Original Contribution

Free Radical Biology & Medicine, Vol. 32, No. 12, pp. 1293–1303, 2002 Copyright © 2002 Elsevier Science Inc. Printed in the USA. All rights reserved 0891-5849/02/$–see front matter PII S0891-5849(02)00831-6

PHOTOPROTECTIVE POTENTIAL OF LYCOPENE, ␤-CAROTENE, VITAMIN E, VITAMIN C AND CARNOSIC ACID IN UVA-IRRADIATED HUMAN SKIN FIBROBLASTS

1 † ‡ ELIZABETH A. OFFORD,* JEAN-CHARLES GAUTIER,* ORNELLA AVANTI,* CORINNE SCALETTA, FRANK RUNGE, ‡ † KLAUS KRAMER¨ , and LEE ANN APPLEGATE *Nestle´ Research Center, Department of Nutrition, Lausanne, Switzerland; †University Hospital of the Canton of Vaud, Department of Gynecology and Obstetrics, Lausanne, Switzerland; and ‡BASF Aktiengesellschaft, Strategic Marketing Fine Chemicals, Ludwigshafen, Germany

(Received 3 December 2001; Revised 1 March 2002; Accepted 1 March 2002)

Abstract—The photoprotective potential of the dietary antioxidants vitamin C, vitamin E, lycopene, ␤-carotene, and the rosemary polyphenol, carnosic acid, was tested in human dermal fibroblasts exposed to ultraviolet-A (UVA) light. The carotenoids were prepared in special nanoparticle formulations together with vitamin C and/or vitamin E. Nanoparticle formulations, in contrast to dimethylsulphoxide, stablized lycopene in the cell culture medium and allowed efficient cellular uptake. The presence of vitamin E in the formulation further increased the stability and cellular uptake of lycopene. UVA irradiation of the human skin fibroblasts led to a 10–15-fold rise in metalloproteinase 1 (MMP-1) mRNA. This rise was suppressed in the presence of low ␮M concentrations of vitamin E, vitamin C, or carnosic acid but not with ␤-carotene or lycopene. Indeed, in the presence of 0.5–1.0 ␮M ␤-carotene or lycopene, the UVA-induced MMP-1 mRNA was further increased by 1.5–2-fold. This increase was totally suppressed when vitamin E was included in the nanoparticle formulation. Heme-oxygenase 1 (HO-1) mRNA expression was strongly induced by UVA irradiation but none of the antioxidants inhibited this effect at the concentrations used in this study. Indeed, ␤-carotene or lycopene (0.5–1.0 ␮M) led to a further 1.5-fold rise in the UVA-induced HO-1 mRNA levels. In conclusion, vitamin C, vitamin E, and carnosic acid showed photoprotective potential. Lycopene and ␤-carotene did not protect on their own but in the presence of vitamin E, their stability in culture was improved and the rise in MMP-1 mRNA expression was suppressed, suggesting a requirement for antioxidant protection of the carotenoids against formation of oxidative derivatives that can influence the cellular and molecular responses. © 2002 Elsevier Science Inc.

Keywords—Oxidative stress, UVA, Heme oxygenase, Metalloproteinase I, Skin, Lycopene, ␤-carotene, Vitamin C, Vitamin E, Rosemary, Carnosic acid, Free radicals

INTRODUCTION stress marker gene, heme-oxygenase I (HO-1) [3,4]. Sin- glet oxygen is strongly implicated as mediator in induc- Ultraviolet (UV) irradiation is well known to induce tion of these two genes [5,6]. photodamage and premature skin aging [1]. Ultravio- While topical application of sun screens provides a let-A (UVA) light is particularly associated with oxida- barrier protection to the skin epithelium, protection of the tive processes involved in photoaging. A cascade of gene more profound dermal layers may be offered by dietary expression is initiated following UVA irradiation, which antioxidants acting from within. Up until now, the most results in the upregulation of interstitial collagenase intensively studied dietary antioxidants for prevention of (metalloproteinase 1 [MMP-1]) [2] and the oxidative skin photodamage have been vitamin C, vitamin E, and ␤ Address correspondence to: Dr. Elizabeth Offord, Nestle´ Research -carotene [7]. Short-term clinical studies on the ability Center, Department of Nutrition, P. O. Box 44, CH-1000 Lausanne 26, of these antioxidants to protect against UV-induced er- Switzerland; Tel: ϩ41 (21) 785 8809; Fax: ϩ41 (21) 785 8925; ythema have shown varying degrees of success [7–11]. E-Mail: [email protected]. ␤ 1Present address: Aventis Pharma, Drug Safety Evaluation, 13, Quai Protection by -carotene is enhanced when combined Jules guesde, BP 14, 94403 Vitry-sur-Seine, France. with vitamin E [9]. Another carotenoid found in toma-

1293 1294 E. A. OFFORD et al.

