See related commentary on pg 1756

ORIGINAL ARTICLE

Irradiation of Skin with Visible Light Induces Reactive Oxygen Species and Matrix-Degrading Enzymes
Frank Liebel1, Simarna Kaur1, Eduardo Ruvolo2, Nikiforos Kollias2 and Michael D. Southall1
Daily skin exposure to solar radiation causes cells to produce reactive oxygen species (ROS), which are a primary factor in skin damage. Although the contribution of the UV component to skin damage has been established, few studies have examined the effects of non-UV solar radiation on skin physiology. Solar radiation comprises o10% of UV, and thus the purpose of this study was to examine the physiological response of skin to visible light (400700 nm). Irradiation of human skin equivalents with visible light induced production of ROS, proinflammatory cytokines, and matrix metalloproteinase (MMP)-1 expression. Commercially available sunscreens were found to have minimal effects on reducing visible lightinduced ROS, suggesting that UVA/ UVB sunscreens do not protect the skin from visible lightinduced responses. Using clinical models to assess the generation of free radicals from oxidative stress, higher levels of free radical activity were found after visible light exposure. Pretreatment with a photostable UVA/UVB sunscreen containing an antioxidant combination significantly reduced the production of ROS, cytokines, and MMP expression in vitro, and decreased oxidative stress in human subjects after visible light irradiation. Taken together, these findings suggest that other portions of the solar spectrum aside from UV, particularly visible light, may also contribute to signs of premature photoaging in skin.
Journal of Investigative Dermatology (2012) 132, 19011907; doi:10.1038/jid.2011.476; published online 9 February 2012

INTRODUCTION The contribution of the UV (290400 nm) component of solar irradiation on the skin has been well studied; more recently, studies have begun to explore the effects of non-UVR on skin physiology. The spectral distribution of the solar energy at the sea level comprises roughly 37% of UVR (290400 nm), 44% of visible light (400700 nm), and 53% of infrared (IR) radiation (7001440 nm; Frederick et al., 1989). The effects of UVR are well documented. UVR from sunlight normally consists of three wavelength regions: UVC (which is absorbed by ozone in the atmosphere), UVB, and UVA. Energy from the shorter-wavelength UVB is absorbed in greater amounts by the epidermis and by keratinocyte DNA, compared with the energy from UVA, which penetrates more deeply into the dermal layers of the skin. Visible and IR light wavelengths penetrate deep into the dermis and have been
1

Preclinical Pharmacology, Johnson & Johnson Skin Research Center, CPPW, a Unit of Johnson & Johnson Consumer Companies, Skillman, New Jersey, USA and 2Measurement Sciences, Johnson & Johnson Skin Research Center, CPPW, a Unit of Johnson & Johnson Consumer Companies, Skillman, New Jersey, USA

Parts of this work have been presented in a poster format at the Society for Investigative Dermatology Annual Meeting, 2010. Correspondence: Michael D. Southall, Preclinical Pharmacology, Johnson & Johnson Skin Research Center, CPPW, a Unit of Johnson & Johnson Consumer Companies, 199 Grandview Road, Skillman, New Jersey 08558, USA. E-mail: msoutha@its.jnj.com Abbreviations: IR, infrared; MMP, matrix metalloproteinase; ROS, reactive oxygen species Received 14 April 2011; revised 1 September 2011; accepted 4 October 2011; published online 9 February 2012

thought to, following absorption, only produce heat. In contrast to the extensive research on the damaging effect of UV, few studies have looked at the effects of visible light on skin. It has been known for decades that reactive oxygen species (ROS) produced in the skin following UV irradiation are key mediators of oxidative damage to the skin. Cell damage from UV also occurs through peroxidation of membrane lipids via generation of lipid peroxides. UV irradiation results in the rapid depletion of several endogenous skin enzymes and antioxidants such as glutathione reductase and catalase, as well as glutathione, tocopherol, and ubiquinone. Exposure to UV has also been shown to induce proinflammatory cytokines, such as IL-1a, and matrix metalloproteinases (MMPs) in skin cells such as keratinocytes and fibroblasts (Wlaschek et al., 1994; Wan et al., 2001). Activation of MMPs leads to the breakdown of collagen, the major structural component of skin, and also inhibits new collagen synthesis. Induction of inflammatory cytokines by UV, leading to overall skin inflammation, is another significant contributor to the photoaging process in skin. In addition to UV, IR has also been shown to cause oxidative stress to skin; exposing human fibroblasts to IR led to increased ROS formation, and enhanced expression of MMPs (Schroeder et al., 2008). Additional studies exposed photoprotected sites of healthy human skin to solar relevant doses of IR radiation, and observed that the MMP-1 expression in the dermis, but not the epidermis, was upregulated in 80% of the tested individuals (Cho et al., 2008).
www.jidonline.org 1901

