Light and Chemical Carcinogens in Human ...

Overlapping Pathways for Repair of Damage from Ultraviolet Light and Chemical Carcinogens in Human Fibroblasts Alex J. Brown, Ted H. Fickel, James E. Cleaver, et al. Cancer Res 1979;39:2522-2527.

Updated version

E-mail alerts Reprints and Subscriptions Permissions

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/39/7_Part_1/2522

Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at [email protected]. To request permission to re-use all or part of this article, contact the AACR Publications Department at [email protected].

Downloaded from cancerres.aacrjournals.org on May 31, 2013. © 1979 American Association for Cancer Research.

[CANCER RESEARCH 39, 2522-2527, July 1979] 0008-5472/79/0039-0000$02.0O

Overlapping Pathways for Repair of Damage from Ultraviolet Light and Chemical Carcinogens in Human Fibroblasts' Alex J. Brown,2 Ted H. Fickel,3 James E. Cleaver,4 P. H. M. Lohman, M. H. Wade, and Raymond Waters TheUniversityof Tennessee-Oak RidgeGraduateSchoolof BiomedicalSciences(A. J. B.) and BiologyDivision,Oak RidgeNationalLaboratory,Oak Ridge, Tennessee37830 (A. J. B., R. WI; Laboratoryof Radiobiology,Universityof California,San Francisco,California94143(T. H. F., J. E. C.); and Medische Biologische Laboratorium, TNO, Rijswijk, The Netherlands (P. H. M. L., M. H. W.)

ABSTRACT DNA excision repair was measured in cultured human fibro blasts after single or dual treatments with ultraviolet radiation, 4-nitroquinoline I -oxide, or N-acetoxy-2-acetylaminofluorene. Three approaches were used to monitor repair: unscheduled DNA synthesis, measured by autoradiography; repair replica tion, measured by the incorporation of a density-labeled DNA precursor into repaired regions; and excision of ultraviolet endonuclease-sensitive sites. When a single repair-saturating dose of one of the three carcinogens was administered, little stimulation of unscheduled DNA synthesis or repair replication could be observed by additional treatment with one ofthe other carcinogens. In no instance was total additivity of repair ob served. These observations were confirmed by showing that the excision of endonuclease-sensitive sites produced by ultra violet damage (i.e. , pyrimidine dimers) was inhibited by expo sure to 4-nitroquinoline 1-oxide and N-acetoxy-2-acetylami nofluorene. The data indicate that the repair of lesions induced by these substances may have common rate-limiting steps, a conclusion previously indicated by the repair deficiency in xeroderma pigmentosum cells in which a single mutation elim mates the repair of damage caused by each of these agents. INTRODUCTION The treatment of human fibrobbasts with mutagens or carcin ogens and the subsequent monitoring of DNA damage and its repair have been the subject of considerable research (7, 18). These studies generally involve exposing cells to a single substance and examining the fate of any DNA lesions induced. The environment, however, often contains several carcinogens, and an organism may be exposed to more than one agent simultaneously. How these agents may interact to influence DNA repair is important, especially in attempts to estimate permissible environmental levels of carcinogens. In human fibrobbasts, 2 broad classifications of DNA damage and its excision repair have been made. One kind of damage is repaired relatively quickly by the insertion of a small number of bases into the repair patch (19); the other requires a longer time for repair, and the patch size involved is considerably larger (19). The first process operates in cells from patients , Supported

by the

Division

of Biomedical

and

Environmental

Research,

United

States Department of Energy, partly under Contract W-7405-eng-26 with the Union Carbide Corporation, and the Medlsche Blologlsche Laboratorium, TNO, Rljswljk, The Netherlands. 2 RecIpient

