0022-202X/ 81 /770 1-0 139$02.00/ 0 THE JOURNAL OF I NVESTIGATIVE DERMATOLOGY , 77:139- 143. 198 1 Copy right © 198 1 by The Williams & Wilkins Co.

Vol. 77, No. I Printed in U.S.A.

Photocarcinogenesis: An overview

PAUL DONALD FORBES, PH.D.

Center (or Photobiology, T emple University H ealth Sciences Center, Philadelphia, Pennsy lvania, U.S.A.

This paper reviews factors that have been reported to influence photocarcinogenesis in laboratory animals. Such factors include the sensitivity of the test animals, the amount of the ultraviolet radiation (UVR) delivered, the mode of its delivery, and interactions of other radia­tions or of chemicals in the process of carcinogenesis. New data are presented in these areas: reduction in the size of each unit dose (and thus an increase in dosing frequency) increases the carcinogenic effectiveness of a given lifetime dose; certain inbred strains of albino hair­less mice exhibit heritable differences in their suscepti­bility; several chemicals are known to enhance photo­carcinogenesis, but they appear to have so little in com­IIlon, either structurally or functionally, that they offer l.iJ:nited guidance about which other compounds may be effective in this way. Prevention of long-term UVR ef­fects on skin is a desirable goal; development of personal UVR dosimeters will aid in defining the quantitative nature of the problem; improved sunscreens should pro­vide the means to achieve significant reduction in the incidence of UVR-induced human skin cancer.

DW'ing the tenUl'e of these symposia, photocarcinogenesis has been the subject of several presentations, including a com­prehensive review at the 1975 meeting [1). The number of conferences and publications since that time attests to a growing awareness of the subject, and even to its controversial nature. Not everyone shares our basic assumptions on the nature of the light source and the atmosphere which is Oul' primary optical filter [2]; some doubt that our ability to modify the environment is cause for alarm [3]; others are unimpressed with the oft-repeated litany relating skin neoplasia to ultraviolet radiation. Nevertheless, most of us agree that rumors on the demise of the skin cancer problem are premature and that continued effort in this area is a worthwhile occupation. With that in mind, we look now at several categories which relate to our interests in photocarcinogenesis.

CARCINOGEN EFFECTIVENESS-UVR

Identifying and measUl'ing the carcinogen is a problem com­mon to physical and chemical carcinogenesis research. In pho­tocarcinogenesis we speak of 3 r elated questions: (1) dose-re­sponse (designation of amounts required to produce specified responses), (2) dose-delivery (how the response to a given dose is influenced by the rate or route of exposUl'e) and (3) action spectrum (an expression of relative effectiveness in producing a defined response, versus wavelength of t he radiation being measured). These concepts al'e intentionally expressed here in familiar terms; the international language of radiation units is also evolving, and guidelines on more precise terminology are part of recent photobiological literatUl'e [4].

The 3 questions outlined above are implicit in any effort, for example, to develop a model for the relationship of solar expo­sure and human skin cancer. Most of the data on human skin

Reprint requests to: P. Donald Forbes, Ph.D., Skin & Cancer Hos­pital, 3322 North Broad Street, Philadelphia, PA 19140.

Abbreviations: 8-MOP: 8-methoxypsoralen RA: all-trans reti noic ac id TPA: tetradecanoyl phorbol acetate

139

cancer incidence come from populations where the affected proportion never exceeds more than a few percent. Extrapolat­ing to responses at increased anticipated doses requil'es that assumptions be made about the natUl'e of the dose-response re lationship. Most of the assumptions cannot yet be tested on data from human populations, for such efforts must await current and future studies on skin cancer epidemiology and on measurement of human exposures, perhaps with personal UVR dosimeters [5,6]. In the meantime, data from animal experi­ments, both classical and more recent, form an important component in the development of models.

Dose-respon.se

HaiJ'less mice exhibit a dose-dependent response to daily whole-body expo Ul'e to fluorescent sunlamps. As the daily dose increases, the mean latent period decreases, and tumor multi­plicity increases [7]. Recent models have incorporated these animal data [8,9].

