Proc. Natl. Acad. Sci. USAVol. 89, pp. 7516-7520, August 1992Immunology

Pyrimidine dimers in DNA initiate systemic immunosuppressionin UV-irradiated mice

(UV-B radiation/delayed hypersensitivty/photoimm uolg/excislon repair/liposomes)

MARGARET L. KRIPKE*t, PATRICIA A. Cox*, LORI G. ALASt, AND DANIEL B. YAROSHO*Department of Immunology, The University of Texas M. D. Anderson Cancer Center, Houston, TX 77030; and tApplied Genetics Inc., Freeport, NY 11520

Communicated by R. B. Setlow, April 30, 1992 (received for review February 12, 1992)

ABSTRACT Exposing the skin of mice to UV radiationinterferes with the induction of delayed and contact hypersen-sitivity immune responses initiated at nonirradiated sites. Theidentity of the molecular target in the skin for these Immuno-suppressive effects of UV radiation remains controversial. Totest the hypothesis that DNA is the target for UV-inducedsystemic immunosuppression, we exposed C3H mice to UVradiation and then used liposomes to deliver a dimer-specificexcision repair enzyme into the epidermis in situ. The appli-cation of T4 endonuclease V enulated in liposomes toUV-irradiated mouse skin decreased the number of cyclobu-tane pyrimidine dimers in the epidermis and prevented sup-pression of both delayed and contact hypersensitivity re-sponses. Moreover, the formation of suppressor lymphoid cellswas inhibited. Control, heat-inactivated endonuclease encap-sulated in liposomes had no effect. These studies demonstratethat DNA is the major target ofUV radiation in the generationof systemic immunosuppression and suggest that the primarymolecular event mediating these types of immunosuppressionby UV radiation is the formation of pyrimidine dimers. Fur-thermore, they illustrate that the delivery of lesion-specificDNA repair enzymes to living skin after UV Irradiation is aneffective tool for restoring Immune function and suggest thatthis approach may be broadly applicable to preventing otheralterations caused by DNA damage.

Wavelengths of UV radiation in the middle, or UV-B (280-320 nm), range can impair a variety of immune responses inhumans and laboratory animals both locally, within UV-irradiated skin, and systemically, at distant sites (1). Expo-sure of mice to UV-B radiation interferes with the rejectionof UV-induced skin cancers and the induction of delayed andcontact hypersensitivity (DHS and CHS) responses initiatedat unirradiated sites; these forms of immune suppression areassociated with the induction of antigen-specific suppressorT lymphocytes (2). How UV-B radiation exerts its systemic,immunosuppressive effects is a question of considerableinterest, both for understanding the regulatory pathwaysgoverning these immune responses and for assessing thepotential effects of UV-B radiation on human health. TheDHS response is particularly important in this regard becausethis T-lymphocyte-mediated immune reaction is responsiblefor protection against many chronic infectious diseases.

Current experimental evidence implicates soluble sub-stances derived from UV-irradiated keratinocytes as theprobable mediators of UV-induced systemic suppression ofDHS and CHS responses (3-5). However, the initial photo-biological reaction responsible for triggering the cascade ofevents leading to activation of the suppressor pathway of theimmune response remains controversial. Based on an in vivoaction spectrum for systemic suppression of CHS in the

mouse, it has been proposed that urocanic acid, a deamina-tion product of histidine, present in the stratum corneum, isthe photoreceptor for this form of UV-induced immunosup-pression (6). Several subsequent studies supported this hy-pothesis (7-9) by demonstrating that the cis isomer of uro-canic acid, which is formed upon UV irradiation ofthe nativemolecule, has immunosuppressive activity in mice. On theother hand, studies with the South American opossum Mon-odelphis domestica implicated DNA damage as the initiatingevent in UV-induced suppression of CHS, at least in thesedistant relatives of eutherian mammals (10). Cells from thesemarsupials have the ability to repair UV-induced pyrimidinedimers in DNA by means of a photoreactivating enzyme,which binds to DNA, forming a complex that absorbs energyin the visible wavelength range. The absorbed energy splitsthe dimers, thereby restoring the DNA to its original config-uration. In these studies, UV-induced suppression of CHSwas prevented by exposing the opossums to photoreactivat-ing light immediately after exposure to UV radiation (10).These results implicated DNA damage as the trigger forUV-induced suppression ofCHS in marsupials; however, itsrole in immune suppression in placental mammals remainedopen to question. Because cells in the skin of adult mice donot have a photoreactivation mechanism for repairing pyrim-idine dimers (11), this approach could not be used to deter-mine the role ofDNA damage in UV-induced immunologicaleffects in the mouse.We have devised an alternative approach for examining the