& 2012 The Society for Investigative Dermatology

F Liebel et al.
Visible Light Induces ROS, Inflammatory Cytokines, and MMPs in Skin

Although sunlight comprises up to 44% of visible light, few studies have sought to determine the effects of visible light on skin. Commercial sunscreens are designed to only block wavelengths up to 380 nm, and thus skin with topical sunscreens is not protected from the effects of visible light. In the current study, we examined the effect of visible light on the ROS, and MMP responses in skin in vitro. We report that visible light can induce significant ROS production, and this ROS mediates the release of proinflammatory cytokines and MMP expression. Skin is exposed to visible light for substantial durations of the day, and as skin contains several chromophores for visible light the cumulative effects of visible light could result in skin damage, which may contribute to premature skin aging. RESULTS
Exposure to visible light induces increases in epidermal ROS, cytokine, and MMP production

from epidermal equivalents. IL-1a release was increased 1.1-, 1.5-, and 2.5-fold after exposure to visible light doses of 65, 130, and 180 J cm2, respectively, with 6.0 J cm2 from a solar simulator increasing IL-1a release by 2.8-fold (Figure 1b). A similar effect was seen with release of IL-1 receptor antagonist, IL-6, GM-CSF, and IL-8 (data not shown). In contrast, UV was found to increase the release of tumor necrosis factor-a (TNFa), whereas visible light, even at doses that induced other proinflammatory mediators, did not increase TNFa release. MMP release was also increased after exposure to visible light. MMP-1 release was increased by 2-fold from visible light doses, compared with 2.8-fold by solar simulator (Figure 1c), and MMP-9 release was similarly increased by approximately 2-fold from visible light doses.
Exposure to visible light induces the EGFRp42/44 MAPK pathway

To assess the role of visible light on skin, human epidermal equivalents were exposed to a doseresponse of visible light, and the production of ROS, inflammatory cytokines, and MMPs were determined. Visible light induced a dosedependent increase in intracellular hydrogen peroxide formation, with visible light doses of 65, 130, and 180 J cm2 increasing ROS by 5-, 9-, and 18-fold, respectively (Figure 1a). In comparison, a dose of 6.0 J cm2 from a solar simulator increased ROS by 19-fold. Visible light was also found to increase the release of proinflammatory cytokines

The EGFR/extracellular signalregulated kinase pathway has been shown to be activated by UV irradiation in keratinocytes and other cell types.2 To examine the effects of visible light on this pathway, primary human keratinocytes were exposed to UV solar light or doseresponse of visible light for specific periods of time. Treatment with the positive control (TNFa) and exposure to UV light or visible light (65180 J cm2) resulted in increased phosphorylation of the EGFR detected as phospho-tyrosine and the downstream marker of proliferation p42/44 MAPK, suggesting activation

a
Reactive oxygen species formation (mean fluorescence units) 250

b
Inflammatory protein Il-1 release (pg ml )

300 250 200 150 100 50 0

*
200 150 100 50 0 No treatment 65 J cm 130 J cm Visible light 180 J cm

* *

* *
6 J cm UVA/UVB

No treatment

65 J cm

130 J cm Visible light

180 J cm

6 J cm UVA/UVB

c
MMP-1 release (pg ml )

800 700 600 500 400 300 200 100 0 No treatment 65 J cm 130 J cm Visible light 180 J cm

* * * *

6 J cm UVA/UVB

Figure 1. Human epidermal skin equivalents were exposed to the indicated dose of visible light or UVA/UVB, and reactive oxygen species, proinflammatory cytokine, or matrix metalloprotease (MMP) production was measured. Exposure to visible light significantly increased (a) reactive oxygen species, (b) IL-1a, and (c) MMP-1 production in a dose-dependent manner. An asterisk indicates a significant difference (Po0.05) compared with unexposed skin using analysis of variance with NewmanKeuls post hoc test.

