-
Arch Toxicol (2008) 82:493512DOI 10.1007/s00204-008-0313-yREVIEW
ARTICLE
Carcinogenic metal compounds: recent insight into molecular and
cellular mechanisms
Detmar Beyersmann Andrea Hartwig
Received: 15 April 2008 / Accepted: 30 April 2008 / Published
online: 22 May 2008 Springer-Verlag 2008
Abstract Mechanisms of carcinogenicity are discussed formetals
and their compounds, classiWed as carcinogenic tohumans or
considered to be carcinogenic to humans: arsenic,antimony,
beryllium, cadmium, chromium, cobalt, lead,nickel and vanadium.
Physicochemical properties governuptake, intracellular distribution
and binding of metal com-pounds. Interactions with proteins (e.g.,
with zinc Wngerstructures) appear to be more relevant for metal
carcinoge-nicity than binding to DNA. In general, metal
genotoxicity iscaused by indirect mechanisms. In spite of diverse
physico-chemical properties of metal compounds, three
predominantmechanisms emerge: (1) interference with cellular redox
reg-ulation and induction of oxidative stress, which may
causeoxidative DNA damage or trigger signaling cascades leadingto
stimulation of cell growth; (2) inhibition of major DNArepair
systems resulting in genomic instability and accumula-tion of
critical mutations; (3) deregulation of cell prolifera-tion by
induction of signaling pathways or inactivation ofgrowth controls
such as tumor suppressor genes. In addition,speciWc metal compounds
exhibit unique mechanisms suchas interruption of cellcell adhesion
by cadmium, directDNA binding of trivalent chromium, and
interaction of vana-date with phosphate binding sites of protein
phosphatases.
Keywords Carcinogenic metals Mechanisms Genotoxicity Oxidative
stress DNA repair Deregulation of cell proliferation
Introduction
A decisive step in the development of human technology
andculture was the discovery of metals below the surface of
ourplanet, their excavation, extraction and use as tools to
fulWllhuman needs. Nowadays, we recognize that the wastes ofmetals
are distributed over the soils and waters of the earthssurface and
exert detrimental eVects on life in the environmentand human
health. Unlike organic waste, metals and theircompounds are not
degraded by living organisms and mayaccumulate up to harmful
levels. Metals are small entitieswhen compared to organic materials
and their reactions withliving matter are seemingly simple to
evaluate. However, thepicture emerging today shows a very complex
pattern of metalinteractions with cellular macromolecules,
metabolic and sig-nal transduction pathways and genetic processes.
Some metalcompounds even undergo metabolic transformation, such
asreduction to lower oxidation state or alkylation. Hence,
thetoxicological assessment of metal eVects is by no means sim-ple,
which is especially true for mechanisms of metal carcino-genicity.
Even for single metal species, the hitherto revealedmechanisms
involved in their carcinogenic action are multi-ple. A special
feature of metal biology is the fact that evenmetals that are
essential for the sustainment of life (such asiron and copper) may
become toxic depending on the oxida-tion state, complex form, dose
and mode of exposure. In thisreview, Wrst we depict the possible
common mechanisms ofmetal carcinogenicity without neglecting the
speciWc diVer-ences and then discuss the peculiarities of speciWc
metals.
Overview over classiWcation as carcinogens
Carcinogens are classiWed by both scientiWc committeesand
regulatory agencies. The following assignments refer to
D. Beyersmann (&)Biochemistry, Department of Biology and
Chemistry, University of Bremen, 28334 Bremen, Germanye-mail:
[email protected]
A. HartwigInstitute of Food Technology and Food Chemistry,
Technical University of Berlin, 13355 Berlin, Germanye-mail:
[email protected]
-
494 Arch Toxicol (2008) 82:493512evaluations of scientiWc
commissions only, since legal clas-siWcations depend on national
policies. Table 1 summarizesthe classiWcation of some metals and
metalloids by theInternational Agency for Research on Cancer (IARC)
andthe German Commission for the Investigation of HealthHazards of
Chemical Compounds in the Work Area (MAKcommission). The IARC
classiWes the carcinogens accord-ing to carcinogenic hazard as
carcinogenic to humans(Group 1), probably carcinogenic to humans
(Group 2A),possibly carcinogenic to humans (Group 2B) or not
classiW-able as to its carcinogenicity to humans. The
GermanMAK-commission also has classiWed several metals eitheras
carcinogenic to humans (category 1), as considered to
becarcinogenic to man based on long-term animal studies(category 2)
or as giving concern to cause cancer, but evi-dence is not suYcient
for classiWcation in group 1 or 2 (cat-egory 3B) (DFG 2007a).
