Portal de Periódicos da CAPES

Benzoyl Peroxide Increases UVA-Induced Plasma Membrane Damage and Lipid Oxidation in Murine Leukemia L1210 Cells

1998; Elsevier BV; Volume: 110; Issue: 1 Linguagem: Inglês

10.1046/j.1523-1747.1998.00091.x

1523-1747

Sally H. Ibbotson, Christopher Lambert, Michael Moran, Mary C. Lynch, Irene E. Kochevar,

Free Radicals and Antioxidants

Ultraviolet A radiation induces oxidative stress and cell damage. The purpose of this investigation was to examine whether ultraviolet A-induced cell injury was amplified by the presence of a non-ultraviolet A absorbing molecule capable of generating free radicals. Benzoyl peroxide was used as a lipid soluble potential radical-generating agent. Plasma membrane permeability assessed by trypan blue uptake was used to measure cell damage in murine leukemia L1210 cells. Cells were irradiated with a pulsed Nd/YAG laser at 355 nm using 0–160 J per cm2. The ratio of the fluence–response slope in the presence of 40 μM benzoyl peroxide to that of irradiated controls was 4.3 ± 2.6. Benzoyl peroxide alone or benzoyl peroxide added after irradiation did not cause increased trypan blue uptake. The ratio of the fluence–response slopes in the presence of 40 μM benzoyl peroxide to that of irradiated controls was 4.7 ± 1.4 when cells were irradiated (0–43 J per cm2) with a xenon lamp, filtered to remove wavelengths 34 J per cm2. The increased trypan blue uptake and thiobarbituric acid reactive substances were inhibited by butylated hydroxytoluene. These results suggest that agents not absorbing ultraviolet A radiation may enhance ultraviolet A-initiated oxidative stress in cells. Ultraviolet A radiation induces oxidative stress and cell damage. The purpose of this investigation was to examine whether ultraviolet A-induced cell injury was amplified by the presence of a non-ultraviolet A absorbing molecule capable of generating free radicals. Benzoyl peroxide was used as a lipid soluble potential radical-generating agent. Plasma membrane permeability assessed by trypan blue uptake was used to measure cell damage in murine leukemia L1210 cells. Cells were irradiated with a pulsed Nd/YAG laser at 355 nm using 0–160 J per cm2. The ratio of the fluence–response slope in the presence of 40 μM benzoyl peroxide to that of irradiated controls was 4.3 ± 2.6. Benzoyl peroxide alone or benzoyl peroxide added after irradiation did not cause increased trypan blue uptake. The ratio of the fluence–response slopes in the presence of 40 μM benzoyl peroxide to that of irradiated controls was 4.7 ± 1.4 when cells were irradiated (0–43 J per cm2) with a xenon lamp, filtered to remove wavelengths 34 J per cm2. The increased trypan blue uptake and thiobarbituric acid reactive substances were inhibited by butylated hydroxytoluene. These results suggest that agents not absorbing ultraviolet A radiation may enhance ultraviolet A-initiated oxidative stress in cells. butylated hydroxytoluene benzoyl peroxide trypan blue thiobarbituric acid reactive substances Ultraviolet A (UVA) causes damage to cells and skin partly by the action of reactive oxygen species. UVA-induced erythema and pigmentation are oxygen-dependent (Auletta et al., 1986Auletta M. Gange R.W. Tan O.T. Matzinger E. Effect of cutaneous hypoxia upon erythema and pigment responses to UVA, UVB, and PUVA (8-MOP + UVA) in human skin.J Invest Dermatol. 1986; 86: 649-652Abstract Full Text PDF PubMed Scopus (44) Google Scholar) and there is ample evidence to support the generation of reactive oxygen species in cells by UVA irradiation (Basu-Modak and Tyrrell, 1993Basu-Modak S. Tyrrell R.M. Singlet oxygen: a primary effector in the ultraviolet A/near-visible light induction of the human heme oxygenase gene.Cancer Res. 1993; 53: 4505-4510PubMed Google Scholar, Vile and Tyrrell, 1995Vile G.F. Tyrrell R.M. UVA radiation-induced oxidative damage to lipids and proteins in vitro and in human skin fibroblasts is dependent on iron and singlet oxygen.Free Radical Biol Med. 1995; 18: 721-730Crossref PubMed Scopus (253) Google Scholar). Oxidative damage to lipids, changes in cutaneous antioxidants, and membrane damage are induced by UVA (Punnonen et al., 1991Punnonen K. Jansen C.T. Puntala A. Ahotupa M. Effects of in vitro UVA irradiation and PUVA treatment on membrane fatty acids and activities of antioxidant enzymes in human keratinocytes.J Invest Dermatol. 