toes, lycopene, has received much attention for its po- ing [1] and of HO-1 as a general marker of cellular tential health properties [12], particularly with respect to oxidative stress [3,4]. protection against gastrointestinal tract, prostate, and lung cancers [13–16]. Lycopene is an acyclic carotenoid MATERIALS AND METHODS with 11 linearly arranged conjugated double bonds and is the most potent singlet oxygen scavenger among the carotenoids [17]. Lycopene is present in skin and could Antioxidant preparations play a role for protecting against UV radiation as it is the Powdered nanoparticles containing various amounts antioxidant most quickly depleted in skin upon exposure of lycopene, ␤-carotene, vitamin E (all-rac-␣-tocopher- to solar radiation [18]. Recently, it was reported for the ol), and sodium ascorbate either alone or in combination, first time that ingestion of tomato paste daily for 10 were provided by BASF (Ludwigshafen, Germany). The weeks, protected against UV light-induced erythema on carotenoid formulations were prepared using a precipi- the dorsal skin [19]. tation process by mixing a carotenoid solution in alcohol The natural spice and flavoring agent, rosemary, is a with an aqueous gelatin solution (so-called mixing cham- potent source of natural antioxidants, such as the poly- ber micronization), as previously described [36–38]. The phenolic diterpenes carnosol and carnosic acid [20–23]. resulting nanoscale carotenoid dispersions were trans- Rosemary extracts or their active constituents have anti- ferred into dry powders by conventional spray drying. carcinogenic properties [24] and reduce skin tumorige- The vitamin E formulations were prepared by an emul- nicity [25–27]. Their mode of action has principally been sification/spray drying process. The corresponding con- described to involve scavenging of free radicals and modulation of xenobiotic-metabolizing enzymes [28– trol powders contained gelatin and glucose syrup mixed 30]. In this study, we were interested in further investi- or not mixed with ascorbyl palmitate, a lipophilic form of ␮ gating whether carnosic acid showed protection against vitamin C (0.5% in powder; 0.7 M final concentration UVA-induced photodamage. in cell culture medium). Ascorbyl palmitate was used as Cell culture systems provide the opportunity to inves- an additive in the mixing chamber micronization process tigate the molecular and cellular processes involved in to achieve the proper stability of the carotenoid nanopar- photoprotection by antioxidants; however, certain tech- ticles. The powders were stored at room temperature nical difficulties may be encountered with substances sealed under argon, and were dissolved in the culture that are either unstable or lipid soluble. Carotenoids are medium. Equivalent amounts of control matrix powders very hydrophobic molecules that are often solubilized in were added in all experiments. Crystalline lycopene was organic solvents such as tetrahydrofuran (THF) or dim- prepared in DMSO as a stock concentration of 1 mM and ethylsulphoxide (DMSO); however, an uncontrolled pre- diluted in cell culture medium preheated to 37°C, to a cipitation process occurs upon addition of these solutions final concentration of 1 ␮M. The stock and final culture to aqueous media. In this process, carotenoid crystals are medium concentrations were checked by HPLC as de- formed with a noncontrollable particle size which is scribed below. Carnosic acid, purifed from rosemary typically in the order of several micrometers, depending leaves [39], was prepared as a stock solution of 1 mg/ml on the experimental conditions. The solubility and up- in DMSO and used at a final concentration of 0.1–1.0 take of these large crystals in the cells is quite limited and ␮g/ml (equivalent to 0.3–3.0 ␮M). Control cells were there is almost no protection against chemical degrada- treated with 0.1% DMSO alone. tion. Alternative ways of delivering lipid-soluble compounds include micelles, microemulsions, nanoparticles, water-dispersible beadlets, artificial liposomes, enriched Cell cultures bovine serum, or specialized formulations, each of which has an influence on the uptake and stability of the com- Human dermal fibroblasts GT-F were prepared from a pounds [31–35]. 26-year-old male with skin type III (brown hair, brown The aim of the present study was to test the dietary eyes) and were used between the 10th and 15th passages, antioxidants, vitamin C, vitamin E, lycopene, ␤-carotene, as previously described [40]. Fibroblasts were plated out and carnosic acid for their ability to protect against in 100-mm-diameter petri dishes in DMEM medium photoaging in primary cultures of human dermal fibro- containing Glutamax I, glucose 4.5 g/l (GibcoBRL, blasts irradiated with UVA. The lipid-soluble com- Basel, Switzerland) supplemented with 10% fetal bovine pounds were formulated in special nanoparticle formu- serum (50 ml for 500 ml medium), and penicillin/strep- lations. The photoprotective effect of the antioxidants tomycin (0.5 ml of 0.1% stock solution for 500 ml

was monitored by studying the expression of MMP-1 as medium). Cells were incubated in a saturated 10% CO2/ a marker of potential collagen degradation and photoag- 90% air atmosphere at 37°C and grown to conﬂuency. Photoprotective effects of dietary antioxidants 1295