Dose-delivery

In the absence of information to the contrary, one assumes time-dose reciprocity, i.e. , that the effect will be proportional to the total dose delivered (or absorbed), where dose is the product of exposure dw-ation times in tensity. This holds quite well for primary photochemical systems, but the time requil'ed to de­liver the dose becomes a significant factor in biological systems [10,11], and many cases of "reciprocity failure" have been demonstrated. For example, the carcinogenic effectiveness of a given dose of ionizing radiation is decreased when the dose is spread over a longer delivery period [12]; in contrast, the carcinogenic effectiveness of DMBA on hail'less mouse skin is in.creased by extending the time between topical applications [13]. Data from ow' animal studies indicate rather consistently that both fractionation and attenuated delivery of UVR in­crease the caJ'cinogenic effectiveness of a given dose [7,11,14]' One illustration of this effect is shown in Fig 1. A specified weekly dose was delivered on 1, 3, or 5 days per week for 40 weeks. The effectiveness was related dij-ectly to the number of exposures (fractions) per week.

For human skin cancer epidemiology a question related to time-dose reciprocity is the possible long-term consequence of episodic overexposure (occasional sunburn). Ow- limi ted expe­rience with a nimal studies indicates that the influence of a re latively large dose "pulse" is less than what would be pre­dicted on the basis of dose additivity alone (Fig 2, Table I) .

Action spectrum

Accurate assessment of envij'onmental impact (atmospheric ozone and the skin cancer question) requil'es information on the UVR-cancer action spectrum which we are unlikely to acquiJ'e for human skin. Some data aj'e available for tumor development in the ear of the mouse [15]; similari t ies to the human erythema action spectra were noted. This study also illustrates the instru­mental a nd biological difficul ty in acquil-ing such data using nominally monochromatic radiation. Additiona l stud ies would be req uil'~d to deteJ'mine whether the individual wavebands were strictly additive in theil' effects. Luckiesh employed an­other approach to the determination of human skin response (erythema and tanning) spectra, incorporating a broad-band light source with a series of short wavelength cutoff filters [16]. The action spectrum was derived from a matrix of simul-

140

1.0

D W >- .8 (J UJ U. u. <! .6 z o ;:: tr: .4 o a.. o if .2 -

FORBES

C UMULATIVE IN CIDENCE : ;; 1 MM TU MORS

10 20 30

WEEKS

40

f7Sb 6 -? -

FIG l. Tumor incidence (affected/survivors) for 3 groups of hairless mice receiving identical weekJy dose of UVR (Westinghouse FS sun- · lamps) . The weekJy dose was given on 1 day (group 3), or one-third on each of 3 days (group 2) or one-fifth on each of 5 days (group 1). Group 3 had a s ignificantly longer latent period than did th e oth er 2 groups.

taneous equations. We have used a similar approach to deter­mining carcinogenic effectiveness in ha irless mice, employing a series of simulated solar UVR spectra [14]. The results indicate that carcinogenic effectiveness increases significan tly with each decrement of simulated atmospheric filtration. Action spectra are currently being evaluated.

Although most of our efforts here have been to define effec­tiveness of the UVB (A 280-320 nm) portion of the spectrum, wavebands outside this region' may contribute independently or interactively to the carcinogenic process. Ionizing radiation may sensitize the skin to the effects of UVR (and vice versa) [17, 18], a nd UV A has been said to produce augmentation of UVB can.:inogenesis [19]. The process of chemical sensitization to pmtions of the UVR spectrum is considered below.

HERITABLE DIFFERENCES

The Cel ts have long been recognized as a population partic­ularly susceptible to UVR-induced skin cancer. Whether this difference is a function of, OT meTely conelated with t his ethnic group's cutaneous characteristics, cannot yet be resolved . Her­itable differences in skjn tumor susceptibility may not be solely a function of pigmentation, cornified tissue thickness and other obvious anatomic features; the nat UTe of the heritable difference is of great practical interest.

There is a limited amount of comparative species information available on photocarcinogenesis. The larger mammals appear to have skin tumor development time' roughly proportional to life expectancy [20]. S maller laboratory rodents (rats, hamsters, mice), particularly those without hair, have shorter lifespa n and shorter skin tumor development time, and are thus preferred for the saving they represent in time and money.

Two strains of albino hairless mice show distinctly different responses to the carcinogenic stimulus of broad spectrum UV A

. plus B-methoxypsoralen(8-MOP) (Grube et al 1977) . In con­h'ast, these strains had similal' dose-response relationships for such acute symptoms as erythema and epidermal cell thymine dimer crosslink formation.

