ORIGINAL ARTICLE

The Total Amount of DNA Damage DeterminesUltraviolet-radiation-induced CytotoxicityAfter Uniformor Localized Irradiation of Human Cells

Kyoko Imoto,nw Nobuhiko Kobayashi,w Sachiko Katsumi,nwYoko Nishiwaki,nwTaka-aki Iwamoto,nwAyaYamamoto,nYukioYamashina,wToshihiko Shirai,w Sachiko Miyagawa,wYoshiko Dohi,z Shigeki Sugiura,yand Toshio MorinnRadioisotope Research Center, wDepartments of Dermatology and zPublic Health, and yMedical Genetics Research Center,Nara Medical University, Kashihara, Nara 634-8521, Japan

We have recently developed a micropore ultravioletirradiation technique. An isopore membrane ¢lter with3 lm diameter pores shields ultraviolet C radiationfrom cultured human ¢broblasts, leading to partial ir-radiation within the cells with an average of about threeexposed areas per nucleus. This study addressed thequestion of whether the spatial distribution of DNA da-mage within a cell nucleus is important in triggeringultraviolet-induced cytotoxicity. We have examinedwhether there are di¡erences in cytotoxicity betweenpartially ultraviolet-irradiated cells and uniformly irra-diated cells after equal amounts of DNA damage wereinduced in the cell nuclei.We ¢rst determined DNA da-mage formation in normal human ¢broblasts using anenzyme-linked immunosorbent assay.We found that 5 Jper m2 ultraviolet irradiation produced an equivalentamount of cyclobutane pyrimidine dimers and (6^4)photoproducts per cell as 100 J per m2 with the mem-

brane ¢lter shield. At those doses, we found that bothtypes of ultraviolet irradiation induced similar levelsof cytotoxicity as assessed by a 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium assay. Both types of ultraviolet-irradiated cellsalso had similar cell-cycle distribution and apoptosis asmeasured by £ow cytometry. Moreover, no signi¢cant dif-ferences in repair kinetics for either type of photolesionwere observed between the two di¡erent ultraviolet treat-ments. Similar results were obtained in Cockayne syn-drome cells that are defective in transcription-couplednucleotide excision repair. Present results indicate that inthe range of photoproducts studied, the spatial distributionof DNA damage within a cell is less important than theamount of damage in triggering ultraviolet-inducedcytotoxicity. Key words: apoptosis/DNA damage/DNArepair/micropore ultraviolet irradiation/ultraviolet cytotoxicity.J Invest Dermatol 119:1177 ^1182, 2002

Short-wave ultraviolet (UV) radiation produces two ma-jor types of photoproducts, cyclobutane pyrimidine di-mers (CPD) and (6^4) photoproducts (6^4PP), betweenadjacent pyrimidine nucleotides on the same strand ofDNA (Cadet et al, 1992). 6^4PP are formed at a rate

15^33% that of CPD (Mitchell, 1988; Clingen et al, 1995; Evenoet al, 1995; Nakagawa et al, 1998). Such DNA damage can result ina permanent arrest of DNA replication and/or transcription,which leads to cell death (Friedberg, 2001). UV-irradiated humancells attempt to remove these helix-distorting DNA lesions bynucleotide excision repair (NER), which consists of creating in-cisions of the damaged strand of the DNA duplex on both sidesof the lesion, releasing the oligonucleotide carrying the lesion,and then ¢lling in the single-stranded gap and ligating the strand(de Boer and Hoeijmakers, 2000). The major NER pathway

operates across the genome (global genome NER), whereas tran-scription-coupled NER preferentially repairs DNA lesions thatblock transcription (Hanawalt et al, 1994). Genetic defects inNER are associated with three diseases characterized by extremesensitivity to UV: xeroderma pigmentosum, Cockayne syndrome(CS), and trichothiodystrophy. CS is defective only in transcrip-tion-coupled NER, and not in global genome NER (van Steegand Kraemer, 1999).The induction of DNA damage activates cellular defense pro-

cesses mediated by accumulation of the tumor suppressor proteinp53 and overexpression of p53-regulated genes, which helps thecell maintain genetic stability (Vogelstein et al, 2000). Indeed,UV-induced DNA damage of an actively transcribed gene blocksRNA polymerase II transcription, which triggers the nuclear ac-cumulation of p53 protein (Yamaizumi and Sugano, 1994) and in-duces cell death by apoptosis (Ljungman and Zhang, 1996).Activated p53 protein inhibits cell-cycle progression by stimulat-ing the expression of p21/WAF1, an inhibitor of cyclin-dependentkinases (Harper et al, 1993; Terada et al, 1995), and induces genesthat regulate DNA repair, which include p48 (Hwang et al, 1999)and p53R2 (Tanaka et al, 2000).We have recently developed a technique that uses a micropore

