- Email: [email protected]
The Alternative Complement Pathway Seems to be a UVA Sensor that Leads to Systemic Immunosuppression
ORIGINAL ARTICLE
The Alternative Complement Pathway Seems to be a UVA Sensor that Leads to Systemic Immunosuppression Michael P.F. Stapelberg1, Rohan B.H. Williams2, Scott N. Byrne1 and Gary M. Halliday1 UV wavebands in sunlight are immunomodulatory. About half the amount of UVA within a minimum erythemal dose of sunlight is systemically immunosuppressive, whereas higher doses protect from UVB immunosuppression in mice. We have earlier shown that these responses to UVA are genetically restricted, as they occur in C57BL/6 but not in Balb/c mice. We used gene set enrichment analysis of microarray data and real-time reverse transcriptase (RT)-PCR confirmation to determine the molecular mechanisms associated with UVA immunomodulation. We found upregulation of mRNA for the alternative complement pathway. The core-enriched genes complement component 3, properdin, and complement factor B were all activated by the immunosuppressive dose of UVA only in UVA-responsive C57BL/6 but not in unresponsive BALB/c mice. This therefore matched the genetic restriction and dose responsiveness of UVA immunosuppression. The immuneprotective higher UVA dose prevented UVB from downregulating chemokine receptor 7 and IL-12B, and decreased IL-10, supporting the earlier identification of IL-12 and IL-10 in high-dose UVA protection from UVB immunosuppression. Our study has identified activation of the alternative complement pathway as a trigger of UVA-induced systemic immunosuppression and suggests that this pathway is likely to be an important sensor of UVA-induced damage to the skin. Journal of Investigative Dermatology (2009) 129, 2694–2701; doi:10.1038/jid.2009.128; published online 11 June 2009
INTRODUCTION The UV wavebands contained in sunlight are the main cause of skin cancer. Gene mutations and immunosuppression are both important biological consequences of UV exposure that lead to skin cancer (Halliday, 2005). UV wavebands in sunlight are composed of approximately 5% UVB (290–320 nm) and 95% UVA (320–400 nm), both of which can cause gene mutations (Agar et al., 2004) and immunosuppression (Poon et al., 2005) in humans. UVA-induced molecular triggers for immunosuppression have received little research attention. Unlike UVB, which has a linear dose–response relationship with immunosuppres1
Discipline of Dermatology, Bosch Institute, Sydney Cancer Centre, at the University of Sydney, Sydney, New South Wales, Australia and 2The John Curtin School of Medical Research, Australian National University Canberra, ACT, and School of Biotechnology and Biomolecular Sciences, University of New South Wales, Sydney, New South Wales, Australia Work performed at the University of Sydney, Sydney, New South Wales, Australia Correspondence: Professor Gary M. Halliday, Dermatology Research Laboratories, Blackburn Building D06, University of Sydney, New South Wales 2006 Australia. E-mail: [email protected] Abbreviations: C1QA, complement component 1, q subcomponent, alpha polypeptide; C3, complement component 3; CCR7, chemokine receptor 7; CFB, complement factor B; CHS, contact hypersensitivity; FDR, false discovery rate; GSEA, gene set enrichment analysis; RT, reverse transcriptase; ssUV, solar-simulated ultraviolet radiation. Received 17 December 2008; revised 17 February 2009; accepted 21 March 2009; published online 11 June 2009
2694 Journal of Investigative Dermatology (2009), Volume 129
sion, UVA causes a complex bell-shaped dose–response curve. Therefore, the molecular signals initiated by UVA that lead to immunosuppression are also likely to have a nonlinear dose response. We have earlier shown that three consecutive exposures of C57BL/6 mice to 1,700 mJ cm2 UVA causes systemic immunosuppression on sensitization at a distal non-irradiated site, whereas twice this dose does not induce immunosuppression. These dose-related effects of UVA are genetically restricted, as they occurred in C57BL/6 but not in Balb/c mice. Furthermore, exposure of C57BL/6 mice to solar-simulated UV (ssUV) consisting of an immunosuppressive UVB dose combined with the higher nonsuppressive dose of UVA does not cause immunosuppression (Byrne et al., 2002). This is consistent with studies that highdose UVA protects mice from UVB immunosuppression (Reeve et al., 1998). Although UVA does not contribute substantially to the erythemal or oedemal components of sunburn, the immunosuppressive dose of UVA in mice and humans (Poon et al., 2005) is about half of what would be delivered to pale skinned Caucasians receiving a dose of sunlight sufficient to cause barely detectable sunburn (minimum erythemal dose). Therefore, these immunosuppressive UVA doses are within the range received by humans during normal daily activities. UVA must therefore affect different molecular mechanisms at different doses to have this complex affect on immunity. As there are no known UVA-induced molecular responses to explain immunosuppression, we took a global approach of & 2009 The Society for Investigative Dermatology
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
microarray analysis and validation with quantitative real-time reverse transcriptase (RT)-PCR combined with the power of the genetic association of the UVA dose response in C57BL/6 but not in Balb/c mice to determine the most prominent molecular pathway changes that were associated with UVA immunosuppression. RESULTS All raw microarray data are available at Gene Expression Omnibus (GEO) with accession GSE15618. UVA causes systemic immunosuppression in C57BL/6 but not in Balb/c mice with a bell-shaped dose response
UVA systemically suppressed the induction of primary immunity in C57BL/6 mice at a lower (1,750 mJ cm2) but not higher (3,500 mJ cm2) dose (Figure 1a). The contact hypersensitivity (CHS) results also showed that UVA did not modulate immunity in Balb/c mice over this dose range
(Figure 1b). We further clarified the immunomodulatory effects of UVA by combining the higher dose with an immunosuppressive UVB dose (285 mJ cm2) chosen so that this mix of UVA and UVB was a good mimic of sunlight (ssUV). UVB suppressed the induction of CHS in both mouse strains. However, the CHS response of C57BL/6 mice was unaffected by this dose of UVB when combined with the higher dose of UVA to produce ssUV (Figure 1a). Therefore, this higher dose of UVA protected C57BL/6 mice from the immunosuppressive effects of UVB. In contrast, ssUV suppressed CHS in Balb/c mice (Figure 1b), showing that the higher UVA dose did not protect from UVB in this mouse strain. Therefore, UVA modulates immunity in C57BL/6 but not in Balb/c mice with a complex dose response. It suppressed immunity at the lower but not the higher dose. The higher dose is not immunologically inert, as it protects from UVB-induced suppression. The immunosuppressive, but not the higher UVA dose activates the alternative complement pathway
C57BL/6 (n)17
Unirradiated
(n)13 P<0.0001
Lower UVA Higher UVA
(n)13 n.s
UVB
(n)12 P<0.0001 (n)16 n.s
ssUV 0.0
0.2
0.4
0.6
0.8
1.0
1.2
Normalized change in ear thickness Balb/c (n)13
Unirradiated Lower UVA
(n)13 n.s
Higher UVA
(n)12 n.s
UVB
(n)17 P<0.0001
ssUV
(n) 9 P<0.0001 0.0
0.2
0.4
0.6
0.8
1.0
1.2
Normalized change in ear thickness Figure 1. UVA causes immunosuppression at the lower and not higher dose tested in C57BL/6 but not in Balb/c mice. Contact hypersensitivity (CHS) was assessed in (a) C57BL/6 and (b) Balb/c mice by measuring ear swelling 24 hours after challenge, which was 7 days after sensitization on the abdomen. Mice had been irradiated on the dorsal trunk for 3 consecutive days with 1,750 (lower UVA) or 3,500 (higher UVA) mJ cm2 UVA, or with 285 mJ cm2 UVB or this UVB dose combined with the higher UVA to produce solarsimulated UV (ssUV). Irritant controls (ear swelling of unsensitized mice) were subtracted from sensitized mice. For each CHS experiment the ear swelling for each mouse that was irradiated with UV was normalized to the mean of the control group that did not receive UV irradiation to remove interexperimental variation in magnitude of the control response. Each experiment was repeated thrice and normalized values were pooled. The mean and SEM of normalized values are shown. The total number of mice for each group is shown (n). An unpaired student’s t-test was used for statistical analysis compared with the unirradiated group.
To search for signaling pathways likely to be involved in UVA-induced immunosuppression, we used microarrays with gene set enrichment analysis (GSEA). This analysis searches for co-ordinated changes in whole pathways and therefore is more robust than searching for changes in single genes and also enables a stringent statistical analysis to be carried out. Enriched pathways (Po0.05 and false discovery rate (FDR)o0.05) were identified for irradiated C57BL/6 and Balb/c mice compared with unirradiated control mice. Our search strategy, on the basis of the CHS responses described above was that a pathway likely to be involved in UVA immunosuppression would be altered by the lower, but not the higher, UVA dose in C57BL/6 but not in Balb/c mice. This would match the genetic restriction and dose response. All statistically significant enriched pathways are listed in Table 1. Furthermore, for each enriched pathway listed in Table 1, core-enriched genes were identified using GSEA. Core-enriched genes contribute most to the enrichment score of the pathway. The core-enriched genes chosen for confirmation by real-time PCR are also listed in Table 1. The 50 most up- and downregulated pathways for each UV spectra, dose and mouse strain are shown in supplementary tables 1–12 to provide additional information. In C57BL/6 mice, only the alternative complement pathway (P ¼ 0.001, FDR ¼ 0.035) was significantly upregulated. No pathways were downregulated. Neither this, nor any other pathway was altered by the lower UVA dose in Balb/c mice, and the higher UVA dose did not modulate this pathway in C57BL/6 mice. To confirm the association of the alternative complement pathway with UVA-induced immunosuppression, a selected number of core-enriched genes were examined in a series of different skin samples by quantitative real-time RT-PCR (Figure 2). In C57BL/6 mouse skin irradiated with the lower UVA dose, real-time RT-PCR showed significant upregulation of the alternative complement pathway genes C3, complement factor B (CFB) and properdin. These genes were not altered by the higher UVA dose, UVB, or ssUV (which contained the higher UVA dose). www.jidonline.org 2695
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
Table 1. Enriched pathways for lower and higher UVA, UVB, and ssUV Enriched phenotype1
Strain
Genes2
0.001
Increased by lower UVA
C57BL/6
C33, CFB3, PFC3
0.014
0.001
Increased by higher UVA
C57BL/6
CD803, CD86
Classical complement
0.039
0.001
Increased by higher UVA
C57BL/6
C1QA3, C1QB
Cytokine
0.010
0.001
Decreased by higher UVA
C57BL/6
IL3, IL9
Cytokine
0.007
0.001
Decreased by higher UVA
Balb/c
IL3, IL9
b alanine metabolism
0.005
0.001
Increased by UVB
C57BL/6
ALDH23, ALDH9A13
Propanoate metabolism
0.011
0.001
Increased by UVB
C57BL/6
ALDH23, ALDH9A13
Valine, leucine, and isoleucine degradation
0.012
0.001
Increased by UVB
C57BL/6
ALDH23, ALDH9A13
nkt
0.052
0.003
Decreased by UVB
C57BL/6
CCR73, IFNg3
IL12
0.023
0.005
Increased by ssUV
C57BL/6
IL12B3, IL12RB2
IL10
0.047
0.001
Decreased by ssUV
C57BL/6
IL63, IL103
Enriched pathway Po0.05
FDR q-value
P-value
Alternative complement
0.035
Inflammatory response
ALDH2, aldehyde dehydrogenase 2 family; ALDH9A1, aldehyde dehydrogenase 9 family, member A1; CCR7, chemokine receptor 7; CFB, complement factor B; FDR, false discovery rate; PFC, properdin; RT, reverse transcriptase. 1 Increased or decreased by UV compared with unirradiated control. 2 Core-enriched and associated genes for enriched pathways. 3 Genes selected for real-time RT-PCR.
