ORIGINAL ARTICLE

UV Radiation-Induced Immunosuppression Is Greater in Men and Prevented by Topical Nicotinamide

Diona L. Damian1, Clare R.S. Patterson1, Michael Stapelberg1, Joohong Park1, Ross St.C. Barnetson1 and Gary M. Halliday1
UV radiation-induced immunosuppression augments cutaneous carcinogenesis. The incidence of skin cancer continues to increase despite increased use of sunscreens, which are less effective at preventing immunosuppression than sunburn. Using the Mantoux reaction as a model of skin immunity, we investigated the effects of solar-simulated (ss) UV and its component UVA and UVB wavebands and tested the ability of topical nicotinamide to protect from UV-induced immunosuppression. Healthy, Mantoux-positive volunteers were UV-irradiated on their backs, with 5% nicotinamide or vehicle applied to different sites in a randomized, double-blinded manner. Subsequent Mantoux testing at irradiated and adjacent unirradiated sites enabled measurement of UV-induced immunosuppression with and without nicotinamide. Suberythemal ssUV caused significant immunosuppression, although component UVB and UVA doses delivered independently did not. Men were immunosuppressed by ssUV doses three times lower than those required to immunosuppress women. This may be an important cause of the higher skin cancer incidence and mortality observed in men. Topical nicotinamide prevented immunosuppression, with gene chip microarrays suggesting that the mechanisms of protection may include alterations in complement, energy metabolism and apoptosis pathways. Nicotinamide is a safe and inexpensive compound that could be added to sunscreens or after-sun lotions to improve protection from immunosuppression. JID JOURNAL CLUB ARTICLE: For questions, answers, and open discussion about this article, please go to http://network.nature.com/group/jidclub
Journal of Investigative Dermatology (2008) 128, 447454; doi:10.1038/sj.jid.5701058; published online 20 September 2007

INTRODUCTION Skin cancer is the most common form of malignancy in Caucasian populations (Diepgen and Mahler, 2002). Immunosuppression enhances the development of skin cancers, as seen most dramatically in transplant recipients whose skin cancer risk is severalfold higher than that of the immunocompetent population (Bordea et al., 2004). UV radiation in sunlight has profound immunosuppressive effects on the skin (Damian et al., 2001), which play a central role in cutaneous carcinogenesis (Noonan et al., 2003; Ullrich, 2005). The exact mechanisms of UV-induced immunosuppression, and the relative contribution of UVB (290320 nm) and UVA (320400 nm), are as yet unclear. Various animal and human models of skin immunity have been used to demonstrate UV-induced immunosuppression
1

Department of Dermatology, Melanoma and Skin Cancer Research Institute, Sydney Cancer Centre, University of Sydney, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia Correspondence: Professor Diona L. Damian, Department of Dermatology, Gloucester House Level 3, Royal Prince Alfred Hospital, Missenden Rd, Camperdown, New South Wales 2050, Australia. E-mail: diona.damian@email.cs.nsw.gov.au Abbreviations: cdmac, cadmium induces DNA synthesis and proliferation in macrophages; EI, erythema index; GSEA, gene set enrichment analysis; hdac, histone deacetylase; MED, minimal erythema dose; PARP, poly-ADPribose polymerase; PPD, purified protein derivative; ss, solar-simulated

Received 6 December 2006; revised 14 June 2007; accepted 10 July 2007; published online 20 September 2007

in vivo. UV irradiation can impair both induction (primary sensitization) and elicitation of immune responses to cutaneous allergens (Damian et al., 2001; Wolf et al., 2003). Systemic UV-induced immunosuppression can be measured following sensitization at a site distant from the irradiated field, whereas local immunosuppression is observed when antigen is applied at the site of UV exposure. Mantoux testing with tuberculin purified protein derivative (PPD) is used in clinical practice to assess tuberculosis immunity and exposure status. Mantoux-positive subjects, who have been either exposed to tuberculosis or vaccinated with Bacille Calmette-Guerin, develop localized induration and erythema 4872 hours after intradermal injection of PPD. The diameter of induration can be measured clinically using the pen method whereby a ballpoint pen is used to mark the outer aspects of the Mantoux response (Bouros et al., 1991). Erythema at reaction sites, assessed via reflectance spectrometry, provides a more sensitive measure of Mantoux intensity, which correlates well with the diameter of induration (Damian and Halliday, 2002). Men have both a higher incidence of and mortality from skin cancer (Foote et al., 2001; Molife et al., 2001) but the effect of gender on susceptibility to UV-induced immunosuppression in humans is unknown. Immunosuppression occurs with UV doses well below the sunburn threshold, and sunscreens may be less effective at preventing immunosuppression than sunburn (Damian et al., 1999; Poon et al., 2003); improved protection from the immune effects of UV is
www.jidonline.org 447

& 2007 The Society for Investigative Dermatology

DL Damian et al.
UV Immunosuppression, Gender, and nicotinamide

thus likely to reduce the incidence of malignant and premalignant skin lesions. Nicotinamide, the active form of vitamin B3, is used clinically in the treatment of a variety of inflammatory skin disorders including bullous pemphigoid, acne, and rosacea (Niren, 2006). Although the exact mechanism of action in these settings is unknown, nicotinamide is a known endogenous inhibitor of the nuclear enzyme poly-ADP-ribose polymerase (PARP), which regulates the expression of immunomodulatory proteins such as inducible nitric oxide synthase, intercellular adhesion molecule 1, major histocompatibility complex class II, and NF-kB (Virag and Szabo, 2002). Nicotinamide has also been shown to prevent UV-induced immunosuppression and carcinogenesis in mice (Gensler, 1997, 1999) but its effects on UV immunosuppression in humans have not yet been investigated. We have previously shown that UV exposure attenuates Mantoux reactions in Bacille Calmette-Guerin-vaccinated volunteers (Damian et al., 1998; Friedmann et al., 2004), and now report the effects of gender and nicotinamide on UV-induced suppression of the Mantoux reaction in humans. RESULTS
Subjects