Physicochemical aspects
Metals may be carcinogenic in the form of free ions,
metalcomplexes, or particles of metals and poorly soluble
com-pounds. The toxicity of metals and their compounds is gov-erned
by their physicochemical properties. Regarding metal
ions, oxidation state, charge and ionic radii are crucial.
Withmetal complexes, the coordination number, the geometry andthe
type of ligands (e.g., their lipophilicity) are important fortoxic
interactions. Regarding metals and their poorly solublecompounds,
particle size and crystal structure are important.Not only toxic
metal cations, but also essential transitionmetal ions bind to
biological ligands of opposite charge, suchas acid amino acid side
chains of proteins and phosphategroups of nucleotides and nucleic
acids, and form complexeswith oxygen, sulfur and nitrogen groups of
proteins, nucleicacids and other biomolecules. If toxic metal ions
have similarphysiochemical properties such as charge and size as
those ofessential ions, they may compete for the biological
bindingsites of the latter and cause structural perturbations
resultingin aberrant function of biochemical macromolecules
(Beyers-mann 1995; Nieboer et al. 1999; Hartwig 2001). Some
exam-ples are discussed in more detail here.
Be2+ carries the same charge as Mg2+ and competes forMg2+ in
biochemical binding sites. However, its ionic radiusin
hexacoordination (0.45 ) is much smaller than that ofMg2+ (0.72 ),
hence it polarizes ligands more strongly, andits bonds with
proteins are more stable than those of Mg2+.
Cd2+ ions have ionic radii very similar to those of Ca2+(in
hexacoordination 0.95 and 1.00 , respectively, in octa-coordination
1.10 and 1.12 , respectively). Although the
Table 1 ClassiWcation of met-als and/or their compounds as
carcinogenic)
Substances IARC carcinogen group
MAK carcinogen category
Antimony and its compounds 2 (except SbH3)Antimony trioxide
(Sb2O3) 2B 2Antimony trisulWde (Sb2S3) 3 2Arsenic and its compounds
1 1Beryllium and its compounds 1 1Cadmium and its compounds 1
1Chromium metal 3 Chromium(VI) compounds 1 2 (except ZnCrO4: cat.
1)Chromium(III) compounds 3 Cobalt and its compounds 2B 2Cobalt
with tungsten carbide
(hard metal)2A 1
Gallium arsenide 1 Indium phosphide 2A 2Lead metal 2Lead
compounds 2A 2Mercury and its compounds 2B 3BNickel metal 2B
1Nickel compounds 1 1Rhodium 3BSelenium and its compounds 3
3BVanadium and its compounds 2Vanadium pentoxide (V2O5) 2B 2
Origin: International Agency for Research on Cancer (IARC)
Monographs, MAK (German Commission for the Investiga-tion of Health
Hazards of Chem-ical Compounds in the Work Area). (, not
classiWed)123
-
Arch Toxicol (2008) 82:493512 495preferred ligand of Ca2+ is
oxygen, whereas Cd2+ preferssulfur, Cd2+ also accepts oxygen and is
able to substituteCa2+ in protein binding sites. Cd2+ interferes
with the func-tions of numerous Ca2+-transport and Ca2+-dependent
sig-naling proteins. In some cases substitution of Ca2+ by Cd2+may
even yield the normal protein function such as in theprotein
calmodulin where substitution of Ca2+ by Cd2+ stillconserves 90%
activity with cyclic phosphodiesterase(Cheung 1984). Cd2+ ions have
an analogous electron con-Wguration with Zn2+ (4d10 vs. 3d10).
