Effect on topical 5-methoxypsoralen on tumorigenesis induced in albino and pigmented hairless mouse skin by UV irradiation

Journal of Photochemistry and Photobiology, B : Biology, 5 (1990) 343 - 357

343

EFFECT ON TOPICAL 5-METHOXYPSORALEN ON TUMORIGENESIS INDUCED IN ALBINO AND PIGMENTED HAIRLESS MOUSE SKIN BY UV IRRADIATION V . E . REEVEt, G . E . GREENOAK, C . BOEHM-WILCOX, P . J . CANFIELD and C . H . GALLAGHER Department of Veterinary Pathology, University of Sydney, Sydney, N.S.W. 2006 (Australia) (Received March 10, 1989 ; accepted September 11, 1989)

Keywords. 5-Methoxypsoralen, hairless mouse, ultraviolet, phototumorigenesis, action spectrum, melanogenesis .

Summary A new line of the Skh :HRII hairless pigmented mouse (black juvenile coat) is described which has been selectively bred for the capacity to respond consistently to simulated solar UV radiation with a continuous and strong tan . This mouse demonstrates a degree of protection from chronic UVinduced tumorigenesis when compared with the Skh :HRI hairless albino mouse, and has been used here to study the effect of induced melanogenesis on phototumorigenesis . Mice were irradiated for 10 weeks with incremental doses of simulated solar UV radiation (UVA + B) from a fluorescent tube source which induced tumours in 100% of albino mice and 93% of black mice by 200 days (minimally oedemal), or with 60% of this dose (sub-oedemal) which induced tumours in 85% of albino mice and 65% of black mice . Mice were also exposed to the UVA component of these radiation sources, obtained by window glass filtration . The effect of topical 5-methoxypsoralen (5-MOP) was examined, at either 0 .003% with minimally oedemal UVA + B or its UVA component alone, or at 0 .01% with sub-oedemal UVA + B or its UVA component alone, in both albino and black mice . The 5-MOP concentrations were selected as the maximum concentration which did not increase the erythema and oedema responses after a single exposure to minimally oedemal or sub-oedemal UVA + B . At 200 days, the tumorigenic response to sub-oedemal UVA + B was significantly increased by topical 5-MOP, to 100% in albinos and 93% in black mice . In contrast, tumorigenesis in response to minimally oedemal UVA + B was unaffected by topical 5-MOP . The UVA component alone of either irradiation regime was not tumorigenic under these conditions . When *Author to whom correspondence should be addressed . 1011-1344/90/$3 .50

© Elsevier Sequoia/Printed in The Netherlands

344 combined with topical 5-MOP, the UVA of minimally oedemal UVA + B became moderately tumorigenic, and resulted in a tumour incidence of 23% in albinos and 14 .5% in black mice . However, the UVA component of suboedemal UVA + B, when combined with topical 5-MOP, was highly tumorigenic specifically in albino mice, inducing tumours in 93% of albino mice but in only 27% of black mice . Tan intensity resulting from minimally oedemal UVA + B was not enhanced by topical 5-MOP, and its UVA component combined with 5-MOP resulted in only a minimal tan . However, the tan intensity resulting from sub-oedemal UVA + B with topical 5-MOP was strongly increased, although its UVA component combined with 5-MOP did not produce a perceptible tan . Protection from tumorigenesis in black mice could not be ascribed to the induction of melanogenesis, since tan intensity, whether or not induced in the presence of 5-MOP, correlated with tumorigenesis . There was no evidence for protection from tumorigenesis by induced pigmentation .