We have fo und that groups of hairless mice, having similar anatomic features but differing in genetic background (Table

. II), can be significa ntly different in th eir susceptibility to tumors induced by simulated sunligh t (Fig 3). These groups were exposed simultaneously to simulated sunlight and their acute reactions were very similar. Two of the groups (HRA/Skh and HRS/J) are from inbred colonies, and we are attempting to identify possible sow'ces ofthei.r dissimilar UVR-cancer suscep­tibility, such as subtle differences in skin structw'e, in epidermal cell kinetics, in vascular responses, or immunologic capability.

en t::: Z :::l

In

~ W (/J

0 D

Vol. 77, NO. 1

Lifetime 16% Incrementa l Pulse at Midpoint 12289

l800t J '00

11 . 600

1800

,00

Daily 1& lifelime)16% Increment 200

IO~+-~~~~~~~~~~~~~~~~~~~~~~IJ~ .• ~Oo~

200

n/~ .!...+'+~f4++"'rI+I-I.j.,..,.j.+j..j..l.,,...fI.j..j..I.,..,f~~+++t-,..,..H+l+r,.1+I-t+n++-t#---.:'.:::O.:::00:!!..0 3 4 5 6

WEEKS 7 9 10

FIG 2. Schematic descrip tion of experimenta l design: "pulse dose" experiment. Fow' groups of mice were exposed to a xenon arc solar simulator, 5 days per week for 10 weeks. Group 1 received 200 RB units of UVR (about half the dose required to produce minimal erythema in untanned human skin) per day. Group 2 received an additional 16%. daily; group 3 received its addi tional 16% lifetime dose (1 ,600 RB uni ts) aU on the 1st day; group 4 received its additional 16% at the beginning of the 6th week. Table I gives a summary of the resul ts.

TABLE I.

Lifetime T", Group dose Description x tumorsb

(R·B uni ts) (wk)"

1 10,000 Baseline dose 55 5.4 2 11,600 Daily 51 8.0

Increment 3 11,600 Initial 66 5.0

" pulse" 4 11,600 Mid-cow'se 56 5.4

Hpulse"

" T r>o = Time in weeks to reach 50% incidence (median latent period~ /, Tumors = average number of tumors per mouse (tumors ~ 1 mm).

IMMUNOLOGY

The mice described above were immunologically evalUated by monitoring their ability to respond to various mitogen:> (concanavalin A, phytoh em agglutjnjn and lipopolysaccharide) and to both T-cell dependent (sheep erythrocytes) and T-cell independent (pneumococcal polysaccharide Type III) antigens. The numbers of T lymphocytes (thymus-processed) and B lymphocytes (surface immunoglobulin-positive) in the thymus, spleen and pooled lymph nodes of individual a nimals were determined using immunofluorescence techniques; the value for each group of animals were found to be comparable. The antibody-forming ability was evaluated by determining the number of pla.que-forming cells in the spleen after immunization with pneumococcal polysaccharide or sheep erythrocytes. All strains of mice responded to both typ es of antigens, and all groups of mice responded to the standard mitogens listed above (submitted for pubJicationi . We are now beginning to examine these and other parameters in UVR-irradiated mice.

Immunosuppressive drugs appear to enhance photocar cino-' genesis clinically and experimentally [22]. The mechanism of effect, however, has not yet been clearl y defined and the overall importance of the immune system in skin cancer development has remained conjectw·al. One test system does clearly dem-

July 1981 PHOTOCARCINOGENESIS 141

T ABL E II. Characteristics of test anim als

Stock or stra in Designa l ion Relevant gene tics!1 Ha il· growth Color Source

Response curve "# (Fig 3)

Stock Skh:HR (type 1) Skh:HR (typ e 2) HRA/ S kh Skh:CRH C3H / HeN-hr HR/ D e/ Hflcr HRS/ J

hr/ h1· hr/ hr hr/ hr + / crh + / hr + / h1" + / hr + / hr + / ab

c/ c S&C Temple Univ. S&C T emple Univ. S&C T emple Univ. S&C T emple Univ. NIH

1 9 2 3 6 5 7 B 4

Strain (FIB) Stock

+ / c, + / b, + / a c/ c + / c

Strain (NB, F3) Strain (F BD) Strain (F54) Stock

pi p, bi b c/ c

Inst. Cancer Res. J ackson Lab. Argonne Nat. La b. Argonne ha irless

BALB/ cSkh-ab c/ c

Str a in (FBI +37) c/ c, bi b Univ. of California, Berkeley to S&C T emple Univ.