C3
PFC
CFB Unirradiated
Unirradiated
∗
Lower UVA
Unirradiated
∗
Lower UVA
Higher UVA
UVB
UVB
UVB
ssUV
ssUV
ssUV
0
1
2 3 4 5 Fold change
0
6
1
2 3 4 5 Fold change
CD80
6
0
∗
ALDH2
∗∗
Higher UVA
Lower UVA
∗
Higher UVA
UVB
UVB
UVB
ssUV
ssUV
ssUV
0 1 2 3 4 5 6 7 Fold change
0
1
2 3 4 5 Fold change
ALDH9A1
0
6
IFN-γ
Lower UVA
Lower UVA
Lower UVA
Higher UVA
Higher UVA
UVB
UVB
ssUV
ssUV 3
4
∗∗
UVB ssUV
1
0
2
3
IL-12B
IL-6
Lower UVA
Lower UVA
Higher UVA
Higher UVA
∗∗
ssUV 0
1
2
1
3
4
Fold change
2
3
Fold change IL-10
Unirradiated
UVB
0
Fold change
Fold change
Unirradiated
4
Unirradiated
Higher UVA
2
1 2 3 Fold change
CCR7 Unirradiated
Unirradiated
5
Unirradiated
Lower UVA
Lower UVA Higher UVA
1
1 2 3 4 Fold change
C1QA Unirradiated
Unirradiated
0
∗
Lower UVA
Higher UVA
Higher UVA
Unirradiated
∗∗
Lower UVA Higher UVA
UVB
UVB
ssUV
ssUV 0
1
2
3
Fold change
4
∗ 0
1
2
3
4
5
Fold change
Figure 2. Confirmation of quantitative gene expression of core enriched and associated genes using real-time reverse transcriptase (RT)-PCR. For statistical analysis a Kolmogorov–Smirnoff test first determined that the data obtained from the six biological replicates were normally distributed. An unpaired student’s t-test was then used for statistical analysis * Po0.05, ** Po0.01 compared with the unirradiated control mice (0 UV). Bar graphs represent means þ SEM of six biological replicates for lower and higher UVA, UVB, and ssUV doses. ALDH2, aldehyde dehydrogenase 2 family; ALDH9A1, aldehyde dehydrogenase 9 family, member A1; CCR7, chemokine receptor 7; CFB, complement factor B; PFC, properdin.
2696 Journal of Investigative Dermatology (2009), Volume 129
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
In UVA unresponsive Balb/c mice, none of these genes were altered by low-dose UVA, and only C3 was significantly downregulated by the higher UVA dose (results not presented). Therefore, increased expression of alternative complement pathway genes correlated with the genetic restriction and dose response for UVA-induced immunosuppression. Higher dose of immunoprotective UVA associated with normalization of IL-12B and CCR7, and decreased IL-10
To examine the possible immunomodulatory signals activated by the higher immunoprotective dose of UVA, this dose was compared with unirradiated mice, and the combination of this UVA dose with UVB (ssUV) was contrasted to UVB-irradiated mice. In C57BL/6, but not in Balb/c mice, the inflammatory response (P ¼ 0.001, FDR ¼ 0.014) and classical complement (P ¼ 0.001, FDR ¼ 0.039) pathways were significantly upregulated. The cytokine pathway was significantly downregulated in both mouse strains. As C57BL/6, but not Balb/c mice were modulated over this UVA dose range, this pathway was regarded as not being involved in immunomodulation and was not considered further. To determine the pathways involved in higher dose UVA protection from UVB, relative effects of UVB compared with ssUV is more insightful to determine the UVB effects modulated by addition of higher dose UVA. In C57BL/6 mice, the b alanine metabolism (P ¼ 0.001, FDR ¼ 0.005), propanoate metabolism (P ¼ 0.001, FDR ¼ 0.011), and valine, leucine, and isoleucine degradation (P ¼ 0.001, FDR ¼ 0.012) pathways were significantly upregulated. No pathways were downregulated by UVB using the qo0.05 FDR cut-off. However, the nkt pathway (P ¼ 0.003, FDR ¼ 0.052) was very close to the qo0.05 FDR cut-off and was therefore considered further, as these criteria have earlier been suggested to be worthy of further study (Subramanian et al., 2005). In C57BL/6 mice exposed to ssUV, only the IL-12 pathway was upregulated (P ¼ 0.005, FDR ¼ 0.023) and the IL-10 (P ¼ 0.001, FDR ¼ 0.047) pathway was downregulated (Table 1). These changes were not induced by ssUV in Balb/c mice. After this screen by pathway analysis of microarray data, core-enriched genes identified in these pathways (Table 1) were compared in a different set of skin from irradiated mice by real-time RT-PCR. The higher UVA dose significantly upregulated the inflammatory response pathway CD80 and classical complement pathway C1QA core-enriched genes, confirming the microarray GSEA. Although these genes were not altered in Balb/c mice (data not shown), they were not increased by the same UVA dose when combined with UVB (ssUV), indicating, not surprisingly, that the combined effects of these wavebands differ from the summation of each waveband alone. Therefore, these changes are unlikely to be important for higher dose UVA protection from UVB. In C57BL/6 mice irradiated with UVB, significant downregulation of the nkt pathway core-enriched gene chemokine receptor 7 (CCR7) determined by real-time RT-PCR confirmed the microarray data. The core-enriched genes for the remaining pathways identified by GSEA (aldehyde dehydro-
genase 2 family and aldehyde dehydrogenase 9 family, member A1) were not confirmed by RT-PCR. In C57BL/6 mice irradiated with ssUV, GSEA identified the IL-12 pathway as being increased and the IL-10 pathway as being decreased. Real-time RT-PCR did not confirm the ssUV-induced increase in the IL-12B gene, as it was not significantly different to the unirradiated group. However, in comparison with the UVB dose, the IL-12 pathway gene IL-12B was no longer downregulated. Therefore, the UVA component of ssUV inhibited the downregulation of IL-12B caused by UVB. Furthermore, IL-10 was confirmed to be significantly downregulated by ssUV but not by UVB in C57BL/6 but not in Balb/c mice (data not shown for Balb/c mice). In summary these data show that genes upregulated by the immunoprotective UVA dose alone do not predict differences between the effects of UVB and ssUV. They further indicate that prevention of UVB-induced downregulation of CCR7 and IL-12B along with reduced expression of IL-10 seem to correlate with higher dose UVA protection from UVB-induced immunosuppression. DISCUSSION In this study, we used pathway analysis of microarray data and quantitative real-time RT-PCR to identify molecular mechanisms associated with systemic UVA immunosuppression. The CHS results in this study were consistent with earlier findings (Byrne et al., 2002). UVA was immunosuppressive at a lower, but not higher, dose in C57BL/6 but not in Balb/c mice. The higher UVA was not immunologically inert, but protected from UVB-induced immunosuppression with the same genetic association. We determined that the alternative complement pathway responds to UVA and may be a key trigger of UVA-induced immunosuppression. This is because mRNA for genes in this pathway was found to be upregulated by the immunosuppressive dose of UVA, but not by the higher immunoprotective dose, nor by UVB alone or combined with the higher protective UVA dose. This upregulation of alternative complement occurred in responsive C57BL/6 but not in unresponsive Balb/c mice, and therefore was associated with the genetic restriction, dose and waveband response relationship with UVA immunosuppression. Our data also showed that immunoprotection by the higher UVA dose was allied with the prevention of UVBinduced depletion of CCR7 and IL-12B, and a reduction in IL-10 mRNA. Three pathways lead to complement activation, the classical, lectin, and alternative pathways. All three pathways result in the production of C3 and a subsequent activation of the final lytic pathway and membrane attack complex. The alternative pathway is unique in complement activation, in that it functions as an amplification loop activating itself with its own product, C3b (Schreiber and Muller-Eberhard, 1978; Schreiber et al., 1978). Complement components C3 and CFB are central to the alternative pathway, as after activation they become key constituents of the serine protease C3bBb (Isenman and Cooper, 1981). C3bBb cleaves C3 molecules producing C3b (Schreiber and Muller-Eberhard, 1978; Schreiber et al., 1978). The cleavage of C3 in the alternative pathway is promoted by properdin (Fearon and Austen, www.jidonline.org 2697
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
1975a, b). Properdin protects C3bBb and C3b bi-products including iC3b from degradation, resulting in an accumulation of C3b and iC3b in tissue. The epidermis of normal skin is a rich source of components of the alternative complement pathway including C3 and CFB (Dovezenski et al., 1992). C3associated molecules of the alternative pathway have also been observed in the basement membrane zone of the epidermis (Yoshida et al., 1998). Studies have shown that epidermally derived cytokines including IL-6 enhance the synthesis of alternate pathway C3 and CFB (Pasch et al., 2000; Stapp et al., 2005). In this respect, it is interesting that the suppressive UVA, but not the higher dose, increased IL-6 expression. Therefore, induction of IL-6 could contribute to UVA-induced immunosuppression by augmenting or initiating activation of the alternative complement pathway. It has not been suggested earlier that the alternative complement pathway is involved in systemic immunosuppression caused by UVA exposure of the skin. However, UVB causes deposition of C3 and later components of the membrane attack complex in the skin (Rauterberg et al., 1993), and UVB-induced activation of C3 to produce iC3b has been shown to be important for local immunosuppression (Hammerberg et al., 1998). Whether the iC3b involved in UVB-induced local immunosuppression was a product of the classical, lectin, or alternative pathways was not clear, but iC3b binding to its receptor CD11b on macrophages infiltrating UVB-exposed skin was identified as being important for local UVB-induced immunosuppression (Hammerberg et al., 1998). Neutralization of C3 with soluble complement receptor 1 prevented UVB-induced local immunosuppression and reduced infiltration by CD11b þ macrophages; in addition, immunosuppression did not occur in C3 knockout mice. This study differed to ours in a number of ways. A dose of UVB was used that induced local but not systemic immunosuppression, and both iC3b and infiltrating CD11b þ macrophages were required at the site of antigen exposure. In our study, activation of the alternative complement pathway in UVA-irradiated skin was at a different site to antigen exposure, and therefore interactions between complement components, any inflammatory cells, and antigen could not have occurred. In addition, the UVA irradiation regime used in these studies systemically suppresses primary but not secondary immunity and therefore does not induce tolerance (Byrne et al., 2002). This contrasts with the studies described above that showed a role for C3 using a UVB irradiation dose that caused locally induced tolerance (Hammerberg et al., 1998).Therefore, the mechanism by which the alternative complement pathway leads to systemic UVA immunosuppression is likely to be different to what was described by Hammerberg et al. Nevertheless, our study supports an important regulatory role for complement in responses of the skin immune system to UV radiation. It is unclear how the immunosuppressive, but not the higher, dose of UVA could induce expression of the alternative complement pathway. This pathway has multiple positive and negative regulators. It is possible that the lower UVA dose could activate one of the positive regulators and 2698 Journal of Investigative Dermatology (2009), Volume 129