¢lter mask to produce UV-induced DNA lesions in localizedareas of the cell nucleus (Katsumi et al, 2001). An isopore

Reprint requests to: Toshio Mori, PhD, Radioisotope Research Center,Nara Medical University, Kashihara, Nara 634-8521, Japan. Email: tmori@naramed-u.ac.jpAbbreviations: CPD, cyclobutane pyrimidine dimers; 6^4PP (6^4) photo-

products; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4 -sulfophenyl)-2H-tetrazolium.

Manuscript received April 11, 2002; revised July 19, 2002; accepted forpublication August 8, 2002

0022-202X/02/$15.00 � Copyright r 2002 by The Society for Investigative Dermatology, Inc.

1177

membrane ¢lter with 3 mm diameter pores shields UV radiationfrom cultured human ¢broblasts, leading to partial irradiationwithin the nuclei in the cells with an average of about three ex-posed areas per nucleus. This micropore UV irradiation techni-que, combined with £uorescent antibody labeling, has enabledus to visualize not only the localized DNA damage, but also therepair of damage and the accumulation of repair proteins at da-maged sites (Katsumi et al, 2001;Volker et al, 2001;Wakasugi et al,2002).While performing those studies, we realized that this is avaluable technique that enables us to irradiate partially with UVlarge numbers of cells at one time. So, we decided to apply thistechnique in addressing an important question that has not yetbeen answered, i.e., are there any di¡erences in cytotoxicity be-tween partially UV-irradiated cells and uniformly irradiated cellsafter an equivalent amount of DNA lesions have been induced intheir nuclei? In other words, do the cells with a localized, high-density distribution of DNA damage in cell nuclei show similarcytotoxicity to the cells with a uniform, low-density distribu-tion? High-density, localized photoproducts could theoreticallyexceed a threshold level that triggers the induction of cell deathby apoptosis (Oda et al, 2000) and/or exceed a level of maximalDNA repair (Klocker et al, 1982).Thus, these photoproducts couldinduce an enhanced sensitivity to UV compared with low-den-sity, uniform ones. In contrast, if the total number of DNA le-sions is critical for UV-induced cytotoxicity, there would be nodi¡erence in cell viability between the two types of UV expo-sure.In this study, we found similar cytotoxicity, cell cycle altera-

tions, and removal of photoproducts by DNA repair ability withor without the membrane shield at UVdoses that yielded similaramounts of photoproducts.

MATERIALS AND METHODS

Cells Normal human ¢broblasts (MSU-2) from newborn foreskinwere kindly provided by Dr James E. Trosko (Michigan State University)(Mori et al, 1989). Fibroblasts (CS2OS) from skin of a CS patient(complementation group A) were kindly provided by Dr Mituo Ikenaga(Kyoto University) (Sugita et al, 1982). MSU-2 (passages 13^19) andCS2OS (passages 13^15) cells were used for present experiments. Cellswere cultured in Dulbecco’s modi¢ed Eagle’s medium (Nissui Seiyaku,Tokyo, Japan) supplemented with 10% fetal bovine serum (DainipponPharmaceutical, Osaka, Japan) and antibiotics.

Micropore UV irradiation Micropore (localized) UV irradiation wasperformed essentially as described previously (Katsumi et al, 2001). Brie£y,cells were cultured in dishes for 1 or 2 d. After washing with Dulbecco’sphosphate-bu¡ered saline, cells were carefully covered with a poly-carbonate isopore membrane ¢lter (pore size, 3 mm; diameter, 25 or 47 mm;Millipore, Bedford, MA). Filter-masked cells were then UV-irradiatedwith ¢ve low-pressure mercury lamps (GL-10, Toshiba, Tokyo, Japan;predominantly 254 nm UV) at a dose rate of 1.67 J per m2 per s at adistance of 1.6 m. For whole-cell (uniform) irradiation, cells without a¢lter, were irradiated with a low-pressure mercury lamp at a dose rate of0.42 J per m2 per s at the same distance. Dose rates were monitored usinga UV radiometer (UVR-1,Topcon,Tokyo, Japan).