role of DNA damage in the initiation of UV-induced immu-nosuppression in mice by using liposomes containing anexcision repair enzyme, T4 endonuclease V, which is specificfor pyrimidine dimers (12). When topically applied to murineskin, T4N5 liposomes penetrate cells of the epidermis (13).The multilamellar lipid vesicles destabilize at low pH,thereby delivering the endonuclease intracellularly and in-creasing the rate of repair of pyrimidine dimers in DNA (14,15). In the present studies, we have used this approach toanalyze the role of pyrimidine dimers, induced in DNA byUV irradiation in vivo, in suppressing the DHS and CHSresponses in mice.

MATERIALS AND METHODSMice. Specific pathogen-free C3H/HeN (mammary tumor

virus negative) female mice were obtained from the FrederickCancer Research Facility Animal Production Area (Freder-ick, MD). Age-matched mice between 10 and 12 wk old werehoused in filter-protected cages, and ambient lighting wascontrolled to provide 12-h light/12-h dark cycles. AutoclavedNational Institutes of Health open formula mouse chow andwater were provided ad libitum. The animal facility is ac-

Abbreviations: CHS, contact hypersensitivity; DHS, delayed hyper-sensitivity; DNFB, 2,4-dinitrofluorobenzene; FITC, fluorescein iso-thiocyanate.tTo whom reprint requests should be addressed.

7516

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 89 (1992) 7517

credited by the American Association for Accreditation ofLaboratory Animal Care; procedures were approved by theInstitutional Animal Care and Use Committee.UV Irradiation. The UV source was a bank of six FS40

sunlamps (National Biological, Twinsburg, OH), which emit=z60% of their radiation within the UV-B (280-320 nm) rangeand have a peak emission at 313 nm. The average irradianceof the source was =9 W/m2, as measured by an IL-700radiometer (International Light, Newburyport, MIA), whichdetects radiation in the wavelength range between 280 and320 nm. The dose rate of the incident radiation received byanimals was 4.5 W/m2 because of screening by the cage lid.Before irradiation, the dorsal fur of the mice was shaved withelectric clippers, and the animals were placed in individualcompartments in cages that were located 20 cm below thelight source. Mice used for measuring CHS had their earscovered with electrical tape during irradiation. Except forexposure to UV radiation, control mice were treated exactlythe same as the irradiated mice.T4N5 Liposomes. T4N5 liposomes were prepared by en-

capsulating purified, recombinant T4 endonuclease V inliposomes composed of phosphatidylcholine/phosphatidyl-ethanolamine/oleic acid/cholesterol hemisuccinate (2:2:1:5molar -ratio) by the detergent dialysis method (14). Theconcentration of the entrapped enzyme was determined byELISA (16) and is expressed as pg of T4 endonuclease V perml of vehicle. The encapsulated activity was assayed bynicking UV-supercoiled DNA with and without dissolution ofthe liposomes (16). Control preparations of liposomes con-tained boiled (enzymatically inactive) T4 endonuclease V(14). The liposomes were mixed into a 1% hydrogel (HypanSS201; Kingston Hydrogels, Dayton, NJ) made with phos-phate-buffered saline and applied to shaved mouse skin witha moist cotton swab. Immediately after UV irradiation, 0.25ml of liposome suspension containing 0.5 pg of T4 endonu-clease V per ml was applied to the UV-irradiated skin.Assay for Pyrimidine Dimers. The 'number of pyrimidine

