Relative Photomutagenicity of Furocoumarins and ... - ACS Publications

Sep 2, 2009 - Furocoumarins are phototoxic and photomutagenic natural plant constituents found in many medicinal plants and food items. Because plants...
0 downloads 0 Views 2MB Size
Download PDF
Chem. Res. Toxicol. 2009, 22, 1639–1647

1639

Relative Photomutagenicity of Furocoumarins and Limettin in the Hypoxanthine Phosphoribosyl Transferase Assay in V79 Cells Nicole Raquet and Dieter Schrenk* Food Chemistry and Toxicology, UniVersity of Kaiserslautern, Erwin-Schroedinger-Strasse 52, D-67663 Kaiserslautern, Germany ReceiVed July 7, 2009

Furocoumarins are phototoxic and photomutagenic natural plant constituents found in many medicinal plants and food items. Because plants contain mixtures of several furocoumarins, there is a need for a comparative risk assessment of a large number of furocoumarins. Little is known about the photomutagenicity of the structurally related family of coumarins, which are also abundant in many plant species. In this study, we analyzed the photomutagenic potency of the linear furocoumarins 5-methoxypsoralen (5-MOP) and 8-methoxypsoralen (8-MOP), the angular furocoumarin angelicin, and the coumarin limettin. Above certain concentrations, all test compounds were more or less phototoxic in the presence of UVA doses between 50 and 200 mJ/cm2, 5-MOP being the most phototoxic compound. At nonphototoxic concentrations, linear correlations were found between concentration and mutagenicity at a UVA dose of 125 mJ/cm2 for all test compounds including limettin. For 5-MOP, strictly linear correlations were also found for the relationships of mutagenicity vs concentration at various UVA doses or vs UVA dose at given concentrations, respectively. These data indicate that the photomutagenicity of 5-MOP is proportional to the UVA dose × concentration product for noncytotoxic combinations of both factors. They also suggest that the slope of the concentration-photomutagenicity correlation at a given UVA dose may provide a basis for comparison between individual compounds. Applying this concept, in vitro photomutagenicity equivalency factors at 125 mJ/cm2 were as follows: 1.0 (5-MOP, reference compound), 0.25 (8-MOP), and 0.02 (angelicin and limettin, respectively). These findings provide a new concept for the description of the relative photomutagenic potency of coumarins and furocoumarins and indicate that, in V79 cells, 8-MOP is less photomutagenic and limettin and angelicin are much less photomutagenic than 5-MOP. Introduction Furocoumarins are natural plant constituents, bearing a coumarin structure attached to a furan ring, while limettin is a naturally occurring coumarin dervative abundant in citrus fruits (Figure 1). Furocoumarins are usually divided into linear (“psoralen type”) and angular (“angelicin type”) compounds. They occur in various plant families such as Apicaceae and Umbelliferae, most notably in species such as Ammi, Pimpinella, Angelica, and Heracleum, in Fabaceae, and in citrus plants belonging to the Rutaceae (1). Most probably, these plants produce furocoumarins as protectants (“phytoalexins”) against microorganisms and/or other external factors. In food, furocoumarins occur, for example, in celery (Apium graVeolens L.), parsnip (Pastinaca satiVa), parsley (Petrosilenum crispum), carrot (Daucus carota L.), orange (Citrus sinensis L.), lemon (Citrus limon), lime (Citrus aurantifolia) (2), and grapefruit (Citrus paradisi) (3). In particular, citrus oils can contain very high levels of furocoumarins, while distillation of the oils may remove furocoumarins due to their low volatility. Furthermore, furocoumarin levels can vary tremendously depending on the conditions of crop cultivation and storage. Likewise, microbial infections of stored parsnips can result in an enormous increase in furocoumarin levels to up to 570 (4) or 2500 mg/kg (5), respectively. The uptake of furocoumarins via food is probably subject to pronounced interindividual and daily variation. High exposure * To whom correspondence should be addressed. Fax: +49-06312054398. E-mail: [email protected].

Figure 1. Chemical structures of angelicin, limettin, 5-MOP, and 8-MOP.

may occur from celery or parsnips infected with microorganisms. Likewise, the consumption of 200 g of infected parsnips, assuming a total furocoumarin content of 500 mg/kg, could lead to an uptake of up to 100 mg total furocoumarins per adult (6). Estimates for the average daily intake of furocoumarins via food in adults were published being in the range of 1.3 mg for the United States (1), 1.2 mg for the United Kingdom (6), and 1.45 mg for Germany (7). In humans and laboratory animals, systemic and dermal application of furocoumarins can lead to phototoxic effects in combination with UVA light, that is, furocoumarins can strongly

10.1021/tx9002287 CCC: $40.75  2009 American Chemical Society Published on Web 09/02/2009

1640

Chem. Res. Toxicol., Vol. 22, No. 9, 2009

enhance the basic skin toxicity of UVA light. Furthermore, photoactivation of furocoumarins can result in DNA damage, mutations, and skin cancer (8, 9). In particular, the occurrence of mutations was reported in various test systems including V79 cells treated with imperatorin or 8-MOP (10), Chinese hamster ovary cells treated with 8-methoxypsoralen (8-MOP), 5-methoxypsoralen (5-MOP), or 5-methylangelicin (11), or in human lymphoblasts treated with 4,5′,8-trimethylpsoralen (12). It has been suggested that furocoumarins, due to their planar structure, may intercalate with DNA in the so-called dark reaction. Excitation of furocoumarins is expected as the initial chemical event to lead to the formation of furan- and pyrone-side monoadducts to pyrimidine bases, thymine particularly at 5′TpA-3′ sequences being the most reactive. While pyroneside monoadducts are no more photoreactive, the furan-side monoadducts can be excited by a second UVA photon, giving rise for linear furocoumarins to interstrand cross-links, whereas angular furocoumarins are supposed to form monoadducts only (12, 13). In this study, we applied the hypoxanthine phosphoribosyl transferase (HPRT) mutagenicity assay in V79 cells to analyze the quantitative relationship between furocoumarin concentration, UVA dose, and phototoxicity and photomutagencity using the prototype furocoumarin 5-MOP. Furthermore, we investigated the phototoxic and photomutagenic properties of the abundant furocoumarins 8-MOP and angelicin and the coumarin limettin.