In situ visualization of localized DNA damage Immediately aftermicropore UV irradiation (100 or 200 J per m2), cells were ¢xed andcellular DNAwas denatured. CPD and 6^4PP were then visualized by animmunologic method using TDM-2 and 64M-2 monoclonal antibodies(Mori et al, 1991; Kobayashi et al, 2001), respectively, as described before(Nakagawa et al, 1998; Katsumi et al, 2001). These antibodies speci¢callybind to CPD or 6^4PP in all dipyrimidine sequences (TT, TC, CT, andTT) (Mori et al, 1991). Nuclear DNA was counterstained with propidiumiodide. Fluorescence microscopy was performed using a Leica DMIRB.The £uorescent images of DNA photoproducts and DNA weresuperimposed using Adobe Photoshop software.

Quantitation of induction and repair of DNA damage Cells werelabeled with 1.85 kBq per ml of [2-14C]thymidine (Amersham, Piscataway,NJ, 2.18 GBq per mmol) in 175 cm2 £asks for 3 d. Cells (42�106) werethen cultured in 100 mm Falcon dishes for 48 h in a radioisotope-freemedium. After washing with Dulbecco’s phosphate-bu¡ered saline, cells

were UV-irradiated partially or uniformly and were then incubated forvarious times to allow repair. Immediately after irradiation or at varioustimes after UV irradiation, cells were harvested by trypsinization. Partiallyirradiated cells, which had been marked on the dish bottom, wereexclusively isolated and harvested using cloning cylinders (diameter,35 mm). Genomic DNA of UV-irradiated cells was puri¢ed using aQIAamp Blood Kit (QIAGEN, Hilden, Germany). DNA concentrationswere calculated from the absorbance at 260 nm, and 14C-radioactivity ofDNAwas measured using a liquid scintillation counter (LSC-5100, Aloka,Tokyo, Japan). CPD and 6^4PP were quantitated by an enzyme-linkedimmunosorbent assay (ELISA) using TDM-2 and 64M-2 monoclonalantibodies. Details of this method have been described previously(Nakagawa et al, 1998). In brief, 96 well polyvinylchloride £at-bottommicrotiter plates, precoated with 0.003% protamine sulfate, were coatedwith sample DNA (1.6 Bq per well for CPD, 32 Bq per well for 6^4PP).In this study, equal amounts of radioactive DNA, instead of equal amountsof DNA calculated from the absorbance at 260 nm, were coated in thewells in order to correct for decreases in photolesions per DNA bypossible DNA replication during the post-UV incubation period; 1.6 and32 Bq DNA corresponded to 15 and 300 ng DNA, respectively, of cellswithout post-UV incubation. The binding of monoclonal antibodies tophotolesions in immobilized DNA in wells (in quadruplicate) wasdetected with biotinylated F(ab0)2 fragment of goat anti-mouse IgG andthen with streptavidin peroxidase. The absorbance of colored productsderived from o-phenylene diamine was measured at 492 nm by TitertekMultiskan Plus MK (Labsystems, Helsinki, Finland). For examining repairkinetics, the percentage of the initial number of photolesions wascalculated at various times after UV irradiation using damage inductionstandard curves.

Measurement of cytotoxicity Cytotoxicity was evaluated using an[3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (MTS) assay (Promega Madison, WI). The MTS assayyields a water-solubleformazan product, which is a great advantage over a[3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] (MTT)assay (Buttke et al, 1993). Cells (106) were cultured in 35 mm dishes for24 h, and then were UV irradiated and harvested by trypsinization.Partially irradiated cells were isolated and harvested using cloningcylinders (diameter, 13 mm). Cells (2�103) after localized or uniform UVirradiation were plated in each well of 96 well tissue culture plates and werecultured for 4 d.Viability was then assessed by the ability of cells to convertMTS into formazan, which was quantitated spectro-photometrically.Cytotoxicity is expressed relative to unirradiated cells.