of

Grant

GM

1 974-09

from

the

National

Medical Sciences, NIH. 3 RecipIent

of NIH

4 To

requests

whom

Training for

of

in less than additive amounts of repair because of overlapping

of the repair pathways for these agents. This has been ob served in human cells exposed to a combined dose of UV and aflatoxin B (21 ) and in V79 hamster cells exposed to UV light and AAAF (2). Other investigators, however, have reported apparent additivity of repair in human cells exposed to UV and AAAF (1). Because of the varying results obtained in these investigations, we repeated the experiments for 3 carcinogens (UV, AAAF, and 4NQO) that form irreparable damage in XP cells (19), using each of the 3 possible pairs to measure UDS and repair replication. We also measured the excision of UV endonuclease-sensitive sites in cells exposed to UV and 4NQO. We concludedthat saturationof repair by a high dose of one agent restricts the induction of more repair by a dose of another agent. MATERIALS AND METHODS UDS. Normal, nonembryonichuman skin fibroblasts(desig nated E-1I ) were plated at 105/35-mm Linbro well containing a glass coverslip and incubated in Eagle's minimal essential medium plus 15% fetal calf serum for 2 days. Cells were prelabeled for 15 mm with 10 @tCi [3H]dThd per ml (specific activity, 60 Ci/mmol) to label S-phase cells heavily and to ensure their identification in autoradiographs. The cells were then irradiated without medium at 254 nm UV at an incident dose rate of 1.3 J/sq m/sec. Then AAAF (dissolved in dimethyl sulfoxide; a gift of A. J. Miller) or 4NQO (dissolved in ethanol; a gift of J. Epler) was added after dilution into larger (at least bOx) volumes of Eagle's minimal essential medium without serum. Dilution in medium with serum was considered inadvis able because of the possibility of reactions with serum proteins

General

5 The Grant reprInts

5TO1GMOD829. should

be

addressed.

Received May 22, 1978; accepted March 28, 1979.

2522

Institute

with XP5 (19), whereas the â€˜ â€˜long-patch repair' â€˜ is defective in most of these cells (5, 6, 19). Attempts have been made to classify DNA-damaging carcinogens by their long or short repair processes (6, 19). For instance, the damage induced by Uv bightand AAAF is repaired in a long-patch manner, and its repair is defective in XP cells (22), whereas the damage in duced by 4NQO is repaired in both long- and short-patch fashion (13, 19); only the bong-patch repair is defective in XP cells (6, 19). The regulatory factors that regulate the maximum amount of repair that normal cells can perform after damage from UV light and chemical carcinogens may be the same factors that are affected by the repair defect(s) in XP cells. Exposure of normal cells to several agents simultaneously should therefore result

abbreviations

used

are:

XP,

xeroderma

plgmentosum;

AAAF,

N-acetoxy

2-acetylamlnofiuorene; 4NQO. 4-nltroqulnollne 1-oxIde; UDS, unscheduled DNA synthesis; [3H)dlhd. tritlated thymidine; BrdUrd, bromodeoxyurldine; FdUrd, fluorodeoxyundlne.