1991; 96: 255-259Abstract Full Text PDF PubMed Google Scholar, Gaboriau et al., 1993Gaboriau F. Morliere P. Marquis I. Moysan A. Geze M. Dubertret L. Membrane damage induced in cultured human skin fibroblasts by UVA irradiation.Photochem Photobiol. 1993; 58: 515-520Crossref PubMed Scopus (61) Google Scholar, Moysan et al., 1993Moysan A. Marquis I. Gaboriau F. Santus R. Dubertret L. Morliere P. Ultraviolet A-induced lipid peroxidation and antioxidant defense systems in cultured human skin fibroblasts.J Invest Dermatol. 1993; 100: 692-698Abstract Full Text PDF PubMed Google Scholar, Cohen and DeLeo, 1993Cohen D. DeLeo V.A. Ultraviolet radiation-induced phospholipase A2 activation in mammalian cell membrane preparations.Photochem Photobiol. 1993; 57: 383-390Crossref PubMed Scopus (37) Google Scholar). For example, singlet oxygen is implicated in the inactivation of human skin fibroblasts and damage to lipid and proteins by UVA (Tyrrell and Pidoux, 1989Tyrrell R.M. Pidoux M. Singlet oxygen involvement in the inactivation of cultured human fibroblasts by UVA (334 nm, 365 nm) and near-visible (405 nm) radiations.Photochem Photobiol. 1989; 49: 407-412Crossref PubMed Scopus (153) Google Scholar, Vile and Tyrrell, 1995Vile G.F. Tyrrell R.M. UVA radiation-induced oxidative damage to lipids and proteins in vitro and in human skin fibroblasts is dependent on iron and singlet oxygen.Free Radical Biol Med. 1995; 18: 721-730Crossref PubMed Scopus (253) Google Scholar). Hydroxyl radical (Keyse and Tyrrell, 1989Keyse S.M. Tyrrell R.M. Induction of the heme oxygenase gene in human skin fibroblasts by hydrogen peroxide and UVA (365 nm) radiation: evidence for the involvement of the hydroxyl radical.Carcinogenesis. 1989; 11: 787-791Crossref Scopus (103) Google Scholar) and singlet oxygen (Basu-Modak and Tyrrell, 1993Basu-Modak S. Tyrrell R.M. Singlet oxygen: a primary effector in the ultraviolet A/near-visible light induction of the human heme oxygenase gene.Cancer Res. 1993; 53: 4505-4510PubMed Google Scholar) are involved in the UVA induction of the heme oxygenase gene. Some of the cytotoxic effects of UVA/UVB are thought to be mediated via hydrogen peroxide (Bertling et al., 1996Bertling C.J. Lin F. Girotti A.W. Role of hydrogen peroxide in the cytotoxic effects of UVA/B radiation on mammalian cells.Photochem Photobiol. 1996; 64: 137-142Crossref PubMed Scopus (44) Google Scholar). Thus, many of the physiologic and pathologic effects of UVA irradiation are mediated by the production of reactive oxygen species. This study was performed to test the hypothesis that UVA-induced damage to membranes could be amplified using an agent with potential for generation of free radicals but which does not absorb UVA radiation. Organic peroxides are good candidates for such agents because they decompose to form free radicals. For this study, benzoyl peroxide (BPO) was chosen because it is highly lipophilic, is expected to localize in cell membranes (a site of UVA-induced effects), and does not absorb UVA radiation. In addition, BPO is frequently used for treatment of acne on facial skin that is exposed to sunlight. In solution, decomposition of BPO to radicals can be induced by reaction with radicals (Denney and Feig, 1959Denney D.B. Feig G. A study of the induced decomposition of benzoyl peroxide in diethylether.J Am Chem Soc. 1959; 81: 5322-5324Crossref Scopus (13) Google Scholar); thus, it can participate in free radical chain reactions. In addition, in the absence of irradiation, BPO generates radicals that are believed to be involved in epidermal proliferation (Klein-Szanto and Slaga, 1982Klein-Szanto A.J.P. Slaga T.J. Effects of peroxides on rodent skin: epidermal hyperplasia and tumor promotion.J Invest Dermatol. 