Measurement of cell-cycle phase distribution and apoptosis by £owcytometry After localized or uniform UV irradiation, cells (2.5�105 or5�105) were plated in 100 mm dishes and were cultured for 4 d. Cells(both £oating and attached) were collected and ¢xed in 70% methanolfor 1 h, and then were treated with propidium iodide (50 mg per ml) andRNase A (1mg per ml) for 30 min. The samples were then analyzed usinga (Becton Dickinson, San Jose, CA) FACScan £ow cytometer for DNAcontent measured as propidium iodide £uorescence. Histograms weregenerated from the £ow cytometric data using Cell Quest software. Thecell-cycle phase distribution was calculated using ModFit LT software.The sub-G1 fraction, which contains lower £uorescent levels than the G1fraction, was considered as the apoptotic cell fraction (Nicoletti et al, 1991;Darzynkiewicz et al, 1992).

RESULTS

Micropore UV irradiation damages DNA within localizedareas of the cell nucleus Localized areas within cell nucleiwere UV irradiated using the micropore UV irradiationtechnique. To ascertain whether localized areas were indeed UVirradiated, two major types of DNA damage were visualizedin situ by the immunologic method. Figure 1 displays thetypical nuclear localization of immuno£uorescent CPD foci inMSU-2 cells, the size and shape of which roughly resemble the¢lter pores. The CPD focus number per nucleus varied from 1 to6 (mean 2.9), depending on the random localization of pores inthe ¢lter. More than 80% of cells had two, three, or four fociper nucleus with a mode of 3. A similar localization wasobserved for 6^4PP foci (data not shown).

Localized UV irradiation with 100 J per m2 causes similarlevels of DNA damage as uniform UV irradiation with 5 Jper m2 We compared the DNA damage formation of cells

1178 IMOTO ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

partially UV irradiated or uniformly UV irradiated using anELISA (Fig 2). After uniform UV irradiation, the formation ofCPD and of 6^4PP increased with increasing UV doses. Bothtypes of DNA damage were also determined after localizedirradiation and compared with those uniformly irradiated. InMSU-2 cells, we found that localized irradiation with 100 or200 J per m2 produced the same amounts of CPD as diduniform irradiation with 5.3 or 8.8 J per m2, respectively.Similarly, localized irradiation with 100 or 200 J per m2 formedthe same amounts of 6^4PP as did uniform irradiation with 4.8or 7.9 J per m2, respectively. Similar results were obtained inCS2OS cells that are defective in transcription-coupled NER.Localized irradiation with 100 or 200 J per m2 produced theequivalent amount of CPD as did uniform irradiation with 4.2or 8.2 J per m2, respectively. Similarly, localized irradiation with100 or 200 J per m2 formed the equivalent amount of 6^4PP asdid uniform irradiation with 5.0 or 9.1 J per m2, respectively.

Cytotoxicity is not signi¢cantly di¡erent between localizedand uniform UV irradiation at levels inducing an equalamount of DNA damage To examine possible di¡erences incytotoxicity between localized and uniform UV irradiationforming equal amounts of DNA damage, we determined cellviability 96 h after irradiation using the MTS assay. Cytotoxicityincreased with increasing UV doses after uniform or localizedirradiation (Fig 3). In MSU-2 cells, we found that localizedirradiation with 100 or 200 J per m2 matches uniform irradiationwith 5.2 or 8.3 J per m2 for cytotoxicity, respectively.These resultsare consistent with those of DNA damage formation (Fig 4A).On the other hand, CS2OS cells showed much higher UVsensitivity than MSU-2 cells after both types of UV irradiationbecause of the defect in transcription-coupled NER (Fig 3). Itwas found that localized irradiation with 100 or 200 J per m2

matches uniform irradiation with 4.5 or 9.2 J per m2 forcytotoxicity, respectively. Importantly, these results are alsoconsistent with those of DNA damage formation (Fig 4B).

No signi¢cant di¡erences in repair kinetics for either type ofphotolesion were observed between localized and uniformUV irradiated cells Localized and uniform UV irradiationcaused similar levels of cytotoxicity, suggesting there were nosigni¢cant di¡erences in DNA repair. To ascertain this, weexamined the di¡erences in DNA damage repair followinglocalized or uniform UV irradiation by an ELISA in MSU-2cells (Fig 5). Two hundred joules per square meter localized UVirradiation or 10 J per m2 uniformUV irradiation were performedin order to produce similar numbers of, but di¡erent distributionof DNA damage within the nuclei. Cells were able to repair 90%of the initial 6^4PP within 2 h and removed 60% of the initial

CPD at 24 h after localized or uniform UV irradiation. Nosigni¢cant di¡erences in repair kinetics for either typephotolesion were observed between the two types of UVtreatment.

