- Email: [email protected]
Dietary intake and risk assessment of phototoxic furocoumarins in humans
Fd Chem. Toxic. Vol. 29, No. 8, pp. 523-530, 1991 Printed in Great Britain.All rights reserved
0278-6915/91 $3.00+ 0.00 Copyright © 1991 PergamonPress pie
DIETARY INTAKE A N D RISK ASSESSMENT OF PHOTOTOXIC F U R O C O U M A R I N S IN H U M A N S SCHLATTER*t,B. ZIMMERLI*,R. DICK*, R. PANIZZON~and CH. SCHLATTER§ *Swiss Federal Office of Public Health, Division of Food Science, CH-3014 Bern, ~;UniversityHospital, Dermatology Department, CH-8091 Ziirich and §Institute of Toxicology, Swiss Federal Institute of Technology and University of Zfirich, CH-8603 Schwerzenbach, Switzerland J.
(Accepted 31 May 1991)
Abstract--The question of whether the furocoumarin content of vegetables is sufficient to induce phototoxic skin reactions after ultraviolet irradiation was examined in two experiments with four human volunteers. In a first experiment, 300 g of celery roots (total phototoxic furocoumarin content 28.2 #g/g) was ingested. No skin reactions were observed after UVA exposure (1.5-9 J/cm2), and the blood levels of the furocoumarins--psoralen, 8-methoxypsoralen (8-MOP) and 5-methoxypsoralen (5-MOP)--were below the analytical detection limit of 2 ng/ml. To investigate the phototoxic effects of a mixture of the two most important furocoumarins in vegetables, after gastro-intestinal uptake, 8-MOP and 5-MOP (15 mg each) were ingested separately in a 50% alcoholic solution. A strong and persistent erythema was induced in three out of the four subjects (UVA dose: 3-25 J/cm2). The blood levels immediately before UVA irradiation varied between 14 and 114 ng/ml (8-MOP), and 17 and 70 ng/ml (5-MOP). In the subject who did not show phototoxicity, the blood levels remained at trace levels (3 ng/ml). Two subjects were also tested with a mixture of 10 mg 8-MOP plus 10 mg 5-MOP; in one subject the mixture induced pigmentation only, while in the other a mild-to-medium erythema was induced. The blood levels of the furocoumarins in the two subjects were similar (12-15 ng/ml for 8-MOP and 5-MOP). It is concluded that in humans the phototoxic threshold dose of furocoumarin mixtures is of the order of 10 mg 8-MOP plus 10 mg 5-MOP, which is equivalent to about 15 mg 8-MOP per person (blood levelsof 8-MOP and 5-MOP at 30 min after ingestion of about 10-15 ng/ml each). This phototoxic threshold dose was not reached by the consumption of celery roots and other conventional vegetables under normal dietary habits (experimental intake of 2-8 mg per subject of the potentially phototoxic furocoumarin mixture). However, the safety factor between the possible actual intake of furocoumarins and the phototoxic threshold dose is about 2-10, which is relatively small.
INTRODUCTION Although there has been less public debate about potentially toxic chemical compounds that occur naturally in foods in comparison with man-made chemicals in the diet, this topic is of considerable importance. An example of such natural substances are the furocoumarins (secondary plant metabolites) with known phototoxic properties. They occur in Rutaceae and Umbelliferae, but also occur widely in members of other plant families, including some edible species such as celery, parsley, parsnips, citrus fruits, figs and certain spices. The major source of dietary furocoumarins is celery (see below). About 150 furocoumarins are known, some of which show a marked phototoxic activity in humans. It has long been known that, upon skin contact, the furocoumarins in celery are capable of inducing local phototoxic lesions in celery harvesters (Garfield, 1985; Scheel et al., 1963) as well as in grocery-store
tTo whom all correspondence should be addressed at: Federal Office of Public Health, Division of Food Science, c/o Institute of Toxicology, Swiss Federal Institute of Technology and University of Ziirich, Schorenstrasse 16, CH-8603 Schwerzenbach, Switzerland. Abbreviations: 5-MOP = 5-methoxypsoralen; 8-MOP = 8methoxypsoralen.
workers (Seligman et al., 1987). A topical or systemic furocoumarin exposure combined with long-wave UV radiation (UVA; wavelength 320--400 nm) of the skin typically leads to an erythematous, sunburn-like response and, in severe cases, to blistering. The onset of this toxic reaction is typically delayed for 12-36 hr after exposure, and the toxic reaction is followed within 1-2 wk by hyperpigmentation at the sites of exposure. Pigmentation only is observed at exposure levels near a phototoxic threshold dose. Furocoumarins are activated by UVA radiation and are able thereafter to bind to macromolecules (e.g. membranes, DNA), leading to cytotoxic (e.g. erythema) and genotoxic responses (Grossweiner, 1984; Yang et al., 1987 and 1988). Wavelengths between 350 and 365 nm (UVA) in the presence of furocoumarins induce the maximal phototoxic skin reactions in humans (Pathak, 1984; Scott et al., 1976). The main source of information about the phototoxicity of furocoumarins is the fact that some derivatives, particularly 8-methoxypsoralen (8-MOP) and to a lesser extent 5-methoxypsoralen (5-MOP), used as therapeutic agents in combination with UVA irradiation in the treatment of psoriasis and vitiligo (PUVA therapy). Typical doses of 0.4--0.7 mg 8-MOP/kg body weight are administered orally, and followed 2 hr later with UVA exposure (dose: 0.5-7 J/cm2). In 1984, the NCI published an extensive
523
J. SCHLATTERet al.
524
discussion of the photobiological, toxicological and pharmacological aspects of PUVA therapy, and a further review dealt with the molecular basis of furocoumarin reactions (Scott et al., 1976). Schlatter (1988) reviewed the data on the toxicity of furocoumarins with respect to the significance of a total dietary intake of a few mg per day, as well as the data on natural UV exposure; he concluded that the estimated natural UVA dose received by inhabitants of Switzerland would be high enough for phototoxic skin reactions in the presence of furocoumarins, even in non-alpine areas and during summer and winter-a period of 5-30 min outdoors between 10 am and 2 pm might be sufficient to reach the minimal phototoxic UVA dose. Monographs on carcinogenicity are available for 8-MOP (IARC, 1980 and 1982) and 5-MOP (IARC, 1986a), and for UV radiation (Bundesgesundheitsblatt, 1987; IARC, 1986b). IARC concluded that there is sufficient evidence for the carcinogenicity of PUVA, and the risk-benefit of a PUVA therapy has been repeatedly discussed (Epstein, 1979; Farber et al., 1983; Gibbs et al., 1986; Henseler and Christopher, 1984; Stern, 1984; Stern and Momtaz, 1984). Although normal levels of furocoumarins in vegetables are low (Beier et al., 1983; Ceska et al., 1986a and 1987), their concentrations can increase greatly following fungai invasion or other stress situations, such as exposure to UV light, cold or chemicals (Beier and Oertly, 1983; Ceska et al., 1986a,b; Chaudhary et al., 1985). Their induction by fungi and their in vitro toxicity to fungi are evidence that furocoumarins may act as antimicrobial defence compounds (i.e. phytoalexins) in some plants (Desjardins et al., 1989). In infected celery or carrots, furocoumarin concentrations greater than 10 times higher than normal have been measured (Ceska et al., 1986a; Chaudhary et al., 1985; Surico et al., 1987). In a recent Swiss survey (Baumann et al., 1988), the highest concentrations were found in celery in the edible part of the roots (up to 25/zg/g), in parsnips (40/ag/g) and in parsley (15/ag/g), while no furocoumarins were detectable in fennels, carrots or figs (detection limit 0.05 #g/g each). Apparently healthy celery-root samples had markedly higher concentrations of phototoxic furocoumarins than those given in the literature for non-infected samples (0.6--l.8/ag/g; Ashwood-Smith et al., 1985; Beier et al., 1983). The reason for this difference is not known, but could be due to the cultivation of different celery varieties. Furocoumarins are fairly thermostable and cooking will not significantly reduce their amounts. Hence, it has been concluded that in
some cases the individual furocoumarin intake may reach several mg per day, which approaches the doses used in photochemotherapy of psoriasis (AshwoodSmith et al., 1985; Chaudhary el aL, 1985; Schlatter, 1988). However, the threshold doses with respect to phototoxicity of mixtures of furocoumarins (both phototoxic and non-phototoxic) are not known at present. It was the aim of the present study to investigate the possibility of phototoxic skin reactions due to the consumption of mixtures of furoeoumarins as they occur usually in celery roots. MATERIALS
AND METHODS
Chemicals. Psoralen (P 8399; Sigma Chemical Co., St Louis, MO, USA), 5-MOP (27, 52-7; Aldrich Chemical Co., Milwaukee, WI, USA), 8-MOP M3501; Sigma) and 5,8-dimethoxypsoralen (2-8854; Carl Roth GmbH and Co., Karlsruhe, Germany), as well as all solvents used, were of analytical grade and used without further purification. Subjects. Four healthy male volunteers (subjects A-D), all academic personnel from the University, agreed to participate both in the experiment with celery roots and in the experiment with 8-MOP plus 5-MOP mixtures. Their respective body weights are listed in Table I. All subjects were classified as belonging to the same skin type (class III, according to Greiter, 1984: fair skin, no freckles, brown hair and grey or brown eyes). For this skin type the threshold dose for UVA alone (32ff-400nm) for pigmentation is expected to be 17-20 J/cm 2, for UVB (280-320nm) 0.021J/cm 2, and for erythema formation 25-35J/cm 2 (UVB 0.036J/cm2; Bundesgesundheitsblatt, 1987). In subjects A, C and D the threshold UVA dose for skin pigmentation was verified to be 1>20 J/cm 2. Experimental procedure. During autumn to spring 1988-1989, 70 samples of celery roots (Apium graveolens L.) were bought at retail stores in Bern and analysed for furocoumarin content (for analytical details see Baumann et al., 1988). For phototoxicity testing, 2340 g of celery roots (eight items) was obtained from a local shop in Schwerzenbach and stored for 2 months (4°C). On the day of the experiment, the celery roots were washed, shredded and mixed. An aliquot of 50 g was frozen for subsequent furocoumarin analysis. In experiments where pure chemicals were used, 8-MOP and 5-MOP were dissolved in 15 ml ethanol and diluted with 15 ml water immediately before consumption. UVA irradiation was performed at the dermatology department of the University Hospital Ziirich, using a Waldmann-800 UVA lamp (verified emission of 7 x 10 -3 W/cm 2 and