Cell treatment and UVA irradiation and 100 ␮l for cell lysates) were injected onto a C18 Hypersil precolumn (20 ϫ 4 mm, 5 ␮m internal diame- Cells were first treated with test compounds 24 h ter, Hewlett Packard) followed by a C18 Nova-pak col- before UVA irradiation to allow penetration into the umn (300 ϫ 3.9 mm, Waters). Tocopherols and carote- ϫ 6 cells. Just prior to the irradiation, 2 10 cells were noids were separated under isocratic conditions with a washed with 10 ml of phosphate buffered saline (PBS) mobile phase consisting of acetonitrile/tetrahydrofuran/ and 5 ml was added. UVA irradiation was performed methanol/1% ammonium acetate (533.5/193.6/53.7/28; with a high-pressure UVA lamp UVASUN 3000 2 by weight) at a flow rate of 1.5 ml/min. Data were (Mu¨tzhas, Munich, Germany) using a dose of 50 J/cm simultaneously acquired on the diode array detector at (25 min exposure time). On average, this dose typically 472 nm for lycopene, 450 nm for ␤-carotene, and on the corresponds to one minimal erythemal dose in human fluorimetric detector at 298 nm excitation, 328 nm emis- skin [41]. Fresh medium containing the test compounds sion for tocopherols. was immediately added to the cells (concentrations and times of incubation are described in the results). For Northern blot experiments, cultures were stopped by Northern blots washing cells twice with 10 ml of PBS buffer, then removing the buffer and storing cells at Ϫ80°C. Cyto- Total RNA was extracted from confluent cultures of toxicity was evaluated with neutral red. human fibroblasts GT-F as described previously [29]. Fifteen micrograms of RNA were electrophoresed on a 1.2% agarose gel under denaturing conditions, trans- Extraction of tocopherols and carotenoids ferred to Hybond-NX membranes (Amersham Pharma- Tocopherols and carotenoids were extracted from the cia Biotech, Du¨bendorf, Switzerland), and hybridized ␥ 32 culture medium and the cell lysates at the time zero or with nick-translated [ - P]-labeled cDNA probes to 24 h after treatment as described previously [42] with MMP-1 (1970 bp fragment from clone pCol 185.2 kindly some modifications. Culture medium (150 ␮l) was mixed provided by Dr. B. Marmer, Washington University with 200 ␮l water, 400 ␮l ethanol, and 2 ␮l deferox- School of Medicine) [43] or HO-1 (500 bp fragment amine mesylate (10 mg/ml water) for 30 s in a colored 2 from clone 1/149), kindly provided by Dr. R. Tyrrell [4]. ␤ ml Eppendorf tube to protect samples from exposure to Blots were also hybridized with cDNA probe to -actin light. Two extractions were performed with 0.9 ml hex- (738 bp fragment amplified by PCR) for standardization. ane containing butylated hydroxytoluene (BHT) (30 mg/ Northern blot bands were visualized and quantified with 100 ml). The solution was mixed for 30 s and centrifuged a Packard Instant Image Analyzer. Northern blots were for 5 min at 1000 ϫ g at room temperature. Both super- always performed at least in duplicate (two independent natants were transferred into a colored Eppendorf tube experiments). and the solvent was evaporated under nitrogen. The organic residues were resuspended with 70 ␮l dioxane/ RESULTS ethanol (1/1; v/v) and 70 ␮l acetonitrile. Cells cultured in 100 mm petri dishes (107 cells) were washed four times High chemical stability and cellular uptake of with 10 ml phosphate buffered saline (PBS) and lysed antioxidants in nanoparticle formulations with 600 ␮l of lysis buffer (150 mM NaCl, 100 mM Nanoparticles were chosen to deliver the carotenoids Tris-HCl pH 8, 2 mM EDTA pH 8, and 1% Tween 20). to the aqueous media of cell cultures. Upon mixing with The cells were scraped and sonicated for 10 s in an water, cold-water-dispersible powders form stable nano- Eppendorf tube. Tocopherol and carotenoids were ex- particulate dispersions (particle size well below 1 ␮m) tracted from 150 ␮l of cell lysates as described above. that do not change in particle size distribution. The The protein contents of the cell lysates were measured carotenoid nanoparticles are stabilized with tocopherol to with Coomassie Plus reagent (Pierce, Rockford, IL, protect against oxidative degradation, and provide a USA) using bovine serum albumin as a reference stan- well-defined form of carotenoids that is physically dard. (against precipitation) and chemically (against degradation) stable. The influence of the nanoparticle formula- HPLC analysis of tocopherols and carotenoids tion, compared to a solution of crystalline lycopene in DMSO, on the chemical stability and the cellular uptake The HPLC system consisted of a Hewlett Packard of carotenoids and vitamin E is shown in Fig. 1. The 1100 equipped with a Hewlett Packard 1040 M Series II stability of the components was assessed by the percent- photodiode array detector and a Perkin Elmer LC 240 age recovery in the cell medium after 24 h exposure. fluorimetric detector. Samples (50 ␮l for supernatants Both vitamin E and ␤-carotene were highly stable when 1296 E. A. OFFORD et al.

DMSO, lycopene crystals were observed on the cell surface). In addition, we found that the presence of vitamin E in the nanoparticles clearly enhanced the cellular uptake of lycopene (8.2%, corresponding to 2.0 nM lycopene/mg protein, Fig. 1B).

UVA-induced MMP-1 and HO-1 in human skin fibroblasts

UVA irradiation of human dermal ﬁbroblasts GT-F with a dose of 50 J/cm2 led to a 10–15-fold rise in MMP-1 mRNA between 18 and 30 h postirradiation (Fig. 2). The 24 h time-point was chosen thereafter for investigating the effect of test compounds on MMP-1 mRNA expression. UVA irradiation of GT-F cells increased HO-1 mRNA by 10–40-fold between 3 and 6 h (Fig. 2). The experimental condition chosen for HO-1 mRNA expression was 4 h post irradiation.