1902 Journal of Investigative Dermatology (2012), Volume 132

F Liebel et al.
Visible Light Induces ROS, Inflammatory Cytokines, and MMPs in Skin

P-Tyrosine Total EGFR Untreated TNF- Sham 4 6 65 130 Visible light (J cm2) 180

Solar light (J cm2)

p-ERK

Total ERK TNF- Untreated EGF 4 6 65 130 (J cm2) 180

of the EGFRERK pathway (Figure 2a and b). Total EGFR and ERK were also measured to show uniform protein loading. Pretreatment of keratinocytes with 100 nM of the specific EGFR inhibitor, Tyrphostin (AG 1478), for 20 minutes before irradiation blocked the downstream effects of visible light on the induction of ERK (Figure 2c), suggesting that the increased phosphorylation of ERK resulted from the intermediate activation of EGFR. These results demonstrate that visible light can induce activation of the EGFR pathway in keratinocytes in a manner similar to UV.
Visible light does not induce thymine dimer formation

Solar light
2 (J cm )

Visible light

p-ERK

Total ERK Sham 160 160+ Tyrphostin

2 Visible light (J cm )

Figure 2. Primary human keratinocytes exposed to visible light were evaluated for the levels of EGFR and ERK activation. (a, b) Primary human keratinocytes were either left untreated, treated with the positive control tumor necrosis factor-a (TNFa; 100 ng ml1), or exposed to UV solar light or visible light for calculated periods of time. The sham group was incubated at room temperature for an equivalent amount of time. Post exposure, the cells were incubated at 371C for 20 minutes, followed by cell lysis. Western blotting with phospho-tyrosine (molecular weight corresponding to EGFR) and phosphoextracellular signalregulated kinase (ERK) antibodies demonstrated activation of the EGFRERK pathway by multiple doses of visible light. (c) Pretreatment with the selective EGFR inhibitor, Tyrphostin (100 nM), for 20 minutes before irradiation blocked the downstream effects of visible light on the induction of ERK.

UV exposure is widely known to cause DNA damage in skin, including the formation of thymine (TT) dimers (Sinha and Hader, 2002; Marrot and Meunier, 2008). To investigate whether visible light also leads to thymine dimer formation, skin equivalents were treated with visible light or the positive control (UV) and stained for TT dimer formation. Whereas UV led to strong induction of TT dimers, visible light did not result in TT dimer formation even at higher doses of visible light (Figure 3).
Antioxidants reduce the ROS, cytokine, and MMP production induced by visible light

To determine whether antioxidants could reduce the damage caused by visible light, an antioxidant combination of Feverfew (Tanacetum parthenium) extract, Soy (Glycine soja) extract, and Gamma Tocopherol were combined into a UVA/UVB Sunscreen and the effect of visible light on ROS, IL-1a, or MMP-1 release was determined. Exposure to 65 J cm2 of visible light resulted in a significant increase in

TT dimers

Untreated

UV (6 J cm2)

Visible light (180 J cm2)

Visible light (240 J cm2)

Figure 3. Human epidermal skin equivalents were either left untreated or exposed to UV or visible light for calculated periods of time and stained for TT dimer formation. Exposure to UV led to induction of TT dimers, whereas visible light did not have the same effect. Bar 50 mm.

www.jidonline.org 1903

F Liebel et al.
Visible Light Induces ROS, Inflammatory Cytokines, and MMPs in Skin

ROS, IL-1a, and MMP-1 release. Pretreatment with a lotion containing UVA/UVB Sunscreen alone had no effect on reducing the damage from visible light. In contrast, the addition of antioxidants into the same UVA/UVB Sunscreen resulted in a significant reduction in the effects of visible light, reducing ROS, IL-1a, and MMP-1 release by 78%, 82%, and 87%, respectively (Figure 4). The direct effect of the antioxidants alone, without sunscreen, was tested and similarly mitigated the ROS, cytokine, and MMP release induced by visible light (data not shown).
Measuring free-radical production in the skin from visible light using chemiluminescence

subjects. Areas of the skin high in porphyrin content such as the forehead responded to low levels of visible light to induce free-radical production, which could be measured by photon emission or chemiluminescence. A 50 J cm2 dose at 150 mW cm2 of visible light was able to significantly increase the amount of free radicals by 85.8% over baseline measurements (Figure 5a). The addition of an antioxidant combination comprising Feverfew (T. parthenium) extract, Soy (G. soja) extract, and Gamma Tocopherol to the sunscreen was able to significantly reduce the free radicals by 54% (Figure 5b). These results are consistent with the in vitro ROS results and clearly demonstrate that visible light exposure induces free-radical production by the skin. DISCUSSION Although visible light comprises about 44% of solar light, few studies have looked at just visible light alone. Fuchs et al. (1989) have observed effects from UVA and visible light on skin, whereas Cho et al. (2008) have looked at visible light and IR; however, only Mahmoud et al. (2010) have discussed the potential effects of visible light on skin. Visible light is the only portion of the spectrum visible to the human eye and responsible for general illumination. If we consider that the solar irradiance in the visible range is about 50 mW cm2, the doses used in this study (40240 J cm2) would be equivalent to outside midsummer sunlight exposure of approximately 1590 minutes in Houston, TX. The data presented in this paper suggest that visible light can produce some of the same physiological effects as UV, including inflammation, ROS, and matrix-degrading enzymes being generated in the skin. It is well known that ROS are produced in skin following UV irradiation (Pathak and Stratton, 1968) and are major mediators of oxidative damage to the skin. Singlet oxygen can be generated from the action of UVA and endogenous photosensitizers, such as porphyrins and flavins (Ravanat et al., 2000), which can produce oxidative damage (Pelle et al., 2003). In the current study, exposing human skin equivalents to increasing doses, 40180 J cm2, of visible