1 . Introduction Topically applied psoralens such as 8- and 5-methoxypsoralen (8-MOP, 5-MOP) are UVA-activated photosensitizers, mutagens and carcinogens in experimental mice [1 - 3] . In addition, photoactivated psoralens induce a delayed hyperpigmentation of the skin, which has been utilized in the treatment of vitiligo [4] . The psoralen tan differs histologically from the UVA tan in human skin, being characterized by melanocyte abnormalities and a dispersed rather than aggregated distribution of enlarged single melanosomes [5, 6] . In the UVA tan, melanin is found mainly in the basal layer, but psoralens result in a diffusion of pigment throughout the epidermal layers including the stratum corneum [7 - 9], more characteristic of the UVB tan . The psoralen tan has been described as intense, diffuse and prolonged [10] . Because fair human skin normally incapable of achieving a sun-tan may respond to UV-activated psoralens with the cosmetically desirable pigmentation, 5-MOP has been incorporated into some UVB blocking sunscreen formulations, with the aim of stimulating melanogenesis by solar UVA in the absence of the erythemogenic UVB wave band . The protective properties of UVA- and UVB-induced pigmentation have been tested in several human systems . The UVB tan, but not the UVA tan, induced in fair-skinned individuals by moderate exposures, has been found to provide protection from a subsequent erythemal dose of UVB or solar UV [8, 11, 12] . However, a more intense UVA tan induced in capable tanners by exposures of at least 5 times greater has been shown to protect successfully from subsequent UVB erythema and sunburn cell formation [7] . Substantial protection from sunburn has also been reported for an 8MOP plus UVA tan [13 - 15] and a 5-MOP tan has been shown to reduce

345 DNA damage, measured as unscheduled DNA synthesis, from a challenge dose of UV [16] . The determination of the action spectra for the various photobiological activities of the psoralens, particularly the erythemogenic, hyperpigmentary and tumorigenic functions in mammalian skin, is a current issue, not only because of the inclusion of 5-MOP in some sunscreen preparations, but also because of the use of 8-MOP plus UVA irradiation (PUVA therapy) in the treatment of hyperproliferative skin disorders such as psoriasis . While the UV absorption maximum of 8-MOP and 5-MOP occurs at 300 run, the biological responses of erythema, sunburn cell formation, inhibition of DNA synthesis and phototumorigenesis in mice are reportedly at a maximum at 320 - 335 nm [17 - 21] . Baydoun and Young [22] have suggested that it is not the parent psoralen (which intercalates readily in the dark with DNA) which is the target molecule for photobiological activity, but rather its 4',5'-monoadduct to pyrimidine bases, which absorbs maximally at 320 - 335 nm . There has been some indication that the action spectrum for the hyperpigmentary effect also lies in the UVA range [23], but this has not been rigorously determined . It seems that the action spectra for the pigmentary and erythemogenic responses may be different, as these reactions have been shown to be independently induced by somes related psoralen compounds [24] . The aim of these experiments was to compare the phototumorigenic and melanogenic responses to simulated solar UV radiation or its UVA component only, both alone and in the presence of topical 5-MOP . Mice were chronically exposed to either minimally oedemal or sub-oedemal daily doses of simulated solar UV radiation (UVA + B), or the UVA component of each, which are two levels of sunlight exposure likely to be encountered daily by humans . The concentration of topical 5-MOP was adjusted so that a single combined treatment with UVA + B was not more phototoxic than a treatment in the absence of 5-MOP . This constitutes a model of UVB sunscreen-attenuated sunlight, with and without 5-MOP, as a potential tan enhancer . A new line of the Skh :HRII pigmented hairless mouse was used which has been selectively bred in our laboratory for the capacity to respond to simulated solar UV irradiation with a strong, diffuse and consistent tan . A comparison of the responses of this pigmented mouse with those of albino mice is therefore an indicator of the contribution of epidermal melanin .

2 . Materials and methods

2.1 .

Animals Hairless mice, which originated from the Skin Cancer and Photobiology Unit, Temple University, Philadelphia, U .S .A ., were bred to provide both inbred Skh :HRI albinos and partially inbred Skh :HRII pigmented mice . A new line of Skh :HRII mice has been established in this department from the original Skh :HRII outbred stock in which a variety of juvenile coat colour

346

phenotypes (brown, black and grey) were segregating . The new line was produced by breeding selectively for non-agouti pigment homozygosity (black juvenile coat) and for the capacity to respond to UV irradiation with a continuous strong tan, which appears to be an autosomal recessive character . Breeding pairs for successive generations were selected by these criteria, after exposure of weaner litters to 10 - 15 daily minimal oedemal doses (MODs) of simulated solar UV radiation . The line is maintained by sibling matings, and is periodically tested for retention of the enhanced tanning capability . Untreated mice of this line are strongly pigmented on the ears and tail and the adults are indistinguishable from other Skh :HRII mice . Female mice from both albino and the black capably tanning lines were evenly age distributed (34 - 114 days) into groups of 15 . They were housed in wire-topped plastic boxes on vermiculite bedding (Boral Ltd ., Camellia, NSW) and were bred and maintained at 25 °C with 12 h light (GEC F40GO gold light which does not emit any UV) followed by 12 h dark. They were fed standard mouse pellets (Doust & Rabbidge, Concord West, NSW) and water ad libitum . The probability of mice surviving the experimental period (200 or 225 days) was 0 .94 .