Regulation of UVA induced MMP-1 by antioxidants

Human dermal fibroblasts were preincubated with the Fig. 1. Stability in culture medium (A) and cellular uptake (B) of antioxidants either individually or as mixtures, then ir- vitamin E (light grey bars), ␤-carotene (white bars) and lycopene (black radiated with 50 J/cm2 UVA and reincubated with the bars) formulated as nanoparticles and incubated with human fibroblast same compounds. The carotenoids were formulated to- cells for 24 h. In (A), the stability of lycopene in DMSO is shown as ␮ a black, patterned bar. The percentage of cellular uptake was calculated gether with vitamin C (14 M) and/or vitamin E (0.5–2 with respect to the amount of antioxidants initially measured in the ␮M). The results for ␤-carotene are shown in Fig. 3. culture medium, namely, 1.0 ␮M lycopene, 0.6 ␮M ␤-carotene, 1.8 UVA irradiation of fibroblast cells led to an 8-fold rise in ␮M vitamin E, and 1.1 ␮M lycopene combined with 0.4 ␮M vitamin E. All formulations contained additional vitamin C (14 ␮M). The MMP-1 mRNA expression (Fig. 3, lanes 3, 4), which results are shown as means of two independent experiments with was reduced by 4-fold in the presence of vitamin E (2 variation less than 10%. ␮M) and by 2.5-fold in the presence of vitamin C (14 ␮M), (Fig. 3, lanes 5, 6, 9, 10). When ␤-carotene (0.1– formulated in nanoparticles as shown by a recovery of 0.5 ␮M) was added, the protection by either vitamin C or 80–100% in the cell medium. Lycopene was signifi- vitamin E was diminished (Fig. 3, lanes 7, 8, 11–4). cantly stabilized in the nanoparticle formulation com- Indeed, in the cells treated with 0.5 ␮M ␤-carotene in the pared to a preparation in DMSO, but was less stable than presence of vitamin C, the level of MMP-1 mRNA was vitamin E or ␤-carotene. Indeed, more than 40% of about 1.5-fold higher than in the cells treated with UVA lycopene was recovered when formulated in nanopar- alone (Fig. 3, lanes 13, 14). The combination of all three ticles whereas less than 10% of lycopene was recovered antioxidants, vitamin E, vitamin C and ␤-carotene, pro- when dissolved in DMSO (Fig. 1A). Lycopene was fur- vided strong protection, reducing the MMP-1 mRNA ther stabilized in the culture medium by the addition of levels to those obtained with vitamin C (Fig. 3, lanes 15, vitamin E to the formulation leading to almost full re- 16). covery in the cell medium. Similar results to ␤-carotene were observed for lyco- Nanoparticles were found to be convenient carriers pene (Fig. 4). Once again, vitamin C alone inhibited the for delivering lycopene and vitamin E into human dermal UVA-induced rise in MMP-1 mRNA by about 2-fold fibroblasts in monolayer cultures (Fig. 1B). Indeed, (Fig. 4, lane 5) but the addition of 0.1 ␮M lycopene nanoparticles were nontoxic (upon evaluation with the restored the MMP-1 mRNA to values seen in the UVA- neutral red assay) and allowed high cellular uptakes of irradiated fibroblasts without antioxidants (Fig. 4, lanes lycopene (4.9%, corresponding to 0.9 nM lycopene/mg 6, 7). The addition of lycopene at 1 ␮M resulted in a protein) and vitamin E (2.1%, corresponding to 1.0 nM reduced level of actin mRNA and an overexpression of vitamin E/mg protein) after 24 h exposure. No visible MMP-1 mRNA (Fig. 4, lanes 8, 9; not quantified due to crystals were found on the cell surface with lycopene the effect on actin). However, both the cytotoxicity and when formulated as nanoparticles (when dissolved in the induction of MMP-1 could be reversed by the addi- Photoprotective effects of dietary antioxidants 1297

Fig. 2. Northern blot showing time kinetic of MMP-1 and HO-1 mRNA expression in human ﬁbroblasts irradiated with UVA (50 J/cm2). The relative levels of MMP-1 (grey bars) and HO-1 (white bars) mRNA were normalized with respect to the control ␤-actin mRNA and the resulting ratio of test mRNA:control mRNA is shown in the bar chart.

Fig. 3. Northern blot showing the effect of vitamin E, vitamin C, and ␤-carotene on MMP-1 mRNA expression in human ﬁbroblasts irradiated with UVA. The initial concentrations of the antioxidants in the culture medium were 2 ␮M vitamin E, 14 ␮M vitamin C, and 0.1–0.5 ␮M ␤-carotene. The MMP-1 mRNA levels were normalized to the control ␤-actin mRNA and illustrated as a bar chart. 1298 E. A. OFFORD et al.

Fig. 4. Northern blot showing the effect of lycopene in combination with vitamin C and/or vitamin E on MMP-1 mRNA expression in human fibroblasts irradiated with UVA. The initial concentrations of the antioxidants in the culture medium were 0.5–2.0 ␮M vitamin E, 14 ␮M vitamin C, and 0.1–1.0 ␮M lycopene. The MMP-1 mRNA levels were normalized to the control ␤-actin mRNA and illustrated as a bar chart. The bar for lanes 8, 9 is not shown as this treatment led to cellular toxicity as shown by the reduced actin mRNA expression. tion of 0.5–2 ␮M vitamin E. Indeed, very strong protec- copene or ␤-carotene were present at 0.1–0.5 ␮M (Fig. tion was seen by the combination of vitamin E, vitamin 6, lanes 9–15). C, and lycopene (Fig. 4, lanes 10–13). The polyphenolic antioxidant, carnosic acid (0.3–3.0 ␮M), showed a dose-dependent inhibition of UVA-in- DISCUSSION duced MMP-1 expression (Fig. 5, lanes 4–9). Indeed, the To overcome some of the technical difficulties of effect of carnosic acid or vitamin E was very similar at 1 studying unstable, lipid-soluble antioxidants in aqueous ␮M concentrations, leading to around 90% inhibition of culture medium, carotenoids were delivered in powdered MMP-1 mRNA expression (data not shown). nanoparticles together with vitamin C and/or vitamin E. The endogenous inhibitor of MMP-1, TIMP-1 [44] The stability, solubility, and cellular uptake of the anti- was expressed at low levels in human fibroblasts GT-F oxidants, especially of lycopene compared to a solution and was only marginally or not induced 24 h after UV of crystalline lycopene in DMSO, was greatly improved irradiation (data not shown). Vitamins and carotenoids (Fig. 1). Importantly, lycopene was further stabilized in did not clearly alter TIMP-1 expression in fibroblasts the presence of vitamin E in the nanoparticles. Carote- irradiated with UVA. noids can undergo oxidative or enzymatic cleavage to form apo-carotenoids [45]. The type and quantity of Regulation of UVA induced HO-1 by antioxidants apo-carotenals/lycopenals formed may vary according to the culture conditions. The observation that the stability The antioxidants vitamin E, vitamin C, and carnosic and cellular uptake of lycopene was improved in the acid, at the low micromolar concentrations used in this presence of vitamin E suggests that vitamin E in the study, had no effect on UVA-induced HO-1 mRNA in nanoparticle protected against oxidative transformation fibroblasts (Fig. 6, lanes 5–8; Fig. 7, lanes 5–12). A of lycopene. small induction of HO-1 was seen (1–2-fold) when ly- In the human dermal skin fibroblast model described Photoprotective effects of dietary antioxidants 1299

Fig. 5. Northern blot showing the effect of carnosic acid (0.3–3.0 ␮M) on MMP-1 mRNA expression in human ﬁbroblasts irradiated with UVA. The MMP-1 mRNA levels were normalized to the control ␤-actin mRNA and illustrated as a bar chart.