We next sought to confirm the in vitro ROS results by studying free-radical production on the skin of human
120 100 Percent response 80 60 40 20 0 ROS IL-1 MMP-1 Visible light + UVA/UVB sunscreen Visible light + UVA/UVB sunscreen with antioxidant combo Visible light alone

Figure 4. UVA/UVB sunscreen with an antioxidant reduces reactive oxygen species (ROS), proinflammatory cytokine (IL-1), and matrix metalloprotease (MMP) production induced by visible light. Exposure to visible light significantly increased H2O2, IL-1a, and MMP-1 production, and this increase was not affected by the UVA/UVB sunscreen. In contrast, the UVA/UVB sunscreen containing an antioxidant blend significantly reduced H2O2, IL-1a, and MMP-1 production. An asterisk indicates a significant difference (Po0.05) compared with unexposed skin using analysis of variance with NewmanKeuls post hoc test.

140 120 Chemiluminescence signal (counts per second) 100 180 60 40 20 0 Untreated 50 J cm2 visible light exposure

100

*
% Change vs. baseline

90 80 70 60 50 40 30 20 10 0 No topical treatment Sunscreen

Sunscreen w/ antioxidants

Figure 5. Human subjects were exposed to visible light, and chemiluminescence was measured as a marker of reactive oxygen species. A 50 J cm2 dose of visible light at 150 mW cm2 significantly increased free-radical production; an asterisk indicates a significant difference (Po0.05) compared with untreated skin (N 40, a). UV sunscreens do not block non-UV longer wavelengths from stimulating free radicals in the skin. The addition of a potent antioxidant combination to the formulation significantly reduced the number of free radicals (b); *Po0.05 compared with sunscreen alone (N 12 per group of sunscreens, N 24 for visible light alone).

1904 Journal of Investigative Dermatology (2012), Volume 132

F Liebel et al.
Visible Light Induces ROS, Inflammatory Cytokines, and MMPs in Skin

light resulted in a dose-dependent increase in ROS production similar to UV (Figure 1a). Visible light has been used in light therapy for various conditions such as atopic dermatitis (Byun et al., 2011), eczema (Krutmann et al., 2005), and also in antimicrobial photochemotherapy (Soukos et al., 1998). These dermatological treatments using visible light tend to be focal applications to the affected lesional skin only to treat the underlying immune response. It is interesting to speculate that the therapeutic effects of visible light for atopic dermatitis and eczema may be mediated by the production of ROS in the skin, similar to the mechanism of psoralen plus UVA treatments for psoriasis. Irradiation of normal human epidermal keratinocytes with UVB can generate a dose-dependent increase in proinflammatory cytokines and expression of MMPs (Figure 1ac), which was also seen with visible light exposure. IR light has also been shown to be a source of oxidative stress for skin. Exposing human skin fibroblasts to near-IR radiation was shown to induce ROS formation and lead to the subsequent increased expression of MMPs (Schroeder et al., 2007). Additional studies exposed photoprotected sites of healthy human skin to solar relevant doses of IR radiation. MMP-1 expression in the dermis, but not in the epidermis, was upregulated in 80% of the tested individuals (Schroeder et al., 2008). Cho et al. (2008) were able to show that, in addition to UV, IR plus the visible light spectrum within natural sunlight increases MMP-1 and MMP-9 expression in vivo. Taken together, these findings suggest that although the UV fraction of the solar spectrum is harmful and can cause skin damage, the visible light and the IR light portions of the spectrum may also induce skin damage through to the dermis, which could lead to premature photoaging of skin through the generation of free radicals. Exposure to UVR has been shown to activate the EGFRERK pathway in keratinocytes (Huang et al., 1996) and in human skin (Knebel et al., 1996; Katiyar, 2001). The rapid phosphorylation of EGFR by UV is known to be reactive oxygen intermediatemediated (Huang et al., 1996; Knebel et al., 1996). To investigate the effects of visible light on this pathway, primary human keratinocytes were exposed to UV solar light or doseresponse of visible light. Exposure to UV light and visible light (65180 J cm2) resulted in increased phosphorylation of the EGFR (detected as phospho-tyrosine) and the downstream marker of cell proliferation p42/44 MAPK, suggesting activation of the EGFRERK pathway (Figure 2a and b). To determine whether EGFR is a prerequisite for ERK activation by visible light, human keratinocytes were pretreated with the specific EGFR kinase inhibitor (Osherov and Levitzki, 1994; Ellis et al., 2006), Tyrphostin (AG1478), before visible light treatment. The presence of Tyrphostin abolished the phosphorylation of ERK, suggesting that EGFR is required for the downstream induction of ERK by visible light (Figure 2c). Activation of the EGFRERK pathway has been implicated in UV-induced epidermal hyperplasia (El-Abaseri et al., 2006), and also in the activation of MMPs that may accelerate skin aging (Kang et al., 2008). Induction of ERK signaling by visible light might be upstream of the MMP-1 activation shown in human skin equivalents (Figure 1c), and potentially any visible lightinduced hyperplasia.