Experimental Procedures Materials. V79 cells were a generous gift from Prof. G. Eisenbrand, Food Chemistry and Toxicology, University of Kaiserslautern (Germany). Angelicin (m.w. 186.16), limettin (citropten, 5,7-dimethoxycoumarin; m.w. 206.19), 8-MOP (m.w. 216.19), DMSO, 6-thioguanine, saponin, and resazurin were from SigmaAldrich (Steinheim, Germany). 5-MOP (m.w. 216.19) and N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) were from TCI Europe (Zwijndrecht, Belgium), and Dulbecco’s modified Eagle’s medium (DMEM) low glucose, fetal calf serum (FCS), penicillin/streptomycin, and accutase were from PAA (Co¨lbe, Germany). HPRT Assay. All cell culture experiments were carried out under the exclusion of visible and/or UV light. Special precautions were taken for any work with genotoxic/mutagenic reference compounds. All cell culture experiments were carried out under sterile conditions. First, 770000 V79 cells were seeded on each 94 mm plastic Petri dish in a medium containing 89% DMEM low glucose, 10% FCS, and 1% penicillin/streptomycin and kept for 24 h under standard conditions in an incubator (37 °C, under air with 5% CO2). Then, the medium was removed, and the test compounds were dissolved in 50 µL of DMSO and added to the plates in 10 mL of FCS-free medium containing 500 mL of DMEM low glucose and 5 mL of penicillin/streptomycin, resulting in a DMSO concentration of 0.5%. After preincubation for 30 min, the medium was removed, and the plates were rinsed with PBS and transferred to a UVirradiation chamber BS-03 (Dr. Gro¨bel, Ettlingen, Germany) equipped with UVA lamps (λmax ) ca. 365 nm) and a UV-MAT dosimeter. After removal of the lid, irradiation was carried out at various UVA doses. Then, 10 mL of DMEM/FCS medium was added, and the plates were kept for another 24 h in an incubator. Negative controls either were not irradiated or were supplemented with DMSO without test substance. Positive controls were treated with 10-20 µM MNNG (moistened with ca. 50% water). The stock solution (1000×) containing 4.41 mg of MNNG in 1 mL of water was freshly prepared before each experiment. After 24 h, and again after 72 h, the medium was removed, the plates were rinsed with phosphate-buffered saline (PBS), and 1 mL of trypsin solution was added to each plate. The cells were suspended in DMEM/FCS medium and counted, and 106 cells were

Raquet and Schrenk transferred with 15 mL of DMEM/FCS medium to 250 mL incubation flasks. After additional 48 h, the medium was removed, the cells were rinsed with PBS, and 1 mL of accutase solution was added. The cells were suspended in DMEM/FCS medium and counted. From the same incubation, 106 cells in 15 mL of TG medium (500 mL of DMEM low glucose, 25 mL of FCS, 5 mL of penicillin/streptomycin, 5 mL of 100 mM sodium pyruvate solution, and 0.5 mL of 6-thioguanine stock solution) each were transferred to three 250 mL flasks. The 6-thioguanine stock solution (53.83 mM) was prepared from 9 mg of 2-amino-6-mercaptopurine in 1 mL of DMSO. As a control for vitality, 240 cells each were suspended in 10 mL of DMEM/FCS medium and transferred to two 94 mm Petri dishes. After 9 days, Petri dishes and incubation flasks were rinsed with 0.9% saline, and 5 mL of ethanol (-20 °C) was added to each vessel and kept for 15 min at -20 °C. After the ethanol was removed and 3 mL of methylene blue solution (0.5% methylene blue in ethanol) was added, the vessels were kept for 30 min at -20 °C. Then, the methylene blue solution was removed, the cell layers were carefully rinsed with tap water, and the plates and flasks were air-dried. The colonies per vessel were counted, and the mean (number) was calculated for each treatment. The mutation frequency (MF) was calculated as MF ) meanflask × 240/meanplate. Cytotoxicity Testing. The cytotoxicity was tested using the Alamar Blue assay according to the manufacturer’s (BioSource International, Camarillo, CA) instructions. Briefly, 104 V79 cells were seeded under sterile conditions, analogue to the hPRT assay, on each of a 24-well plastic dish in a medium containing 89% DMEM low glucose, 10% FCS, and 1% penicillin/streptomycin and kept for 24 h under standard conditions in an incubator (37 °C, under air with 5% CO2). Then, the medium was removed, and the test compounds were dissolved in DMSO and added to the wells in duplicates in 1 mL of FCS-free medium containing 500 mL of DMEM low glucose and 5 mL of penicillin/ streptomycin resulting in a DMSO concentration of 0.5%. After preincubation for 30 min, the medium was removed, and the plates were rinsed with PBS and transferred to a BS-03 UV irradiation chamber (Dr. Gro¨bel). After the lid was removed, irradiation was carried out at various UVA doses. Then, 1 mL of DMEM/FCS medium per well was added, and the plates were kept for another 3 days in an incubator. Negative controls either were not irradiated or were supplemented with DMSO without test substance. Positive controls were treated with saponin (0.1% final concentration, stock solution dissolved in water). After 24 h, the medium was removed, and the plates were rinsed with PBS and incubated for 1 h with 1 mL per well resazurin-containing DMEM. The fluorescence was measured in a fluorescence spectrophotometer (Labsystems, Dreieich, Germany) using excitation at 544 nm and detection of emission at 590 nm. The reducing capacity, a measure for the number of vital cells, was expressed as percent of vehicle (DMSO)-treated controls. Statistical Analysis. From several independent measurements means and standard deviations were calculated. Testing for significant differences between means was carried out using Dunnett’s post test at a probability of error of 5 and 1%. Linear fitting and regression analysis were carried out with the Origin G8 software (Origin Lab Corp., United States), defining a 95% confidence interval for the linear model.