the higher UVA dose activates one of the inhibitors, or the lower UVA dose could relieve the pathway of one of its inhibitory mechanisms, enabling it to become activated. For example, properdin binding to lipopolysaccharide activates the alternative complement pathway, and the yeast cell wall component zymosan can activate alternative complement through a mechanism dependent on CFB (Kimura et al., 2008). These complement components therefore seem to be pattern recognition molecules. UVA is a potent oxidizer of lipids and other cellular macromolecules (Halliday, 2005), and it is possible that alternative complement components may recognize and be activated by these UVA-altered cell structures. It is unlikely that activation of the membraneattack complex is required for the alternative complement pathway to trigger UVA immunosuppression, as these later steps in the complement pathway, components C4–C9, are part of the classical pathway, which were not found to be involved in UVA immunosuppression in our study. It is more likely that one of the alternative complement proteins acts as a sensor of UVA-induced damage by binding to an unknown cell surface receptor or triggering a sequence of events culminating in systemic immunosuppression. Elucidation of these details will require extensive biochemical and functional characterization. Higher dose UVA protection from UVB immunosuppression was associated with UVA prevention of UVB-induced reduction in CCR7 and IL-12B, and with a reduction in IL-10 expression. Activated dendritic cells, including Langerhans cells, express the CCR7 (Ohl et al., 2004). UV has earlier been shown to impair the induction of CCR7 on dendritic cells during maturation, thus preventing their migration to draining lymph nodes (Mittelbrunn et al., 2005). In these experiments, antigen was applied to normal unirradiated skin and dendritic cells function and numbers in the draining lymph nodes have been found not to be altered during systemic immunosuppression (Byrne and Halliday, 2005; Gorman et al., 2005). It is therefore unlikely that high-dose UVA protection of CCR7 expression in the irradiated skin would prevent systemic immunosuppression. A UVA dose about 10-fold higher than what was used in these studies, that has previously been shown to protect from systemic UVB immunosuppression in Skh:HR-1 hairless mice, has been shown to abrogate UVB-enhanced expression of IL-10, and to increase IL-12 protein production (Shen et al., 1999). This study further showed that IFN-g was initially upregulated by high-dose UVA, which then lead to an increased expression of IL-12. The genome-wide expression analysis in this study, using different mouse strains and a lower UVA dose supports this earlier study by Shen et al. that high-dose UVA inhibition of IL-10 and prevention of IL-12 reduction by UVB are likely to be important mediators of UVA immunoprotection. This is also consistent with earlier studies that UVB-induced production of IL-10 contributes to UV immunosuppression (Nishigori et al., 1996) and that IL-12 can reverse UVB-induced immunosuppression by blocking IL-10 secretion (Schwarz et al., 1996). Macrophages infiltrating UVB-exposed skin migrate to lymph nodes in response to contact sensitization in which they produce
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
IL-10, and UV also reduces IL-12 production by lymph node CD11c þ cells in response to contact sensitization (Toichi et al., 2008). Thus, inflammatory cells and keratinocytes (Shen et al., 1999) show a similar pattern of IL-12 and IL-10 production that regulates responses to UV. We examined skin after exposure to low doses of UVA, which does not cause observable inflammation and therefore the cytokines and complement components were likely to have been produced by resident skin cells. In summary, our studies have identified that the alternative complement pathway seems to be a sensor of UVA-induced damage to the skin, leading to systemic immunosuppression. The immunosuppressive dose of UVA is only half the amount contained in a minimum erythemal dose of sunlight. Therefore, this UVA dose can be achieved by sunlight exposure during normal daily activities and is physiologically relevant for a large proportion of humans. Alternative complement is a complex pathway with multiple proteins that could be activated or inhibited by recognition of UVA-altered molecular structures in the skin and may be responsible for the complex bell-shaped dose response for UVA immunosuppression. This was the only pathway identified by our global genome approach and pathway analysis to be associated with UVA-induced immunosuppression. This approach was limited to detection of events leading to a change in mRNA expression, and therefore other events that we did not detect are likely to be involved. Our studies also confirmed earlier observations that high-dose UVA protection from UVBinduced immunosuppression is mediated, at least in part, by protection of IL-12 and a reduction in IL-10. Complement seems to be an important trigger of UV immunosuppression and requires further research.
measured regularly using a radiometer (Solar Light Company, PA), which had been calibrated against the source with an OL-754 spectroradiometer (Optronics Laboratories, Orlando, FL), and the timing of UV exposure was adjusted using an automated timing device to deliver consistent UV doses. The spectral output was regularly measured using the OL-754 spectroradiometer calibrated against standard lamps. The UVA doses were 1,750 and 3,500 mJ cm2, respectively, and for simplicity will be referred to as the lower and higher UVA doses. The UVB and ssUV doses were 285 and 3,860 mJ cm2, respectively, (comprising 285 UVB with 3,575 mJ cm2 UVA) and for simplicity will be referred to as the UVB and ssUV doses. The UV doses used in this study are physiological and representative of typical human sunlight exposure. The UVB and higher UVA doses are the amounts in sunlight that would cause barely detectable sunburn. The ssUV dose is the same amount of UVB combined with UVA to approximate the solar spectrum and give a barely detectable sunburn.
Contact hypersensitivity response The experimental design and protocol for all CHS experiments was as described (Byrne et al., 2002). Mouse abdomens were shaved and sensitized with a 50 ml epicutaneous application of 2% (w/v) oxazolone (4-ethoxymethylene-2-pheyl-2-oxazolin-5-one; Sigma, St Louis, MN) in acetone 3 days after the last UVA, UVB, and ssUV irradiation, which had been on the dorsal trunk. After 7 days, the mouse’s ears were challenged topically with 10 ml of 2% (w/v) oxazolone in acetone per ear. Changes in ear thickness were then measured using micrometer calipers (Interapid, Tesa, Bergdietikon Switzerland).
RNA isolation for the microarray study and real-time RT-PCR
Female C57BL/6 and Balb/c mice aged 8–9 weeks at the start of irradiations were used for these experiments (Animal Resource Centre, Perth, WA, Australia), which were conducted with the approval of the University of Sydney animal ethics committee. Before each irradiation, the dorsal trunk was shaved and the mice were then allowed to rest for 24 hours. Mice to be irradiated were then placed in an animal-restraining cage covered with a quartz glass lid. All irradiation times were o2 minutes and air-conditioning was used to ensure that the mouse body temperature did not increase during irradiation. Single irradiation for microarray and RT-PCR experiments and three irradiations on consecutive days for contact sensitivity experiments were given to each mouse and all irradiations were always carried out at the same time of the day.
At 24 hours after UVA, UVB, and ssUV irradiation a 1-cm2 uniform section of skin was excised from the dorsal surface of irradiated and control mice. Total RNA was then extracted from the whole skin using TRIzol reagent (Gibco Invitrogen Life Technologies, Carlsbad, CA) according to the manufacturer’s instructions, purified, DNase treated, and reverse transcribed into cDNA. For the microarray study, a direct incorporation of Cyanine 3-dCTP and Cyanine 5dCTP fluorescent dyes (Perkin Elmer Life Sciences, Inc. Boston, MA) was used for cDNA synthesis. For each UV dose, a reference design was used to compare an unirradiated control against an irradiated sample. Microarray experiments used compugen 22k mouse oligonucleotide microarray slides (The Clive and Vera Ramaciotti Centre for Gene Function Analysis, Sydney, Australia (http://www. ramaciotti.unsw.edu.au) and were replicated six times for each UV dose. A fluorescent dye swap was done for each alternate hybridization to reduce systematic dye bias of incorporated fluorescent dyes.