Table 1. Skin reactions of four human subjectsafter oral administrationof a mixture of 8-MOP plus 5-MOP Skin reaction after UVA dose (J/cm 2) of Subject 8-MOP + (body weight, kg) 5-MOP dose (mg) Day of maximalresponse 3 6 9 12 18 25 A (65) 10 + 10 7 -El El El E2 E2 15 + 15 7 El E2 E2 E3 E3 E3 B (63) 10 + 10 2 PI P2 P3 P3 P3 P3 15 + 15 4 -El E2 El E3b E3b c (85) 15 + 15 . . . . . . . D(74) 15 + 15 8 -El E2 E2 E3 E3 b = blistering
E = erythema
P = pigmentation 1, 2, 3 = slight, medium or strong reaction, respectively
Phototoxicity of furocoumarins in humans
525
UVB emission of 1 x 10-6W/cm 2 at a distance of (pglg) 20cm). Six non-pigmented circular skin areas of a0 O 2.5 cm diameter, in the gluteal region (lower back), were exposed to six UVA doses (maximal UVB dose O O O was 0.00006 J/cm 2, which is far below the threshold dose for pigmentation--see above). In the first experz0 O O iment, all subjects ingested 300 g of celery roots after a fasting period of at least 4 hr. Venous blood 8 8 o samples (10 ml, heparin glass vials; B-D Vacutainer) 0 O0 O÷ - C o c o were drawn immediately before ingestion of the 1o o oo~_ celery, and 2 and 4 hr thereafter. Blood samples were o a~o frozen and stored at - 2 0 ° C until chemical analysis. O O UVA irradiation was started 2 hr after ingestion of 0 i i i i i i ¢ ¢ i / i the celery (1.5, 2, 3, 5, 7 and 9 J/cm2). Skin readings were performed after 10 min, and after 4, 24, 48 and Fig. 1. Time-course of the content of phototoxic furocou72 hr. In the second experiment, all four subjects marins (sum of psoralens 8-MOP and 5-MOP) in celery-root ingested a mixture of 15mg 8-MOP plus 15rag items (ready-made tubers) from autumn 1988 to spring 1989. Arithmetic mean (--). 5-MOP in 30 ml 50% ethanol, after a fasting period of at least 4 hr. Venous blood samples were drawn after 30 and 90 min, and the UVA irradiation was by small superficial spots. The concentrations of started 30 min after ingestion of the mixture (UVA psoralen were generally the lowest (median 0.26, dose: 3, 6, 9, 12, 18 and 25 J/cm2). Skin readings were mean 1.07, range <0.1 to 10.29#g/g). The concenperformed after 10 min, and after 4, 24, 48 and 72 hr. trations of 8-MOP and 5-MOP varied widely but In cases where the intensity of the erythema was still tended to be about equal (8-MOP: median 2.28, mean increasing, daily readings were continued until the 3.81, range 0.43 to 21.62 #g/g; 5-MOP: median 1.92, maximal response was reached. In subjects A and B, mean 2.46, range 0.65 to 12.53 gg/g). Some nona dose of 10 mg 8-MOP plus 10 mg 5-MOP was also phototoxic 5,8-dimethoxypsoralen (isopimpinellin) tested under the same experimental conditions. The was also present in the samples, but not quantified time interval between the different doses and exper- because of interference. It is obvious from Fig. 1 that iments was at least I wk. For a comparison of the the mean concentrations of the phototoxic furofurocoumarin blood levels after ingestion of an alco- coumarins (mainly 5-MOP and 8-MOP) did not holic solution with those after ingesting furocou- , increase in the time between harvesting (August) and matins in celery, subjects A and B ingested mixtures the date of sampling, as had been expected at first. of 8-MOP plus 5-MOP at two dose levels (5 mg and Instead, the mean concentrations remained relatively 7.5mg each) in 30ml 50% ethanol without sub- constant throughout the whole season (median sequent UVA irradiation. Blood samples were taken 4.6#g/g, arithmetic mean 7.3#g/g, range 1.1 to at 30 and 90 min after ingestion. The 'non-responder' 28.3/~g/g; Figs 1 and 2). of experiment 2, subject C, ingested a mixture of 15mg 8-MOP plus 15mg 5-MOP in 40ml 50% Blood analysis of furocoumarins ethanol for a second time without subsequent UVA Addition of psoralen, 5-MOP, 8-MOP and 5,8irradiation. Blood samples (5 ml) were taken 15, 30, dimethoxypsoralen to bovine blood (20 ng/g each) 60, 90, 120 and 180 min after ingestion of the mixture resulted in mean recoveries ( + S D , n = 3) of 91 + 3, through a venous catheter, using disposable syringes 93 + 3, 95 + 2 and 91 + 6%, respectively. Therelative (Omnifix, Braun Melsungen AG, Germany). error of a single measurement can be assumed to be Analyticalprocedure. The procedure developed for of the order of 10%. Figure 3 presents some typical the analysis of furocoumarins in vegetables, as de- chromatograms of the reagent blank, standard solscribed previously (Baumann et al., 1988), has been ution and human plasma, after ingestion of furocouadapted for blood. Instead of the addition of water marins. The detection limits for psoralen, 5-MOP and to the sample, physiological sodium chloride solution 8-MOP in blood were in the range of 0.5 to 2 ng/ml was used. The final extract was concentrated to 50 #1 instead of 5 ml, and 20 #! was injected into the HPLC (,) system (plus UV-detector). All the precautions ]6.r usually followed in trace analysis of organic corn]. pounds, such as blanks (i.e. of reagents and syringes and tubes used for drawing blood), were strictly ~2 observed. RESULTS
Occurrence of furocoumarins in ready-made celery roots
In Fig. 1, the analytical result of 70 samples of apparently healthy celery roots (fresh material), offered at the retail level, were plotted against the month of sampling. Probable fungal infection of the celery roots in about 7% of the items was indicated
'hllllflt_.
0
0
S
10
lS
20
25
(pg/g)
Fig. 2. Frequency distribution of the concentrations of total phototoxic furocoumarins in celery from autumn 1988 to spring 1989.
J. SCHLATTERet al.
526
I
I
0
5
I
10
I'
I
I
I
15
0
5
10
m
I
l
I
s
15
0
5
10
15
mln
m|n
mln
Fig. 3. Chromatograms of: (A) reagent blank; (B) standard solution (34 ng 8-MOP plus 24 ng 5-MOP, injected); (C) blood plasma of subject B, 90 min after ingestion of 10mg 8-MOP plus 10 mg 5-MOP (5.8 ng/ml 8-MOP plus 12.5 ng/ml 5-MOP). each, depending on the chromatography system and the respective sensitivity of the detector.
Phototoxicity after ingestion of celery roots The celery roots consumed by the four subjects revealed a total furocoumarin content of 28.2 pg/g (psoralen, 2.3 pg/g; 8-MOP, 3.8 yg/g; 5-MOP, 2 2 . 1 # g / g fresh material; 5,8-dimethoxypsoralen, < 0 . 0 5 y g / g ) and the total furocoumarin dose ingested was, therefore, 8.4 mg/subject (0.100.13 mg/kg body weight). Throughout the whole observation period after U V A irradiation (up to 9 J / c m 2) in all subjects, there was no observable pigmentation or erythema formation. The furocoumarin blood levels were below the analytical detection limit ( < 2 ng/ml each).
Phototoxicity after ingestion of mixtures of 8,MOP plus 5-MOP in ethanol The results of the second experiment are summarized in Tables 1 and 2. The ingestion of a mixture of 15 mg 8-MOP plus 15 mg 5-MOP clearly induced skin erythemas in subjects A, B and D, while neither erythema formation nor pigmentation was observed in subject C (Table 1). In subject A, the response became visible after 24 hr with t> 9 J U V A / c m 2. The intensity of the erythema was UVA-dose-dependent, and the subject complained of strong itching at the skin areas of higher U V A doses. In subjects B and D,
the first effects were already visible 4 hr after treatment with a U V A dose of >_-9 (subject B) or >t 18 J/cm 2 (subject D). In subject D, cortisone was topically applied to the skin areas of the higher U V A doses because of strong discomfort. The blood levels of 8-MOP and 5-MOP immediately before and after U V A irradiation showed individual patterns (Table 2). The blood levels were highest in subject A, followed by subjects B and D. In subject C, the blood levels of 8-MOP and 5-MOP remained very low, just slightly above the analytical detection limit. The mixture of 10 mg 8-MOP plus 10 mg 5-MOP ingested by subject B induced pigmentation after 2 4 h r at U V A doses of >19J/cm 2. N o erythema was induced and the blood levels of 8-MOP and 5-MOP were initially about 12 ng/ml each. In subject A, given 10mg each of 8-MOP and 5-MOP, a slight pigmentation of the skin areas exposed to the three highest U V A doses was seen after 4 and 24 hr, and this developed into an erythema within 48 hr. Blood levels of 8-MOP and 5-MOP were similar to those measured earlier in subject B. Also, similar blood levels in volunteer A compared with those of volunteer B were observed after the ingestion of a mixture of 5 mg 8-MOP plus 5 mg 5-MOP, as well as after the ingestion of a mixture of 7.5 mg 8-MOP plus 7.5 mg 5-MOP, both at 30 and 90 min.
Table 2. Blood levelsof 8-MOP and 5-MOP (ng/ml) after oral administration of a mixture of 8-MOP plus 5-MOP in 50oVoethanol solution Blood level of 5-MOP (ng/ml) Blood level of 8-MOP (ng/ml) Subject 8-MOP + (body weight, kg) 5-MOP dose (mg) 30 rain 90 min 30 min 90 min A (65) 5+ 5 3.7 <0.5 6.2 1.4 7.5 + 7.5 23.4 4.2 27.2 9.9 10+ 10 14.2 5.1 15.4 11.4 15 + 15 113.9 18.7 69.6 28.1 B (63) 5+ 5 3.7 <0.5 4.2 2.4 7.5 + 7.5 26.0 3.2 32.7 10.3 (PI 10.6)* (5.8) 12.0 (12.9) 9.1 (12.5) 10 + 10 11.9 (EC 13.4) 4.6 (3.3) (11.0) (5.2) 15 + 15 64.7 24.1 50.4 64.0 ( ~ 3.0) ~ 3.0) ,,~3.0) ~2.5 ((~2.0) ~2.5 ((~2.0) C (85) 15 + 15 ~3.0 (Pl (EC~~3.0) 3.0) ~,2.5 (~2.0) D(74)
15+15
(PI 20.5) 13.7(EC6.1 )
(14.6) 12.2 (9.4)
(24.7) 16.8 ( 7 . 8 )
(21.3) 16"5(11.1)
*In several experiments, plasma (PI) and erythrocyte concentrations (EC) were determined separately (values given in parentheses). The concentration in whole blood was then calculated, assuming 53% plasma and 47% erythrocytes.