Mantoux-induced erythema increases in proportion to PPD dose

Although all subjects had initial Mantoux tests performed the week before UV irradiation, only 12 volunteers from various study groups were initially tested with the PPD dose combination of 1, 2, and 5 U (Figure S1). The erythema index (EI) of subsequent Mantoux reactions was significantly correlated with PPD dose, as shown by a repeated measures linear regression (r 0.705, n 36, Po0.001).
UV irradiation did not cause systemic immunosuppression

Of 75 people who started the protocol, 5 did not complete due to unrelated reasons. None of the subjects who completed the study were excluded from the results and none suffered significant adverse effects from the study. Table 1 shows the similar characteristics of volunteers who completed the various studies.

The initial, pre-irradiation Mantoux tests (lateral lower back) were compared with those of unirradiated nicotinamideand vehicle-treated controls adjacent to the irradiated areas (medial lower back). Subjects were excluded from this analysis if initial PPD doses were different from those subsequently used. The mean EI of the initial Mantoux tests did not significantly differ from subsequent control tests in any of the groups. Therefore, solar-simulated (ss) UV, UVB, or UVA exposure did not cause systemic suppression of Mantoux reactions. Nicotinamide in the absence of UV had no immune effects in any study, as the EI of unirradiated control sites treated with nicotinamide did not differ from the EI of unirradiated initial or vehicle-treated sites. Similarly, there was no significant difference in the unirradiated control EI of men and women.
Topical 5% nicotinamide prevents ssUV-induced immunosuppression but not sunburn

The effect of 5% nicotinamide applied before ssUV exposure in 20 volunteers (10 men, 10 women) is shown in Figure 1a.

Table 1. Baseline characteristics of volunteers

Nicotinamide and ssUV
Characteristic Age (years) (range) MED J/cm ssUV (range)
2

Nicotinamide before and after combined1

Females (n=18) 41 (2563) 2.15 (1.113.75) 79 (123 to 36) Males (n=18) 32 (2356) 2.05 (0.743.7) 60 (129 to 5) UVA only (n=15; 7 men) 30 (2054) 2.64 (0.797.95) 35 (130 to 148) UVB only (n=15; 8 men) 31 (2050) not determined 40 (130 to 25) Microarray (n=5; 4 men) 33 (2150) 4.22 (2.697.68) 58 (105 to 14)

Before (n=20) 35 (2263) 2.15 (1.113.01) 76 (129 to 5)

After (n=20) 38 (2362) 2.06 (0.743.72) 59 (117 to 20)

Skin color Melanin index in relative units (range) Skin type I/II/III/IV2 Current smokers Number of subjects Previous NMSC3
1 2

5:9:6:0

5:10:5:0

4:11:3:0

6:6:6:0

6:4:4:1

3:5:7:0

2:2:0:1

2 0

2 2

2 1

2 1

3 1

2 0

2 0

MED, minimal erythema dose; NMSC, non-melanoma skin cancer; ssUV, solar-simulated UV. Two women and two men participated in both studies therefore one set of each of their results was excluded from analysis. Fitzpatricks skin types (Fitzpatrick, 1998): I, constitutive skin color pale, does not tan; II, pale, tans with difficulty; III, pale, tans readily; IV, light brown, tans profusely. 3 No patients had previous melanoma.

448

Journal of Investigative Dermatology (2008), Volume 128

DL Damian et al.
UV Immunosuppression, Gender, and nicotinamide

Exposure of vehicle-treated skin to 1.48 or 2.22 J/cm2 ssUV for 3 consecutive days significantly suppressed Mantoux reactions by 19.4% (mean DEI 13.8, SEM 3.7; Po0.01) and 27.7% (mean DEI 19.8, SEM 2.8; Po0.001), respectively. When irradiated nicotinamide-treated sites were compared with unirradiated nicotinamide-treated control sites, significant immunosuppression no longer occurred. Hence, 5% nicotinamide applied before exposure prevented ssUV-induced immunosuppression. This protective effect of nicotinamide treatment was confirmed by repeated measures analysis of variance (Po0.0001). The effect of nicotinamide applied immediately after irradiation in the second group of 20 volunteers (10 men, 10 women) is shown in Figure 1b. There were no significant differences in baseline characteristics between the two groups. Three consecutive daily exposures to 0.74 J/cm2 suppressed Mantoux reactions by 24.8% (mean DEI 16.3, SEM 6.2; Po0.05) and exposures to 1.48 and 2.22 J/cm2 caused suppression of 29.1% (mean DEI 19.2, SEM 6.2; Po0.05) and 43.8% (mean DEI 28.8, SEM 5.7; Po0.001), respectively. At each of the three fixed doses, significant immunosuppression was prevented by application of 5% nicotinamide after irradiation. The protective effect of nicotinamide treatment was again confirmed by repeated measures analysis of variance (Po0.0001).
40
Vehicle

The minimal erythema dose (MED) was unchanged by nicotinamide applied before or after ssUV exposure (Table 2; paired two-tailed Students t-test). Mean erythema (EI) at vehicle-treated MED sites was not significantly different to the EI at the nicotinamide-treated site exposed to the same UV dose whether nicotinamide was applied before or after ssUV exposure. Hence, nicotinamide did not prevent sunburn. Spectrophotometry (Ultrospec 2100 pro UV/Visible Spectrophotometer, Biochrom Ltd, Cambridge, UK) of 5% nicotinamide in vehicle confirmed negligible absorption between 290 and 400 nm.
Men are more susceptible to ssUV-induced immunosuppression than women

