AECL-6627

ATOMIC ENERGY • f f i S L'ÉNERGIE ATOMIQUE

OF CANADA UMITED • j ^ S j f DU CANADA LIMITÉE

DNA REPAIR, HUMAN CANCER AND ASSESSMENTOF RADIATION HAZARDS

Réparation de l'ADN, cancer chez les êtres humainset évaluation des risques d'irradiation

M.C. PATERSON and D.K. MYERS

Chalk River Nuclear Laboratories Laboratoires nucléaires de Chalk River

Chalk River, Ontario

September 1979 septembre

ATOMIC ENERGY OF CANADA LIMITED

DNA REPAIR, HUMAN CANCER AND ASSESSMENT

OF RADIATION HAZARDS

by

M.C. Paterson and D.K. Myers

Health Sciences DivisionChalk River Nuclear Laboratories

Chalk River, Ontario1979 Septattoer

AECL-6627

Réparation de l'ADN, cancer chez les êtres humainset evaluation des risques d'irradiation

par

M.C. Paterson et D.K. Myers

Résumé

On pense que les cancers/comme les anomalies généti-ques, sont principalement dus à des changements se produisantdans l'ADN, matière génétique présente dans toutes les cellulesvivantes. Une partie de la preuve à l'appui de cette hypothèsedécoule de l'étude de certains troubles héréditaires trouvéschez les êtres humains auxquels ils font courir un grand risque decancer. Des cellules provenant de malades souffrant,en tous cas,de l'un de ces troubles appelé "ataxia telangiectasia", semblentavoir une déficience qui les empêche de réparer les dommagescausés par des rayonnements et/ou par certains autres agents environ-nementaux. Des études relatives aux conséquences de la réparationde l'ADN permettent de penser que les évaluations courantes touchantles risques cancérogènes des rayonnements de faible radioactivitésont en grande partie correctes. Il semble y avoir une certainemarge de sécurité en jeu dans ces évaluations de risque pour lamajorité de la population. Il est, cependant, déconseillé deréduire dans une large mesure les évaluations de risque couram-ment acceptées, par suite de l'existence des sous-groupes poten-tiellement radiosensibles qui forment une minorité dans l'ensemblede la population.

L'Energie Atomique du Canada, LimitéeLaboratoires nucléaires de Chalk River

Chalk River, Ontario

Septembre 197 9

AECL-6627

ATOMIC ENERGY OF CANADA LIMITED

DNA REPAIR, HUMAN CANCER AND ASSESSMENT

OF RADIATION HAZARDS

by

M.C. Paterson and D.K. Myers

ABSTRACT

Cancers, like genetic defects, are thought to becaused primarily by changes in DNA, the genetic materialof all living cells. Part of the evidence in support ofthis hypothesis derives from the study of certain rarehereditary disorders in man associated with high risk ofcancer. Cells derived from patients suffering from atleast one of these disorders, ataxia telangiectasia, appearto be defective in their ability to repair the damage causedby radiation and/or by certain other environmental agents.Studies of the consequences of DNA repair suggest thatcurrently accepted estimates of the carcinogenic hazardsof low level radiation are substantially correct. Therewould appear to be some margin of safety involved in theserisk estimates for the majority of the population, butany major reduction in the currently accepted risk estimatesappears inadvisable in view of the existence of potentiallyradiosensitive subgroups forming a minority in the generalpopulation.

Chalk River Nuclear LaboratoriesChalk River, Ontario

1979 September

AECL-6627

ENVIRONMENTALLY INDUCED DNA DAMAGE, ITSDEFECTIVE REPAIR, AND HUMAN CANCER

Malcolm C. Paterson

Biology and Health Physics DivisionAtomic Energy of Canada LimitedChalk River Nuclear Laboratories

ABSTRACT

Deoxyribonucleic acid (DNA) is the re-pository of the genetic information indis-pensable to the well-being of living cells.Damage to cellular DNA by cancer-causingagents in the environment is consideredto De a major factor in the development ofhuman cancer. Evidence in support of thishypothesis derives in part îrom the labor-atory study of ataxia telangiectasia (AT),a rare hereditary disorder in man in whichafflicted patients are at high risk of can-cer and respond adversely to radiotherapy.Skin cells cultured from such donors arealso hypersensitive to inactivation by ra-diation, owing to a defect in an enzymaticmechanism for the repair of radiogenic dam-age to DNA. AT cells are also hypersensi-tive to the ultraviolet component of sun-light and certain cancer-causing chemicalssuggesting that individuals unusually sen-sitive to irradiation may also display in-creased sensitivity to other environmentalagents. Cells cultured from parents of ATpatients are moderately radiosensitive.Persons of this genetic type are estimatedto comprise %1% of the total population andare known to be cancer-prone. We have alsoobserved enhanced radiosensitivity in cellscultured from three cancer-stricken membersof a family with an unusually high inci-dence of acute myelogenous leukemia, sug-gesting that these family members are ge-netically predisposed to cancer (possiblydue to hypersensitivity to irradiation).These and other clinical examples showpromise as models for assessing the role ofgenetic factors (e.g. faulty genetic in-formation needed for the production of DNArepair enzymes) in cancer induction byradiation and other extrinsic agents.

ABBREVIATIONS

DNA, deoxyribonucleic acid; UV, ultra-violet; Dlo, radiation dose (in rads) re-ducing the survival of a population ofcells to 10% of the unirradiated level; AT,ataxia telangiectasia; DRF, dose reduction

factor [Dlo (control cells)/Dio (indicatedcells)); ENU, ethylnitrosourea; MMS,methylmethanesulfonate; MNNG, N-methyl-N'-nitro-N-nitrosoguanidine; 4-NQO, 4 nitro-quinoline-1-oxide; RTS, Rothmund-Thomsonsyndrome; PE, plating efficiency (percentof the unirradiated cells plated out thatgive rise to colonies)

INTRODUCTION

Research in radiation biology at CRNLis directed toward an improved understand-ing of the harmful effects of ionizing ra-diation on living organisms, with the aimof applying this knowledge to the assess-ment of the health hazards of low-levelradiation to both occupational workers andsociety-at-large. To this end, a numberof different research projects on radia-tion effects are being carried out, rang-ing from the induction of genetic changesin microorganisms to the induction of can-cer in rodents. The main emphasis through-out is the enzymatic repair of radiationdamage to the genetic material of livingorganisms (for details, see ref. 1 and thesucceeding article by D.K. Myers).

THE GENETIC MATERIAL - DNA

The repository of the genetic infor-mation is DNA; it consists of two strandstwisted around each other similar to twospiral staircases (see Fig. 1 in articleby Myers and ref. 2, p. 208-215). Eachstaircase is comprised of alternatingsugars and phosphates; at each sugar, oneof four bases (adenine, cytosine, guanine,or thymine) projects into the stairwelland is paired by hydrogen bonding with asecond base projecting from the oppositestaircase. Only specific pairs of basesfit properly into the stairwell, adeninewith thymine and cytosine with guanine,and hence the sequence of bases along onestrand is said to be complementary to thesequence of bases along the opposite

- 2 -

strand. It is this sequence of bases alongeach strand which determines the geneticinformation; a unit of information iscalled a gene. The genetic material inliving organisms contains an enormousnumber of base pairs, in a human, for ex-ample, there are 10'° base pairs per so-matic cell, and there are ^6 x 10'5 somaticcells in an adult, giving a total of -\,6 x1023 base pairs per individual (1,2).(Note: germ cells, the other type of cellsin the human body, are involved in sexualreproduction and are numerically negligiblecompared to somatic cells.)

The genetic material is distributedwithin cells differently in differentorganisms. In each somatic cell in man,it is divided among 46 chromosomes in twohomologous sets of 23 (thus called a di-ploid cell), one set inherited via a germcell (thus called a haploid cell) from thefather and the other set via a germ cellfrom the mother (3). Each chromosome of ahomologous set contains one DHA moleculeWith an unique structure determined by thesequence of the four bases along each ofthe two complementary strands. Residingwithin the DMA of each set of 23 chromo-somes are all the instructions needed forthe somatic cell: (i) to synthesize, ina highly controlled manner, many differentproteins (including enzymes) and other geneproducts needed to survive and carry outits specific function in the body; (ii) toeventually duplicate its 46 chromosomes anddivide into two daughter cells, each con-taining an identical complement of geneticinformation accurately preserved; and (iii)to produce germ cells (sex organs only).

DNA DAMAGE AND ITS BIOLOGICAL CONSEQUENCES

Man and all other living organisms areroutinely exposed to deleterious agents inthe natural habitat (4,5). These include:ionizing radiation of both natural (e.g.cosmic rays) and man-made origin (e.g.radiation sources used in diagnostic medi-cine and fallout from nuclear weapons test-ing) ; the UV component of sunlight, andnumerous chemicals. Many of these chemi-cals, such as benzene and 2-naphthylamine,are found in the workplace, but some, suchas components of cigarette smoke and vari-ous nitrosamines, are related to life-styleand dietary factors (5).