CANCERRESEARCHVOL. 39


DNA Repair in Human Fibrob!asts that would reduce the effective concentration reaching DNA. fractions, the DNA was precipitated with trichboroacetic acid, This is particularly important with AAAF, which is extremely and radioactivity was determined with a liquid scintillation spec reactive and does not require activation (1) and may contribute trometer. The ratio of[3H]dThd activity banding with 32P-labebed parental DNA was taken as an estimate of repair replication. to differences between the absolute amounts of repair ob served in our experiments and those of others in which serum The value obtained after irradiation with UV, 20 J/sq m (ap was present during AAAF exposure (1). The medium was then proximately 3 x 1O@cpm 3H/800 cpm 32Pin comparison to removed, and the cells were washed once with 0.9% NaCI 200 to 400 cpm 3H/800 cpm 32Pfrom unirradiated cells) was solution (10) and reincubated for 1 hr in medium containing 10 normalized to 100% to facilitate comparisons between experi j@Ci[3H@dThdper ml. The cells were then incubated for an ments done on different occasions. additional hr with fresh medium containing 10@ M unlabeled UV Endonuclease Assay. Monolayers of normal, nonem dThd. After this medium was removed, the cells were fixed to bryonic human skin fibroblasts (designated AH) (16) were the coverslips with 25% acetic acid in ethanol. The coverslips obtained by adding about 1O@cells to growth medium that was [3HJdThdper ml (17 Ci/ were removed from the wells and glued to the microscope supplemented 72 hr later with 0.3 @sCi mmol). The media were removed 24 hr later, the dishes were slides, which were then dipped in photographic emulsion (NTB 2) and allowed to stand for 4 to 6 days. After developing, the rinsed once with phosphate-buffered saline, and the cells were slides were stained with Giemsa, coverslips were placed over irradiated. Immediately afterwards, AAAF or 4NQO dissolved them, and grains from 40 randomly selected nuclei were in dimethyl sulfoxide (final concentration of solvent was 1% or counted. Hydroxyurea was not required in these experiments less in media containing serum) was added to medium on the because the S-phase cells (20 to 40% of the population) were dishes for 1 hr. The media were then replaced, and, at times black with grains because of the 15-mm prebabel, and the up to 24 hr, samples were taken by rinsing the plates once number of cells entering S during the 1-hr labeling period after with phosphate-buffered saline and storing them at _800 for high UV and chemical doses would be negligible. S-phase cells at beast 30 mm. Cells were then lysed. DNA was extracted, could therefore be recognized and ignored in the grain count incubated with UV endonuclease (4), and centrifuged in alka ing, which was confined to G1 and G2 cells undergoing UDS. line sucrose gradients at 40,000 rpm for 80 mm (16). Molecular The average number of grains in untreated cells was 0.6 to weight of unirradiated DNA was 65 to 75 x 106, DNA samples were analyzed on alkaline sucrose gradients 1.7, and in cells exposed to 20 J/sq m or more of UV radiation, it was 30 to 75. In each experiment, the value of UDS at 20 J/ to determine the number of single-strand breaks, i.e. , the sq m UV was normalized to 100% to facilitate comparisons number of UV endonuclease-sensitive sites introduced by ir between experiments done on different occasions. radiation, as described previously (16, 26). Investigators at 2 Repair Replication via Density Labeling. Normal, nonem of the laboratories collaborating in this study (Oak Ridge Na bryonic human skin fibroblasts (designated HSBP) (1 1), pre tional Laboratory and Medische Biologische Laboratorium), labeled with 0.5 Ci 32P per ml, were plated in 100-mm dishes who used micrococcal UV endonuclease preparations made at 2 x 105/dish and incubated in Dulbecco's modified Eagle's separately, estimated the number of UV endonuclease-sensi minimal essential medium plus 10% fetal calf serum at 37Â°in tive sites introduced into DNA to be 2.5 sites/i O@daltons at a 5% CO2 atmosphere. Forty hr later, BrdUrd (3 jig/mI) and 10 J UV per sq m. Investigators at the Laboratory of Radio FdUrd (0.25 @g/ml)were added for 1 hr. Dishes were then biology obtained values of 1.4 sites/i O@daltons at i 0 J/sq m rinsed, and mutagen treatment was carried out in the presence using T-4 UV endonuclease V (a gift of E. C. Friedberg) (28) of BrdUrd and FdUrd. Cells were irradiated with UV light after and of 2.0 sites/i 0@daltons at i 0 J/sq m using thin-layer being rinsed with warm phosphate-buffered saline (10). AAAF chromatography (9). These values indicate that the dosimetry and 4NQO were administered in medium without serum and in this investigation was consistent from one laboratory to incubated at 37Â°for 30 mm. Afterwards, the dishes were another, and the incident dose and dose rate at each were incubated at 37Â°for 3 hr in medium plus BrdUrd, FdUrd, 2.5 used without correction. DNA isolated from cells exposed to mM hydroxyurea,and 10 @Ci [3HldThdper ml (55 Ci/mmol). In 4NQO or AAAF alone was insensitive to attack by UV endo these experiments, hydroxyurea was used to improve the mes nuclease under the above conditions. olution of the isopyknic gradients, but the results were essen RESULTS tially similar to those obtained in the autoradiographic experi ments, indicating thatthe drug did not bias results significantly. Low doses of UV radiation are known to induce pyrimidine In experiments to determine the amount of drug bound to DNA, dimemslinearly with dose (8). However, some chemical carcin cells were exposed to various concentrations of [3HJ4NQO(28 ogens, such as 4NQO, require activation before causing DNA mCi/mmob; a gift of V. Kawazoe) prepared as described above damage (i 4, 25) (AAAF is already in activated form) and for 4NQO, and the DNA was isolated as described below and undergo many competing and branching reactions with media and cellular components; a linear relationship between expo purified on isopyknic gradients. Cells were then scraped into 3 ml of cold 0.01 M Tris-HCI, sure dose and damage to DNA cannot be assumed. We there 0.04 M NaCI, and 0.001 M EDTA (pH 8.0) to which 1 ml of 0.5 fore determined the amount of [3HJ4NQOthat bound to DNA at M EDTA, 0.5% Sarkosyl (Geigy Pharmaceuticals, Ardsley, N. various drug doses, using the repair-deficient cell line XPI 2RO V.), and 200 @sg of proteinase-K(E. M. Biochemicabs,Elmsford, (Group A, with less than 5% of normal amounts of UDS and N. V.) per ml (pH 8.0) was added. Each sample was vortexed repair replication of UV and 4N00 damage) (6) to minimize the at high speed for 1 mm and incubated at 37Â°for at least 4 hr. effect of excision of 4NQO-DNA adducts during the exposure CsCl (1 .25 g/mI) was dissolved in each sample, and cells were time. In subsequent experiments, the amount of 3H bound was centrifuged in a No. 40 rotor of a Beckman L5-50 centrifuge at a linear function of dose, which also implies linearity for the 37,000 rpm for 40 hr at 200. Gradients were collected as 40 amount of 4NQO-induced DNA damage (Chart i). JULY 1979


2523

A. J. Brown et a!. o@o z 01 0

>1

100-

7

0 0

U w 0.