30

Nicotinamide

***
20

**

10

0 0.74 1.48 ssUV (J/cm2) 2.22

40
Vehicle

Immunosuppression (EI)

*** *

Nicotinamide

30

To analyze the effects of gender, subjects from both ssUV studies, in which the irradiation protocols were identical, were combined. When the studies were analyzed separately, the results were similar and therefore we present the results of the combined studies only. Four subjects participated in both studies. One set of their results was therefore excluded, based on matching MEDs as evenly as possible for male and female volunteers. Four of the female volunteers were taking oral contraceptives, 5 were postmenopausal, and 9 were premenopausal. Because of the small numbers of women in each group, it was not possible to correlate hormonal status with susceptibility to immunosuppression. When ssUV-induced immunosuppression at vehicle-treated sites was compared between men and women, men were significantly more immunosuppressed than women (repeated measures analysis of variance P 0.0224) and the threshold dose for immunosuppression was at least three times lower in men than women (Figure 2). Among male volunteers, significant immunosuppression was observed at all doses of ssUV, whereas female volunteers were only significantly immunosuppressed by the highest dose. Although both men and women were significantly immunosuppressed by the highest ssUV dose, immunosuppression was 40% greater in men. Nicotinamide prevented suppression of Mantoux responses in both men and women (results not shown). In this combined group of volunteers, immunosuppression following exposure to the two highest ssUV doses correlated to MED (r 0.587, Po0.001 and r 0.353, Po0.05, respectively, using linear regression analysis); subjects who were more easily sunburned were also more susceptible to UV immunosuppression.
Component UVB and UVA doses were not independently immunosuppressive

Immunosuppression (EI)

*
20

10

0 0.74 1.48 ssUV (J/cm2) 2.22

Figure 1. Topical 5% nicotinamide protects from immunosuppression when applied before (a) or after (b) ssUV exposure. Results are presented as mean SEM; *Po0.05, **Po0.01, and ***Po0.001 (compared with unirradiated site, paired two-tailed Students t-test; Bonferroni correction applied; each n 20).

The effect of irradiation with UVA or UVB alone on the Mantoux response was investigated in separate groups of 15 volunteers. Subjects were exposed to UVA or UVB on 3 consecutive days at doses equivalent to the UVA and UVB components of the previously used ssUV doses (0.75, 1.50, and 2.24 J/cm2 of UVA and 0.075, 0.150, and 0.225 J/cm2 of UVB). Comparison of Mantoux-induced erythema at irradiated and unirradiated sites showed no significant immunosuppression at any dose of UVB or UVA. Nicotinamide had no immune effects in this experiment (Figure S2).
www.jidonline.org 449

DL Damian et al.
UV Immunosuppression, Gender, and nicotinamide

Table 2. Application of 5% nicotinamide before or after ssUV exposure has no effect on MED
Nicotinamide application
Before ssUV MED EI at MED MED EI at MED

Vehicle (Jcm2)
2.1570.11 8576.82 2.0670.19 8976.85

Nicotinamide (J/cm2)
2.1270.12 7975.38 2.1270.21 8576.68

P-value
0.61 0.214 0.46 0.148

After ssUV

EI, erythema index; MED, minimal erythema dose; ssUV, solar-simulated UV. In 20 volunteers, nicotinamide was applied 15 minutes before MED testing on one side of the back whereas at the same time base lotion was applied on the other side of the back. There was no significant difference between the mean MED (shown7SEM), nor in the mean EI at skin irradiated with one vehicletreated MED (EI at MED) between base lotion-treated sites and nicotinamide-treated sites. In 20 volunteers, nicotinamide was applied immediately after MED testing on one side of the back and base lotion was applied on the other side of the back. Again, there was no significant difference between the mean MED between base lotion-treated sites and nicotinamide-treated sites.

40 Immunosuppression (EI)

Men Women

**

** *

30

20

10

0 0.74 1.48 ssUV (J/cm2) 2.22

Figure 2. Men are more immunosuppressed by ssUV than women. Results are presented as mean7SEM; *Po0.01 and **Po0.001 (compared with unirradiated site, paired two-tailed Students t-test; Bonferroni correction applied; n 36).

Genes involved in complement, energy metabolism and apoptosis pathways are differentially regulated in vehicle- and nicotinamide-treated irradiated skin

Six healthy volunteers (five men and one woman) were recruited for ssUV irradiation, biopsy, and microarray analysis. In each volunteer, two microarrays were compared and analyzed (vehicle-treated unirradiated vs vehicle-treated irradiated sites; and nicotinamide-treated unirradiated vs nicotinamide-treated irradiated sites). Results from one (male) volunteer were excluded because of technical inability to extract the microarray data due to inadequate hybridization. Gene set enrichment analysis (GSEA) (false discovery rate p0.25, Po0.05) was applied to identify differentially regulated functional gene sets. In vehicle-treated skin, no significantly enriched pathways were upregulated by ssUV irradiation. Instead, a number of pathways were downregulated, including those relating to apoptosis (10 pathways), energy metabolism (5 pathways), and immune function including complement (4 pathways) (Tables S1 and S2). Highly overlapping genes in apoptosis-related pathways were protein kinase C (PRKCA), cytochrome c (CYCS), insulin-like growth factor 1 receptor (IGF1R), BAD, BCL2, AKT1, caspase 3 (CASP3), HRAS, and TP53. In energy-related pathways, the leading edge subset of genes in the electron transport chain pathway (i.e., the genes most responsible for
450 Journal of Investigative Dermatology (2008), Volume 128