Fig. 6. Northern blot showing the effect of antioxidants on HO-1 mRNA expression in human ﬁbroblasts irradiated with UVA. The initial concentrations of antioxidants in the culture medium were 2 ␮M vitamin E, 14 ␮M vitamin C, and 0.1–0.5 ␮M lycopene or ␤-carotene. The HO-1 mRNA levels were normalized to the control ␤-actin mRNA and illustrated as a bar chart. 1300 E. A. OFFORD et al.

Fig. 7. Northern blot showing the effect of carnosic acid (1–3 ␮M) and vitamin E (0.5–1.0 ␮M) on HO-1 mRNA expression in human fibroblasts irradiated with UVA. The HO-1 mRNA levels were normalized to the control ␤-actin mRNA and illustrated as a bar chart. here, using UVA irradiation at 50 J/cm2 (which typically carotenoids in the presence of vitamin E and the protec- corresponds to an average dose inducing one MED in tion against MMP-1 induction, suggesting that vitamin E humans [41]), we observed a significant induction of is acting as an antioxidant to prevent oxidative degrada- HO-1 mRNA from 3 to 6 h and of MMP-1 mRNA from tion of lycopene or ␤-carotene and that the oxidative 24 h to 30 h post irradiation (Fig. 2), in accordance with metabolites may be responsible for the exacerbated UV- previously published literature [2,4]. As the MMP-1 induction of MMP-1. The potential concomitant expres- mRNA levels are known to be correlated with MMP-1 sion of tissue inhibitors of metalloproteinase-1 (TIMP-1) protein expression, the physiological consequence of in- [44] was tested, but there was no correlation (either up- creased MMP-1 expression would be expected to be an or downregulation) with MMP-1 expression, nor with increase of collagen degradation in skin and subsequent respect to the UVA response or with antioxidant treat- formation of wrinkles [1,46]. HO-1, a general stress gene ment (data not shown). that is particularly responsive to oxidative stress, has UVA-induced HO-1 mRNA expression was not af- been shown to be induced following UVA radiation fected by vitamin C, vitamin E, or carnosic acid at the [3,4]. This radiation penetrates readily to dermal tissue same concentrations that were effective for inhibition of and is highly implicated in skin aging. Therefore, human MMP-1 expression (Figs. 6 and 7). Therefore, the dermal skin fibroblasts are a useful model to investigate MMP-1 response appears to be more sensitive to regu- the effect of antioxidants on UVA-induced skin photo- lation by these dietary antioxidants than the HO-1 re- aging. sponse. Probably, the concentrations used were too low Protection against UVA-induction of MMP-1 was to have an effect on HO-1 expression as a previously seen with low micromolar concentrations of vitamin C, published study reported that vitamin C at 50–100 ␮M vitamin E, or carnosic acid. In contrast, ␤-carotene or had only a moderate suppressing effect on HO-1 induc- lycopene (0.1–1.0 ␮M), stimulated MMP-1 expression tion while vitamin E at 10–25 ␮M had a significant (Figs. 3 and 4). This effect was completely suppressed suppressive effect when skin fibroblasts were irradiated when vitamin E was included in the nanoparticle formu- with UVA at 20 J/cm2 [49]. In the same study, HO-1 lation. Beta-carotene and lycopene can act as antioxi- expression was stimulated by ␤-carotene when given dants or pro-oxidants, depending on the experimental alone, but the combination of ␤-carotene (0.5 ␮M), vi- conditions, such as oxygen tension, carotenoid concen- tamin C (100 ␮M), and vitamin E (25 ␮M) suppressed tration, and interactions with other antioxidants, and con- this induction. In the results presented here, HO-1 ex- sequently influence the cellular and molecular responses pression was slightly stimulated by 0.5 ␮M lycopene or [35,47,48]. Our results show an improved stability of ␤-carotene. However, vitamin C or vitamin E had no Photoprotective effects of dietary antioxidants 1301 effect on their own at the concentrations used, nor did the level in plasma. The carotenoids ␤-carotene or lycopene polyphenolic antioxidant, carnosic acid. Physiologically, must be delivered together with vitamin E to prevent HO-1 represents an early response to stress and is in- formation of oxidative derivatives, which may influence duced in a similar way to heat shock proteins as part of the cellular and molecular responses. In the future, com- the cellular defense to environmental and chemical stress binations of natural compounds may be envisaged for [3,4]. The result is to increase the cell’s capacity to deal effective photoprotection. with oxidative damage. Therefore, although the induction of HO-1 mRNA appears to indicate higher oxidative Acknowledgement — The Swiss League Against Cancer (grant KFS- stress, it is not yet clear whether it is an advantage to 695-7-1998) and the Erwin Braun Foundation are thanked for their stimulate or inhibit its expression. contribution to the support of this work. Another polyphenolic antioxidant, epigallocatechin (15 ␮g/ml) from green tea, was shown to inhibit UVA- REFERENCES induced HO-1 mRNA expression in the adult fibroblast 2 [1] Fisher, G. J.; Wang, Z. Q.; Datta, S. C.; Varani, J.; Kang, S.; skin cell line, FEK4, after doses of 250–400 kJ/m UVA Voorhees, J. J. Pathophysiology