@Az 50

> 4

-J 4NQO DOSE(SM)

Chart 1. Binding of [3H)4N00 to the DNA of human excision-deficient xP1 2RO fibroblasts. Cells were treated with [3H]4NQO(28 mCi/mmol) in serum free medium for 45 (0) or 120 (â€¢) mm. DNA was isolated from the cells after treatment with RNaseand pronase and after deproteinization with chloroformamyl isoalcohol. The specific activity was then determined as cpm 3H/[email protected] values are expressed so that a value of 10 corresponds to 1 < 3Hcpm/@gDNA.

In experiments combining UV and 4NQO, the amount of UDS (Chart 2) and of repair replication (Chart 3) approached satu ration at 20 J UV per sq m (24). However, a higher dose of UV (30 J/sq m) was used to ensure saturation to determine whether additional doses of 4NQO could increase repair fur them.Some additivity of UDS (Chart 2) was apparent during a 1-hr labeling period after saturating UV doses plus low (i @sM) 4NQO doses, but the amount of UDS did not approach the sum of 2 separate treatments. After higher 4NQO doses (2 zMand above), some additivity in UDS and repair replication was observed, but only when the UV dose was not saturating (i.e., 5 or 10 J/sq m). In experiments combining UV and AAAF, 5 or 20 @M AAAF alone resulted in amounts of UDS that were the same, indicating that these doses of AAAF saturate the amount of UDS resulting from the drug. Exposure to UV plus AAAF increased the UV saturation level by only i 0 to 20% (Chart 4). Repair replication

measured during the 3 hr after treatment with UV and 2 or i 0 @LMAAAF

also

resulted

in small

increases,

but much

less than

that required for complete additivity (Chart 5). In experiments combining 4NQO and AAAF, the amount of repair replication from 4N00 alone saturated at 4 @zM. Doses of AAAF that resulted in relatively large amounts of repair replication when administered alone did not result in an in crease when combined with saturating doses of 4NQO (Chart 6). These observations indicate that, when high doses of 2 agents are used, the amount of repair observed is less than the sum of that measured from each agent alone. We therefore determined whether similar interference in repair could be detected in measurements of an earlier step in excision repair, the removal of damaged sites. After exposure to a low UV dose (3 J/sq m), about one-half of the UV endonuclease-sensitive sites were removed in 8 to i 0 hr. Exposure to 4NQO inhibited this removal (Chart 7). A dose of 3 @LM 4N00, which is below saturation, had a small effect detectable i 5 to 24 hr after irradiation. A dose of i 0 @sM, which saturates repair for 4N00 alone, completely prevented excision of UV endonuclease-sen sitive sites. Similarly, when nonsaturating doses of UV plus AAAF were used, no inhibitionof the excision of UV endonu clease-sensitive

sites could be detected but, at higher AAAF

doses (i 0 ELM),excision was inhibited distinctly (Table 1). It 2524

/

0 0

10

20

30

UVDOSE (J.m@2)

Chart 2. UDS in normal human fibroblasts (E-1 1) after single or dual treat ments with uv and 4NQO. Cells were treated with UV alone (f'), Uv plus 1 ga.i 4NQO (0). UV plus 5 oM4NQO (â€¢), or uv plus 10 @oM 4NQO (0). Cells were incubated at 37Â°in medium containing [3H]dThd for 1 hr, followed by a 1-hr incubation in unlabeled dThd, and UDS was measured. Data were normalized by taking the saturating level with uv alone (20 J/sq m to be 100%).

z 0

4 U

a w 0. 0. uJ 0.

0

10

0

20

30

UVDOSE (J.m2)

Chart 3. Repair replication In normal human fibroblasts (HSBP) after single or dual treatments with uv and 4NQO. @P-preIabeIed cells were treated with UV alone (Es),UV plus 0.5 @D@4 4NQO (0), UV plus 2 @o,i 4NQO (0), or UV plus 6 @tM 4N00

(â€¢).They were then incubated at 37Â° with hydroxyurea,

[3H]dThd, and

BrdUrd as a density label for 3 hr before DNA precipitation and isopyknic CsCl centrifugation. The ratio of [3H)dThdbanding with parental 32P-labeledDNA was taken as a measure of repair replication. Data were normalized as in Chart 2 and represent averages of at least 2 experiments. Maximum S.D., Â±10%.