Results In this study, we wanted to analyze the relative phototoxicity and photomutagenicity of three furocoumarins, the two linear furocoumarins 5-MOP and 8-MOP, and an angular furocoumarin, angelicin, (Figure 1), in the HPRT mutagenicity assay in V79 cells. Furthermore, the same experimental setting was used to investigate a possible phototoxicity/photomutagenicity of limettin, a coumarin found in many food items, particularly in citrus fruits. As a first step, we analyzed if the test compounds, either under exclusion or in the presence of UVA light, were cytotoxic in

Photomutagenicity of Furocoumarins and Limettin

Chem. Res. Toxicol., Vol. 22, No. 9, 2009 1641

Figure 2. Cytotoxicity of 5-MOP in V79 cells treated with various concentrations (µg/mL) at UVA doses (mJ/cm2) as indicated. After the addition of 5-MOP, cells were incubated for 30 min, rinsed, irradiated with UVA, rinsed again, and incubated for another 72 h. Then, resazurin reduction was measured as described in the Experimental Procedures. Bars represent means ( SDs from n ) 4 independent experiments. *Significantly different (p e 0.05) or **very different (p e 0.01) from the untreated control. Saponin (0.1%) served as a positive control.

Figure 3. Cytotoxicity of angelicin in V79 cells treated with various concentrations (µg/mL) at UVA doses (mJ/cm2) as indicated. After the addition of angelicin, cells were incubated for 30 min, rinsed, irradiated with UVA, rinsed again, and incubated for another 72 h. Then, resazurin reduction was measured as described in the Experimental Procedures. Bars represent means ( SDs from n ) 4 independent experiments. *Significantly different (p e 0.05) or **very different (p e 0.01) from the untreated control. Saponin (0.1%) served as a positive control.

the V79 cells used for the HPRT assay. For the determination of phototoxicity, various concentrations of the test compounds were combined with different UVA doses, and cytotoxicity was tested using the resazurin reduction assay (Figure 2-4). It was found that, in the absence of UVA, none of the test compounds was cytotoxic up to concentrations of 25 µg/mL for 8-MOP, 50 µg/mL for 5-MOP, and 100 µg/mL for angelicin and limettin. The highest concentrations used were different, because of the different water solubilities of the test compounds. With 5-MOP, UV irradiation caused significant cytotoxicity, for example, at the lowest UVA dose used (50 mJ/cm2) in combination with 20 µg/mL, while at 200 mJ/cm2, 1.0 µg/mL was sufficient for a significant cytotoxic effect (Figure 2). In contrast, angelicin was almost noncytotoxic (Figure 3), showing some slight toxic effect at the combination of 100 mJ/cm2 and 100 µg/mL only. For 8-MOP and limettin (Figure 4), only relatively high concentrations g25 or 50 µg/mL, respectively, were able to elicit cytotoxicity when combined with a UVA dose of 125 mJ/cm2. The latter UVA dose was further used for the photomutagenicity tests of these compounds. 5-MOP was selected as a reference compound because of its abundance in food and medicinal plants and its strong photomutagenicity, reported in previous studies. The photomutagenicity testing was carried out in detail for 5-MOP, to better understand the quantitative roles of both the concentration of the test compound and the UVA dose for the mutagenic outcome.

Figure 4. Cytotoxicity of limettin or 8-MOP in V79 cells treated with various concentrations (µg/mL). After the addition of limettin or 8-MOP, cells were incubated for 30 min, rinsed, either kept in the dark or irradiated with 125 mJ/cm2 UVA, rinsed again, and incubated for another 72 h. Then, resazurin reduction was measured as described in the Experimental Procedures. Bars represent means ( SDs from n ) 4 independent experiments. *Significantly different (p e 0.05) or **very different (p e 0.01) from the untreated control. Saponin (0.1%) served as a positive control.

For the concentration-mutagenicity correlations, clear linear regressions were found at fixed UVA doses of 50, 75, and 100 mJ/cm2 for a concentration range between 0.5 and 3.0 µg/mL 5-MOP (Figure 5A-C). At higher UVA doses, that is, 125 and 200 mJ/cm2, linear correlations were also found (Figure 5D,E) for the noncytotoxic concentration ranges, which were selected based on the cytotoxicity testing. The slope of the concentration-response correlation increased with increasing UVA dose. At a UVA dose of 125 mJ/cm2, linear correlations between photomutagenicity and concentration were also obtained for angelicin, 8-MOP, and limettin (Figure 6-8). In the next step, the impact of different UVA doses was analyzed at fixed concentrations of 5-MOP, namely, at 0.5, 1.0, and 3.0 µg/mL (Figure 9A-C). It was found that the mutagenicity increased in a linear fashion with increasing UVA dose at each of the three given concentrations. The slope of the UVA dose-response correlation increased with increasing 5-MOP concentration. These findings led us to suggest that the mutagenicity of 5-MOP is directly proportional to the product of UVA dose and concentration, at least in the noncytotoxic range. Analysis ofthecorrelationbetweentheslopesoftheconcentration-response relationships for different UVA doses and the UVA doses revealed a linear correlation (Figure 10). The actual values