dimers in epidermalDNA was measured by the endonucleasesensitive-site assay using alkaline agarose gels (17). In thisassay, the frequency of dimers in DNA is measured by thechange in average single-stranded DNA length in alkaline gelsafter treatment with T4 endonuclease V. The lengths ofDNAin the control and endonuclease-digested samples are calcu-lated by using DNA restriction fiagments of defined lengthrun in adjacent lanes of the gel, and the average molecularlength is computed by the same (weight - average molecularweight)/2 calculation used in the well-established alkalinesucrose gradient method (18). Under these conditions, dimerinduction is directly proportional to UV-B fluence up to 100dimers per million bases (17), and all measurements reportedhere are within this range. The dimer measurements in thealkaline agarose gel assay agree closely with results from thealkaline sucrose gradient method (18).The epidermis was isolated by overnight digestion of

excised dorsal skin in 0.25% trypsin on ice, and the DNA waspurified by two rounds of proteinase K digestion (100 pg/mlin 1% SDS; 370C for 30 min); phenol, phenol/chloroform (1:1,vol/vol), and chloroform extraction; and ethanol precipita-tion. The DNA was then treated with purified T4 endonu-clease V (10 pg/ml) to produce breaks at all dimer sites, andthe single strands were separated by alkaline agarose gelelectrophoresis. Photographic negatives of the ethidium bro-mide-stained gels were scanned by a Hoefer GS300 scanningdensitometer, and the output was digitized by a Data Trans-lations 2805 analog/digital converting board in a CompaqDeskpro computer. The stored files were used to calculatethe average molecular length for each sample, and the changein average molecular length was used to calculate the dimerfrequency. The approximation of linear film response wasused to calculate average molecular length because the T4

endonuclease V-treated samples showed fluorescence thatfell within the linear portion of the film response curve; theuntreated DNA samples showed fluorescence of single highmolecular weight bands, most affected by a nonlinear filmresponse, but nearly identical in all samples and subject to thesame small error. At each UV dose, the average dimernumbers per unit length of DNA from at least two sampleswere compared in skin treated with either active or inactiveliposomes or not treated with liposomes.DRS Response. Formalin-fixed Candida albicans was pre-

pared as described (19). Mice were sensitized by injecting 0.1ml of formalin-fixed C. albicans cells (1 x 107 cells) s.c. intoeach flank. Nine days later, the hind footpad thickness wasmeasured with a spring-loaded micrometer (Swiss PrecisionInstruments, Los Angeles, CA), and the animals were chal-lenged by intradermal injection of 50 A1 of Candida antigen(Antigen Supply House, Northridge, CA) in both hind foot-pads. Footpad thickness was measured again 24 h later, andthe swelling was determined by subtracting prechallengefrom postchallenge measurements. The thickness of thefootpads before challenge ranged from 1.7 to 2.0 mm. Forboth D4S and CHS, percentage suppression was calculatedfrom Eq. 1 as follows:

T-N% suppression = 1 - ~~x100,

P-N[1]

where N = negative control (response of unsensitized miceto challenge), P = positive control (response of sensitizedmice ± liposomes to challenge), and T = test group (responseof mice given UV radiation ± liposomes before sensitizationand challenge).CHS Response. Animals were sensitized by the epicutane-

ous application of 2,4-dinitrofluorobenzene (DNFB) (50 ,ul ofa 0.3% solution in acetone) on shaved abdominal skin. Fivedays later, the mice were challenged by applying 5 1 of a0.2% solution ofDNFB onto each ear surface. Alternatively,mice were sensitized with fluorescein isothiocyanate (FITC)according to the method of Thomas et al. (20). A 0.5% FITCsolution (Aldrich) was prepared in a solvent of equal volumesof acetone and dibutyl phthalate (Aldrich). For sensitization,0.4 ml of the solution was applied to shaved abdominal skin.Five days later, the CHS reaction was elicited by applying 5;l ofthe 0.5% FITC solution to each surface ofboth ears. Thethickness of each ear was measured immediately beforechallenge and 24 h later. The ear thickness before challengeranged from 0.24 to 0.30 mm.