0 3

0

10

20 UVDOSE (J.m2)

30

40

Chart 4. UDS In normal human fibroblasts (E-11) after single or dual treat ments with U@and AAAF. Cells were treated with U@Ialone (/@),UV plus 5 pt.i AAAF (0), or uv plus 20 @LM AAAF (â€¢).Incubations and normalization of data were carried out as in Chart 2.



@

-@

DNA Repair in Human Fibroblasts should be noted that, whereas UDS and repair replication measure the incorporation of new bases into DNA and have greatest resolution of small differences at early times after exposure, excision measures the loss of a fraction of damaged sites, and its resolution is greater at long times after irradiation. For this reason, a small inhibition of excision after 3 J/sq m and 3 @LM 4NQO was detectable at long times after exposure by UV endonuclease techniques but not by UDS or repair replication.

z 0 I-

4

U @0 0. 0. 0. 4 0. LU

0.

DISCUSSION UVDOSE(J.m'2) Chart 5. Repair replication in normal human fibroblasts (HSBP) after single or dual treatments with UV and AAAF. Cells were treated wIth UV alone (ti), UV plus 2 @a.i AAAF (â€¢),or Uv plus 10 p@ AAAF (0). Incubatlons and estimation of repair

replication were as in Chart 3.

100 z

In this study, we measured excision repair in DNA of normal human fibroblasts treated singly or with combinations of car cinogens, using UDS as measured by autoradiography, repair replication as measured by isopyknic gradient centrifugation, and excision of UV endonuclease-sensitive sites. Every corn bination that we investigated with UV, 4NQO, and AAAF gave similar results. The amount of repair observed after combined high

@

@i 0 @..

@-.---_, 4NQO

DOSE (p41)

Chart 6. Repair replication in normal human fibroblasts (HSBP) after single or dual treatments with 4NOO and AAAF. Cells were treated with 4N00 alone (a), 4NQO plus 2 tIM AAAF (0). or 4N00 plus 10 @o,.i AAAF (â€¢).Incubations and estimation of repair replication were carried out as in Chart 3.

to â€˜p

5

10

15

20

25

HOURS

@

Chart 7. Removal of UV endonuclease-sensitive sites. Normal human fibro blasts (AH) were irradiated with 3 J/sq m; exposed to 0 (â€¢), 3 pt,i (0), or 10 gLM (A)

4N00

for

1 hr;

and

then

incubated

in

medium

at

37Â°

for

various

times

before Isolation of the DNA and determination of the relative number of endonu clease-sensitive sites.

Table1 Relativenumberof DNAof UVendonuclease-sensitive sitesremainingin doseof humanfibroblasts(AHstrain)irradiatedwitha nonsaturating UVlight (3 J/sq m), exposedto AAAF,and variousconcentrationsof hrFraction grownfor 24 of initial no. ofat24 UV endonuclease sites hrTreatment

SE.Uv

Indi@duaIvalues

0.05UV alone 0.015UV + AAAF (5 ELM) 0.04a+ AAAF (10 pM)

Value from Chart 7. JULY 1979

Mean Â±

0.28,0.198,0.092@

0.19Â±

0.1 2. 0.1 29, 0.181 , 0.167 0.40, 0.545, 0.481

0.10 Â± 0.48 Â±

(saturating)

doses

of 2 agents

was

much

less

than

the

sum of the repairs that resulted from each agent separately. These results indicate that there is considerable overlap in the mechanisms that regulate the total amount of repair performed for each of these agents; this was expected because the repair deficiency in XP cells reduces their capacity to repair damage from each of these agents coordinately (6, i 9). After combined treatments of UV and 4NQO, a small increase of repair by UDS was observed during I hr after treatment at high UV and low 4NQO doses, but this increase diminished at

higher doses of 4NQO. Similar treatments analyzed by repair replication for 3 hr after exposure indicated that some additivity of UV and 4NQO repair occurred at nonsaturating UV doses, but this additivity diminished after higher, saturating UV doses. Regardless of the techniques used, total additivity of the repair seen after single, saturating UV and 4NQO doses was not seen. For example, the UV dose of 20 J/sq m resulted in i 00% repair by UDS and 1 @LM 4N00 resulted in 80% of this, but only i 50% repair was observed after a combined treatment. This was the highest amount of additivity observed. When higher 4NQO doses were used after 20 J/sq m, 120% repair was obtained as measured by both UDS and repair replication. After 20 J UV per sq m and a low 4NQO dose, there was a discrepancy of about 25% between UDS at 1 hr and repair replication at 3 hr. This may have been due to the different lengths of assay times and the types of lesions involved. Two classes of 4NQO damage have been identified: (a) 4NQO purine adducts that are stable during acid hydrolysis but are not repaired in excision-deficient bacteria or XP fibroblasts (i 2, i 3); (b) a 4N00-guanine adduct that is unstable during acid hydrolysis and is repaired in excision-defective bacteria (i 2) and XP fibrobbasts (13). These classes are induced in a ratio of about 4/i (acid stable/acid unstable) (12). The acid unstable lesion is repaired in XP and normal fibroblasts by a short-patch, quick-repair mechanism (13, i 9), whereas the acid-stable product is repaired by a long-patch process, which takes a longer time to remove the lesions (i 3, i 9). Hence, at i hr, the contribution of the short-patch, quick-repair mecha nism to the total repair synthesis would be greater than at 3 hr. This may account for the small difference between the UDS (i hr) and repair replication (3 hr) estimates.