Regardless of their origin, these phy-sical and chemical agents are believed toexert their harmful biological effects pri-marily as a consequence of interacting with,and structurally altering, DNA (6-8). [onepossible exception is asbestos; this phy-sical agent is chemically inert, and itsmode of action is not known (8). ] Let ustake ionizing radiation as a case in point.At the molecular level, the alterationsproduced in DNA by ionizing radiation areof two major classes: (i) structural

changes in individual bases, particularlythymine and cytosine, which disrupt hydro-gen bonding and base stacking and therebycause localized structural distortions inthe two "spiral staircases"; and (ii)breaks arising from cleavage of the sugar-phosphate backbone of individual "stair-cases" (single-strand breaks) (9). Thesetypes of lesions are formed in consider-able number during radiation exposure. Inthe case of human diploid cells culturedin the laboratory, for example, a biolog-ically relevant dose of 385 rads (3.85 Gy)60Co y-irradiation (i.e. the Oi0 value)will induce in the DNA of each cell -x.4600single-strand breaks and ^2500 basedefects (10).

The presence of such lesions in DNAcan result in either death of the cell ora viable cell carrying altered genetic in-formation which would then be passed on toall subsequent generations of daughtercells. The two types of biological changesassociated with chronic exposure to lowlevels of radiation are referred to as ge-netic (i.e. hereditary defects in the de-scendants of the irradiated person) andsomatic (i.e. changes in the tissues ororgans of the irradiated person presumablyarising from the propagation of genetical-ly altered somatic cells). The most ser-ious somatic changes are those leading tothe appearance of leukemia and other formsof cancer.

ENZYMATIC REPAIR OF DAMAGED DNA

Given the formidable array of DNA-damaging agents present in the environmentand the high premium necessarily placed onmaintaining the accuracy of the geneticscript encoded in DNA, it is not surpris-ing that all living organisms, rangingfrom bacteria to mammals, possess multipleenzymatic mechanisms whose joint actionspromote the restoration of damaged sitesin DNA to a normal structure (10-13).

Two repair mechanisms operating onDNA damaged by many extrinsic agents in-cluding ionizing radiation are termedexcision repair and strand-rejoining; theformer process acts on base defects andthe latter process acts on single-strandbreaks. In the excision-repair mechanism,remedial action is accomplished by the re-moval of a single-strand segment containingthe altered base followed by replacementwith a segment containing the correct se-quence of bases (13-16). In the conven-tional model of excision repair (14), asillustrated in Fig. 1, the following fourenzyme-mediated reactions are carried outin a coordinated fashion:

(i) an incision is introduced intothe sugar-phosphate backbone near thealtered base by an endonuclease which"recognizes" the site as having an abnor-mal structure;

- 3 -

EICIS1QN t I P I I I STIMO-ICJOINIRC

(•I HClSldl

>) MPII1 SïttlHISIS

L

'> L ICaltON

Figure 1 Schematic representation of en-zymatic mechanisms for correcting base de-fects (left) and single-strand breaks(right) induced in DNA of human cells byionizing radiation (see text for details)

(ii) a second single-strand nick ismade on the opposite side of the defect byan exonuclease, thereby excising the dam-aged site from the DNA; the resulting gapis often extensively widened by additionalexonucleolytic activity;

(iii) a DNA polymerase then insertsnew material into the gap (a process termedrepair synthesis), using the opposite in-tact strand for base-pairing instruction;and finally

(iv) the newly synthesized and pre-existing strand segments are joined by aDNA ligase, thus restoring strand contin-uity.

It has recently been shown that someclasses of altered bases are released asfree bases rather than within single-strandsegments as in the traditional model of ex-cision repair (15). In the newer model,the covalent bond joining the damaged basewith the sugar moiety is first cleaved by aDNA glycosylase, and the modified site con-taining the denuded sugar is then correctedin the same fashion as a base defect in thetraditional model—that is, strand incision,excision of the modified site, repair syn-thesis, and strand ligation. The classesof base defects acted upon by the newermode of excision repair are believed tocause only minor loéffLized distortions inthe DNA molecule; hence the pre-incisionstep involving the removal of an alteredbase as a free entity may serve as a meansof preventing abortive repair, i.e. ensurethat the site, once incised, is not rejoin-ed before the defective base is removed.It would now appear that most base defectsinduced by ionizing radiation are correctedby this mode of excision repair whereasbase defects produced by UV light, for ex-ample, are corrected by the conventional

excision-repair mode (13).The mechanism mediating the rejoining

of single-strand breaks is less well under-stood, although this mechanism may be as-sumed to be less complex than excision re-pair of base damage because the sugar-phosphate backbone has been severed di-rectly by the radiation treatment, thusobviating the need to carry out the inci-sion reaction. The restitution of asingle-strand break is often accompaniedby the release of nearby base and sugarmoieties; thus its repair is presumablyachieved by an abbreviated form of exci-sion repair involving limited exonucleo-lytic digestion to "clean the frayed ends",repair synthesis and strand ligation (9,16) .

A key property of the excision-repairmechanism is its ability to correct aseemingly limitless spectrum of chemicallydistinct modifications in the four bases,whether induced by extrinsic agents orarising spontaneously. This versatilityis given by a family of different endo-nucleases and DNA glycosylases (13).Whether the three types of reactions com-mon to the excision-repair and strand-rejoining mechanisms are carried out bythe same enzymes has yet to be determined.

CAUSES OF HUMAH CANCER: OUR APPROACH TOTHE PROBLEM

One research project at CRNL is de-signed to clarify our understanding of theunderlying causes of cancer in man. Twofactors are thought to predominate: en-vironmental and genetic. It is now widelyheld that most, if not all, human cancersare caused, at least in part, by some ex-trinsic factor, e.g. a cultural agent,such as cigarette smoke, or a chemical inthe workplace, such as vinyl chloride (seerefs. 4 and 5). Recently it has also be-come evident that certain genetic alter-ations may predispose an individual to thedevelopment of cancer. It is thereforebecoming increasingly important to assessthe interplay of these environmental andgenetic factors in cancer induction. Weare particularly interested in determininghow defective genes involved in the syn-thesis of DNA repair enzymes may interactwith ionizing radiation in cancer develop-ment.

Our basic experimental approach is toexpose cultured human cells to a givencancer-causing agent (carcinogen), usuallyCo60 Y-radiation» a n d measure:

1. the ability of single cells toundergo many successive cell divisions(>6) and thereby form macroscopic colonies,each containing an aggregate of 100 ormore cells (biological endpoint); and

- 4 -

2. the initial yield and subsequentfate (i.e. repair) of single-strand breaksand base defects (molecular endpoint).(Note: these cells grown in the laboratoryare derived from a skin biopsy of a humandonor; such cells serve as useful test ma-terial for studying the effects of ionizingradiation on man because they respond toradiation exposure in a manner similar tothat of most types of somatic cells in thebody.)

ATAXIA TELANGIECTASIA: HUMAN DISORDERLINKING RADIOSENSITIVITY WITH CANCERPROMENES?

He have been greatly aided in our re-search by the availability of skin cellsderived from patients with the rare hered-itary disease, ataxia telangiectasia (10,16). This is a complex, single-gene dis-order whose incidence is ̂ 24 per millionlive-births. The disease is transmitted ina recessive fashion--that is, affected in-dividuals inherit two defective copies ofan AT gene, one copy of maternal, and theother of paternal origin. Its major clin-ical characteristics include: (i) muscu-lar incoordination; (ii) blood vessel di-lation (eyes and skin); (iii) defectiveimmunity; (iv) proneness to cancer (par-ticularly lymphomas and lymphatic leuke-mias); and (v) hypersensitivity to radio-therapy. One in 10 AT patients developscancer, a risk which is ^1200 times higherthan in an age-matched control population.Afflicted persons typically die beforeadulthood from pneumonia (due to immunedeficiency) or cancer or both. In threewell-documented cases, AT patients havedied within six months upon receiving con-ventional radiotherapy for the treatment ofsolid malignant tumors; hence, an unusuallysevere reaction to ionizing radiation isobserved at the clinical level (16).

Our goal has been to determine theradiosensitivity and DNA repair propertiesof skin cells from AT donors and therebyexplain the basis of the hypersensitivityto irradiation associated with the disorder.That abnormal radiosensitivity extends tothe cellular level can be seen in Fig. 2.Cells from nine unrelated AT patients alldisplay increased sensitivity to the lethaleffects of y-irradiation compared to con-trol cells from five unrelated normal per-sons; the sensitivity increase correspondsto a DRF of '«3. This enhanced radiosensi-tivity of the AT strains is observed wheth-er the radiation exposure is given underoxic or hypoxic conditions (i.e. in air orin N2) (see Table 1). Furthermore, cellsfrom one AT donor, while proficient in re-joining strand breaks, are "v.3 times slowerin the removal (i.e. repair) of base de-fects than are cells from a clinically nor-mal donor (Fig. 3). Two other AT strains,AT2BE and AT81CTO, are also defective in

NORMAL(S SWAINS) -

ZOO 400 «00 100 1000 1200/•RAY DOSE (rod, in N,)

Figure 2 Range of Y-ray survival curvesof cultured cells from nine AT and fiveclinically normal subjects Co60 Y-irradiation was administered under hypoxicconditions.