Phototoxicity of furocoumarins in humans Table 3. Time-courseof blood levelsof 8-MOP and 5-MOP (ng/ml) in subjectC after oral administrationof a mixture of 15mg 8-MOP plus 15mg 5-MOP in 50*/, ethanol Time (rain) Furocoumarin 8-MOP* 5-MOP
15 <3 <2
30 60 90 120 180 <3 ~<3 10.2 5.6 -~<2 2.7 13.4 11.4 <2 = not measured *The detectionlimitwas about 3 ng/ml becauseof an eluantchange in HPLC, which resulted in higher retention time of 8-MOP. This change was necessaryin order to separate8-MOP from an interfering substanceoriginatingfrom the piston of the disposable syringe,type Omnifix(see Experimentalprocedure). -
-
In subject C, the possibility of a different kinetic or metabolic behaviour was tested after an oral dose of 15 mg 8-MOP plus 15 mg 5-MOP by monitoring blood levels over a longer period. The data (Table 3) again showed low blood levels during the first hour after ingestion of a mixture of 15 mg 8-MOP plus 15rag 5-MOP. A slight increase in 8-MOP and 5-MOP blood levels was observed after 90 min, and declined again thereafter. DISCUSSION
Blood analysis o f furocoumarins Most of the published analytical methods for plasma or serum use reverse-phase HPLC with UV detection (Gasparro et al., 1988; Herfst et al., 1980; Kornhauser et al., 1984; Monbaliu et al., 1981; Stolk et al., 1987; Susanto et aL, 1986) or fluorescence detection (Prognon et aL, 1983). Gas chromatography has also been used, and electron capture detection results in an increased sensitivity and selectivity compared with flame ionization detection (Ehrsson et al., 1977; Schmid and Koss, 1978). Although different authors (e.g. Carter and Goldstein, 1984; Stolk et al., 1987) have described analytical problems such as loss of analyte by adherence to container walls or by volatilization during the procedure (i.e. concentration steps), the recoveries obtained in the present work were acceptable in spite of the fact that the procedure includes several evaporation steps. Phototoxicity after ingestion o f celery roots The metabolism of furocoumarins (Bickers and Pathak, 1984; De Wolff and Thomas, 1986; Rogers et al., 1983) shows a saturable first-pass effect in the liver (Schmid et al., 1980). Only oral doses that significantly elevate plasma levels lead to phototoxic effects (Brickl et al., 1984a,b). At these plasma levels, furocoumarins are found in skin suction blister fluid (Herfst and De Wolff, 1982; Kornhauser et al., 1984; Korting et al., 1982). In our experiment with the ingestion of 300 g of celery roots (roughly double or triple the amount of a normal portion, per capita consumption 4 g)--equivalent to about 10mg of phototoxic furocoumarins--the blood levels remained below the analytical detection limit, which explains the absence of phototoxic skin reactions. However, a similar amount of total furocoumarins (5 mg each of 8-MOP and 5-MOP) ingested in an alcoholic solution led to measurable blood levels. Therefore, almost all furo-
527
coumarins must have been cleared from the blood during the first liver passage after the ingestion of celery roots, probably because of a protracted uptake of the furocoumarins from the gastro-intestinal tract in comparison with the ingestion as an alcoholic solution, or because of a low bioavailability of the furocoumarins in celery. It is known from PUVA therapy that concomitant food intake can influence 8-MOP and S-MOP blood levels (Herfst and De Wolff, 1982 and 1983; Mosher et aL, 1983), as can both the galenic form of the furocoumarin application (Siddiqui et aL, 1984; Stolk et al., 1985) and the mode of administration (Kornhauser et aL, 1984). Phototoxicity after ingestion o f mixtures o f 8 - M O P plus 5 - M O P in ethanol To eliminate interfering factors with furocoumarin uptake and to determine the safety factor between dietary furocoumarin uptake and the phototoxic threshold dose, we tried to maximize the uptake by using ingestion of an alcoholic solution of a mixture of the pure substances of the main components in celery roots, namely 8-MOP and 5-MOP, after a fasting period. 10, 20 or 40rag 8-MOP per person gave rise to peak plasma levels of 8-MOP of about 12, 180 and 550ng/ml after 30rain (Schmid et aL, 1980). A minimal 8-MOP serum concentration of 10-25ng/ml combined with UVA doses of 10-20J/cm 2 is phototoxic (Ljunggren et aL, 1981). Phototoxicity seems to be slightly delayed compared with plasma levels in the absorption phase, but it persists much longer than plasma levels. It has been recommended that subjects be irradiated at or shortly after peak plasma levels (Brickl et aL, 1984), and a threshold dose for erythema formation from orally administered 8-MOP alone plus UVA irradiation was calculated to be of the order of 14rag/person (0.24 mg/kg body weight). However, the potency of the different furocoumarin derivatives in inducing phototoxicity after oral administration varies (in PUVA therapy, approximately a two-fold dose of 5-MOP is required to induce the same effect as 8-MOP), and there is some evidence that the different derivatives might act multiplicatively with respect to phototoxicity; if 5-MOP is administered 30 rain before 8-MOP (Bricki et aL, 1984a,b) plasma levels of 8-MOP are increased, and the combination is much more phototoxic than the individual compounds at the same dose. Therefore, we gave a dose of 10 mg 8-MOP plus 10 rag S-MOP to volunteer B. His blood level of 8-MOP was in good agreement with the published data. However, the skin response was clearly only pigmentation and not erythema, even at the highest UVA dose of 25 J/cm 2. This was the first indication that the simultaneous administration of the two furocoumarins act additively rather than multiplicatively as was assumed by Schlatter (1988), based on the experiments of Brickl et al. (1984a,b). Therefore, the 8-MOP plus 5-MOP dose was increased to 15 mg each for further tests. This produced a strong phototoxic skin reaction in all but one subject, and the measured blood levels in the responding subjects were in agreement with the published data. Again, the simultaneous ingestion of the two furocoumarins produced only an additive response. The interindividual differences in blood levels of
528
J. SCHLATTERet al.
the three responding subjects may be due to differences in pharmacokinetics but, interestingly, the phototoxicity in these three subjects was similar. This is in agreement with the observation of De Bersaques et al. (1984); the rate of furocoumarin absorption does not seem to influence the phototoxic skin reaction much, once a critical blood level is exceeded. The dose mixture of 10mg 8-MOP plus 10mg 5-MOP induced clear erythema in subject A, but in subject B only pigmentation was observed, although the furocoumarin blood levels of subject A were only marginally higher. Thus, this dose mixture is very close to the phototoxic threshold dose. In subject C, the blood levels of 8-MOP and 5-MOP were much lower than in the other subjects. In addition, the peak blood levels were delayed to approximately 90 min. The reason for this unexpectedly low and late rise in blood furocoumarin levels is not known. However, it explains the absence of phototoxicity in this subject upon UVA exposure between 30 and 90 min after ingesting the mixture. This demonstrates again the sometimes unpredictable pharmacokinetic behaviour of the furocoumarins as is known from PUVA therapy (Carter and Goldstein, 1984; De Wolff and Thomas, 1986), and that the intraindividual variation is generally low (factor 1.3-5.3) compared with the interindividual variation (Schfifer and Korting, 1982). The apparently higher initial blood furocoumarin level in subjects A and B (at 30 min, during the expected steep rise of blood levels, but not at 90 min) after ingestion of the 7.5 mg dose mixture compared with the 10 mg dose mixture may also be due to such intraindividual variations. From all these data, we conclude that in humans the approximate value of a phototoxic threshold dose of furocoumarin mixtures is of the order of 10 mg 8-MOP plus 10mg 5-MOP--equivalent to about 15 mg 8-MOP per person (blood levels of 8-MOP and 5-MOP at 30min after ingestion of about 10-15 ng/ml each)--which is in good agreement with the calculated threshold dose for 8-MOP alone of 14 mg/person (Brickl et al., 1984a) when taking into account the weaker potency of orally administered 5-MOP as described earlier. This phototoxic threshold dose is not reached by the consumption of celery roots at the Swiss retail level (maximum 200 g, maximal concentration of the sum of phototoxic furocoumarins 30-50#g/g). However, the safety factor between the possible actual intake of furocoumarins and the phototoxic threshold dose is relatively small, being of the order of about 2-10. In this light, the only recently published case report of a severe phototoxic burn after ingestion of unusually large amounts of celery with an exceptionally high furocoumarin level (estimated total furocoumarin concentration equivalent to 100#g 8-MOP/g) is not surprising (Ljunggren, 1990): a 65-year-old vegetarian (skin type III) ingested about 450 g celery one hour before a visit to a suntan parlour (estimated UVA dose about 20J/cm2), and suffered 48 hours later from a generalized, intense erythema with multiple large blisters, oedema and fever. We are aware of all the limitations of our experimental conditions (e.g. limited number of volunteers of one skin type, fixed furocoumarin mixture tested, fixed time of UVA irradiation, no separate phototox-
icity testing of the single furocoumarin components, no continuous blood level monitoring), but because of ethical reasons experimentation with a human carcinogen has to be kept to a minimum. Nevertheless, the risk assessment presented above, with all its inherent uncertainties, allows the following conclusions: (i) under normal dietary habits there is presently no risk of phototoxic burn due to the ingestion of celery and other conventional vegetables; (ii) extreme habitual consumption must be avoided, at least when subsequent UVA exposure can not be excluded; and (iii) when breeding attempts are made to improve the quality, resistance against diseases or other properties of the celery plant, a possible increase in the furocoumarin content must be avoided. REFERENCES
Ashwood-Smith M. J., Ceska O. and Chaudhary S. K. (1985) Mechanism of photosensitivity reactions to diseased celery. British Medical Journal 290, 1249 Baumann U., Dick R. and Zimmerli B. (1988) Orientierende Untersuchung zum Vorkommen yon Furocoumarinen in pflanzlichen Lebensmitteln und Kosmetika. Mitteilungen aus dem Gebiet der Lebensmitteluntersuchung und Hygiene 79, 112-129. Beier R. C., Ivie G. W., Oertli E. H. and Holt D. L. (1983) HPLC analysis of linear furocoumarins (psoralens) in healthy celery (Apium graveolens). Food and Chemical Toxicology 21, 163-165. Beier R. C. and Oertli E. H. (1983) Psoralens and other linear furocoumarins as phytoalexins in celery. Phytochemistry 22, 2595-2597. Bickers D. R. and Pathack M. A. (1984) Psoralen pharmacology, studies on metabolism and enzyme induction. National Cancer Institute Monographs 66, 77-84. Brickl R., Schmid J. and Koss F. W. (1984a) Pharmacokinetics and pharmacodynamics of psoralens after oral administration: considerations and conclusions. National Cancer Institute Monographs 66, 63-67. Brickl R., Schmid J. and Koss F. W. (1984b) Clinical pharmacology of oral psoralen drugs. Photodermatology 1, 174-186. Bundesgesundheitsblatt (1987) Empfehlungen zur Begrenzung gesundheitlicherStrahlenrisiken bei der Anwendung von Solarien und Heimsonnen. Bundesgesundheitsblatt 30, 19-30. Carter D. M. and Goldstein D. P. (1984) Quantitative tests for psoralens in the blood; methods and uses in monitoring psoralen and longwave radiation therapy. National Cancer Institute Monographs 66, 69-72. Ceska O., Chaudhary S. K., Warrington P. J. and AshwoodSmith M. J. (1986a) Furocourmarins in the cultivated carrot, Daucus carota. Phytochemistry 25, 81-83. Ceska O., Chaudhary S. K., Warrington P. J. and Ashwood M. J. (1987) Photoactive furocoumarins in fruits of some umbellifers. Phytochemistry 26, 165-169. Ceska O., Chaudhary S, K., Warrington P. J., Poulton G. and Ashwood-Smith M. J. (1986b) Naturally-occurring crystals of photocarcinogenic furocoumarins on the surface of parsnip roots sold as food. Experientia 42, 1302-1304. Chaudhary S. K., Ceska O., Warrington P. J. and AshwoodSmith M. J. (1985) Increased furocoumarin content of celery during storage. Journal of Agricultural and Food Chemistry 33, 1153-1157. De Bersaques J., Monbaliu J., Belpaire F. and Bogaert M. (1984) 8-Methoxy-psoralen plasma levels and phototoxic effect. Acta Dermato-Venereologica, Stockholm Suppl. 113, 166-167.