Transfer of Spleen Cells. On day 10 after sensitization (day6 for CHS), spleens were removed from mice that exhibiteda decreased DHS or CHS response and from appropriatecontrol animals. Single-cell suspensions were prepared frompooled spleens by teasing the spleens apart in Hanks' bal-anced salt solution. Clumps were removed by filtrationthrough nylon mesh. The cells were washed once and refil-tered, and 108 viable, nucleated cells were injected i.v. intosyngeneic recipients. Immediately thereafter, the recipientswere sensitized with C. albicans or FITC; the DHS or CHSresponse was measured 10 or 6 days later, respectively, asdescribed above, and percentage suppression was calculatedaccording to Eq. 1.

Statistical Analysis. The significance of differences in theDHS or CHS response between control and treatment groupswas determined by ANOVA for repeated measures. Eachgroup contained at least five mice, and each experiment wasperformed at least twice and usually three times.

RESULTSEffect ofT4N5 Liposomes on UV-Induced Suppression of the

Induction of DHS to Candida. Exposure of mice to UV-B

Immunology: Kripke et al.

Proc. Natl. Acad. Sci. USA 89 (1992)

radiation before immunization inhibits the induction of theDHS response to C. albicans injected s.c. at an unirradiatedsite; this treatment also causes the formation of suppressor Tlymphocytes that suppress the induction ofDHS to Candida(19). To determine whether liposomes containing T4 endo-nuclease V would abrogate these immunosuppressive effectsof UV irradiation, groups of mice were exposed to a singledose of 5 kJ/m2 UV-B radiation, and liposomes containingactive (T4N5) or heat-inactivated enzyme were immediatelyapplied to the UV-irradiated dorsal skin. The mice wereimmunized 3 days later by s.c. injection of formalin-fixedCandida. As shown in Fig. 1A, application of T4N5 lipo-somes completely abrogated the suppressive effect of UVirradiation; in contrast, liposomes containing heat-inacti-vated endonuclease had no effect. This experiment wasperformed six times with very similar results. Transfer ofspleen cells from the immunized, UV-irradiated mice tonormal, syngeneic recipients rendered the recipients unre-sponsive to immunization with Candida, indicating that sup-pressor cells were present in the spleen cell suspension (Fig.1B). Treatment of the spleen cell donors with T4N5 lipo-somes after UV irradiation prevented the development ofsuppressor cells; transfer of spleen cells from these animals

A. DHS

IMM UV Liposomes

Hi

T4N5

+ + Hi

+ + T4N5

4 8 1 2 1 6 20 24 28

B. SPLEEN CELL TRANSFER

Spleen Cell Donors

None (Negative Control)None (Positive Control)1MM UV Liposomes

+ - Hi+ T4N5

+ + Hi+ + T4N5

to normal mice did not affect the magnitude of their DHSresponse to Candida.Although unlikely, it is conceivable that the T4N5 lipo-

somes could be blocking the release or action of the immu-nosuppressive mediators produced by UV-irradiated cells(3-5) or activating an immunostimulatory mechanism, ratherthan acting at the level of DNA repair. To test these possi-bilities, control experiments were performed in which theT4N5 liposomes were applied to ventral, unirradiated skin orwere applied to dorsal, UV-irradiated skin 2 days after UVtreatment. Application of T4N5 liposomes on ventral (unir-radiated) skin had no effect on UV-induced suppression ofthe DHS response: 51% suppression in mice treated onventral skin with T4N5 liposomes versus 57% in UV-irradiated mice and 50%o in mice treated on ventral skin withliposomes containing heat-inactivated enzyme. In a secondexperiment, mice were exposed to UV radiation and T4N5liposomes were applied immediately or 2 days after UVradiation; all groups were immunized with Candida 5 daysafter UV radiation. As before, applying T4N5 liposomesimmediately after UV irradiation completely restored theDHS response (0%1 suppression). In contrast, applying theactive liposomes 2 days after UV radiation had no effect onUV-induced suppression (68%) compared with mice givenUV radiation alone (68% suppression) or given inactiveliposomes immediately (70% suppression) or 2 days afterUVirradiation (72% suppression).UV Dose-Response. It is also possible that at different doses

of UV radiation, different mechanisms might be involved inimmunosuppression. We therefore performed a dose-response study to determine whether liposome treatmentwould completely restore the DHS response across a widerange of UV doses. As illustrated in Fig. 2, the T4N5liposomes abrogated the UV-induced suppression ofDHS atall doses ofUV radiation tested. In contrast, the response ofmice treated with liposomes containing heat-inactivated en-donuclease did not differ significantly from that ofmice givenUV irradiation without liposomes as determined by linearregression analysis.