2.2. UV irradiation Simulated solar UV radiation, here designated UVA + B, was provided by a bank of seven fluorescent tubes (length, 120 cm) (six Sylvania F40BL UVA tubes flanking a single Oliphant FL40SE UVB tube) housed in a planar arrangement in a reflective batten and filtered through a layer of cellulose acetate film (0.125 mm) (this reduces the radiation sharply below 290 nm (Celcast, Chatswood, NSW)) . The UVA component was isolated from this source by filtering through a sheet of 4 mm window glass, which attenuated the UVA radiation by 27% as measured radiometrically, and removed any radiation below 305 nm . The relative spectral irradiances were measured between 250 and 400 nm using a computer-controlled doublegrating monochromator as previously described [25] ; they are illustrated in Fig. 1, normalized at 350 nm for comparative purposes, together with global UV . Integrated irradiances for the spectral bands 250- 315 nm and 315 400 nm were measured at the target distance using an International Light IL 1500 radiometer with two detectors for which sensitivity factors had been calculated on the basis of the relative spectral irradiances of the source and the measured spectral responses of the detectors (F . J . Wilkinson, Division of Applied Physics, National Measurement Laboratory, Commonwealth Scientific and Industrial Research Organization) . The integrated irradiances were 2 .7 W in-' (250 - 315 nm ; UVB detectable only) and 52 W M-2 (315 400 nm ; UVA) . The MOD was determined from a graded series of single exposures to UVA + B in groups of ten albino mice as the UVA + B dose resulting in a minimal increase in the double skinfold thickness of 30% at 24 h postirradiation (measured with a spring micrometer ; Mercer, U .K .) . This ap-

347

a 10 0 u C a 1

v

i 9 e

10-5 250

275

300 325 350 Wavolangth (nm)

375

40D

Fig . 1 . Relative spectral irradiance, compared with global sunlight, of simulated solar UV (UVA + B) and its UVA component obtained by filtration through 4 mm window glass (UVA (WG)) . The spectra are normalized at 350 nm . peared to coincide with the first perceptible erythema, which is, however, difficult to assess consistently in albino mice . Mice were exposed unrestrained with the wire cage tops removed, at 45 cm from the lights, on 5 days in a week . They received initially either 1 MOD of UVA + B (10 min) or its window-glass-filtered UVA component alone, calculated to correct for the window glass attenuation (12 min 42 s), or a sub-oedemal dose of UVA + B (6 min) or the window-glass-filtered UVA component of this (7 min 37 s) . Exposure times were increased every week by 20% of the initial exposure time to overcome acquired tolerance . The increments were 2 min or 1 min 12 s for UVA + B and 2 min 32 s or 1 min 31 s for UVA alone at the minimally oedemal and sub-oedemal exposures respectively . At the minimally oedemal exposure the cumulative doses were 153 .9 kJ in' UVB and 2964 kJ M-2 UVA, and at the sub-oedemal exposures the doses were 92 .3 kJ m -2 UVB and 1778 kJ in - ' UVA .

2.3. Tumour

induction Crystalline 5-MOP, kindly donated by Dr . M . M . Lane-Brown, was purified by recrystallization from ethanol-diethyl ether, and its melting point was determined to be 189 °C . Solutions were prepared in A. R . ethanol and stored at room temperature in brown glass bottles with aluminium-foillined caps . The maximum concentrations of 5-MOP which did not augment the 24 h skinfold thickness or cause a visible erythema were determined for both a single MOD and a single sub-oedemal dose of UVA + B . The concentrations were found to be 0 .003% and 0 .01% w/v respectively . The experiments therefore compare the responses to two combination treatment regimes, both subphototoxic after a single treatment, namely sub-oedemal