1982; 79: 30-34Abstract Full Text PDF PubMed Scopus (177) Google Scholar) as well as tumor promotion in murine skin after initiation with a chemical carcinogen (Slaga et al., 1981Slaga T.J. Klein-Szanto A.J.P. Triplett L.L. Yotti L.P. Trosko J.E. Skin tumor-promoting activity of benzoyl peroxide, a widely used free radical generating compound.Sci. 1981; 213: 1023-1025Crossref PubMed Scopus (497) Google Scholar). In this study, we compared the effect of UVA radiation with that produced by nontoxic concentrations of BPO plus UVA radiation on plasma membrane damage and lipid oxidation in murine leukemia L1210 cells. Alpha minimal essential medium, horse serum, Hanks' balanced salt solution (HBSS) and trypan blue (TB) (0.4%) were obtained from Gibco BRL (Grand Island, NY). Trolox and vitamin E were from Aldrich (Milwaukee, WI). Butylated hydroxytoluene (BHT), 1,1,3,3-tetramethoxypropane, and 2-thiobarbituric acid were from Sigma (St. Louis, MO). Benzoyl peroxide (BPO) (75% in water) was obtained from Lancaster Synthesis (Windham, NH). Two light sources were used for this study. Initial experiments were performed using the frequency-tripled output of a Quantel YG660 A Nd/YAG laser at 355 nm, 5 Hz (Continuum, Santa Clara, CA). The unfocused incident beam (8 ns pulse duration) was transmitted through a UG11 filter, attenuated to 10 mJ per pulse by neutral density filters, unless stated otherwise, and passed through a 6 mm diameter aperture to irradiate the 1 cm cuvette. The irradiance at the sample surface was monitored using a Scientech 365 power and energy meter. Fluences of 0–160 J per cm2 were delivered by varying the number of pulses. Subsequent experiments were performed using a 1000 W xenon lamp (Oriel, Stratford, CT). The spectral output was initially filtered through 1% copper sulfate/1% cobalt sulfate to eliminate wavelengths <320 nm and later through water and glass with a Schott glass KG-11 heat filter. Figure 1 shows the spectra of the irradiation sources. The filtered irradiances at the cuvette were 11–18 mW per cm2through the copper sulfate/cobalt sulfate filter and 2 mW per cm2 through the water and glass filters. Irradiance was measured using an International Light SED 033 probe with an IL1700 radiometer, with wavelength sensitivity of 310–390 nm and maximum sensitivity at 360 nm (Newburyport, MA). Spectral output of the lamp was monitored using a spectroradiometer (Model 742, Optronics, Orlando, Florida). Fluences of 0–43 J per cm2 were used. Murine L1210 lymphocytic leukemia cells, obtained from the American Type Culture Collection (Rockville, MD), were grown in alpha minimal essential medium supplemented with 10% horse serum in plastic tissue culture flasks (175 cm2), at 37°C in a humidified atmosphere of 95% air and 5% carbon dioxide. Medium was changed every 1–2 d to maintain cell concentrations between 0.2 and 1.0 × 106 per ml, with cell doubling times of 8–12 h. Experiments were performed on cells in exponential growth phase. BPO and antioxidant solutions were prepared on the day of usage. BPO was dissolved in HBSS and filtered solutions were diluted to the desired concentration using an absorption coefficient of 1.33 × 104 M−1 cm−1 at 224 nm. BHT and vitamin E were dissolved in ethanol and diluted into the culture medium; the maximum final concentration of ethanol was 0.1%. Trolox was dissolved in cell medium. For dark toxicity studies, cells were seeded at 0.2 × 106 per ml in medium with 10% horse serum and separately in the presence of serial dilutions of ethanol (0–10%), BPO (0–80 mM), BHT (0–151 mM), vitamin E (0–151 mM), or trolox (0–1 mM), and incubated for 24 h at 37°C prior to assessment of plasma membrane permeability by the TB exclusion assay. Subsequent experiments were performed at concentrations of each agent that caused <5% TB uptake. In experiments testing the effect of antioxidants, cells were cultured in medium supplemented with the antioxidant for 24 h before irradiation. Cells were harvested, washed twice with HBSS, resuspended in HBSS with or without BPO (40 or 80 mM) at 3 × 106 per ml (5–6 × 106 per ml for lipid oxidation experiments), and kept at room temperature for 30 min before irradiation. Aliquots of cell suspensions (1.2 ml for TB assay and 2 ml for lipid oxidation studies) in HBSS with or without BPO were transferred to a 1 cm2 quartz cuvette containing a magnetic stir bar. Irradiations were performed on stirred samples at room temperature and nonirradiated stirred samples were used as controls. Immediately after irradiation, the cell samples were transferred to 1.5 ml