Cell-cycle pro¢les and apoptosis do not di¡er signi¢cantlybetween localized and uniform UV irradiation at dosesinducing similar amounts of DNA damage Factorsa¡ecting the UV-induced cytotoxicity detected by the 4 d MTSassay can include inhibition of cell-cycle progression and/or cellkilling. So, we used £ow cytometry to compare cell-cycle phasedistribution and apoptosis 96 h after UV irradiation by thetwo di¡erent exposure techniques in MSU-2 cells (Fig 6 andTable I). With increasing doses of uniform UV irradiation, themost signi¢cant change was the decreased number of cells in theG1 phase, in addition to the increase in number of cells in theG2/M phase. Localized irradiation with 100 or 200 J per m2

displayed patterns of cell-cycle distribution that were similarto those of uniform UV irradiation with 5 or 10 J per m2,respectively. The fraction of sub-G1 (apoptotic cells) increasedwith increasing UV doses after uniform irradiation. The fractionof sub-G1 cells was 13.1 or 16.8% after uniform irradiation with 5or 10 J per m2, respectively. Similar data of 12.2 or 16.3% were

Figure1. Micropore UV irradiation produces CPD in localizedareas of the nucleus.The nucleus on the left has two CPD foci, whereasthe nucleus on the right has three CPD foci.

Figure 2. Determination of DNA photoproducts after localized oruniform UV irradiation using an ELISA. (A) MSU-2 cells. LocalizedUV irradiation with 100 or 200 J per m2 produced the same amounts ofCPD as did uniform irradiation with 5.3 or 8.8 J per m2, respectively. Simi-larly, localized UV irradiation with 100 or 200 J per m2 produced the sameamounts of 6^4PP as did uniform irradiation with 4.8 or 7.9 J per m2,respectively. (B) CS2OS cells. Localized irradiation with 100 or 200 J perm2 produced the equivalent amount of CPD as did uniform irradiationwith 4.2 or 8.2 J per m2, respectively. Similarly, localized irradiation with100 or 200 J per m2 formed the equivalent amount of 6^4PP as did uni-form irradiation with 5.0 or 9.1 J per m2, respectively. Each point showsthe mean 7SD of four independent experiments. L100 and L200 show lo-calized UV irradiation with 100 or 200 J per m2, respectively.

UNIFORM- OR LOCALIZED UV IRRADIATION 1179VOL. 119, NO. 5 NOVEMBER 2002

obtained after localized UV irradiation with 100 or 200 J per m2,respectively.

DISCUSSION

This study addressed the question of whether the spatial distribu-tion of DNA damage within a cell nucleus is important in trig-gering UV-induced cytotoxicity. The micropore UV irradiationtechnique was used to examine e¡ects on the cytotoxicity of

equal amounts of DNA damage between cells given severeDNA damage in localized areas of the nucleus and cells givenmild damage throughout the nucleus. Comparisons of DNA da-mage formation and cytotoxicity between localized and uniformUV irradiation in MSU-2 cells are shown in Fig 4(A), whichclearly shows that 100 J per m2 localized UV irradiation matches5 J per m2 uniform UV irradiation not only for CPD and 6^4PPformation but also for cytotoxicity. Similarly, 200 J per m2 loca-lized UV irradiation matches 8.3 J per m2 uniformUV irradiationnot only for CPD and 6^4PP induction but also for cytotoxicity.These results indicate that 100 J per m2 localized UV irradiationor 5 J per m2 uniform irradiation (or 200 J per m2 localized UVirradiation or 8.3 J per m2 uniform irradiation) produce similaramounts of CPD and 6^4PP photolesions and cause similar levelsof cytotoxicity. In other words, there are no signi¢cant di¡er-ences in cytotoxicity between partially UV-irradiated cells anduniformly irradiated cells, when equal amounts of DNA damageare induced in the cell nuclei. This is quite consistent with theresult showing that no signi¢cant di¡erences in repair kineticsfor CPD or 6^4PP were observed between localized and uniform

Figure 3. Localized UV irradiation with 100 or 200 J per m2

matched e¡ects on cytotoxicity in MSU-2 cells induced by uniformirradiation with 5.2 or 8.3 J per m2, respectively. Similarly, localizedirradiation with 100 or 200 J per m2 matched e¡ects on cytotoxicity inCS2OS cells induced by uniform irradiation with 4.5 or 9.2 J per m2, re-spectively. Cytotoxicity was evaluated using a 4 d MTS assay. Each pointshows the mean7SD of three independent experiments.