the enrichment signal) were mitochondrial ATP synthase (ATP5A1), succinate dehydrogenase complex subunits a and b (SDHA, SDHB), and cytochrome c (CYCS), which are key genes in energy production. In the citrate pathway, the nicotinamide adenine dinucleotide (NAD )-dependent enzymes isocitrate dehydrogenase (IDH1, IDH3B) and succinate-CoA ligase (SUCLG1, SUCLA2) were also downregulated. The leading edge subset of genes in complement pathways were C1Q, C1R, C1S, C4b, and C3. In nicotinamide-treated skin, significant downregulation of these pathways was no longer observed. Hence, nicotinamide inhibited ssUV-induced dysregulation of these genes. In nicotinamide-treated irradiated skin, there was upregulation of the cadmium induces DNA synthesis and proliferation in macrophages (cdmac) pathway with leading edge subset genes comprising PLCB1, PRKCA, PRKCB1, HRAS, CUZD1, MAP2K1 (MEK1 or MKK1), MAPK3 (ERK1), TNF, and RELA. The histone deacetylase (hdac) pathway (involved in calcium regulation; McKinsey et al., 2000) was downregulated by ssUV in the presence of both vehicle and nicotinamide (leading edge genes in nicotinamide skin AKT1, CALM2, CAMK1, HDAC5, MEF2A, MEF2C, and NFATC2). Pathways for arginine and proline metabolism were also downregulated in nicotinamide-treated irradiated skin. The genes in these pathways, involved in energy metabolism such as providing precursors to the citrate cycle (Davis, 1986) were NOS3, GOT1, CKM, ASS, SMS, AMD1, CPS1, AOC3, OAT, and CKMT2. DISCUSSION Topical application of nicotinamide prevented UV-induced immunosuppression when applied either before or after UV exposure. The ssUV used in this study closely approximated the spectrum of natural sunlight, and was delivered at doses equivalent to those readily encountered on a daily basis. ssUV equivalent to 8 minutes of sun exposure (COLIPA, 1994) on 3 consecutive days caused more than 40% immunosuppression, and immunosuppression of 25% was caused by UV doses well below the average MED. Hence, significant immunosuppression occurred in the absence of sunburn. Immunosuppression was prevented at all doses tested, with no adverse effects, when nicotinamide was applied

DL Damian et al.
UV Immunosuppression, Gender, and nicotinamide

before or after UV exposure. The protective effects of nicotinamide therefore do not result from sunscreening activity. This was confirmed by the lack of effect on MED and negligible UV absorbance by spectrophotometry. Although the baseline characteristics of the two ssUV groups were similar, the second group showed higher levels of immunosuppression. This is likely to reflect normal interindividual variation in susceptibility to UV-induced immunosuppression (Damian et al., 1998; Kelly et al., 2000). Individuals who are more susceptible to UV-induced immunosuppression are at a greater risk of both melanoma and non-melanoma skin cancer (Yoshikawa et al., 1990); however, identifying these individuals is difficult as there are no clinically recognizable signs of cutaneous immunosuppression. As previously reported by Kelly et al. (2000), we found that immunosuppression was significantly correlated to sunburn threshold; paler-skinned individuals were more susceptible to not only sunburn, but also UV immunosuppression. We analyzed our microarray data using the innovative, supervised computational method GSEA (Subramanian et al., 2005). Compared with unsupervised methods of analysis such as gene clustering and grouping based solely on gene expression, the pathway-oriented approach of GSEA is better able to detect relatively subtle differences between sample classes and examines entire pathways rather than single genes. For example, differential activity can be identified with a mean difference in gene expression of only 20% (Mootha et al., 2003). Leading edge analysis, showing the gene sets most responsible for the enrichment signal, provided additional clues to identify the most critical subpathways. Analysis of ssUV-regulated gene expression in vehicletreated human skin revealed several downregulated enriched biological pathways mainly related to complement, apoptosis, and energy metabolism. Complement has been shown to be involved early in the elicitation phase of DTH reactions in mice (Tsuji et al., 1997), and is also known to be modulated by UV irradiation (Hammerberg et al., 1998). In humans exposed to 3 MED of UVB, C3 activation was observed 312 hours after irradiation but not at later time points (Terui et al., 2001). In our study, there was downregulation of complement pathways in vehicle-treated skin, prevented by topical nicotinamide. It is possible that ssUV may have upregulated complement in vehicle-treated skin if biopsies were performed at an earlier time point, or in response to higher UV doses. UV radiation-induced apoptosis is a protective response to cellular injury that removes potentially genetically damaged cells from the skin through the formation of sunburn cells (Melnikova and Ananthaswamy, 2005), although imunohistochemistry of human skin irradiated with very low doses of ssUV (0.51 MED; within the range used in these studies) may not show any increase in apoptotic cell counts (Murphy et al., 2002). Our gene array analysis of vehicle-treated skin showed ssUV-induced downregulation of the IGF-1R signaling pathway including the antiapoptotic genes IGF1R and AKT1. Also downregulated in vehicle-treated skin was the