Phototoxicity of furocoumarins in humans Desjardins A. E., Spencer G. F. and Plattner R. D. (1989) Tolerance and metabolism of furocoumarins by the phytopathogenic fungus Gibberella pulicaris (Fusarium sambicinum ). Phytochemistry 28, 2963-2969. De Wolff F. A. and Thomas T. V. (1986) Clinical pharmacokinetics of methoxalen and other psoralens. Clinical Pharmacokinetics 11, 62-75. Ehrsson H., Eksborg S., Wallin I., Kallberg N. and Swanbeck G. (I 977) Determination of 8-methoxypsoralen in plasma by electron capture gas chromatography. Journal of Chromatography 140, 157-164. Epstein J. H. (1979) Risk and benefits of the treatment of psoriasis. New England Journal of Medicine 300, 852-853. Farber E. M., Abel E. A. and Cox A. J. (1983) Long-term risks of psoralen and UV-A therapy for psoriasis. Archives of Dermatology 119, 426-431. Garfield E. (1985) Current comments. From tonic to psoriasis: stalking celery's secrets. Current Contents 18, 3-12. Gasparro F. P., Battista J., Song J. and Edelson R. L. (1988) Rapid and sensitive analysis of 8-methoxypsoralen in plasma. Journal of lnvestigative Dermatology 90, 234-236. Gibbs N. K., H6nigsmann H. and Young A. R. (1986) PUVA treatment strategies and cancer risk. Lancet i, 150-151. Greiter F. (1984) Sonne und Gesundheit. Gustav Fischer Verlag, Stuttgart. Grossweiner L. I. (1984) Mechanisms of photosensitization by furocoumarins. National Cancer Institute Monographs 66, 47-54. Henseler T. and Christopher E. (1984) Risk of skin tumors in psoralen- and ultraviolet A-treated patients. National Cancer Institute Monographs 66, 217-219. Herfst M. J. and De Wolff F. A. (1982) Influence of food on the kinetics of 8-methoxypsoralen in serum and suction blister fluid in psoriatic patients. European Journal of Clinical Pharmacology 23, 75-80. Herfst M. J. and De Wolff F. A. (1983) Intraindividual and interindividual variability in 8-methoxypsoralen kinetics and effect in psoriatic patients. Clinical Pharmacology and Therapeutics 34, 117-124. Herfst M. J., Edelbroek P. M. and De Wolff F. A. (1980) Determination of 8-methoxypsoralen in suction-blister fluid and serum by liquid chromatography. Clinical Chemistry 26, 1825-1828. IARC Working Party (1980) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Some Pharmaceutical Drugs. Vol. 24, Methoxsalen. pp. 101-124. International Agency for Research on Cancer, Lyon. IARC Working Party (1982) 1ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Chemicals, Industrial Processes and Industries Associated with Cancer in Humans. Suppl. 4. Methoxsalen with Ultra-violet A Therapy (PUVA) (group 1). pp. 158-160. International Agency for Research on Cancer, Lyon, IARC Working Party (1986a) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Some Naturally Occurring and Synthetic Food Components, Furocourmarins and Ultraviolet Radiation. Vol. 40. 5-Methoxypsoralen. pp. 327-347. International Agency for Research on Cancer, Lyon. IARC Working Party (1986b) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Some Naturally Occurring and Synthetic Food Components, Furocoumarins and Ultraviolet Radiation. Vol. 40. Appendix: Ultraviolet Radiation. pp. 379-404. International Agency for Research on Cancer, Lyon. Kornhauser A., Wamer W. G. and Giles A. L., Jr (1984) Difference in topical and systemic reactivity of pseralens: determinations of epidermal and serum levels. National Cancer Institute Monographs 66, 97-101.
529
Korting H. Ch., Sch/ifer M., Roser E. and Mutschler E. (1982) Determination of 8-methoxypsoralen levels in plasma and skin suction blister fluid by a new sensitive fluorodensitometric method. Archives of Dermatological Research 272, 9-20. Ljunggren B. (1990) Severe phototoxic burn following celery ingestion. Archives of Dermatology 126, 1334--1336. Ljunggren B., Bjellerup M. and Carter M. (1981) Dose-response relations in phototoxicity due to 8methoxypsoralen and UV-A in man. Journal of Investigative Dermatology 76, 73-75. Monbaliu J. G., Rosseel M. T. and Bogaert M. G. (1981) Analysis of 8-methoxypsoralen in plasma by reversedphase high-performance liquid chromatography. Journal of Pharmaceutical Sciences 70, 965-966. Mosher D., Momtaz K., Borowska Z., Caldwell D. and Carter D, M. (1983) Bioavailability of 8-methoxypsoralen: another determinant. Clinical Research 31, 590A. NCI (1984) Photobiologic, toxicologic, and pharmacologic aspects of psoralens, National Cancer Institute Monograph 66, NIH publication no. 84-2692. Pathak M. A. (1984) Mechanisms of psoralen photosensitization reactions. National Cancer Institute Monographs 66, 41-46. Prognon P. M., Simon G. and Mahuzier G. (1983) Dosage du mrthoxy-5-psoralrne dans le plasma par chromatographie en phase liquide et drtection spectrofluorimrtrique. Journal of Chromatography 272, 193-199. Rogers S. L., Staubus A. E., Bianchine J. R. and Gerber N. (1983) Michaelis-Menten pharmacokinetics of 8methoxypsoralen. Federation Proceedings. Federation of the American Societies for Experimental Biology 42, 1139. Sch/ifer M. and Korting H. Ch. (1982) Intraindividual variations of 8-methoxypsoralen plasma levels. Archives of Dermatological Research 272, 1-7. Scheel L. D., Perone V. B., Larkin R. L. and Kupel R. E. (1963) The isolation and characterization of two phototoxic furanocoumarins (psoralens) from diseased celery. Biochemistry, New York 2, 1127-1131. Schlatter J. (1988) Die toxikologische Bedeutung von Furocoumarinen in pflanzlichen Lebensmitteln. Mitteilungen aus dem Gebiet der Lebersmitteluntersuchung und Hygiene 79, 130-143. Schmid J. and Koss F. W. (1978) Rapid, sensitive gas chromatographic analysis of 8-methoxypsoralen in human plasma. Journal of Chromatography 146, 498-502. Schmid J., Prox A., Zipp H. and Koss F. W. (1980) The use of stable isotopes to prove the saturable first-pass effect of methoxalen. Biomedical Mass Spectrometry 7, 560-564. Scott B. R., Pathak M. A. and Mohn G. R. (1976) Molecular and genetic basis of furocoumarin reactions. Mutation Research 39, 29-74. Seligman P. J., Mathias T. C. G., O'Malley M. A., Beier R. C., Fehrs L. J., Serrill W. S. and Halperin W. E. (1987) Phytophotodermatitis from celery among grocery store workers. Archives of Dermatology 123, 1478-1482. Siddiqui A. H., Stolk L. M. L. and Cormane R. H. (1984) Comparison of serum levels and clinical results of PUVA therapy with three different dosage forms of 8-methoxypsoralen. Archives of Dermatological Research 276, 343-345. Stern R. S. (1984) Carcinogenic risk of psoralen plus ultraviolet radiation therapy: evidence in humans. National Cancer Institute Monographs 66, 211-216. Stern R. S. and Momtaz K. (1984) Skin typing for assessment of skin cancer risk and acute response to UV-B and oral methoxsalen photochemotherapy. Archives oJ Dermatology 120, 869-873. Stolk L. M. L., De Ruiter R., Saadawi A., Siddiqui A. H. and Cormane R. H. (1987) Determination of psoralen
530
J. SCHLATTERet al.
in serum by reversed-phase high-performance liquid chromatography. Journal of Chromatography 423, 383-386. Stolk L. M. L., Siddiqui A. H., Westerhof W. and Cormane R. H. (1985) Comparison of bioavailability and phototoxicity of two oral preparations of 5-methoxypsoralen. British Journal of Dermatology 112, 469--473. Surico G., Varvaro L. and Solfrizzo M. (1987) Linear furocoumarin accumulation in celery plants infected with Erwinia carotovora pv. carotovora. Journal of Agricultural and Food Chemistry 35, 406--409.
Susanto F., Humfeld S. and Reinauer H. (1986) Highperformance liquid chromatographic measurement of 8-methoxypsoralen in plasma. Chromatographia 21, 443-446. Yang X. Y., DeLeo V. and Santell R. M. (1987) Immunological detection and visualization of 8-methoxypsoralenDNA photoadducts. Cancer Research 47, 2451-2455. Yang X. Y., Delohery T. and Santella R. M. (1988) Flow cytometric analysis of 8-methoxypsoralen-DNA photoadducts in human keratinocytes. Cancer Research 48, 7013-7017.
