(CANCER RESEARCH (SUPPL.) 54. ]890s-18'Ms. April 1, 1994)

Reactive Oxygen Species in Tumorigenesis1

Daniel I. Feig, Thomas M. Reid, and Lawrence A. Loeb

Joseph (¡ottstein Memorial Cancer Research Laboratory, Departments (if Biochemistry ¡D. I. F.] and Pathology /T. M. R., L. A. L.¡, University of Washington,Seattle. Washington 98195

Abstract

In this review we will summarize recent data on reactive oxygenspecies-induced mutagenesis and consider its relationship to tumorigene-sis in humans. With the use of a single-stranded DNA template it has beenpossible to correlate oxygen radical-induced chemical alterations at spe

cific nucleotides with the types of mutations that occur when these alteredbases are copied by DNA polymerases. This has allowed us to identify thetypes of mutations that occur secondary to a variety of oxidative stressesand study several of the mechanisms by which they arise. The mostfrequent mutations that result from reactive oxygen species-induced dam

age to UNA in bacteria are C to T transitions. These mutations, however,are not pathoneumonic for mutagenesis by oxygen-free radicals since they

result from UNA damage caused by other genotoxic agents as well as byUNA polymerase errors. One type of mutation, a tandem CC to TT doublesubstitution, has been shown to be induced by reactive oxygen speciesgenerated by a variety of systems and may be diagnostic for such damage.In studies with mammalian UNA polymerases, UNA damaged by reactiveoxygen species yields mutations different from those observed in Esche-

richia coli. This diversity of mutagenic changes in these in vitro studieshighlights the role of UNA replicating enzymes in specifying the types ofmutations produced by reactive oxygen species. In conclusion, we willconsider the role of reactive oxygen species in the pathogenesis of threecommon tumors, carcinoma of the liver, lung, and prostate with consideration on the possible use of antioxidant preventive therapy to slowtumorigenesis sufficiently to prevent clinical presentation of these cancersduring the life span of a patient.

Introduction

Damage to DNA by oxygen-free radicals is frequently postulated to

cause mutations that are associated with the initiation and progressionof human cancers. It has been difficult to prove this relationship,however, due to the large number of reactive oxygen species that havethe potential to damage DNA. These highly reactive chemical molecules include H2O2, O2--, 'O2, HO-, NO2, RO-, and ROO-, and while

each has its own distinct chemistry and cellular distribution, all ofthem have the potential to alter nucleotide residues. Upon reactionwith DNA, oxygen radicals produce more than 30 different adducts(1) and this number excludes protein and lipid addition products aswell as inter- and intra-strand cross-links. Thus, there are potentially

hundreds of different types of chemical changes in DNA resultingfrom oxygen-free radicals that could be mutagenic lesions involved in

the etiology of cancer.Our focus has been to catalogue the types of mutations that arise

secondary to reactive oxygen species damage to DNA and to study themechanisms by which DNA damage leads to mutations. The purposeis to identify mutations that are diagnostic of oxygen-free radical,

damage to DNA, to determine the chemical lesions responsible forthese mutations, and to analyze the association of these mutations withhuman cancer. This review will describe the mutations that result fromreactive oxygen species damage to DNA, some of the mechanisms bywhich they arise, the mutations that result, and the possibility thatreactive oxygen species are involved in the pathogenesis of three

' Presented at the 4lh International Conference on Anticarcinogenesis & RadiationProteelion, April 18-23. 1993. Baltimore. MD. These studies were supported by grantsfrom the NIH, OIG-R35-CA-39903 and AG-01751 lo L. A. L. and F32-CA-08855 toT. M. R. D. I. F. was supported by the Medical Scientist Training Program. GM-07266,and a grant from the Catherine Wilkins Cancer Research Foundation.

common tumors: hepatocellular carcinoma; smoking-associated lung

cancers; and carcinoma of the prostate.

Mutations Induced by Reactive Oxygen Species

In order to determine the spectrum of mutations produced byoxygen-free radicals we have utilized a forward mutation assay.M13mp2 circular single-stranded DNA, containing the lacZa reporter

gene, was exposed to reactive oxygen species followed by either invivo replication in E. coli, or in vitro replication using purified DNApolymerases. Use of a single-stranded template avoids the complex

ities associated with DNA repair and provides a direct correlationbetween the alteration produced at a specific site and the mutation thatresults from copying that altered base during DNA replication. In ourinitial studies, we used reactive oxygen species-generating-systems

containing either Fe, Cu, or Ni ions, stimulated neutrophils, or thephoto-excitable dye méthylèneblue to generate reactive oxygen species (2-5). These in vitro systems allow the study of the mutageniceffects of DNA damage by H2O2, 'O2, O2_-, HOC1, and HO-. Mutants

can be identified phenotypically, by reduced a-complementation ofj3-galactosidase activity in an indicator strain of E. coli. DNA isolatedfrom phage-exhibiting mutant phenotypes is sequenced, thereby es

tablishing the frequency and types of mutations produced at eachnucleotide residue within the target gene. Our results indicate thatdamage to DNA by oxygen-free radicals produces a variety of mutagenic alterations. The damage-induced mutation spectra are dependent

upon both the source of reactive oxygen species and the DNA replicative apparatus that encounters the lesion.

Mutations Produced in E. coli

Replication of DNA damaged by incubation with FeSO4 in thepresence of oxygen induces mainly C to T, G to T, and G to Csubstitutions (2), whereas incubation with CuCl plus H2O2 induces Cto T and G to T substitutions and CC to TT tandem double transitions(4). The mutations that result from exposure to both of these oxygenradical-generating systems are not distributed randomly but rather

clustered in hot spots that are characteristic for each system. Sinceboth metal ions catalyze the production of O2_- by the one electronreduction of O2 and HO- via the Fenton reaction (1, 6), it is not

surprising that the types of mutations produced by both metal ions aresimilar. The differences in the distribution of mutants are probablydue to the liganding properties of each metal ion so as to direct thereactive oxygen species to certain bases and sequence motifs. Cu ionsalso form reactive Cu-oxygen complexes which damage DNA (7) and

may account for some of the mutations specific to that system. Eventhough these systems define a spectrum of mutations produced byoxygen damage to DNA, under precisely controlled conditions, thereare important limitations. The mutations observed are only thosecaused by lesions that escape DNA repair in E. coli, and the types andextent of repair can not be equated with that which occurs in humancells. Furthermore, the observed mutations are dependent on theinduction of the SOS response and on the properties of the E. coliDNA replicative apparatus (2, 4). Thus, while studies using thissystem yield clues about the types of mutations caused by oxygen-free

radicals in human cells, they cannot establish the spectrum of mutations that result from oxygen-free radical generation in human cells.

ISWs

on August 6, 2021. © 1994 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

REACTIVE OXYGEN SPECIES IN TUMORIGENESIS

Tandem CC-»TT Substitutions

Despite their limitations, these in vitro studies have identified anintriguing candidate as a mutation diagnostic for DNA damage byreactive oxygen species. Tandem double CC—>TTmutations were

originally detected in the forward mutation assay in the presence ofboth Cu ions and neutrophil-induced oxidants. More recently, in areversion assay, CC—>TT mutations have been observed to result

from damage induced by Fe, Cu, Ni ions, and gamma irradiation (8).This mutation had previously only been reported in vitro followingexposure to UV light and has been thought to be a specific indicatorof DNA damage by UV irradiation (9). These mutations are notinduced by a variety of other genotoxic chemicals or as a result oferrors by DNA polymcrases in copying of undamaged DNA templates.

A reversion assay that is specific for damage to cytosine residuesand that can detect both single and tandem double mutations at thesame locus has been used to measure the frequency of tandem doubleCC^> TT mutations induced by reactive oxygen species produced bya number of systems (8). We initially compared oxygen radical-

induced mutagenesis with that produced by UV light. In these experiments, reactive oxygen species produced by Fe and Cu ions inducedsingle mutations at a frequency of approximately 1/2000 while tandem double CC *TT mutations were about 30-fold less frequent.

When the hydroxyl radical scavenger mannitol was included in theincubation of DNA and Fe, the frequency of CC-> TT mutations was

decreased. In contrast, mannitol has no effect on the frequency ofCC—»TTmutations produced by UV irradiation (5).2 The diminution

of the Fe-induced mutations by the addition of mannitol provides

evidence that the formation of these mutations is dependent on theproduction of reactive oxygen species. The fact that mannitol did notinhibit the production of tandem double mutations by UV irradiationsuggests that there is probably not a common reactive oxygen speciesinvolved in UV and oxidatively induced double mutations (10). However, the chemical alteration responsible for the CC—»TTsubstitu

tions produced by both UV and reactive oxygen species could besimilar or identical.

Recent studies indicate that reactive oxygen species produced bygamma irradiation and Ni, a human carcinogen, also produce tandemdouble CC »TT mutations. In the case of Ni, the presence of atripeptide, glycyl-glycyl-i.-histidine, known to stimulate the produc

tion of reactive oxygen species by Ni, increases the frequency ofdouble CC—>TTmutations.1 The ability of a tripeptide cofactor to

enhance mutagenesis by Ni suggests that in vim mutations generatedby Ni may be localized at particular sites on DNA by the associationof Ni with specific nucleoproteins. This association may be a majorfactor in determining the sites for nickel-induced mutations in human

cells and dictating the tissue specificity for nickel associated humancancers.

The ability of Fé,Cu, Ni, gamma irradiation, and phorbol ester-stimulated neutrophils to produce CC—»TTmutations suggests that

these mutations are a general manifestation of oxygen damage toDNA. The tandem double CC—>TTmutation is a unique mutation that

provides a marker of oxygen-free radical-induced mutagenesis in cells

that are not exposed to UV irradiation, and may serve as an indicatorfor assessing the involvement of oxidative damage to DNA in agingand tumor progression.

Reactive Oxygen Species-induced Mutations by Mammalian

DNA Polymerases

As a first step to elucidating the mechanisms of mutagenesis byreactive oxygen species in mammalian cells, we studied the fidelity ofmammalian DNA polymerases a and ßon oxidatively damaged DNAtemplates. In these experiments, the duplex DNA with a singlestranded gap spanning the lacla target reporter gene was treated withFeSO4 plus H2O, or CuCI plus H2O2 followed by gap-filling DNAsynthesis using purified enzymes. Mutants were identified phenotyp-

ically following transfections of the copied DNA into E. coli andagain the mutations were clustered in hot spots. Damage catalyzed byFe yielded an increased frequency of C to A and G to A substitutionsby pol-a, and C to A, G to A and C to T substitutions, as well asdeletions of C, by pol-ß.Damage catalyzed by Cu yielded A to C, andC to A substitutions, and deletions of C by pol-a. and A to G, C to A,C to T, G to C and G to T substitutions, and deletions of G by pol-ß(11).4 As observed in the E. coli system, the specificity of the

mutations secondary to DNA damage from the two sources is similaralthough Cu induces additional types of single base substitutions thatmay be caused by Cu-oxygen complexes (7). The unexpected finding

is that when the same template (i.e., damaged by incubation with thesame system for generating reactive oxygen species) is copied by twodifferent DNA polymerases the resulting mutations occur at differentpositions (see Fig. 1). For example, damage catalyzed by Fe inducesa variety of substitutions in place of C if pol-ßis used to copy theDNA whereas only C to A substitutions are produced by pol-a (11).4

Thus, pol-ßexhibits less specificity at poorly instructional, modified

cytosines. In addition, certain lesions, or lesions in certain sequencecontexts, may be mutagenic for one polymerase and not the other. Forexample, in the case of mutations secondary to Cu-mediated damagebase substitutions opposite Gs are observed only with DNA pol-ß

(11). Even in the cases in which the same types of substitutions areintroduced by the two polymerases, the locations arc different. Fig. 1shows the sequence position, in the lacla gene of M13mp2, of thedamage-induced mutational hot spots for polymerases a and ß.There

is little overlap for the two polymerases. When the same template ispretreated with Cu there are only two positions that are sites ofmutation for both polymerases, 64 and 70, whereas for templatestreated with Fe only one site is common to both polymerases, position103. Clearly DNA polymerases have an important role in determiningthe sites and the types of mutations produced as a result of DNAdamage by reactive oxygen species.

Mechanisms of Reactive Oxygen Species-induced Mutations

An analysis of the mutation spectrum suggests that there are at leastfour different pathways by which oxidative DNA damage leads tomutations. The first and simplest mechanism is chemical modificationof nucleotide moieties in DNA causing an alteration in their hydrogenbonding or "coding" specificity. A few oxidative lesions have been

studied for their miscoding potential in E. coli, most notably, thymineglycol and 8-hydroxyguanine, which have been shown to code for T

to C and G to T substitutions, respectively (12, 13). Several otherlesions, including 8-hydroxyadenine, 2-6-diamino-4-5-iV-formami-

dopyrimidine, urea and several thymine fragmentation products (14)have been shown, also in E. coli, to be blocks to replication or poorlymutagenic. Because of the differences between organisms, not tomention DNA polymerases within an organism, it is not possible toconclude from the E. coli data which lesions are responsible for themutations introduced by polymerases a and ß.It is likely, however,that the observed elevations of C to T and C to A substitutions are the

J Unpublished results.1Tkeshclashvili cl ni. unpublished results. 4 Unpublished observations.

1891s

on August 6, 2021. © 1994 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

REACTIVE OXYGEN SPECIES IN TUMORIGENESIS

result of oxidative modifications of cytosines. The identity of thelesions responsible for these mutations is currently under study in ourlaboratory. Pol-ßis also prone to damage-induced errors at guanines,the most prevalent of which are 8-hydroxyguanine and 2-6-diamino-4-5-yV-formamidopyrimidine (FapyG).

A second mechanism, a damage-mediated exacerbation of poly-merase-specific hot spots, appears to be a major contributor to mutagen-esis using DNA pol-ß.In the region of the lacZa gene shown in Fig. 1,

T to G transversions are frequently observed at positions +70 and +103when pol-ßis used to copy the undamaged template. Fe2+-catalyzed

damage induces a 6-10-fold elevation in the frequency of transversions

at each of these sites (11) and Cu ion catalyzed damage causes anelevated frequency of mutations at +103 but not at +70 (11). There aretwo direct explanations for the increase in mutations at this position. Thetransversions observed on the undamaged template could actually becaused by oxidative damage to the DNA template before or duringisolation or during storage. If this is the case, pol-ßmust be exquisitely

sensitive to oxidative damage at these sites because the DNA wasprepared and stored in ways to minimize such damage. This mechanismimplies the background mutations we observed in the undamaged template were the result of endogenous or inadvertent oxygen-mediated

damage. The second explanation is the sequences, which in these twocases are runs of Ts, are inherently mutagenic for pol-ßand can accord

ingly enhance the miscoding potential of an adjacent lesion that resultsfrom oxidative damage.

A third mechanism is a damage-induced conformational change inthe DNA template that prevents accurate replication by DNA poly-merases. This mechanism is suggested by the observation that Fe-

catalyzed DNA damage induces a high frequency of mutants eachcontaining substitutions at both positions +95 and +103 (11). Thefrequency of this double mutation in the pol-ßspectrum is four orders

of magnitude higher than would be expected from two independentmutational events. Oxidative template damage causes an alteration inthe pausing pattern for pol-ßat the two sites of mutation which

implies a change in the structure of the DNA template (11). Exposure

Table 1 Prevention of cancer by delay

TumorHepatomaLung

carcinomaProstaticcarcinomaAge

ofinitiation103540Ageofdeath(yr)505585Age

ofdeathifmutationrateis

reduced2-fold9075130

of the DNA template to reactive oxygen species also tremendouslyincreases the frequency of mutations that can be ascribed to mechanisms which require nascent strand-template rearrangements. Such

rearrangements would be enhanced by conformational changes leading to increased polymerase pausing or decreased processivity. Because of the inherent heterogeneity of the products of oxidative DNAdamage, direct studies of DNA conformation have not been possiblebut many of the known oxidized base lesions are nonplanar and wouldbe likely to alter local DNA strand structure.

Another potential mechanism is the induction of a DNA polymerase conformation that is error prone. This could be the result ofoxygen radical damage to DNA polymerase, which would not beobserved in these studies, or by a change in the conformation of DNApolymerase as a result of encountering an oxygen-induced alteration

in the DNA template. A change in the conformation of DNA polymerase has been invoked by a number of researchers to explain thefrequency of closely spaced double mutation that arise secondary to avariety of genotoxic stresses (15-18). This hypothesis suggests that inorder to perform trans-lesion synthesis a polymerase may assume a

low fidelity conformation that would not immediately revert to thenormal conformation. Mutations likely to arise from this mechanismare closely spaced double substitutions. On template DNA damagedby a Cu-catalyzed reactive oxygen species pol-a generates 10-fold

more mutants sustaining multiple substitutions than it does on undamaged DNA. Furthermore, the average separation between the substitutions on damaged DNA is 14 nucleotides whereas it is 56 nucleo-tides on the undamaged template. The damage-induced substitutions

10,

Fig. I. Location of metal-induced mutation hotspots. The sequence position numbers (horizontalaxis) refer to the IncZa gene of M13mp2 where + Iis the transcription start site. The vertical axis is thenumber of independent mutants observed sustaininga substitution at the indicated sequence position.Top. mutations by pol-a; bottom, mutations bypol-ß.Mutations made on templates damaged byFe- and Cu-bascd reactive oxygen species generating systems are indicated in the figure. Only hotspots (substitutions observed in four or more independent mutants) are shown for simplicity.

Number ofMutationsObserved

6 -

a-Fe

a-Cu

vi ir, tr, v, v, <f, v, ir, ir, 5 sC ^ -i

40,

Number ofMutations

Observed

20

10

l l

B ß-Fe

0 ß-cu

, .i/-, >r,i/-, w, v, v, w-,w, W. Nv, -c r~-x j- O —rii"^. "

-

Sequence Position

»•v, •£r~-oc o* ~

1892s

on August 6, 2021. © 1994 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

REACTIVE OXYGEN SPECIES IN TUMORIOENESIS

arc close enough to one another that they are likely to have occurredduring a single DNA polymerase-template binding event.

Clinical Correlates

Currently, there are no unambiguous data that link mutagenesis byoxygen-free radicals with the initiation of any human cancers or to the

accumulation of mutations during the progression of cancer. A key toassessing oxygen-free radicals as a source of the mutations found in

human cancers would be the detection and quantification of characteristic mutations in DNA obtained from human cancer cells. Oxygen-

free radicals and related species have the potential to damage anynucleotide sequence in DNA and the mutations that result from thisdamage could occur at divergent loci throughout the genome. A verysmall fraction of these mutations might occur in oncogenes that imparta proliferative advantage and result in clonal expansion of cancer cellscontaining the mutation. The majority of oxygen-induced mutations,

however, need not lead to proliferation. Thus, within a given tumorcell, there might be a large number of phenotypically silent, oxygenradical-induced mutations. Since these mutations would be dispersed

throughout the genome, assays with exceptionally high sensitivity willbe needed for their detection. A change in DNA sequence at anyparticular nucleotide could be a very rare occurrence. We will consider three human cancers that have characteristics suggesting theinvolvement of oxygen-free radicals in the malignant process. In each

of these common cancers, a progression of cellular alterations occursover many years suggesting the accumulation of multiple mutations(19) (see Table 1).

Hepatocellular carcinoma is the major cause of cancer mortality inboth Africa and Asia. In these countries, primary hepatoma is associated with either ingestion of aflatoxin (20) or with chronic infectionby hepatitis viruses B (21) or C (21). In areas with a high level ofexposure to aflatoxin, mutations involving codon 249 of the p53 genehave been extensively documented in primary hepatocellular carcinomas (22). In areas lacking aflatoxin, a spectrum of different nucleotidesubstitutions has been detected in p53 in primary hepatomas (23).Since p53 mutations are present in most or all of the malignant cellswithin these tumors, it is presumed that they confer or are associatedwith changes that confer a proliferative advantage. G to T transversions, which have been shown to be one of the more common typesof mutations produced by aflatoxin lesions and oxygen-free radical

damage to DNA (2, 4), are the predominant substitution seen at codon249 in the aflatoxin associated tumors. The same type of transversion,however, has been shown to arise secondary to replication of DNAcontaining abasic sites which result from a variety of DNA damagingagents including oxygen-free radicals (24). Moreover, trans-lesion

bypass of bulky adducts during DNA replication is also likely to resultin the insertion of deoxyadenosine, yielding G to T transversions (24).Thus, the presence of G to T transversions at position 249 in p53 inprimary hepatomas in areas with aflatoxin is not diagnostic for aspecific mutagenic mechanism.

The distribution of p53 mutations in hepatomas found in regionswith a high prevalence of chronic hepatitis, but lacking aflatoxinexposure, are not localized to codon 249 (23) and are not necessarilydue to the production of oxygen-reactive species during chronicinfection. However, a number of factors make hepatitis-associatedprimary hepatomas potential candidates for an oxygen-free radical-

induced human cancer. These include a long incubation time, a lack ofa demonstrable oncogene or the integration of a viral genome, and anassociation with persistent chronic liver infection. The findings thathepatocellular carcinomas result from the overproduction of hepatitisB virus large envelope polypeptide in transgenic mice is also compatible with a nongenetic mechanism for carcinogenesis that couldinvolve the production of oxygen-free radicals (25).

In Asia and Africa the onset of chronic hepatitis occurs in earlychildhood. Beasley and others (26, 27) have demonstrated that thetime between initial infection with hepatitis B virus as measured bythe presence of the Australia antigen and the detection of hepatocellular carcinomas is frequently as long as 40 years. During thisinterval there is a persistent viral infection associated with the presence of inflammatory cells that generate reactive oxygen species.Thus, measurements of DNA damage and mutations in human liver asfunction of persistence of chronic hepatitis might be predictive of theonset of liver cancer. It would be important to determine the types ofmutations that are found in DNA from hepatomas since a knowledgeof these might point to the agent associated with these mutations. Ifreactive oxygen species are etiological agents, treatment with agentsthat reduce the inflammatory response or scavenge oxygen-free rad

icals may delay the onset of hepatoma sufficiently to prevent itsclinical presentation (Table 1).

Cancer of the lung is the most frequent cancer-related cause of

death in both males and females in the United States and possiblythroughout the world. The fact that 90% of lung cancer is the result ofsmoking has made it possible to document the time of exposure to theagent that causes the cancer. For most individuals, there is at least a30-year interval between the commencement of smoking and the

onset of a clinically demonstrable tumor (28. 29). The requirement forcontinuous exposure to agents in smoke is further reinforced by thefindings that cessation of smoking is correlated with a rapid diminution in the risk of lung cancer (30). Presumably there ¡sa series ofsmoking-associated events required for the production of lung can

cers. These events could involve the production of mutations bycomponents of cigarette smoke some of which might be mediated byDNA damage via oxygen-free radicals. In fact, cigarette smoke hasbeen estimated to contain some IO17oxidant molecules per puff (31)

and the exposure of animal cells in culture to cigarette smoke causesthe accumulation of 8-hydroxy-deoxyguanosine (32). Urine obtainedfrom smokers also has a 4- to 10-fold elevation in altered nucleotidesthat are known to be produced by oxygen-free radicals (33). Furthermore, 8-OH-dG is highly elevated, in the sperm of smokers, and can

be reduced by a diet consisting of a high intake of green vegetableswith oxygen radical scavengers such as ascorbic acid and ß-carotene

(33). For any individual it seems obvious that the most importantpreventive medicine is smoking cessation, but in the absence of this,an increased intake of food with potent reducing agents might delaythe progression of lung cancer.

Prostate cancer is characteristically diagnosed in old age or atautopsy and it constitutes the third most frequent cancer-related cause

of death in males in the United States. Chronic prostate hypertrophycommences in most males by the age of 40 years but the latepresentation of prostatic carcinoma suggests that a multistep processis involved in tumorigenesis. Some hypertrophie foci could containcells that progress to overt cancers. The situation could be analogousto colon cancer in which malignant foci arise in premalignant, hyper-

trophic polyps. Since prostate cancer is most prevalent in the seventhand eighth decade of life, it can be inferred that the events causing thisdisease accumulate over a 40-year span. The paucity of known chem

ical agents associated with prostate cancer and the absence of anyapparent environmental or occupational risk factors suggests thatprostate cancer is associated with endogenous cellular processes. Themost reasonable candidates for endogenously formed gcnotoxins thataccumulate in later life are the reactive oxygen species. If one couldidentify the DNA lesions associated with this cancer and determinethe mechanism for their formation, then it might be feasible tosignificantly reduce the deaths caused by prostate cancer. Agents thatreduce the formation of the rate-limiting event in the progression of

prostate cancer by a factor of 2 could significantly diminish the

1893s

on August 6, 2021. © 1994 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

REACTIVE OXYGEN SPECIES IN TUMORIOENESIS

incidence of prostate cancer. Reduction in the rate of mutagenesismight delay the onset of prostate cancer beyond the life span of mostmale patients (Table 1). If one considers the onset of the diseaseoccurs with the start of prostatic hypertrophy at age 40, and theaverage age of death from prostate cancer is 85, then a 2-fold reduc

tion in the progression of the disease would delay death from prostatecancer until age 130.

Primary hepatoma, lung cancer, and prostate cancer are examplesof common cancers that are associated with malignant progressionover many years. In each of these cancers there is evidence for theaccumulation of chromosomal rearrangements that encompass hundreds of thousands or even millions of nucleotides. Conceivably, largerearrangements are associated with a high frequency of DNA sequence changes in cancer cells. There are several normal cellularprocesses that occur at a sufficiently high frequency to generate thesemutations. Among these processes are those that produce errors inDNA replication, enhance the instability of the DNA molecule resulting in deamination or depurination, alter the chemical structure ofnucleotides by alkylation, or damage DNA bases by the production ofoxygen-free radicals (19). The focus on oxygen-free radicals as a

source of mutations during tumor progression is partly a reflection ofour ability to quantitate the characteristic mutations as well as thefeasibility of significantly reducing damage by these free radicalsthrough drugs or dietary manipulations.

References

1. 11.illi«i II. B.. and Aruoma. O. DNA damage by oxygen-derived species. FEBS Let..281: 9-19. 1991.

2. Mcliride, T. J.. Presión, B. D., and Loeb, L. A. Mutagenic spectrum resulting fromDNA damage by oxygen radicals. Biochemistry, 30: 207-213. 1991.

3. McBride, T. !.. Schnieder. J. E.. Floyd, R. A., and Loeb, L. A. Mutations induced byméthylèneblue plus light in single-stranded M13mp2. Proc. Nati. Acad. Sci. USA.89: 6866-6870. 1992.

4. Tkeshelashvili, L. K., McBride. T. J.. Spence. K.. and Loeb. L. A. Mutation spectrumof copper-induced DNA damage. J. Biol. Chem., 266: 6401-6406, 1991.

5. Reid. T. M.. and Loeb, L. A. Mutagenic specificity of oxygen radicals produced byhuman leukemia cells. Cancer Res.. S3: 1082-1086, 1992.

6. Snow. E. T. Metal carcinogenesis: mechanistic implications. Pharmacol. Ther.. 53:31-65, 1992.

7. Yamamoto. K., and Kawanishi. S. Hydroxyl free radical is not the main active speciesin site-specific DNA damage induced by copper(II) and hydrogen peroxide. J. Biol.Chem.. 264: 15435-15440, 1989.

8. Reid. T. M.. and Loeb. L. A. Tandem double CCTT mutations are produced byreactive oxygen species. Proc. Nati. Acad. Sci. USA. 90: 3905-3907, 1993a.

9. Brash. D. E.. Rudolph, J. A.. Simon. J. A., Lin, A., McKenna, G. J., Halpem, A. J.,and Ponten, J. A role for sunlight in skin cancer: UV-induced p53 mutations insquamous cell carcinoma. Proc. Nati. Acad. Sci. USA, 88: 10124-10128, 1991.

10. Reid. T. M.. and Loeb. L. A. The effect of DNA repair enzymes on mutagenesis byoxygen radicals. Mutât.Res.. 289: 181-186, 1993.

11. Feig, D. !.. and Loeb, L. A. Mechanisms of mutation by oxidative DNA damage: reducedfidelity of mammalian DNA polymerase-ß.Biochemistry. 32: 4466-4473. 1993.

12. Basu, A. K., Loechler, E. L., Leadon. S. A., and Essigmann, J. M. Genetic effects ofthymine glycol: site-specific mutagenesis and molecular modeling studies. Proc. Nail.Acad. Sci. USA, 86: 7677-7681, 1989.

13. Cheng, K. C, Cahill, D. S., Kasai, H.. Nishimura, S.. and Loeb, L. A. 8-hydroxygua-nine, an abundant form of oxidative DNA damage, causes G to T and A to Csubstitutions. J. Biol. Chem.. 267.- 166-172, 1992.

14. Evans. J.. Maccabee, M., Hatahet, Z., Courcele, J.. Bockrath, R., Ide. H., and Wallace,S. Thymine ring saturation and fragmentation products: lesion bypass, misinsertionand implication for mutagenesis. Mutag. Res., 299: 147-156. 1993.

15. Hauser. J.. Seidman. M. M., Sidur, K.. and Dixon, K. Sequence specificity of pointmutations induced during passage of a UV-irradiated shuttle vector plasmid inmonkey cells. Mol. Cell. Biol.. 6: 277-285, 1986.

16. Hauser, J., Levine, A. S.. and Dixon, K. Unique pattern of point mutations arisingafter gene transfer into mammalian cells. EMBO J.. 6: 63-67, 1987.

17. Seidman. M. M.. Bredberg, A., Seetharam. S.. and Kraemer, K. H. Multiple pointmutations in a shuttle vector propagated in human cells: evidence for an error-proneDNA polymerase activity. Proc. Nati. Acad. Sci. USA. 84: 4944-4948, 1987.

18. Madzak, C., and Sarasin. A. Mutation spectrum following transfection of ultraviolet-irradiated single-stranded or double-stranded shuttle vector DNA into monkey cells.J. Mol. Biol., 218: 667-673, 1991.

19. Loeb, L. A. Mutator phenotype may be required for multistage carcinogenesis. CancerRes., 51: 3075-3079. 1991.

20. Ross, R. K., Yuan, J-M., Yu, M. C.. Wogan, G. N., Qian, G-S., Tu, J-T., Groopman,J. D., Gao, Y-T., and Henderson, B. E. Urinary aflatoxin biomarkers and risk ofhepatocellular carcinoma. Lancet, 339: 943-946, 1992.

21. Blumberg, B. S.. Larouze, B., London, W. T., Werner, B., Hesser, J. E., Millman, I.,Saimot. G.. and Payet, M. The relation of infection with the hepatitis B agent toprimary hepatic carcinoma. Am. J. Pathol., 81: 669, 1975.

22. Hsu, 1. C.. Metcalf, R. A.. Sun, T.. Welsh, J. A.. Wang, N. J., and Harris. C. C.Mutational hot spot in the p53 gene in human hepatocellular carcinomas. Nature(Lond.), 350: 429-431. 1991.

23. Nishida. N., Fukuda, Y., Kokuryu, H.. Toguchida. J.. Yandell. D. W., Ikenega, M.,Imura. H., and Ishizaki. K. Role and mutational heterogeneity of the p53 gene inhepatocellular carcinoma. Cancer Res., 53: 368-372, 1993.

24. Loeb. L. A., and Preston, B. D. Mutagenesis by apurinic/apyrimidinic sites. Annu.Rev. Genet.. 20: 210-230, 1986.

25. Chirsari. F. V.. Klopchin, K.. Moriyama. T.. Pasquinelli, C., Ducnsford, H. A., Sell,S.. Pinkert, C. A., Brinster. R. L.. and Palmiter. R. D. Molecular pathogenesis ofhepatocellular carcinoma in hepatitis B virus transgenic mice. Cell, 59: 1145-1156.1989.

26. Beasley. R. P. Hepatitis B virus. The major etiology of hepatocellular carcinoma.Cancer (Phila.), 61: 1942-1956, 1988.

27. Stein, W. D. Analysis of cancer incidence data on the basis of multistage and clonalgrowth models. Adv. Cancer Res.. 56: 161-212, 1991.

28. Boring, C. C., Squires, T. S., and Tong, T. Cancer Statistics, 1993. Cancer (Phila.).43: 7-26. 1993.

29. Pierce, J. P.. Thurmond. L.. and Rosbrook, B. Projecting international lung cancermortality rates: first approximations with tobacco-consumption data. Monogr. J. Nail.Cancer Inst.. 12: 45-49, 1992.

30. Loeb, L. A.. Ernster. V. L., Warner. K. E.. Abbotts, J., and Laszio, J. Smoking andlung cancer: an overview. Cancer Res.. 44: 5940-5958, 1984.

31. Pryor. W. A.. Prier. D. G., and Church, D. F. An electron spin resonance study ofmainstream and side stream cigarette smoke: the nature of the free radicals ingas-phase smoke and in cigarette tar. Environ. Health Perpect.. 47: 345-355, 1983.

32. Fraga, C. G., Shigenaga, M. K., Park, J. W.. Degan. P.. and Ames, B. N. Oxidativedamage to DNA during aging: 8-hydroxy-2'-deoxyguanine in rat organ DNA and

urine. Proc. Nati. Acad. Sci. USA, 87: 4533-4547, 1990.33. Fraga, C. G., Molchnik, P. A.. Shigenaga, M. K., Helbock. H. J., Jacob, R. A., and

Ames, B. N. Ascorbic acid protects against endogenous DNA damage in humansperm. Proc. Nati. Acad. Sei. USA, ÄS:11(X)3-110Ü6,1991.

1894s

on August 6, 2021. © 1994 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1994;54:1890s-1894s. Cancer Res   Daniel I. Feig, Thomas M. Reid and Lawrence A. Loeb  Reactive Oxygen Species in Tumorigenesis

  Updated version

  http://cancerres.aacrjournals.org/content/54/7_Supplement/1890s

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/54/7_Supplement/1890sTo request permission to re-use all or part of this article, use this link

on August 6, 2021. © 1994 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from