a Genotype symbols, c-a lbino, ab-asebia , b-brown, a-non agout i, crh-crytpothrix, hr-hairless, p-pink eyed.

o n stra te t umor growth under immunologic control. UV-irradia­tion of C3H haired mice apparently induces appearance of cells w hich inhibit the animals' a bili ty to reject UV -induced tumors [23]. Through a series of elegant experiments these investigators h a ve demonstra ted the involvement of T-Iymphocyte suppres­s or cells. The system apparently can react different ially to UVR alone or to photosensitization (UVA plus psoralen) [24).

D e Gruijl and van der Leun have also demonstrated a sys­temically-mediated effect of UVR on photocar cinogenesis [25]. Preexposing one side of hairless mice to UVR makes the oth er side more suscept ible to subsequent tumor induction by U VR. Whether this systemic effect is immunologically me­diated remains to be determined.

CHEMICAL INTERACTIONS

The inflammatory and carcinogenisic proper ties of UVR have been recognized for several decades; the interaction of exoge­n o us chemicals in photocarcinogenesis is a more recent concern, and has been discussed in reviews by several authors [1,26-28].

This sequence actually begins wi th the observat ion that UVR can profoundly influence chemical carcinogenesis. To generalize from a wealth of sometimes conflicting experimental da ta, it n ow appears (at least in the case of mice t reated with aromatic hyd rocarbon carcinogens) that light may contribute in 2 oppos­ing ways: (1) by degradat ion of the car cinogen to noncarcino­genic products and (2) by stimula ting a phototoxic response which appears to be coincident wi th an increased tumor yield [29]. In theory, cal·cinogens may also be genera ted photochem­ically in situ [29].

A problem of increasing magnitude concerns photo-induced carcinogenesis following application of agents to the skin which mayor may not be phototoxic, but in themselves not carcino­genic. Among the phototoxic agents, 8-MOP is the most widely studied in terms of photocarcinogenic interaction [1 ,21]'

R etinoids (vitamin A and its analogues) have been known to be of importa nce for cell differentia tion in many epithelial tissues. In recent years, there has been great interest in the possibili ty that retinoids may affect preneoplastic cell differen­tiation, and can even reverse early epithelial neoplasms induced by chemical carcinogens [30-32). Davies [33] was one of the first to show tha t mice fed a vi tamin A supplemented diet developed fewer skin papillomas following a single dose of 7,12 D MBA than did similar mice fed a diet defi cient in vitamin A. However, Epstein reported enha nced photocarcinogenesis in th e presence of retinoic acid (28). Utilizing topical application of retinoic acid in methanol, preceded by simulated sunligh t irradia tion, we also found marked enhancement of skin photo­car cinogenesis [34]. It is not known at this time whether topical application of retinoic acid acts differently from systemic ad­m inistration, or whether photocal·cinogenesis is a ffected differ­ently from chemical carcinogenesis. Retinoids are suspected of affecting DNA metabolism [32], certa inly cause epithelial hy-

perplasia and affect cell different iation [35). Whether any of these phenomena are related to the mechanism of the observed enhancing effect of topical all -trans retinoic acid applica t ion on UV photocarcinogenesis is yet to be determined.

Our current studies indicate (1) that RA, applied either intercurrent ly with UVR, or applied after a COUl"se of UVR ("promot ion" technique), can enhance the development of tu­mors; (2) that RA in the diet can produce a similar (though reduced) effect systemically; and (3) that RA can enhance the development of DMBA-induced tumors (see T a ble III). T he various treatment regimes did not all resul t in a uniform tumor type. This aspect will be discussed in the next section. Har t­mann and Bollag have also found enhancement ofphotocru·cin ­ogenesis by retinoic acid (36). Verma, Conrad, and Boutwell report that retinoic acid, under cer tain cil·cumstances, poten­tiated the formation of skin tumors by DMBA [37].

T UMOR CRITERIA

UVR is though t. to induce a variety of tumor types in human skin, wi th squamous cell and basal cell cru-cinomas predominat­ing [38). Solar keratoses, carcinoma in-situ, and keratoaca n­t homata probably relate developmentally or anatomically to squamous cell carcinoma; UVR apparently induces few sarco­mas in human skin . The relationship of UVR to melanoma is still conjectural [39].