0278-6915/91 $3.00+ 0.00 Copyright © 1991 PergamonPress pie
DIETARY INTAKE A N D RISK ASSESSMENT OF PHOTOTOXIC F U R O C O U M A R I N S IN H U M A N S SCHLATTER*t,B. ZIMMERLI*,R. DICK*, R. PANIZZON~and CH. SCHLATTER§ *Swiss Federal Office of Public Health, Division of Food Science, CH-3014 Bern, ~;UniversityHospital, Dermatology Department, CH-8091 Ziirich and §Institute of Toxicology, Swiss Federal Institute of Technology and University of Zfirich, CH-8603 Schwerzenbach, Switzerland J.
(Accepted 31 May 1991)
Abstract--The question of whether the furocoumarin content of vegetables is sufficient to induce phototoxic skin reactions after ultraviolet irradiation was examined in two experiments with four human volunteers. In a first experiment, 300 g of celery roots (total phototoxic furocoumarin content 28.2 #g/g) was ingested. No skin reactions were observed after UVA exposure (1.5-9 J/cm2), and the blood levels of the furocoumarins--psoralen, 8-methoxypsoralen (8-MOP) and 5-methoxypsoralen (5-MOP)--were below the analytical detection limit of 2 ng/ml. To investigate the phototoxic effects of a mixture of the two most important furocoumarins in vegetables, after gastro-intestinal uptake, 8-MOP and 5-MOP (15 mg each) were ingested separately in a 50% alcoholic solution. A strong and persistent erythema was induced in three out of the four subjects (UVA dose: 3-25 J/cm2). The blood levels immediately before UVA irradiation varied between 14 and 114 ng/ml (8-MOP), and 17 and 70 ng/ml (5-MOP). In the subject who did not show phototoxicity, the blood levels remained at trace levels (3 ng/ml). Two subjects were also tested with a mixture of 10 mg 8-MOP plus 10 mg 5-MOP; in one subject the mixture induced pigmentation only, while in the other a mild-to-medium erythema was induced. The blood levels of the furocoumarins in the two subjects were similar (12-15 ng/ml for 8-MOP and 5-MOP). It is concluded that in humans the phototoxic threshold dose of furocoumarin mixtures is of the order of 10 mg 8-MOP plus 10 mg 5-MOP, which is equivalent to about 15 mg 8-MOP per person (blood levelsof 8-MOP and 5-MOP at 30 min after ingestion of about 10-15 ng/ml each). This phototoxic threshold dose was not reached by the consumption of celery roots and other conventional vegetables under normal dietary habits (experimental intake of 2-8 mg per subject of the potentially phototoxic furocoumarin mixture). However, the safety factor between the possible actual intake of furocoumarins and the phototoxic threshold dose is about 2-10, which is relatively small.
INTRODUCTION Although there has been less public debate about potentially toxic chemical compounds that occur naturally in foods in comparison with man-made chemicals in the diet, this topic is of considerable importance. An example of such natural substances are the furocoumarins (secondary plant metabolites) with known phototoxic properties. They occur in Rutaceae and Umbelliferae, but also occur widely in members of other plant families, including some edible species such as celery, parsley, parsnips, citrus fruits, figs and certain spices. The major source of dietary furocoumarins is celery (see below). About 150 furocoumarins are known, some of which show a marked phototoxic activity in humans. It has long been known that, upon skin contact, the furocoumarins in celery are capable of inducing local phototoxic lesions in celery harvesters (Garfield, 1985; Scheel et al., 1963) as well as in grocery-store
tTo whom all correspondence should be addressed at: Federal Office of Public Health, Division of Food Science, c/o Institute of Toxicology, Swiss Federal Institute of Technology and University of Ziirich, Schorenstrasse 16, CH-8603 Schwerzenbach, Switzerland. Abbreviations: 5-MOP = 5-methoxypsoralen; 8-MOP = 8methoxypsoralen.
workers (Seligman et al., 1987). A topical or systemic furocoumarin exposure combined with long-wave UV radiation (UVA; wavelength 320--400 nm) of the skin typically leads to an erythematous, sunburn-like response and, in severe cases, to blistering. The onset of this toxic reaction is typically delayed for 12-36 hr after exposure, and the toxic reaction is followed within 1-2 wk by hyperpigmentation at the sites of exposure. Pigmentation only is observed at exposure levels near a phototoxic threshold dose. Furocoumarins are activated by UVA radiation and are able thereafter to bind to macromolecules (e.g. membranes, DNA), leading to cytotoxic (e.g. erythema) and genotoxic responses (Grossweiner, 1984; Yang et al., 1987 and 1988). Wavelengths between 350 and 365 nm (UVA) in the presence of furocoumarins induce the maximal phototoxic skin reactions in humans (Pathak, 1984; Scott et al., 1976). The main source of information about the phototoxicity of furocoumarins is the fact that some derivatives, particularly 8-methoxypsoralen (8-MOP) and to a lesser extent 5-methoxypsoralen (5-MOP), used as therapeutic agents in combination with UVA irradiation in the treatment of psoriasis and vitiligo (PUVA therapy). Typical doses of 0.4--0.7 mg 8-MOP/kg body weight are administered orally, and followed 2 hr later with UVA exposure (dose: 0.5-7 J/cm2). In 1984, the NCI published an extensive
523
J. SCHLATTERet al.
524
discussion of the photobiological, toxicological and pharmacological aspects of PUVA therapy, and a further review dealt with the molecular basis of furocoumarin reactions (Scott et al., 1976). Schlatter (1988) reviewed the data on the toxicity of furocoumarins with respect to the significance of a total dietary intake of a few mg per day, as well as the data on natural UV exposure; he concluded that the estimated natural UVA dose received by inhabitants of Switzerland would be high enough for phototoxic skin reactions in the presence of furocoumarins, even in non-alpine areas and during summer and winter-a period of 5-30 min outdoors between 10 am and 2 pm might be sufficient to reach the minimal phototoxic UVA dose. Monographs on carcinogenicity are available for 8-MOP (IARC, 1980 and 1982) and 5-MOP (IARC, 1986a), and for UV radiation (Bundesgesundheitsblatt, 1987; IARC, 1986b). IARC concluded that there is sufficient evidence for the carcinogenicity of PUVA, and the risk-benefit of a PUVA therapy has been repeatedly discussed (Epstein, 1979; Farber et al., 1983; Gibbs et al., 1986; Henseler and Christopher, 1984; Stern, 1984; Stern and Momtaz, 1984). Although normal levels of furocoumarins in vegetables are low (Beier et al., 1983; Ceska et al., 1986a and 1987), their concentrations can increase greatly following fungai invasion or other stress situations, such as exposure to UV light, cold or chemicals (Beier and Oertly, 1983; Ceska et al., 1986a,b; Chaudhary et al., 1985). Their induction by fungi and their in vitro toxicity to fungi are evidence that furocoumarins may act as antimicrobial defence compounds (i.e. phytoalexins) in some plants (Desjardins et al., 1989). In infected celery or carrots, furocoumarin concentrations greater than 10 times higher than normal have been measured (Ceska et al., 1986a; Chaudhary et al., 1985; Surico et al., 1987). In a recent Swiss survey (Baumann et al., 1988), the highest concentrations were found in celery in the edible part of the roots (up to 25/zg/g), in parsnips (40/ag/g) and in parsley (15/ag/g), while no furocoumarins were detectable in fennels, carrots or figs (detection limit 0.05 #g/g each). Apparently healthy celery-root samples had markedly higher concentrations of phototoxic furocoumarins than those given in the literature for non-infected samples (0.6--l.8/ag/g; Ashwood-Smith et al., 1985; Beier et al., 1983). The reason for this difference is not known, but could be due to the cultivation of different celery varieties. Furocoumarins are fairly thermostable and cooking will not significantly reduce their amounts. Hence, it has been concluded that in
some cases the individual furocoumarin intake may reach several mg per day, which approaches the doses used in photochemotherapy of psoriasis (AshwoodSmith et al., 1985; Chaudhary el aL, 1985; Schlatter, 1988). However, the threshold doses with respect to phototoxicity of mixtures of furocoumarins (both phototoxic and non-phototoxic) are not known at present. It was the aim of the present study to investigate the possibility of phototoxic skin reactions due to the consumption of mixtures of furoeoumarins as they occur usually in celery roots. MATERIALS
AND METHODS
Chemicals. Psoralen (P 8399; Sigma Chemical Co., St Louis, MO, USA), 5-MOP (27, 52-7; Aldrich Chemical Co., Milwaukee, WI, USA), 8-MOP M3501; Sigma) and 5,8-dimethoxypsoralen (2-8854; Carl Roth GmbH and Co., Karlsruhe, Germany), as well as all solvents used, were of analytical grade and used without further purification. Subjects. Four healthy male volunteers (subjects A-D), all academic personnel from the University, agreed to participate both in the experiment with celery roots and in the experiment with 8-MOP plus 5-MOP mixtures. Their respective body weights are listed in Table I. All subjects were classified as belonging to the same skin type (class III, according to Greiter, 1984: fair skin, no freckles, brown hair and grey or brown eyes). For this skin type the threshold dose for UVA alone (32ff-400nm) for pigmentation is expected to be 17-20 J/cm 2, for UVB (280-320nm) 0.021J/cm 2, and for erythema formation 25-35J/cm 2 (UVB 0.036J/cm2; Bundesgesundheitsblatt, 1987). In subjects A, C and D the threshold UVA dose for skin pigmentation was verified to be 1>20 J/cm 2. Experimental procedure. During autumn to spring 1988-1989, 70 samples of celery roots (Apium graveolens L.) were bought at retail stores in Bern and analysed for furocoumarin content (for analytical details see Baumann et al., 1988). For phototoxicity testing, 2340 g of celery roots (eight items) was obtained from a local shop in Schwerzenbach and stored for 2 months (4°C). On the day of the experiment, the celery roots were washed, shredded and mixed. An aliquot of 50 g was frozen for subsequent furocoumarin analysis. In experiments where pure chemicals were used, 8-MOP and 5-MOP were dissolved in 15 ml ethanol and diluted with 15 ml water immediately before consumption. UVA irradiation was performed at the dermatology department of the University Hospital Ziirich, using a Waldmann-800 UVA lamp (verified emission of 7 x 10 -3 W/cm 2 and
