There has been considerable recent interest in the chemopreventive properties of the polyphenols or catechins of green tea that have been shown to inhibit tumorigenesis in a variety of organs in rodent models.1, 2, 3 The major polyphenol component by mass, (−)-epigallocatechin-3-gallate (EGCG) is an effective protectant against the mutagenicity of a number of chemical carcinogens including benzo[a]pyrene (B[a]P), aflatoxin B1 and 3-hydroxyamino-1-methyl-5H-pyrido[4,3-b]indole.4, 5, 6 EGCG, together with other green tea polyphenols also exhibits growth inhibitory properties in a variety of tumour cell lines.7, 8, 9 Several mechanisms have been proposed by which these compounds exert their anti-tumorigenic action: these include the blockade of growth factors binding to their receptors,10 phosphorylation (activation) of mitogen-activated protein kinases (MAPKs)11, 12 possibly resulting in the observed induction of Phase II drug-metabolizing enzymes,13, 14 and the generation of oxidative stress leading to apoptosis.7, 9
In the face of an oxidative challenge, however, it is now well documented that green tea catechins act as antioxidants. For example, EGCG abrogates oxidation by hydrogen peroxide both in a cell free system15 and in terms of DNA single-strand break damage.16 In addition, it protects against generation of the mutagenic base 8-hydroxyguanine17 and genotoxic effects of the superoxide generator paraquat.18 Although the ability of EGCG to prevent oxidative DNA damage may reflect its direct scavenging of ROS as reported for hydroxyl and superoxide radicals,19 it has been proposed recently that its mode of action may also involve chemical repair of radicals formed within the DNA structure itself.20
Although skin carcinogenesis has long been associated with the UVB portion of the solar spectrum, there is an increasing appreciation of the potential contribution of UVA to tumorigenesis particularly with regard to malignant melanoma.21, 22 Moreover, the increased abundance (90% of UV radiation reaching the earth's surface) and penetration of UVA to the actively dividing basal layer of skin tissue relative to UVB points to a greater relative mutagenic efficacy than that observed in cell monolayers in the laboratory.23 In our study, we have examined the role of EGCG in protecting against the damaging effects of UVA solar radiation that exerts its effects on the skin via oxidative mechanisms. We and others have shown previously that ablation of the endogenous antioxidant glutathione enhances the level of intracellular ROS24 and the cytotoxicity generated by UVA irradiation.25
Using an immortalized human keratinocyte line (HaCaT), we show that the green tea polyphenol EGCG reduces the production of certain UVA-induced ROS, can decrease UVA cell-killing and protect DNA from the oxidative damage of UVA and its mutagenic consequences. Our findings can be dissociated from any “sunscreen” effect of EGCG because it does not exhibit any absorbance within the UVA wavebands of our radiation sources.
MATERIAL AND METHODS
The green tea polyphenol, EGCG and 6-thioguanine (6-TG) were purchased from Sigma (St. Louis, MO). SYBR Green I stain (10,000 × concentration, Molecular Probes, Eugene, OR) was obtained from Flowgen (Leicestershire, UK). The fluorogenic probes, 2′,7′,-dichlorodihydrofluorescein diacetate (DCFH-DA) and dihydrorhodamine 123 (DHR) (Molecular Probes) were obtained from Cambridge Biosciences (Cambridge, UK).
Cell culture and treatment
HaCaT cells (a kind gift from N.E. Fusenig) were maintained in DMEM, supplemented with 5% (v/v) FCS and penicillin (100 U/ml)/streptomycin (100 μg/ml). For EGCG treatment, cells were plated at 30–40% confluence in 35 mm dishes (Corning, Corning, NY) and, 24 hr later, the culture medium was replaced with fresh DMEM containing EGCG (5 mM stock solution in 0.1% (v/v) DMSO) for a further 24-hr period.
Clonogenic survival assays
EGCG-containing medium was removed and replaced with Ca++/Mg++-free Dulbecco PBS. Irradiation was carried out on a cooled plate at 4°C under UVA tubes (Phillips) at a dose rate of 100 W/m.2 Mylar film (Lee filters) was used to cut out wavelengths in the UVB region of the spectrum. Only 0.052% of total irradiance was below 340 nm. Cells were then harvested and plated at appropriate cloning densities. Seven days after plating, colonies were rinsed in PBS, fixed in 70% (v/v) ethanol and stained with Giemsa (5%,v/v).
