tumor formation via favoring a proinflammatory environment and reducing the amount of infiltrating CD8 þ cytotoxic T cells (Langowski et al., 2006). The authors demonstrated that depletion of IL-23 inhibited tumor growth of transplanted carcinomas. Their data are in accordance with ours as we also observed a decreased rate of carcinomas in IL-23p19/ mice. IL-12 and IL-23 can be expressed in human skin (Mu¨ller et al., 1994; Yawalkar et al., 2009). In addition, the reducing effect of IL-12 on CPDs has been confirmed in human cells (Schwarz et al., 2002). Whether this applies for IL-23 also remains to be determined, but in our view is highly likely. IL-12 and IL-23 are attractive candidates for immune intervention. Antibodies blocking the p40 unit are approved for the therapy of psoriasis and have turned out to be highly effective (Papp et al., 2008). Initially, IL-12 was the primary target cytokine, but with time it turned out that IL-23 may be even more important in the pathogenesis of psoriasis, as increased p19 mRNA but not p35 mRNA was found in lesional skin compared with nonlesional skin of patients with psoriasis (Lee et al., 2004). Hence, antibodies blocking the p19 unit are developed and will be tested in clinical studies soon. Ustekinumab, which blocks p40, is well tolerated. The most frequent side effects include microbial infections. Whether long-term therapy with ustekinumab increases the risk of skin cancer can only be answered by long-term followup studies. As IL-12 and IL-23 appear to affect the NER, the question is obvious of whether blocking either of these cytokines increases the amounts of UVR-induced DNA damage and in the long term enhances the risk of skin cancer. The primary targets of the anti-IL-12 therapy are definitely T lymphocytes and not keratinocytes. It is unclear whether relevant amounts of these antibodies that are administered systemically reach the skin except the area of injection. Hence, it is thoroughly possible that these antibodies acting in a systemic fashion do not affect keratinocytes and fibroblasts, respectively. Nevertheless, this could be easily clarified by comparing the amounts of CPDs in biopsies of UVR-exposed skin before and during antibody therapy. Unless such studies have been performed, consequent UVR protection is recommended for patients undergoing therapy with biologics that target either IL-12 or IL-23. In addition, combination with phototherapy has to be discouraged. MATERIALS AND METHODS Mice Female C57BL/6J mice were purchased from Charles River Laboratories (Sulzfeld, Germany). Female IL12atm1Jm/J mice were obtained from Jackson Laboratories (Bar Harbor, ME) via Charles River Laboratories. These mice (C57BL/6J background) lack the p35 chain of IL-12 and hence do not produce functional IL-12 (Mattner et al., 1996). Female IL23a mice (C57BL/6J background) were obtained from Genentech (San Francisco, CA). IL23a mice (C57BL/6J background) lack the p19 chain of IL-23 and therefore have no

C Jantschitsch et al. IL-23, IL-12, and Photocarcinogenesis

functional IL-23 (Ghilardi et al., 2004). All mice were 8 weeks of age when the irradiation protocol was started. Animal care was provided by expert personnel in compliance with the relevant laws and institutional guidelines.

Irradiation and tumor induction UVR exposures were performed using a bank of Philips TL12 fluorescent lamps (Philips, Eindhoven, The Netherlands), which emit most of their energy within the UVB range (290–320 nm) with an emission peak at 313 nm. Mice were placed 20 cm below the light bulbs for irradiation. The mice (22 in each group) were shaved with electric clippers on the entire dorsum once weekly. Mice were irradiated thrice weekly for the first 4 weeks with 2.5 kJ m–2, for the next 4 weeks with 5 kJ m–2, and finally for 16 weeks with 10 kJ/m–2 (cumulative dose: 570 kJ m–2). Untreated mice of each strain served as controls. Tumor development was monitored twice weekly during the irradiation period and thereafter for 56 weeks. The location and growth of each tumor exceeding 2 mm in diameter were recorded and photographically documented. The average growth rate in mm3 per week was calculated from all tumors pooled of each group. Mice were killed by cervical dislocation when the volume of a tumor exceeded 1 cm3 or the tumor was necrotic. Tumors were removed surgically and one part was fixed in 4% formalin and embedded in paraffin; the other part was used to establish tumor cell lines (see below). Sections were stained with hematoxylin and eosin and were diagnosed in a blinded fashion by a veterinary pathologist (Dr Frantisek Jelinek, AnLab, Prague, Czech Republic), a dermatopathologist (EP), and a pathologist specialized in soft tissue tumors (Professor Ivo Leuschner, Institute for Pathology, University Clinics of Schleswig-Holstein, Kiel, Germany). For evaluation of the degree of proliferative activity, sections were stained with an antibody directed against Ki-67 (clone SP6, Neomarker via Thermo Fisher Scientific, Fremont, CA) using a standard immunoperoxidase technique. The amount of Ki-67-positive cells was estimated in a blinded fashion and tumors were classified as low (0–10%), medium (11–30%), and highly positive (431% Ki-67 positive cells).