2525


A. J. Brown et a!. After combined treatments of UV and AAAF, the amount of repair was again less than the sum of that resulting from the action of the 2 agents independently. Single treatments with AAAF gave about 40 to 50% of the repair seen by UDS and repair replication after 20 J/sq m. Treatments with both UV and AAAF gave only i i 5% of the amount of repair seen by UV alone. In experiments using UV endonuclease assays to detect the excision of UV-induced damage, the results were consistent with those obtained with UDS and repair replication. High doses of 4NQO (Chart 7) or AAAF (Table i ) interfered with excision of endonuclease-sensitive sites. After combined treatments of 4NQO and AAAF, the clearest example of nonadditivity was observed (Chart 6). At saturating

dosesof 4N00, additionof variousdosesof AAAFresultedin no increase in repair replication, indicating that the repair of damage from these agents completely overlapped and that saturation for one resulted in saturation for the other. This is

not due merely to saturation of available sites for DNA damage, because the maximum repair after 4NQO alone occurred at 4 @LM and above

(Chart

6); but chemical

binding

to DNA was a

taken as supportive evidence because the excision kinetics for pyrimidine dimers in that report (3) is anomalous. These inves tigators found that dimer excision, as measured chromato graphically, was completed within i hr of irradiation; all other reports, including that of Ahmed and Setbow(1), indicate that excision occurs more slowly and progressively over much longer time periods and that little excision is detectable during the first 4â€”5hr after irradiation (i i , i 5, 16, 20, 23, 27). In view of this disparity, we feel we must supend judgment on the significance of the observations of Amacher et a!. (3) and should concentrate on possible explanations for the different conclusions reached by Ahrned and Setbow(1). In our observations (Charts 2 to 5), saturation of UDS and repair replication was just attained at 20 J/sq m, but the convergence of the UV-plus-carcinogen exposure curves was difficult to interpret conclusively without proceeding to higher UV doses to be certain that saturation had been reached. When we used higherdoses of 30 J/sq m, we were able to ascertain that the amounts of UDS and repair replication were less than additive when chemical and UV damages were combined. Consideration of the corresponding curves in the report of Ahmed and Setbow(1) shows that they recorded UDS at only 5, i 0, and 20 J/sq m and that UDS was still increasing at the highest dose. If we had not extended our own observations to higher doses, we would have had difficulty in deciding whether there was complete or only partial additivity, because the repair system for UV alone was not saturated. We therefore believe that our results are consistent with theirs but that their experi ments were not performed at a sufficiently high UV dose to validate their conclusions. Further evidence supporting our interpretation is that their dosimetry with UV endonuclease from Micrococcus luteus indicated a yield of i .8 sites/i O@ daltons at 10 J/sq m in contrast to our 2.5 sites, so that their highest dose of 20 J/sq m corresponds to our dose of i 4 J/sq m; however, subtle differences in endonucleolytic assays make this argument less substantial than the one discussed above. Other interpretations of the differences between Ahmed and Setbow (i ) and ourselves, based on different cell lines or the use of hydroxyurea, are unlikely. Both they and we used