Despite its larger radius(0.95 vs. 0.74 ), Cd2+ can often
substitute for Zn2+ in zincenzymes and transcription factors and
disturb or abolish thebiochemical functions of such proteins.
Pb2+ has ionic radii of 1.19 in hexacoordination and1.29 in
octacoordination. These are suYciently close tothat of Ca2+, and
Pb2+ interferes with many types of Ca2+-regulated physiological
processes, especially the Ca2+-sig-naling system.
Ni2+ ions have nearly the same radius as Mg2+ ions (0.69and 0.66
, respectively) and similar ligand preferences, thatis for oxygen
and nitrogen. Ni2+ ions can interfere with Mg2+functions in enzymes
of nucleic acid synthesis and repair.
Co2+ ions have the same charge and size as Zn2+ ions(both 0.74
A). Both Co2+ and Zn2+ ions bind to the sametypes of ligands, that
is, oxygen, nitrogen and sulfur. In sev-eral instances, the
substitution of Zn2+ by Co2+ in zincenzymes results in proteins
with modiWed catalytic activi-ties. A prominent example is the
eVect of the carcinogenicmetal ions Cd2+, Co2+ and Ni2+ on
structure and function ofzinc Wnger domains of several
transcription factors andenzymes (Hartwig 2001; Kopera et al.
2004).
The toxicity of metals and their compounds largelydepends on
their bioavailability, i.e. the mechanisms ofuptake through cell
membranes, intracellular distributionand binding to cellular
macromolecules (Fig. 1). Theanionic compounds of chromium and
vanadium smoothlypenetrate into cells via the general anion channel
of theplasma membrane. Hence, anionic chromium(VI) is readilytaken
up into cells and, however, as soon as it is reduced to
chromium(III) by intracellular reductants, the metal istrapped
and accumulates (Connett and Wetterhahn 1983).Several toxic metals
occuring as divalent cations can passthrough plasma membranes via
cation transporters. In con-trast, cellular plasma membranes are
impermeable to thetrivalent Cr3+ ion. At variance, sparingly
soluble metalcompounds may be taken up by phagocytosis, which
maylead to considerable intracellular accumulation of thesemetals
after gradual dissolution in the lysosomes.
General mechanisms of metal genotoxicity and carcinogenicity
The interactions of diverse metal carcinogens with livingmatter
are complex, and at Wrst sight it seems daring toassume that there
were common mechanisms of action.However, a closer look reveals
three major mechanismswhich apply for the majority of carcinogenic
metal com-pounds besides some distinct reactions of speciWc
metalcompounds, namely oxidative stress, DNA repair modula-tion and
disturbances of signal transduction pathways.
Induction of oxidative stress
The induction of oxidative stress is an attractive hypothesisto
explain mutagenic and carcinogenic eVects of metals. Ionsof the
carcinogenic metals, such asantimony, arsenic, chro-mium, cobalt,
nickel and vanadium, are capable of perform-ing redox reactions in
biological systems. They have beenshown to induce the formation of
reactive oxygen and nitro-gen species in vivo and in vitro in
mammalian cells. Fre-quently the formation of hydroxyl radicals,
most probablyby Fenton- and HaberWeiss-type reactions, has
beendetected. These radicals are known to cause oxidative dam-age
to lipids, proteins and DNA (Fig. 2). Although the ionsof the
carcinogenic metal cadmium are not capable of exert-ing redox
reactions in biological systems, they have been
Fig. 1 Cellular uptake, intracellular distribution and binding
of solu-ble and particulate metal compounds
Men+
Plasma Membrane
Nucleus
Lysosome pH 4.5
Phagocytosis
ParticulateMetal Compound
SolubleMetal Compound
Men+
Men+Ion Channel
Protein
Men+
DNA
Men+
Fig. 2 Metal ions and oxidative stress (modiWed from Hartwig
2007)
OxidativePhosphorylation
Activated Phagocytes Environmental factors
O21
Super Oxide Dismutase
Cr(V), Fe(II), Co(II), Ni(II), Cu(II)
OH
Catalase
Peroxidase
H2O + O2
H2O
Lipid Peroxidation Oxidative Protein Damage DNA Damage
H2O2123
-
496 Arch Toxicol (2008) 82:493512found to generate oxidative
stress too. The reason for thisproperty of cadmium seems to be the
inhibition of antioxida-tive enzymes in vitro and in vivo. Cadmium
has been shownto inhibit catalase, superoxide dismutase,
glutathione reduc-tase, and glutathione peroxidase (see below). In
addition tothe metals classiWed as carcinogens, iron and copper are
alsoeVective catalysts for Fenton and Fenton-type
reactions.Nevertheless, in living systems iron and copper are
tightlyregulated with respect to uptake, transport, storage,
mobili-zation, transfer to target molecules and excretion,
ensuringthat increased deliberation of free ions is restricted to
condi-tions of extreme overload, genetic defects in metal
homeo-stasis and/or metabolic stress. Besides generating DNAdamage
directly, reactive oxygen species at low concentra-tions function
as mitogenic signals and activate redox-sensi-tive transcription
factors (Genestra 2007). Hence, oxidativestress may not only
initiate tumor development by mutagen-esis but also deregulate cell
growth and promote tumorgrowth depending on extent and time of
interference.