1.6 0 2 4Timi afttr 7-ny incubation (h)

Figure 3 Time-course of the disappear-ance of single-strand breaks (left panel)and sites containing base defects (rightpanel) from the DMA of normal (CRL 1141)and AT (AT3BI) cells exposed to 50 krad(0.5 kGy) of hypoxic y-irradiation[From Paterson et al. (réf. 25) with per-mission of HacMillan Publishing Company]

site removal but rejoin single-strandbreaks at a normal rate (10,13,16). Thesimplest conclusion is that these ATstrains lack a fully functional enzyme,presumably an endonuclease or DNA glyco-sylase, needed to initiate the excision ofsome type of base damage. These combinedbiological and molecular data, besides of-fering new insight into the basic cause ofthe disease, provide one of the bestpieces of evidence available to date fora causal relation between defective repairof DNA damage and predisposition to can-cer. In short, they dramatically illus-trate the importance of DNA repair

- 5 -

Table 1 Sensitivity of AT Strains to Cell Killing by Various Carcinogens*

Strain Physical Agent

AT2BE

AT3BI

AT4BI

Y-raysOxxa Hypoxia

+++ +++

NeutronsOxia Hypoxia

+ +

+ + + +

++ ++

254

N

N

N

UV Linm

ght313 nm

+

N

+ +

Chemical AgentMMÎ3 MNNG INU 4-NQO

313 nm UV FLUENCE Jm'* i K)'4

10 20 30254 nm UV FLUENCE JflT

From ref. 16 and unpublished data (M.C. Paterson, P.J. Smith, B.P. Smith,M.V. Middlestadt, N.T. Bech-Hansen and B.M. Sell)

The following symbols are used to denote relative sensitivity: N, normal (0.8<DRF<1.2);+, slightly sensitive (1.2<DRF<1.5); ++, moderately sensitive (1.5<DRF<2.0); +++, mark-edly sensitive (2.0<DRF<4.0).

mechanisms to the well-being of man.We are currently attempting to define

further the repair defect in AT. Two as-pects of this problem are under investi-gation: (i) identification of the repairenzyme, endonuclease or DNA glycosylase,mal functional in AT cells; and (ii) de-termination of the chemical nature of thealtered base acted upon by this repair en-zyme. We are fortunate to have severalradiation chemists and enzymologists col-laborating with us on these studies, in-cluding Drs. J. Cadet (Centre D'EtudesNucléaires de Grenoble), N.E. Gentner(CRNL), and P.V. Hariharan (Roswell ParkMemorial Institute).

Strains from various AT patients arealso abnormally sensitive to 14 Mev neu-trons; however, the AT strains are only-̂ 1.8 times more sensitive to this denselyionizing radiation than are control strainsfrom normal persons, if treated in air orin N2 (Table 1). These results indicatethat the abnormality in AT cells is lesscritical for the lethal effects of denselycompared to sparsely ionizing radiation.The findings are also consistent with ano-malous DNA repair as a basic defect in ATcells; the fraction of the total DNA damageamenable to correction by enzymatic repairis smaller for densely than for sparselyionizing radiation (17) , and thereforefaulty repair should be less of a handicapfor cell recovery from neutron exposurecompared to y-ray exposure.

It is of interest to determine whetherAT cells are also abnormally sensitive tothe lethal effects of other carcinogens inour environment. Fig. 4 shows that onestrain, AT4BI, is hypersensitive to near UVlight (313-nm wavelength), but exhibits anormal sensitivity to. kill ing by far UV light(254-nm wavelength). AT2BE cells respondsimilarly to 313-nm and 254-nm light (Table1).' This marks the first time that humanstrains have been found that are hypersen-sitive to near but not far UV light. This

Figure 4 Survival curves of AT (AT4BI)and normal (GM38) cells upon exposure tofar (left panel) and neat (right panel)UV light

enhanced sensitivity to 313-nm light ob-served in AT4BI and AT2BE cells providesa rational explanation for why blood ves-sel dilation is particularly pronouncedover sunlight-exposed regions of the skinin some AT patients. Furthermore, thisobservation implies that near UV lightnot absorbed by the upper atmosphere mayconfer a biological effect which partiallymimics that produced by ionizing radiation,thus modifying the terms in which thisubiquitous carcinogen should be consider-ed. (Note: far UV light has littlepractical relevance; it is absorbed byozone in the stratosphere and thus doesnot penetrate to the surface of the earth.Over the years it has been extensivelyused in the laboratory largely becausegermicidal lamps are relatively cheap andreadily available sources of this light,and its biological effects are for themost part similar to thosa produced bynear UV light.)

- 6 -

As a general rule, AT strains alsodisplay hypersensitivity to those chemicalcarcinogens whose biological effects mimicthose of ionizing radiation; included inthis list is ENU (Fig. 5), an agent notedfor its ability to induce brain tumors inrats (18) . There is, however, much morevariability among different AT strains inresponse to treatment with radiomimeticchamicals, such as MMS and MNNG, than isfound for ionizing radiation (Table 1).The data are compatible with faulty DNArepair as a root cause. Radiation isknown to produce a wide spectrum of les-ions in DNA (9), and hence efficient re-pair presumably requires that a number ofdifferent repair pathways all be fullyoperational; conversely, a malfunction inany one of these pathways would be expect-ed to result in the enhanced radiosensi-tivity observed for all AT strains. Onthe other hand, radiomiinetic chemicals arethought to induce a relatively narrowspectrum of reaction products in DNA, andthe relative yields of these productsoften differ greatly from one chemicalagent to another (19) . Consequently, ifdifferent AT strains are defective in thesynthesis of repair enzymes mediating dif-ferent pathways, it is not surprising thatthere are differences in the sensitivityof these strains to specific radiomimeticcarcinogens. In brief, these findings onthe response of skin cells from differentAT donors to killing by various physicaland chemical carcinogens serve to illus-trate two general principles: (i) thereare multiple genes in man which can leadto the genetic disorder AT [a possibilityin keeping with the heterogeneous natureof most human genetic diseases (3)]; and(ii) persons found to be hypersensitiveto ionizing radiation will in all pro-bability also prove to be hypersensitiveto certain other environmental carcino-gens. A ready explanation for the secondprinciple may lie in the likelihood thatportions of the DNA damage induced byionizing radiation and by various radio-mimetic carcinogens are handled by re-pair pathways mediated by one or moreidentical enzymes (e.g. endonucleases orDNA glycosylases).

PERSONS CARRYING ONE DEFECTIVE AT GENE:RADIOSENSITIVE AND CANCER-PRONESUBPOPULATION

Let us now turn to a matter of somepractical importance. Although personsafflicted with ataxia telangiectasia (i.e.individuals inheriting two altered copiesof an AT gene) are too rare to constitutea serious public health problem, it can beestimated by basic genetic principles thatpersons carrying one defective copy (plusone normal copy) of an AT gene comprise"•*!% of the general population (20,21) .

1 .U

FR

AC

TIO

N

O

'IVIN

G

j- 001n

0001

-

- • OM30* AT3B1a AT4BI

1 I

i i I 1 1 :

\ \ ;

\ \ ~

NORMAL \ \ J

, , , ,0 1 Z J 4 S 6 7

ENU DOSE <AiM '

Figure 5 Survival curves of two ATstrains and one normal strain upon one-hour treatment with indicated concentra-tions of ENU

Morover, Swift and co-workers (20) havereported that blood relatives (below theage of 45 years) of AT patients are a fewtimes more likely to die from cancer thanare members of an age-matched controlpopulation; this increased risk of cancermortality presumably reflects the exist-ence of a high frequency of AT carriersamong these blood relatives. It istherefore of interest to measure theradxosensitivity of skin cells from pre-sumed AT carriers—that is, the parentsin AT families, each of whom presumablytransmitted one defective copy of an ATgene to their children afflicted with thedisease. Fig. 6 shows that cells culturedfrom the two parents in ore AT familydisplay y~ray sensitivity intermediatebetween that exhibited by cells fromtheir AT child and control cells fromfive normal subjects. These and otherresults (10,21) suggest that a signifi-cant fraction of the general populationknown to be at increased risk of canceris also moderately radiosensitive.Clearly, continued efforts are needed todetect AT carriers in society-at-largeand to evaluate the role of the inter-action between this genetic state andrelevant extrinsic carcinogens in the ap-pearance of common cancers.

- 7 -

5 0 -

o 10S

OS

SIHIH t tO UH4SKT0 4 31

D "H414CI0 I ] 2»

, l)l4»CTa t II

200 400 600 600 1000 1200r-RAY DOSE (rod. in N2)

Figure 6 Hypoxic 7-ray survival curvesof cell strains derived from three members[affected child (A) plus parents (0,0)] ofan AT family The shaded area representsthe range of survival of control strainsfrom five normal donors. The symbols anderror bars are the means and their stan-dard errors of multiple experiments.Range of PE and number of independent ex-periments conducted for each strain areindicated.

ROTHMUND-THOMSON SYNDROME: HUMAN DISORDERLINKING SUN SENSITIVITY WITH HIGH RISK OFSKIN CANCER

AT is but one of the human diseasesassociated with cancer predisposition thatis under investigation in our laboratory.For example, Dr. P.J. Smith, a postdoctor-al fellow, has been studying the Rothmund-Thomson syndrome. This is another raregenetic disease whose clinical hallmarksinclude: (i) sensitivity of the skin tosunlight (e.g. excessive blistering andpigmentation changes); (ii) cataractformation in both eyes at an early age;(iii) cancer-proneness (e.g. skin carcin-omas); (iv) stunted growth; and (v)underdevelopment of sex organs. Skincells from one RTS patient have been foundto respond to UV light in a fashion simi-lar to AT4BI cells, i.e. abnormally sen-sitive to killing by 313-nm but not 254-nmlight (Fig. 7). This RTS donor had apositive clinical history of sunlightsensitivity and had developed basal cellcarcinoma of the eyelid. Hence, we seehere an association between hypersensi-tivity of cultured cells to the UV compon-ent of sunlight and the presence of skin

JlSnm UV FIU£N«

Figure 7 Survival curves of culturedcells from one RTS patient (•) and onenormal person (o) upon exposure to far(left panel) and near (right panel) UVlight

cancer in the donor (presumably arisingfrom damage inflicted by solar light).Strains from other RTS patients are nowunder study to test the generality ofthis association.