348 UVA + B with and without 0 .01% 5-MOP and minimally oedemal UVA + B with and without 0 .003% 5-MOP (as well as the UVA component alone of sub-oedemal and minimally oedemal UVA + B with and without 0 .01% 5-MOP or 0 .003% 5-MOP respectively) . The sub-oedemal and minimally oedemal treatment regimes were administered as consecutive experiments . Mice in groups of 15 were painted with 50 pl (Gilson Pipetman) of a 5-MOP solution or ethanol on the dorsal skin . After 30 min the mice were exposed to either UVA + B or to a dose of UVA equal to the UVA component of UVA + B ; these treatments were repeated 5 days per week for 10 weeks . Tumour appearance was monitored and recorded for each mouse until day 200 or 225 and is expressed graphically for each treatment group as the progressive tumour incidence (percentage of mice bearing at least one tumour of diameter 1 mm or larger) and as the progressive tumour multiplicity (average number of tumours per mouse) . Progressive tumour incidence was analysed statistically by the Mantel Haenszel test [26], which is a logrank test for comparing time to first tumour . Final tumour multiplicity was analysed by Student's t test, and the standard error of the mean is shown by bars in the figures . Relative tan intensity between the groups of black mice was assessed visually by unanimous consensus between six observers at the end of the treatment regime (day 70), and was scored 0, +, ++ or +++, the maximum score indicating the most intense, continuous and consistent pigmentation of the dorsal skin . Histological examination of the tan was made in a separate group of black mice exposed daily to UVA + B . Three mice were killed after 9, 12 or 16 daily MOD exposures of UVA + B, mid-dorsal skin (0 .5 em') was excised, fixed in 10% phosphate-buffered formalin and embedded in paraffin . Sections (5 - 6 µm) were stained (Masson-Fontana) for light microscopic examination of melanin deposition .

3 . Results

3.1 . Induction of tan The dorsal skin of the capably tanning Skh :HRII black mouse responds after 10 - 15 daily MODs of UVA + B with a wide area of hyperpigmentation from the shoulder blades to the tail, as well as the dorsal aspect of the head and feet . This tan is uniform and intense, and quite dissimilar to the common response of Skh :HRII mice, which is a mottled pigmentation of varying intensity resembling freckling . The superior induced tan exhibits maximum intensity after 15 - 20 MODs, is maintained during the irradiation regime, and gradually fades after its completion . Histologically, melanin deposition can be seen in the basal layer of the epidermis after 9 MODs (Fig . 2(a)) . The pigmentation becomes dispersed through the epidermal layers as UVA + B irradiation continues and the tan develops . After 12 MODs, melanin can be seen in the stratum comeum (Fig .

349

(a)

(b)

(c) Fig . 2 . Histological sections of black pigmented hairless mouse skin following 9 (a), 12 (b) and 16 (c) minimally oedemal daily exposures to simulated solar UVA + B . MassonFontana, x30.

2(b)), and is also deposited in the dermis with increasing exposure (16

MODs ; Fig . 2(c)) . No qualitative difference is observed if the tan is induced in the presence of 5-MOP, whether activated by UVA + B or by UVA . A background concentration of melanin can be observed in both the epidermis and dermis of untreated skin .

3 .2 . Sub-oedemal UV, 0.01% 5-MOP The progressive tumour incidence is illustrated in Fig, 3 and tumour multiplicity in Fig . 4 . The results demonstrate that UVA + B induces tumours in 85% of the albino mice and 65% of the black mice by day 200 (Fig . 3) . Although the albinos seem to be more susceptible, from this higher tumour incidence and a slightly shorter latent period to first tumour appear . ance (99 days) compared with black mice (106 days), their response is not statistically significantly different from the black mice . Tumour multiplicity (Fig . 4) is similar between albinos and blacks (2 .5 and 2 .0 tumours per mouse respectively at day 200), and although albinos appear to be more susceptible, again this is not statistically significant . Treatment with topical 5-MOP enhances UVA + B tumour incidence significantly in both albinos (P= 0 .055) and in blacks (P = 0 .049), to 100% and 93% respectively (Fig . 3), and reduces the latent period from 99 days