1642

Chem. Res. Toxicol., Vol. 22, No. 9, 2009

Raquet and Schrenk

Figure 5. Photomutagenicity of 5-MOP in V79 cells using the HPRT assay. After the addition of 5-MOP as indicated, cells were incubated for 30 min, rinsed, irradiated with 50 (A), 75 (B), 100 (C), 125 (D), or 200 mJ/cm2 (E) UVA, and further incubated. Mutagenicity was analyzed as described in the Experimental Procedures. Symbols represent means ( SDs from n ) 3 independent experiments. Lines show linear regression fitting with 95% confidence intervals. Regression coefficients are provided in Table 1.

were all within the 95th percentile of the linear regression. Furthermore, a test for curvilinearity (data not shown) did not reveal better fitting. Our findings suggest that in fact the concept of proportionality between the UVA dose × concentration product and the mutagenicity in the system V79/ HPRT is correct for a wide range of 5-MOP concentrations and UVA doses.

We then decided to compare the slopes for all test compounds at a given UVA dose of 125 mJ/cm2 to get some information about their relative photomutagenicities. For this comparison, we applied the concept proven for 5-MOP, which suggests that relative photomutagenicity of this class of compounds can be estimated from the slopes of the mutagenicity-concentration curves at the same UVA dose.

Photomutagenicity of Furocoumarins and Limettin

Chem. Res. Toxicol., Vol. 22, No. 9, 2009 1643

Figure 6. Photomutagenicity of angelicin in V79 cells using the HPRT assay. After the addition of angelicin as indicated, cells were incubated for 30 min, rinsed, irradiated with 125 mJ/cm2 UVA, and further incubated. Mutagenicity was analyzed as described in the Experimental Procedures. Symbols represent means ( SDs from n ) 3 independent experiments. The line shows linear regression fitting with a 95% confidence interval. The regression coefficient is provided in Table 1.

Figure 8. Photomutagenicity of 8-MOP in V79 cells using the HPRT assay. After the addition of 8-MOP as indicated, cells were incubated for 30 min, rinsed, irradiated with 125 mJ/cm2 UVA, and further incubated. Mutagenicity was analyzed as described in the Experimental Procedures. Symbols represent means ( SDs from n ) 3 independent experiments. The line shows linear regression fitting with a 95% confidence interval. The regression coefficient is provided in Table 1.

Figure 7. Photomutagenicity of limettin in V79 cells using the HPRT assay. After the addition of limettin as indicated, cells were incubated for 30 min, rinsed, irradiated with 125 mJ/cm2 UVA, and further incubated. Mutagenicity was analyzed as described in the Experimental Procedures. Symbols represent means ( SDs from n ) 3 independent experiments. The line shows linear regression fitting with a 95% confidence interval. The regression coefficient is provided in Table 1.

Their phototoxicity is known from incidences of dermal lesions after contact with the plant or food in combination with exposure to sunlight or artificial UV irradiation (18, 19). It appears more difficult to estimate the risk of furocoumarin exposure via the oral route. Here, a number of questions are still open with respect to the toxicological risk assessment for this pattern of effects. First, it is unknown how the photomutagenic potency of plants and herbs should be assessed since furocoumarins are never found in plants as single compounds but always as complex mixtures of many individual furocoumarins (4, 20). The (relative) phototoxic and photomutagenic potencies of most of these individual compounds have never been tested and are mostly unknown. Second, it is unknown if furocoumarins are the only photomutagens in food or if there are other relevant photomutagens, for example, among the larger family of coumarins and related compounds. Third, it is unknown how the photomutagenic potency of certain furocoumarins can be used as a basis for cancer risk assessment. It is well-known that the combination of psoralen or 8-MOP with UVA light, used in PUVA therapy of psoriasis, can lead to an increased risk of various types of skin cancer such as keratocarcinoma, sarcoma, and melanoma (21-23). Exposure toward furocoumarins via food has been assessed coming to the conclusion that a “threshold dose” of oral exposure may exist for the development of erythema under sunlight exposure (6, 24). This threshold is also discussed as a possible “practical threshold” for an additional cancer risk. However, other authors discuss the possibility that average dietary exposure to furocoumarins, under certain conditions of sunlight exposure and skin type, may already result in a higher cancer risk (25). 5-MOP and 8-MOP were reported to form noncovalent DNA complexes in the dark (26, 27), which, upon UVA radiation, bind covalently (28-30) and can form interstrand cross-links (31). In bacterial or yeast mutagenicity tests, a number of furocoumarins were shown to be weakly or not mutagenic in the absence of light but exhibited strong mutagenicity when combined with UVA light (13, 32-35). In mammalian cells, some studies report on a weak mutagenicity of furocoumarins in the absence of light (36-38). In combination with UVA,

From these assumptions, photomutagenicity equivalency factors (PMEFs) at a UVA dose of 125 mJ/cm2 (PMEF125s) were derived as 0.25 for 8-MOP and 0.02 for angelicin and limettin, respectively (Table 1). The PMEF of 5-MOP was set as 1.0.

Discussion For risk assessment of dietary furocoumarins, certain major pharmacotoxicological properties of this class of compounds have to be considered. These include the UV-independent interference of furocoumarins with drug metabolism (14-16) and their UV-dependent phototoxicity and photomutagenicity (8, 9). Furocomarins widely occur in many plant families and in individual plants used as food and/or herbal medicine (1, 17).