microcentrifuge tubes, centrifuged, and the pellet resuspended in 1 ml of medium with 10% horse serum. The samples were transferred to 6 well plates and a further 2 ml of medium was added to each well. The cell suspensions were gently mixed before incubation at 37°C, 95% air and 5% carbon dioxide. TB uptake by cells was used to measure plasma membrane disruption by UVA, BPO, and UVA plus BPO. Optimal experimental conditions to detect changes in TB uptake were determined by varying the incubation time at 37°C, the incubation temperature, and the incubation time in TB. On the basis of these findings, assessments were performed after incubation for 4.5 h at 37°C and after 4 min incubation in TB. Assay of thiobarbituric acid-reactive substances (TBARS) was used as an index of lipid oxidation. Aliquots of cell suspension (2 ml at 5–6 × 106 per ml in HBSS with or without 80 mM BPO) were irradiated with UVA from a filtered xenon source as described earlier. Immediately after irradiation, three 0.5 ml aliquots of cell suspension were transferred to glass tubes and incubated at 37°C for 1 h. Thiobarbituric acid (0.03% wt/vol) in 0.9 M sodium acetate buffer (pH 3.5) with 3% sodium dodecyl sulfate (wt/vol) and BHT (0.01% wt/vol) (1.5 ml) was added to each sample at the end of the incubation and tubes kept on ice until all samples were ready for assay. Tubes were heated at 100°C for 30 min in a circulating water bath, prior to cooling by agitation in water at room temperature. After centrifugation, fluorescence of the supernatant was measured using a SPEX fluoromax spectrofluorometer (SPEX Industries, Edison, NJ) lex 532 nm, lem 548 nm, 1 mm slit widths, and excitation at 90°C. A calibration curve was made using 1,1,3,3-tetramethoxypropane and measurements were expressed in terms of pmol malondialdehyde normalized to cell protein content measured by the Micro BCA protein assay (Pierce, Rockford, IL). Values are given as mean ± SD unless stated otherwise. The significance of the data was determined by Student's t test (two-tailed). As shown in Fig 1, the absorption spectrum of BPO in buffer does not overlap with either the laser wavelength at 355 nm or the spectral output of the xenon lamp with either filter. The absorption spectrum of BPO was not altered by irradiation at 355 nm up to a fluence of 330 J per cm2, nor by incubation at 37°C for 24 h. The relationship between TB uptake and fluence of 355 nm irradiation in the presence and absence of BPO (40 or 80 μM) was established. In these experiments, one sample was irradiated at each fluence. The long exposure times (e.g., 15 min to deliver 150 J per cm2) precluded multiple samples at each fluence. Three fluence–response experiments were performed using 40 μM BPO and two using 80 μM. Figure 2 shows a representative fluence-response plot using 40 μM BPO and a large range of fluences. In experiments using a smaller range of fluences, a linear fluence–response relationship was found for fluences producing uptake of TB in less than 20% of the cells. The slopes at low fluences were calculated for samples with and without BPO and the ratio of these slopes was determined for each experiment. The mean ± SD for these ratios was 4.3 ± 2.6 (n = 3) for experiments using 40 μM BPO; as expected, the ratios of the slopes were higher, 11.4 and 24, for the two experiments using 80 μM BPO. Dark controls containing BPO and samples without BPO showed the same TB uptake. To determine whether BPO associated with the cells or BPO in solution was responsible for the enhanced plasma membrane damage upon irradiation, the effect of removal of BPO from the solution before irradiation was examined. Cells were incubated with BPO in the usual manner, and then centrifuged and resuspended in HBSS without BPO before irradiation. BPO-enhanced uptake of TB was of a similar magnitude when BPO was removed from the solution immediately before irradiation and when BPO was present in solution during irradiation (single experiment; data not shown). Additionally, when BPO (80 μM) was added immediately after irradiation to samples not containing BPO, TB uptake was the same as in control samples at all of the fluences examined (data not shown). The effect of varying the irradiance