of premature skin aging induced stimulated MMP-1 expression [50]. The authors suggest by ultraviolet light. N. Engl. J. Med. 337:1419–1428; 1997. that the antioxidant properties of epigallocatechin may [2] Scharffetter, K.; Wlaschek, M.; Hogg, A.; Bolsen, K.; Schothorst, A.; Goerz, G.; Krieg, T.; Plewig, G. UVA irradiation induces be responsible for the decrease in HO-1 expression due collagenase in human dermal fibroblasts in vitro and in vivo. to lowering of oxidative intermediates while the induc- Arch. Dermatol. Res. 283:506–511; 1991. tion of MMP-1 may be due to stimulation of signalling [3] Applegate, L. A.; Luscher, P.; Tyrell, R. M. Induction of heme oxygenase: a general response to oxidant stress in cultured mam- pathways such as c-jun, c-fos and MAP kinases. MMP-1 malian cells. Cancer Res. 51:974–978; 1991. is regulated by several factors including c-jun, c-fos, and [4] Keyse, S. M.; Tyrell, R. M. Heme oxygenase is the major 32-kDa cytokines [51–53]. HO-1 is also transcriptionally regu- stress protein induced in human skin fibroblasts by UVA radiation, hydrogene peroxide, and sodium arsenite. Proc. Natl. Acad. lated, but the factors controlling its regulation are less Sci. USA 86:99–103; 1989. well characterized [54]. [5] Scharffetter-Kochanek, K.; Wlaschek, M.; Briviba, K.; Sies, H. Although the compounds studied here all have pow- Singlet oxygen induces collagenase expression in human skin fibroblasts. FEBS Lett. 331:304–306; 1993. erful antioxidant properties in vitro, their overall biolog- [6] Basu-Modak, S.; Tyrrell, R. M. Singlet oxygen: a primary effector ical functions in vivo have yet to be elucidated. Whether in the ultraviolet A/near-visible light induction of the human they act as antioxidant, anti-inflammatory, enzyme in- heme oxygenase gene. Cancer Res. 53:4505–4510; 1993. [7] Fuchs, J. Potential and limitation of the natural antioxidants RRR hibitor or inducers, regulators of gene expression, or alpha tocopherol, L-ascorbic acid and ␤-carotene in cutaneaous whether they contribute to redox regulation in cells in- photoprotection. Free Radic. Biol. Med. 25:848–873; 1998. dependently of their antioxidant properties, has yet to be [8] Fuchs, J.; Kern, H. Modulation of UV-light-induced skin inflam- mation by D-alpha-tocopherol and L-ascorbic acid: a clinical clarified and will partly depend on their metabolism and study using solar simulated radiation. Free Radic. Biol. Med. local concentration in plasma and tissues [55]. Carote- 25:1006–1012; 1998. noids, beyond their antioxidant properties, stimulate gap [9] Stahl, W.; Heinrich, U.; Jungmann, H.; Sies, H.; Tronnier, H. Carotenoids and carotenoids plus vitamin E protect against ultra- junctional communication in a differential and dose- violet light-induced erythema in humans. Am. J. Clin. Nutr. 71: dependent manner [56–58]. It has been hypothesized that 795–798; 2000. lycopene derivatives may have hormone-like actions at [10] Lee, J.; Jiang, S.; Levine, N.; Watson, R. R. Carotenoid supplementation reduces erythema in human skin after simulated solar low concentrations by acting as ligands for a nuclear radiation exposure. Proc. Soc. Exp. Biol. Med. 223:170–174; receptor, similar to retinoic acid derived from ␤-carotene 2000. [59]. Recent studies with lycopene in human mammary [11] Garmyn, M.; Ribaya-Mercado, J. D.; Russell, R. M.; Bhawan, J.; Gilchrest, B. A. Effect of ␤-carotene supplementation on the cancer cells showed a stimulation of the basal level of human sunburn reaction. Exp. Dermatol. 4:101–111; 1995. AP-1 but an inhibition of its induction by insulin-like [12] Clinton, S. K. Lycopene: chemistry, biology, and implications for growth factor-1 [60]. At the low micromolar concentra- human health and disease. Nutr. Rev. 56:35–51; 1998. [13] Franceschi, S.; Favero, A.; La Vecchia, C.; Negri, E.; Conti, E.; tions used in this study, we observe differential effects of Montella, M.; Giacosa, A.; Nanni, O.; Decarli, A. Food groups the carotenoids, vitamins, and polyphenols on the ex- and risk of colorectal cancer in Italy. Int. J. Cancer 72:56–61; pression of MMP-1 and HO-1 genes. Further studies are 1997. [14] Franceschi, S.; Parpinel, M.; La Vecchia, C.; Favero, A.; Ta- required to investigate the fine signaling pathways af- lamini, R.; Negri, E. Role of different types of vegetables and fruit fected by these different compounds and whether or not in the prevention of cancer of the colon, rectum, and breast. the gene expression effects are linked to their antioxidant Epidemiology 9:338–341; 1998. [15] Giovannucci, E. Tomatoes, tomato-based products, lycopene, and properties. cancer: review of the epidemiologic literature. J. Natl. Cancer In conclusion, our results suggest that vitamin C, Inst. 91:317–331; 1999. vitamin E, and carnosic acid have the potential to be [16] Giovannucci, E.; Ascherio, A.; Rimm, E. B.; Meir, J.; Stampfer, M. J.; Colditz, G. A.; Willett, W. C. Intake of carotenoids and effective photoprotectors. These compounds are effec- retinol in relation to risk of prostate cancer. J. Natl. Cancer Inst. tive at low micromolar concentrations, similar to their 87:1767–1776; 1995. 1302 E. A. OFFORD et al.