1644

Chem. Res. Toxicol., Vol. 22, No. 9, 2009

Raquet and Schrenk

Table 1. Slopes of Linear Regression for the Correlation between Furocoumarin (Coumarin) Concentration or UVA Dose and Mutagenicity in the HPRT Assay compound 5-MOP

angelicin limettin 8-MOP 5-MOP

concentration (µg/mL)

UVA dose (mJ/cm2)

correlation coefficient r2

variable variable variable variable variable variable variable variable 0.5 1.0 3.0

50 75 100 125 200 125 125 125 variable variable variable

0.98 0.82 0.94 0.97 0.99 0.87 0.98 0.98 0.69 0.82 0.93

slope of linear regression (colonies/µg mL)

5-MOP and 8-MOP bind to cellular DNA and cause interstrand cross-links (39), are strongly mutagenic, and result in increased sister chromatide exchange (11, 40-42). The HPRT assay in V79 cells has been shown to be highly responsive to furocoumarin/UVA combinations. At the HPRT locus, mainly base substitutions at A-T base pairs were observed. The human HPRT gene comprises more than 190 thymines on the coding strand, 165 of which are located in coding regions and at least 25 at splicing sites. Some of the thymines in the context 5′TpA, in particular those on the noncoding strand, seem to be particularly vulnerable against such base substitutions. It was suggested that furocoumarins may thus lead to a special signature of base substitution in the human HPRT locus (12). Chiou and Yang (43) reported an increase in mutagenicity with addition of 7 µM (1.51 µg/mL) 8-MOP combined with a single dose of 6 mJ/cm2 UVA in the HPRT assay in HFW diploid human fibroblasts. With higher 8-MOP concentrations, the mutagenicity increased continuously. The authors found a similar signature, that is, mainly T to A conversions in the 5′TA context on the noncoding strand, as Laquerbe et al. (12). Here, we describe that 5-MOP, assumed to be one of the most potent photomutagens among the furocoumarins, is mutagenic in the HPRT assay in V79 cells, in the presence of UVA light. The strictly linear relationships between the 5-MOP concentration in the assay and the photomutagenicity at a given UVA dose and between the UVA dose and the photomutagenicity at a given 5-MOP concentration are in agreement with the assumption of a simple proportionality. Similarly, Averbeck (13) found strictly linear relationships between the UVA dose and the induction of DNA photoadducts or the frequency of mutants in yeast for constant concentrations of 5-MOP and 8-MOP. Proportionality was affected in our experiments by cytotoxic effects of 5-MOP, which were detected at combinations of higher 5-MOP concentrations and higher UVA doses. Consequently, these were excluded from photomutagenicity analysis. The correlation between photomutagenicity, UVA dose, and 5-MOP concentration can thus be expressed using a proportionality factor f

[photomutagenicity]5-MOP ) f × [UVA] × [c]5-MOP leading to

f ) slope[UVA]const /[UVA] which was found to be relatively stable over a wide range of UVA doses; that is, it ranged between 0.41 and 0.49 without obvious influence of the UVA dose applied. Thus, any of the UVA doses tested could be used for the determination of f. It

24.3 32.6 40.9 52.5 95.5 1.2 1.1 12.8

proportionality factor f (colonies cm2 mL/mJ µg)

slope of linear regression (colonies/mJ cm2)

0.49 0.43 0.41 0.42 0.48

in vitro PMEF

1.00 0.02 0.02 0.25 0.18 0.31 0.93

thus can serve as a direct measure of the photomutagenic potency of 5-MOP. For the other compounds tested, a weaker photomutagenic potency was found in a number of screening experiments (not shown). Analogous to our results with 5-MOP, we assumed that any UVA dose within the range tested can also be used for the determination of an f value specific for each of the other test compounds. Therefore, we decided to set a UVA dose of 125 mJ/cm2 as a reference dose. Under these conditions, linear relationships between concentration and photomutagenicity were also found for angelicin, 8-MOP, and limettin. If the slope of the concentration-response relationship for 5-MOP at 125 mJ/ cm2 is set as 1.0, we could derive preliminary in vitro PMEFs of 0.25 for 8-MOP and 0.02 for angelicin and limettin, respectively (Table 1). A similar rank order of the photomutagenic/clastogenic potencies (5-MOP > 8-MOP) was reported by Averbeck (13) for yeast and by Abel (42) for human lymphocytes in vitro. Rodighero et al. (29) found a rank order of 8-MOP > 5-MOP > angelicin for the photoreactivity with calf thymus DNA. Furthermore, the UVA source, that is, the spectral properties of the irradiation, may explain contradicting results on the relative photomutagenicity of both compounds (44). Our approach, if also valuable for additional furocoumarins and coumarins, could allow the calculation of a combined photomutagenic potency of complex mixtures of these compounds. For this purpose, the issue of additivity of various photomutagenic compounds of the same structural class has to be investigated. In addition, further work is needed to test the hypothesis that the same PMEF values can in fact be derived over a wide range of UVA reference doses. Further work is also needed with respect to the more complicated question of additivity of photomutagenic effects of mixtures of compounds belonging to different groups of chemicals. The relationship between overall photomutagenicity and the chemical nature of the photoproducts warrants further investigation. In preliminary experiments using a modified Comet assay, we obtained indication for the formation of DNA interstrand cross-links for 5-MOP > 8-MOP but not for angelicin or limettin (data not shown). The formation of interstrand cross-links was previously reported by Wu et al. (45) for 8-MOP/UVA in HaCaT cells using a modified Comet assay. Other test methods for analysis of photoproducts include a quantitative polymerase chain reaction method and a thermal denaturation and renaturation assay (46), indicating the formation of biadducts with 5-MOP, 8-MOP, and psoralen, while angelicin was negative. Furthermore, a relevant HPLC-tandem mass spectrometric method is available for the measurement of mono- and biadducts of 8-MOP to cellular DNA (47, 48).