One major objection against the oxidative-stress-hypoth-esis of
metal carcinogenicity is the discrepancy between thecomparatively
high, often cytotoxic doses of metal com-pounds that are required
to evoke the formation of reactiveoxygen species and/or measurable
increase in damage tocellular macromolecules and the often very low
doses ofmetals that induce tumors. Hence, it seems that
metal-induced oxidative stress is not the sole cause for metal
car-cinogenesis but still contributes to the development
ofmalignancy in a potentiating manner.
Interference with DNA repair
With the exception of chromium(VI), carcinogenic metalsare only
weak mutagens in mammalian cells and often inac-
tive in bacterial assays. Since mutagenicity in bacterialassays
is an indicator of reactivity of a substance withDNA, metals are
thought to exert genotoxicity by indirectmechanisms. Carcinogenic
metal compounds often arecomutagenic, that is, they enhance the
mutagenicity ofother genotoxic agents. Indeed, many carcinogenic
metalcompounds at low concentrations have been identiWed
asinhibitors of the repair of DNA damage that is caused eitherby
other xenobiotics or by endogeneous factors (Hartwig2007). The four
main, partly overlapping DNA repair path-ways operating in
mammalian cells are base excision, mis-match, nucleotide excision
and recombinational repair.Figure 3 gives an overview over sources
of DNA damageand the four major repair pathways involved in the
removalof the respective DNA lesions. Inherited or acquired
deW-ciencies in such pathways can contribute to the onset
ofmalignant growth. DNA repair mechanisms are frequenttargets for
interference by toxic metals as discussed indetail for the
individual metals below. Inhibition of repairand persistent DNA
damage results in genomic instability,which may become especially
deleterious under conditionsof accelerated cell proliferation
and/or impaired apoptosis.
Deregulation of cell proliferation
Tumor development is characterized by a deregulation ofcell
growth and diVerentiation. Carcinogenic metal com-pounds may alter
cell growth by several distinct mecha-nisms, either aVecting the
expression of growth stimulatingfactors or inactivating growth
control mechanisms. Withrespect to the former, some metal ions are
found to activatemitogenic signaling pathways and induce the
expression ofcellular proto-oncogenes. Furthermore, epigenetic
mecha-nisms, such as hypo- or hyper-methylation of DNA or dis-
Fig. 3 Sources of DNA damage and major repair pathways Ionizing
Radiation
Reactive OxygenSpecies,
Alkylating Agents
UV Irradiation,PolycyclicAromatic
Hydrocarbons
ReplicationErrors
Ionizing RadiationAntitumor Agents,
e.g. Cisplatin,Mitomycin C
Abasic Sites,Base Modification,
Single Strand Breaks
(6-4)-Photoproducts,
Pyridine-Dimers,Bulky Adducts
A-G Mismatch,T-C Mismatch
InterstrandCrosslinks,
Double Strand Breaks
Base ExcisionRepair
NucleotideExcision Repair
Base MismatchRepair
RecombinationalRepair123
-
Arch Toxicol (2008) 82:493512 497turbed histone acetylation, may
contribute to modiWedpatterns of gene expression. Changes in gene
regulation areobserved prior to manifestation of tumors. Initially,
they arenot Wxed by mutation, and the agent must be present for
anextended time period to cause persistent modiWcations,which can
be genetically Wxed during tumor development.Concerning the
interference with cellular growth control,some metal carcinogens
have been shown to inactivate thetumor suppressor proteins p53
and/or downregulate theexpression of tumor suppressor genes Fhit,
p16, p53 and ofsenescence genes. Finally, metal ions may deregulate
cellproliferation by inactivating apoptotic processes resultingin
adaptation to the cytotoxicity of the metal.