UVR-exposed animals develop essent ially no distinguishable basal cell car cinomas; they have not yet provided a useful model for malignant melanoma in skin; they produce a variety of epi thelia lly-derived tumors, including benign epitheliomas and carcinomas; and t hey can yield sarcomas under some circumstances (40).

In mice, skin tumor types are varied but not entil·ely random. The distribu tion of tumor types can be influenced not only by skin type (hail·ed vs. hail-less, etc.), bu t also by vru·iables in the treatment . For example, brief exposure to DMBA or UVR leads pr imarily to the formation of benign epithiliomas; more pro­longed exposure increases the liklihood of carcinoma develop­ment . It is of interest to us that either tetradecanoyl phorbol acetate (TP A) or all trans retinoic acid (RA) can enhance the development of UVR-induced papillomas or cru·cinomas, de­pending on the treatment conditions; in tercurrent treatments favor cal"cinoma development, whereas sequential " promotion" favors papillomas. Usefuln ess as a chemical probe in the carci­nogenic process is apparent ly a featm e to be added to RA's repetoil· of biologic effects.

PREVENTION

If most of human skin cancer is caused by UVR exposure, then most such cancers ar e theoretically preventable [6]. In practice, it · appears unlikely that most susceptible people will be convinced to abstain fro m UVR exposure, or to substitute skin paint ing for a sun-induced tan.

The tumor risk may at least be reduced by decreasing the

142

1.0

. 8

uu () .6 z uu 0 ()

~ .4

. 2

0

1.0

. 8

uu .6 () z uu 0 () .4 z

.2

0

1.0

.8

uu . 6 ()

z UU 0 () z .4

.2

0

FORBES

TUMORS > 1MM

A 10

B 10

C 10

20

WEEKS

• •

20

WEEKS

20

WEEKS

30

•

• 7

• •

12740-2

30

• •

30

FI G 3. A ·C, Tumor incidence (affected/survivors) in nonhaired mice from differen t genet ic backgrounds. All animals exposed s imultaneously tu a xenon arc solar s imulatur, 400 RB cuunts/day. Gruups #5 and 8 a re sepa rated out for cla ri ty. The groups show significant differences in photocarcinogenic sensit ivity. See T able II fur descrip t ion of stocks and strains.

Vol. 77, No. 1

TABLE III. S ummary of all-tran.s R etin.oic A cid (RAJ studies (Temple Un.iuersit»)

Study Chemical UVR RA HA 10# " initiator" exposure exposure effect

#77-50 Xenon lamp In tercu rrent, More tu mo r . 7 days/ wk; top ica l shorte r mean 40 wk latent pe riod

#77-54 Xenon lamp In tercurre nt; More tumo rs, 5 days/wk; top ical shorter mean 40 wk late nt period

#78-95 FS lamp 5 wk 7-20 More tumors, days/wk: 6 ("promo· shorter mean wk tion") la tent period

#78-97 benzo(alpyrene wk 7-20 No significant ("promo· effect tion")

#78-84 Xenon lum p Intercurren t; More tumors. 5 days/wk: topical shorter mean 40 wk latent period

#78-84 Xenon lump In diet S light increase 5 days/ wk; in tumors, 40 wk shorter la·

tent period #79-36 FS sunlamp Topical wk More tumors,

Day 1-6 6-20 (" pro· shorter mean (newborn) motion") late nt pc·

rioel!! #80-29 DMBA Topical wk More tumors,

7-20 ("p ro· shorter mean molion") late nt pe·

riodll

#80-29 FS sunlamp T opical wk More tumors, 5 days/ wk; 7-20 ("pro· shorte r mean 3 wk mo tion") latent pe·

riod"

" Mostly pedunculated epitheliomas.

quantity of UVR reaching the target layer. Knox et al ~'eported 20 yr ago that sunscreen lotion can reduce the UVR-tumor yield in mouse ears [41], and recently Kligman, Akin, and Kligman have confIrmed this finding on the backs of hail-less mice [42]. Further studies could undoubtedly result in improved UV absorption and substantivity by sunscreen prepal"ations. Whether a "systemic sunscreen" is a feasible technique awaits extensive research.