350

albino Aback _ UVA+B +5-MOP .albino UVA + 6-MOP

Claek

40

80

120 T1m+(Daya)

160

200

Fig . 3 . Progressive tumour incidence in albino (full lines) and black (broken lines) hairless mice following exposure to sub-oedemal UVA + B or UVA alone, and after treatment with topical 0 .01% 5-MOP . (albino) or 106 days (blacks) to 78 days . No significant difference is detected between albinos and blacks in this larger response . Final tumour multiplicity is also significantly (P < 0 .025) increased to more than 13 tumours per mouse and again there is no significant difference between albinos and blacks (Fig . 4) . Irradiation with UVA alone is not tumorigenic . However, the combination of topical 5-MOP and UVA treatment induces tumours in 93% of albino mice but only 27% of black mice by day 200 (Fig . 3), with a multiplicity of 5 .2 and 0 .3 respectively (Fig . 4) . This treatment is clearly significantly more tumorigenic in albinos than in blacks in terms of both a greater incidence (P = 0 .005) and a greater multiplicity (P < 0 .005) . In albino mice, the UVA + 5-MOP tumour incidence is similar to the response to UVA + B (no statistically significant difference), and tumour multiplicity at day 200 is even greater than for UVA + B (P < 0 .025) . However, in black mice, tumour incidence and tumour multiplicity are significantly less following UVA + 5MOP than following UVA + B (P < 0 .005) . It is also clear that in albinos (P < 0 .025), and more so in black mice (P < 0.005), treatment with UVA + B + 5-MOP is more tumorigenic than UVA + 5-MOP . In terms of tumour multiplicity, where there is no upper limit to the measurement of the carcinogenic response, it can be seen that the elevated multiplicity derived from UVA + B + 5-MOP is greater than the additive effects of UVA + B plus UVA + 5-MOP .

351

14 13 12 11 10

T

u n

9

as

a

0E Ea

7

1

Tlma(Dayt)

Fig. 4 . Progressive tumour multiplicity in albino (full lines) and black (broken lines) hairless mice following exposure to sub-oedemal UVA + B or UVA alone, and after treatment with topical 0 .01% 5-MOP . Bars indicate standard error of the mean .

Visual assessment of tan intensity at day 70 (Table 1) indicates a perceptible tan (+) produced by UVA + B, a strong tan by UVA + B + 5-MOP (+++), but no tan by UVA or UVA + 5-MOP . Tan intensity appears to parallel the degree of tumorigenicity .

3.3. Minimally oedemal UV, 0 .003% 5-MOP The progressive tumour incidence is illustrated in Fig . 5 and tumour multiplicity in Fig . 6 . By day 225, UVA + B has induced tumours in 100% of albino mice and 93% of black mice . The comparative protection from UVA + B tumorigenesis in black mouse skin is statistically significant (P = 0 .020) and is also demonstrated by an increased latent period to first tumour appearance from 92 to 110 days (Fig . 5) . However, there is no difference in tumour multiplicity at day 225 (albinos 3 .0, blacks 3 .12, Fig . 6) . Treatment with topical 5-MOP under these conditions does not enhance UVA + B tumorigenesis in either albinos or blacks, in contrast with the



35 2 TABLE I Assessment of tan intensity in black mice at day 70 following various topical treatments Treatment

Tan intensity

Sub-oedemal UV UVA+B UVA UVA + B + 5-MOP (0 .01%) UVA + 5-MOP (0 .01%)

+ 0 +++ 0

Minimally oedemal UV UVA + B UVA UVA + B + 5-MOP (0 .003%) UVA + 5-MOP (0 .003%)

100

8o eu e 9 Y

60

e

E f

r 40

i (

20

•

•'

-tjbo o-©-o-rgir' a0

8o

120

180

,

200

Time(Oays) Fig . 5 . Progressive tumour incidence in albino (full lines) and black (broken lines) hairless mice following exposure to minimally oedemal UVA + B or UVA alone, and after treatment with topical 0 .003% 5-MOP .

response to a higher 5-MOP concentration and sub-oedemal UVA + B . However, as with the response to UVA + B, albinos are at significantly greater risk than blacks after UVA + B + 5-MOP treatment, illustrated by a comparatively rapid increase in tumour incidence (Fig . 5 ; P = 0 .021) and an increase in multiplicity (Fig . 6 ; albinos 2 .92, blacks 1 .95) . Irradiation with UVA alone is again not tumorigenic, but becomes mildly tumorigenic in the presence of topical 5-MOP, inducing tumours in



353

a F

-440

BO

120

180

200

Time(days)

Fig . 6 . Progressive tumour multiplicity in albino (full lines) and black (broken lines) hairless mice following exposure to minimally oedemal UVA + B or UVA alone, and after treatment with topical 0.003% 5-MOP . Bars indicate standard error of the mean .