programmed cell death pathway, including the antiapoptotic gene BCL-2. Hence, ssUV exposure had proapoptotic effects in our transcriptional study. This is consistent with the results of Enk et al. (2006), who used unsupervised gene clustering and functional grouping to report microarray findings in human skin irradiated in vivo with supraerythemal UVB doses (4 MED) . In contrast with our findings, Enk et al. (2006) found that UVB differentially regulated several pathways including those relating to S100 protein, serine protease inhibitors, and extracellular matrix factors such as metalloproteinases as well as apoptosis. This may reflect the much higher UV dose used. Nicotinamide prevented all of these effects. The apoptotic genes BCL2, TP53, IGF1R, PRKCA, and AKT1, comprising the leading edge subset of genes in the Tel pathway, are involved in regulation of telomerase activity, which enables the addition of telomeric repeats to telomeres and is thought to be involved in the early stages of cutaneous carcinogenesis (Ueda, 2000). UV irradiation increases the rate of telomere shortening in skin cells (Kosmadaki and Gilchrest, 2004), consistent with our findings that ssUV downregulated the telomerase pathway, and suggesting a protective effect of nicotinamide on this pathway. Our observed downregulation of genes for energy production in several pathways in vehicle-treated, irradiated skin is consistent with ssUV-induced cellular energy depletion (Jacobsen et al., 2001). This pathway was no longer downregulated in nicotinamide-treated skin, suggesting that nicotinamide induces changes in energy metabolism as well as apoptosis. Intracellular nicotinamide is metabolized to NAD , the essential coenzyme in ATP production and the sole substrate of the nuclear enzyme PARP (Hageman and Stierum, 2001). PARP is involved in multiple cellular functions, including regulation of p53 expression (Wieler et al., 2003), and is rapidly activated in response to DNA damage. Although modest PARP activation enhances DNA repair, overactivation can cause cell death. Nicotinamide is an endogenous PARP inhibitor (Hageman and Stierum, 2001); hence PARP inhibition by nicotinamide may have reduced cellular damage. Like Enk et al. (2006), we found that ssUV downregulated p53, whereas in the presence of nicotinamide, p53 downregulation was prevented. Nicotinamide has also been shown to alter p53 regulation directly, independent of PARP (McLure et al., 2004). The gene expression profile of nicotinamide-treated, irradiated skin showed upregulation of the enriched pathway cdmac, which includes genes that overlap with UVA-induced ERK activation (protein kinase C; PRKCA, PRKCB1, phospholipase C, RAS and the extracellular signal-regulated kinases 1 and 2, ERK, or MAPK3) (He et al., 2004). Direct NADH insertion is known to increase intracellular calcium (Kaplin et al., 1996). Phospholipase activation and calcium increase are required for ERK activation, which plays an important role in cell proliferation and differentiation. It is possible that a nicotinamide-induced increase in intracellular calcium might be an essential upstream signal regulator for the ERK activation (He et al., 2004). For the first time, we have shown that ssUV-induced immunosuppression is greater in men. Only the highest ssUV
www.jidonline.org 451

DL Damian et al.
UV Immunosuppression, Gender, and nicotinamide

dose caused significant immunosuppression in female volunteers. In contrast, men were significantly immunosuppressed by ssUV doses three times less than the sunburn threshold, and were more immunosuppressed than women at all doses tested. Male and female volunteers were well matched for MED and skin color and were of similar ages. The incidence of smoking, which not only suppresses systemic immunity but is also associated with cutaneous squamous-cell carcinoma (Freiman et al., 2004), was the same in men and women as was previous history of skin cancer. Hence, these factors were not a cause of greater immunosuppression in men. Similar findings have been reported in mice, where UVBinduced immunosuppression was significantly greater in male than in female mice and this was reversed by injecting males with 17-b-estradiol before irradiation (Hiramoto et al., 2004). In female mice, ssUV-induced immunosuppression was exacerbated by topical application of an estrogen receptor antagonist and attenuated by topical 17-b-estradiol (Widyarini et al., 2006). Immune cells such as lymphocytes and macrophages are known to bear sex steroid receptors, which implies that estrogen may influence cytokine production by these cells (Kovacs and Messingham, 2002). Male gender is an independent risk factor for immunosuppression following surgical procedures (Wichmann et al., 2003), and murine studies suggest that testosterone is responsible (Wichmann et al., 1996). Our findings may explain the higher female incidence of cutaneous disorders such as polymorphous light eruption (Mastalier et al., 1998), as well as the greater skin cancer incidence and mortality observed in men (Foote et al., 2001; Molife et al., 2001). In our study, low doses of ssUV (UVA UVB) were immunosuppressive but when the UVA and UVB components of various ssUV doses were studied individually, we found that neither UVA alone nor UVB alone was immunosuppressive. This suggests that at the very low UV doses used here, an interaction between the effects of UVA and UVB leads to immunosuppression, as has been described in studies of ssUV-induced suppression of contact hypersensitivity to nickel in nickel-allergic volunteers (Poon et al., 2005). Immunosuppressive effects may have been detected with higher individual component doses of UVA and UVB. In conclusion, we have shown that a likely interactive effect of UVB and UVA produces ssUV-induced immunosuppression at doses well below the sunburn threshold, and that men are far more susceptible to this immunosuppression than women. Nicotinamide appeared to mediate its immuno protective effects by preventing ssUV-induced changes in complement, apoptosis, and cellular energy pathways, consistent with its known roles as a PARP inhibitor, ATP precursor, and anti-inflammatory agent. Nicotinamide also influenced ssUV-induced transcriptional regulation of genes in the cdmac and arginine catabolic pathways. Whether this novel finding is related to the protective effects of nicotinamide on the skin requires further investigation with functional studies. Skin cancer incidence is rising and sunscreens can be improved by ensuring protection from immunosuppression as well as sunburn.
452 Journal of Investigative Dermatology (2008), Volume 128

MATERIALS AND METHODS

Participants
Healthy Mantoux-positive volunteers who had previously been vaccinated with BCG were recruited and were required to cease any vitamin supplementation and avoid sun exposure on their back at least 4 weeks before and during the study. The study was conducted according to the Declaration of Helsinki Principles, and approval was obtained from the Sydney University and Central Sydney Area Health Service Ethics Committees. All volunteers provided written informed consent before inclusion.