[17] Di Mascio, P.; Kaiser, S.; Sies, H. Lycopene as the most efficient [35] Junghans, A.; Sies, H.; Stahl, W. Carotenoid-containing unilamel- biological carotenoid singlet oxygen quencher. Arch. Biochem. lar liposomes loaded with glutathione: a model to study hydro- Biophys. 274:532–538; 1989. phobic-hydrophilic antioxidant interactions. Free Radic. Res. 33: [18] Ribaya-Mercado, J. D.; Garmyn, M.; Gilchrest, B. A.; Russell, 801–808; 2001. R. M. Skin Lycopene is destroyed preferentially over ␤-carotene [36] Horn, D. Preparation and characterization of microdisperse bio- during ultraviolet irradiation in humans. J. Nutr. 125:1854–1858; available carotenoid hydrosols. Angew. Makromol. Chem. 166/ 1995. 167:139–153; 1989. [19] Stahl, W.; Heinrich, U.; Wiseman, S.; Eichler, O.; Sies, H.; [37] Horn, D.; Lueddecke, E. Preparation and characterization of Tronnier, H. Dietary tomato paste protects against ultraviolet nano-sized carotenoid hydrosols. In: Fine Particles Science and light-induced erythema in humans. J. Nutr. 131:1449–1451; Technology. NATO ASI, Ser 3, 12; 1996:761–775. 2001. [38] Auweter, H.; Haberkorn, H.; Heckmann, W.; Horn, D.; Lued- [20] Aruoma, O. I.; Halliwell, B.; Aeschbach, R.; Lo¨ligers, J. Antiox- decke, E.; Rieger, J.; Weiss, H. Supramolecular structure of idant and pro-oxidant properties of active rosemary constituents: precipitated nanosize ␤-carotene particles. Angew. Chem. 111: carnosol and carnosic acid. Xenobiotica 22:257–268; 1992. 2325–2328; 1999. [21] Aruoma, O. I.; Spencer, J. P. E.; Rossi, R.; Aeschbach, R.; Khan, [39] Aeschbach, R.; Philippossian, G. Carnosic acid obtension and A.; Mahmood, N.; Munoz, A.; Murcia, A.; Butler, J.; Halliwell, B. uses. USP 5 256 700. 1993. An evaluation of the antioxidant and antiviral action of extracts of [40] Applegate, L. A.; Frenk, E. Oxidative defense in cultured human rosemary and provencal herbs. Food Chem. Toxicol. 34:449– skin fibroblasts and keratinocytes from sun-exposed and non- 456; 1996. exposed skin. Photodermatol. Photoimmunol. Photomed. 11:95– [22] Huang, S. W.; Frankel, E. N.; Schwarz, K.; Aeschbach, R.; 101; 1995. German, J. B. Antioxidant activity of carnosic acid and methyl [41] Luy, H.; Frenk, E.; Applegate, L. A. Ultraviolet A-induced cel- carnosate in bulk oils and oil-in-water emulsions. J. Agric. Food lular membrane damage in the photosensitivity dermatitis/actinic Chem. 44:2951–2956; 1996. reticuloid syndrome. Photodermatol. Photoimmunol. Photomed. [23] Haraguchi, H.; Saito, T.; Okamura, N.; Yagi, A. Inhibition of lipid 10:126–133; 1994. peroxidation and superoxide generation by diterpenoids from [42] Hess, D.; Keller, H. E.; Oberlin, B.; Bonfati, R.; Schuëp, W. Rosmarinus officinalis. Planta Medica 61:333–336; 1995. Simultaneous determination of retinol, tocopherols, carotenes and [24] Ho, C.-T.; Ferraro, T.; Chen, Q.; Rosen, R. T.; Huang, M.-T. lycopene in plasma by means of high-performance liquid chro- Phytochemicals in teas and rosemary and their cancer-preventive matography on reversed phase. Int. J. Nutr. Res. 61:232–238; properties. In: Ho, C.-T.; Toshihiko, O.; Huang, M.-T.; Rosen, 1991. R. T., eds. Food phytochemicals for cancer prevention II. ACS [43] Goldberg, G. I.; Wilhelm, S. M.; Kronberger, A.; Bauer, E. A.; Symposium Series 547, USA; 1994:2–19. Grant, G. A.; Eisen, A. Z. Human fibroblast collagenase. Com- [25] Huang, M.-T.; Ho, C.-T.; Wang, Z. Y.; Ferraro, T.; Lou, Y.-R.; plete primary structure and homology to an oncogene transfor- Stauber, K.; Ma, W.; Gerogiadis, C.; Laskin, J. D.; Conney, A. H. mation-induced rat protein. J. Biol. Chem. 261:6600–6605; 1986. Inhibition of skin tumorigenesis by rosemary and its constituents [44] Docherty, A. J. P.; Lyons, A.; Smith, B. J.; Wright, E. M.; carnosol and ursolic acid. Cancer Res. 54:701–708; 1994. Stephens, P. E.; Harris, T. J. R. Sequence of human tissue inhib- [26] Tokuda, H.; Ohigashi, H.; Koshimizu, K.; Ito, Y. Inhibitory itor of metalloproteinases and its identity to erythroid-potentiating effects of ursolic acid and oleanic acid on skin tumor promotion activity. Nature 318:66–69; 1985. by 12-O-tetradecanoylphorbol-13-acetate. Cancer Lett. 33:279– [45] Tang, G. W.; Wang, X. D.; Russell, R. M.; Krinsky, N. I. 285; 1986. Characterization of ␤-apo-13-carotenone and ␤-apo-14Ј-carotenal [27] Yasukawa, K.; Takido, M.; Matsumoto, T.; Takeuchi, M.; Naka- as enzymatic products of the excentric cleavage of ␤-carotene. gawa, S. Sterol and triterpene derivatives from plants inhibit the Biochemistry 30:9829–9834; 1991. effects of a tumor promoter, and sitosterol and betulinic acid [46] Fisher, G. J.; Datta, S. C.; Talwar, H. S.; Wang, Z.-Q.; Varani, J.; inhibit tumor formation in mouse skin two-stage carcinogenesis. Kang, S.; Voorhees, J. J. Molecular basis of sun-induced prema- Oncology 48:72–76; 1991. ture skin ageing and retinoid antagonism. Nature 379:335–339; [28] Offord, E. A.; Guillot, F.; Aeschbach, R.; Lo¨liger, J.; Pfeifer, 1996. A. M. A. Antioxidant and biological properties of rosemary [47] Lowe, G. M.; Booth, L. A.; Young, A. J.; Bilton, R. F. Lycopene components: implications for food and health. In: Shahidi, F., ed. and ␤-carotene protect against oxidative damage in HT29 cells at Natural antioxidants: chemistry, health effects, and applications. low concentrations but rapidly lose this capacity at higher doses. Champaign, IL: AOCS Press; 1997:88–96. Free Radic. Res. 30:141–151; 1999. [29] Offord, E. A.; Mace´, K.; Ruffieux, C.; Malnoe¨, A.; Pfeifer, [48] Palozza, P. Prooxidant actions of carotenoids in biological sys- A. M. A. Rosemary components inhibit benzo(a)pyrene-induced tems. Nutr. Rev. 56:257–265; 1998. genotoxicity in human bronchial cells. Carcinogenesis 16(9): [49] Obermu¨ler-Jevic, U. C.; Francz, P. I.; Franck, J.; Flaccus, A.; 2057–2062; 1995. Biesalski, H. K. Enhancement of the UVA induction of haem [30] Offord, E. A.; Mace´, K.; Avanti, O.; Pfeifer, A. Mechanisms oxygenase-1 expression by ␤-carotene in human skin fibroblasts. involved in the chemoprotective effects of rosemary extract stud- FEBS Lett. 460:212–216.; 1999. ied in human liver and bronchial cells. Cancer Lett. 114:1–7; [50] Soriani, M.; Rice-Evans, C.; Tyrrell, R. M. Modulation of the 1997. UVA activation of heme oxygenase, collagenase, and cyclooxy- [31] Xu, X.; Wang, Y.; Constantinou, A. I.; Stacewicz-Sapuntzakis, genase gene expression by epigallocatechin in human skin cells. M.; Bowen, P. E.; Van Breemen, R. B. Solubilization and stabi- FEBS Lett. 439:253–257; 1998. lization of carotenoids using micelles: delivery of lycopene to [51] Stein, B.; Rahmsdorf, H. J.; Steffen, A.; Litfin, M.; Herrlich, P. cells in culture. Lipids 34:1031–1036; 1999. UV-induced DNA damage is an intermediate step in UV-induced [32] Auge´, N.; Santanam, N.; Parthasarathy, S. An efficient method for expression of human immunodeficiency virus type I, collagenase, solubilizing ␤-carotene in aqueous solutions. J. Med. Food 1:39– c-fos and metallothionein. Mol. Cell Biol. 9:5169–5181; 1989. 43; 1998. [52] Wlaschek, M.; Bolsen, K.; Herrmann, G.; Schwarz, A.; Wilmroth, [33] Williams, A. W.; Boileau, T. W.; Clinton, S. K.; Erdman, J. W. Jr. F.; Heinrich, P. C.; Goerz, G.; Scharffetter-Kochanek, K. UVA- ␤-Carotene stability and uptake by prostate cancer cells are de- induced autocrine stimulation of fibroblast-derived-collagenase pendent on delivery vehicles. Nutr. Cancer 36:185–190; 2000. by IL-6: a possible mechanism in dermal photodamage. J. Invest. [34] Pfitzner, I.; Francz, P. I.; Biesalski, H. K. Carotenoid: methyl- Dermatol. 101:164–168; 1993. beta-cyclodextrin formula, an improved method for supplemen- [53] Fisher, G. J.; Voorhees, J. J. Molecular mechanisms of photoag- tation of cultured cells. Biochim. Biophys. Acta 1474:163–168; ing and its prevention by retinoic acid: ultraviolet irradiation 2001. induces MAP kinase signal transduction cascades that induce Photoprotective effects of dietary antioxidants 1303