23% of albinos and 14 .5% of blacks at day 225 (Fig . 5), with a final multiplicity of 0 .45 (albino) and 0 .1 (black) (Fig. 6) . Although albinos appear to be more susceptible than blacks, this is not statistically significant . This response to UVA + 5-MOP is markedly less than UVA + B alone (P = 0 .007 for albinos, P < 0 .002 for blacks) . With this treatment regime, as with the response to a higher 5-MOP concentration and sub-oedemal UVA + B, UVA + B + 5-MOP is more tumorigenic than UVA + 5-MOP (P = 0 .007 for albinos ; P < 0 .002 for blacks) ; however, in contrast, UVA + B + 5-MOP is not more tumorigenic than UVA + B . Visual assessment of tan intensity at the termination of the treatment regime (Table 1) indicates that maximum pigment induction is produced by UVA + B alone (+++), but only a moderate tan by UVA + B + 5-MOP (++) and a perceptible tan (+) by UVA + 5-MOP (no tan with UVA alone) . Again tan intensity appears to parallel the tumorigenicity of these treatments in black mice .

4 . Discussion We have described a new line of the Skh :HRII black hairless mouse which responds consistently to UV irradiation of the dorsal skin with an evenly diffuse area of pigmentation . At chronic sub-oedemal exposures to UVA + B and, significantly, at minimally oedemal exposures to UVA + B, this mouse demonstrates protection from tumorigenesis in comparison with

354 the Skh :HRI albino hairless mouse, and is therefore a useful model for studies examining the role of pigment induction on skin tumorigenesis . We have observed that as the line becomes further inbred, the protection afforded by the pigmentation is increasingly obvious and consistently statistically significant at sub-oedemal exposures also . However, this mouse is not yet available as a line congenic with the Skh :HRI strain, and there is a possibility that its reduced phototumorigenic susceptibility may be determined by genetic differences additional to the capacity for melanogenesis . A congenic line of Skh :HRII (grey juvenile coat) is currently available, but this mouse is an inefficient tanner, and in our hands has not demonstrated protection from phototumorigenesis compared with Skh :HRI mice . Under conditions of daily sub-oedemal exposure to simulated solar UV (UVA + B), we have shown that 0 .01% 5-MOP topically applied is hazardous, markedly increasing the tumorigenicity of solar UV in both albino and black mice . This is especially obvious in terms of tumour multiplicity . The enhanced risk cannot be fully accounted for by UVA activation of the 5-MOP, and it seems possible that some UVB, or the wave band less than 315 nm, may also be active in promoting 5-MOP tumorigenesis . In these respects, our results are similar to those of Young et al . [27], who found that UVA + B + 5-MOP, at 0 .0025% or 0.005% (applied in a volume three times greater than used here), was much more tumorigenic than UVA + B in albino mice . The increase in risk was attributed to that portion of the spectrum absorbed by a UVB sunscreen mixture with a maximum absorption around 310 nm . Therefore the action spectrum for 5-MOP tumorigenesis may peak in the longer UVB wave band . Alternatively, UVA-activated 5-MOP may increase the sensitivity of the epidermis to UVB phototumorigenesis . Young et al. [21] have recently indicated that for 8-MOP, using twice the dose described here, the action spectrum peaks between 320 and 335 nm from tumorigenesis responses to a series of quartz glass optical filters transparent to increasing proportions of shorter wavelengths . However, the more opaque filters permit the passage of more UVB than the window glass filter used here and the distinction above and below 320 nm is less clear . It remains possible that the maximum activity may be below 320 run . These findings suggest that the tumorigenic risk associated with PUVA therapy might be reduced by filtering the 'UVA' emission through window glass to remove all wavelengths less than 320 run . Commonly used PUVA fluorescent tubes are known to emit a substantial proportion of UVB . The antiproliferative function of PUVA would hopefully be retained in the transmitted UVA irradiation . At the lower 5-MOP dose of 0.003% no enhancement of minimally oedemal UVA + B tumorigenesis is observed . There does not appear to be a disproportionate increase in the 5-MOP response in the presence of UVB. In the absence of UVB, the UVA + 5-MOP treatment is less tumorigenic than the higher dosage of 0 .01% 5-MOP plus sub-oedemal UVA, and while more hazardous to albino than black mice, is distinctly less tumorigenic than minimally oedemal solar UV (UVA + B) . Therefore there is at least one com-