Table 1. Skin reactions of four human subjectsafter oral administrationof a mixture of 8-MOP plus 5-MOP Skin reaction after UVA dose (J/cm 2) of Subject 8-MOP + (body weight, kg) 5-MOP dose (mg) Day of maximalresponse 3 6 9 12 18 25 A (65) 10 + 10 7 -El El El E2 E2 15 + 15 7 El E2 E2 E3 E3 E3 B (63) 10 + 10 2 PI P2 P3 P3 P3 P3 15 + 15 4 -El E2 El E3b E3b c (85) 15 + 15 . . . . . . . D(74) 15 + 15 8 -El E2 E2 E3 E3 b = blistering
E = erythema
P = pigmentation 1, 2, 3 = slight, medium or strong reaction, respectively
Phototoxicity of furocoumarins in humans
525
UVB emission of 1 x 10-6W/cm 2 at a distance of (pglg) 20cm). Six non-pigmented circular skin areas of a0 O 2.5 cm diameter, in the gluteal region (lower back), were exposed to six UVA doses (maximal UVB dose O O O was 0.00006 J/cm 2, which is far below the threshold dose for pigmentation--see above). In the first experz0 O O iment, all subjects ingested 300 g of celery roots after a fasting period of at least 4 hr. Venous blood 8 8 o samples (10 ml, heparin glass vials; B-D Vacutainer) 0 O0 O÷ - C o c o were drawn immediately before ingestion of the 1o o oo~_ celery, and 2 and 4 hr thereafter. Blood samples were o a~o frozen and stored at - 2 0 ° C until chemical analysis. O O UVA irradiation was started 2 hr after ingestion of 0 i i i i i i ¢ ¢ i / i the celery (1.5, 2, 3, 5, 7 and 9 J/cm2). Skin readings were performed after 10 min, and after 4, 24, 48 and Fig. 1. Time-course of the content of phototoxic furocou72 hr. In the second experiment, all four subjects marins (sum of psoralens 8-MOP and 5-MOP) in celery-root ingested a mixture of 15mg 8-MOP plus 15rag items (ready-made tubers) from autumn 1988 to spring 1989. Arithmetic mean (--). 5-MOP in 30 ml 50% ethanol, after a fasting period of at least 4 hr. Venous blood samples were drawn after 30 and 90 min, and the UVA irradiation was by small superficial spots. The concentrations of started 30 min after ingestion of the mixture (UVA psoralen were generally the lowest (median 0.26, dose: 3, 6, 9, 12, 18 and 25 J/cm2). Skin readings were mean 1.07, range <0.1 to 10.29#g/g). The concenperformed after 10 min, and after 4, 24, 48 and 72 hr. trations of 8-MOP and 5-MOP varied widely but In cases where the intensity of the erythema was still tended to be about equal (8-MOP: median 2.28, mean increasing, daily readings were continued until the 3.81, range 0.43 to 21.62 #g/g; 5-MOP: median 1.92, maximal response was reached. In subjects A and B, mean 2.46, range 0.65 to 12.53 gg/g). Some nona dose of 10 mg 8-MOP plus 10 mg 5-MOP was also phototoxic 5,8-dimethoxypsoralen (isopimpinellin) tested under the same experimental conditions. The was also present in the samples, but not quantified time interval between the different doses and exper- because of interference. It is obvious from Fig. 1 that iments was at least I wk. For a comparison of the the mean concentrations of the phototoxic furofurocoumarin blood levels after ingestion of an alco- coumarins (mainly 5-MOP and 8-MOP) did not holic solution with those after ingesting furocou- , increase in the time between harvesting (August) and matins in celery, subjects A and B ingested mixtures the date of sampling, as had been expected at first. of 8-MOP plus 5-MOP at two dose levels (5 mg and Instead, the mean concentrations remained relatively 7.5mg each) in 30ml 50% ethanol without sub- constant throughout the whole season (median sequent UVA irradiation. Blood samples were taken 4.6#g/g, arithmetic mean 7.3#g/g, range 1.1 to at 30 and 90 min after ingestion. The 'non-responder' 28.3/~g/g; Figs 1 and 2). of experiment 2, subject C, ingested a mixture of 15mg 8-MOP plus 15mg 5-MOP in 40ml 50% Blood analysis of furocoumarins ethanol for a second time without subsequent UVA Addition of psoralen, 5-MOP, 8-MOP and 5,8irradiation. Blood samples (5 ml) were taken 15, 30, dimethoxypsoralen to bovine blood (20 ng/g each) 60, 90, 120 and 180 min after ingestion of the mixture resulted in mean recoveries ( + S D , n = 3) of 91 + 3, through a venous catheter, using disposable syringes 93 + 3, 95 + 2 and 91 + 6%, respectively. Therelative (Omnifix, Braun Melsungen AG, Germany). error of a single measurement can be assumed to be Analyticalprocedure. The procedure developed for of the order of 10%. Figure 3 presents some typical the analysis of furocoumarins in vegetables, as de- chromatograms of the reagent blank, standard solscribed previously (Baumann et al., 1988), has been ution and human plasma, after ingestion of furocouadapted for blood. Instead of the addition of water marins. The detection limits for psoralen, 5-MOP and to the sample, physiological sodium chloride solution 8-MOP in blood were in the range of 0.5 to 2 ng/ml was used. The final extract was concentrated to 50 #1 instead of 5 ml, and 20 #! was injected into the HPLC (,) system (plus UV-detector). All the precautions ]6.r usually followed in trace analysis of organic corn]. pounds, such as blanks (i.e. of reagents and syringes and tubes used for drawing blood), were strictly ~2 observed. RESULTS
Occurrence of furocoumarins in ready-made celery roots
In Fig. 1, the analytical result of 70 samples of apparently healthy celery roots (fresh material), offered at the retail level, were plotted against the month of sampling. Probable fungal infection of the celery roots in about 7% of the items was indicated
'hllllflt_.
0
0
S
10
lS
20
25
(pg/g)
Fig. 2. Frequency distribution of the concentrations of total phototoxic furocoumarins in celery from autumn 1988 to spring 1989.
J. SCHLATTERet al.
526
I
I
0
5
I
10
I'
I
I
I
15
0
5
10
m
I
l
I
s
15
0
5
10
15
mln
m|n
mln
Fig. 3. Chromatograms of: (A) reagent blank; (B) standard solution (34 ng 8-MOP plus 24 ng 5-MOP, injected); (C) blood plasma of subject B, 90 min after ingestion of 10mg 8-MOP plus 10 mg 5-MOP (5.8 ng/ml 8-MOP plus 12.5 ng/ml 5-MOP). each, depending on the chromatography system and the respective sensitivity of the detector.
Phototoxicity after ingestion of celery roots The celery roots consumed by the four subjects revealed a total furocoumarin content of 28.2 pg/g (psoralen, 2.3 pg/g; 8-MOP, 3.8 yg/g; 5-MOP, 2 2 . 1 # g / g fresh material; 5,8-dimethoxypsoralen, < 0 . 0 5 y g / g ) and the total furocoumarin dose ingested was, therefore, 8.4 mg/subject (0.100.13 mg/kg body weight). Throughout the whole observation period after U V A irradiation (up to 9 J / c m 2) in all subjects, there was no observable pigmentation or erythema formation. The furocoumarin blood levels were below the analytical detection limit ( < 2 ng/ml each).
Phototoxicity after ingestion of mixtures of 8,MOP plus 5-MOP in ethanol The results of the second experiment are summarized in Tables 1 and 2. The ingestion of a mixture of 15 mg 8-MOP plus 15 mg 5-MOP clearly induced skin erythemas in subjects A, B and D, while neither erythema formation nor pigmentation was observed in subject C (Table 1). In subject A, the response became visible after 24 hr with t> 9 J U V A / c m 2. The intensity of the erythema was UVA-dose-dependent, and the subject complained of strong itching at the skin areas of higher U V A doses. In subjects B and D,
the first effects were already visible 4 hr after treatment with a U V A dose of >_-9 (subject B) or >t 18 J/cm 2 (subject D). In subject D, cortisone was topically applied to the skin areas of the higher U V A doses because of strong discomfort. The blood levels of 8-MOP and 5-MOP immediately before and after U V A irradiation showed individual patterns (Table 2). The blood levels were highest in subject A, followed by subjects B and D. In subject C, the blood levels of 8-MOP and 5-MOP remained very low, just slightly above the analytical detection limit. The mixture of 10 mg 8-MOP plus 10 mg 5-MOP ingested by subject B induced pigmentation after 2 4 h r at U V A doses of >19J/cm 2. N o erythema was induced and the blood levels of 8-MOP and 5-MOP were initially about 12 ng/ml each. In subject A, given 10mg each of 8-MOP and 5-MOP, a slight pigmentation of the skin areas exposed to the three highest U V A doses was seen after 4 and 24 hr, and this developed into an erythema within 48 hr. Blood levels of 8-MOP and 5-MOP were similar to those measured earlier in subject B. Also, similar blood levels in volunteer A compared with those of volunteer B were observed after the ingestion of a mixture of 5 mg 8-MOP plus 5 mg 5-MOP, as well as after the ingestion of a mixture of 7.5 mg 8-MOP plus 7.5 mg 5-MOP, both at 30 and 90 min.
Table 2. Blood levelsof 8-MOP and 5-MOP (ng/ml) after oral administration of a mixture of 8-MOP plus 5-MOP in 50oVoethanol solution Blood level of 5-MOP (ng/ml) Blood level of 8-MOP (ng/ml) Subject 8-MOP + (body weight, kg) 5-MOP dose (mg) 30 rain 90 min 30 min 90 min A (65) 5+ 5 3.7 <0.5 6.2 1.4 7.5 + 7.5 23.4 4.2 27.2 9.9 10+ 10 14.2 5.1 15.4 11.4 15 + 15 113.9 18.7 69.6 28.1 B (63) 5+ 5 3.7 <0.5 4.2 2.4 7.5 + 7.5 26.0 3.2 32.7 10.3 (PI 10.6)* (5.8) 12.0 (12.9) 9.1 (12.5) 10 + 10 11.9 (EC 13.4) 4.6 (3.3) (11.0) (5.2) 15 + 15 64.7 24.1 50.4 64.0 ( ~ 3.0) ~ 3.0) ,,~3.0) ~2.5 ((~2.0) ~2.5 ((~2.0) C (85) 15 + 15 ~3.0 (Pl (EC~~3.0) 3.0) ~,2.5 (~2.0) D(74)
15+15
(PI 20.5) 13.7(EC6.1 )
(14.6) 12.2 (9.4)
(24.7) 16.8 ( 7 . 8 )
(21.3) 16"5(11.1)
*In several experiments, plasma (PI) and erythrocyte concentrations (EC) were determined separately (values given in parentheses). The concentration in whole blood was then calculated, assuming 53% plasma and 47% erythrocytes.