Effect of T4N5 Liposomes on UV-Induced Inhibition of theElicitation ofDHS toCad. Previous studies demonstratedthat exposing mice to UV radiation after immunization also

100 r

80

j.5 i 60

-I1I---4 8 12 16 20 24 28 f

Footpad Swelling SEM C 40(10-2 mm) at

FIG. 1. Effect ofT4N5 liposomes on suppression of induction ofDHS to Candida. C3H mice were exposed to UV-B radiation (5kJ/m2). Liposomes containing T4 endonuclease V (T4N5) or heat-inactivated (HI) endonuclease were applied to UV-irradiated skinimmediately thereafter. On day 3, mice were immunized s.c. at anunirradiated site with 2 x 107 C. albicans (IMM). All mice werechallenged 9 days later by footpad injection of Candida antigen;footpad swelling was measured 24 h later. (A) Numbers in paren-theses = % suppression, calculated by Eq. 1, relative to the appro-priate control group (e.g., for test group treated with HI liposomesafter UV irradiation, positive control was that treated with HIliposomes without UV, etc.). (B) Spleen cells from these mice wereinjected i.v. into normal recipients, which were immediately immu-nized with C. albicans; recipients were challenged 9 days later.Numbers in parentheses = % suppression calculated by Eq. 1,relative to positive control group (mice immunized but not injectedwith spleen cells).

20

0

0300 UV

* uv-I* UV-HI

A UV-T4N5 0

a U~~~~~0 *

A

t

5 10

UV-B Dose (kJ/m2)

FIG. 2. UV dose-response for suppression of DHS: Effect ofT4N5 and heat-inactivated (HI) liposomes. Data are pooled resultsfrom eight separate experiments; results are expressed as UV dose(logarithmic scale) versus % suppression of DHS to Candida (cal-culated by Eq. 1) relative to appropriate control groups (see legendto Fig. 1 and Materials and Methods for details). Each data point iscalculated from mean footpad swelling of five test and five controlmice. Lines were obtained by linear regression analysis. ---, UV;

, UV-HI; --, UV-T4N5.

7518 Immunology: Kripke et A

Proc. Natl. Acad. Sci. USA 89 (1992) 7519

reduces the DHS response; however, higher doses of UVradiation are required, and suppressor lymphocytes do notseem to be involved. To determine whether T4N5 liposomeswould also prevent this form of UV-induced immunosup-pression, experiments were performed in which mice weretreated with UV radiation and liposomes 5 days after immu-nization. Treatment with T4N5 liposomes, but not liposomescontaining heat-inactivated enzyme, completely abrogatedthe inhibitory effect ofUV radiation on elicitation of the DHSresponse (data not shown).

Effect ofT4N5 Liposomes on UV-Induced Suppression of theInduction of CHS. It has been postulated (6) that UV-inducedsystemic suppression of CHS in the mouse is mediated byurocanic acid, which, when it absorbs UV radiation, under-goes a trans-to-cis isomerization. If this hypothesis is correct,an increase in the repair of UV-induced pyrimidine dimers inDNA should not abrogate UV-induced suppression of CHS.To test this possibility, we examined the effect of T4N5liposomes on UV-induced suppression of the induction ofCHS to DNFB. Mice were exposed to a single dose (10kJ/m2) of UV-B radiation on shaved dorsal skin. Three dayslater, they were sensitized by application of DNFB ontoshaved ventral skin. Some groups of mice were treated withT4N5 liposomes or liposomes containing heat-inactivatedendonuclease on UV-irradiated skin immediately after irra-diation; unirradiated control groups were also treated withboth liposome preparations. As shown in Fig. 3A, treatmentof UV-irradiated mice with active liposomes significantlyincreased the CHS response to nearly the level observed inunirradiated, T4N5 liposome-treated animals (P = 0.027 forUV T4N5 group vs. UV heat-inactivated group; P = 0.036 forUV T4N5 group vs. UV group), but the response was notcompletely restored (P = 0.04 for UV T4N5 group vs.nonirradiated T4N5 group).