355

bination of 5-MOP with UVA + B which does not increase the tumorigenic risk . However, if 5-MOP is to be added to sunscreens for use on non-tanning skin, its concentration must be kept at levels which are sufficiently low to allow erythemal doses of sunlight to be absorbed without phototoxicity so that the tumorigenic risk remains less than UVA + B . Pigmentation by UVA + B is enhanced by 5-MOP in the sub-oedemal exposure regime, but not in the minimally oedemal exposure regime where less pigmentation is observed . This reduction by 5-MOP of melanogenesis with minimally oedemal UVA + B is difficult to explain, but we have observed this phenomenon twice in other experiments, and suspect that it may reflect the sensitivity of melanocyte function at certain combinations of UV radiation and 5-MOP . Furthermore, quantitative analysis of the induced tans is hoped to clarify this unexpected observation . Nevertheless, the greater the tan intensity induced by UVA + B, whether in the presence or absence of 5-MOP, the greater the tumorigenic risk . A perceptible tan is achieved with minimally oedemal UVA + 0 .003% 5-MOP but not with sub-oedemal UVA + 0 .01% 5-MOP. This suggests that perceptible tan induction will result from minimally oedemal sunlight exposures filtered by a UVB sunscreen, if 5-MOP is present . However, our results indicate that the induction of melanogenesis is of cosmetic advantage only, and is irrelevant as a protective mechanism from tumour induction . In these black mice, the intensity of the tan achieved represents the degree of tumorigenic potential . Recently, this conclusion has also been reached by others who have compared the action spectra for UV carcinogenesis in mouse skin and for pigmentation in human skin [28] . Acknowledgments This work was supported by a research grant from the National Health and Medical Research Council of Australia . We thank Lyn Blyth, Mares Godhard and Pamela Kenny for assistance in the breeding and maintenance of the hairless mice, and Beverley Horsburgh for her histological expertise . We are also grateful to Dr. P. J . Cooke, Department of Mathematical Statistics, University of New South Wales . References 1 M . J . Ashwood-Smith, G . A . Poulton, M . Barber and M . Mildenberger, 5-Methoxypsoralen, an ingredient in several suntan preparations, has lethal, mutagenic and clastogenic properties, Nature, 285 (1980) 407 - 409 . 2 F. Zajdela and E . Bisagni, 5-MOP, the melanogenic additive in suntan preparations, is tumorigenic in mice exposed to 265 nm UV radiation, Carcinogenesis, 2 (1981) 121 127 . 3 A . R . Young, I . A . Magnus, A . C . Davis and N . P. Smith, A comparison of the phototumorigenic potential of 8-MOP and 5-MOP in hairless albino mice exposed to solar simulated radiation, Br . J. DermatoL, 108 (1983) 507 -518 .