DNA damage by UVB irradiation results from photochemical reactions consequent to direct absorption of photons by DNA bases. The UV-induced DNA lesions that have been studied in most detail are the cyclobutane pyrimidine dimer and the 64 pyrimidinepyrimidone photoproduct at adjacent pyrimidines (Nakajima et al., 2004). Nuclear DNA strand breaks are also readily produced by incubation of keratinocytes with hydrogen peroxide (Armeni et al., 2001), and hydroxyl radicals can be generated from hydrogen peroxide through Fe2 -mediated Fenton-type reactions (Stewart et al., 1996). Our studies showed that visible lightirradiated tissues did not induce thymine dimer formation (Figure 3) even at concentrations sufficient to induce significant increases in ROS. In comparison, UV irradiation induced a pronounced DNA damage (Figure 3). These results suggest that, in contrast to UV, visible light photons may not be directly absorbed by DNA bases and therefore may not result in thymine dimer formation. Visible light could contribute to other forms of DNA damage, such as 8-Oxoguanine, which were not measured in the current study. Indeed, exposure of AS52 Chinese hamster cells to a visible light source was reported to produce 8-Oxoguanine DNA damage with a relative maximum between 400 and 450 nm (Kielbassa et al., 1997). To confirm the in vitro results, we evaluated the effects of visible light on oxidative stress in a clinical model. UVA (320 nm400 nnm) exposure has been shown to be a source of oxidative stress in the skin clinically (Ou-Yang et al., 2004), but no studies have examined ROS from visible light alone. To study the effects of visible light (400700 nm) on the skin of subjects, we used a photon detector to measure the rate of photon emission or chemiluminescence. A visible light dose of 50 J cm2 was able to significantly increase the photon emission on the skin by 85.8% compared with baseline measurements (Figure 5a). A broad UVA/UVB sunscreen was tested for its ability to inhibit free radicals generated from visible light, and this resulted in no change in the amount of photon emission. These data show that the UV sunscreens do not block the longer non-UV wavelengths from stimulating free radicals in the skin. When an antioxidant combination including a Parthenolide-Depleted Feverfew (T. parthenium) extract (Martin et al., 2008) was added to the UV sunscreen, a 54% decrease in the amount of generated free radicals from visible light was observed (Figure 5b). These results are consistent with in vitro study that demonstrated that a UVA/UVB sunscreen did not affect the visible lightinduced release of ROS, MMP-1, or proinflammatory cytokines from epidermal tissues; however, the addition of a potent antioxidant combination to a UVA/UVB sunscreen significantly reduced the release of all these mediators (Figure 4). These data demonstrate that sunscreens containing UVA/UVB sunfilters alone do not prevent the free-radical production from non-UV longer wavelengths of the solar spectrum. When UV contacts the skin, the energy of the photons is absorbed by chromophores, which can mediate the cellular effects of UV exposure. A variety of chromophores, including DNA and some aromatic amino acids, are known to absorb UVR; however, biomolecules that can absorb longer wavewww.jidonline.org 1905