UV source
Microarray analysis
A 1,000 W xenon arc solar simulator (Oriel, Stratford, CT) with two dichroic mirrors that allow wavelengths between 200 and 400 nm to pass through were used with a filter that blocks UVB to produce a UVA spectrum. For ssUV and UVB, filters were used to attenuate the visible and infrared wavelengths and remove UVC (o290 nm) so that it simulated sunlight (ssUV), or comprised predominantly UVB. The spectra have earlier been described (Byrne et al., 2002). The intensity (mW cm2) of the UVA, UVB, and ssUV output was
The first component of the microarray analysis was done using the R software package Linear Models for Microarray Data (Bioconductor) (Wettenhall and Smyth, 2004) and involved no background correction, within-array print-tip loess normalization, followed by scale normalization for each unirradiated control and UV-irradiated group (Yang et al., 2002). To determine the biological pathways likely to be playing a role in UVA immunomodulation, UVB immunosuppression and the
MATERIALS AND METHODS Mice
www.jidonline.org 2699
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
interaction between UVB and UVA in ssUV, all genes expressed on the microarrays from each unirradiated control and UV-irradiated group were analyzed using GSEA (Subramanian et al., 2005). From each unirradiated control versus UV-irradiated group comparison, we used normalized red and green channel data in identifying pathways that were differentially expressed using GSEA. GSEA was carried out as described earlier by Subramanian et al., 2005 using a total of 489 gene sets containing genes whose products are involved in specific metabolic and signaling pathways (Subramanian et al., 2005). We used the default setting signal to noise ranking metric in GSEA to rank genes. The statistical significance (P-value) of each pathway was estimated by using phenotype-based permutation analysis using 1,000 random permutations. We chose pathways with a significant enrichment score (Po0.05) that had been adjusted for multiple comparisons using the FDR (qo0.05; or very close to qo0.05) for further interpretation. The FDR is a quantity that describes, for a set of tests that are called significant at or above a given level, what proportion of them are likely to be falsepositive findings (Storey and Tibshirani, 2003). For each unirradiated control versus UV-irradiated comparison, GSEA identified significantly enriched pathways that were either up- or downregulated by UVA, UVB, and ssUV. Each of these significantly enriched pathways contained core-enriched genes. These genes contribute most to the enrichment score of the pathway, and therefore are most likely to participate in a biological response. To determine which pathways were associated with the doserelated effects of UVA on immunomodulation in C57BL/6 mice, a cross-comparison was made between statistically enriched results in C57BL/6 and Balb/c mice. Any pathways that had a significant enrichment score (Po0.05) and that met the FDR qo0.05 criterion in both strains were regarded as unlikely to be associated with effects of UVA on immunomodulation. The remaining pathways were then suggested to be playing a role in the immunomodulating effects caused by each specific dose of UVA in C57BL/6 mice.
Real-time RT-PCR confirmation of candidate genes Primers used in this study were designed using the PCR simulation program Primer3 (Whitehead Institute for Biomedical Research, Cambridge, MA). Forward and reverse primers are shown in Supplementary Table 13. BLASTN searches confirmed the total gene specificity of the nucleotide sequences chosen for the primers. Real-time RT-PCR using 2 platinum SYBR Green qPCR SuperMixUDB (Invitrogen) was carried out on a Rotor-Gene RG-3000 PCR machine using Rotor-Gene 6.0 software for analysis (Corbett Research, NSW, Australia). Each experiment was repeated six times using six mice for each unirradiated control and UV-irradiated group. Data obtained from the six replicate values were pooled. For each real-time RT-PCR experiment the unirradiated control and UVirradiated group were normalized to the housekeeping gene HPRT. For data analysis the 2(-Delta Delta C(T)) method was used (Livak and Schmittgen, 2001). For statistical analysis, a Kolmogorov–Smirnoff test determined that the data obtained from the six biological replicates was normally distributed. An unpaired Student’s t-test was then used with Po0.05 compared with the unirradiated control mice (0 UV) being regarded as significant.
CONFLICTS OF INTEREST The authors state no conflicts of interest. 2700 Journal of Investigative Dermatology (2009), Volume 129
ACKNOWLEDGMENTS We would like to acknowledge Aravind Subramanian for providing us with the C2 mouse functional gene sets. Financial assistance was received from NHMRC project grant 302021 and Epiderm. SN Byrne was supported by the National Health and Medical Research Council CJ Martin Fellowship (307726) and the Cancer Institute NSW Career Development and Support Fellowship (07-CDF-1/07). R Williams was supported by an NH&MRC Peter Doherty fellowship (#307715).
SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid
REFERENCES Agar NS, Halliday GM, Barnetson RS, Ananthaswamy HN, Wheeler M, Jones AM (2004) The basal layer in human squamous tumors harbors more UVA than UVB fingerprint mutations: A role for UVA in human skin carcinogenesis. Proc Nat Acad Sci USA 101:4954–9 Byrne SN, Halliday GM (2005) B cells activated in lymph nodes in response to ultraviolet irradiation or by interleukin-10 inhibit dendritic cell induction of immunity. J Invest Dermatol 124:570–8 Byrne SN, Spinks N, Halliday GM (2002) Ultraviolet a irradiation of C57BL/6 mice suppresses systemic contact hypersensitivity or enhances secondary immunity depending on dose. J Invest Dermatol 119:858–64 Dovezenski N, Billetta R, Gigli I (1992) Expression and localization of proteins of the complement system in human skin. J Clin Invest 90:2000–12 Fearon DT, Austen KF (1975a) Properdin: binding to C3b and stabilization of the C3b-dependent C3 convertase. J Exp Med 142:856–63 Fearon DT, Austen KF (1975b) Properdin: initiation of alternative complement pathway. Proc Nat Acad Sci USA 72:3220–4 Gorman S, Tan JWY, Thomas JA, Townley SL, Stumbles PA, Finlay-Jones JJ et al. (2005) Primary defect in UVB-induced systemic immunomodulation does not relate to immature or funtionally impaired APCs in regional lymph nodes. J Immunol 174:6677–85 Halliday GM (2005) Inflammation, gene mutation and photoimmunosuppression in response to UVR-induced oxidative damage contributes to photocarcinogenesis. Mutation Res 571:107–20 Hammerberg C, Katiyar SK, Carroll MC, Cooper KD (1998) Activated complement component 3 (C3) is required for ultraviolet induction of immunosuppression and antigenic tolerance. J Exp Med 187:1133–8 Isenman DE, Cooper NR (1981) The structure and function of the third component of human complement–I. The nature and extent of conformational changes accompanying C3 activation. Mol Immunol 18:331–9 Kimura Y, Miwa T, Zhou L, Song WC (2008) Activator-specific requirement of properdin in the initiation and amplification of the alternative pathway complement. Blood 111:732–40 Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402–8 Mittelbrunn M, Tejedor R, de la Fuente H, Garcia-Lopez MA, Ursa A, Penas PF et al. (2005) Solar-simulated ultraviolet radiation induces abnormal maturation and defective chemotaxis of dendritic cells. J Invest Dermatol 125:334–42 Nishigori C, Yarosh DB, Ullrich SE, Vink AA, Bucana CD, Roza L et al. (1996) Evidence that DNA damage triggers interleukin 10 cytokine production in UV-irradiated murine keratinocytes. Proc Nat Acad Sci USA 93:10354 Ohl L, Mohaupt M, Czeloth N, Hintzen G, Kiafard Z, Zwirner J et al. (2004) CCR7 governs skin dendritic cell migration under inflammatory and steady-state conditions. Immunity 21:279–88 Pasch MC, Van Den Bosch NH, Daha MR, Bos JD, Asghar SS (2000) Synthesis of complement components C3 and factor B in human keratinocytes is differentially regulated by cytokines. J Invest Dermatol 114:78–82
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
Poon TS, Barnetson RS, Halliday GM (2005) Sunlight-induced immunosuppression in humans is initially because of UVB, then UVA, followed by interactive effects. J Invest Dermatol 125:840–6 Rauterberg A, Jung EG, Rauterberg EW (1993) Complement deposits in epidermal cells after ultraviolet B exposure. Photodermatol Photoimmunol Photomed 9:135–43 Reeve VE, Bosnic M, Boehm-Wilcox C, Nishimura N, Ley RD (1998) Ultraviolet A radiation (320–400 nm) protects hairless mice from immunosuppression induced by ultraviolet B radiation (280–320 nm) or cis-urocanic acid. Int Arch Allergy Appl Immunol 115:316–22 Schreiber RD, Muller-Eberhard HJ (1978) Assembly of the cytolytic alternative pathway of complement from 11 isolated plasma proteins. J Exp Med 148:1722–7 Schreiber RD, Pangburn MK, Lesavre PH, Muller-Eberhard HJ (1978) Initiation of the alternative pathway of complement: recognition of activators by bound C3b and assembly of the entire pathway from six isolated proteins. Proc Nat Acad Sci USA 75:3948–52
Stapp JM, Sjoelund V, Lassiter HA, Feldhoff RC, Feldhoff PW (2005) Recombinant rat IL-1beta and IL-6 synergistically enhance C3 mRNA levels and complement component C3 secretion by H-35 rat hepatoma cells. Cytokine 30:78–85 Storey JD, Tibshirani R (2003) Statistical significance for genomewide studies. Proc Nat Acad Sci USA 100:9440–5 Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Nat Acad Sci USA 102:15545–50 Toichi E, Lu KQ, Swick AR, McCormick TS, Cooper KD (2008) Skininfiltrating monocytes/macrophages migrate to draining lymph nodes and produce IL-10 after contact sensitizer exposure to UV-irradiated skin. J Invest Dermatol 128:2705–15 Wettenhall JM, Smyth GK (2004) limmaGUI: a graphical user interface for linear modeling of microarray data. Bioinformatics 20:3705–6
Schwarz A, Grabbe S, Aragane Y, Sandkuhl K, Riemann H, Luger TA et al. (1996) Interleukin-12 prevents ultraviolet B-induced local immunosuppression and overcomes UVB-induced tolerance. J Invest Dermatol 106:1187–91
Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J et al. (2002) Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res 30:e15
Shen J, Bao S, Reeve VE (1999) Modulation of IL-10, IL-12 and IFN-gamma in the Epidermis of Hairless Mice by UVA (320–400 nm) and UVB (280–320 nm) Radiation. J Invest Dermatol 113:1059–64
Yoshida Y, Kang K, Berger M, Chen G, Gilliam AC, Moser A et al. (1998) Monocyte induction of IL-10 and down-regulation of IL-12 by iC3b deposited in ultraviolet-exposed human skin. J Immunol 161:5873–9
www.jidonline.org 2701