linear function of dose to at least 200 @sM (Chart i). Our results indicate that, at saturating doses of UV, additional exposure to 4NQO or AAAF results in limited increases in the amount of repair. This indicates that most of the rate-limiting elements in the regulation of excision repair of damage from these disparate agents are shared, even though UV predomi nantly damages pyrimidines, and these chemical carcinogens damage purines. When combined doses of 4NQO and AAAF were used, the rate-limiting elements appeared to be corn pletely shared, because no additivity was observed. Our results are consistent with predictions made from the common classi fication of these agents with respect to the irreparability of their damage in XP cells (6, i 9). Our results are also consistent with observations of human cells exposed to UV and aflatoxin (21) and of V79 cells exposed to UV and AAAF (2), in which repair after combined exposure was less than additive and in some cases mutually inhibitory. Our data from the UV endonuclease assay are also consistent with those obtained by the same method in human cells exposed to UV and AAAF (i )@in which primaryhumanfibroblasts,and we obtainedconsistentresults with cultures from 3 separate donors. We also obtained essen a dose of 20 @sM AAAF inhibited the excision of UV endonucle tially similar results using UDS without hydroxyurea and repair ase-sensitive sites to a small extent over a 24-hr period, al though the effect was not measurable at 6 hr. We observed an replication with hydroxyurea. Results with other combinations inhibitory effect at 10 @M AAAF, which was lower than that in of agents (2i ) and another cell type (2) are also consistent with Ahmed and Setbow's experiments (1), because (see â€˜ â€˜Materials our results. and Methods' â€˜) our chemical carcinogen exposures were done In conclusion, we find that repair synthesis following corn in the absenceof serumproteinswhichcan interactwith these bined doses of UV, AAAF, and 4N00 is not totally additive in reactive chemicals and reduce the effective dose to cellular the 2 normal cell lines used. In most instances, repair after a DNA. Mutually inhibitoryeffects between combined doses of saturating dose of UV light was stimulated only about 20% by carcinogens may, in fact, be a general phenomenon, because exposure to chemical carcinogens. These results may indicate inhibition of excision of methylnitrosourea-induced O6-methyl that repair of DNA damage from these agents involves common rate-limiting steps, as was suggested by their classification into guanine by exposure to dimethylnitrosamine has been ob a common group of agents whose repair is defective in XP cells servedin vivo (17). The one exceptional result among various studies with corn (6). Because a small amount of stimulation was observed when the 2 agents in combination damaged different bases in DNA bined doses of UV and chemical carcinogens (Refs. 2 and 2i (pyrimidines by UV and purines by AAAF or 4NQO), there may and the present report) is the study by Ahmed and Setbow(i) reporting complete additivity between apparently saturating be slight differences in some of the steps of repair for these doses of UV and AAAF. A companion paper by Amacher et a!. agents as well as a high degree of overlap. When 2 different (3), which also reported different rate-limiting steps for the agents damaged purines, there appeared to be no stimulation excision of pyrimidine dimers and AAAF lesions, cannot be and hence a higher degree of common regulation for the repair

2526



DNA Repair in Human Fibrob!asts

@

of damage from these agents. These results should be borne in mind when calculating the environmental risks of individual carcinogens. Exposure to low concentrations of a large number of DNA-damaging compounds that are repaired by a common pathway may correspond to exposure to a saturating bevelof one compound and may result in the accumulation of unre paired damage that increases the relative risk of exposure. REFERENCES 1. Ahmed, F. E., and Setlow, A. B. Different rate-limiting steps in excision repair of ultraviolet- and N-acetoxy-2-acetylamlnofluorene-damaged DNA in normal human fibroblasts. Proc. Nat. Acad. Sci. U. S. A., 74: 1548-1552, 1977.

2. Ahmed, F. E., and Setlow, A. B. DNA repair in V-79 cells treated with combinations of ulfraviolet radiation and N-acetoxy-2-acetylaminofluorene. Cancer Res., 37: 3414â€”3419,1977. 3. Amacher, D. E., Elliott, J. A., and Lieberman, M. W. Differences in removal of acetylaminofluorene and pyrimldlne dimers from the DNA of cultured mammalian cells. Proc. Nat. Acad. Si. U. S. A., 74: 1553â€”1 557, 1977. 4. CarrIer, W. L, and Setlow, R. B. Endonucleasefrom Micrococcusluteus which has activity toward ultraviolet-irradiated deoxyribonucleic acid: pun ficatlon and properties. J. Bacteniol., 102: 178â€”1 86, 1970. 5. Cleaver, J. E. Defective repair replication of DNA in xeroderma pigmentosum. Nature (Lond.), 218: 652â€”656,1968. 6. Cleaver, J. E. DNA repair with punines and pynimidines in radiation- and carcinogen-damaged normal and xeroderma pigmentosum human cells. Cancer Ass., 33: 362-369, 1973. 7. Cleaver, J. E., and Bootsma, D. Xeroderma pigmentosum: biochemical and genetic characteristics. Annu. Rev. Genet., 9: 19â€”38,1975. 8. Cleaver, J. E., and Trosko, J. E. Absence of excision of ultraviolet-induced cyclobutane dimers in xeroderma pigmentosum. Photochem. Photobiol., 11: 547-550, 1970.