F Liebel et al.
Visible Light Induces ROS, Inflammatory Cytokines, and MMPs in Skin

lengths such as visible light are not fully understood. A few potential targets for the actions of visible light may include riboflavin, hemoglobin, and bilirubin, all of which show absorbance peaks in the visible range of the spectrum, and are found in the skin. Melanin and b-carotene have been shown to absorb UV light, but can also act as endogenous chromophores for visible light (Anderson and Parrish, 1981; Mahmoud et al., 2008). It is also known that the depth of penetration of optical radiation in skin is wavelength dependent, and affected by position and absorption spectrum of the corresponding chromophore in skin (Mahmoud et al., 2008). In the dermis, optical scattering of light is considered to be inversely proportional to wavelength and affects the depth of optical penetration (Anderson and Parrish, 1981). The approximate depth for penetration of visible light (400700 nm) in a fair-skinned Caucasian individual was estimated to be between 90 and 750 mm by Anderson and Parrish (1981), compared with a depth of 1.590 mm for UVR (Anderson and Parrish, 1981). Thus, even though visible light photons are less energetic than UV photons, due to the deeper dermal penetration visible light may still have a substantial effect on skin. Taken together, these results demonstrate that visible light exposure can induce ROS, which can lead to the release of proinflammatory cytokines and MMPs in the skin, similar to the effects of UV, and therefore visible light may contribute to the signs of premature aging in the skin. MATERIALS AND METHODS
UV and visible lightinduced cytokine release
The UV light source used as a positive control was an 1,000 W Oriel solar UV simulator (Oriel, Stratford, CT), equipped with an atmospheric attenuation filter (Schott WG 320, Schott Optical, Elmsford, NY; 1 mm thick) and a visibleIR filter (Schott UG 11; 1 mm thick). This filtered xenon light source provided a simulated solar UVR spectrum (290400 nm) that was nearly devoid of visible and IR radiation as determined using an IL-1400 Research Radiometer (International Light, Newburyport, MA). The fluence rate of the light sources was 7.52 mW cm2, measured using a calibrated Oriel Thermopile Model 71767. The visible light source used was the Fiber-Lite Model 170-D (Dolan-Jenner Industries, Boxborough, MA) with a 150 W quartz halogen lamp; a straight 8-mm Dolan-Jenner glass optical fiber was used for irradiation. The source was filtered with three KG5/3-mm Schott glass filters and one GG400/3-mm (Schott Optical Company) was used to filter IR and UVR from the light source, respectively. The light source was characterized spectrally by a calibrated Optronics OL750 spectroradiometer, Optronics Laboratories, Orlando, FL. From the spectroradiometric measurements (from 350 to 1800 nm), the light source has 0.14% of UVA (350400 nm), 98.3% of visible light, 1.7% of IRA (7001400 nm), and 0.3% of IRB (Supplementary Figure S1 online). For the doses used in the studies presented in this paper, the contribution from UVA and IR (A and B) are negligible. For the highest visible light dose used in this study, 240 J cm2, we would have less than 0.34 J cm2 of UVA and about 4 J cm2 of IRA being delivered to the irradiated sample. The spectral irradiance of the visible light source is described in Mahmoud et al (2010). Skin surface temperature of skin equivalents and media temperature
1906 Journal of Investigative Dermatology (2012), Volume 132

exposed to either UV or visible light sources were measured during exposure periods and did not increase by 411C.

Cell culture and skin equivalents

Normal human epidermal neonatal keratinocytes were obtained from Lifeline Cell Technologies (Frederick, MD) and maintained in Dermalife keratinocyte medium (Lifeline Cell Technologies, Frederick, MD) with supplements. Reconstituted human epidermis (EPI-200-HCF) was purchased from MatTek (Ashland, MA). After reception, epidermal equivalents were incubated with a phenol-free and hydrocortisone-free maintenance medium (MatTek at 371C for 24 h.

Assessment of UV and visible lightinduced ROS and mediator release from reconstituted epidermis
MatTek epidermal equivalent tissues were topically treated for 45 minutes with 5 mM of the hydrogen peroxidesensitive fluorescent probe 5-(and-6)-chloromethyl-20 ,70 -dichlorodihydro-fluorescein diacetate, acetyl ester (Invitrogen, Carlsbad, CA). After incubation, the plate was rinsed with warm phosphate-buffered saline to remove excess probe. A volume of 6 ml of sunscreen formulation or Feverfew extract (10 mg ml1, wt/volt) was topically applied and spread uniformly across the surface of the epidermal equivalent. Tissues were then incubated for 1 hour before UV or visible light exposure. Fluorescence was then measured using a CytoFluor Fluorescence Plate Reader (PerSeptive Biosystems, Framingham, MA) set with the following filter combination: excitation at 485 nm and emission at 530 nm. The tissues were then exposed to UV or visible light, and immediately after exposure the tissues were measured for fluorescence. For measurement of cytokines, the equivalents were transferred back to the maintenance medium and incubated at 371C for 24 hours. After 24 hours, the medium below each equivalent was collected and analyzed for secreted IL-1a, IL-1 receptor antagonist, IL-6, TNFa, GM-CSF, IL-8, MMP-1, and MMP-9 by ELISA, using commercially available immunoassay multiplex kits (Millipore, Bedford, MA) on a Luminex L100 (Luminex, Austin, TX).