Photomutagenicity of Furocoumarins and Limettin

Chem. Res. Toxicol., Vol. 22, No. 9, 2009 1645

Figure 10. Correlation between slopes of concentration-MF relationships for 5-MOP at different UVA doses (Figure 5A-E), according to Table 1, and the respective UVA doses. The line shows linear regression fitting with a 95% confidence interval. The regression coefficient is 0.98.

Figure 9. Photomutagenicity of 5-MOP in V79 cells using the HPRT assay. After the addition of 5-MOP at concentrations of 0.5 (A), 1.0 (B), or 3.0 µg/mL (C), cells were incubated for 30 min, irradiated with various UVA doses as indicated, rinsed again, and further incubated. Mutagenicity was analyzed as described in the Experimental Procedures. Symbols represent means ( SDs from n ) 3 independent experiments. Lines show linear regression fitting with 95% confidence intervals. Regression coefficients are provided in Table 1.

The issue of cytotoxicity also warrants further investigation. In a complex organ such as the skin, the interplay between many cells types and factors is needed for a photomutagenic and -carcinogenic response. Toxic cell damage may be part of this response, being mediated by reactive (photoactivated) intermediates. These include the photoactivated furocoumarins as well

as secondary products such as singlet oxygen. However, 5-MOP as the most cytotoxic furocoumarin in our experiments has been described to cause a relatively small yield of singlet oxygen in model systems (49, 50). Inflammation and erythema are considered as a typical response of the skin toward UVA irradiation (51), for example, in combination with UV sensitizers such as furocoumarins. It has been suggested that erythema may result from or at least be tightly linked to DNA damage (52) and may thus be a good indicator for a photomutagenic threshold dose. Likewise, Schlatter et al. (24), the SKLM (7), and others (6) concluded that the ingestion of furocoumarin doses causing no erythema at a relevant UVA dose can be considered as safe with respect to carcinogenicity. In fact, Brickl et al. (53) and Schlatter et al. (24) found indications for threshold doses of 5-MOP or 8-MOP, respectively, not leading to erythema in humans and correlated these to certain blood levels, that is, in the range of 10-15 ng/ mL. It appears very difficult to draw any conclusions on a “pratical threshold” for photomutagenicity from our data. Because it was demonstrated for 8-MOP that serum concentrations in guinea pigs can provide a good estimate for the target concentrations in the skin (54), these levels may represent an adequate target dose measure for “practical threshold” considerations. The usefulness of a simple V79 cell culture model, however, to address the issue of threshold effects in a complex organ like the skin is very limited. Likewise, complex tissue reactions such as erythema cannot be mimicked. It appears unlikely that cytotoxicity measured in our experiments is related to DNA damage on the one hand or can serve as a surrogate for erythema on the other hand. In fact, a comparison between cytotoxicity and photomutagenicity data, for example, for 8-MOP or angelicin, currently suggests a lack of correlation between both. The model presented here was especially designed as a tool in risk assessment of complex furocoumarin mixtures present in phytomedicines and food. Although synthetic furocoumarins such as trioxalen (4,5,8′-trimethylpsoralen) may be more potent photomutagens than 5-MOP, the latter was selected as a reference compound because of its relevant occurrence in these materials. Psoralen, used in phototherapy, was found in preliminary experiments to be less photomutagenic than 5-MOP (data not shown). The HPRT UVA assay is able to analyze DNA

1646

Chem. Res. Toxicol., Vol. 22, No. 9, 2009

damage-related photomutagenicity in a quantitative manner. These data may not be used for a quantitative analysis of a critical oral exposure for in vivo considerations. Major reasons are the uptake and metabolism of furocoumarins, which is independent of their photomutagenic potency but may have a strong influence on actual blood (24, 51, 52) or skin levels after oral exposure. In contrast, there is a good chance to use these data as a possible indicator for critical skin (or blood) concentrations of individual furocoumarins. Considerable differences between individuals, for example, dependent on their skin type, with respect to adverse effects were found, however, and need further consideration in risk assessment. The approach chosen here may be helpful for an assessment of the relative photomutagenic potencies of individual coumarins and furocoumarins, possibly also of mixtures thereof in vitro. An extrapolation to a photomutagenic risk in vivo can only be drawn on the basis of actual target levels. Acknowledgment. We thank Prof. Rolf Diller, Department of Biophysics, University of Kaiserslautern, for very helpful discussions and Prof. Gerhard Eisenbrand, Department of Food Chemistry and Toxicology, University of Kaiserslautern, for providing V79 cells. This work was supported financially by a grant from Steigerwald Inc. (Darmstadt, Germany).