The Alternative Complement Pathway Seems to be a UVA Sensor that Leads to Systemic Immunosuppression Michael P.F. Stapelberg1, Rohan B.H. Williams2, Scott N. Byrne1 and Gary M. Halliday1 UV wavebands in sunlight are immunomodulatory. About half the amount of UVA within a minimum erythemal dose of sunlight is systemically immunosuppressive, whereas higher doses protect from UVB immunosuppression in mice. We have earlier shown that these responses to UVA are genetically restricted, as they occur in C57BL/6 but not in Balb/c mice. We used gene set enrichment analysis of microarray data and real-time reverse transcriptase (RT)-PCR confirmation to determine the molecular mechanisms associated with UVA immunomodulation. We found upregulation of mRNA for the alternative complement pathway. The core-enriched genes complement component 3, properdin, and complement factor B were all activated by the immunosuppressive dose of UVA only in UVA-responsive C57BL/6 but not in unresponsive BALB/c mice. This therefore matched the genetic restriction and dose responsiveness of UVA immunosuppression. The immuneprotective higher UVA dose prevented UVB from downregulating chemokine receptor 7 and IL-12B, and decreased IL-10, supporting the earlier identification of IL-12 and IL-10 in high-dose UVA protection from UVB immunosuppression. Our study has identified activation of the alternative complement pathway as a trigger of UVA-induced systemic immunosuppression and suggests that this pathway is likely to be an important sensor of UVA-induced damage to the skin. Journal of Investigative Dermatology (2009) 129, 2694–2701; doi:10.1038/jid.2009.128; published online 11 June 2009
INTRODUCTION The UV wavebands contained in sunlight are the main cause of skin cancer. Gene mutations and immunosuppression are both important biological consequences of UV exposure that lead to skin cancer (Halliday, 2005). UV wavebands in sunlight are composed of approximately 5% UVB (290–320 nm) and 95% UVA (320–400 nm), both of which can cause gene mutations (Agar et al., 2004) and immunosuppression (Poon et al., 2005) in humans. UVA-induced molecular triggers for immunosuppression have received little research attention. Unlike UVB, which has a linear dose–response relationship with immunosuppres1
Discipline of Dermatology, Bosch Institute, Sydney Cancer Centre, at the University of Sydney, Sydney, New South Wales, Australia and 2The John Curtin School of Medical Research, Australian National University Canberra, ACT, and School of Biotechnology and Biomolecular Sciences, University of New South Wales, Sydney, New South Wales, Australia Work performed at the University of Sydney, Sydney, New South Wales, Australia Correspondence: Professor Gary M. Halliday, Dermatology Research Laboratories, Blackburn Building D06, University of Sydney, New South Wales 2006 Australia. E-mail: [email protected] Abbreviations: C1QA, complement component 1, q subcomponent, alpha polypeptide; C3, complement component 3; CCR7, chemokine receptor 7; CFB, complement factor B; CHS, contact hypersensitivity; FDR, false discovery rate; GSEA, gene set enrichment analysis; RT, reverse transcriptase; ssUV, solar-simulated ultraviolet radiation. Received 17 December 2008; revised 17 February 2009; accepted 21 March 2009; published online 11 June 2009
2694 Journal of Investigative Dermatology (2009), Volume 129
sion, UVA causes a complex bell-shaped dose–response curve. Therefore, the molecular signals initiated by UVA that lead to immunosuppression are also likely to have a nonlinear dose response. We have earlier shown that three consecutive exposures of C57BL/6 mice to 1,700 mJ cm2 UVA causes systemic immunosuppression on sensitization at a distal non-irradiated site, whereas twice this dose does not induce immunosuppression. These dose-related effects of UVA are genetically restricted, as they occurred in C57BL/6 but not in Balb/c mice. Furthermore, exposure of C57BL/6 mice to solar-simulated UV (ssUV) consisting of an immunosuppressive UVB dose combined with the higher nonsuppressive dose of UVA does not cause immunosuppression (Byrne et al., 2002). This is consistent with studies that highdose UVA protects mice from UVB immunosuppression (Reeve et al., 1998). Although UVA does not contribute substantially to the erythemal or oedemal components of sunburn, the immunosuppressive dose of UVA in mice and humans (Poon et al., 2005) is about half of what would be delivered to pale skinned Caucasians receiving a dose of sunlight sufficient to cause barely detectable sunburn (minimum erythemal dose). Therefore, these immunosuppressive UVA doses are within the range received by humans during normal daily activities. UVA must therefore affect different molecular mechanisms at different doses to have this complex affect on immunity. As there are no known UVA-induced molecular responses to explain immunosuppression, we took a global approach of & 2009 The Society for Investigative Dermatology
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
microarray analysis and validation with quantitative real-time reverse transcriptase (RT)-PCR combined with the power of the genetic association of the UVA dose response in C57BL/6 but not in Balb/c mice to determine the most prominent molecular pathway changes that were associated with UVA immunosuppression. RESULTS All raw microarray data are available at Gene Expression Omnibus (GEO) with accession GSE15618. UVA causes systemic immunosuppression in C57BL/6 but not in Balb/c mice with a bell-shaped dose response
UVA systemically suppressed the induction of primary immunity in C57BL/6 mice at a lower (1,750 mJ cm2) but not higher (3,500 mJ cm2) dose (Figure 1a). The contact hypersensitivity (CHS) results also showed that UVA did not modulate immunity in Balb/c mice over this dose range
(Figure 1b). We further clarified the immunomodulatory effects of UVA by combining the higher dose with an immunosuppressive UVB dose (285 mJ cm2) chosen so that this mix of UVA and UVB was a good mimic of sunlight (ssUV). UVB suppressed the induction of CHS in both mouse strains. However, the CHS response of C57BL/6 mice was unaffected by this dose of UVB when combined with the higher dose of UVA to produce ssUV (Figure 1a). Therefore, this higher dose of UVA protected C57BL/6 mice from the immunosuppressive effects of UVB. In contrast, ssUV suppressed CHS in Balb/c mice (Figure 1b), showing that the higher UVA dose did not protect from UVB in this mouse strain. Therefore, UVA modulates immunity in C57BL/6 but not in Balb/c mice with a complex dose response. It suppressed immunity at the lower but not the higher dose. The higher dose is not immunologically inert, as it protects from UVB-induced suppression. The immunosuppressive, but not the higher UVA dose activates the alternative complement pathway
C57BL/6 (n)17
Unirradiated
(n)13 P<0.0001
Lower UVA Higher UVA
(n)13 n.s
UVB
(n)12 P<0.0001 (n)16 n.s
ssUV 0.0
0.2
0.4
0.6
0.8
1.0
1.2
Normalized change in ear thickness Balb/c (n)13
Unirradiated Lower UVA
(n)13 n.s
Higher UVA
(n)12 n.s
UVB
(n)17 P<0.0001
ssUV
(n) 9 P<0.0001 0.0
0.2
0.4
0.6
0.8
1.0
1.2
Normalized change in ear thickness Figure 1. UVA causes immunosuppression at the lower and not higher dose tested in C57BL/6 but not in Balb/c mice. Contact hypersensitivity (CHS) was assessed in (a) C57BL/6 and (b) Balb/c mice by measuring ear swelling 24 hours after challenge, which was 7 days after sensitization on the abdomen. Mice had been irradiated on the dorsal trunk for 3 consecutive days with 1,750 (lower UVA) or 3,500 (higher UVA) mJ cm2 UVA, or with 285 mJ cm2 UVB or this UVB dose combined with the higher UVA to produce solarsimulated UV (ssUV). Irritant controls (ear swelling of unsensitized mice) were subtracted from sensitized mice. For each CHS experiment the ear swelling for each mouse that was irradiated with UV was normalized to the mean of the control group that did not receive UV irradiation to remove interexperimental variation in magnitude of the control response. Each experiment was repeated thrice and normalized values were pooled. The mean and SEM of normalized values are shown. The total number of mice for each group is shown (n). An unpaired student’s t-test was used for statistical analysis compared with the unirradiated group.
To search for signaling pathways likely to be involved in UVA-induced immunosuppression, we used microarrays with gene set enrichment analysis (GSEA). This analysis searches for co-ordinated changes in whole pathways and therefore is more robust than searching for changes in single genes and also enables a stringent statistical analysis to be carried out. Enriched pathways (Po0.05 and false discovery rate (FDR)o0.05) were identified for irradiated C57BL/6 and Balb/c mice compared with unirradiated control mice. Our search strategy, on the basis of the CHS responses described above was that a pathway likely to be involved in UVA immunosuppression would be altered by the lower, but not the higher, UVA dose in C57BL/6 but not in Balb/c mice. This would match the genetic restriction and dose response. All statistically significant enriched pathways are listed in Table 1. Furthermore, for each enriched pathway listed in Table 1, core-enriched genes were identified using GSEA. Core-enriched genes contribute most to the enrichment score of the pathway. The core-enriched genes chosen for confirmation by real-time PCR are also listed in Table 1. The 50 most up- and downregulated pathways for each UV spectra, dose and mouse strain are shown in supplementary tables 1–12 to provide additional information. In C57BL/6 mice, only the alternative complement pathway (P ¼ 0.001, FDR ¼ 0.035) was significantly upregulated. No pathways were downregulated. Neither this, nor any other pathway was altered by the lower UVA dose in Balb/c mice, and the higher UVA dose did not modulate this pathway in C57BL/6 mice. To confirm the association of the alternative complement pathway with UVA-induced immunosuppression, a selected number of core-enriched genes were examined in a series of different skin samples by quantitative real-time RT-PCR (Figure 2). In C57BL/6 mouse skin irradiated with the lower UVA dose, real-time RT-PCR showed significant upregulation of the alternative complement pathway genes C3, complement factor B (CFB) and properdin. These genes were not altered by the higher UVA dose, UVB, or ssUV (which contained the higher UVA dose). www.jidonline.org 2695
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
Table 1. Enriched pathways for lower and higher UVA, UVB, and ssUV Enriched phenotype1
Strain
Genes2
0.001
Increased by lower UVA
C57BL/6
C33, CFB3, PFC3
0.014
0.001
Increased by higher UVA
C57BL/6
CD803, CD86
Classical complement
0.039
0.001
Increased by higher UVA
C57BL/6
C1QA3, C1QB
Cytokine
0.010
0.001
Decreased by higher UVA
C57BL/6
IL3, IL9
Cytokine
0.007
0.001
Decreased by higher UVA
Balb/c
IL3, IL9
b alanine metabolism
0.005
0.001
Increased by UVB
C57BL/6
ALDH23, ALDH9A13
Propanoate metabolism
0.011
0.001
Increased by UVB
C57BL/6
ALDH23, ALDH9A13
Valine, leucine, and isoleucine degradation
0.012
0.001
Increased by UVB
C57BL/6
ALDH23, ALDH9A13
nkt
0.052
0.003
Decreased by UVB
C57BL/6
CCR73, IFNg3
IL12
0.023
0.005
Increased by ssUV
C57BL/6
IL12B3, IL12RB2
IL10
0.047
0.001
Decreased by ssUV
C57BL/6
IL63, IL103
Enriched pathway Po0.05
FDR q-value
P-value
Alternative complement
0.035
Inflammatory response
ALDH2, aldehyde dehydrogenase 2 family; ALDH9A1, aldehyde dehydrogenase 9 family, member A1; CCR7, chemokine receptor 7; CFB, complement factor B; FDR, false discovery rate; PFC, properdin; RT, reverse transcriptase. 1 Increased or decreased by UV compared with unirradiated control. 2 Core-enriched and associated genes for enriched pathways. 3 Genes selected for real-time RT-PCR.