Western blotting
Primary keratinocytes grown in 12-well plates were either left untreated, treated with TNFa (100 ng ml1) or exposed to UV solar light or multiple doses of visible light for specific periods of time corresponding to the doses. The sham group was left at room temperature for an equivalent amount of time. Post exposure, cells were incubated at 371C for 20 minutes, then washed with phosphate-buffered saline and lysed with radioimmunoprecipitation lysis buffer containing 65 mM Tris (pH 7.4), 150 mM NaCl, 1 mM EDTA (pH 8), 1% nonidet P-40, 0.25% sodium deoxycholate, 50 mM NaF, 1 mM Na3VO4, 1 mM phenylmethylsulfonyl fluoride, and 1 protease inhibitor cocktail (Sigma, St Louis, MO). Lysates were centrifuged and total protein was estimated in the supernatants using a BCA protein assay kit (PierceThermo Fisher Scientific, Rockford, IL). Protein (30 mg) was loaded on SDS-PAGE followed by immunoblotting with the specific antibodies (incubation with primary antibodies diluted 1:1,000 for overnight at 41C) and detection using the ECL chemiluminescence detection system (Amersham Life SciencesGE Healthcare, Piscataway, NJ).

Assessing DNA damage (thymine dimer staining)

Thymine (TT) dimer staining of human skin equivalents was performed as described previously (Martin et al., 2008). In brief, skin equivalents

F Liebel et al.
Visible Light Induces ROS, Inflammatory Cytokines, and MMPs in Skin

were either left untreated or exposed to UV or visible light, after which they were immediately fixed in formalin and sent to Paragon BioServices (Baltimore, MD) for TT dimer staining.

Frederick JE, Snell HE, Haywood EK (1989) Solar ultraviolet radiation at the earths surface. Photochem Photobiol 50:44350 Fuchs J, Huflejt ME, Rothfuss LM et al. (1989) Acute effects of near ultraviolet and visible light on the cutaneous antioxidant defense system. Photochem Photobiol 50:73944 Huang RP, Wu JX, Fan Y et al. (1996) UV activates growth factor receptors via reactive oxygen intermediates. J Cell Biol 133:21120 Kang KA, Zhang R, Piao MJ et al. (2008) Inhibitory effects of triphlorethol-A on MMP-1 induced by oxidative stress in human keratinocytes via ERK and AP-1 inhibition. J Toxicol Environ Health A 71:9929 Katiyar SK (2001) A single physiologic dose of ultraviolet light exposure to human skin in vivo induces phosphorylation of epidermal growth factor receptor. Int J Oncol 19:45964 Kielbassa C, Roza L, Epe B et al. (1997) Wavelength dependence of oxidative DNA damage induced by UV and visible light. Carcinogenesis 18:8116 Knebel A, Rahmsdorf HJ, Ullrich A et al. (1996) Dephosphorylation of receptor tyrosine kinases as target of regulation by radiation, oxidants or alkylating agents. EMBO J 15:531425 Krutmann J, Medve-Koenigs K, Ruzicka T et al. (2005) Ultraviolet-free phototherapy. Photodermatol Photoimmunol Photomed 21:5961 Mahmoud BH, Hexsel CL, Hamzavi IJ et al. (2008) Effects of visible light on the skin. Photochem Photobiol 84:45062 Mahmoud BH, Ruvolo E, Hexsel CL et al. (2010) Impact of long-wavelength UVA and visible light on melanocompentant skin. J Invest Dermatol 130:20927 Marrot L, Meunier JR (2008) Skin DNA photodamage and its biological consequences. J Am Acad Dermatol 58(5 Suppl 2):S13948 Martin K, Sur R, Liebel F et al. (2008) Parthenolide-depleted Feverfew (Tanacetum parthenium) protects skin from UV irradiation and external aggression. Arch Dermatol Res 300:6980 Nakajima S, Lan L, Kanno S et al. (2004) UV light-induced DNA damage and tolerance for the survival of nucleotide excision repair-deficient human cells. J Biol Chem 279:466747 Osherov N, Levitzki A (1994) Epidermal-growth-factor-dependent activation of the src-family kinases. Eur J Biochem 225:104753 Ou-Yang H, Stamatas G, Saliou C et al. (2004) A chemiluminescence study of UVA-induced oxidative stress in human skin in vivo. J Invest Dermatol 122:10209 Pathak MA, Stratton K (1968) Free radicals in human skin before and after exposure to light. Arch Biochem Biophys 123:46876 Pelle E, Huang X, Mammone T et al. (2003) Ultraviolet-B-induced oxidative DNA base damage in primary normal human epidermal keratinocytes and inhibition by a hydroxyl radical scavenger. J Invest Dermatol 121:17783 Ravanat JL, Di Mascio P, Martinez GR et al. (2000) Singlet oxygen induces oxidation of cellular DNA. J Biol Chem 275:406014 Schroeder P, Lademann J, Darvin ME et al. (2008) Infrared radiation-induced matrix metalloproteinase in human skin: implications for protection. J Invest Dermatol 128:24917 Schroeder P, Pohl C, Calles C et al. (2007) Cellular response to infrared radiation involves retrograde mitochondrial signaling. Free Radic Biol Med 43:12835 Sinha RP, Hader DP (2002) UV-induced DNA damage and repair: a review. Photochem Photobiol Sci 1:22536 Soukos NS, Ximenz-Fyvie L, Hamblin MR et al. (1998) Targeted antimicribal photochemotherapy. Antimicrob Agents Chemother 42:2595601 Stewart MS, Cameron GS, Pence BC et al. (1996) Antioxidant nutrients protect against UVB-induced oxidative damage to DNA of mouse keratinocytes in culture. J Invest Dermatol 106:10869 Wan Y, Belt A, Wang Z et al. (2001) Transmodulation of epidermal growth factor receptor mediates IL-1 beta-induced MMP-1 expression in cultured human keratinocytes. Int J Mol Med 7:32934 Wlaschek M, Heinen G, Poswig A et al. (1994) UVA-induced autocrine stimulation of fibroblast-derived collagenase/MMP-1 by interrelated loops of interleukin-1 and interleukin-6. Photochem Photobiol 59:5506