Figure 4. Cytotoxicity is not signi¢cantly di¡erent between loca-lized or uniform UV irradiation at doses inducing similar amountsof DNA damage both in MSU-2 and CS2OS cells. Comparisons ofDNA damage formation and cytotoxicity between localized and uniformUV irradiation were made using data from Figs 2 and 3.

Figure 5. No signi¢cant di¡erences in repair kinetics for CPD or 6^4PP were observed following localized or uniform UV irradiation.Two hundred joules per m2 localized UV irradiation (�) or 10 J per m2

uniform UV irradiation (J) were performed in order to produce similarnumbers of DNA damage. CPD (A) and 6^4PP (B) were then quantitatedby ELISA immediately after irradiation and after UV incubation. Eachpoint shows the mean 7SD of four independent experiments.

Figure 6. Cell-cycle pro¢les do not di¡er signi¢cantly following lo-calized or uniform UV irradiation at doses inducing similaramounts of DNA damage. Cell-cycle phase distribution was determined96 h after localized or uniform UV irradiation using £ow cytometry.Typi-cal cell-cycle pro¢les are presented. Upper, uniform UV irradiation; lower,localized UV irradiation.

1180 IMOTO ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

UV exposure methods (Fig 5). Thus, these combined resultssuggest that it is the total amount of DNA damage in the cellnucleus that is critical for UV-induced cytotoxicity, not thelocal density.Cytotoxicity, evaluated 96 h after UVusing the MTS assay, can

result from inhibition of cell-cycle progression and from apopto-tic and nonapoptotic cell death. Flow cytometry data indicatedthat 100 J per m2 localized UV irradiation or 5 J per m2 uniformirradiation displayed similar patterns of cell-cycle distributionand induced apoptosis with similar frequencies. Similar resultswere also obtained from the comparison between 200 J per m2

localized UV irradiation and 10 J per m2 uniform irradiation.Thus, it is suggested that the total amount of DNA damage, notthe spatial distribution of DNA damage within a cell, is an im-portant parameter for the inhibition of cell-cycle progression andthe induction of apoptosis following UV irradiation.In MSU-2 cells, however, cell death by apoptosis (sub-G1 frac-

tion) was less than 20% of the total cytotoxicity (MTS assay) byUV irradiation. In contrast, CS cells are known to be killedmainly by apoptosis after UV, because they are unable to repairDNA lesions that block transcription (Ljungman and Zhang,1996). Comparisons of DNA damage formation and cytotoxicitybetween localized and uniform UV irradiation have been doneusing CS2OS cells (Fig 4B).There were no signi¢cant di¡erencesin cytotoxicity between partially UV-irradiated cells and uni-formly irradiated cells, when similar amounts of DNA damagewere induced in the cell nuclei, indicating that in the range ofDNA damage studied, transcription-coupled NER de¢ciencydoes not play a part in producing a di¡erence in triggering UV-induced cytotoxicity (mainly apoptosis) between the two di¡er-ent UV treatments. Thus again, it is suggested that the spatialdistribution of DNA damage within a cell is less important thanthe amount of damage in triggering UV-induced apoptosis.Ljungman and Zhang (1996) have suggested that it is not thenumber of lesions in bulk DNA that is critical for the inductionof apoptosis following UV irradiation, but rather that it is thepresence and persistence of lesions in the transcribed strands ofactive genes that is responsible for the triggering event. Thus, itis conceivable that localized or uniform UV irradiation at dosesinducing an equivalent amount of DNA damage may producesimilar numbers of DNA lesions not only in bulk DNA but alsoin the transcribed strands of active genes, as similar levels ofapoptosis were induced following both types of UV treatmentsin normal and CS cells.The present results reveal that such a high UVdose as 100 J per

m2 for localized irradiation is necessary to induce the same num-ber of DNA lesions as does 5 J per m2 uniform irradiation. As the