CELLULAR RADIOSENSITIVITY IN A LEUKEMIA-PRONE FAMILY

Me have recently received a contractfrom the United States National CancerInstitute to expand our DNA repairstudies on cultured cells derived frompersons with known or suspected geneticpredisposition to malignancy.Dr. N.T. Bech-Hansen, a Research Associ-ate, and Ms. B.M. Sell, a Research As-sistant, are supported by the contract.This project is beginning to yield in-teresting result». An example pertainsto a family displaying an unusualclustering of acute myelogenous leukemia(22). The well-known ability of ionizingradiation to induce leukemia (23) prompt-ed us to assess the radiosensitivity ofskin cells from six members of thisfamily. The results are summarized inTable 2. Strains from three family mem-bers with cancer (2649T from the mother,and 409T and 2642T from two daughters)are all significantly more radiosensitivethan the control strain from a normaldonor (GM38), but are not as sensitive asthe strain from an AT patient (AT2BE).The radiosensitivity of the strains fromthe father (2650T) and from one son(2647T) are not significantly differentfrom that of the control strain, whilethe strain from a second son (2648T) doesexhibit increased Y-ray sensitivity.Hence, there appears to be an associationbetween cancer proneness in an individualand enhanced radiosensitivity of his cul-tured cells. Given an abnormal clustering

- 8 -

Table 2 Sensitivity to Oxic ylrradiation of Skin Cells Cultured from Members of aFamily with an Abnormal clustering of Acute-Myelogenous Leukemia*

Strain

GM38AT2BE

2649T2650T409T2642T2647T2648T

Family Relation

_

-

motherfather

daughterdaughter

sonson

DonorClinical Description

normalataxia telangiectasia

uterine cervical cancernormalacute myelogenous leukemiaacute myelogenous leukemianormalnormal

Age

97

414612201616

No. ofExperiments

62

454423

D,0±S.E.

407+15169± 9

301±12425*19354+18290±31417+30321+25

P

*

*

*

*

Prom ref. 24 and unpublished data (N.T. Bech-Hansen, B.M. Sell and M.C. Paterson)

•P<0.05, comparing Dio value to that of GM38, using the standard error of a differenceas a statistical test

of malignancies in a family, studies suchas radiosensitivity of cultured cells maybe of value in predicting the predispo-sition of an individual to acute myelo-genous leukemia or other cancers. Inves-tigations are now underway to test thisprediction.

It is perhaps not too surprising thatone family member, although his culturedcells display abnormal Y-ray sensitivity,should be free of cancer at the presenttime. This person seems to possess a re-duced tolerance to the harmful effects ofextrinsic radiation and presumably certainradiomimetic agents. Our results havebeen passed on to the physician attendingto the family, and this family member isnow being carefully watched for the ap-pearance of any malignancy.

SUMMARY AND CONCLUDING REMARKS

In summary, our studies have providedinformation on (i) who is more suscepti-ble to cancer and why; and (ii) the roleof inefficient or incomplete repair ofdamage to DNA in the underlying mecha-nism (s) of cancel induction by extrinsicagents in our biosphere.

Our work is pertinent to both thenuclear power industry and the biomédicalsciences. Its significance can be sum-marized as follows: (i) identificationof persons unusually sensitive to radia-tion exposure; (ii) provision of infor-mation relevant to assessment of the riskof chronic exposure to low-level radia-tion; (iii) clarification of the basicmechanism (s) involved in the developmentof cancer; and (iv) assessment of thecontribution of genetic factors in cancerpredisposition.

In short, our research should helpexplain why certain persons tend to de-velop certain types of cancer. This in-formation will contribute to the develop-ment of unifying concepts concerning theorigin of cancer which will in turn as-sist in the formulation of better strate-gies for the prevention and early detec-tion of cancer.

ACKNOWLEDGMENTS

Numerous persons have been involvedin the studies described here. B.P. Smithand P.A. Knight have provided experttechnical help from the onset;Dr. P.J. Smith, Dr. P.H.M. Lohman, avisiting scientist from TNO, Rijswijk,Holland, and A.K. Anderson, L. Fishman,and M.V. Middlestadt, three undergraduacesummer students, have contributed greatlyon a temporary basis. More recently,Dr. N.T. Bech-Hansen and B.M. Sell, inaddition to Drs. R.B. Miller,J.J. Mulvihill and J.M. Lamon of theClinical Epidemiology Branch, U.S.National Cancer Institute, have mademajor contributions to the program. Botha multidisciplinary approach and a widerange of biological systems are beingused in the research ongoing in the Bio-logy Branch, and we have profited greatlyfrom the expertise and encouragementfreely offered by our colleagues in theBranch, particularly Dr. D.K. Myers, itsHead, as well as Dr. A.M. Marko, Directorof the Biology and Health PhysicsDivision. We gratefully acknowledge par-tial support by U.S. NCI Contract N01-CP-81002 with the Clinical EpidemiologyBranch, U.S. National Cancer Institute.Finally, I am indebted to Drs.

- 9 -

N.T. Bech-Hansen, N.E. Gentner, andP.J. Smith for their constructive com-mentary on the manuscript and toC.I. Walters for her patience and forbear-ance during the typing of the manuscript.

REFERENCES

1. Objectives of Research Activities inBiology Branch, Chalk River NuclearLaboratories, 1976. Atomic Energyof Canada Limited, Report AECL-5613,1977. See also: Myers, D.K.,Radiation Biology for the Non-Biolo-gist. Atomic Energy of CanadaLimited, Report AECL-5721, 1978.

2. Watson, J.D., Molecular Biology ofthe Gene, 3rd edition, W.A. Benjamin,Menlo Park, Ca., 1976.

3. Levitan, M., and Montagu, A., Text-book of Human Genetics, 2nd edition(revised by Levitan, M.), OxfordUniversity Press, New York, 1977.

4. Doll, R., Nature, Vol. 265, p. 589-596, 1977.

5. Higginson, J., Carcinogens: Iden-tification and Mechanisms of Action(The University of Texas SystemCancer Center 31st Annual Symposiumon Fundamental Cancer Research,1978), p. 187-208, Raven Press, NewYork, 1979.

6. Cerutti, P., DNA Repair Mechanisms(ICN-UCLA Symposia on Molecular andCellular Biology, Vol. 9), p. 1-14,Academic Press, New York, 1978.

7. Miller, E.C., Cancer Res., Vol. 38,p. 1479-1496. 1978.

8. Harnden, D.G., Knudson, A.G., Miller,R.W., Brookes, P., Cairns, J., Croce,CM., Doerfler, W. , Fialkow, P.J.,Green, H., Magee, P.N., Muir, C.S.,Paterson., H.C., Rajewsky, M.F., andWinnacker, F..L., Neoplastic Trans-formation: Mechanisms and Con-sequences, p. 247-268, Dahlem Kon-ferenzen, Berlin, 1977.

9. Ward, J.F., Adv. Radiation Biol.,Vol. 5, p. 182-239, 1975.

10. Paterson, M.C., DNA Repair Mechanisms(ICN-UCLA Symposia on Molecular andCellular Biology, Vol. 9), p. 637-650,Academic Press, New York, 1978.

11. Cleaver, J.E., Progress in GeneticToxicology, p. 29-42, Elsevier/NorthHolland Biomédical Press, Amsterdam,1977.

12. Bootsma, D., DNA Repair Mechanisms(ICN-UCLA Symposia on Molecular andCellular Biology, Vol. 9), p. 589-601,Academic Press, New York, 1978. Seealso: German, J., ibid., p. 625-631.

13. Paterson, M.C., Carcinogens: Identi-fication and Mechanisms of Action(The University of Texas System CancerCenter 31st Annual Symposium on Funda-mental Cancer Research, 1978), p. 251-276, Raven Press, New York, 1979.

(AECL-6369).14. Grossman, L., and Riazuddin, S., DNA

Repair Mechanisms (ICN-UCLA Symposiaon Molecular and Cellular Biology,Vol. 9), p. 205-217, Academic Press,New York, 1978. See also:Waldstein, S., ibid., p. 219-224.

15. Linn, S., DNA Repair Mechanisms(ICN-UCLA Symposia on Molecular andCellular Biology, Vol. 9), p. 175-178, Academic Press, New York, 1978.

16. Paterson, M.C., and Smith, P.J.,Annu. Rev. Genet., vol. 13, in press.

17. Ritter, M.A., Cleaver, J.E., andTobias, C.A., Nature, Vol. 266, p.653-655, 1977.

18. Goth, R., and Rajewsky, M.F., Proc.Natl. Acad. Sci. USA, Vol. 71, p.639-643, 1974.

19. Strauss, B.S., Aging, Carcinogenesis,and Radiation Biology, p. 287-314,Plenum Press, New York, 1976.

20. Swift, M., Sholman, L., Perry, M.,and Chase, C , Cancer Res., Vol. 36,p. 209-215, 1976.

21. Paterson, M.C., Anderson, A.K.,Smith, B.P., and Smith, P.J., CancerRes., Vol. 39, in press.

22. Snyder, A.L., Li, F.P., Henderson,E.S., and Todaro, G.J., Lancet, Vol.1, p. 586-589, 1970.

23. Miller, R.W., Biology of RadiationCarcinogenesis, p. 45-50, RavenPress, New York, 1976.

24. Bech-Hansen, N.T., Sell, B.M.,Mulvihill, J.J., and Anderson, A.K.,Proceedings Annual Joint Meeting,American Association for CancerResearch and American Society ofClinical Oncology, p. 78, 1979.

25. Paterson, M.C., Smith, B.P., Lohman,P.H.M., Anderson, A.K., and Fishman,L., Nature, Vol. 260, p. 444-447,1976. (AECL-5410) .

- 10 -

RISK ESTIMATES AND DNA REPAIR

D.K. Myers

Biology Branch, Chalk River Nuclear LaboratoriesAtomic Energy of Canada Limited

ABSTRACT

Cancers, like genetic defects, arethought to be caused by changes in DNA, thegenetic material. Different types ofchanges in DNA are mentioned which may beinduced either by a single damaging eventor by two coexistent events. In the firstcase, the biological effects are directlyproportional to the total dose of radiationand are independent of the dose-rate. Inthe second case, the effects tend to dis-appear as the time for delivery of a givendose of y~radiation is prolonged; this isattributed to the operation of DNA repairsystems in the living organism. It seemsprobable that currently accepted estimatesof the carcinogenic hazards of low levelradiation based on the linear dose-effectmodel are substantially correct. There maywell be some margin of safety involved inthese risk estimates, but any major re-duction in the currently accepted riskestimates appears inadvisable.

THE ACCEPTED RISK ESTIMATES

Data reviewed most recently by theUnited Nations Scientific Committee on theEffects of Atomic Radiation (UNSCEAR) (1)suggest that about 1 in every 200 of thefatal cancers and about 1 in every 200 ofthe genetic defects which occur normallyin human populations might be caused by thenatural background radiation level of 100millirem (1000 ySv) per year. The Inter-national Commission on Radiological Pro-tection (ICRP) (2) has independently pub-lished risk estimates which agree withthose given in UNSCEAR and has recently re-confirmed (3) that "published informationon the epidemiological and radiobiologicalevidence of radiation risks to man...up toMay 1978 does not call for changes in therisk factors" published in 1977. Variousother national agencies have from time tptime issued reports with risk estimateswhich, by and large, agree with those sug-gested by UNSCEAR and ICRP (4-6).

The above risk estimates are basedprimj rily on epidemiological studies ofcancai frequency in human populations ex-posed to unusual levels of ionizing radi-ation and, in the absence of human data,on laboratory studies of genetic defectsinduced in mice and other organisms byionizing radiation. However, present riskestimates are necessarily derived from re-sults observed at high radiation doses andhigh dose-rates where measurable effectsof radiation can be observed. The appro-priateness of the standards of radiationexposure which have been set for protec-tion of the health of radiation workersand of the general public depends uponthe reliability of the estimates of bio-logical effects of low-level radiation.There are some uncertainties involved inthe extrapolation of estimates from highdoses down to very low radiation doses atlow dose-rates where measurable effectscannot be observed. A continuing programin basic research is therefore being car-ried out at the Chalk River laboratoriesto make certain that we know as much aspossible about the long-term biologicaleffects of low radiation doses (7). Inorder to understand these effects, wewould like to understand the basic mech-anisms responsible. For this purpose, wehave concentrated on radiation damage toDNA in the living organism and the repairof this damage.

DNA AND DNA REPAIR

About 26 years ago, biologists re-cognized that the information for all lifeprocesses is contained in a long thread-like molecule called deoxyribonucleic acid(DNA). The DNA molecule represents theblueprint or coding tape for the construc-tion and function of living things; itstores all the genetic information that ispassed from one generation to the next.In this manner it ensures that the majorcharacteristics of living things remainessentially constant over many genera-tions. A small portion of the DNA mole-cule is illustrated in Fig. 1; other more

- 11 -

DNA Structure

0-0

0t

Figure 1

detailed illustrations can be found in mostof the recent textbooks on molecular cellbiology. On the outside of the moleculeare two separate continuous strands of aparticular sugar (deoxyribose) connected byalternating phosphate moieties; in thecentre of the DNA molecule is a series ofpaired bases, whose sequence along thelength of the DNA molecule provides thecoded information for the intricate chem-ical processes carried out by a livingcell. Most cells of the human body contain46 DNA molecules of this kind, each ofwhich is about 2 nanometres across and con-tains roughly one hundred million of basepairs along some 4 centimetres (4 x 10'nanometres) of length. The germ cells(sperm or ovum) each contain 23 differentmolecules of DNA; the 46 molecules presentin most cells of the body represent 23molecules inherited from the father plus 23molecules with similar, but not necessarilyidentical, information from the mother.Thus most cells in the human body containtwo non-identical but similar versions ofthe information for some 100,000 separateoperations that are required for develop-ment and normal function.

Current data suggest that up to 50alterations per minute (or several milliona year) may occur spontaneously in the DNAof each living cell in the human body (8,9)(Table 1). Changes in DNA structure mayalso be caused by radiation, ultravioletlight, viruses and many environmentalchemical agents.

Since the DNA coding tape, like anyother organic molecule, is not inherentlystable, life as we know it would probably-be impossible if the living organism didnot contain built-in mechanisms which re-pair the DNA molecule. The instructionsfor these repair mechanisms are also ineluded in the DNA as part of the informa-tion necessary for life processes.

These repair systems identify and cor-rect changes in the DNA coding tape, wheth-er these changes arise spontaneously or as

a result of exposure to environmentalagents. The cancers and genetic defectswhich do occur naturally are believed tobe the result of a small number of errorswhich are overlooked and therefore notrepaired, or are repaired incorrectly.

Heritable genetic defects are knownto be caused by changes in the geneticinformation, i.e., by changes in DNA.There are perhaps three basic reasons forbelieving that cancers are also caused byalterations in the DNA. First, any changein normal cell development, such as a can-cer, presumably involves some kind ofchange in the geii.*-ic information control-ling the normal functions of the livingcell. Second, DNA-damaging agents whichcause genetic defects in lower organismsusually cause cancer in mammals, and viceversa (10-12). Third, human individualswho suffer from deficiencies in theirability to repair DNA damage are also moreprone than the average person to cancerdevelopment (13-16).

There are many different types ofchanges which can and do occur in the DNAmolecule. Not all of these changes areequally likely to result in cancer. Froma biochemist's point of view, radiationbiologists are attempting to identify thelong-term biological effects of two orthree DNA defects per year due to naturalradiation in a background of several mil-lion similar defects per year due to othercauses (Table 1). However, it is veryobvious to most radiation biologists thatwe still have a great deal to learn aboi":the exact mechanisms by which DNA lesionslead to development of cancers many yearslater. The process of cancer developmentcan undoubtedly be accelerated by "pro-moting agents" and is probably hinderedby itnmunological factors in the healthybody.

A large part of our biology researchprogram at Chalk River is currently con-cerned with an understanding of the fun-ction of DNA, with the nature of the er-rors introduced into the DNA by radiation,and with an understanding of the mecha-nisms by which this damage is repaired(7). Radiation effects are being studiedwith a wide variety of living organismsusing advanced techniques in molecularbiology, in order to be certain that weunderstand the effects upon other livingthings in our environment as well as theeffects upon humans.

FACTORS WHICH AFFECT DNA REPAIR

Different types of radiationdamage to DNA can be measured by avariety of physico-chemical techniques.The most prevalent types of damageappear to be single-strand breaks,various kinds of base damage includingloss of a base, cross-linking of

- 12 -

Table 1 Estimates of the Rate at Which Defects are Introduced into the DNA of a TypicalHuman Cell

Cause of defects DNAdefects/cell/minute

Background radiation (0.1 rem/year)

Maximum occupational exposures (5 rem/working year)

Doubling of mutation rate in male mice(10 rem/week for 10 weeks)

50 percent life shortening in mice(200 rem/week for life)

Spontaneous depurination

Total spontaneous degradation

Bright summer sunlight (fair skin only)

Acute exposure to 100 rem in one minute

0.000004-0.00001

0.001-0.002

0.02-0.05

0.4-1

2-10

10-50

1,000-2,000

2,200-5,000

*Data from Myers (44).dividing by 100.

Dose equivalents in rem can be converted into sieverts by

strands and double-strand breaks (13).Regardless of the type of agent whichcaused the initial damage, most of thisdamage disappears from the DNA of normalorganisms within a few hours as a result ofactive repair systems in the living organ-ism (unless the exposure to radiation orchemical agents is so high that the repairsystem is unable to cope). One example ofthis is shown in Fig. 2. Any factor whichinterferes with the normal DNA repair sys-tems might well be expected to alter themagnitude of the biological hazards ofradiation. Two such factors are consideredbelow.(a) Hereditary deficiencies in the enzy-matic DNA repair systems have been studiedin microorganisms for many years. Fig. 3illustrates the marked increase in radia-tion sensitivity caused by hereditary de-fects in two distinct repair systems in aparticular radiation-resistant microorgan-ism which is being studied at Chalk River.Much of our knowledge of the role of DNArepair mechanisms stems from this type ofexperiment with microorganisms.

More recently, analogous defects havebeen identified in humans who suffer fromcertain rare hereditary diseases (13-16).The clinical abnormalities associated withthese hereditary diseases emphasize theimportance of DNA repair systems for ournormal health and development. This topicis considered in more detail in the ac-companying paper by M.C. Paterson, but somefurther points might be noted. The datafrom this line of research are not expected

1 2 3TIHE AFTER IRRADIATION (hours)

Figure 2 Repair of double strand breaksin the DNA of M. radiodurans (unpublisheddata from D.K. Myers and L.D. Johnson).Note that 1000 krad is lethal, 500 kradis not.

- 13 -

i i.o i-

0.001

0. 10 -r

o.oi -:

0 500 1000 1500

X-R»Y EXPOSURE, trad2000

Figure 3 Radiosensitivity of normal cells(WT) of M. radiodurans and of mutant cells(UV-17, rec 30) defective in DNA repairsystems (unpublished data from N.E. Gentner;note 100 krads equals I kGy).

to alter the risk estimates for the averagepersons in any way; in fact, currently ac-cepted risk estimates are thought to besufficiently conservative to provide somemargin of safety even for a populationwhich contains a minor proportion of radio-sensitive individuals (13,17). However, itis probable that this research will helpto identify those individuals in the popu-lation who are more sensitive than normalto induction of cancer by radiation and byother environmental agents (e.g., chemi-cals, ultraviolet light) and, even more im-portantly perhaps, will increase our under-standing of the basic mechanisms involvedin the induction of cancer by environmentalagents.

(b) Chemical agents such as caffeine oracriflavine which inhibit DNA repair pro-cesses have been shown to cause an appreci-able increase in the lethal effects of Y~radiation on microorganisms (Fig. 4). Syn-ergistic interactions between y-radiationand ultraviolet light have also been demon-strated in microorganisms (18); this ap-pears to be due to overloading of one par-ticular repair system with different typesof DNA damage arising from exposure tothese two agents. Presumably a similarsynergism could occur with other carcino-genic agents or other chemicals whichinterfere with DNA repair.

EHH - 2 CEILS

sor R l l EIPOSURE.

Figure 4 Radiosensitivity of yeast cellsin the presence and absence of 0.1% caf-feine (data from reference 45; note 100krads equals 1 kGy).

We have therefore undertaken animalexperiments to explore the induction ofcancer by radiation alone and by radiationin combination with various chemicalagents. To date, these studies have con-centrated on caffeine (an inhibitor ofDNA repair), urethane (a weak carcinogenand a promoting agent), cigarette smokecondensate (another carcinogen and pro-moting agent), and anthramine (a ratherpotent carcinogen for humans). These arelong term experiments and the data aretherefore still incomplete. However, itis already fairly clear that chemicalagents may, under certain conditions, in-crease radiation effects in one system,i.e., induction of skin cancers (19), butnot in another system, i.e., induction ofbreast cancers (13) (Pig. 5). Synergisticinteractions between cigarette smoking andeither alpha-radiation (from radon daugh-ters) or asbestos have also been observedin humans (20,21).

In order to obtain some indicationof the overall effects on the wholeanimal, we have further carried out ex-periments with high doses of X-radiationand of urethane (22) . The results ob-tained with four different strains of ratsindicate that the overall carcinogenic andlife-shortening effects of radiation plusurethane are not greater than additive(Fig. 6).

Continued attention to the inter-action of radiation and other environmen-tal agents is obviously advisable. Anysynergistic interactions occurring inhuman populations as they have lived overthe past 30-40 years are already taken in-to account in the accepted estimates

- 14 -

1 / -RADIATION

2 , URETHANE/«t (MINUS CONTROL VALUES!

1 r -e*OUTlON FOLLOIEDB» 2g URETHANE/»!

100 ran s 200 ' " "

Figure 5 Induction of breast cancers infemale Sprague-Dawley rats by radiationand urethane (data adapted from reference13; note 100 rads equals 1 Gy).

I 1 825 R X-RAOIAT10N

^ H H - 5 * URETHANE/*S (MINUS CONTROL VALUES)

B251 X-RADIATION PLUS UREIHANE

Figure 6 Cumulative mortality in fourstrains of rats 16 months after exposureto radiation and urethane (unpublished datafrom D.K. Myers; methods described inreference 22).

of the carcinogenic hazards of ionizingradiation to humans. The major problemsinvolved appear to be, first, possiblesynergistic reactions between radiation andnew environmental chemicals, as illustratedby data obtained at Brookhaven (with onestrain of rats only and not with other

strains) for cancer induction by radiationplus diethylstilbestrol (23), and, second,identification of synergistic hazardswhich can be minimized, as illustrated bythe synergistic effects of cigarettesmoking in combination with exposure toeither radon daughters or asbestos inhumans (20,21).

DNA REPAIR AND INDUCTION OF GENETICDEFECTS

As noted above, most of the moleculardefects in DNA that are induced by ioniz-ing radiation, by chemicals or by ultra-violet light can no longer be detected atthe molecular level after the living cellshave been allowed to recuperate from theinitial insult for a few hours. The dis-appearance of molecular defects does notnecessarily mean that the DNA coding tapehas, in all cases, been restored to nor-mal. Consider, for example, what mighthappen during the repair of two doublestrand breaks which were present simul-taneously and in close physical proximityto one another within the living cell.Depending upon the circumstances, repairprocesses could lead to (i) restorationof the normal DNA structure, (ii) in-version of a portion of the coding in-formation within the DNA molecule, (iii)deletion of a portion of the coding tapeand loss of a small fragment or its trans-fer to another DNA molecule, (iv) dele-tion of a portion of the coding tape to-gether with the formation of a ring struc-ture from the deleted fragment, (v) ex-change of genetic information between twoDNA molecules, and (vi) the formation ofbridges between different DMA molecules.All of the latter radiâtion-induced ab-normalities have been known for many yearsand can, indeed, frequently be seen by eyewith the aid of a microscope (24,25).

Other types of incorrect repair ofDNA damaç/o are also known to occur andfrequently lead to invisible minor changesor "point mutations" in the coding tape.In fact, it is suspected that most of thegenetic changes which are observed afterexposure of the organism to ionizing ra-diation, to ultraviolet light or to chem-ical agents are the result of incorrectrepair of the initial damage produced inthe DNA by these agents. There is stillmuch to be learned about these phenomena.One of the more puzzling observations isthe occurrence of "hot spots" in the DNA,i.e., certain small portions of the codingtape are more frequently altered than areother portions after exposure to a givenenvironmental insult. There is, as yet,no known chemical difference in differentportions of the DNA molecule which couldaccount for this observation.

- 15 -

Some recent work at Chalk River on aparticular type of genetic alterationwhich is induced by radiation in yeastcells may be of interest. The DNA in"diploid" yeast cells is organized intodiscrete, paired chromosomes which arelocalized within a cell nucleus; in thisrespect at least, yeast cells resemblehuman cells and differ from bacteria.However, yeast cells, in contrast to humancells, are easy to grow in a test tubeand can be used for some highly sophisti-cated genetic experiments. Once the ap-propriate and rather intricate geneticmanipulations have been carried out andappropriate yeast strains have been con-structed, it becomes fairly simple tomeasure "gene conversions" induced by ra-diation or other agents. Gene conversionoccurs when the DNA damage caused by en-vironmental agents results in exchange ofpart of one DNA molecule with that of itspartner; moreover, in this particularcase, the exchange must occur within agiven gene or coding instruction that iscarried by both of the pairs of chromo-somes. This exchange, it might be noted,results frort the activity of one of theDNA repair systems and does not normallyoccur in mutant cell lines with a heredi-tary deficiency in this particular ("re-combinational") repair system.

Radiation gives rise to this type ofgenetic change in yeast much more fre-quently than it induces the minor changesin DNA coding that are known as pointmutations (13). Our interest in this par-ticular effect stems not only from itsrelatively high sensitivity to environ-mental insults but also from the distinctpossibility that this system will providea more reliable rapid screening assay forcancer-producing agents than do the moreconventional screening tests based on theability of the agent in question to inducepoint mutations in microorganisms.

P. Unrau and D. Morrison have investeda considerable amount of their time re-cently exploring radiation effects withthis particular genetic assay at the ChalkRiver laboratories. The main conclusionsare as follows (Table 2): The number ofgenetic changes induced by ionizing radia-tion is proportional to the total dose ofX- or -y-radiation over the range from oneup to 10,000 rads (0.01 to 100 Gy). Thereis no indication of a "safe" or thresholdradiation dose; there is also no evidencethat effects at low doses are in any waygreater than those predicted by a simplelinear extrapolation from effects observedat high doses (13). The number of geneticchanges induced by radiation is virtuallyindependent of dose rate over a wide rangefrom 0.8 up to 20,000 rads (0.008 to 200Gy) per minute (Table 2). With this par-ticular test system, about in every tenmillion "spontaneous" genetic changescould be ascribed to a background radia-tion level of 100 millirem (1000 uSv) per

year.A recent paper by Ito and Kobayashi

(26) suggested peculiar effects of tritiumg-radiation in this system. Their experi-ment has been repeated by Unrau andMorrison with appreciably greater preci-sion and a wider range of tritium concen-trations. The results to date do not sup-port the preliminary report by Ito et al_. ;tritium g-radiation is not very much~~moreeffective than X-radiation and no signi-ficant dose-rate effects are observed withtritium (Table 2). The reasons for thediscrepancy with the initial report by Itoet al. are being investigated.

DOSE-EFFECT CURVES AND DNA REPAIR

As noted above, the number of geneticchanges i.nduced by ionizing radiation can,in certain cases, be shown to be strictlyproportional to the total accumulatedradiation dose in microorganisms, and tobe virtually independent of the dose-rate.Although the data are somewhat less pre-cise due to the difficulty of working withlarge numbers of higher organisms, similarconclusions appear to be valid for themutations that produced heritable eye-color changes in the wasp Dahlbominus.Previous studies by W.F. Baldwin at theChalk River laboratories (27,28) showedthat the dose-effect curve for this systemwas linear from 15 to 500 rads (0.15-5 Gy)and was again virtually independent of thedose-rate between 0.01 and 100 rads(0.0001-1 Gy) per minute. In both cases,therefore, the genetic change in questionappears to be caused by a single damagingevent in the DNA.

Other types of genetic change maydepend upon two coexistent events in theDNA, as suggested earlier. In these par-ticular circumstances, the number of bio-logical changes induced can be highlydependent on the dose-rate of sparsely-ionizing radiations such as X- or Y-rays.The distance between "damaging events"along the track of a Y-ray is usuallylarge. Thus, more than one Y-ray trackis usually required to induce two co-existent events in the DNA which arephysically close enough (e.g., less than500 nanometers apart) to interact. . Atlow dose rates, the first site of damageproduced in the DNA by one Y-ray trackmay have been repaired and have disap-peared before the second site of damageis produced by another Y-ray, so thatinteraction between the two sites cannotoccur at low dose-rates. Densely ioniz-ing particles such as alpha-particlesproduce many damaging events within ashort distance along their track; thereis thus a high probability that a singlealpha-particle will produce two or morephysically adjacent lesions in the DNA.The dose-rate becomes largely irrelevant

- 16 -

Table 2 Effect of Variations in Dose and in Dose-Rate on the Induction of GeneticChanges in Yeast (unpublished data from P. TJnrau and D. Morrison, using the diploidyeast strain D7-rad52a).

Radiationsource

Dose rate(rads/min)

Total dose(rads)

Gene convertants/106 cells/rad

Co-60Y-rays

150 kv X-rays

00

86323.21.60.8

5.80.580.058

5,000-10,000

150-5,00025-5001.0-50

12.5-5012.5-50

44-8,63043-1,73035-176

5.1

5.4.4.3.3.

2.2.1 .

142*66

608

Tritiatedwater**

•Average for ten measurements at different total doses, but the increments in gene con-version at 1 and 2.5 rads total dose were not statistically significant. The average forthe remaining eight measurements at 5 to 50 rads total dose was 4.4+1.2 gene convertants/108 cells/rad.

**Data obtained with tritiated water are preliminary and should not be used to calculateprecise values for relative biological effectiveness. The relative biological effective-ness of tritium B-rays is close to one when the experiments with X-rays ~.nd with tritiat-ed water are carried out under identical conditions.

Doses in rads can be converted into greys by dividing by 100. One rad (or 0.01 Gy)of X-radiation is equivalent to one rem (or 0.01 Sv) of X-radiation.

in this latter case (Fig. 7).

20

alteredcell.)

ACUTEY » »D + pD"1

Y(falteredcells)

ACUTE ORCHRONIC

Y-Radiation dose

a -Radiation doae

Figure 7 Theoretical effects ofsparsely-ionizing (left) and densely-ionizing (top) radiations at high and lowdose-rates.

- 17 -

However, all sparsely-ionizing radia-tion contains a densely-ionizing componentowing to the nature of its physical inter-actions with matter (29,30). That is tosay, occasionally a single Y-ray track maydamage two sites in the DNA that are closeenough to interact. A simple mathematicalexpression can be used to describe the ob-served results: Y •» aD + bD2, where Y isthe number of induced changes, D is ra-diation dose and a and b are proportion-ality factors. The value of 'b' dependsupon the biological test system used, thetype of radiation, and the dose rate. Inthe case of- sparsely ionizing radiations,e.g., eradiation, the 'bD2' componentwill usually predominate at high doses(e.g., 100 rads or 1 Gy) and high doserates (e.g., 100 rads or 1 Gy per minute),but only the 'aD' component remains atvery low doses or at low dose rates (Fig.7). The non-zero value of 'a1 in thisequation, which is caused by the densely-ionizing component of the y-ray track,provides for a linear component in thedose-effect curve even at very low doses(down to zero) and low dose rates (e.g.,0.1 rad or 0.001 Gy per year).

Other complications need to be kept inmind when interpreting results obtained atmuch higher doses (e.g., 1000 rads or 10Gy) at high dose rates. It is frequentlyobserved that the yield of induced bio-logical effects may decrease at these high-er doses, due to the radiation-induceddeath or "sterilization" of the damagedcells (Fig. 8). This complication rarelyaffects data obtained at doses below 100rads (1 Gy) and will, therefore, not beconsidered further here.

The experimental data obtained forradiation-induced somatic "mutations"(possibly small deletions) in the stamenhairs of Tradescantia (spiderwort) are invery good agreement with above consider-ations (31) (Fig. 8). The number ofhair color changes produced by high doses(up to 100 rads or 1 Gy) of X- or Y-radia-tion are predominantly determined by thesquare of the dose at high dose rates(30-300 rad or 0.3-3 Gy per minute). Asthe dose-rate is decreased, the yield ofinduced effects at total doses of 60-100rads (0.6-1 Gy) also decreases until itapproaches a minimum value which is thesame as that predicted from the 'aD1 com-ponent of the dose-effect curve at highdose-rates. Rather similar effects ofdose-rate are observed for induction ofcoat-color mutations in mice (32). Anallowance for dose-rate effects is incor-porated into the accepted risk estimatesfor the genetic hazards of low-level radia-tion in humans (1).

Risk estimates for induction of cancerin humans by low-level radiation are basedupon the effects of radiation at high dose-rates (1) and do not make any similar al-lowance for a possible decrease in effectsof sparsely-ionizing radiation at low

EfFECl OF BBSE *«B BOSE-«IU 0» IHQUCUQHOF GENETIC CHANGES IN TR»DESC»NTI«

t 100 u t i l - 1 G»>

30 ram/a l i i

100 200TOTAL RtDMTION OOSE ( r . d i )

300

Figure 831).

(data adapted from reference

dose-rates. Animal experiments indicatethat x- or y-radiation at low dose-ratesis frequently much less carcinogenic thanat high dose-rates (Fig. 9). However,the percent decrease in efficiency at lowdose-rates varies greatly from one typeof cancer to another in the animalstudies (33); the range extends from adecrease of a few per -;ent in the case ofbreast cancer to a decrease of more thantwenty-fold in the case of skin cancer.Most scientists who have considered thisproblem would probably agree that the ac-cepted risk estimates for radiation-induced cancer in humans over-estimatethe average risks of low levels of sparse-ly-ionizing radiation by perhaps two tofive times. Both ICRP and UNSCEAR havein fact suggested something like this inrecent publications (1,2), without put-ting any number on the extent of over-estimation.

One should be extremely wary of ex-tending this conclusion to the carcino-genic hazards of radiation generally.First, as judged both from animal experi-ments and from the limited human dataavailable, there is probably little or nosafety factor in the accepted risk esti-mates for radiation-induced breast cancerin humans (1,2,4-6,33) and breast cancersdo account for an appreciable portion ofthe total carcinogenic hazards of lowlevel radiation. Second, the above con-siderations suggest that there is littleor no safety factor in the accepted riskestimates for the carcinogenic hazards oflow levels of densely-ionizing radiationssuch as alpha-particles or neutrons (34)(Figs. 7.and 9 ) . If there are any safetyfactors involved in this latter case, theywould have to stem from two other sources.

- 18 -

FEMILE «CE.OVARIAN TUNDtS.7-RA0IA110H

KALE «ICE. « L E KICE.MYELOID LEUKEMA. MVELDIO LEUKEMIA.I - 01 r-HDUTIOK NEUIUK MDKI IOK

tflO 2DO » 0 100 2M> XO

RIDUT1GH OOSE IN KAOS

IDO 20D 30D

Figure 9 Numbers of cancers induced inmice by irradiation at. high dose-rate(o o) and low dose-rate (• •) (dataadapted from reference 46; note 100 radsequals 1 Gy),

0.01 0.I I

HICBOCUBUS PLUtomuH-239 PEU KILOGRAM

INJECTED INTO DOGS

(DAIA FWa IOTAH MD

stum»

0.001 0.01 O.I 1 10

MILLIGRAMS 0I8ENZANTHRACENE PER HOUSE

Figure 10 Variations in the latent periodbetween exposure to a carcinogenic agentand the development of cancer (data adaptedfrom references 47 and 48).

notably: (i) There may in some cases bean overly-generous allowance in the magni-tude of the quality factors used to con-vert absorbed doses in rads (or grays) todose equivalents in rems (or sieverts).(ii) As the radiation dose decreases, thelatent period between exposure and cancerdevelopment may increase to a point whereit exceeds the normal life span (Fig. 10);this phenomenon is poorly understood and

is not well-documented.It seems probable that estimates of

the carcinogenic hazards of low levels ofionizing radiation based on the lineardose-effect model are substantially cor-rect. There may well be some margin ofsafety involved in these risk estimates.However, the uncertainties involved inthese risk estimates (1,32), the discoveryof radiosensitive subgroups in the humanpopulation (14-16) and the possibility ofsynergistic interactions between radiationand new environmental agents (23) allargue against any major reduction in cur-rent risk estimates.

A few words concerning the theory of"supra-linearity" may be useful in con-clusion. It has been alleged on thebasis of some very shaky epidemiologicaldata on radiation workers at Hanford thatlow doses of radiation cause more cancersin humans than would be predicted on thebasis of the linear dose-effect model (3 5,36). However, it is generally agreed thatradiation workers, whether at Windscale(U.K.) (37), Hanford (U.S.A.) (38), orPickering (Canada) (39), appear to behealthier than the average person of thesame age in the general population. Asnoted in a recent publication, the origi-nal proponents of the above theory "didnot claim that cancer was a major hazardof the nuclear industry or even that thecancer mortality of Hanford workers wassignificantly raised" (40). The originalstatistical analysis of the Hanford data(35) has been severely criticized by otherscientists (38,41) and scientific discus-sion of these data is still continuing.The epidemiology of radiation workers isconsidered in more detail in the accom-panying paper by T.W. Anderson.

Meanwhile, it might be noted that,with one possible exception (42), dataon induction of cancer in animals byradiation do not provide any support forthe concept of "supra-linearity". Dataon induction of genetic changes in theDNA coding by radiation fail to supportthis concept. Our understanding of DNArepair mechanisms provides no reason tobelieve that low level radiation would bemore effective per unit dose than higherradiation levels. This does not meanthat "supra-linearity" is thereforeimpossible; research on possible alter-native mechanisms is in fact supported byAECL at the Whiteshell Nuclear ResearchEstablishment (43). A more correct con-clusion from the accumulated results ofthirty years of research in radiationbiology would be that the concept of"supra-linearity" is high improbable.

- 19 -

ACKNOWLEDGEMENT

The author is indebted to J.D. Childs,N.E. Gentner, G.C. Hanna, R.V. Osborne,A. Petkau, and P. Unrau for valuableadvice, and to Colleen Walters for pre-paration of this manuscript.

REFERENCES

1. United Nations Scientific Committeeon the Effects of Atomic Radiation,Sources and Effects of IonizingRadiation, United Nations, NewYork, 1977.

2. ICRP Publication 26, Annals of theICRP, Vol. 1, Ho. 3, 1977.

3. ICRP Publication 28, Annals of theICRP, Vol. 2, No. 1, 1978.

4. Medical Research Council, The Hazardsto Man of Nuclear and Allied Radia-tions, H.M.S.O., London, 1960.

5. Advisory Committee on the BiologicalEffects of Ionizing Radiations, TheEffects on Populations of Exposureto Low Levels of Ionizing Radiation,National Academy of Sciences,Washington, 1972.

6. Final Reports of the National CancerInstitute Ad Hoc Working Groups onMammography in Screening for BreastCancer and a Summary Report of TheirJoint Findings and Recommendations,J. Natl. Cancer Inst., Vol. 59, p.496-541, 1977.

7. objectives of Research Activities inBiology Branch, Chalk River NuclearLaboratories 1976. Atomic Energy ofCanada Limited, Report AECL-5613,1977. See also: Radiation Biologyfor the Non-Biologist. Atomic Energyof Canada Limited, Report AECL-5721,1978.

8. Lindhal, T., and Nyberg, B., Bio-chemistry, Vol. 11, p. 3610-3618(197.2) .

9. Crine, P., and Verly, W.G., Biochim.Biophys. Acta, Vol. 442, p. 50-57,1976.

10. Ames, B.N., McCann, J., and Yamasaki,E., Mutation Res., Vol. 31, p. 347-?64, 1?75.

11. Bridges, B.A., Nature, Vol. 261, p.195-200, 1976.

12. Meselson, M., and Russell, K., Originsof Human Cancer, Cold Spring HarborConferences on Cell Proliferation,p. 1473-1482, 1977.

13. Myers, D.K., Paterson, M.C., Gentner,N.E., Unrau, P., and Zimmermann, F.K.,Late Biological Effects of IonizingRadiation, Vol. 2, p. 447-459, Inter-national Atomic Energy Agency, Vienna,1978.

14. Paterson, M.C., Smith, B.P., Knight,P.A., and Anderson, A.K., Research inPhotobiology, p. 207-218, Plenum Press,New York, 1977.

15. Cleaver, J.E., Progress in GeneticToxicology, p. 29-42, Elsevier/North Holland Press, Amsterdam, 1977.

16. Paterson, M.C., Carcinogens:Identification and Mechanisms ofAction, p. 251-276, Raven Press,New York, 1979.

17. Myers, D.K., Summary and Proceedingsof a Biology Workshop on BiologicalRepair Mechanisms and Exposure Stand-ards, p. 106-144, Institute forEnergy Analysis, Oak Ridge, 1978.

18. Gentner, N.E., and Werner, M.M.,Molec. gen. Genetics, Vol. 164, p.31-37, 1978.

19. McGregor, J.F., J. Natl. Cancer Inst.,Vol. 56, p. 429-430, 1976.

20. Archer, V.E., Gillam, J.D., andWagoner, J.K., Annals N.Y. Acad.Sciences, Vol. 271, p. 280-293, 1976.

21. Selikoff, I.J., Hammond, E.C., andChurg, J., J. Amer. Med. Assoc,Vol. 204, p. 106-112, 1968.

22. Myers, D.K., Radiation Res., Vol. 65,p. 292-303, 1976.

23. Shellabarger, C.J., Stone, J.P., andHoltzman, S., Cancer Res., Vol. 36,p. 1019-1022, 1976.

24. Newcombe, H.B., Encyclopedia ofMedical Radiology, Vol. 2, part 1,p. 487-632, Springer-Verlag, 1966.

25. Grosch, D.S., Biological Effects ofRadiations, Blaisdell PublishingCompany, New York, 1965.

26. Ito, T., and Kobayashi, K., RadiationRes., Vol. 76, p. 139-144, 1978.

27. Baldwin, W.F., Radiation Res., Vol.49, p. 190-196, 1972.

28. Baldwin, W.F., and Knight, P.A.,Radiation Res., Vol. 63, p. 320-325,1975.

29. Bacq, Z.M., and Alexander, P.,Fundamentals of Radiobiology, 2ndedition, Pergamon Press, New York,1961.

30. Johns, H.E., The Physics of Radi-ology, revised 2nd edition, CharlesC. Thomas Publisher, Springfield,1964.

31. Nauman, C.H., Underbrink, A.G., andSparrow, A.H., Radiation Res., Vol.62, p. 79-96, 1975.

32. United Nations Scientific Committeeon the Effects of Atomic Radiation,Ionizing Radiation: Levels and Ef-fects, United Nations, New York,1972.

33. Mays, C.W., Lloyd, R.D., and Marshall,J.H., Proc. Third Intl. Congr., p.417-428, International RadiationProtection Association, 1973.

34. Rossi, H.H., and Mays, C.W., HealthPhysics, Vol. 34, p. 353-360, 1978.

35. Mancuso, T.F., Stewart, A., andKneale, G., Health Physics, Vol. 33,p. 369-385, 1977.

36. Kneale, G.w., Stewart, A.M., andMancuso, T.F., Late Biological Ef-fects of Ionizing Radiation, Vol. 1,p. 387-412, International Atomic

- 20 -

Energy Agency, Vienna, 1978.37. Dolphin, G.W., Technical Report No.

NRPB-R54, National RadiologicalProtection Board, Harwell, 1976.

38. Marks, S., Gilbert, E.S., andBreitenstein, R.D., Late BiologicalEffects of Ionizing Radiation, Vol. 1,p. 369-386, International AtomicEnergy Agency, Vienna, 1978.

39. Anderson, T.W., Ontario HydroMortality 1970-1977 (4th annualreport), Ontario Hydro, Toronto,1978.

40. Kneale, G.W., Stewart, A.M., andMancuso, T.F., Health Physics,Vol. 36, p. 87, 1979.

41. See, for example, the following:Mole, R.H., Lancet, Vol. 1, p. 1155-1156, 1978. Sanders, B.S., Lancet,Vol. 2, p. 840, 1978. Sanders, B.S.,Health Physics, Vol. 34, p. 521-538,1978. Anderson, T.W., Health Physics,Vol. 35, p. 743-750, 1978. Gertz,S.M., Health Physics, Vol. 35, p.723-724, 1978. Cohen, B.L., HealthPhysics, Vol. 35, p. 582-584, 1978.Reissland, J.A., Technical Report No.NRPB-R79, National Radiological Pro-tection Board, Harwell, 1978.

42. Rossi, H.H., Biological and Environ-mental Effects of Low-Level Radiation,Vol. 1, p. 245-251, InternationalAtomic Energy Agency, Vienna, 1976.

43. Petkau, A., and Chelack, W.S., Bio-chim. Biophys. Acta, Vol. 433, p.445-456, 1976.

44. Myers, D.K., Atomic Energy of CanadaLimited, Report AECL-5715, ChalkRiver, 1977.

45. Gentner, N.E., Molec. gen. Genetics,Vol. 154, p. 129-133, 1977.

46. Upton, A.C., Randolph, M.L., andConklin, J.W., Radiation Res., Vol.41, p. 467-491, 1970.

47. Bair, W.J., and Thompson, R.C.,Science, Vol. 183, p. 715-722, 1974.

48. Bryan, W.R., and Shimkin, M.B., J.Natl. Cancer Inst., Vol. 1, p. 807-833, 1941.

ISSN 0067 - 0367

To identify individual documents in the series

we have assigned an AECL- number to each.

Please refer to the AECL- number when re-

questing additional copies of this document

from

Scientific Document Distribution Office

Atomic Energy of Canada Limited

Chalk River, Ontario, Canada

KOJ 1JO

ISSN 0067 - 0367

Pour identifier les rapports individuels faisant

partie de cette série nous avons assigné

un numéro AECL- à chacun.

Veuillez faire mention du numéro AECL- si

vous demandez d'autres exemplaires de ce

rapport

au

Service da Distribution des Documents Officiels

L'Energie Atomique du Canada Limitée

Chalk River, Ontario, Canada

KOJ 1JO

Price $3.00 per copy Prix $3.00 par exemplaire

2049-79