Measurements of UVA-induced radical species
Cells were harvested, washed twice in PBS containing 2.1 mM EDTA, 5.2 mM dextrose and 0.1% (v/v) gelatin and resuspended at 2 × 106/ml in Hank's balanced salt solution (HBSS). Each sample was loaded with either DCFH-DA or DHR (2 mM stock solutions in DMSO) to 20 μM for 20 min at 37°C. Because photo oxidation of the fluorescent probes becomes an increasing problem closer to the UVB region of the spectrum, for these experiments, irradiations were carried out using a 360 ± 5 nm interference filter fitted to a 200 W Hg(Xe) arc lamp (Oriel, Stratford, CT). At this wavelength, we have demonstrated that oxidation of the dyes reflects their reaction with free radical species.24 Immediately after irradiation, samples were run on a FACSCALIBUR flow cytometer (Becton Dickinson, Franklin Lakes, NJ) and geometric means for histograms calculated using CellQuest software.
Single cell gel electrophoresis (comet) assay
Cells were harvested, washed twice in PBS and mixed with 1% low melting point (LMP) agarose in a 1:1 ratio to a density of 2.5 × 104/ml. A 300 μl volume of cells in agarose was applied to slides and coverslips were laid on top to form a thin layer. After 10 min on ice to allow the agarose to solidify, the coverslips were removed and the samples irradiated using UVA tubes and mylar film as for the cell survival experiments. After 1 hr in lysis buffer (60 mM NaOH, 1 M NaCl, 0.5%(w/v) N-lauryl sarcosine, >pH 12.5) and a further 1 hr in DNA unwinding solution (40 mM NaOH, 2 mM Na2 EDTA, pH 13), samples were electrophoresed at 20 V for 25 min. After staining with SYBR Green (0.01% in TBE buffer) for 25 min, comets were analyzed using fluorescence microscopy linked to Optimas software.
HPRT mutation assay
Cells were cultured in HAT medium (100 μM hypoxanthine, 0.4 μM aminopterin, 16 μM thymidine) for at least 14 days to remove existing HPRT mutants and plated at 30% confluence in 10 cm dishes. After 24 hr, the HAT-containing medium was replaced with normal medium containing 50 μM EGCG or carrier (0.1% DMSO). After a 24-hr treatment, cells were UVA irradiated (500 kJ/m2) and cultured for a further 12–14 days to allow expression of HPRT mutants before plating in selective medium containing 30 μM 6-thioguanine (3 × 106 cells/T180 cm2 flask). Cells were also plated in regular medium at cloning density to evaluate plating efficiency for each treatment. Colonies were fixed in 70% ethanol and stained in 5% Giemsa 14 days later.
Annexin V/propidium iodide staining
Cells undergoing early apoptosis or late apoptosis or necrosis were analyzed by flow cytometry after processing using an Apoptosis Detection Kit (Oncogene Research Products, The Netherlands). Briefly, cells were harvested as normal except that 0.5× trypsin was used for cell detachment and cell media, potentially containing dead cells, was retained on ice. Cells and media were spun down (220g, 4°C) and resuspended in 1× cold binding buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 2.5 mM CaCl2, 1 mM MgCl2, 4% BSA) at a concentration of 1 × 106/ml. Cells (5 × 105) were incubated with Annexin V-FITC (1 μl, 15 min, room temp), spun down (1,000g, 5 min) and resuspended in 0.5 ml binding buffer. Propidium iodide (2.5 μl; 500 μg/ml) was added before flow cytometric analysis to identify late apoptotic or necrotic cells.
All data were analyzed with Sigma Stat software (SPSS, Chicago, IL) using Student's t-test for pairwise comparisons.