In a similar experiment using FITC as a contact sensitizer,the CHS response was also increased in UV-irradiated T4N5-treated mice but not completely restored; however, thedifference between the UV T4N5-treated group and thenonirradiated T4N5-treated group was not statistically sig-nificant (Fig. 3B). Transfer of spleen cells from the UV-irradiated, FITC-sensitized mice to normal, syngeneic recip-ients confirmed the presence of suppressor cells that inhib-ited the induction of CHS to FITC; such suppressor cellscould not be detected in mice given T4N5 liposomes imme-diately after UV irradiation (Fig. 3C).

Effect of T4N5 Liposomes on the Number of PyrimidineDimers. T4N5 liposomes were shown previously to reducethe number of pyrimidine dimers in DNA isolated from theepidermis of UV-irradiated hairless mice (14). These animalshave an endogenous DNA repair mechanism that removes upto 40% of the cyclobutane pyrimidine dimers from epidermalDNA during the first 6 h after UV irradiation (17), and T4N5liposomes reduce the number remaining by an additional 40%during that time period (14). To ensure that the liposomeswere also increasing the repair of such lesions in the epider-mis of the inbred C3H mice used in these experiments,pyrimidine dimers were measured 6 h after UV irradiation inmice taken at random from treatment groups of the experi-ments depicted in Fig. 2. The pooled results of these mea-surements are presented in Fig. 4. The number of pyrimidinedimers per 106 bases was directly proportional to the UV dose(11 ± 1 dimers per 106 bases per kJ per m2 ofUV-B; r = 0.986)and is expressed as a percentage of the dimers in epidermalDNA from mice treated with UV radiation only. Dimers inthe epidermis of mice treated with inactive liposomes rangedfrom 97% to 135% of that in control animals; in contrast,dimers in the epidermis of mice treated with T4N5 liposomesranged from 44% to 74% of the control values across therange of UV doses tested. These results demonstrate that 6h after UV irradiation, fewer pyrimidine dimers were present

A. CHS Response to DNFBIMM UV Liposomes

HiT4N5

+ + Hi

+ + T4N5

B. CHS Response to FITCIMM UV Liposomes

+ Hi+ T4N5

+ + Hi+ + T4N5

C. Spleen Cell TransferNone (Negative Control)None (Positive Control)IMM UV Liposomes+ Hi+ - T4N5

+ + Hi

+ + T4N5

_ r~~~~~~~~53i

119)

0o .8 16 24 32-- -.. ............

t79:

(67 23i

o 2 4 10 12 14 16

__.. I. --

o 2 4 6 8 1O 12 14 16

Ear Swelling.: SEM(10-2 mm)

FIG. 3. Effect ofT4N5 liposomes on UV-induced suppression ofthe induction ofCHS. C3H mice were exposed to UV-B radiation (10kJ/m2) on shaved dorsal skin and immediately treated with liposomescontaining T4 endonuclease V (T4N5) or heat-inactivated (HI) en-zyme. Three days later, they were painted with DNFB or FITC onshaved ventral skin. Five days after immunization, they were chal-lenged with the same hapten on the ears; ear swelling was measured24 h later. Numbers in parentheses = % suppression (calculated byEq. 1) relative to the appropriate control group (A and B, see legendto Fig. 1 and Materials and Methods for details). Spleen cells frommice in B were transferred to normal recipients i.v., which wereimmediately immunized by epicutaneous application of FITC. Re-cipients were challenged 5 days later; ear swelling was measured 24h after challenge. Numbers in parentheses = % suppression relativeto positive control group (C).

in the epidermis of mice treated with T4N5 liposomes than inuntreated mice or in mice treated with inactive liposomes.