was examined by changing the energy/pulse at which a fluence of 55 J per cm2 was delivered. The TB uptake of irradiated cells (21 + 3%) in the presence of BPO was higher than that in irradiated controls (4 + 3%) and was not influenced by the peak irradiance, at seven pulse energies from 1.5 × 106 W per cm2 (2.5 mJ per pulse) to 21 × 106 W per cm2 (36 mJ per pulse) (n = 2) at which the fluence was delivered. A filtered xenon lamp was used to establish a fluence– response relationship using a continuous wave source with lower irradiance. Fluences of 0–43 J per cm2 were used and, as for irradiation at 355 nm, linear fluence–response plots were obtained at fluences producing TB uptake in less than 20% of the cells. The ratio of the fluence–response slopes in the presence of BPO (40 μM) compared with irradiated control samples was 4.7 ± 1.4 (n = 4) (data not shown). In contrast to irradiation of control samples at 355 nm where less than 5% of the cells showed TB uptake at a fluence of 20 J per cm2, the broad-band UVA source typically produced 12% TB uptake at the same fluence. To determine whether antioxidants altered the TB uptake of irradiated L1210 cells in the presence and absence of BPO (40 μM), three agents were used. Two lipid-soluble antioxidants, BHT (23 and 45 μM) and vitamin E (23 and 45 μM), and a water-soluble vitamin E derivative, trolox (150 and 300 μM), were used. Cells were incubated with the appropriate concentration of antioxidant for 24 h before irradiation. The antioxidant concentrations used were nontoxic and did not affect cell doubling time. Three independent irradiations were employed for each condition. All three antioxidants independently inhibited the TB uptake by cells treated with BPO and 63–93 J per cm2 of 355-nm radiation. Results from one experiment with each antioxidant are shown in Fig 3. The inhibition increased with concentration of antioxidant. In the experiments shown, BHT reduced the TB uptake by cells irradiated in the presence of BPO by 57 and 86% at 23 and 45 μM, respectively; vitamin E reduced the TB uptake by 50 and 78% at 23 and 45 μM, respectively; and Trolox reduced TB uptake by 63 and 79% at 150 and 300 μM, respectively. These experiments were repeated at the higher antioxidant concentrations and showed reductions in TB uptake, by cells irradiated in the presence of BPO, of 91% with BHT (45 μM), 77% with vitamin E (45 μM), and 81% with trolox (300 μM). A similar effect was obtained for inhibition by BHT when samples were irradiated with the xenon lamp (41 J per cm2). BHT inhibited TB uptake by cells irradiated in the presence of BPO (41 and 24% TB uptake by cells in the absence of BHT and 10 and 10%, respectively, when 45 μM BHT was present in the two experiments performed). An increase in TBARS was detectable in cell suspensions 1 h after irradiation with broad-band UVA in the presence of BPO in all nine experiments performed. In each experiment, samples in the presence and absence of BPO were exposed to one or two fluences of UVA, and one sample contained BPO but was unirradiated. Untreated samples (no BPO, no UVA) showed a low and variable level of TBARS (0.25 ± 0.07 pmol malondialdehyde per μg of protein) (n = 3). Samples incubated with BPO but not exposed to UVA showed an average increase in TBARS of 28 ± 14% (p = 0.03, n = 11). Exposure of cell samples to UVA in the absence of BPO did not significantly increase the TBARS (23 ± 30%, n = 11). Samples treated with BPO and exposed to UVA showed greater increases in TBARS than the sum of the separate BPO and UVA-induced increases. These differences increased as the UVA fluence increased and were nonsignificant at doses <34 J per cm2 (n = 5), but were significant at higher fluences (p = 0.01, n = 6) (Fig 4). In two experiments, the effect of BHT on formation of TBARS was evaluated. The results of one experiment are summarized in Fig 5. In both experiments, UVA exposure alone increased TBARS (17 and 66%) compared with the unirradiated controls. BPO also increased TBARS (25 and 56%) compared with the untreated samples. Treatment with BPO plus UVA (41 J per cm2) increased the TBARS by 234 and 110% compared with the irradiated control samples, and 45 μM BHT abolished this increase. In the