Black et ai, have reported protection against photocarcino­genesis by dietary antioxidants [43,44]. We fed hairless mice a semi purifIed diet containing Black's antioxidant formula, and did not find a reduction in photocarcinogenic sensitivity (un­published data). The potent ial importance of this approach is evident, and further studies are clearly needed.

CONCLUSION

Experimental photobiology provides the opportunity to deal with an ubiquitous and familial' car cinogen. The ultimate con­sequences of significant sunlight exposure may exhibit them­selves for all to see, but most of the operating forces are hidden from sight. We have a great deal yet to learn a bout factors which influence the skin's response to ultraviolet radiation, and about the preventitive measures that might reduce the risks associated with sunlight exposure. Fortunately, many clues are now available, and several techniques for investigation ar e in hand.

REFERENCES 1. Urbach F, Forbes PD, Davies RE, Berger D: Cutaneous photobiol­

ogy: Past, Present and Fu ture. J Invest Dermato l 67:209-224, 1976

2. J ohnson CK: The sun does not set. Flat Earth News (In te rnationa l Flat Earth Research Society, Lancaster, California 93534) Ma rch 1978

3. Allaby M, Lovelock J : Spray cans: the threat that never was. New Scientist 17 July, 1980

4. Rupert CS, Latru-jet R: Towru'd a nomenclature and dos imetric scheme applicable to a ll radiations. Photochem Photobiul 28:3-5, 1978

July 1981

5. Scotto J , Fears 1', Lisiecki 0 , Rado evich S: Incidence of nomela­noma skin can cer in the Uni ted States, (Pre limina ry report) DHEW publication No. (NIH)80-2154 Wa hington D.C., 1980

6. Ur bach F, Spikes JD, Epstein JH: R ecommendations for fu ture r esearch on ultrav iolet radiation carcinogenesis. Natl Cancer Inst Monograph 50:203-205, 1978

7. Forbes PD, Blum HF, Davies HE: Photocarc inogenesis in ha irless mice: Dose-response a nd the influence of dose-delivery. Photo­chem P hotobiol, 1981

8. Rundel RD, NachLwey DS: S kin cancer and ul travio let rad iation. Photochem Photobiol 28:345-354, 1978

9. D e Gruijl FR, van der Leun JC: A dose- response model for skin cancer induction by chronic UV exposure of a human population. J Theor B ioi 83:487-504, 1980

10. LeG rand Y: An introduction to photobiology. New York, Amer. E lsevier Publishing Co., Inc., 1967

11. Forbes PD, Davies RE, Urbach F: Aging, environmenta l influences, and photocarcinogenesis . J Invest Dennatol 73: 131- 134, 1979

12. Ullrich RI, Storer JB: Influence of gamma irradiation on t he development of neoplastic disease in mice III . D ose- rate effects. Radiation Res 80:325-342, 1979

13. Iversen U, Iversen OH , Bjerknes R: A comparison of the tumouri ­genic effect of five graded doses of 3-methyl-cholanthrene applied to the skin of ha irless mice at in terva ls of 3 or 14 days. Acta P athol Microbiol Scandinav 73:502- 520, 1968

14. Forbes PO, Davies RE, Urbach F: Experimenta l ul traviolet pho­tocarcinogenesis: Wavelength interactions and time-dose rela­tionships. Nat! Cancer Inst Monogr 50:31-38, 1978

15. Freeman R G: Data on the action spectrum fo r ul trav iolet carcino­genesis. J Natl Cancer Inst 55: 111 9- 1121, 1975

16. Luckiesh M, HoUaday LL, Taylor AH: React ion of un tanned hu ­man skin to ul trav iolet radiation. J Optical Soc Amer 20:423-432, 1930

17. Burns FJ , Alber t RE: Tumorigenic action of beta, proton, a lpha and electron radiation on the rat skin . Repor t U.S. Dept of Energy #AT(Il-l)3380-30, 1980

18. Stern RS: Risk of cutaneous carcinoma in pa tien t t reated wi th oral methoxalen photochemotherapy for psoriasi . N E ngl J Med 360:809-813, 1979

19. Willis I , Minter J , Whyte HJ : The rapid induction of cancer in th e hairless mouse utilizing the principle of photoaugmentation. J Invest Dermato l, in press, 1981