Chemiluminescence
Photons emitted from the skin were measured by a red-sensitive, tri-alkali photomultiplier with a 2.5-cm diameter photocathode surface (Electron Tubes, Rockaway, NJ, model 9828SA). The photomultiplier was in a Peltier-cooled housing unit (Electron Tubes, model CDM30) and was routinely cooled to 18 to 201C to reduce detector noise. The housing unit can be held with one hand as a probe and placed on the skin area to be measured. The surface of the photocathode was protected by a manual shutter (Products for Research, Danvers, Ma, model PR318). The shutter is opened for measurements and closed for instrument background. The distance between the skin and the photocathode surface was approximately 1.5 cm. The photomultiplier was supplied with a potential of 1.15 kV of high voltage (Brandenburg, Brierley Hill, UK, model 477) with a variation less than 2 V. After going through an amplifier and discriminator (Electronic Tubes, model AD6), the output signals from the photomultiplier were measured in a photon-counting module (Electronic Tubes, model CT1) and recorded by a computer. All measurements were recorded in a positively dark room; otherwise, the background would overwhelm the signal (Ou-Yang et al., 2004). The forehead was used for all the measurements.

Statistical analysis
All data are presented as meanSD. For in vitro studies, a one-way analysis of variance with NewmanKeuls post hoc test was used to determine significance. A value of Po0.05 was considered significant. For chemiluminescence results, data were analyzed using a Students t-test with significance set at Po0.05.
CONFLICT OF INTEREST
All authors are employees of Johnson & Johnson.

ACKNOWLEDGMENTS
We thank Jean Krutmann for discussions and suggestions on the effect of non-UV light on skin. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid

REFERENCES
Anderson RR, Parrish JA (1981) The optics of human skin. J Invest Dermatol 77:139 Armeni T, Battino M, Stronati A et al. (2001) Total antioxidant capacity and nuclear DNA damage in keratinocytes after exposure to H2O2. Biol Chem 382:1697705 Byun HJ, Lee HI, Kim B et al. (2011) Full-spectrum light phototherapy for atopic dermatitis. Int J Dermatol 50:94101 Cho S, Lee MJ, Kim MS et al. (2008) Infrared plus visible light and heat from natural sunlight participate in the expression of MMPs and type I procollagen as well as infiltration of inflammatory cell in human skin in vivo. J Dermatol Sci 50:12333 El-Abaseri TB, Putta S, Hansen LA et al. (2006) Ultraviolet irradiation induces keratinocyte proliferation and epidermal hyperplasia through the activation of the epidermal growth factor receptor. Carcinogenesis 27:22531 Ellis AG, Doherty MM, Walker F et al. (2006) Preclinical analysis of the analinoquinazoline AG1478, a specific small molecule inhibitor of EGF receptor tyrosine kinase. Biochem Pharmacol 71:142234

www.jidonline.org 1907