References (1) Wagstaff, D. J. (1991) Dietary exposure to furocoumarins. Regul. Toxicol. Pharmacol. 14, 261–272. (2) Stanley, W. L., and Vannier, S. H. (1967) Psoralens and substituted coumarins from expressed oil of lime. Phytochemistry 16, 1091–1092. (3) Feger, W., Brandauer, H., Gabris, P., and Ziegler, H. (2006) Nonvolatiles of commercial lime and grapefruit oils separated by highspeed countercurrent chromatography. J. Agric. Food Chem. 54, 2242– 2252. (4) Ostertag, E., Becker, T., Ammon, J., Bauer-Aymanns, H., and Schrenk, D. (2002) Effects of storage conditions on furocoumarin levels in intact, chopped or homogenized parsnips (Pastinaca satiVa L.). J. Agric. Food Chem. 50, 2565–2570. (5) Mongeau, R., Brassard, R., Cerkauskas, R., Chiba, M., Lok, E., Nera, E. A., Jee, P., McMullen, E., and Clayson, D. B. (1994) Effect of addition of dried healthy or diseased parsnip root tissue to a modified AIN-76A diet on cell proliferation nd histopathology in the liver, oesophagus and forestomach of male Swiss Webster mice. Food Chem. Toxicol. 32, 265–271. (6) COT (1996) Toxicity, Mutagenicity and Carcinogenicity Report 1996, Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment, http://www.archive.official-documents.co.uk/ document/doh/toxicity/chap-1c.htm. (7) SKLM (2006) Toxicological Assessment of Furocoumarins in Foodstuffs, DFG Senate Commission on Food Safety, Kaiserslautern, Germany. (8) IARC (1980) Some pharmaceutical drugs. IARC Monogr. EVal. Carcinog. Risk Chem. Hum. 24, 101-124. (9) IARC (1986) Some naturally occurring and synthetic food components, furocoumarins and ultraviolet radiation. IARC Monogr. EVal. Carcinog. Risk Chem. Hum. 40, 1-415. (10) Uwaifo, A. O., Billings, P. C., and Heidelberger, C. (1983) Mutation of Chinese hamster V79 cells and transformation and mutation of mouse fibroblast C3H/10T1/2 clone 8 cells by aflatoxin B1 and four other furocoumarins isolated from two Nigerian medicinal plants. Cancer Res. 43, 1054–1058. (11) Loveday, K. S., and Donahue, B. A. (1984) Induction of sister chromatid exchanges and gene mutations in Chinese hamster ovary cells by psoralens. Nat. Cancer Inst. Monogr. 66, 149–155. (12) Laquerbe, A., Guillof, C., Moustacchi, E., and Papadopoulo, D. (1995) The mutagenic processing of psoralen photolesions leaves a highly specific signature at an endogenous human locus. J. Mol. Biol. 254, 38–49. (13) Averbeck, D. (1985) Relationship between lesions photoinduced by mono- and bifunctional furocoumarins in DNA and genotoxic effects in diploid yeast. Mutat. Res. 151, 217–233. (14) Bailey, D. G., Malcolm, J., Arnold, O., and Spence, J. D. (1998) Grapefruit juicesDrug interactions. Br. J. Clin. Pharmacol. 52, 216– 217.

Raquet and Schrenk (15) Evans, A. M. (2000) Influence of dietary compounds on the gastrointestinal metabolism and transport of drugs. Ther. Drug Monit. 22, 131–136. (16) Baumgart, A., Schmidt, M., Schmitz, H.-J., and Schrenk, D. (2005) Natural furocoumarins as inducers and inhibitors of cytochrome P450 1A1 in rat hepatocytes. Biochem. Pharmacol. 69, 657–667. (17) Ramaswamy, S. (1975) Psoralens in foods. Ind. Food Pack. 8, 37– 46. (18) Bowers, A. G. (1999) Phytophotodermatitis. Am. J. Contact Dermatitis 10, 89–93. (19) Weber, I. C., Davis, C. P., and Greeson, D. M. (1999) Phytophotodermatitis: The other “lime” disease. J. Emerging Med. 17, 235–237. (20) Lombaert, G. A., Siemens, K. H., Pellaers, P., Mankotia, M., and Ng, W. (2001) Furanocoumarins in celery and parsnips: Method and multiyear Canadian survey. J. AOAC Int. 84, 1135–1143. (21) Stern, R. S., Thibodeau, L. A., Kleinerman, R. A., Parrish, J, A., and Fitzpatrick, T. B. (1979) Risk of cutaneous carcinoma in patients treated with oral methoxsalen photochemotherapy for psoriasis. N. Engl. J. Med. 300, 809–813. (22) Stern, R. S., and Lange, R. (1988) Non-melanoma skin cancer occurring in patients treated with PUVA five to ten years after first treatment. J. InVest. Dermatol. 91, 120–124. (23) Stern, R. S. (2001) The risk of melanoma in association with longterm exposure to PUVA. J. Am. Acad. Dermatol. 44, 755–761. (24) Schlatter, J., Zimmerli, B., Dick, R., Panizzon, R., and Schlatter, C. (1991) Dietary intake and risk assessment of phototoxic furocoumarins in humans. Food Chem. Toxicol. 29, 523–530. (25) Sayre, R. M., and Dowdy, J. C. (2008) The increase in melanoma: Are dietary furocoumarins responsible? Med. Hypotheses 70, 855– 859. (26) Dall’Acqua, F., Terbojevich, M., Marciani, S., Vedaldi, D., and Recher, M. (1978) Investigation of the dark interaction between furocoumarins and DNA. Chem.-Biol. Interact. 21, 103–115. (27) Isaacs, S. T., Wiesehahn, G., and Hallick, L. M. (1984) In vitro characterization of the reaction of four psoralen derivatives with DNA. Nat. Cancer Inst. Monogr. 66, 21–30. (28) Musajo, L., Rodighiero, G., Breccia, A., Dall’Acqua, F., and Malesani, G. (1966) The photoreaction between DNA and the skin-photosensitizing furocoumarins studied using labeled bergapten. Experientia 22, 75. (29) Rodighiero, G., Musajo, L., Dall’Acqua, F., Marciani, S., Carporale, G., and Ciavetta, L. (1970) Mechanisms of skin photosensitization by furocoumarins. Photoreactivity of various furocoumarins with native DNA and with ribosomal RNA. Biochim. Biophys. Acta 217, 40–49. (30) Musajo, L., and Rodighiero, G. (1972) Mode of photosensitizing action of furocoumarins. Photophysiology 7, 115–147. (31) Dall’Acqua, F., Vedaldi, D., Bordin, F., and Rodighiero, G. (1979) New studies on the interaction between 8-methoxypsoralen and DNA in vitro. J. InVest. Dermatol. 73, 191–197. (32) Ashwood-Smith, M. J. (1978) Frameshift mutations in bacteria produced in the dark by several furocoumarins; absence of activity of 4,5′,8-trimethylpsoralen. Mutat. Res. 58, 23–27. (33) Ashwood-Smith, M. J., Poulton, G. A., Barker, M., and Mildenberger, M. (1980) 5-Methoxypsoralen, an ingredient in several suntan preparations, has lethal, mutagenic and clastogenic properties. Nature 285, 407–409. (34) Pool, B. L., Klein, R., and Deutsch-Wenzel, R. P. (1982) Genotoxicity of 5- methoxypsoralen and near ultraviolet light in repair-deficient strains of Escherichia coli WP2. Food Chem. Toxicol. 20, 177–181. (35) Schimmer, O., and Abel, G. (1986) Mutagenicity of a furocoumarin epoxide, heraclenin, in Chlamydomonas reinhardii. Mutat. Res. 169, 47–50. (36) Bridges, B. A., and Mottershead, R. P. (1977) Frameshift mutagenicity in bacteria by 8-methoxypsoralen (methoxalen) in the dark. Mutat. Res. 44, 305–312. (37) NTP (1989) National Toxicology Program technical report on the toxicology and carcinogenesis studies of 8-methoxypsoralen (CAS No. 298-81-7) in F344/N rats (gavage studies). NIH Publication No. 892814, National Toxicology Program Technical Report Series, U.S. Department of Health and Human Services, Public Health Services, National Institutes of Health, Vol. 359, pp 1-130. (38) Abel, G., and Schimmer, O. (1986) Chromosome-damaging effects of heraclenin in human lymphocytes in vitro. Mutat. Res. 169, 51– 54. (39) Papadopoulo, D., and Moustacchi, E. (1990) Mutagenic effects photoinduced in normal human lymphoblasts by a monofunctional pyridopsoralen in comparison to 8-methoxypsoralen. Mutat. Res. 245, 259–266. (40) Natarajan, A. T., Verdegaal-Immerzeel, E. A., Ashwood-Smith, M. J., and Poulton, G. A. (1981) Chromosomal damage induced by furocoumarins and UVA in hamster and human cells including cells from patients with ataxia teleangiectasia and xeroderma pigmentosum. Mutat. Res. 84, 113–124.