same two experiments UVA exposure, but not BPO, increased the percentage of cells showing TB uptake (from 4 to 7% and 1 to 21%, respectively). This effect was increased by treatment with the combination of BPO and UVA (24 and 41%, respectively). BHT decreased the percentage of TB positive cells to 10% in both experiments (Fig 5). The results of this study indicate that the presence of BPO enhances UVA-induced damage to cells as measured by permeability of the plasma membrane to a charged dye, trypan blue, and oxidation of unsaturated cellular lipids. BPO is expected to accumulate in all cell membranes, not only the plasma membrane, because it is highly lipophilic. Thus, treatment with UVA plus BPO may alter other cell functions not assessed in this study. Our results also indicate that cell-associated BPO and not BPO in the external medium is responsible for these UVA-induced effects. It is well established that exposure of cells and skin to UVA produces many responses ranging from induction of gene products, such as collagenase (Wlaschek et al., 1995Wlaschek M. Briviba K. Stricklin G.P. Sies H. Scharffetter-Kochanek K. Singlet oxygen may mediate the ultraviolet A-induced synthesis of interstitial collagenase.Invest Dermatol. 1995; 104: 194-198Crossref PubMed Scopus (181) Google Scholar) and heme oxygenase (Basu-Modak and Tyrrell, 1993Basu-Modak S. Tyrrell R.M. Singlet oxygen: a primary effector in the ultraviolet A/near-visible light induction of the human heme oxygenase gene.Cancer Res. 1993; 53: 4505-4510PubMed Google Scholar), to chronic photodamage (Lavker and Kaidbey, 1997Lavker R. Kaidbey K. The spectral dependence for UVA-induced cumulative damage in human skin.J Invest Dermatol. 1997; 108: 17-21Abstract Full Text PDF PubMed Scopus (108) Google Scholar) and photocarcinogenesis (Staberg et al., 1983Staberg B. Wulf H.C. Klemp P. Poulsen T. Brodthagen H. The carcinogenic effect of UVA irradiation.J Invest Dermatol. 1983; 81: 517-519Abstract Full Text PDF PubMed Scopus (149) Google Scholar). Reactive oxygen species and/or other radicals appear to be involved in the mechanisms for UVA-induced responses in vitro because they are inhibited by radical quenchers and antioxidants (Gaboriau et al., 1993Gaboriau F. Morliere P. Marquis I. Moysan A. Geze M. Dubertret L. Membrane damage induced in cultured human skin fibroblasts by UVA irradiation.Photochem Photobiol. 1993; 58: 515-520Crossref PubMed Scopus (61) Google Scholar). In our studies, polychromatic UVA radiation, in the absence of BPO, produced greater membrane permeability than 355-nm laser radiation at the same fluences. This difference is not caused by the much higher intensity of the laser radiation because varying the laser intensity over a 14-fold range at a constant fluence did not alter the percentage of cells excluding dye. The greater effectiveness of the polychromatic UVA to cause membrane permeability may be attributed to the inclusion of wavelengths shorter than 355 nm in this waveband because the action spectrum for lipid oxidation, measured as TBARS, is highest at the shortest UV wavelengths (Morliere et al., 1995Morliere P. Moysan A. Tirache I. Action spectrum for UV-induced lipid peroxidation in cultured human skin fibroblasts.Free Radical Biol Med. 1995; 19: 365-371Crossref PubMed Scopus (91) Google Scholar). An action spectrum for membrane permeability is not available but lipid oxidation is associated with membrane permeability. According to the action spectrum for TBARS formation, 325-nm radiation is approximately twice as effective as 350-nm radiation in cultured human fibroblasts. The fluences used in our study (0–43 J per cm2) are similar to those employed by Morliere et al., 1995Morliere P. Moysan A. Tirache I. Action spectrum for UV-induced lipid peroxidation in cultured human skin fibroblasts.Free Radical Biol Med. 1995; 19: 365-371Crossref PubMed Scopus (91) Google Scholar, who found measurable lipid oxidation at UVA doses <10 J per cm2. Antioxidants protect biomolecules against oxidative damage induced by UVA (Hu and Tappel, 1992Hu M.-L. Tappel A.L. Potentiation of oxidative damage to proteins by ultraviolet-A and protection by antioxidants.Photochem Photobiol. 