DISCUSSIONBased on these studies, we conclude that UV-induced sup-pression of the DHS response to Candida in C3H mice istriggered entirely by the DNA-damaging effects of UV radi-ation. We have attempted to rule out other possible expla-nations for the effect of the T4N5 liposomes by using controlliposome preparations containing heat-inactivated endonu-clease and by performing control experiments in whichliposomes were applied to unirradiated skin or were applied2 days after UV irradiation. Because the endonuclease usedto repair DNA damage is highly specific for pyrimidinedimers, these lesions are the most likely initiators of thisimmunosuppressive effect of UV irradiation. Regardless ofthe specific DNA lesion involved, however, the initiation ofsuppression of the DHS response by UV radiation can beaccounted for entirely by DNA damage, and there is no

Immunology: Kripke et al.

+

Proc. Natl. Acad. Sci. USA 89 (1992)

r

125

E

CD

0

100

75

50 t

--- ---------t-----------+---------- t--0 4

o

*UV-T4N5

a I I a a . . *

25

0

5 101

UV-B Dose (kJ/m2)

FIG. 4. Repair of pyrimidine dimers in mouse skin treated withT4N5 liposomes. Some C3H mice from the experiments depicted inFig. 2 were killed 6 h after UV-B irradiation and DNA from theepidermis was analyzed for pyrimidine dimers by the endonucleasesensitive-site assay. Dimer frequency is plotted vs. UV dose on alogarithmic scale. The dimer frequency in epidermal DNA fromUV-irradiated, control mice was compared to that from UV-irradiated mice treated with active (T4N5) (e) or heat-inactivated(HI) (o) liposomes. Each point represents the average of two to sixmeasurements ± SEM, and the lines were obtained by linearregression analysis.

evidence that other molecules altered by UV radiation, suchas urocanic acid, can by themselves trigger photoimmuno-suppression.Systemic suppression of CHS in the mouse after exposure

to UV-B radiation is thought to involve mechanisms otherthan those responsible for suppression of DHS (21). Ourstudies indicated, however, that systemic suppression of theCHS response in UV-irradiated C3H mice was also mediatedmainly by DNA damage in the form of pyrimidine dimers.Unlike the studies with DHS, the CHS response could not berestored completely by treatment with liposomes. A smallcomponent of the suppression, comprising 10-25%, was notreversed by treatment with T4N5 liposomes, suggesting thatan additional mechanism contributes to UV-induced sys-temic suppression of the CHS response.One question raised by our studies is why the T4N5

liposomes cause complete, or nearly complete, restoration ofimmunological function, while they appear in Fig. 4 to repaironly 40-50%o ofthe pyrimidine dimers remaining in epidermalDNA. Most likely, this is due to the fact that in the studiesdepicted in Fig. 4, dimers were measured 6 h after UVirradiation, whereas immunization occurred 3 days later. Inthese experiments, DNA repair measurements were made atthis early time point to avoid the complication introduced bysubsequent UV-induced epidermal hyperplasia. However,we performed an experiment in hairless mice under the sameconditions ofUV irradiation, in which dimers were measuredin DNA prelabeled with [3H]thymidine in vivo. This studydemonstrated that repair of 80%6 of the dimers occurred overa 30-h period after exposure to UV-B radiation (5 kJ/m2)(D.Y., unpublished data), confirming that complete excisionrepair of dimers occurs slowly, over a period of days.Alternatively, target DNA involved in UV-induced immuno-suppression may be repaired more rapidly than other DNAregions, since it is known that actively transcribed DNA isrepaired preferentially after UV irradiation (22).Our experiments demonstrated that skin treated with active

T4N5 liposomes exhibited repair of -z50%o of the remainingdimers, over a wide range of UV-B doses. This result is

puzzling because one would expect a larger proportion of thedimers to be repaired at the lower UV-B doses where fewerdimers are induced. Perhaps not all regions of the genome areequally accessible to T4 endonuclease V; alternatively, someother factor released or activated in proportion to the UV dose(e.g., a DNA ligase or helicase) may become rate limiting afterT4 endonuclease V has incised all the remaining dimers.These studies are consistent with those carried out in the