Study techniques
The UV source, described in detail previously (Damian and Halliday, 2002), was a 1,000 W xenon arc solar simulator (Oriel, Stratford, CT), filtered to attenuate visible and infrared wavelengths, remove UVC (o290 nm), and modify the spectral output so that it simulated sunlight (ssUV), or comprised predominantly UVB or UVA only (Damian and Halliday, 2002). Spectra were measured at 1 nm intervals with an OL754 spectroradiometer (Optronics, Orlando, FL), which had been calibrated with standard lamps. An IL-1350 broadband radiometer with interchangeable UVA (SED 038) and UVB (SED 240) detectors (International Light, Newburyport, MA) was calibrated against the source using the spectroradiometer and this was used to enable rapid measurement of irradiance before each irradiation. One week before irradiation, volunteers were tested with three different concentrations of tuberculin PPD (CSL Limited, Parkville, Victoria, Australia), diluted in 0.9% normal saline to a volume of 0.05 ml and injected intradermally on the lower back to confirm positivity and allow selection of the most appropriate PPD dose for further testing in each volunteer. Different PPD dose ranges were used depending on the subjects previous reactivity. The intensity of subsequent Mantoux reactions was assessed 72 hours after injection using the pen method for measuring Mantoux diameter (Bouros et al., 1991). Only participants with a Mantoux diameter X4 mm were eligible for the study. Mantoux reactions were also measured with a reflectance erythema meter (Diastron, Hampshire, UK), which we have previously found to provide a more sensitive and reliable measure of changes in skin immunity than diameter (Damian et al., 1998; Friedmann et al., 2004). Erythema induced by PPD was calculated as the difference between the EI of the Mantoux response and the EI of adjacent skin. Erythema meter readings were taken in triplicate and an average EI for each site was recorded. Nicotinamide (Fluka Chemie AG, Buchs, Switzerland) was prepared in 1:2:1 propylene glycol, ethanol, and distilled water vehicle at a concentration of 5%. The clear, colorless solutions, prepared fortnightly in our laboratory, were applied (2 ml/ cm2) topically to test sites on the mid- to lower back in a randomized, double-blinded manner. Subjects were asked to avoid washing the test sites for at least 8 hours after each application.

Study design
The week after initial Mantoux testing, separate 2.0 2.5 cm areas on each side of the lower back were irradiated with one of three fixed doses of ssUV (0.74, 1.48, and 2.22 J/cm2) daily for 3 consecutive days. Adjacent, unirradiated sites on each side served as immunologically intact controls. Nicotinamide lotion was applied to test sites on one side of the back, and vehicle to the opposite side,

DL Damian et al.
UV Immunosuppression, Gender, and nicotinamide

15 minutes before or immediately after each ssUV exposure in separate studies. Mantoux tests were performed at irradiated and unirradiated sites immediately after the final irradiation and measured 72 hours later. Each volunteers MED or sunburn susceptibility was determined by exposing ten 1 cm squares on the upper back to graded ssUV doses and noting the dose that caused threshold erythema 24 hours later. The effects of nicotinamide and vehicle on MED were assessed in all subjects both visually and with the erythema meter. A separate melanin index function in the erythema meter was used to measure the skin color of the mid-back. Additional groups of Mantoux-positive volunteers were recruited to assess the immune effects of UVB and UVA doses approximating those contained within the ssUV doses, with nicotinamide applied to one side of the back and vehicle to the other immediately after irradiation as described above.

given level what proportion of them are likely to be false-positive findings (Storey and Tibshirani, 2003).

Statistical analysis
Immunosuppression at each site (DEI) was calculated as the difference between the EI of unirradiated control sites and the EI of irradiated test sites. Statistical comparisons were made using paired two-tailed Students t-tests, with DEI considered significant if Po0.05 following Bonferroni correction. All data sets in the immunosuppression and MED experiments showed normal distribution (KolmogorovSmirnov test). Results are presented as mean7SEM. We analyzed intervention and immunosuppression using repeated measures analysis of variance. Correlations were performed using linear regression analysis or repeated measures linear regression (a random intercept linear mixed model fitted using STATA v 9.2) as stated.
CONFLICT OF INTEREST
The authors state no conflict of interest.

Microarray study
To investigate possible mechanisms of the immune effects of ssUV and nicotinamide, six volunteers were irradiated with a single, fixed dose of ssUV (2.22 J/cm2; B0.5 average MED of this group) onto two areas of their lower backs. These sites were treated with nicotinamide or vehicle immediately after irradiation, as were two adjacent, unirradiated sites. Nicotinamide was applied after irradiation to exclude completely any theoretical possibility of a sunscreening effect of nicotinamide. Twenty-four hours later, a 3 mm punch biopsy was taken from each of the four sites. Total RNA was extracted using TRIzol reagent (Gibco Invitrogen Life Technologies, Carlsbad, CA), purified and DNase-treated using the RNeasy Fibrous Tissue Mini Kit (Qiagen Sciences, Valencia, CA), and amplified using the MessageAmp II aRNA Kit (Ambion Inc., Austin, TX). Total RNA (6070 mg) was then reverse-transcribed into cDNA with direct incorporation of cyanine 3-dCTP and cyanine 5-dCTP fluorescent dyes (Perkin Elmer Life Sciences, Boston, MA). Following hybridization onto Compugen 19,000 human oligonucleotide microarray slides (Adelaide Microarray Facility, The University of Adelaide, Australia), microarrays were scanned using an Axon scanner (Axon Instruments, Sunnyvale, CA) with GenePix Pro 5.0 software. Preprocessing and normalization were applied to all microarray expression data before the assessment of differential gene expression, using GSEA methodology (Subramanian et al., 2005). To input microarray data into GSEA, the accession numbers from the microarrays of normalized treated groups were mapped to corresponding gene symbols using the EASE software in the Database for Annotation, Visualization, and Integration Discovery (DAVID) (Dennis et al., 2003; Hosack et al., 2003). GSEA was performed as previously described using a C2 human-specific gene set library containing 522 gene sets 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 estimation using 1,000 random permutations. We chose pathways with a significant enrichment score (Po0.05) that had been adjusted for multiple comparisons using the false discovery rate (qo0.25) for further interpretation. The false discovery rate is a quantity that describes for a set of tests that are deemed significant at or below a