1992; 56: 357-363Crossref PubMed Scopus (46) Google Scholar). For example, lipid oxidation and immune suppression caused by UVA irradiation of human epidermal cells in vitro have been reported to be reduced by vitamin E (Clement-Lacroix et al., 1996Clement-Lacroix P. Michel L. Moysan A. Morliere P. Dubertret L. UVA-induced immune suppression in human skin: protective effect of vitamin E in human epidermal cells in vitro.Br J Dermatol. 1996; 134: 77-84Crossref PubMed Scopus (66) Google Scholar). Our results are similar; BHT protected cells against UVA-induced increases in membrane permeability and lipid oxidation in the absence of BPO (Fig 5). The mechanism for the enhancement of UVA-induced effects in L1210 cells by BPO does not involve absorption of UV radiation by BPO in the cells because the absorption spectrum of BPO does not overlap with the emission spectra of the light sources used (Fig 1). In cells, where BPO is expected to be in a lipophilic environment, it is even less capable of absorbing UVA radiation because the BPO spectrum in nonpolar organic solvents shifts to shorter wavelengths than in aqueous solution (data not shown). BPO enhancement of UVA-induced cell damage involves oxidative processes because the increase in plasma membrane permeability was inhibited by antioxidants (Fig 3). BHT (45 μM), vitamin E (45 μM), and trolox (300 μM) reduced the uptake of TB by up to 7-fold. The observation that a water-soluble antioxidant, trolox, as well as lipid-soluble antioxidants, was effective indicates that radicals in both the bilayer and the cytoplasm are important for the effect of BPO. Although the radicals formed from BPO (and the lipid radicals subsequently formed) are expected to reside in the membranes, the ability of water-soluble trolox to inhibit oxidation suggests that it interacts at the membrane surface with the stable radicals formed from BHT and vitamin E. This process regenerates the antioxidant capacity of the membrane-bound antioxidants. This type of transfer of antioxidant capacity from membranes to the aqueous phase has been well studied (Doba et al., 1985Doba T. Burton G.W. Ingold K.U. Antioxidant and co-antioxidant activity of vitamin C. The effect of vitamin C, either alone or in the presence of vitamin E or a water soluble vitamin E analogue, upon the peroxidation of aqueous multilamellar phospholipid liposomes.Biochim Biophys Acta. 1985; 835: 298-303Crossref PubMed Scopus (391) Google Scholar, Buettner, 1993Buettner G.R. The pecking order of free radicals and antioxidants: lipid peroxidation, alpha-tocopherol, and ascorbate.Arch Biochem Biophys. 1993; 300: 535-543Crossref PubMed Scopus (2058) Google Scholar). A relevant example is the decrease in the ratio of oxidized to reduced gluathione, another water-soluble antioxidant, that occurs in cells incubated with BPO (Perchellet et al., 1986Perchellet J.P. Perchellet E.M. Orten D.K. Schneider B.A. Decreased ratio of reduced/oxidized glutathione in mouse epidermal cells treated with tumor promoters.Carcinogenesis. 1986; 7: 503-506Crossref PubMed Scopus (50) Google Scholar). The effectiveness of trolox at inhibiting UVA plus BPO-induced uptake of TB also indicates that BHT and vitamin E do not inhibit an increase in membrane permeability simply by physically stabilizing the bilayer. BHT also reduced the formation of TBARS in samples irradiated in the presence of BPO; the decrease was greater than the effect of BHT either on unirradiated BPO samples or on samples irradiated in the absence of BPO (Fig 5). Similar concentrations of antioxidants have been shown to inhibit UVA-induced plasma membrane damage in fibroblasts (Gaboriau et al., 1993Gaboriau F. Morliere P. Marquis I. Moysan A. Geze M. Dubertret L. Membrane damage induced in cultured human skin fibroblasts by UVA irradiation.Photochem Photobiol. 1993; 58: 515-520Crossref PubMed Scopus (61) Google Scholar) and UVB-induced oxidative damage to DNA in keratinocytes (Stewart et al., 1996Stewart M.S. Cameron G.S. Pence B.C. Antioxidant nutrients protect against UVB-induced oxidative damage to DNA of mouse keratinocytes in culture.J Invest Dermatol. 