opossum, which demonstrated by means ofphotoreactivationthat DNA damage in the skin, and particularly pyrimidinedimers, are primarily responsible for generating systemicsuppression ofCHS (10). The target cell ofUV irradiation hasnot been identified; however, the keratinocyte seems to be themost likely candidate. These cells produce a variety of cyto-kines involved in immune and inflammatory reactions (23),and they release immunosuppressive mediators after exposureto UV radiation in vitro (3-5). Therefore, we propose thatUV-induced damage to the DNA of keratinocytes triggers thesynthesis and release of the immunosuppressive cytokinesthat inhibit the DHS and CHS responses and cause theinduction of suppressor T lymphocytes.The studies also demonstrate that the use of liposomes to

deliver lesion-specific DNA repair enzymes to the epidermisin situ provides an effective means ofpreventing UV-inducedimmunosuppression. In principle, this approach could beapplied to the study of other biological end points induced byDNA damage and may have practical applications in pre-venting pathological effects resulting from damage to DNA.

This work was supported by National Institutes of Health GrantsR5 ES04875 (National Institute of Environmental Health Sciences),RO1-CA52457 (National Cancer Institute), R44-CA52401 (NationalCancer Institute), and CA16672 (National Cancer Institute).

1. Parrish, J. A., Kripke, M. L. & Morison, W. L., eds. (1983) Pho-toimmunology (Plenum, New York).

2. Kripke, M. L. (1984) Immunol. Rev. 80, 87-102.3. Schwarz, T., Urbanska, A., Gschnait, F. & Luger, T. A. (1986) J.

Invest. Dermatol. 87, 289-291.4. Kim, T.-Y., Kripke, M. L. & Ullrich, S. E. (1990) J. Invest.

Dermatol. 94, 26-32.5. Ullrich, S. E., McIntyre, B. W. & Rivas, J. M. (1990) J. Immunol.

145, 489-498.6. DeFabo, E. C. & Noonan, F. P. (1983) J. Exp. Med. 1S7, 84-98.7. Ross, J. A., Howie, S. E. M., Norval, M. & Maingay, J. P. (1987)

J. Invest. Dermatol. 89, 230-233.8. Harriott-Smith, T. G. & Halliday, W. J. (1988) Clin. Exp. Immunol.

72, 174-177.9. Norval, M., Gilmour, J. W. & Simpson, T. J. (1990) Photoderma-

tol. Photoimmunol. Photomed. 7, 243-248.10. Applegate, L. A., Ley, R. D., Alcalay, J. & Kripke, M. L. (1989)

J. Exp. Med. 170,1117-1131.11. Ananthaswamy, H. & Fisher, M. S. (1981) Cancer Res. 41, 1829-

1833.12. Gordon, L. & Haseltine, W. (1980)J. Biol. Chem. 255,12047-12050.13. Yarosh, D. (1992) in Liposome Dermatics, eds. Braun-Falco, O.,

Korting, H. & Maibach, H. (Springer, Heidelberg), pp. 151-155.14. Yarosh, D. B., Tsimis, J. & Yee, V. (1990) J. Soc. Cosmet. Chem.

41, 85-92.15. Yarosh, D. B., Kibitel, J., Green, L. & Spinowitz, A. (1991) J.

Invest. Dermatol. 97, 147-150.16. Ceccoli, J., Rosales, N., Tsimis, J. & Yarosh, D. B. (1989)J. Invest.

Dermatol. 93, 190-194.17. Yarosh, D. & Yee, V. (1990) J. Photochem. Photobiol. 7,173-179.18. Freeman, S., Blackett, B., Moneteleone, D., Setlow, R., Suther-

land, B. & Sutherland, J. (1986) Anal. Biochem. 158, 119-129.19. Denkins, Y., Fidler, I. J. & Kripke, M. L. (1989) Photochem.

Photobiol. 49, 615-619.20. Thomas, W. R., Edwards, A. J., Watkins, M. C. & Asherson,

G. L. (1980) Immunology 39, 21-27.21. Kripke, M. L. & Morison, W. L. (1986) J. Invest. Dermatol. 86,

543-549.22. Bohr, V., Okumoto, D. & Hanawalt, P. (1986) Proc. Nat!. Acad.

Sci. USA 83, 3830-3833.23. Luger, T. A. & Schwarz, T. (1990) J. Invest. Dermatol. 95, loos-

104s.

7520 Immunology: Kripke et al.

I