induction rate of cyclobutane thymine dimers is known to be0.0021^0.0024% of total thymines per J per m2 in human skin¢broblasts (Klocker et al, 1982; Niggli and Rothlisberger, 1988), itis calculated that 5 J per m2 uniform irradiation produces an aver-age of 3.1�105 cyclobutane thymine dimers per cell, assumingthat each cell contains 6 pg DNA (Ehmann et al, 1978). Althoughthe micropore irradiation produced 2.9 average damage spots pernucleus, which occupied 10% of the nuclear area, our confocalmicroscopy revealed that DNA photoproducts are produced inonly 5% of the nuclear volume because of attenuated UV pene-tration through membrane pores (data not shown).We used ¢veUV lamps for localized irradiation, of which occupation is 41 cmlong and 41 cmwide.This wide distribution of UVemission mayresult in forming angles to small pores and making attenuatedUV penetration. As a result, 20-fold di¡erences occur in the den-sity of DNA lesions induced by 5 J per m2 uniform or 100 J perm2 localized UV irradiation, Thus, we had predicted that high-density, localized photoproducts could theoretically exceed athreshold level that triggers the induction of cell death by apop-tosis (Oda et al, 2000), and/or would exceed a level of maximalDNA repair (Klocker et al, 1982), which would result in an en-hanced sensitivity to UV compared with low-density, uniformones; however, the fact that we found no di¡erences stronglyargues against these possibilities.In summary, at UV doses that yielded similar amounts of

photoproducts, we found similar cytotoxicity, cell cycle altera-tions and removal of photoproducts by DNA repair ability withor without the membrane shield.These results indicate that in therange of photoproducts studied, the spatial distribution of DNAdamage within a cell nucleus is less important than the amount ofdamage in triggering UV-induced cytotoxicity.

We thank Drs James E. Trosko (Michigan State University) and Mituo Ikenaga(Kyoto University) for human ¢broblasts. This work was supported by Grants-in-Aids for Scienti¢c Research fromThe Ministry of Education, Culture, Sports, Scienceand Technology of Japan to NK and TM, and from The Japanese DermatologicalAssociation to SK.

REFERENCES

de Boer J, Hoeijmakers JHJ: Nucleotide excision repair and human syndromes.Carcinogenesis 21:453^460, 2000

ButtkeTM, McCubrey JA, OwenTC: Use of an aqueous soluble tetrazolium/forma-zan assay to measure viability and proliferation of lymphokine-dependent celllines. J Immunol Methods 157:233^240, 1993

Cadet J, Anselmino C, Douki T,Voituriez L: Photochemistry of nucleic acids in cells.J Photochem Photobiol B Biol 15:277^298, 1992

Clingen PH, Arlett CF, Cole J, et al: Correlation of UVC and UVB cytotoxicitywith the induction of speci¢c photoproducts in T-lymphocytes and ¢broblastsfrom normal human donors. Photochem Photobiol 61:163^170, 1995

Darzynkiewicz Z, Bruno S, Del Bino G, GorczycaW, Hotz MA, Lassota P,TraganosF: Features of apoptotic cells measured by £ow cytometry. Cytometry 13:795^808, 1992

Ehmann UK, Cook KH, Friedberg EC: The kinetics of thymine dimer excision inultraviolet-irradiated human cells. Biophys J 22:249^264, 1978

Eveno E, Bourre F, Quilliet X, et al: Di¡erent removal of ultraviolet photoproductsin genetically related xeroderma pigmentosum and trichothiodystrophy dis-eases. Cancer Res 55:4325^4332, 1995

Friedberg EC: How nucleotide excision repair protects against cancer. Nature RevCancer 1:22^33, 2001

Hanawalt PC, Donahue BA, Sweder KS: Repair and transcription. Collision orcollusion? Curr Biol 4:518^521, 1994

Harper JW, Adami GR,Wei N, Keyomarsi K, Elledge SJ: The p21 Cdk-interactingprotein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases. Cell 75:805^816, 1993

Hwang BJ, Ford JM, Hanawalt PC, Chu G: Expression of the p48 xeroderma pig-mentosum gene is p53-dependent and is involved in global genomic repair.Proc Natl Acad Sci USA 96:424^428, 1999

Katsumi S, Kobayashi N, Imoto K, et al: In situ visualization of ultraviolet-light-induced DNA damage repair in locally irradiated human ¢broblasts. J InvestDermatol 117:1156^1161, 2001

Klocker H, Auer B, Burtscher HJ, Hirsch-Kau¡mann M, Schweiger M: Repairrate in human ¢broblasts measured by thymine dimer excorporation. MolGen Genet 188:309^312, 1982

Table I. Cell-cycle phase distribution and apoptosis do notdi¡er signi¢cantly following localized or uniform UV

irradiation at doses eliciting similar amounts of DNA damage

Uniform UV irradiation

Cell-cycle phase (%) 0 J per m2 5 J per m2 10 J per m2 15 J per m2

Sub-G1 10.773.1 13.174.1 16.872.4 26.178.2G1 76.676.4 72.878.4 68.973.8 52.377.4S 4.771.6 5.172.3 4.971.0 6.870.3G2/M 6.772.3 7.571.9 8.570.6 12.971.1

Localized UV irradiation

Cell-cycle phase (%) 100 J per m2 200 J per m2

Sub-G1 12.273.5 16.372.3G1 74.976.1 66.176.0S 4.470.7 6.473.2G2/M 7.572.0 9.170.8

Each value shows the mean 7 SD of three independent experiments.