Establishment of tumor cell lines Tumors were surgically removed and rinsed twice with phosphatebuffered saline supplemented with 1% antibiotic/antimycotic solution (PAA Laboratories, Linz, Austria). Small pieces (o2 mm3) were seeded in 25 cm2 cell culture flasks containing RPMI-1640 supplemented with 20% fetal calf serum, 1% glutamine and 1% nonessential amino acids, and 1% HEPES buffer solution (all from PAA Laboratories). All cells were maintained at 37 1C in a humidified atmosphere containing 5% CO2. When the cultures had reached B80% confluence, the adherent cells were detached with 0.1% trypsin/0.05% EDTA (PAA Laboratories) and used for subsequent passage. A total of 28 tumors were utilized to establish tumor cell lines. In all, 22 tumor cell lines could be established, which corresponds to a 79% success rate. The outgrowth rate was higher in sarcomas (84%) than in epithelial tumors (67%).

Proliferation assay The proliferative capacity of tumor cell lines was assessed using the tetrazolium salt (MTT) proliferation assay (Sigma-Aldrich, St Louis, MO). The MTT method involves the conversion of MTT to colored formazan by cells and serves as an indirect measurement of cell

growth. Tumor cells were seeded in triplicates into 96-well plates at a density of 2.5  103 cells per well. At 0 hours and 72 hours after seeding, MTT solution (25 ml; 5 mg ml–1 in phosphate-buffered saline) was added to each well followed by incubation at 37 1C. After 4 hours, SDS-formamide buffer was added to dissolve the dark formazan crystals (10% SDS, 50% formamide, pH 4.7, 100 ml per well). After overnight incubation at 37 1C, spectrophotometric absorbance of each sample was measured at 595 nm using a microplate reader (Bio-Rad, Hercules, CA). The values at 0 hours were set as 1 and the increase of proliferation after 72 hours was calculated. The mean values of the MTT assays of 16 different sarcoma lines (7 lines from IL-23p19/ mice, 6 from IL-12p35/ mice, and 3 from WT mice) were determined in 32 independent experiments (each cell line was tested at least twice). To determine the colony-forming efficiency, the tumor cell lines were seeded at low concentrations (350 cells cm–2) into 6 cm Petri dishes. The outgrowth of colonies was analyzed 14 days after seeding by visualizing the colonies with crystal violet.

Statistical analysis The method of Kaplan and Meier was used to describe the probability of development of a malignant tumor (tumor-free survival) in the carcinogenesis study (Figures 1 and 3). This is a life-table analysis and takes into account animals that die before developing a malignant tumor. The differences in tumor latent periods were analyzed by the log-rank test. The differences in Figure 4d (proportions of highly proliferating tumors) were analyzed using w2 test. All other differences were analyzed using Student’s t-test. Differences were considered significant at Po0.05. CONFLICT OF INTEREST The authors state no conflict of interest.

ACKNOWLEDGMENTS This study was supported by grants from the German Research Foundation (SCHW1177/1-3, SCHW625/4-1) given to AS and TS. We are grateful to Susanne Dentel, Nadine Tu¨xen, and Martina Wedler for excellent technical assistance. We thank Professor Ivo Leuschner (Institute for Pathology, University Clinics of Schleswig-Holstein, Kiel, Germany) for help in evaluation of the histological slides. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid

REFERENCES Ackland ML, Newgreen DF, Fridman M et al. (2003) Epidermal growth factorinduced epithelio-mesenchymal transition in human breast carcinoma cells. Lab Invest 83:435–48 Aragane Y, Riemann H, Bhardwaj RS et al. (1994) IL-12 is expressed and released by human keratinocytes and epidermoid carcinoma cell lines. J Immunol 153:5366–72 Beissert S, Ru¨hlemann D, Mohammad T et al. (2001) IL-12 prevents the inhibitory effects of cis-urocanic acid on tumor antigen presentation by Langerhans cells: implications for photocarcinogenesis. J Immunol 167:6232–8 Bo¨hm M, Wolff I, Scholzen TE et al. (2005) Alpha-Melanocyte-stimulating hormone protects from ultraviolet radiation-induced apoptosis and DNA damage. J Biol Chem 280:5795–802 de Grujil FR (1999) Skin cancer and solar UV radiation. Eur J Cancer 35:2003–9

www.jidonline.org 1485

C Jantschitsch et al. IL-23, IL-12, and Photocarcinogenesis

de Laat WL, Jaspers NG, Hoeijmakers JH (1999) Molecular mechanism of nucleotide excision repair. Genes Dev 13:768–85

patients with psoriasis: 52-week results from a randomised, doubleblind, placebo-controlled trial (PHOENIX 2). Lancet 371:1675–84

Ghilardi N, Kljavin N, Chen Q et al. (2004) Compromised humoral and delayed-type hypersensitivity responses in IL-23-deficient mice. J Immunol 172:2827–33

Patrick MH (1977) Studies on thymine-derived UV photoproducts in DNA I. Formation and biological role of pyrimidine adducts in DNA. Photochem Photobiol 25:357–72

Hunter CA (2005) New IL-12-family members: IL-23 and IL-27, cytokines with divergent functions. Nat Rev Immunol 5:521–31

Schmitt DA, Owen-Schaub L, Ullrich SE (1995) Effect of IL-12 on immune suppression and suppressor cell induction by ultraviolet radiation. J Immunol 154:5114–20

Kastelein RA, Hunter CA, Cua DJ (2007) Discovery and biology of IL-23 and IL-27: related but functionally distinct regulators of inflammation. Annu Rev Immunol 25:221–42 Kraemer KH, Levy DD, Parris CN et al. (1994) Xeroderma pigmentosum and related disorders: examining the linkage between defective DNA repair and cancer. J Invest Dermatol 103:96S–101S Kripke ML, Cox PA, Alas LG et al. (1992) Pyrimidine dimers in DNA initiate systemic immunosuppression in UV-irradiated mice. Proc Natl Acad Sci USA 89:7516–20 Langowski JL, Zhang X, Wu L et al. (2006) IL-23 promotes tumour incidence and growth. Nature 442:461–5 Lee E, Trepicchio WL, Oestreicher JL et al. (2004) Increased expression of interleukin 23 p19 and p40 in lesional skin of patients with psoriasis vulgaris. J Exp Med 199:125–30

Schwarz A, Grabbe S, Aragane Y et al. (1996) Interleukin-12 prevents UVB-induced local immunosuppression and overcomes UVB-induced tolerance. J Invest Dermatol 106:1187–91 Schwarz A, Maeda A, Gan D et al. (2008) Green tea phenol extracts reduce UVB-induced DNA damage in human cells via interleukin-12. Photochem Photobiol 84:350–5 Schwarz A, Maeda A, Kernebeck K et al. (2005) Prevention of UV radiationinduced immunosuppression by IL-12 is dependent on DNA repair. J Exp Med 201:173–9 Schwarz A, Maeda A, Sta¨nder S et al. (2006) IL-18 reduces ultraviolet radiation-induced DNA damage and thereby affects photoimmunosuppression. J Immunol 176:2896–901

Maeda A, Schneider SW, Kojima M et al. (2006) Enhanced photocarcinogenesis in interleukin-12-deficient mice. Cancer Res 66:2962–9

Schwarz A, Sta¨nder A, Berneburg M et al. (2002) Interleukin-12 suppresses ultraviolet radiation-induced apoptosis by inducing DNA repair. Nat Cell Biol 4:26–31

Majewski S, Jantschitsch C, Maeda A et al. (2010) IL-23 antagonizes UVRinduced immunosuppression through two mechanisms: reduction of UVR-Induced DNA damage and inhibition of UVR-induced regulatory T cells. J Invest Dermatol 130:554–62

Takenaka N, Edamatsu H, Suzuki N et al. (2011) Overexpression of phospholipase Ce in keratinocytes upregulates cytokine expression and causes dermatitis with acanthosis and T-cell infiltration. Eur J Immunol 41:202–13

Mattner F, Magram J, Ferrante J et al. (1996) Genetically resistant mice lacking interleukin-12 are susceptible to infection with Leishmania major and mount a polarized Th2 cell response. Eur J Immunol 26:1553–9

Trinchieri G (1998) Proinflammatory and immunoregulatory functions of interleukin-12. Int Rev Immunol 16:365–96

Meeran SM, Mantena SK, Meleth S et al. (2006) Interleukin-12-deficient mice are at greater risk of UV radiation-induced skin tumors and malignant transformation of papillomas to carcinomas. Mol Cancer Ther 5:825–32 Mills KH (2008) Induction, function and regulation of IL-17-producing T cells. Eur J Immunol 38:2636–49 Mu¨ller G, Saloga J, Germann T et al. (1994) Identification and induction of human keratinocyte-derived IL-12. J Clin Invest 94:1799–805 Mu¨ller G, Saloga J, Germann T et al. (1995) IL-12 as mediator and adjuvant for the induction of contact sensitivity in vivo. J Immunol 155:4661–8 Papp KA, Langley RG, Lebwohl M et al. (2008) Efficacy and safety of ustekinumab, a human interleukin-12/23 monoclonal antibody, in

1486 Journal of Investigative Dermatology (2012), Volume 132

Verschooten L, Claerhout S, Van Laethem A et al. (2006) New strategies of photoprotection. Photochem Photobiol 82:1016–23 Wendt MK, Smith JA, Schiemann WP (2010) Transforming growth factor-binduced epithelial-mesenchymal transition facilitates epidermal growth factor-dependent breast cancer progression. Oncogene 29:6485–98 Yarosh D, Klein J, O’Connor A et al. (2001) Effect of topically applied T4 endonuclease V in liposomes on skin cancer in xeroderma pigmentosum: a randomised study. Xeroderma Pigmentosum Study Group. Lancet 357:926–9 Yawalkar N, Tscharner GG, Hunger RE et al. (2009) Increased expression of IL-12p70 and IL-23 by multiple dendritic cell and macrophage subsets in plaque psoriasis. J Dermatol Sci 54:99–105