EGCG inhibits UVA-generated ROS detected by DHR
Consistent with previous reports,44 we confirmed that a solution of EGCG (0.6 mM) in 0.1% DMSO, pH 7.5 produced an absorbance spectrum with a λmax below 300 nm (292 nm) and falling to 1% of this value by 340 nm. The use of UVA tubes and an arc lamp that emitted in the 340–400 nm region and at 365 ± 5 nm, respectively, therefore ensured that the effects of EGCG on UVA irradiated cells were not attributable to UV absorbance by the compound.
The basal and UVA-induced levels of fluorescence for DHR and DCFH-DA loaded HaCaT cells after EGCG treatment are shown in Figures 1 and 2 respectively. In the case of both DHR and DCFH, EGCG treatment has no effect on the levels of fluorescence in unirradiated cells up to 20 μM; however after 50 μM EGCG treatment there are increases of 21% and 29% in the levels of fluorescent rhodamine 123 and DCF respectively indicating that at sufficiently high concentrations, EGCG induces cellular ROS. In the absence of fluorescent probe, 50 μM EGCG-treated cells did not produce any autofluorescent profile on the flow cytometer.
As reported previously,24 UVA exposure (100 kJ/m2) produced an increase in fluorescence in both DHR and DCFH-loaded cells (increases of, on average, 161 and 152% over basal levels, respectively). A prior 24-hr exposure to an EGCG concentration as low as 10 μM, however, resulted in an increase of only 126% (p < 0.05) over the corresponding unirradiated level. Moreover, a 50 μM EGCG treatment, despite increasing the unirradiated level of ROS, reduced the UVA-induced fluorescence increase to, on average, 104% over the basal value. In contrast, no such reduction in UVA signal was observed for DCFH-loaded cells. In fact, EGCG treatment resulted in an increase in UVA-induced DCF oxidation of up to 32% (p < 0.05, 50 μM EGCG).
EGCG protects against the production of DNA single strand breaks and alkali-labile sites by UVA
In agreement with other studies,26 UVA (50 kJ/m2) induced immediate DNA damage in HaCaT cells as measured by the “comet” assay (Fig. 3). Pretreatment with 10 μM EGCG reduced the level of DNA damage in UVA-irradiated cells to 62% of control levels, on average. Exposing the cells to a 5-fold higher concentration of EGCG i.e., 50 μM did not enhance the level of protection significantly (58% of control). In unirradiated cells, treatment with EGCG produced a small increase in the comet moment compared to untreated cells particularly at 10 μM (comet moment of 6.69 ± 1.1 at 10 μM EGCG vs. 4.01 ± 0.67 in control, not statistically significant). The fact that 10 μM EGCG produced a slightly higher proportion of cells in S phase compared to controls (46% vs. 41 %) after a 24-hr treatment may have contributed to this increase in damage because there are reports of S-phase cells exhibiting higher basal levels of damage than those in G1 in the comet assay.27 Treatment with higher concentrations of EGCG up to 100 μM did not result in any changes in the proportion of HaCaT cells in G1, S or G2/M (data not shown).
In the absence of UVA, EGCG is cytotoxic to HaCaT cells
Twenty-four hour EGCG treatment alone decreased cell survival in HaCaT cells (Fig. 4) with a 50 μM exposure reducing it to 50%. In support of a cytotoxic action of EGCG, annexin V/propidium iodide staining of HaCaT cells 4 hr after the end of EGCG treatment indicated an increase in the non-viable annexin+/PI+ fraction of cells from 9% in controls to 15% in cells treated with 50 μM EGCG. At 4 hr post EGCG treatment, there was, however, no increase over the untreated control in the fraction of cells staining for annexin V alone that is indicative of apoptosis.
EGCG modulates cell survival after UVA irradiation
UVA irradiation is cytotoxic and Figure 5a shows a reduction in HaCaT cell survival with increasing UVA dose in the absence of EGCG pretreatment. The individual cytotoxic effects of UVA and EGCG, however, are not additive when treatments are combined. As Figure 5a indicates, there is an increase in survival from the initial cytotoxic effects of EGCG in the absence of irradiation to overall protection against UVA. For example, 50 μM EGCG pretreatment results in 74% survival after 200 kJ/m2 in contrast to 50.8% in untreated cells. A similar pattern is observed for 20 μM EGCG (data not shown) with survival at 96% (from an unirradiated level of 66%) after 50 kJ/m2 UVA, as compared to 89.7% in the absence of EGCG. Expressing the data using the unirradiated survival values with the appropriate level of EGCG as controls (100%) demonstrates the polyphenol protection against UVA cell-killing more clearly (Fig. 5b). The survival curves illustrate how increasing concentrations of EGCG effectively shifts the survival curve “shoulder” out to higher doses.