C3
PFC
CFB Unirradiated
Unirradiated
∗
Lower UVA
Unirradiated
∗
Lower UVA
Higher UVA
UVB
UVB
UVB
ssUV
ssUV
ssUV
0
1
2 3 4 5 Fold change
0
6
1
2 3 4 5 Fold change
CD80
6
0
∗
ALDH2
∗∗
Higher UVA
Lower UVA
∗
Higher UVA
UVB
UVB
UVB
ssUV
ssUV
ssUV
0 1 2 3 4 5 6 7 Fold change
0
1
2 3 4 5 Fold change
ALDH9A1
0
6
IFN-γ
Lower UVA
Lower UVA
Lower UVA
Higher UVA
Higher UVA
UVB
UVB
ssUV
ssUV 3
4
∗∗
UVB ssUV
1
0
2
3
IL-12B
IL-6
Lower UVA
Lower UVA
Higher UVA
Higher UVA
∗∗
ssUV 0
1
2
1
3
4
Fold change
2
3
Fold change IL-10
Unirradiated
UVB
0
Fold change
Fold change
Unirradiated
4
Unirradiated
Higher UVA
2
1 2 3 Fold change
CCR7 Unirradiated
Unirradiated
5
Unirradiated
Lower UVA
Lower UVA Higher UVA
1
1 2 3 4 Fold change
C1QA Unirradiated
Unirradiated
0
∗
Lower UVA
Higher UVA
Higher UVA
Unirradiated
∗∗
Lower UVA Higher UVA
UVB
UVB
ssUV
ssUV 0
1
2
3
Fold change
4
∗ 0
1
2
3
4
5
Fold change
Figure 2. Confirmation of quantitative gene expression of core enriched and associated genes using real-time reverse transcriptase (RT)-PCR. For statistical analysis a Kolmogorov–Smirnoff test first determined that the data obtained from the six biological replicates were normally distributed. An unpaired student’s t-test was then used for statistical analysis * Po0.05, ** Po0.01 compared with the unirradiated control mice (0 UV). Bar graphs represent means þ SEM of six biological replicates for lower and higher UVA, UVB, and ssUV doses. ALDH2, aldehyde dehydrogenase 2 family; ALDH9A1, aldehyde dehydrogenase 9 family, member A1; CCR7, chemokine receptor 7; CFB, complement factor B; PFC, properdin.
2696 Journal of Investigative Dermatology (2009), Volume 129
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
In UVA unresponsive Balb/c mice, none of these genes were altered by low-dose UVA, and only C3 was significantly downregulated by the higher UVA dose (results not presented). Therefore, increased expression of alternative complement pathway genes correlated with the genetic restriction and dose response for UVA-induced immunosuppression. Higher dose of immunoprotective UVA associated with normalization of IL-12B and CCR7, and decreased IL-10
To examine the possible immunomodulatory signals activated by the higher immunoprotective dose of UVA, this dose was compared with unirradiated mice, and the combination of this UVA dose with UVB (ssUV) was contrasted to UVB-irradiated mice. In C57BL/6, but not in Balb/c mice, the inflammatory response (P ¼ 0.001, FDR ¼ 0.014) and classical complement (P ¼ 0.001, FDR ¼ 0.039) pathways were significantly upregulated. The cytokine pathway was significantly downregulated in both mouse strains. As C57BL/6, but not Balb/c mice were modulated over this UVA dose range, this pathway was regarded as not being involved in immunomodulation and was not considered further. To determine the pathways involved in higher dose UVA protection from UVB, relative effects of UVB compared with ssUV is more insightful to determine the UVB effects modulated by addition of higher dose UVA. In C57BL/6 mice, the b alanine metabolism (P ¼ 0.001, FDR ¼ 0.005), propanoate metabolism (P ¼ 0.001, FDR ¼ 0.011), and valine, leucine, and isoleucine degradation (P ¼ 0.001, FDR ¼ 0.012) pathways were significantly upregulated. No pathways were downregulated by UVB using the qo0.05 FDR cut-off. However, the nkt pathway (P ¼ 0.003, FDR ¼ 0.052) was very close to the qo0.05 FDR cut-off and was therefore considered further, as these criteria have earlier been suggested to be worthy of further study (Subramanian et al., 2005). In C57BL/6 mice exposed to ssUV, only the IL-12 pathway was upregulated (P ¼ 0.005, FDR ¼ 0.023) and the IL-10 (P ¼ 0.001, FDR ¼ 0.047) pathway was downregulated (Table 1). These changes were not induced by ssUV in Balb/c mice. After this screen by pathway analysis of microarray data, core-enriched genes identified in these pathways (Table 1) were compared in a different set of skin from irradiated mice by real-time RT-PCR. The higher UVA dose significantly upregulated the inflammatory response pathway CD80 and classical complement pathway C1QA core-enriched genes, confirming the microarray GSEA. Although these genes were not altered in Balb/c mice (data not shown), they were not increased by the same UVA dose when combined with UVB (ssUV), indicating, not surprisingly, that the combined effects of these wavebands differ from the summation of each waveband alone. Therefore, these changes are unlikely to be important for higher dose UVA protection from UVB. In C57BL/6 mice irradiated with UVB, significant downregulation of the nkt pathway core-enriched gene chemokine receptor 7 (CCR7) determined by real-time RT-PCR confirmed the microarray data. The core-enriched genes for the remaining pathways identified by GSEA (aldehyde dehydro-
genase 2 family and aldehyde dehydrogenase 9 family, member A1) were not confirmed by RT-PCR. In C57BL/6 mice irradiated with ssUV, GSEA identified the IL-12 pathway as being increased and the IL-10 pathway as being decreased. Real-time RT-PCR did not confirm the ssUV-induced increase in the IL-12B gene, as it was not significantly different to the unirradiated group. However, in comparison with the UVB dose, the IL-12 pathway gene IL-12B was no longer downregulated. Therefore, the UVA component of ssUV inhibited the downregulation of IL-12B caused by UVB. Furthermore, IL-10 was confirmed to be significantly downregulated by ssUV but not by UVB in C57BL/6 but not in Balb/c mice (data not shown for Balb/c mice). In summary these data show that genes upregulated by the immunoprotective UVA dose alone do not predict differences between the effects of UVB and ssUV. They further indicate that prevention of UVB-induced downregulation of CCR7 and IL-12B along with reduced expression of IL-10 seem to correlate with higher dose UVA protection from UVB-induced immunosuppression. DISCUSSION In this study, we used pathway analysis of microarray data and quantitative real-time RT-PCR to identify molecular mechanisms associated with systemic UVA immunosuppression. The CHS results in this study were consistent with earlier findings (Byrne et al., 2002). UVA was immunosuppressive at a lower, but not higher, dose in C57BL/6 but not in Balb/c mice. The higher UVA was not immunologically inert, but protected from UVB-induced immunosuppression with the same genetic association. We determined that the alternative complement pathway responds to UVA and may be a key trigger of UVA-induced immunosuppression. This is because mRNA for genes in this pathway was found to be upregulated by the immunosuppressive dose of UVA, but not by the higher immunoprotective dose, nor by UVB alone or combined with the higher protective UVA dose. This upregulation of alternative complement occurred in responsive C57BL/6 but not in unresponsive Balb/c mice, and therefore was associated with the genetic restriction, dose and waveband response relationship with UVA immunosuppression. Our data also showed that immunoprotection by the higher UVA dose was allied with the prevention of UVBinduced depletion of CCR7 and IL-12B, and a reduction in IL-10 mRNA. Three pathways lead to complement activation, the classical, lectin, and alternative pathways. All three pathways result in the production of C3 and a subsequent activation of the final lytic pathway and membrane attack complex. The alternative pathway is unique in complement activation, in that it functions as an amplification loop activating itself with its own product, C3b (Schreiber and Muller-Eberhard, 1978; Schreiber et al., 1978). Complement components C3 and CFB are central to the alternative pathway, as after activation they become key constituents of the serine protease C3bBb (Isenman and Cooper, 1981). C3bBb cleaves C3 molecules producing C3b (Schreiber and Muller-Eberhard, 1978; Schreiber et al., 1978). The cleavage of C3 in the alternative pathway is promoted by properdin (Fearon and Austen, www.jidonline.org 2697
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
1975a, b). Properdin protects C3bBb and C3b bi-products including iC3b from degradation, resulting in an accumulation of C3b and iC3b in tissue. The epidermis of normal skin is a rich source of components of the alternative complement pathway including C3 and CFB (Dovezenski et al., 1992). C3associated molecules of the alternative pathway have also been observed in the basement membrane zone of the epidermis (Yoshida et al., 1998). Studies have shown that epidermally derived cytokines including IL-6 enhance the synthesis of alternate pathway C3 and CFB (Pasch et al., 2000; Stapp et al., 2005). In this respect, it is interesting that the suppressive UVA, but not the higher dose, increased IL-6 expression. Therefore, induction of IL-6 could contribute to UVA-induced immunosuppression by augmenting or initiating activation of the alternative complement pathway. It has not been suggested earlier that the alternative complement pathway is involved in systemic immunosuppression caused by UVA exposure of the skin. However, UVB causes deposition of C3 and later components of the membrane attack complex in the skin (Rauterberg et al., 1993), and UVB-induced activation of C3 to produce iC3b has been shown to be important for local immunosuppression (Hammerberg et al., 1998). Whether the iC3b involved in UVB-induced local immunosuppression was a product of the classical, lectin, or alternative pathways was not clear, but iC3b binding to its receptor CD11b on macrophages infiltrating UVB-exposed skin was identified as being important for local UVB-induced immunosuppression (Hammerberg et al., 1998). Neutralization of C3 with soluble complement receptor 1 prevented UVB-induced local immunosuppression and reduced infiltration by CD11b þ macrophages; in addition, immunosuppression did not occur in C3 knockout mice. This study differed to ours in a number of ways. A dose of UVB was used that induced local but not systemic immunosuppression, and both iC3b and infiltrating CD11b þ macrophages were required at the site of antigen exposure. In our study, activation of the alternative complement pathway in UVA-irradiated skin was at a different site to antigen exposure, and therefore interactions between complement components, any inflammatory cells, and antigen could not have occurred. In addition, the UVA irradiation regime used in these studies systemically suppresses primary but not secondary immunity and therefore does not induce tolerance (Byrne et al., 2002). This contrasts with the studies described above that showed a role for C3 using a UVB irradiation dose that caused locally induced tolerance (Hammerberg et al., 1998).Therefore, the mechanism by which the alternative complement pathway leads to systemic UVA immunosuppression is likely to be different to what was described by Hammerberg et al. Nevertheless, our study supports an important regulatory role for complement in responses of the skin immune system to UV radiation. It is unclear how the immunosuppressive, but not the higher, dose of UVA could induce expression of the alternative complement pathway. This pathway has multiple positive and negative regulators. It is possible that the lower UVA dose could activate one of the positive regulators and 2698 Journal of Investigative Dermatology (2009), Volume 129