Mechanisms of action of speciWc metals
Arsenic
Arsenic is a well-documented human carcinogen. Numer-ous
epidemiological studies have shown that arsenic cancause diVerent
types of cancer via exposure to contami-nated drinking water and/or
ambient air (reviewed by Yos-hida et al. 2004). In humans and many
other mammals,inorganic arsenic prevalent in drinking water as
arsenite orarsenate is metabolised in the liver through successive
oxi-dative methylation and reduction steps to its trivalent
andpentavalent mono- and di-methylated metabolites (Fig. 4;reviewed
by Aposhian and Aposhian 2006).
While previously methylation has been considered as
adetoxiWcation process, recent studies have shown that incontrast
to the pentavalent methylated species, the trivalentmethylated
metabolites monomethylarsonous [MMA(III)]and dimethylarsinous
[DMA(III)] acid in various test sys-tems are at least as or even
more genotoxic when comparedto inorganic arsenic (Styblo et al.
2000; Schwerdtle et al.2003a; Kligerman et al. 2003). Therefore,
they may con-tribute to inorganic arsenic-induced carcinogenicity.
Sev-eral modes of action have been proposed to be involved in
arsenic carcinogenicity, including the induction of oxida-tive
stress, diminished DNA repair, altered DNA methyla-tion patterns,
enhanced cell proliferation and suppression ofp53 (reviewed by
Schoen et al. 2004). However, it is notentirely clear which
mechanisms prevail in the carcinoge-nicity of arsenic.
Genotoxic eVects
Arsenite does not induce point mutations in bacterial
ormammalian test systems. However, it increases the mutage-nicity
of other DNA damaging agents, such as UVC radia-tion (Rossman et
al.1986), which may be explained byinterference with DNA repair
processes (see below). Incontrast, the induction of micronuclei,
chromosomal aber-rations, DNA strand breaks and oxidative DNA base
dam-age is well documented and has been observed atcomparatively
low concentrations in cultured mammaliancell lines such as V79,
CHO, A549, and in human periphe-ral lymphocytes (Schoen et al.
2004). With respect to theinorganic species, arsenate (with
pentavalent As) and arse-nite (with trivalent As), similar
genotoxic eVects have beenobserved, albeit at about tenfold higher
concentrations ofarsenate as compared to arsenite. Regarding the
methylatedspecies, MMA(III) and DMA(III) are genotoxic at
lowerconcentrations than arsenite at all endpoints, while
geno-toxic eVects of MMA(V) and DMA(V) are either absent
orrestricted to much higher concentrations (Styblo et al.2000;
Schwerdtle et al. 2003a; Kligerman et al. 2003).Micronuclei and
chromosomal aberrations have also beenobserved in mice after oral
exposure to comparatively lowconcentrations of arsenite.
Furthermore, chromosomalaberrations and micronuclei were found in
peripheral lym-phocytes, buccal and bladder cells of humans
exposedtowards elevated concentrations of arsenite via
drinkingwater (Schoen et al. 2004). Underlying mechanisms may bethe
induction of oxidative stress, inhibition of DNA repairsystems and
changes in DNA methylation patterns, all ofwhich may lead to
genomic instability.