EGCG protects against UVA-induced mutagenesis
In agreement with previous work,28, 29 a UVA dose of less than 500 kJ/m2 was insufficient to generate mutant levels in mammalian cells above spontaneous levels. Exposing HaCaT cells to a UVA dose of 500 kJ/m2 produced, on average, a 3.1-fold increase in mutation frequency (p < 0.05) (Fig. 6). Twenty-four hour treatment with 50 μM EGCG alone actually increased the frequency of hprt mutants (p = 0.048) in the absence of UVA but in contrast clearly suppressed UVA-induced mutagenesis (p < 0.05) to spontaneous levels.
Our study demonstrates that the green tea polyphenol, EGCG, is able to inhibit UVA-induced DNA damage and mutagenesis in human cells and provides evidence, using the ROS-sensitive probe DHR, that it exerts an antioxidant activity by reducing the level of certain UVA-induced ROS. EGCG is also able to modulate UVA cytotoxicity resulting in overall protection at certain doses. Interestingly, inhibition of UVA-induced ROS (DHR oxidation) and DNA damage were achieved at the lowest concentration of EGCG used (10 μM) that is of the same order as concentrations attained in human plasma after diet supplementation with green tea extracts (4.4 μM).30, 31 Although the UVA radical species that determine the endpoints of survival, DNA damage and mutagenesis and against which EGCG acts, are not known, they are clearly of a different makeup to those detected by the probe DCFH because it shows a pro-oxidant effect of EGCG after UVA.
The only previous study examining the effects of EGCG on UVA radiation reported its reduction of UVA-induced AP-1 and NF-KB transcription factor activities,32 processes known to be affected by redox status33, 34 thus providing additional evidence for EGCG antioxidant action against UVA. Wei et al.15 also reported the protection afforded by a green tea fraction containing a number of polyphenols against UVA-induced oxidative damage but only in a DNA solution.
There have been a number of studies citing the inhibitory effects of EGCG against UV radiation of shorter wavelengths.15, 35, 36, 37 In contrast to UVA, interpreting its protective role against UVC or UVB needs to take into account its strong absorption in the UVC and well into the UVB region (280–320 nm). Nevertheless, by application of EGCG post-irradiation, a few workers have shown that its action against UVB is not merely due to a “sunscreen” effect.35, 38 The UV sources used here emit at wavelengths beyond the absorbance spectrum of EGCG so the observations can be solely ascribed to the biological properties of the compound.