20. Kral F , J ovak B, Beck JD: Veterinary D ermato logy. JB Lippincott Co., Philadelphia 1953

21. Grube DD, Ley RD, Fry RJM: P hotosensitizing effects of 8-me­thoxypsoralen on the skin of ha irless mice II . Stra in and spectral differences for t.umorigenesis. Photochem Photo bioi 25:269-276, 1977

22. Forbes I'D, Davies RE, Urbach F: Experimenta l ultravio let cru'ci­nogenesis, Research in P hotobiology. Edited by A Castellani Plenum. New York, 1977, p 469

23. KJ'ipke ML, Lofgreen J S, Beru'd J : In. vivo immune responses of mice during ca'rcinogenesis by ul t raviolet radiat ion. J Nat! Cancer Inst 59: 1227-1230, 1977

PHOTOCARCINOGENESIS 143

24 . Kripke ML, Morison W, Parrish J: Differences in the immunologic reactivity of mice treated with UVB or methoxsalen plus UV A rad iation . J Invest Dennatol, in press 1981

25. deG ruijl FR, van der Leun JC: On the interdependence of UV induced skin tumors. In ternat Congr Photobiology Strasbourg (France) abs# P158, 1980

26. Emmett EA: Ultrav iolet radiation as a cause of skin tumors. CRC Cri t Rev in Toxicol 2:211 , 1973

27. B lack HS, Chan JT: Experimenta l ul traviolet ligh t. cru·cinogenesis. P hotochem P hotobiol 26: 183-1 99, 1977

28. Epste in JH: Chemicals and photocarcinogenesis. Aust J Dermatol 18:57-61, 1977

29. Davies R E: In te raction of light and chemicals in carcinogenesis. Natl Cancer Inst Monogr 50:45-50, 1978

30. Mulay DN, Urbach F: Local therapy of ora l leukoplakia with vitamin A. Arch D ermatol 78:637- 638, 1958

31. Bollag W: Prophylax is of chemica lly induced papillomas a nd cru'­cinomas of mouse skin by vitamin A acid. Eur J Cancer 8:689-693, 1973

32. Sporn MB, Dunlop NM, Newton DL, S mi th JM: Prevention of chemical carcinogenesis by vit.amin A and it.s synthetic ana logs. Fed P roc 35:1 332-1338, 1976

33. Davies RE: Effect of vitamin A on 7,12-d imethylbenz(a)anthracene­induced papillomas in rhino mouse skin. Cancer Res 27:237- 241 , 1967

34. Forbes PO, Urbach F, Davies RE: E nhancement of experimental photocarcinogenesis by topical retinoic ac id. Cancer Letters 7: 85-90, 1979

35. Zil JS: Vitamin A acid effects on epidermal mi totic activity, t hick­ness and ce llu larity in the ha irless mouse. J Invest Dermatol 59: 228-232, 1972

36. Hartmann HR, Bolag W: Persona l communication. 37. Verma AK, Conrad EA, Boutwell RJ<: Induction of mouse epide r­

mal ornithine decarboxylase activity and skin tumors by 7,12-dimethylbenz(a)anthracene: modulation by ret inoic acid and 7,8-benzoflavone. Carcinogenesis 1:607-6 11 , 1980

38. Steigleder CJ<: Classification of skin tumors and assoc iated lesions. Nat! Cancer Ins! Monograph 50:221-222, 1978

39. Cutchis P: On the linkage of solar ul trav iolet radiation to skin cancer. U.S. D ept of Transportat ion Report No. FAA-EQ-78-19 (IDA paper #P-1342) Washington, D .C., 1978

40. Stenbiick F: Life history and histopathology of ul trav iolet ligh t­induced skin tumors. Natl Cancer Inst Monograph 50:57-73, 1978

41. K nox JM, Gri ffin AC, Hakim HE : Protection from ultraviolet carcinogenesi-. J Invest Dermatol 34:51-58, 1960

42. Kligman LH, Akin F J , Kligman AM: Sunscreens prevent ultraviolet photocarcinogenesis. J Am Acad Dermatol 3:30-35, 1980

43. Black HS, Chan JT, Brown GE: Effects of dietar y constituen ts on ul t raviolet ligh t-mediated carcinogene is. Cancer Res 38:1384-1387, 1978

44. B lack HS, Kle inhans CM , Hudson HT, Sayre RM , Agin PP: Studies on the mechanism of protection by buty lated hydroxy toluene to UV-radiation dam age. Photobiochem P hotobiophys 1:119- 123, 1980