35 6 4 A . M . El Mofti, A preliminary clinical report on the treatment of leukoderma with Ammi majus lien ., J. R . Egypt. Med . Assoc ., 31 (1948) 651 - 665 . 5 M . A . Pathak, K . Jimbow and T. B . Fitzpatrick, in M . Seiji (ed .), Pigment Cell : Phenotypic Expression in Pigment Cells, University of Tokyo Press, Tokyo, 1981, pp . 655 670 . 6 A . S . Zelickson, J . H . Mottaz and S . A . Muller, Melanocyte changes following PUVA therapy, J. Am . Acad . Dermatol ., 1 (1979) 422 - 430 . 7 E . Roser-Maass, E . Holzle and G . Plewig, Protection against UVB by UVA-induced tan, Arch. Dermatol ., 118 (1982) 483 - 486 . 8 R . W . Gange, E . A . Matzinger, B . M . Sutherland and I . E . Kochevar, Comparative protection efficiency of UVA- and UVB-induced tans against erythema and formation of endonuclease-sensitive sites in DNA by UVB in human skin, J . Invest. Dermatol., 85 (1986) 362 - 364 . 9 J . A . Parrish, K . F . Jaenicke and R . R . Anderson, Erythema and melanogenesis action spectra of normal human skin, Photochem . Photobiol., 36 (1982) 187 - 191 . 10 K . Hashimoto, H . Kohda, M . Kumakiri, S . L . Blender and I . Willis, Psoralen-UVAtreated psoriatic lesions, Arch . Dermatol ., 114 (1978) 711 -722 . 11 K . H . Kaidbey and A . M . Kligman, Sunburn protection by long wave UV radiationinduced pigmentation, Arch . Dermatol ., 114 (1978) 46 - 48 . 12 G . Black, E. Matzinger and R . W . Gange, Lack of photoprotection against UVBinduced erythema by immediate pigmentation induced by 382 nm radiation, J . Invest. Dermatol ., 85 (1985) 448 - 449 . 13 F . Gschnait, W . Brenner and K . Wolff, Photoprotective effect of a psoralen UVAinduced tan, Arch . Dermatol., Res ., 263 (1978) 181 - 188 . 14 F . Gschnait, H . Honigsmann, W . Brenner, P. Fritsch and K . Wolff, Induction of UV light tolerance by PUVA in patients with polymorphous light eruption, Br . J. Dermatol, 99 (1978) 293 - 295 . 15 D . J . Cripps, Natural and artificial photoprotection, J . Invest . Dermatol., 76 (1981) 154-157 . 16 A . R . Young, C . S . Potten, C . A . Chadwick, G . M. Murphy and A, J . Cohen, Inhibition of UV radiation-induced DNA damage by a 5-MOP tan in human skin, Pigm . Cell Res., 1 (1988) 350 - 354 . 17 D . J . Cripps, N . J . Lowe and A, B . Lerner, Action spectra of topical psoralens ; a reevaluation, Br. J. Dermatol, 107 (1982) 77 -82 . 18 A . R . Young and I . A . Magnus, An action spectrum for 8-MOP induced sunburn cells in mammalian epidermis, Br. J. Dermatol., 104 (1981) 541 - 548 . 19 K . H . Kaidbey, An action spectrum for 8-MOP-sensitized inhibition of DNA synthesis in vivo, J. Invest . Dermatol., 85 (1985) 98 - 101 . 20 D . D . Grube, R . D . Ley and R . J . M . Fry, Photosensitizing effects of 8-MOP on the skin of hairless mice-II . Strain and spectral differences for tumorigenesis, Photochem . Photobiol., 25 (1977) 269 - 276 . 21 A . R . Young, S . L. Walker and M . Garmyn, A first approach to an action spectrum for 8-MOP phototumorigenesis in mice, J. Invest. Dermatol ., 90 (1988) 175 - 178 . 22 S . A . Baydoun and A . R . Young, An action spectrum for lethal photosensitization of Candida albicans by 8-MOP after low-dose broad-band UVA irradiation ; an action spectrum for 8-MOP 4',5'-monoadducts, Photochem . Photobiol., 46 (1987) 311314 . 23 K . H . Kaidbey, Ultraviolet-induced melanogenesis in man : quantitative aspects, in J . Cahn, P . Forlot, C . Grupper, A . Meybeck and F . Urbach (eds .), Psoralens in Cosmetics and Dermatology, Proc . Int. Psoralen SIR, France, Pergamon, Oxford, 1981, pp . 41 50 . 24 G . Rodighiero, Hyperpigmentation induced by furocoumarins, Farmaco, Ed . Prat ., 40 (1985) 173 - 186 . 25 V . E . Reeve, G . E. Greenoak, C . H . Gallagher, P . J . Canfield and F . J . Wilkinson, Effect of immunosuppressive agents and sunscreens on UV carcinogenesis in the

3 57 hairless mouse, Aust . J. Exp. Biol. Med. Sci., 63 (1985) 655 - 665 . 26 N . Mantel and W . Haenszel, Statistical aspects of the analysis of data from retrospective studies of disease, J. Natl . Cancer Inst., 22 (1959) 719 - 748 . 27 A . R. Young, N. K . Gibbs and I . A. Magnus, Modification of 5-methoxypsoralen phototumorigenesis by UVB sunscreens ; a statistical and histological study in the hairless albino mouse, J. Invest. Dermatol., 89 (1987) 611 - 617 . 28 L. Roza, R . A . Baan, J . C . van der Leun, L . Kligman and A . R . Young, UVA hazards in skin associated with the use of tanning equipment, J. Photochem. Photobiol., B: Biology, 3 (1989) 281 - 287 .