the higher UVA dose activates one of the inhibitors, or the lower UVA dose could relieve the pathway of one of its inhibitory mechanisms, enabling it to become activated. For example, properdin binding to lipopolysaccharide activates the alternative complement pathway, and the yeast cell wall component zymosan can activate alternative complement through a mechanism dependent on CFB (Kimura et al., 2008). These complement components therefore seem to be pattern recognition molecules. UVA is a potent oxidizer of lipids and other cellular macromolecules (Halliday, 2005), and it is possible that alternative complement components may recognize and be activated by these UVA-altered cell structures. It is unlikely that activation of the membraneattack complex is required for the alternative complement pathway to trigger UVA immunosuppression, as these later steps in the complement pathway, components C4–C9, are part of the classical pathway, which were not found to be involved in UVA immunosuppression in our study. It is more likely that one of the alternative complement proteins acts as a sensor of UVA-induced damage by binding to an unknown cell surface receptor or triggering a sequence of events culminating in systemic immunosuppression. Elucidation of these details will require extensive biochemical and functional characterization. Higher dose UVA protection from UVB immunosuppression was associated with UVA prevention of UVB-induced reduction in CCR7 and IL-12B, and with a reduction in IL-10 expression. Activated dendritic cells, including Langerhans cells, express the CCR7 (Ohl et al., 2004). UV has earlier been shown to impair the induction of CCR7 on dendritic cells during maturation, thus preventing their migration to draining lymph nodes (Mittelbrunn et al., 2005). In these experiments, antigen was applied to normal unirradiated skin and dendritic cells function and numbers in the draining lymph nodes have been found not to be altered during systemic immunosuppression (Byrne and Halliday, 2005; Gorman et al., 2005). It is therefore unlikely that high-dose UVA protection of CCR7 expression in the irradiated skin would prevent systemic immunosuppression. A UVA dose about 10-fold higher than what was used in these studies, that has previously been shown to protect from systemic UVB immunosuppression in Skh:HR-1 hairless mice, has been shown to abrogate UVB-enhanced expression of IL-10, and to increase IL-12 protein production (Shen et al., 1999). This study further showed that IFN-g was initially upregulated by high-dose UVA, which then lead to an increased expression of IL-12. The genome-wide expression analysis in this study, using different mouse strains and a lower UVA dose supports this earlier study by Shen et al. that high-dose UVA inhibition of IL-10 and prevention of IL-12 reduction by UVB are likely to be important mediators of UVA immunoprotection. This is also consistent with earlier studies that UVB-induced production of IL-10 contributes to UV immunosuppression (Nishigori et al., 1996) and that IL-12 can reverse UVB-induced immunosuppression by blocking IL-10 secretion (Schwarz et al., 1996). Macrophages infiltrating UVB-exposed skin migrate to lymph nodes in response to contact sensitization in which they produce
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
IL-10, and UV also reduces IL-12 production by lymph node CD11c þ cells in response to contact sensitization (Toichi et al., 2008). Thus, inflammatory cells and keratinocytes (Shen et al., 1999) show a similar pattern of IL-12 and IL-10 production that regulates responses to UV. We examined skin after exposure to low doses of UVA, which does not cause observable inflammation and therefore the cytokines and complement components were likely to have been produced by resident skin cells. In summary, our studies have identified that the alternative complement pathway seems to be a sensor of UVA-induced damage to the skin, leading to systemic immunosuppression. The immunosuppressive dose of UVA is only half the amount contained in a minimum erythemal dose of sunlight. Therefore, this UVA dose can be achieved by sunlight exposure during normal daily activities and is physiologically relevant for a large proportion of humans. Alternative complement is a complex pathway with multiple proteins that could be activated or inhibited by recognition of UVA-altered molecular structures in the skin and may be responsible for the complex bell-shaped dose response for UVA immunosuppression. This was the only pathway identified by our global genome approach and pathway analysis to be associated with UVA-induced immunosuppression. This approach was limited to detection of events leading to a change in mRNA expression, and therefore other events that we did not detect are likely to be involved. Our studies also confirmed earlier observations that high-dose UVA protection from UVBinduced immunosuppression is mediated, at least in part, by protection of IL-12 and a reduction in IL-10. Complement seems to be an important trigger of UV immunosuppression and requires further research.
measured regularly using a radiometer (Solar Light Company, PA), which had been calibrated against the source with an OL-754 spectroradiometer (Optronics Laboratories, Orlando, FL), and the timing of UV exposure was adjusted using an automated timing device to deliver consistent UV doses. The spectral output was regularly measured using the OL-754 spectroradiometer calibrated against standard lamps. The UVA doses were 1,750 and 3,500 mJ cm2, respectively, and for simplicity will be referred to as the lower and higher UVA doses. The UVB and ssUV doses were 285 and 3,860 mJ cm2, respectively, (comprising 285 UVB with 3,575 mJ cm2 UVA) and for simplicity will be referred to as the UVB and ssUV doses. The UV doses used in this study are physiological and representative of typical human sunlight exposure. The UVB and higher UVA doses are the amounts in sunlight that would cause barely detectable sunburn. The ssUV dose is the same amount of UVB combined with UVA to approximate the solar spectrum and give a barely detectable sunburn.
Contact hypersensitivity response The experimental design and protocol for all CHS experiments was as described (Byrne et al., 2002). Mouse abdomens were shaved and sensitized with a 50 ml epicutaneous application of 2% (w/v) oxazolone (4-ethoxymethylene-2-pheyl-2-oxazolin-5-one; Sigma, St Louis, MN) in acetone 3 days after the last UVA, UVB, and ssUV irradiation, which had been on the dorsal trunk. After 7 days, the mouse’s ears were challenged topically with 10 ml of 2% (w/v) oxazolone in acetone per ear. Changes in ear thickness were then measured using micrometer calipers (Interapid, Tesa, Bergdietikon Switzerland).
RNA isolation for the microarray study and real-time RT-PCR
Female C57BL/6 and Balb/c mice aged 8–9 weeks at the start of irradiations were used for these experiments (Animal Resource Centre, Perth, WA, Australia), which were conducted with the approval of the University of Sydney animal ethics committee. Before each irradiation, the dorsal trunk was shaved and the mice were then allowed to rest for 24 hours. Mice to be irradiated were then placed in an animal-restraining cage covered with a quartz glass lid. All irradiation times were o2 minutes and air-conditioning was used to ensure that the mouse body temperature did not increase during irradiation. Single irradiation for microarray and RT-PCR experiments and three irradiations on consecutive days for contact sensitivity experiments were given to each mouse and all irradiations were always carried out at the same time of the day.
At 24 hours after UVA, UVB, and ssUV irradiation a 1-cm2 uniform section of skin was excised from the dorsal surface of irradiated and control mice. Total RNA was then extracted from the whole skin using TRIzol reagent (Gibco Invitrogen Life Technologies, Carlsbad, CA) according to the manufacturer’s instructions, purified, DNase treated, and reverse transcribed into cDNA. For the microarray study, a direct incorporation of Cyanine 3-dCTP and Cyanine 5dCTP fluorescent dyes (Perkin Elmer Life Sciences, Inc. Boston, MA) was used for cDNA synthesis. For each UV dose, a reference design was used to compare an unirradiated control against an irradiated sample. Microarray experiments used compugen 22k mouse oligonucleotide microarray slides (The Clive and Vera Ramaciotti Centre for Gene Function Analysis, Sydney, Australia (http://www. ramaciotti.unsw.edu.au) and were replicated six times for each UV dose. A fluorescent dye swap was done for each alternate hybridization to reduce systematic dye bias of incorporated fluorescent dyes.