Although the data presented indicate that EGCG acts against UVA-induced radical species detected by DHR, the same pattern of inhibition is not seen with DCFH. A difference in the response of the two probes to changes in cellular glutathione status has been reported previously24 and further suggests that DHR and DCFH possess differing ranges of reactivity. There is an increasing trend to use DCFH or DHR oxidation as evidence of H2O2 production7, 37 and although the probes do react with peroxide, they are sensitive to a variety of other free radical species.39 The fact that there is an increase in DCFH oxidation in UVA-irradiated, EGCG-treated cells suggests that DCFH may react with reactive species generated as a result of the oxidation of EGCG through scavenging UVA ROS or with an EGCG radical itself. Although the scavenging mechanisms for EGCG are still poorly understood, there is recent evidence pointing to the production of superoxide (O2−) by EGCG in the process of scavenging peroxyl radicals.40 It is of interest that DCFH is readily oxidized by superoxide formation whereas DHR shows little reactivity to this species.39
EGCG had an antiproliferative effect on unirradiated HaCaT cells as expressed by a fall in cell survival upon subsequent plating. EGCG has been reported to exert a growth inhibitory effect on a number of tumour cell lines7, 8, 41 and on immortalized but non-tumorigenic bronchial epithelial cells.9 At least part of the inhibition has been ascribed to apoptosis in some of these cell lines.7, 9 It is shown that EGCG is cytotoxic to HaCaT cells as shown by an increase in the fraction of non-viable Annexin V+/Propodium iodide + cells, but there was no increase in the proportion of Annexin only + (i.e., apoptotic) cells within the time frame of our experiments. Ahmad et al.42 reported that HaCaT cells do undergo apoptosis, as assayed by DNA laddering, after a 48-hr EGCG treatment albeit at a concentration of 175 μM (80 μg/ml). The degree of DNA fragmentation, however, appears to be far less than in a human epidermoid (A431) or prostate carcinoma (DU145) cell line. In agreement with other studies, however, it is shown here that EGCG has a pro-oxidant effect because a 50 μM treatment induces ROS as detected by both the DHR and DCFH probes (Figs. 2,3). Furthermore, the comet and mutation data presented point to EGCG causing an increase in DNA damage. It has been postulated that this ROS generation underlies the cytotoxic action of EGCG.7, 9, 43
Prior treatment of HaCaT cells with EGCG clearly modifies their survival after UVA irradiation. At lower UVA doses, not only does EGCG prevent further cell killing by UVA but there is a clear increase in the level of survival. Interpreting the data in terms of EGCG antioxidant action against UVA-induced ROS, the initial increase in survival is the result of the inactivation of EGCG (and thereby its own growth inhibitory effect) through its scavenging of UVA free radicals. A similar effect has been observed by Cutter et al.44 where oxidative degradation of EGCG via the Fenton reaction by incubation with hydrogen peroxide and FeSO4 abolished the polyphenol inhibitory effect on the growth of HeLa cells. At sufficiently high UVA doses, survival starts to fall as, presumably, the oxidative stress overwhelms the scavenging capacity of EGCG. In practical terms, a 50 μM EGCG treatment offers overall protection to keratinocytes against a UVA dose (500 kJ/m2) equivalent to 2.5 hr in the summer midday sun at latitude 48′ north.45 The influence, on survival, of factors that modulate the cell's capacity to combat UVA-induced oxidative stress is well known.25 The relative contributions of unrepaired DNA damage and membrane perturbations to UVA cytotoxicity remain unclear. EGCG lowers the level of initial DNA ssb after UVA, but the effect on the repair capacity of EGCG-treated cells is not known. Alterations in ssb repair after UVA have been shown to affect cell survival.46 EGCG may, through its antioxidant action, affect UVA cell signaling events that impinge on cell survival. Indeed, a related green tea polyphenol epigallocatechin (EGC) is reported to prevent UVA induction of cyclooxygenase 2 in keratinocytes.47 Inhibition of cyclooxygenase is known to reduce UVA stimulated release of arachidonic acid from the cell membrane that seems to be an important cellular signal for halting cell proliferation and inducing cell death.48, 49
The decrease observed in UVA-induced DNA damage after EGCG treatment is mirrored by data pointing to its protective action against UVA-induced mutagenesis. EGCG has been reported to have an antimutagenic effect against UVAB light (295–400 nm) in E. coli although its activity may be related to effects on DNA repair rather than ROS because the reduction is abolished in excision-repair deficient mutants.50 EGCG reduction of 4-nitroquinoline 1-oxide (4-NQO)-induced mutagenesis in Chinese hamster cells51 may be based on its antioxidant activity because there is evidence that 4-NQO produces oxidative DNA damage aside from the well established “bulky” quinoline-purine adducts.52, 53 In contrast, there is evidence that spontaneous mutagenesis is not suppressed by green tea polyphenols in a bacterial system.54 Our data indicate a marginally significant increase in hprt spontaneous mutations after EGCG treatment whereas the combination of UVA and EGCG is not significantly different from EGCG treatment alone.
In conclusion, we have demonstrated that EGCG affords protection against the oxidative stress of UVA radiation in terms of biological endpoints relevant to carcinogenesis.
We thank Dr. N.E. Fusenig for the kind gift of HaCaT cells.