Photomutagenicity of Furocoumarins and Limettin (41) Abel, G., and Mannschedel, A. (1987) The clastogenic effect of 5 methoxypsoralene plus UVA-A in human lymphocytes in vitro and its modification by the anticlastogenic beta-aminoethylosithiuronium. Hum. Genet. 76, 181–185. (42) Abel, G. (1987) Chromosome damage induced in human lymphocytes by 5-methoxypsoralen and 8-methoxypsoralen plus UV-A. Mutat. Res. 190, 63–68. (43) Chiou, C. C., and Yang, J. L. (1995) Mutagenicity and specific mutation spectrum induced by 8-methoxypsoralen plus a low dose of UVA in the hprt gene in diploid human fibroblasts. Carcinogenesis 16, 1357–1362. (44) Gruenert, D. C., Ashwood-Smith, M., Mitchell, R. H., and Cleaver, J. E. (1985) Induction of DNA-DNA cross-link formation in human cells by various psoralen derivatives. Cancer Res. 45, 5394–5398. (45) Wu, J. H., Wilson, J. B., Wolfreys, A. M., Scott, A., and Jones, N. J. (2009) Optimization of the comet assay for the sensitive detection of PUVA-induced DNA interstrand cross-links. Mutagenesis 24, 173– 181. (46) Bissonnette, L., Arnason, J. T., and Smith, M. L. (2008) Real-time fluorescence-based detection of furanocoumarin photoadducts of DNA. Phytochem. Anal. 19, 342–347. (47) Cao, H., Hearst, J. E., Corash, L., and Wang, Y. (2008) LC-MS/MS for the detection of DNA interstrand cross links formed by 8-methoxyporalem and UVA irradiation in human cells. Anal. Chem. 80, 2932–2938.

Chem. Res. Toxicol., Vol. 22, No. 9, 2009 1647 (48) Lai, C., Cao, H., Hearst, J. E., Corash, L., Luo, H., and Wang, Y. (2008) Quantitative analysis of DNA interstrand cross-links and monoadducts formed in human cells induced by psoralens and UVA irradiation. Anal. Chem. 80, 8790–8798. (49) Pathak, M. A., and Joshi, P. C. (1984) Production of active oxygen species (1O2 and O2-•) by psoralens and ultraviolet radiation (320400 nm). Biochim. Biophys. Acta 798, 115–126. (50) Jones, S. G., Young, A. R., and Truscott, T. G. (1993) Single oxygen yields of furocoumarins and related moleculessThe effect of excitation wavelength. J. Photochem. Photobiol. 21, 223–227. (51) Svobodova, A., Walterova, D., and Vostalova, J. (2006) Ultraviolet light induced alteration in the skin. Biomedical Papers of the Medical Faculty, Vol. 150, pp 25-38, University Palacky, Olomouc, Czech Republic. (52) Marrot, L., and Meunier, J.-R. (2007) Skin DNA photodamage and its biological consequences. J. Am. Acad. Dermatol. 58, S139–148. (53) Brickl, R., Schmid, J., and Koss, F. W. (1984) Pharmacokinetics and pharmacodynamics of psoralens after oral administration: Considerations and conclusions. Nat. Cancer Inst. Monogr. 66, 63–67. (54) Kornhauser, A., Wamer, W. G., and Giles, A. L. (1984) Difference in topical and systemic reactivity of psoralens: Determination of epidermal and serum levels. Nat. Cancer Inst. Monogr. 66, 97–101.

TX9002287