Phototoxicity of furocoumarins in humans Table 3. Time-courseof blood levelsof 8-MOP and 5-MOP (ng/ml) in subjectC after oral administrationof a mixture of 15mg 8-MOP plus 15mg 5-MOP in 50*/, ethanol Time (rain) Furocoumarin 8-MOP* 5-MOP
15 <3 <2
30 60 90 120 180 <3 ~<3 10.2 5.6 -~<2 2.7 13.4 11.4 <2 = not measured *The detectionlimitwas about 3 ng/ml becauseof an eluantchange in HPLC, which resulted in higher retention time of 8-MOP. This change was necessaryin order to separate8-MOP from an interfering substanceoriginatingfrom the piston of the disposable syringe,type Omnifix(see Experimentalprocedure). -
-
In subject C, the possibility of a different kinetic or metabolic behaviour was tested after an oral dose of 15 mg 8-MOP plus 15 mg 5-MOP by monitoring blood levels over a longer period. The data (Table 3) again showed low blood levels during the first hour after ingestion of a mixture of 15 mg 8-MOP plus 15rag 5-MOP. A slight increase in 8-MOP and 5-MOP blood levels was observed after 90 min, and declined again thereafter. DISCUSSION
Blood analysis o f furocoumarins Most of the published analytical methods for plasma or serum use reverse-phase HPLC with UV detection (Gasparro et al., 1988; Herfst et al., 1980; Kornhauser et al., 1984; Monbaliu et al., 1981; Stolk et al., 1987; Susanto et aL, 1986) or fluorescence detection (Prognon et aL, 1983). Gas chromatography has also been used, and electron capture detection results in an increased sensitivity and selectivity compared with flame ionization detection (Ehrsson et al., 1977; Schmid and Koss, 1978). Although different authors (e.g. Carter and Goldstein, 1984; Stolk et al., 1987) have described analytical problems such as loss of analyte by adherence to container walls or by volatilization during the procedure (i.e. concentration steps), the recoveries obtained in the present work were acceptable in spite of the fact that the procedure includes several evaporation steps. Phototoxicity after ingestion o f celery roots The metabolism of furocoumarins (Bickers and Pathak, 1984; De Wolff and Thomas, 1986; Rogers et al., 1983) shows a saturable first-pass effect in the liver (Schmid et al., 1980). Only oral doses that significantly elevate plasma levels lead to phototoxic effects (Brickl et al., 1984a,b). At these plasma levels, furocoumarins are found in skin suction blister fluid (Herfst and De Wolff, 1982; Kornhauser et al., 1984; Korting et al., 1982). In our experiment with the ingestion of 300 g of celery roots (roughly double or triple the amount of a normal portion, per capita consumption 4 g)--equivalent to about 10mg of phototoxic furocoumarins--the blood levels remained below the analytical detection limit, which explains the absence of phototoxic skin reactions. However, a similar amount of total furocoumarins (5 mg each of 8-MOP and 5-MOP) ingested in an alcoholic solution led to measurable blood levels. Therefore, almost all furo-
527
coumarins must have been cleared from the blood during the first liver passage after the ingestion of celery roots, probably because of a protracted uptake of the furocoumarins from the gastro-intestinal tract in comparison with the ingestion as an alcoholic solution, or because of a low bioavailability of the furocoumarins in celery. It is known from PUVA therapy that concomitant food intake can influence 8-MOP and S-MOP blood levels (Herfst and De Wolff, 1982 and 1983; Mosher et aL, 1983), as can both the galenic form of the furocoumarin application (Siddiqui et aL, 1984; Stolk et al., 1985) and the mode of administration (Kornhauser et aL, 1984). Phototoxicity after ingestion o f mixtures o f 8 - M O P plus 5 - M O P in ethanol To eliminate interfering factors with furocoumarin uptake and to determine the safety factor between dietary furocoumarin uptake and the phototoxic threshold dose, we tried to maximize the uptake by using ingestion of an alcoholic solution of a mixture of the pure substances of the main components in celery roots, namely 8-MOP and 5-MOP, after a fasting period. 10, 20 or 40rag 8-MOP per person gave rise to peak plasma levels of 8-MOP of about 12, 180 and 550ng/ml after 30rain (Schmid et aL, 1980). A minimal 8-MOP serum concentration of 10-25ng/ml combined with UVA doses of 10-20J/cm 2 is phototoxic (Ljunggren et aL, 1981). Phototoxicity seems to be slightly delayed compared with plasma levels in the absorption phase, but it persists much longer than plasma levels. It has been recommended that subjects be irradiated at or shortly after peak plasma levels (Brickl et aL, 1984), and a threshold dose for erythema formation from orally administered 8-MOP alone plus UVA irradiation was calculated to be of the order of 14rag/person (0.24 mg/kg body weight). However, the potency of the different furocoumarin derivatives in inducing phototoxicity after oral administration varies (in PUVA therapy, approximately a two-fold dose of 5-MOP is required to induce the same effect as 8-MOP), and there is some evidence that the different derivatives might act multiplicatively with respect to phototoxicity; if 5-MOP is administered 30 rain before 8-MOP (Bricki et aL, 1984a,b) plasma levels of 8-MOP are increased, and the combination is much more phototoxic than the individual compounds at the same dose. Therefore, we gave a dose of 10 mg 8-MOP plus 10 rag S-MOP to volunteer B. His blood level of 8-MOP was in good agreement with the published data. However, the skin response was clearly only pigmentation and not erythema, even at the highest UVA dose of 25 J/cm 2. This was the first indication that the simultaneous administration of the two furocoumarins act additively rather than multiplicatively as was assumed by Schlatter (1988), based on the experiments of Brickl et al. (1984a,b). Therefore, the 8-MOP plus 5-MOP dose was increased to 15 mg each for further tests. This produced a strong phototoxic skin reaction in all but one subject, and the measured blood levels in the responding subjects were in agreement with the published data. Again, the simultaneous ingestion of the two furocoumarins produced only an additive response. The interindividual differences in blood levels of
528
J. SCHLATTERet al.
the three responding subjects may be due to differences in pharmacokinetics but, interestingly, the phototoxicity in these three subjects was similar. This is in agreement with the observation of De Bersaques et al. (1984); the rate of furocoumarin absorption does not seem to influence the phototoxic skin reaction much, once a critical blood level is exceeded. The dose mixture of 10mg 8-MOP plus 10mg 5-MOP induced clear erythema in subject A, but in subject B only pigmentation was observed, although the furocoumarin blood levels of subject A were only marginally higher. Thus, this dose mixture is very close to the phototoxic threshold dose. In subject C, the blood levels of 8-MOP and 5-MOP were much lower than in the other subjects. In addition, the peak blood levels were delayed to approximately 90 min. The reason for this unexpectedly low and late rise in blood furocoumarin levels is not known. However, it explains the absence of phototoxicity in this subject upon UVA exposure between 30 and 90 min after ingesting the mixture. This demonstrates again the sometimes unpredictable pharmacokinetic behaviour of the furocoumarins as is known from PUVA therapy (Carter and Goldstein, 1984; De Wolff and Thomas, 1986), and that the intraindividual variation is generally low (factor 1.3-5.3) compared with the interindividual variation (Schfifer and Korting, 1982). The apparently higher initial blood furocoumarin level in subjects A and B (at 30 min, during the expected steep rise of blood levels, but not at 90 min) after ingestion of the 7.5 mg dose mixture compared with the 10 mg dose mixture may also be due to such intraindividual variations. From all these data, we conclude that in humans the approximate value of a phototoxic threshold dose of furocoumarin mixtures is of the order of 10 mg 8-MOP plus 10mg 5-MOP--equivalent to about 15 mg 8-MOP per person (blood levels of 8-MOP and 5-MOP at 30min after ingestion of about 10-15 ng/ml each)--which is in good agreement with the calculated threshold dose for 8-MOP alone of 14 mg/person (Brickl et al., 1984a) when taking into account the weaker potency of orally administered 5-MOP as described earlier. This phototoxic threshold dose is not reached by the consumption of celery roots at the Swiss retail level (maximum 200 g, maximal concentration of the sum of phototoxic furocoumarins 30-50#g/g). However, the safety factor between the possible actual intake of furocoumarins and the phototoxic threshold dose is relatively small, being of the order of about 2-10. In this light, the only recently published case report of a severe phototoxic burn after ingestion of unusually large amounts of celery with an exceptionally high furocoumarin level (estimated total furocoumarin concentration equivalent to 100#g 8-MOP/g) is not surprising (Ljunggren, 1990): a 65-year-old vegetarian (skin type III) ingested about 450 g celery one hour before a visit to a suntan parlour (estimated UVA dose about 20J/cm2), and suffered 48 hours later from a generalized, intense erythema with multiple large blisters, oedema and fever. We are aware of all the limitations of our experimental conditions (e.g. limited number of volunteers of one skin type, fixed furocoumarin mixture tested, fixed time of UVA irradiation, no separate phototox-
icity testing of the single furocoumarin components, no continuous blood level monitoring), but because of ethical reasons experimentation with a human carcinogen has to be kept to a minimum. Nevertheless, the risk assessment presented above, with all its inherent uncertainties, allows the following conclusions: (i) under normal dietary habits there is presently no risk of phototoxic burn due to the ingestion of celery and other conventional vegetables; (ii) extreme habitual consumption must be avoided, at least when subsequent UVA exposure can not be excluded; and (iii) when breeding attempts are made to improve the quality, resistance against diseases or other properties of the celery plant, a possible increase in the furocoumarin content must be avoided. REFERENCES
Ashwood-Smith M. J., Ceska O. and Chaudhary S. K. (1985) Mechanism of photosensitivity reactions to diseased celery. British Medical Journal 290, 1249 Baumann U., Dick R. and Zimmerli B. (1988) Orientierende Untersuchung zum Vorkommen yon Furocoumarinen in pflanzlichen Lebensmitteln und Kosmetika. Mitteilungen aus dem Gebiet der Lebensmitteluntersuchung und Hygiene 79, 112-129. Beier R. C., Ivie G. W., Oertli E. H. and Holt D. L. (1983) HPLC analysis of linear furocoumarins (psoralens) in healthy celery (Apium graveolens). Food and Chemical Toxicology 21, 163-165. Beier R. C. and Oertli E. H. (1983) Psoralens and other linear furocoumarins as phytoalexins in celery. Phytochemistry 22, 2595-2597. Bickers D. R. and Pathack M. A. (1984) Psoralen pharmacology, studies on metabolism and enzyme induction. National Cancer Institute Monographs 66, 77-84. Brickl R., Schmid J. and Koss F. W. (1984a) Pharmacokinetics and pharmacodynamics of psoralens after oral administration: considerations and conclusions. National Cancer Institute Monographs 66, 63-67. Brickl R., Schmid J. and Koss F. W. (1984b) Clinical pharmacology of oral psoralen drugs. Photodermatology 1, 174-186. Bundesgesundheitsblatt (1987) Empfehlungen zur Begrenzung gesundheitlicherStrahlenrisiken bei der Anwendung von Solarien und Heimsonnen. Bundesgesundheitsblatt 30, 19-30. Carter D. M. and Goldstein D. P. (1984) Quantitative tests for psoralens in the blood; methods and uses in monitoring psoralen and longwave radiation therapy. National Cancer Institute Monographs 66, 69-72. Ceska O., Chaudhary S. K., Warrington P. J. and AshwoodSmith M. J. (1986a) Furocourmarins in the cultivated carrot, Daucus carota. Phytochemistry 25, 81-83. Ceska O., Chaudhary S. K., Warrington P. J. and Ashwood M. J. (1987) Photoactive furocoumarins in fruits of some umbellifers. Phytochemistry 26, 165-169. Ceska O., Chaudhary S, K., Warrington P. J., Poulton G. and Ashwood-Smith M. J. (1986b) Naturally-occurring crystals of photocarcinogenic furocoumarins on the surface of parsnip roots sold as food. Experientia 42, 1302-1304. Chaudhary S. K., Ceska O., Warrington P. J. and AshwoodSmith M. J. (1985) Increased furocoumarin content of celery during storage. Journal of Agricultural and Food Chemistry 33, 1153-1157. De Bersaques J., Monbaliu J., Belpaire F. and Bogaert M. (1984) 8-Methoxy-psoralen plasma levels and phototoxic effect. Acta Dermato-Venereologica, Stockholm Suppl. 113, 166-167.