UV source
Microarray analysis
A 1,000 W xenon arc solar simulator (Oriel, Stratford, CT) with two dichroic mirrors that allow wavelengths between 200 and 400 nm to pass through were used with a filter that blocks UVB to produce a UVA spectrum. For ssUV and UVB, filters were used to attenuate the visible and infrared wavelengths and remove UVC (o290 nm) so that it simulated sunlight (ssUV), or comprised predominantly UVB. The spectra have earlier been described (Byrne et al., 2002). The intensity (mW cm2) of the UVA, UVB, and ssUV output was
The first component of the microarray analysis was done using the R software package Linear Models for Microarray Data (Bioconductor) (Wettenhall and Smyth, 2004) and involved no background correction, within-array print-tip loess normalization, followed by scale normalization for each unirradiated control and UV-irradiated group (Yang et al., 2002). To determine the biological pathways likely to be playing a role in UVA immunomodulation, UVB immunosuppression and the
MATERIALS AND METHODS Mice
www.jidonline.org 2699
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
interaction between UVB and UVA in ssUV, all genes expressed on the microarrays from each unirradiated control and UV-irradiated group were analyzed using GSEA (Subramanian et al., 2005). From each unirradiated control versus UV-irradiated group comparison, we used normalized red and green channel data in identifying pathways that were differentially expressed using GSEA. GSEA was carried out as described earlier by Subramanian et al., 2005 using a total of 489 gene sets containing genes whose products are involved in specific metabolic and signaling pathways (Subramanian et al., 2005). We used the default setting signal to noise ranking metric in GSEA to rank genes. The statistical significance (P-value) of each pathway was estimated by using phenotype-based permutation analysis using 1,000 random permutations. We chose pathways with a significant enrichment score (Po0.05) that had been adjusted for multiple comparisons using the FDR (qo0.05; or very close to qo0.05) for further interpretation. The FDR is a quantity that describes, for a set of tests that are called significant at or above a given level, what proportion of them are likely to be falsepositive findings (Storey and Tibshirani, 2003). For each unirradiated control versus UV-irradiated comparison, GSEA identified significantly enriched pathways that were either up- or downregulated by UVA, UVB, and ssUV. Each of these significantly enriched pathways contained core-enriched genes. These genes contribute most to the enrichment score of the pathway, and therefore are most likely to participate in a biological response. To determine which pathways were associated with the doserelated effects of UVA on immunomodulation in C57BL/6 mice, a cross-comparison was made between statistically enriched results in C57BL/6 and Balb/c mice. Any pathways that had a significant enrichment score (Po0.05) and that met the FDR qo0.05 criterion in both strains were regarded as unlikely to be associated with effects of UVA on immunomodulation. The remaining pathways were then suggested to be playing a role in the immunomodulating effects caused by each specific dose of UVA in C57BL/6 mice.
Real-time RT-PCR confirmation of candidate genes Primers used in this study were designed using the PCR simulation program Primer3 (Whitehead Institute for Biomedical Research, Cambridge, MA). Forward and reverse primers are shown in Supplementary Table 13. BLASTN searches confirmed the total gene specificity of the nucleotide sequences chosen for the primers. Real-time RT-PCR using 2 platinum SYBR Green qPCR SuperMixUDB (Invitrogen) was carried out on a Rotor-Gene RG-3000 PCR machine using Rotor-Gene 6.0 software for analysis (Corbett Research, NSW, Australia). Each experiment was repeated six times using six mice for each unirradiated control and UV-irradiated group. Data obtained from the six replicate values were pooled. For each real-time RT-PCR experiment the unirradiated control and UVirradiated group were normalized to the housekeeping gene HPRT. For data analysis the 2(-Delta Delta C(T)) method was used (Livak and Schmittgen, 2001). For statistical analysis, a Kolmogorov–Smirnoff test determined that the data obtained from the six biological replicates was normally distributed. An unpaired Student’s t-test was then used with Po0.05 compared with the unirradiated control mice (0 UV) being regarded as significant.
CONFLICTS OF INTEREST The authors state no conflicts of interest. 2700 Journal of Investigative Dermatology (2009), Volume 129
ACKNOWLEDGMENTS We would like to acknowledge Aravind Subramanian for providing us with the C2 mouse functional gene sets. Financial assistance was received from NHMRC project grant 302021 and Epiderm. SN Byrne was supported by the National Health and Medical Research Council CJ Martin Fellowship (307726) and the Cancer Institute NSW Career Development and Support Fellowship (07-CDF-1/07). R Williams was supported by an NH&MRC Peter Doherty fellowship (#307715).
SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid
REFERENCES Agar NS, Halliday GM, Barnetson RS, Ananthaswamy HN, Wheeler M, Jones AM (2004) The basal layer in human squamous tumors harbors more UVA than UVB fingerprint mutations: A role for UVA in human skin carcinogenesis. Proc Nat Acad Sci USA 101:4954–9 Byrne SN, Halliday GM (2005) B cells activated in lymph nodes in response to ultraviolet irradiation or by interleukin-10 inhibit dendritic cell induction of immunity. J Invest Dermatol 124:570–8 Byrne SN, Spinks N, Halliday GM (2002) Ultraviolet a irradiation of C57BL/6 mice suppresses systemic contact hypersensitivity or enhances secondary immunity depending on dose. J Invest Dermatol 119:858–64 Dovezenski N, Billetta R, Gigli I (1992) Expression and localization of proteins of the complement system in human skin. J Clin Invest 90:2000–12 Fearon DT, Austen KF (1975a) Properdin: binding to C3b and stabilization of the C3b-dependent C3 convertase. J Exp Med 142:856–63 Fearon DT, Austen KF (1975b) Properdin: initiation of alternative complement pathway. Proc Nat Acad Sci USA 72:3220–4 Gorman S, Tan JWY, Thomas JA, Townley SL, Stumbles PA, Finlay-Jones JJ et al. (2005) Primary defect in UVB-induced systemic immunomodulation does not relate to immature or funtionally impaired APCs in regional lymph nodes. J Immunol 174:6677–85 Halliday GM (2005) Inflammation, gene mutation and photoimmunosuppression in response to UVR-induced oxidative damage contributes to photocarcinogenesis. Mutation Res 571:107–20 Hammerberg C, Katiyar SK, Carroll MC, Cooper KD (1998) Activated complement component 3 (C3) is required for ultraviolet induction of immunosuppression and antigenic tolerance. J Exp Med 187:1133–8 Isenman DE, Cooper NR (1981) The structure and function of the third component of human complement–I. The nature and extent of conformational changes accompanying C3 activation. Mol Immunol 18:331–9 Kimura Y, Miwa T, Zhou L, Song WC (2008) Activator-specific requirement of properdin in the initiation and amplification of the alternative pathway complement. Blood 111:732–40 Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402–8 Mittelbrunn M, Tejedor R, de la Fuente H, Garcia-Lopez MA, Ursa A, Penas PF et al. (2005) Solar-simulated ultraviolet radiation induces abnormal maturation and defective chemotaxis of dendritic cells. J Invest Dermatol 125:334–42 Nishigori C, Yarosh DB, Ullrich SE, Vink AA, Bucana CD, Roza L et al. (1996) Evidence that DNA damage triggers interleukin 10 cytokine production in UV-irradiated murine keratinocytes. Proc Nat Acad Sci USA 93:10354 Ohl L, Mohaupt M, Czeloth N, Hintzen G, Kiafard Z, Zwirner J et al. (2004) CCR7 governs skin dendritic cell migration under inflammatory and steady-state conditions. Immunity 21:279–88 Pasch MC, Van Den Bosch NH, Daha MR, Bos JD, Asghar SS (2000) Synthesis of complement components C3 and factor B in human keratinocytes is differentially regulated by cytokines. J Invest Dermatol 114:78–82
MPF Stapelberg et al. Complement Mediates UVA Immunosuppression
Poon TS, Barnetson RS, Halliday GM (2005) Sunlight-induced immunosuppression in humans is initially because of UVB, then UVA, followed by interactive effects. J Invest Dermatol 125:840–6 Rauterberg A, Jung EG, Rauterberg EW (1993) Complement deposits in epidermal cells after ultraviolet B exposure. Photodermatol Photoimmunol Photomed 9:135–43 Reeve VE, Bosnic M, Boehm-Wilcox C, Nishimura N, Ley RD (1998) Ultraviolet A radiation (320–400 nm) protects hairless mice from immunosuppression induced by ultraviolet B radiation (280–320 nm) or cis-urocanic acid. Int Arch Allergy Appl Immunol 115:316–22 Schreiber RD, Muller-Eberhard HJ (1978) Assembly of the cytolytic alternative pathway of complement from 11 isolated plasma proteins. J Exp Med 148:1722–7 Schreiber RD, Pangburn MK, Lesavre PH, Muller-Eberhard HJ (1978) Initiation of the alternative pathway of complement: recognition of activators by bound C3b and assembly of the entire pathway from six isolated proteins. Proc Nat Acad Sci USA 75:3948–52
Stapp JM, Sjoelund V, Lassiter HA, Feldhoff RC, Feldhoff PW (2005) Recombinant rat IL-1beta and IL-6 synergistically enhance C3 mRNA levels and complement component C3 secretion by H-35 rat hepatoma cells. Cytokine 30:78–85 Storey JD, Tibshirani R (2003) Statistical significance for genomewide studies. Proc Nat Acad Sci USA 100:9440–5 Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Nat Acad Sci USA 102:15545–50 Toichi E, Lu KQ, Swick AR, McCormick TS, Cooper KD (2008) Skininfiltrating monocytes/macrophages migrate to draining lymph nodes and produce IL-10 after contact sensitizer exposure to UV-irradiated skin. J Invest Dermatol 128:2705–15 Wettenhall JM, Smyth GK (2004) limmaGUI: a graphical user interface for linear modeling of microarray data. Bioinformatics 20:3705–6
Schwarz A, Grabbe S, Aragane Y, Sandkuhl K, Riemann H, Luger TA et al. (1996) Interleukin-12 prevents ultraviolet B-induced local immunosuppression and overcomes UVB-induced tolerance. J Invest Dermatol 106:1187–91
Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J et al. (2002) Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res 30:e15
Shen J, Bao S, Reeve VE (1999) Modulation of IL-10, IL-12 and IFN-gamma in the Epidermis of Hairless Mice by UVA (320–400 nm) and UVB (280–320 nm) Radiation. J Invest Dermatol 113:1059–64
Yoshida Y, Kang K, Berger M, Chen G, Gilliam AC, Moser A et al. (1998) Monocyte induction of IL-10 and down-regulation of IL-12 by iC3b deposited in ultraviolet-exposed human skin. J Immunol 161:5873–9
www.jidonline.org 2701