UNIFORM- OR LOCALIZED UV IRRADIATION 1181VOL. 119, NO. 5 NOVEMBER 2002

Kobayashi N, Katsumi S, Imoto K, Nakagawa A, Miyagawa S, Furumura M, Mori T:Quantitation and visualization of ultraviolet-induced DNA damage usingspeci¢c antibodies: application to pigment cell biology. Pigment Cell Res 14:94^102, 2001

Ljungman M, Zhang F: Blockage of RNA polymerase as a possible trigger for u.v.light-induced apoptosis. Oncogene 13:823^831, 1996

Mitchell DL: The relative cytotoxicity of (6^4) photoproducts and cyclobutane di-mers in mammalian cells. Photochem Photobiol 48:51^57, 1988

Mori T,Wani AA, D’Ambrosio SM, Chang C-C,Trosko JE: In situ pyrimidine dimerdetermination by laser cytometry. Photochem Photobiol 49:523^526, 1989

Mori T, Nakane M, Hattori T, Matsunaga T, Ihara M, Nikaido O: Simultaneous es-tablishment of monoclonal antibodies speci¢c for either cyclobutane pyrimi-dine dimer or (6^4) photoproduct from the same mouse immunized withultraviolet-irradiated DNA. Photochem Photobiol 54:225^232, 1991

Nakagawa A, Kobayashi N, Muramatsu T, et al: Three-dimensional visualization ofultraviolet-induced DNA damage and its repair in human cell nuclei. J InvestDermatol 110:143^148, 1998

Nicoletti I, Migliorati G, Pagliacci MC, Grignani F, Riccardi C: A rapid and simplemethod for measuring thymocyte apoptosis by propidium iodide staining and£ow cytometry. J Immunol Methods 139:271^279, 1991

Niggli HJ, Rothlisberger R: Sunlight-induced pyrimidine dimers in human skin¢broblasts in comparison with dimerization after arti¢cial UV-irradiation.Photochem Photobiol 48:353^356, 1988

Oda K, Arakawa H,TanakaT, et al: p53AIP1, a potential mediator of p53-dependentapoptosis, and its regulation by Ser-46-phosphorylated p53. Cell 102:849^862,2000

van Steeg H, Kraemer KH: Xeroderma pigmentosum and the role of UV-inducedDNA damage in skin cancer. Mol MedToday 5:86^94, 1999

Sugita T, Ikenaga M, Suehara N, Kozuka T, Furuyama J,Yabuuchi H: Prenatal diag-nosis of Cockayne syndrome using assay of colony-forming ability in ultravio-let light irradiated cells. Clin Genet 22:137^142, 1982

Tanaka H, Arakawa H,Yamaguchi T, et al: A ribonucleotide reductase gene involvedin a p53-dependent cell-cycle checkpoint for DNA damage. Nature 404:42^49,2000

TeradaY, Tatsuka M, Jinno S, Okayama H: Requirement for tyrosine phosphoryla-tion of Cdk4 in G1 arrest induced by ultraviolet irradiation. Nature 376:358^362,1995

Vogelstein B, Lane D, Levine AJ: Sur¢ng the p53 network. Nature 408:307^310, 2000Volker M, Mone MJ, Karmakar P, et al: Sequential assembly of the nucleotide exci-

sion repair factors in vivo. Mol Cell 8:213^224, 2001Wakasugi M, Kawashima A, Morioka H, et al: DDB accumulates at DNA damage

sites immediately after UV irradiation and directly stimulates nucleotide exci-sion repair. J Biol Chem 277:1637^1640, 2002

Yamaizumi M, Sugano T: U.v.-induced nuclear accumulation of p53 is evokedthrough DNA damage of actively transcribed genes independent of the cellcycle. Oncogene 9:2775^2784, 1994

1182 IMOTO ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY