- Email: [email protected]
DNA damage and repair: consequences on dose-responses
Mutation Research 464 Ž2000. 35–40 www.elsevier.comrlocatergentox Community address: www.elsevier.comrlocatermutres
DNA damage and repair: consequences on dose-responses Ethel Moustacchi
)
Institut Curie-Recherche, UMR 218 CNRS, LRC no. 1 CEA, 26 rue d’Ulm, 75248 Paris cedex 05, France Received 12 February 1999; received in revised form 10 May 1999; accepted 11 June 1999
Abstract Damage to DNA is considered to be the main initiating event by which genotoxins cause hereditary effects and cancer. Single or double strand breaks, bases modifications or deletions, intra- or interstrand DNA–DNA or DNA–protein cross-links constitute the major lesions formed in different proportions according to agents and to DNA sequence context. They can result in cell death or in mutational events which in turn may initiate malignant transformation. Normal cells are able to repair these lesions with fidelity or by introducing errors. Base excision ŽBER. and nucleotide excision ŽNER. repair are error-free processes acting on the simpler forms of DNA damage. A specialized form of BER involves the removal of mismatched DNA bases occurring as errors of DNA replication or from miscoding properties of damaged bases. Severe damage will be repaired according to several types of recombinational processes: homologous, illegitimate and site-specific recombination pathways. The loss of repair capacity as seen in a number of human genetic diseases and mutant cell lines leads to hypersensitivity to environmental agents. Repair-defective cells show qualitative Žmutation spectrum. and quantitative alterations in dose–effect relationships. For such repair-deficient systems, direct measurements at low doses are possible and the extrapolation from large to low doses fits well with the linear or the linear-quadratic no-threshold models. Extensive debate still takes place as to the shape of the dose–response relationships in the region at which genetic effects are not directly detectable in repair-proficient normal cells. Comparison of repair mutants and wild-type organisms pragmatically suggests that, for many genotoxins and tissues, very low doses may have no effect at all in normal cells. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Mutagenesis; Threshold; Repair pathways; Adaptive response
1. Introduction The biotope is exposed to genotoxic stress induced by exposures to chemical agents of the environment, to solar and to ionizing radiations as well as to endogenous natural damage due to DNA repli-
) Tel.: q33-1-42-34-67-10; fax: q33-1-46-33-30-16; e-mail: [email protected]
cation errors and accumulation of metabolic ‘‘cellular garbage’’. Organisms respond by different strategies which include error-free and error-prone repair of the damaged genome, cell death by cytotoxicity andror by apoptosis, and modulation of genes expression in relation, among various processes, with the cell cycle control. Cellular responses to chemical and physical agents result from the combination of these factors, the genetic background of the specific populations playing a predominant role in the dose– effects relationships.
1383-5718r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 3 - 5 7 1 8 Ž 9 9 . 0 0 1 6 4 - 3
36
E. Moustacchir Mutation Research 464 (2000) 35–40
2. The types of lesions Damage to DNA is considered to be the main initiating event by which chemical and physical genotoxins cause lethality, hereditary effects and cancer. Their interactions with DNA produces numerous types of modifications, many of these being chemically characterized and their relative abundance quantified on the basis of in vitro and in vivo measurements. Ideally, such studies lead to establish the precise number of defects per cell per unit dose of a given agent. It is, however, well-known that all agents produce simultaneously different types of lesions the frequencies of which and their biological consequences may vary according to a number of internal and external parameters ŽpH, temperature, etc... Damages are classified on the basis of their chemical structure and site, e.g., base modifications and loss Žwith the production of apurinic and apyrimidic sites., DNA–DNA intrastrand and interstrand crosslinks, DNA–proteins cross-links, single strand and double strand breaks. It is well-established that within a given segment of DNA, the sequence context influences both the chemical nature and the relative frequency of a given type of lesion Žfor review, see Ref. w1x.. In the case of radiations ŽUVC, UVA and UVB for the solar spectrum or high and low linear energy transfer or LET for ionizing radiations., the nature and relative proportion of DNA lesions are closely related to the physical characteristics of the radiations. For instance, the track structure of ionizing radiations plays an important role; one g-ray electron track crossing a standard 8 mm diameter nucleus will lead to 70 ionizations in average and is equivalent to 1 mGy absorbed dose. In contrast, in the same sized nucleus, one 4 MeV a-particle track will be associated to about 23.000 ionizations and is equivalent to 370 mGy. The complexity of the DNA damage increases with the density of ionizations reflected by the LET. As a consequence, in contrast to sparsely ionizing radiations Ž g-rays., densely ionizing radiations Ž a-particles. show more sensitive response, e.g., higher slope in survival and mutagenesis curves and loss of curve shouldered Žthreshold. region. The precise structure of the damage and its relative abundance Žfor a given genotoxin, sparsely dis-
tributed single lesions for low doses vs. clusters of lesions for high doses. will influence reparability. Indeed, to monitor damage and to maintain the genes without significant alterations constitute major functions in cells. In general, lesions located on one strand such as single strand breaks, intrastrand cross-links and base modifications are relatively easily removed and resynthesis of the lost material, using the undamaged strand as a template, takes place. In contrast, damage that affects both strands such as double strand breaks, cross-links and clustered lesions will be more difficult to repair and will require other enzymatic pathways for its resolution. The biological consequences and the outcome on dose–effect relationships will be closely dependent on the severity of the damages in relation with their reparability. It should be kept in mind that mammalian cells, on average, undergo about 10,000 measurable DNA modification events per cell per hour. This implies that organisms could not survive if they did not develop mechanism to respond to such heavy burden due to endogenous Ž500 g per day of internally produced oxygen for an adult human for instance. and exogenous exposures.
3. The enzymatic repair systems It is well-established that genetically controlled repair processes are generally common to all organisms from bacteria to man. Since the structure and function of repair genes are highly conserved, wellcharacterized bacterial and yeast mutants compared to isogenic wild-type micro-organisms have provided powerful tools to unravel analogous repair mechanisms in mammalian cells including human. It is important, however, to notice that in spite of the structural and functional analogies between species, the regulation of the repair genes activities and their interactions in metabolic networks may differ in different organisms. Indeed, in high eukaryotes, the final response to damage will be determined not only by the repair processes but also by other cellular functions which will optimize the possibility for cells restoration from damage. A typical example can be found when the lesions induce arrest in the cell cycle which will allow time for repair and thus
E. Moustacchir Mutation Research 464 (2000) 35–40
reduce the lethal and genetic consequences for a given dose of genotoxin. The tumour suppressor p53 protein, for example, appears to play a central role in this process w2,3x. The correct functioning of p53-dependent apoptotic pathway can be viewed as a complementary process to the repair of DNA lesions, eliminating damaged cells from the population hence, reducing the probability of accumulating cancer-promoting events w4x. Loss of one or both copies of the p53 gene in knockout mice has significant effects on sensitivity to spontaneous and lesion-induced cancers w5x. DNA double strand breakage produced either directly Žbleomycin, ionizing radiations, DNA topoisomerase inhibitors. or indirectly through the processing of base damage into transiently produced breaks ŽUV light. appear to be the signal for p53-dependent G1 arrest; only one to two double strand breaks may be sufficient for arrest w6x. The loss of repair capacity due to a reduction in damage detection or to an enzymatic deficiency in repair processes is observed in a number of human genetic disorders. This is associated to hypersensitivity to environmental agents, neurological anomalies, immune dysfunction and high predisposition to cancer. High spontaneous frequencies of chromosomal aberrations or of glycophorin A mutants as in ataxia telangiectasia and in Fanconi anemia or of sister chromatid exchanges as in Bloom syndrome reflect a genetic instability reminding one of the «mutator» phenotypes reported in bacteria and yeast in relation with mismatch repair deficiency Žfor review, see Ref. w1x.. This higher level of spontaneous genetic changes and the increased incidence of spontaneously arising neoplasms are amplified following exposures to genotoxins leading to survival and mutagenesis dose-responses with modified shapes Žloss of shoulders and steeper slopes. in comparison with responses from normal individuals. The simplest repair processes involve direct reversal of the damage in a one-step reaction. For instance, the O 6-methylguanine-methyltransferase directly removes methyl groups induced in DNA by alkylating carcinogens w7x or the photolyase in combination with visible light will directly monomerize the UV-induced pyrimidine dimers w8x. Most damage types require, however, the concerted action of a number of enzymes generally forming protein complexes which include specific interacting partners.
37
3.1. The error-free repair Relative to the classes of damages described above, three major error-free pathways have been identified. They eliminate lesions with high fidelity and consequently lead in the low dose range of exposures to undetectable effects. 3.1.1. The base excision repair (BER) pathway When a single base is damaged by monofunctional alkylating or oxidative agents for example, the concerted action of five enzymes is necessary to lead to correct restitution of the original genetic information. A DNA-glycosylase specific for different types of base damage removes the modified base, a DNA endonuclease cuts the DNA backbone at the abasic site, a phosphodiesterase removes the remaining sugar-phosphate moieties, a DNA polymerase fills the gap using the opposite complementary base as a template and finally, a DNA ligase allows rejoining Žfor review, see Ref. w9x.. The last three steps may also be used to repair single strand breaks directly produced by ionizing radiations or by radiomimetic antitumoral drugs Žbleomycin, neocarzinostatin, etc..; the ‘‘cleaning’’ of the break site is indeed necessary before the ligation step. This repair process limited to a unique damaged base or to single strand breaks is efficient and very rapid. 3.1.2. The nucleotide excision repair (NER) DNA containing a bulky adduct produced, for example, by polycyclic hydrocarbons or UV light Žpyrimidine dimers. will be processed by at least 13 enzymes. Recognition of the lesion, cutting of the strand at a defined distance on both sides of the damage, unwinding, removal of the altered strand followed, as in BER final steps, by polymerization and ligation are the different functions required to achieve NER. The coupling between transcription and NER is well-documented: actively expressing genes are repaired much faster than the non-expressing parts of the genome w10–12x. Several genes involved in NER have been found to be mutated in man giving rise to disorders such as xeroderma pigmentosum ŽXP. and Cockayne syndrome w13x. The XP individuals are extremely sensitive to sunlight and to chemical agents producing bulky adducts in DNA in terms of lethality, muta-
38
E. Moustacchir Mutation Research 464 (2000) 35–40
genicity and cancer-proneness. After exposures, the dose–effects relationships for mutation induction in cells derived from such patients do not demonstrate thresholds in contrast to normal cells w14x. Moreover, the deficiency in repair of DNA photoproducts in XP individuals is associated with an excess of cancer in sun-exposed regions whereas in age-matched healthy individuals these neoplasms are extremely rare Žfor review, see Ref. w15x.. The high level of protection afforded by high fidelity DNA repair processes is likely to lead for certain agents and tissues to a negligible risk, pragmatically interpreted as a threshold response. Mice that have been genetically manipulated to be deficient in specific DNA repair genes w16x will allow when compared to isogenic wild-type animals to draw more quantitative conclusions. 3.1.3. Homologous recombination Severe DNA damages may be produced directly by an agent causing for example DNA interstrand cross-links Žcisplatinum, mitomycin C for purine– purine or psoralens plus UVA for pyrimidine– pyrimidine cross-links. or may occur from the interaction with DNA replication of unrepaired single stranded damage or from production of two opposite strand breaks. In such cases after ‘‘cleaning’’ of the cuts by exonuclease excision, recombination enzymes will bring together the damaged site with the homologous undamaged DNA duplex. The broken 3X end of the DNA strand invades an unbroken double stranded homologue and resynthesis on this intact template reforms the damaged strand. Separation of the patched joint product of this reaction will then require the activity of enzymes cutting and rejoining the newly synthesized DNA strands. In yeast, homologous recombination which is under the control of the RAD52 group of genes is the predominant method for DNA double strand break repair: rad52 mutants are defective in such repair process and in homologous recombination of genetic markers w17x. 3.2. The error-prone repair The LEXA, RECA dependent inducible SOS repair described in bacteria constitute the paradigm of a mutagenic repair pathway. The SOS response to genotoxic stress leads to the induction of at least 20
genes involved in mutagenesis, recombination, NER and strand breaks repair as well as genes necessary for cell division and respiratory functions Žfor review, see Ref. w1x.. An analogous coordinated response has not been confirmed in eukaryotes. Although more than 50 genes are induced in mammalian cells by damaging signals including transcription factors, certain proto-oncogenes and oncogene suppressors, detoxification enzymes, proteins involved in DNA metabolism, etc., the precise mechanistic relationship with mutagenic repair is not understood and a coordinated pathway does not yet emerge. Non-homologous illegitimate recombination is the predominant alternative pathway for rejoining altered DNA sequences in human cells. Analysis of genomic breakpoints of induced deletions and rearrangement demonstrates little sequence homology at such genomic sites Žfor review, see Ref. w18x.. An important discovery is that the enzymes involved in repairing DNA double strand breaks also participate in the site-specific recombination process, VŽD.J immunesystem recombination, which assembles functional immune genes Žimmunoglobulins and T-cell receptors. from separate genomic regions through somatic genes rearrangements. Mice homozygous for the scid Ž severe combined immune deficient. mutation lack functional T- and B-cells because of a defect in VŽD.J recombination and they demonstrate hypersensitivity to DNA strand breaks inducing agents such as ionizing radiations and radiomimetic agents. The identification of the genes and genes products complementing this type of defect as well as their mode of action are presently under study by numerous teams. More specifically, a misrejoining process has been found to occur by a non-conservative mechanism which entails deletion of DNA bases between short direct repeat sequences such that either one of the repeats is lost or occasionally one insertion takes place at the deletion site. Such misrejoining mechanism appears to prevail when the RAD52-dependent homologous recombination and the Ku-dependent non-homologous recombination are knocked out w19,20x. This last error-prone pathway appears to be defective in at least two human genetic diseases associated to hypersensitivity to genotoxic agents such as ataxia telangiectasia w21x and Fanconi anemia w22x. An increased level of non-conservative recombination leads to enhancement in deletions and
E. Moustacchir Mutation Research 464 (2000) 35–40
rearrangements in both syndromes and tends to ultimately lead to cancer-proneness.
4. Influence of repair on dose–effect relationships It is evident that repair processes influence the sensitivity of organisms and their derived cell lines to chemical and physical genotoxic agents. If repair is defective, the slope of the survival curve is increased and where normal cells demonstrate a shouldered curve, the repair-deficient cell lines often show a loss of this characteristic. There is very good evidence in yeast, based on comparison of responses of mutants in well-defined repair pathways and isogenic wild-type strains, that these quantitative aspects of not only survival but also mutagenesis and genetic recombination Žconversion and crossing over. are accountable primarily in terms of the repair of DNA lesions. In mammalian cells, it is still difficult to establish similar quantitative correlations. Pre-exposure to low doses before subsequent treatment with high doses has been shown to decrease the frequency of chromosomal damage, mutation and morphological transformation Žfor review, see Ref. w23x.. The features of this adaptive response indicate that low levels of DNA damage Žessentially directly or indirectly induced breaks. act as a signal for accelerated detoxification and repair in mammalian cells. If a threshold should exist, the most coherent explanation is likely to rely on adaptive effects w24x. In view of the possible consequences for occupationally exposed individuals and the incidence on the debated threshold effects, it is clear that the adaptive response merits more studies. Reduction in the efficiency or fidelity of repair clearly leads to an increase in genetic change as exemplified by the analysis of one of the most common human cancer and probably the most frequent form of hereditary neoplasia that is the hereditary non-polyposis colon cancer. A defect in mismatch repair ŽMut HLS system., a form of BER, has been clearly established for this type of cancer and is associated to high mutation rates w25,26x. The majority of DNA damage is repaired or fixed as mutations Ži.e., misrepair. within a few hours or days after exposure. However, there are more and more indications that cellular responses including
39
genetic changes continue to occur for much longer periods and even over many cell generations w27x. More knowledge of this genetic instability related to repair functions is required. Additionally, studies of spectrum of repair capacities in the human population as reflected by interindividual variations in responses to exposures w28x and the presence of heterozygous carriers of defective repair genes will have to be seriously considered to assess mutability on a population and on an individual basis.
5. Conclusion To estimate the effects of low doses of genotoxins, a linear extrapolation from high doses to zero is generally used. This approach implies that pathways of damage of high and low doses are identical and that organisms have no defense against agents that damage DNA. This last assumption disregards the existence of genetically governed error-free repair of lesions and assumes that no dose, even small, is safe. This costly position has been adopted by regulatory international committees for risk estimates related to low radiation doses exposures. It does not mean that it is based on the soundest possible science but reflects acknowledgement of the current limits to scientific certainty w29x. The best that can be done in the future is to narrow the limits on uncertainties by unraveling the mechanisms underlying the dose responses to genotoxic agents as well as by monitoring for long-term health effects.
Acknowledgements Work in Ethel Moustacchi laboratory is supported by grants from CNRS, Institut Curie, CEA, CEE ŽBruxelles., Association pour la Recherche sur le Cancer, Ligue contre le Cancer and ACC SV no. 8 ŽMinistere ` de la Recherche, France.. The author thanks D. Chardonnieras for technical assistance.
References w1x E.C. Friedberg, G.C. Walker, W. Siede, DNA Repair and Mutagenesis, Am. Soc. for Microbiology Press, Washington, DC, 1995.
40
E. Moustacchir Mutation Research 464 (2000) 35–40
w2x M.B. Kastan, O. Onyekwere, D. Sidransky et al., Participation of p53 protein in the cellular response to DNA damage, Cancer Res. 51 Ž1991. 6304–6311. w3x D.P. Lane, p53, guardian of the genome, Nature 358 Ž1992. 15–16. w4x G.T. Williams, Programmed cell death-apoptosis and oncogenesis, Cell 65 Ž1991. 1097–1098. w5x L.A. Donehower, M. Harvey, B.L. Slagle, Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours, Nature 356 Ž1992. 215–221. w6x L.C. Huangn, K.C. Clarkin, G.M. Wahl, Sensitivity and selectivity of the DNA damage sensor responsible for activating p53-dependent G1 arrest, Proc. Natl. Acad. Sci. U.S.A. 93 Ž1996. 4827–4832. w7x A.E. Pegg, Properties of O 6-alkylguanine-DNA transferases, Mutat. Res. 233 Ž1990. 165–175. w8x A. Sancar, G.B. Sancar, DNA repair enzymes, Annu. Rev. Biochem. 57 Ž1988. 29–67. w9x E. Seeberg, L. Eide, M. Bjoras, The base excision repair pathway, Trends Biochem. Sci. 20 Ž1995. 391–397. w10x A. Sancar, DNA excision repair, Annu. Rev. Biochem. 65 Ž1996. 43–81. w11x R.D. Wood, DNA repair in eukaryotes, Annu. Rev. Biochem. 65 Ž1996. 135–167. w12x P.C. Hanawalt, DNA repair comes of age, Mutat. Res. 336 Ž1995. 101–113. w13x J.H.J. Hoeijmakers, Nucleotide excision repair: II. From yeast to mammals, Trends Genet. 9 Ž1993. 211–217. w14x V.M. Maher, J.J. McCormick, Effect of DNA repair on the cytotoxicity and mutagenicity of UV irradiation and chemical carcinogens in normal and xeroderma pigmentosum cells, in: J.M. Yuhas, R.W.X. Tennant, J.D. Regan ŽEds.., Biological of Radiation Carcinogenesis, Raven Press, New York, 1976, pp. 129–145. w15x K.H. Kraemer, D.D. Levy, C.N. Parris, Xeroderma pigmentosum and related disorders: examining the linkage between defective DNA repair and cancer, J. Invest. Dermatol. 103 Ž1994. 96–101. w16x E.C. Friedberg, L.B. Meira, Database of mouse strains carrying targeted mutations in genes affecting cellular responses to DNA damage: version 3, Mutat. Res., DNA Repair 433 Ž1999. 69–87. w17x J.C. Game, DNA double strand breaks and the RAD50 – RAD57 genes in Saccharomyces cereÕisiae, Cancer Biol. 4 Ž1993. 73–83.
w18x S. Vamvakas, E.H. Vock, W.K. Lutz, On the role of DNA double-strand breaks in toxicity and carcinogenesis, Crit. Rev. Toxicol. 27 Ž1997. 155–174. w19x S.J. Boulton, S.P. Jackson, Saccharomyces cereÕisiae Ku70 potentiates illegitimate DNA double strand break repair and serves as a barrier to error-prone DNA repair pathways, EMBO J. 15 Ž1996. 5093–5103. w20x S.J. Boulton, S.P. Jackson, Identification of a Saccharomyces cereÕisiae Ku80 homologue: roles in DNA maintenance, Nucleic Acids Res. 24 Ž1996. 4639–4648. w21x J. Thacker, J. Chalk, A. Ganesh et al., A mechanism for deletion formation in DNA by human cell extracts: the involvement of short sequence repeats, Nucleic Acids Res. 20 Ž1992. 6183–6188. w22x M. Escarceller, M. Buchwald, B.K. Singleton, P.A. Jeggo, S.P. Jackson, E. Moustacchi, D. Papadopoulo, Fanconi anemia C gene product plays a role in the fidelity of blunt DNA end-joining, J. Mol. Biol. 279 Ž1998. 375–385. w23x O. Rigaud, E. Moustacchi, Radioadaptation for gene mutation and the possible molecular mechanisms of the adaptive response, Mutat. Res. 358 Ž1996. 127–134. w24x X.C. Le, J.Z. Xing, J. Lee, S.A. Leadon, M. Weinfeld, Inducible repair of thymine glycol detected by an ultrasensitive assay for DNA damage, Science 280 Ž1998. 1066–1068. w25x F.S. Leach, N.C. Nicolaides, N. Papadopoulos et al., Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer, Cell 75 Ž1993. 1215–1225. w26x G. Marra, C.R. Boland, Hereditary non-polyposis colorectal cancer: the syndrome, the genes, and historical perspectives, J. Natl. Cancer Inst. 87 Ž1995. 1114–1125. w27x K. Holmberg, S. Falt, A. Johansson et al., Clonal chromosome aberrations and genomic instability in X-irradiated human T-lymphocyte cultures, Mutat. Res. 286 Ž1993. 321– 330. w28x F. Leprat, C. Alapetite, F. Rosselli, A. Ridet, M. Schlumberger, A. Sarasin, H.G. Suarez, E. Moustacchi, Impaired DNA repair as assessed by the ‘‘comet’’ assay in patients with thyroid tumors after a history of radiation therapy: a preliminary study, Int. J. Rad. Oncol. Biol. Phys. 40 Ž1998. 1019–1026. w29x UNSCEAR Report ArAC.82R.598, Biological effects at low radiation doses: models, mechanisms and uncertainties. 48th Session of United Nations Scientific Committee on the Effects of Atomic Radiation, 12–16 April 1999 Žin press..
DNA damage and repair: consequences on dose-responses Ethel Moustacchi
)
Institut Curie-Recherche, UMR 218 CNRS, LRC no. 1 CEA, 26 rue d’Ulm, 75248 Paris cedex 05, France Received 12 February 1999; received in revised form 10 May 1999; accepted 11 June 1999
Abstract Damage to DNA is considered to be the main initiating event by which genotoxins cause hereditary effects and cancer. Single or double strand breaks, bases modifications or deletions, intra- or interstrand DNA–DNA or DNA–protein cross-links constitute the major lesions formed in different proportions according to agents and to DNA sequence context. They can result in cell death or in mutational events which in turn may initiate malignant transformation. Normal cells are able to repair these lesions with fidelity or by introducing errors. Base excision ŽBER. and nucleotide excision ŽNER. repair are error-free processes acting on the simpler forms of DNA damage. A specialized form of BER involves the removal of mismatched DNA bases occurring as errors of DNA replication or from miscoding properties of damaged bases. Severe damage will be repaired according to several types of recombinational processes: homologous, illegitimate and site-specific recombination pathways. The loss of repair capacity as seen in a number of human genetic diseases and mutant cell lines leads to hypersensitivity to environmental agents. Repair-defective cells show qualitative Žmutation spectrum. and quantitative alterations in dose–effect relationships. For such repair-deficient systems, direct measurements at low doses are possible and the extrapolation from large to low doses fits well with the linear or the linear-quadratic no-threshold models. Extensive debate still takes place as to the shape of the dose–response relationships in the region at which genetic effects are not directly detectable in repair-proficient normal cells. Comparison of repair mutants and wild-type organisms pragmatically suggests that, for many genotoxins and tissues, very low doses may have no effect at all in normal cells. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Mutagenesis; Threshold; Repair pathways; Adaptive response
1. Introduction The biotope is exposed to genotoxic stress induced by exposures to chemical agents of the environment, to solar and to ionizing radiations as well as to endogenous natural damage due to DNA repli-
) Tel.: q33-1-42-34-67-10; fax: q33-1-46-33-30-16; e-mail: [email protected]
cation errors and accumulation of metabolic ‘‘cellular garbage’’. Organisms respond by different strategies which include error-free and error-prone repair of the damaged genome, cell death by cytotoxicity andror by apoptosis, and modulation of genes expression in relation, among various processes, with the cell cycle control. Cellular responses to chemical and physical agents result from the combination of these factors, the genetic background of the specific populations playing a predominant role in the dose– effects relationships.
1383-5718r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 3 - 5 7 1 8 Ž 9 9 . 0 0 1 6 4 - 3
36
E. Moustacchir Mutation Research 464 (2000) 35–40
2. The types of lesions Damage to DNA is considered to be the main initiating event by which chemical and physical genotoxins cause lethality, hereditary effects and cancer. Their interactions with DNA produces numerous types of modifications, many of these being chemically characterized and their relative abundance quantified on the basis of in vitro and in vivo measurements. Ideally, such studies lead to establish the precise number of defects per cell per unit dose of a given agent. It is, however, well-known that all agents produce simultaneously different types of lesions the frequencies of which and their biological consequences may vary according to a number of internal and external parameters ŽpH, temperature, etc... Damages are classified on the basis of their chemical structure and site, e.g., base modifications and loss Žwith the production of apurinic and apyrimidic sites., DNA–DNA intrastrand and interstrand crosslinks, DNA–proteins cross-links, single strand and double strand breaks. It is well-established that within a given segment of DNA, the sequence context influences both the chemical nature and the relative frequency of a given type of lesion Žfor review, see Ref. w1x.. In the case of radiations ŽUVC, UVA and UVB for the solar spectrum or high and low linear energy transfer or LET for ionizing radiations., the nature and relative proportion of DNA lesions are closely related to the physical characteristics of the radiations. For instance, the track structure of ionizing radiations plays an important role; one g-ray electron track crossing a standard 8 mm diameter nucleus will lead to 70 ionizations in average and is equivalent to 1 mGy absorbed dose. In contrast, in the same sized nucleus, one 4 MeV a-particle track will be associated to about 23.000 ionizations and is equivalent to 370 mGy. The complexity of the DNA damage increases with the density of ionizations reflected by the LET. As a consequence, in contrast to sparsely ionizing radiations Ž g-rays., densely ionizing radiations Ž a-particles. show more sensitive response, e.g., higher slope in survival and mutagenesis curves and loss of curve shouldered Žthreshold. region. The precise structure of the damage and its relative abundance Žfor a given genotoxin, sparsely dis-
tributed single lesions for low doses vs. clusters of lesions for high doses. will influence reparability. Indeed, to monitor damage and to maintain the genes without significant alterations constitute major functions in cells. In general, lesions located on one strand such as single strand breaks, intrastrand cross-links and base modifications are relatively easily removed and resynthesis of the lost material, using the undamaged strand as a template, takes place. In contrast, damage that affects both strands such as double strand breaks, cross-links and clustered lesions will be more difficult to repair and will require other enzymatic pathways for its resolution. The biological consequences and the outcome on dose–effect relationships will be closely dependent on the severity of the damages in relation with their reparability. It should be kept in mind that mammalian cells, on average, undergo about 10,000 measurable DNA modification events per cell per hour. This implies that organisms could not survive if they did not develop mechanism to respond to such heavy burden due to endogenous Ž500 g per day of internally produced oxygen for an adult human for instance. and exogenous exposures.
3. The enzymatic repair systems It is well-established that genetically controlled repair processes are generally common to all organisms from bacteria to man. Since the structure and function of repair genes are highly conserved, wellcharacterized bacterial and yeast mutants compared to isogenic wild-type micro-organisms have provided powerful tools to unravel analogous repair mechanisms in mammalian cells including human. It is important, however, to notice that in spite of the structural and functional analogies between species, the regulation of the repair genes activities and their interactions in metabolic networks may differ in different organisms. Indeed, in high eukaryotes, the final response to damage will be determined not only by the repair processes but also by other cellular functions which will optimize the possibility for cells restoration from damage. A typical example can be found when the lesions induce arrest in the cell cycle which will allow time for repair and thus
E. Moustacchir Mutation Research 464 (2000) 35–40
reduce the lethal and genetic consequences for a given dose of genotoxin. The tumour suppressor p53 protein, for example, appears to play a central role in this process w2,3x. The correct functioning of p53-dependent apoptotic pathway can be viewed as a complementary process to the repair of DNA lesions, eliminating damaged cells from the population hence, reducing the probability of accumulating cancer-promoting events w4x. Loss of one or both copies of the p53 gene in knockout mice has significant effects on sensitivity to spontaneous and lesion-induced cancers w5x. DNA double strand breakage produced either directly Žbleomycin, ionizing radiations, DNA topoisomerase inhibitors. or indirectly through the processing of base damage into transiently produced breaks ŽUV light. appear to be the signal for p53-dependent G1 arrest; only one to two double strand breaks may be sufficient for arrest w6x. The loss of repair capacity due to a reduction in damage detection or to an enzymatic deficiency in repair processes is observed in a number of human genetic disorders. This is associated to hypersensitivity to environmental agents, neurological anomalies, immune dysfunction and high predisposition to cancer. High spontaneous frequencies of chromosomal aberrations or of glycophorin A mutants as in ataxia telangiectasia and in Fanconi anemia or of sister chromatid exchanges as in Bloom syndrome reflect a genetic instability reminding one of the «mutator» phenotypes reported in bacteria and yeast in relation with mismatch repair deficiency Žfor review, see Ref. w1x.. This higher level of spontaneous genetic changes and the increased incidence of spontaneously arising neoplasms are amplified following exposures to genotoxins leading to survival and mutagenesis dose-responses with modified shapes Žloss of shoulders and steeper slopes. in comparison with responses from normal individuals. The simplest repair processes involve direct reversal of the damage in a one-step reaction. For instance, the O 6-methylguanine-methyltransferase directly removes methyl groups induced in DNA by alkylating carcinogens w7x or the photolyase in combination with visible light will directly monomerize the UV-induced pyrimidine dimers w8x. Most damage types require, however, the concerted action of a number of enzymes generally forming protein complexes which include specific interacting partners.
37
3.1. The error-free repair Relative to the classes of damages described above, three major error-free pathways have been identified. They eliminate lesions with high fidelity and consequently lead in the low dose range of exposures to undetectable effects. 3.1.1. The base excision repair (BER) pathway When a single base is damaged by monofunctional alkylating or oxidative agents for example, the concerted action of five enzymes is necessary to lead to correct restitution of the original genetic information. A DNA-glycosylase specific for different types of base damage removes the modified base, a DNA endonuclease cuts the DNA backbone at the abasic site, a phosphodiesterase removes the remaining sugar-phosphate moieties, a DNA polymerase fills the gap using the opposite complementary base as a template and finally, a DNA ligase allows rejoining Žfor review, see Ref. w9x.. The last three steps may also be used to repair single strand breaks directly produced by ionizing radiations or by radiomimetic antitumoral drugs Žbleomycin, neocarzinostatin, etc..; the ‘‘cleaning’’ of the break site is indeed necessary before the ligation step. This repair process limited to a unique damaged base or to single strand breaks is efficient and very rapid. 3.1.2. The nucleotide excision repair (NER) DNA containing a bulky adduct produced, for example, by polycyclic hydrocarbons or UV light Žpyrimidine dimers. will be processed by at least 13 enzymes. Recognition of the lesion, cutting of the strand at a defined distance on both sides of the damage, unwinding, removal of the altered strand followed, as in BER final steps, by polymerization and ligation are the different functions required to achieve NER. The coupling between transcription and NER is well-documented: actively expressing genes are repaired much faster than the non-expressing parts of the genome w10–12x. Several genes involved in NER have been found to be mutated in man giving rise to disorders such as xeroderma pigmentosum ŽXP. and Cockayne syndrome w13x. The XP individuals are extremely sensitive to sunlight and to chemical agents producing bulky adducts in DNA in terms of lethality, muta-
38
E. Moustacchir Mutation Research 464 (2000) 35–40
genicity and cancer-proneness. After exposures, the dose–effects relationships for mutation induction in cells derived from such patients do not demonstrate thresholds in contrast to normal cells w14x. Moreover, the deficiency in repair of DNA photoproducts in XP individuals is associated with an excess of cancer in sun-exposed regions whereas in age-matched healthy individuals these neoplasms are extremely rare Žfor review, see Ref. w15x.. The high level of protection afforded by high fidelity DNA repair processes is likely to lead for certain agents and tissues to a negligible risk, pragmatically interpreted as a threshold response. Mice that have been genetically manipulated to be deficient in specific DNA repair genes w16x will allow when compared to isogenic wild-type animals to draw more quantitative conclusions. 3.1.3. Homologous recombination Severe DNA damages may be produced directly by an agent causing for example DNA interstrand cross-links Žcisplatinum, mitomycin C for purine– purine or psoralens plus UVA for pyrimidine– pyrimidine cross-links. or may occur from the interaction with DNA replication of unrepaired single stranded damage or from production of two opposite strand breaks. In such cases after ‘‘cleaning’’ of the cuts by exonuclease excision, recombination enzymes will bring together the damaged site with the homologous undamaged DNA duplex. The broken 3X end of the DNA strand invades an unbroken double stranded homologue and resynthesis on this intact template reforms the damaged strand. Separation of the patched joint product of this reaction will then require the activity of enzymes cutting and rejoining the newly synthesized DNA strands. In yeast, homologous recombination which is under the control of the RAD52 group of genes is the predominant method for DNA double strand break repair: rad52 mutants are defective in such repair process and in homologous recombination of genetic markers w17x. 3.2. The error-prone repair The LEXA, RECA dependent inducible SOS repair described in bacteria constitute the paradigm of a mutagenic repair pathway. The SOS response to genotoxic stress leads to the induction of at least 20
genes involved in mutagenesis, recombination, NER and strand breaks repair as well as genes necessary for cell division and respiratory functions Žfor review, see Ref. w1x.. An analogous coordinated response has not been confirmed in eukaryotes. Although more than 50 genes are induced in mammalian cells by damaging signals including transcription factors, certain proto-oncogenes and oncogene suppressors, detoxification enzymes, proteins involved in DNA metabolism, etc., the precise mechanistic relationship with mutagenic repair is not understood and a coordinated pathway does not yet emerge. Non-homologous illegitimate recombination is the predominant alternative pathway for rejoining altered DNA sequences in human cells. Analysis of genomic breakpoints of induced deletions and rearrangement demonstrates little sequence homology at such genomic sites Žfor review, see Ref. w18x.. An important discovery is that the enzymes involved in repairing DNA double strand breaks also participate in the site-specific recombination process, VŽD.J immunesystem recombination, which assembles functional immune genes Žimmunoglobulins and T-cell receptors. from separate genomic regions through somatic genes rearrangements. Mice homozygous for the scid Ž severe combined immune deficient. mutation lack functional T- and B-cells because of a defect in VŽD.J recombination and they demonstrate hypersensitivity to DNA strand breaks inducing agents such as ionizing radiations and radiomimetic agents. The identification of the genes and genes products complementing this type of defect as well as their mode of action are presently under study by numerous teams. More specifically, a misrejoining process has been found to occur by a non-conservative mechanism which entails deletion of DNA bases between short direct repeat sequences such that either one of the repeats is lost or occasionally one insertion takes place at the deletion site. Such misrejoining mechanism appears to prevail when the RAD52-dependent homologous recombination and the Ku-dependent non-homologous recombination are knocked out w19,20x. This last error-prone pathway appears to be defective in at least two human genetic diseases associated to hypersensitivity to genotoxic agents such as ataxia telangiectasia w21x and Fanconi anemia w22x. An increased level of non-conservative recombination leads to enhancement in deletions and
E. Moustacchir Mutation Research 464 (2000) 35–40
rearrangements in both syndromes and tends to ultimately lead to cancer-proneness.
4. Influence of repair on dose–effect relationships It is evident that repair processes influence the sensitivity of organisms and their derived cell lines to chemical and physical genotoxic agents. If repair is defective, the slope of the survival curve is increased and where normal cells demonstrate a shouldered curve, the repair-deficient cell lines often show a loss of this characteristic. There is very good evidence in yeast, based on comparison of responses of mutants in well-defined repair pathways and isogenic wild-type strains, that these quantitative aspects of not only survival but also mutagenesis and genetic recombination Žconversion and crossing over. are accountable primarily in terms of the repair of DNA lesions. In mammalian cells, it is still difficult to establish similar quantitative correlations. Pre-exposure to low doses before subsequent treatment with high doses has been shown to decrease the frequency of chromosomal damage, mutation and morphological transformation Žfor review, see Ref. w23x.. The features of this adaptive response indicate that low levels of DNA damage Žessentially directly or indirectly induced breaks. act as a signal for accelerated detoxification and repair in mammalian cells. If a threshold should exist, the most coherent explanation is likely to rely on adaptive effects w24x. In view of the possible consequences for occupationally exposed individuals and the incidence on the debated threshold effects, it is clear that the adaptive response merits more studies. Reduction in the efficiency or fidelity of repair clearly leads to an increase in genetic change as exemplified by the analysis of one of the most common human cancer and probably the most frequent form of hereditary neoplasia that is the hereditary non-polyposis colon cancer. A defect in mismatch repair ŽMut HLS system., a form of BER, has been clearly established for this type of cancer and is associated to high mutation rates w25,26x. The majority of DNA damage is repaired or fixed as mutations Ži.e., misrepair. within a few hours or days after exposure. However, there are more and more indications that cellular responses including
39
genetic changes continue to occur for much longer periods and even over many cell generations w27x. More knowledge of this genetic instability related to repair functions is required. Additionally, studies of spectrum of repair capacities in the human population as reflected by interindividual variations in responses to exposures w28x and the presence of heterozygous carriers of defective repair genes will have to be seriously considered to assess mutability on a population and on an individual basis.
5. Conclusion To estimate the effects of low doses of genotoxins, a linear extrapolation from high doses to zero is generally used. This approach implies that pathways of damage of high and low doses are identical and that organisms have no defense against agents that damage DNA. This last assumption disregards the existence of genetically governed error-free repair of lesions and assumes that no dose, even small, is safe. This costly position has been adopted by regulatory international committees for risk estimates related to low radiation doses exposures. It does not mean that it is based on the soundest possible science but reflects acknowledgement of the current limits to scientific certainty w29x. The best that can be done in the future is to narrow the limits on uncertainties by unraveling the mechanisms underlying the dose responses to genotoxic agents as well as by monitoring for long-term health effects.
Acknowledgements Work in Ethel Moustacchi laboratory is supported by grants from CNRS, Institut Curie, CEA, CEE ŽBruxelles., Association pour la Recherche sur le Cancer, Ligue contre le Cancer and ACC SV no. 8 ŽMinistere ` de la Recherche, France.. The author thanks D. Chardonnieras for technical assistance.
References w1x E.C. Friedberg, G.C. Walker, W. Siede, DNA Repair and Mutagenesis, Am. Soc. for Microbiology Press, Washington, DC, 1995.
40
E. Moustacchir Mutation Research 464 (2000) 35–40
w2x M.B. Kastan, O. Onyekwere, D. Sidransky et al., Participation of p53 protein in the cellular response to DNA damage, Cancer Res. 51 Ž1991. 6304–6311. w3x D.P. Lane, p53, guardian of the genome, Nature 358 Ž1992. 15–16. w4x G.T. Williams, Programmed cell death-apoptosis and oncogenesis, Cell 65 Ž1991. 1097–1098. w5x L.A. Donehower, M. Harvey, B.L. Slagle, Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours, Nature 356 Ž1992. 215–221. w6x L.C. Huangn, K.C. Clarkin, G.M. Wahl, Sensitivity and selectivity of the DNA damage sensor responsible for activating p53-dependent G1 arrest, Proc. Natl. Acad. Sci. U.S.A. 93 Ž1996. 4827–4832. w7x A.E. Pegg, Properties of O 6-alkylguanine-DNA transferases, Mutat. Res. 233 Ž1990. 165–175. w8x A. Sancar, G.B. Sancar, DNA repair enzymes, Annu. Rev. Biochem. 57 Ž1988. 29–67. w9x E. Seeberg, L. Eide, M. Bjoras, The base excision repair pathway, Trends Biochem. Sci. 20 Ž1995. 391–397. w10x A. Sancar, DNA excision repair, Annu. Rev. Biochem. 65 Ž1996. 43–81. w11x R.D. Wood, DNA repair in eukaryotes, Annu. Rev. Biochem. 65 Ž1996. 135–167. w12x P.C. Hanawalt, DNA repair comes of age, Mutat. Res. 336 Ž1995. 101–113. w13x J.H.J. Hoeijmakers, Nucleotide excision repair: II. From yeast to mammals, Trends Genet. 9 Ž1993. 211–217. w14x V.M. Maher, J.J. McCormick, Effect of DNA repair on the cytotoxicity and mutagenicity of UV irradiation and chemical carcinogens in normal and xeroderma pigmentosum cells, in: J.M. Yuhas, R.W.X. Tennant, J.D. Regan ŽEds.., Biological of Radiation Carcinogenesis, Raven Press, New York, 1976, pp. 129–145. w15x K.H. Kraemer, D.D. Levy, C.N. Parris, Xeroderma pigmentosum and related disorders: examining the linkage between defective DNA repair and cancer, J. Invest. Dermatol. 103 Ž1994. 96–101. w16x E.C. Friedberg, L.B. Meira, Database of mouse strains carrying targeted mutations in genes affecting cellular responses to DNA damage: version 3, Mutat. Res., DNA Repair 433 Ž1999. 69–87. w17x J.C. Game, DNA double strand breaks and the RAD50 – RAD57 genes in Saccharomyces cereÕisiae, Cancer Biol. 4 Ž1993. 73–83.
w18x S. Vamvakas, E.H. Vock, W.K. Lutz, On the role of DNA double-strand breaks in toxicity and carcinogenesis, Crit. Rev. Toxicol. 27 Ž1997. 155–174. w19x S.J. Boulton, S.P. Jackson, Saccharomyces cereÕisiae Ku70 potentiates illegitimate DNA double strand break repair and serves as a barrier to error-prone DNA repair pathways, EMBO J. 15 Ž1996. 5093–5103. w20x S.J. Boulton, S.P. Jackson, Identification of a Saccharomyces cereÕisiae Ku80 homologue: roles in DNA maintenance, Nucleic Acids Res. 24 Ž1996. 4639–4648. w21x J. Thacker, J. Chalk, A. Ganesh et al., A mechanism for deletion formation in DNA by human cell extracts: the involvement of short sequence repeats, Nucleic Acids Res. 20 Ž1992. 6183–6188. w22x M. Escarceller, M. Buchwald, B.K. Singleton, P.A. Jeggo, S.P. Jackson, E. Moustacchi, D. Papadopoulo, Fanconi anemia C gene product plays a role in the fidelity of blunt DNA end-joining, J. Mol. Biol. 279 Ž1998. 375–385. w23x O. Rigaud, E. Moustacchi, Radioadaptation for gene mutation and the possible molecular mechanisms of the adaptive response, Mutat. Res. 358 Ž1996. 127–134. w24x X.C. Le, J.Z. Xing, J. Lee, S.A. Leadon, M. Weinfeld, Inducible repair of thymine glycol detected by an ultrasensitive assay for DNA damage, Science 280 Ž1998. 1066–1068. w25x F.S. Leach, N.C. Nicolaides, N. Papadopoulos et al., Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer, Cell 75 Ž1993. 1215–1225. w26x G. Marra, C.R. Boland, Hereditary non-polyposis colorectal cancer: the syndrome, the genes, and historical perspectives, J. Natl. Cancer Inst. 87 Ž1995. 1114–1125. w27x K. Holmberg, S. Falt, A. Johansson et al., Clonal chromosome aberrations and genomic instability in X-irradiated human T-lymphocyte cultures, Mutat. Res. 286 Ž1993. 321– 330. w28x F. Leprat, C. Alapetite, F. Rosselli, A. Ridet, M. Schlumberger, A. Sarasin, H.G. Suarez, E. Moustacchi, Impaired DNA repair as assessed by the ‘‘comet’’ assay in patients with thyroid tumors after a history of radiation therapy: a preliminary study, Int. J. Rad. Oncol. Biol. Phys. 40 Ž1998. 1019–1026. w29x UNSCEAR Report ArAC.82R.598, Biological effects at low radiation doses: models, mechanisms and uncertainties. 48th Session of United Nations Scientific Committee on the Effects of Atomic Radiation, 12–16 April 1999 Žin press..