AP-1 regulated matrix metalloproteinases that degrade human [59] Ben-Dor, A.; Nahum, A.; Danilenko, M.; Giat, Y.; Stahl, W.; skin in vivo. J. Invest. Dermatol. Symposium Proceedings 3:61– Martin, H. D.; Emmerich, T.; Levy, J.; Noy, N.; Levy, J.; Sharoni, 68; 1998. Y. Effects of acyclo-retinoic acid and lycopene on activation of [54] Keyse, S. M.; Applegate, L. A.; Tromvoukis, Y.; Tyrrell, R. M. the retinoic acid receptor and proliferation of mammary cancer Oxidant stress leads to transcriptional activation of the human cells. Arch. Biochem. Biophys. 391:295–302; 2001. heme oxygenase gene in cultured skin fibroblasts. Mol. Cell Mol. [60] Karas, M.; Amir, H.; Fishman, D.; Danilenko, M.; Segal, S.; Biol. 10:4967–4969; 1990. Nahum, A.; Koifmann, A.; Giat, Y.; Levy, J.; Sharoni, Y. Lyco- [55] Rice-Evans, C.; Spencer, J. P. E.; Schroeter, H.; Rechner, A. R. pene interferes with cell cycle progression and insulin-like growth Bioavailability of flavonoids and potential bioactive forms in factor I signaling in mammary cancer cells. Nutr. Cancer 36:101– vivo. Drug Metabol. Drug Interact. 17:291–310; 2000. 111; 2000. [56] Krutovskikh, V.; Asamoto, M.; Takasuka, N.; Murakoshi, M. Differential dose-dependent effects of alpha-carotenes and lycopene on gap-junctional intercellular communication in rat liver in ABBREVIATIONS vivo. Jpn. J. Cancer Res. 88:1121–1124; 1997. [57] Stahl, W.; von Laar, J.; Martin, H. D.; Emmerich, T.; Sies, H. DMSO—dimethylsulphoxide Stimulation of gap junctional communication: comparison of GT-F—human dermal skin fibroblasts acyclo-retinoic acid and lycopene. Arch. Biochem. Biophys. 373: HO-1—heme-oxygenase 1 271–274; 2000. [58] Bertram, J. S. Carotenoids and gene regulation. Nutr. Rev. 57: MMP-1—metalloproteinase 1 182–191; 1999. UVA—ultraviolet radiation A