Phototoxicity of furocoumarins in humans Desjardins A. E., Spencer G. F. and Plattner R. D. (1989) Tolerance and metabolism of furocoumarins by the phytopathogenic fungus Gibberella pulicaris (Fusarium sambicinum ). Phytochemistry 28, 2963-2969. De Wolff F. A. and Thomas T. V. (1986) Clinical pharmacokinetics of methoxalen and other psoralens. Clinical Pharmacokinetics 11, 62-75. Ehrsson H., Eksborg S., Wallin I., Kallberg N. and Swanbeck G. (I 977) Determination of 8-methoxypsoralen in plasma by electron capture gas chromatography. Journal of Chromatography 140, 157-164. Epstein J. H. (1979) Risk and benefits of the treatment of psoriasis. New England Journal of Medicine 300, 852-853. Farber E. M., Abel E. A. and Cox A. J. (1983) Long-term risks of psoralen and UV-A therapy for psoriasis. Archives of Dermatology 119, 426-431. Garfield E. (1985) Current comments. From tonic to psoriasis: stalking celery's secrets. Current Contents 18, 3-12. Gasparro F. P., Battista J., Song J. and Edelson R. L. (1988) Rapid and sensitive analysis of 8-methoxypsoralen in plasma. Journal of lnvestigative Dermatology 90, 234-236. Gibbs N. K., H6nigsmann H. and Young A. R. (1986) PUVA treatment strategies and cancer risk. Lancet i, 150-151. Greiter F. (1984) Sonne und Gesundheit. Gustav Fischer Verlag, Stuttgart. Grossweiner L. I. (1984) Mechanisms of photosensitization by furocoumarins. National Cancer Institute Monographs 66, 47-54. Henseler T. and Christopher E. (1984) Risk of skin tumors in psoralen- and ultraviolet A-treated patients. National Cancer Institute Monographs 66, 217-219. Herfst M. J. and De Wolff F. A. (1982) Influence of food on the kinetics of 8-methoxypsoralen in serum and suction blister fluid in psoriatic patients. European Journal of Clinical Pharmacology 23, 75-80. Herfst M. J. and De Wolff F. A. (1983) Intraindividual and interindividual variability in 8-methoxypsoralen kinetics and effect in psoriatic patients. Clinical Pharmacology and Therapeutics 34, 117-124. Herfst M. J., Edelbroek P. M. and De Wolff F. A. (1980) Determination of 8-methoxypsoralen in suction-blister fluid and serum by liquid chromatography. Clinical Chemistry 26, 1825-1828. IARC Working Party (1980) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Some Pharmaceutical Drugs. Vol. 24, Methoxsalen. pp. 101-124. International Agency for Research on Cancer, Lyon. IARC Working Party (1982) 1ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Chemicals, Industrial Processes and Industries Associated with Cancer in Humans. Suppl. 4. Methoxsalen with Ultra-violet A Therapy (PUVA) (group 1). pp. 158-160. International Agency for Research on Cancer, Lyon, IARC Working Party (1986a) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Some Naturally Occurring and Synthetic Food Components, Furocourmarins and Ultraviolet Radiation. Vol. 40. 5-Methoxypsoralen. pp. 327-347. International Agency for Research on Cancer, Lyon. IARC Working Party (1986b) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Some Naturally Occurring and Synthetic Food Components, Furocoumarins and Ultraviolet Radiation. Vol. 40. Appendix: Ultraviolet Radiation. pp. 379-404. International Agency for Research on Cancer, Lyon. Kornhauser A., Wamer W. G. and Giles A. L., Jr (1984) Difference in topical and systemic reactivity of pseralens: determinations of epidermal and serum levels. National Cancer Institute Monographs 66, 97-101.
529
Korting H. Ch., Sch/ifer M., Roser E. and Mutschler E. (1982) Determination of 8-methoxypsoralen levels in plasma and skin suction blister fluid by a new sensitive fluorodensitometric method. Archives of Dermatological Research 272, 9-20. Ljunggren B. (1990) Severe phototoxic burn following celery ingestion. Archives of Dermatology 126, 1334--1336. Ljunggren B., Bjellerup M. and Carter M. (1981) Dose-response relations in phototoxicity due to 8methoxypsoralen and UV-A in man. Journal of Investigative Dermatology 76, 73-75. Monbaliu J. G., Rosseel M. T. and Bogaert M. G. (1981) Analysis of 8-methoxypsoralen in plasma by reversedphase high-performance liquid chromatography. Journal of Pharmaceutical Sciences 70, 965-966. Mosher D., Momtaz K., Borowska Z., Caldwell D. and Carter D, M. (1983) Bioavailability of 8-methoxypsoralen: another determinant. Clinical Research 31, 590A. NCI (1984) Photobiologic, toxicologic, and pharmacologic aspects of psoralens, National Cancer Institute Monograph 66, NIH publication no. 84-2692. Pathak M. A. (1984) Mechanisms of psoralen photosensitization reactions. National Cancer Institute Monographs 66, 41-46. Prognon P. M., Simon G. and Mahuzier G. (1983) Dosage du mrthoxy-5-psoralrne dans le plasma par chromatographie en phase liquide et drtection spectrofluorimrtrique. Journal of Chromatography 272, 193-199. Rogers S. L., Staubus A. E., Bianchine J. R. and Gerber N. (1983) Michaelis-Menten pharmacokinetics of 8methoxypsoralen. Federation Proceedings. Federation of the American Societies for Experimental Biology 42, 1139. Sch/ifer M. and Korting H. Ch. (1982) Intraindividual variations of 8-methoxypsoralen plasma levels. Archives of Dermatological Research 272, 1-7. Scheel L. D., Perone V. B., Larkin R. L. and Kupel R. E. (1963) The isolation and characterization of two phototoxic furanocoumarins (psoralens) from diseased celery. Biochemistry, New York 2, 1127-1131. Schlatter J. (1988) Die toxikologische Bedeutung von Furocoumarinen in pflanzlichen Lebensmitteln. Mitteilungen aus dem Gebiet der Lebersmitteluntersuchung und Hygiene 79, 130-143. Schmid J. and Koss F. W. (1978) Rapid, sensitive gas chromatographic analysis of 8-methoxypsoralen in human plasma. Journal of Chromatography 146, 498-502. Schmid J., Prox A., Zipp H. and Koss F. W. (1980) The use of stable isotopes to prove the saturable first-pass effect of methoxalen. Biomedical Mass Spectrometry 7, 560-564. Scott B. R., Pathak M. A. and Mohn G. R. (1976) Molecular and genetic basis of furocoumarin reactions. Mutation Research 39, 29-74. Seligman P. J., Mathias T. C. G., O'Malley M. A., Beier R. C., Fehrs L. J., Serrill W. S. and Halperin W. E. (1987) Phytophotodermatitis from celery among grocery store workers. Archives of Dermatology 123, 1478-1482. Siddiqui A. H., Stolk L. M. L. and Cormane R. H. (1984) Comparison of serum levels and clinical results of PUVA therapy with three different dosage forms of 8-methoxypsoralen. Archives of Dermatological Research 276, 343-345. Stern R. S. (1984) Carcinogenic risk of psoralen plus ultraviolet radiation therapy: evidence in humans. National Cancer Institute Monographs 66, 211-216. Stern R. S. and Momtaz K. (1984) Skin typing for assessment of skin cancer risk and acute response to UV-B and oral methoxsalen photochemotherapy. Archives oJ Dermatology 120, 869-873. Stolk L. M. L., De Ruiter R., Saadawi A., Siddiqui A. H. and Cormane R. H. (1987) Determination of psoralen
530
J. SCHLATTERet al.
in serum by reversed-phase high-performance liquid chromatography. Journal of Chromatography 423, 383-386. Stolk L. M. L., Siddiqui A. H., Westerhof W. and Cormane R. H. (1985) Comparison of bioavailability and phototoxicity of two oral preparations of 5-methoxypsoralen. British Journal of Dermatology 112, 469--473. Surico G., Varvaro L. and Solfrizzo M. (1987) Linear furocoumarin accumulation in celery plants infected with Erwinia carotovora pv. carotovora. Journal of Agricultural and Food Chemistry 35, 406--409.
Susanto F., Humfeld S. and Reinauer H. (1986) Highperformance liquid chromatographic measurement of 8-methoxypsoralen in plasma. Chromatographia 21, 443-446. Yang X. Y., DeLeo V. and Santell R. M. (1987) Immunological detection and visualization of 8-methoxypsoralenDNA photoadducts. Cancer Research 47, 2451-2455. Yang X. Y., Delohery T. and Santella R. M. (1988) Flow cytometric analysis of 8-methoxypsoralen-DNA photoadducts in human keratinocytes. Cancer Research 48, 7013-7017.