ACKNOWLEDGMENTS
We are grateful to our participating volunteers, and to Drs Lesley Francis and Patrick FitzGerald for statistical advice and analysis. This study was funded by the Cancer Institute New South Wales, the University of Sydney Cancer Research Fund, and the Dermatology Research Foundation. Associate Professor Damian is supported by a Cancer Institute New South Wales Fellowship.

SUPPLEMENTARY MATERIAL
Table S1. Summary of GSEA results with FDRo0.25. Table S2. Genes regulated by ssUV and nicotinamide: rank in gene list. Figure S1. Mantoux-induced erythema index (EI) increases with the dose of tuberculin purified protein derivative (PPD). Figure S2. Low-dose UVB (A) or UVA (B) did not suppress Mantoux-induced erythema.

REFERENCES
Bordea C, Wojnarowska F, Millard PR, Doll H, Welsh K, Morris PJ (2004) Skin cancers in renal transplant patients occur more frequently than previously recognised in a temperate climate. Transplantation 77:5749 Bouros D, Zeros G, Panaretos C, Vassilatos C, Siafakas N (1991) Palpation vs pen method for the measurement of skin tuberculin reaction (Mantoux test). Chest 99:4169 COLIPA (1994) COLIPA sun protection factor test method. The European Cosmetic, Toiletry and Perfumery Association, ref 94/289, 21 Damian DL, Barnetson RS, Halliday GM (1999) Measurement of in vivo sunscreen immune protection factors in humans. Photochem Photobiol 70:9105 Damian DL, Barnetson RS, Halliday GM (2001) Effects of low-dose ultraviolet radiation on in vivo human cutaneous recall responses. Australas J Dermatol 42:1617 Damian DL, Halliday GM, Taylor CA, Barnetson RS (1998) Ultraviolet radiation induced suppression of Mantoux reactions in humans. J Invest Dermatol 110:8247 Damian DL, Halliday GM (2002) Measurement of ultraviolet radiationinduced suppression of recall contact and delayed-type hypersensitivity in humans. Methods 28:3445 Davis RH (1986) Compartmental and regulatory mechanisms in the arginine pathways of Neurospora crassa and Saccharomyces cerevisiae. Microbiol Rev 50:280313 Dennis GJ, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC et al. (2003) DAVID: database for annotation, visualization and integrated discovery. Genome Biol 9:3

www.jidonline.org

453

DL Damian et al.
UV Immunosuppression, Gender, and nicotinamide

Diepgen TL, Mahler V (2002) The epidemiology of skin cancer. Br J Dermatol 146:16 Enk CD, Jacob-Hirsch J, Gal H, Verbovetski I, Amariglio N, Mevorach D et al. (2006) The UVB-induced gene expression profile of human epidermis in vivo is different from that of cultured keratinocytes. Oncogene 25:260114 Fitzpatrick TB (1998) The validity and practicality of sun-reactive skin types I through VI. Arch Dermatol 124:86971 Foote JA, Harris RB, Giuliano AR, Roe DJ, Moon TE, Cartmel B et al. (2001) Predictors for cutaneous basal- and squamous-cell carcinoma among actinically damaged adults. Int J Cancer 95:711 Freiman A, Bird G, Metelitsa AI, Barankin B, Lauzon GJ (2004) Cutaneous effects of smoking. J Cutan Med Surg 8:41523 Friedmann AC, Halliday GM, Barnetson RS, Reeve VE, Walker C, Patterson CR et al. (2004) The topical isoflavonoid NV-07a reduces solarsimulated UV-induced suppression of Mantoux reactions in humans. Photochem Photobiol 80:41621 Gensler HL (1997) Prevention of photoimmunosuppression and photocarcinogenesis by topical nicotinamide. Nutr Cancer 29:15762 Gensler HL, Williams T, Huang AC, Jacobsen EL (1999) Oral niacin prevents photocarcinogenesis and photoimmunosuppression in mice. Nutr Cancer 34:3641 Hageman GJ, Stierum RH (2001) Niacin, poly(ADP-ribose) polymerase-1 and genomic stability. Mutat Res 475:4556 Hammerberg C, Katiyar SK, Carroll MC, Cooper KD (1998) Activated complement component 3 (C3) is required for ultraviolet induction of immunosuppresion and antigenic tolerance. J Exp Med 187:11338 He Y-Y, Huang J-L, Chignell CF (2004) Delayed and sustained activation of extracellular signal-regulated kinase in human keratinocytes by UVA: implications in carcinogenesis. J Biol Chem 279:5386774 Hiramoto K, Tanaka H, Yanagihara N, Sato EF, Inoue M (2004) Effect of 17 beta-estradiol on immunosuppression induced by ultraviolet B irradiation. Arch Dermatol Res 295:30711 Hosack DA, Dennis GJ, Sherman BT, Lane HC, Lempicki RA (2003) Identifying biological themes within lists of genes with EASE. Genome Biol 4:R70 Jacobsen EL, Giacomoni PU, Roberts MJ, Wondrak GT, Jacobsen MK (2001) Optimizing the energy status of skin cells during solar radiation. J Photochem Photobiol B 63:1417 Kaplin AI, Snyder SH, Linden DJ (1996) Reduced nicotinamide adenine dinucleotide-selective stimulation of inositol 1,4,5 trisphosphate receptors mediates Hypoxic mobilization of calcium. J Neurosci 16:200211 Kelly DA, Young AR, McGregor JM, Seed PM, Potten CS, Walker SL (2000) Sensitivity to sunburn is associated with susceptibility to ultraviolet radiation-induced suppression of cutaneous cell-mediated immunity. J Exp Med 191:5616 Kosmadaki MG, Gilchrest BA (2004) The role of telomeres in skin aging/ photoaging. Micron 35:1559 Kovacs EJ, Messingham KA (2002) Influence of alcohol and gender on immune response. Alcohol Res Health 26:25763 Mastalier U, Kerl H, Wolf P (1998) Clinical, laboratory, phototest and phototherapy findings in polymorphic light eruptions: a retrospective study of 133 patients. Eur J Dermatol 8:5549 McKinsey TA, Zhang CL, Olson EN (2000) Activation of the myocyte enhancer factor-2 transcription factor by calcium calmodulin dependent protein kinase-stimulated binding of 1433 to histone deacetylase 5. Proc Natl Acad Sci USA 97:144005 McLure KG, Takagi M, Kastan MB (2004) NAD+ modulates p53 DNA binding specificity and function. Mol Cell Biol 24:995867 Melnikova VO, Ananthaswamy HN (2005) Cellular and molecular events leading to the development of skin cancer. Mutat Res 571:91106