9. Cook. K. H., and Friedberg, E. C. Measurement of thymine dimers in DNA by thin-layer chromatography. II. The use ofone-dimensional systems. Anal. Biochem., 73: 41 1-418, 1976. 10. Dulbecco, R., and Vogt, M. Plaque formation and isolation of pure lines with pollomyelltis viruses. J. Exp. Med.. 99: 167-182, 1954. 11. Ehmann, U. K., Cook, K. H., and Fnledberg, E. C. The kinetics of thymine dimer excision in ultraviolet Irradiated human cells. Blophys. J., 22: 249â€” 264, 1978. 12. Ikenaga, M., Ichikawa-Ryo, H., and Kondo, 5. The major cause of Inactiva tlon and mutation by 4-nitroquinoline 1-oxide in Escherichia coli: excisable 4N00-punine adducts. J. Mol. Blol., 9: 341 -356, 1975. 13. Ikenaga, M., Takebe, H., and Ishil, V. Excision repair of DNA base damage

in human cells treated with the chemical carcinogen 4-nitroqulnoline 1oxide. Mutat. Res., 43: 415â€”427,1977. 14. lshii, V., and Kondo, S. Differential inactivation of transforming DNA in vitro and in vivo by 4-hydroxyaminoquinoline 1-oxIde. Mutat. Res., 13: 193â€”198, 1971. Mortelmans, K., Cleaver, J. E., Fnledberg, E. C., Paterson, M. C., Smith, B. P., and Thomas, G. H. Photoreactivation of thymine dimers in UV-irradlated human cells: unique dependence on culture conditions. Mutat. Res., 44:

433-466, 1977. 16. Paterson, M. C., Lohman, P. H. M., and Sluyter, M. L. Use of a UV endonuclease from Micrococcus Iuteus to monitor the progress of DNA repair in UV-irradlated human cells. Mutat. Res., 19: 245â€”256,1973. 17. Pegg, A. E. Dimethylnitrosamine inhibits enzymatic removal of 0Â°-methyl guanine from DNA. Nature (Lond.), 274: 182â€”184,1978. 18. Regan, J. D., and Setlow, R. B. Repair of chemical damage to human DNA. In: A. Hollaender (ed.), Chemical Mutagens.â€”Pnlnclplesand Methods for Their Detection, vol. 3, pp. 151-1 70. New York: Plenum Publishing Corp., 1973.

19. Re9an, J. 0., and Setlow, R. B. Two forms of repair In the DNA of human cells damaged by chemical carcinogens and mutagens. Cancer Res., 34: 3318â€”3325,1974.

20. Regan, J. D., Trosko, J. E., and Carrier, W. L. Evidence for excision of ultraviolet-induced pynlmldine dimers from the DNA of human cells in vitro. Biophys. J., 8: 319-325, 1968. 21 . Sarasin, A. R., Smith, C. A., and Hanawalt, P. C. Repair of DNA In human cells after treatment with activated aflatoxin B. Cancer Res., 37: 1786â€” 1793, 1977.

22. Setlow, R. B., and Regan, J. D. Defective repair of N-acetoxy-2-acetylaml nofluorene-induced lesions in the DNA of xeroderma pigmentosum cells. Blochem. Biophys. Res. Commun., 46: 1019â€”1024,1972. 23. Setlow, R. B., Regan, J. D., German, J., and Carrier, W. L Evidence that xeroderma pigmentosum cells do not perform the first step In the repair of ultraviolet damage to their DNA. Proc. Nat. Acad. Sci. U. S. A., 64: 1035â€” 1041, 1969.

24. Smith,C. A., and Hanawalt,P. C. RepairreplicationIn humancells:simplIfIed determination utilizing hydroxyurea. Blochim. Blophys. Acta, 432:336-347,

1976. 25. Tada, M., and Tada, M. Enzymatic activation of the carcinogen 4-hydroxy amlnoqulnoline-1-oxide and Its Interaction with cellular macromolecules. Biochem. Biophys. Res. Commun., 46: 1025-1032, 1972. 26. Waters, R., and Regan, J. D. Recombinatlon of UV-induced pynimidine dimers In human flbroblasts. Biochem. Blophys. Res. Commun., 72: 803â€” 807,1976. 27. Williams, J. I., and Cleaver, J. E. Excision repair of ultraviolet damage in mammaliancells. Evidence for two steps In the excision of pynimidinedimers. Biophys. J., 22: 265-279, 1978. 28. Williams, J. I., and Cleaver, J. E. Removal of T4 endonuclease-sensitive sites from SV4O DNA after exposure to ultraviolet light. Biochim. Biophys. Acta, in press, 1979.

JULY 1979


2527