1996; 106: 1086-1089Crossref PubMed Scopus (132) Google Scholar). One previous study of the effects of BPO and UVA on human keratinocytes also reported increased lipid oxidation but employed BPO doses (1 mM) that were toxic under the experimental conditions, making the results difficult to interpret (Kappus and Artuc, 1987Kappus H. Artuc M. Effects of organic peroxides on human epidermal keratinocytes.Bioelectrochem Bioenerget. 1987; 18: 263-270Crossref Scopus (8) Google Scholar). A possible mechanism for the BPO enhancement of UVA-induced damage, which does not require photon absorption by BPO, is that the radicals (R·) produced in cells by UVA from endogenous chromophores react with BPO (eqn 1). The resulting benzoyloxy radical (BO·) or its decomposition product, the phenyl radical (Ph·) (eqn 2), may facilitate a chain oxidation reaction. In this mechanism, the number of radicals is not increased (one R· radical is converted into one phenyl radical) but the phenyl radical is very reactive, diffuses in the bilayer more easily than phospholipid-derived radicals, and may initiate an oxidative chain reaction by reaction with the unsaturated lipids in the bilayer leading to many altered lipid molecules.R⋅+BPO→BO⋅+BO‐R(1) BO⋅→Ph⋅+CO2(2) An alternative molecular mechanism is that radicals produced by decomposition of BPO without light deplete the antioxidant capacity of the cells. BPO may also decompose by a metal ion catalyzed reaction which produces BO· (eqn 3). The oxidative stress initiated by the radicals may be responsible for the decrease in intracellular ratio of reduced to oxidized glutathione reported after incubation of cells with BPO (Perchellet et al., 1986Perchellet J.P. Perchellet E.M. Orten D.K. Schneider B.A. Decreased ratio of reduced/oxidized glutathione in mouse epidermal cells treated with tumor promoters.Carcinogenesis. 1986; 7: 503-506Crossref PubMed Scopus (50) Google Scholar), and for the observation that cells pretreated with the glutathione-depleting agent, chlorodinitrobenzene, were more sensitive to BPO (50 μM) (Babich et al., 1996Babich H. Zuckerbraun H.L. Wurzburger B.J. Rubin Y.L. Borenfreund E. Blau L. Benzoyl peroxide cytotoxicity evaluated in vitro with the human keratinocyte cell line, RHEK-1.Toxicol. 1996; 106: 187-196Crossref PubMed Scopus (55) Google Scholar).BPO+M2+→BO⋅+BO–+M3+(3) Exposure of cells to UVA alone also depletes cellular antioxidants (Fuchs et al., 1989Fuchs J. Huflejt M.E. Rothfuss L.M. Wilson D.S. Carcamo G. Packer L. Acute effects of near ultraviolet and visible light on the cutaneous antioxidant defense system.Photochem Photobiol. 1989; 50: 739-744Crossref PubMed Scopus (128) Google Scholar). Thus, the net effect of exposing cells to nontoxic doses of UVA and BPO (i.e., doses that do not separately cause significant plasma membrane permeability and lipid oxidation) in combination, may be to overwhelm the cellular antioxidant systems. In our experiments, the preincubation period was 30 min at room temperature. The amount of BPO decomposition occurring under these conditions that could cause oxidation stress may be very low, although lipid oxidation in the presence of BPO was greater than that in untreated control samples during incubation in the dark (Fig 5). It is not possible from our results to discriminate between these mechanisms (in fact, both may operate); however, the basic phenomena is clear, namely that an agent that does not absorb UVA radiation may enhance UVA-induced damage in cells. This mechanism for UVA-induced tissue damage should be considered for agents that are being developed for treatment of sun-exposed skin; although for BPO, which is used extensively for treatment of acne on facial skin, considerable investigation has not demonstrated long-term consequences (Iversen, 1988Iversen O.H. Skin tumorigenesis and carcinogenesis studies with 7,12-dimethylbenz[a]anthracene, ultraviolet light, benzoyl peroxide (Panoxyl gel 5%) and ointment gel.Carcinogenesis. 1988; 9: 803-809Crossref PubMed Scopus (20) Google Scholar, Binder et al., 1995Binder R.L. Aardema M.J. Thompson E.D. Benzoyl peroxide: review of experimental carcinogenesis and human safety data.Prog Clin Biol Res. 1995; 391: 245-294PubMed Google Scholar). This work was supported by the National Institute of General Medical Sciences, National Institutes of Health, Grant R01 GM30755 to IEK.