Molife R, Lorigan P, MacNeil S (2001) Gender and survival in malignant melanoma. Cancer Treat Rev 27:2019 Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J et al. (2003) PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet 34:26773 Murphy M, Mabruk MJEMF, Lenane P, Liew A, Mccann P, Buckley A et al. (2002) The expression of p53, p21, Bax and induction of apoptosis in normal volunteers in response to different doses of ultraviolet radiation. Br J Dermatol 147:1107 Niren NM (2006) Pharmacologic doses of nicotinamide in the treatment of inflammatory skin conditions: a review. Cutis 11:116 Noonan FP, Muller HK, Fears TR, Kusewitt DF, Johnson TM, De Fabo EC (2003) Mice with genetically determined high susceptibility to ultraviolet (UV)-induced immunosuppression show enhanced UV carcinogenesis. J Invest Dermatol 121:117581 Poon TS, Barnetson RS, Halliday GM (2003) Prevention of immunosuppression by sunscreens in humans is unrelated to protection from erythema and dependent on protection from ultraviolet A in the face of constant ultraviolet B protection. J Invest Dermatol 121:18490 Poon TSC, Barnetson RS, Halliday GM (2005) Sunlight-induced immunosuppression in humans is initially because of UVB, then UVA, followed by interactive effects. J Invest Dermatol 125:8406 Storey JB, Tibshirani R (2003) Statistical methods for identifying differentially expressed genes in DNA microarrays. Methods Mol Biol 224:14957 Subramanian A, Tamayo P, Mootha MK, Mukherjee S, Ebert BL, Gillette MA et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 102:1554550 Terui T, Takahashi K, Funayama M, Terunuma A, Ozawa M, Sasai S et al. (2001) Occurrence of neutrophils and activated Th1 cells in UVBinduced erythema. Acta Derm Venereol 81:813 Tsuji RF, Geba GP, Wang Y, Kawamoto K, Matis LA, Askenase PW (1997) Required early complement activation in contact sensitivity with generation of local C5-dependent chemotactic activity, and late T cell interferon g: a possible initiating role of B cells. J Exp Med 186:101526 Ueda M (2000) Telomerase in cutaneous carcinogenesis. J Dermatol Sci 23:S3740 Ullrich SE (2005) Mechanisms underlying UV-induced immunosuppression. Mutat Res 571:185205 Virag L, Szabo C (2002) The therapeutic potential of poly(ADP-ribose) polymerase inhibitors. Pharmacol Rev 54:375429 Wichmann MW, Muller C, Meyer G, Adam M, Angele MK, Eisenmenger SJ et al. (2003) Different immune responses to abdominal surgery in men and women. Langenbecks Arch Surg 387:397401 Wichmann MW, Zellweger R, DeMaso CM, Ayala A, Chaudry IH (1996) Mechanism of immunosuppression in males following trauma-hemorrhage. Critical role of testosterone. Arch Surg 131:118692 Widyarini S, Domanski D, Painter N, Reeve VE (2006) Estrogen receptor signaling protects against immune suppression by UV radiation exposure. Proc Natl Acad Sci USA 103:1283742 Wieler S, Gagne J-P, Vaziri H, Poirier GG (2003) Poly(ADP-ribose) polymerase-1 is a positive regulator of the p53-mediated G1 arrest response following ionizing radiation. J Biol Chem 278:1891421 Wolf P, Hoffmenn C, Quehenberger F, Grinschgl S, Kerl H (2003) Immune protection factors of chemical sunscreens measured in the local contact hypersensitivity model in humans. J Invest Dermatol 121:10807 Yoshikawa T, Rae V, Bruins-Slot W, Van den Berg JW, Taylor JR, Streilein JW (1990) Susceptibility to effects of UVB radiation on induction of contact hypersensitivity as a risk factor for skin cancer in humans. J Invest Dermatol 95:5306

454

Journal of Investigative Dermatology (2008), Volume 128