Induction of oxidative stress
Several lines of evidence suggest the involvement of oxida-tive
stress mediated by increased levels of reactive oxygenspecies in
arsenic-induced genotoxicity (reviewed by Shiet al. 2004). Thus,
arsenite increases the generation ofsuperoxide anions (O2) and
hydrogen peroxide (H2O2) indiverse cellular systems, while the
modulation of nitricoxid (NO) production appears to be restricted
to higherconcentrations. With respect to genotoxicity, the
applica-tion of radical scavengers revealed the involvement of
arse-nite-induced ROS in the induction of lipid peroxidation aswell
as in DNA damage. Furthermore, the ROS can activate
Fig. 4 Proposed mammalian arsenic metabolism. Note that the
exactreaction sequence and enzymes involved are still debated (For
detailssee Aposhian and Aposhian 2006)
Arsenate (V) Arsenite (III)Monomethyl-arsonic Acid[ MMA (V)
]
Reductase
Methyltransferase
Reductase
Monomethyl-arsonous Acid[ MMA (III) ]
Dimethyl-arsinic Acid[ DMA (V) ]
Dimethyl-arsinous Acid[ DMA (III) ] Reductase
Methyltransferase123
-
498 Arch Toxicol (2008) 82:493512signal cascades involving
mitogen-activated proteinkinases (MAPKs) and transcription factors
AP-1 and NFB(Leonard et al. 2004). Several origins of elevated
levels ofROS are possible and have been suggested. They
includeinteractions with the respiratory chain, the generation
ofROS during oxidation of trivalent to pentavalent species
asevident in liver metabolism as well as the release of ironfrom
ferritin by trivalent arsenic species. Furthermore,interactions of
arsenic with cellular protection mechanismsagainst ROS, in
particular a decrease in glutathione levelsand the disturbance of
DNA repair systems, contribute toincreased levels of oxidative
damage in cells (reviewed byShi et al. 2004).
DNA repair inhibition
Arsenite as well as MMA(III) and DMA(III) have beenshown to
inhibit the repair of UVC- and benzo[a]pyrene-diolepoxide
(BPDE)-induced DNA damage in the lowmicromolar, non-cytotoxic
concentration range (see forexample, Schwerdtle et al. 2003b). As
one explanation, adecrease of poly(ADP-ribosyl)ation in HeLa S3
cells afterincubation with nanomolar concentrations of
arsenite,MMA(III) or DMA(III) was observed (Hartwig et al.
2003;Walter et al. 2007), a reaction involved in DNA
damagesignaling and the recruitment of DNA repair proteins to
thesite of DNA damage. A possible mechanism of arsenic tox-icity
may lie in its ability to react with thiols, for example inzinc
binding structures prevalent in many transcription fac-tors, DNA
repair and cell cycle control proteins. Recentstudies applying a 37
amino acid thiol-containing zincWnger peptide of XPA (XPAzf), a
critical component of thenucleotide excision repair (NER) complex,
where zinc iscomplexed to four cysteines, revealed diVerential
eVects ofarsenite and MMA(III). Interestingly, reaction of
arsenitewith the apopeptide resulted in thiol oxidation of two
orfour cysteine residues, producing one or two disulWdebonds,
respectively. In contrast, reaction of MMA(III) withXPAzf produced
complexes containing two MMAs, or oneMMA with or without oxidation
of the remaining two cys-teines. Thus, zinc-binding structures may
be sensitive tar-gets for arsenicals, even though the actual
species involvedin the speciWc interactions diVers (Piatek et al.
2008).
Deregulation of cell proliferation
Accumulating evidence from cell culture studies, experi-mental
animals and also from arsenic-exposed humans sug-gests that arsenic
alters the DNA methylation pattern,thereby aVecting the expression
of oncogenes and tumorsuppressor genes. Interestingly, both hypo-
and hyper-methylation have been observed. For example,
increasedcytosine methylation in the p53 promotor was detected
in
A549 cells, and hypermethylation with the consequence
ofdiminished gene expression of tumor suppressor genes suchas
p16INK4a and RASSF1A were found in arsenic-exposed A/J mice (Cui et
al. 2006). With respect tohumans, a dose-dependent hypermethylation
of the p53gene was observed in blood samples of arsenic-exposedskin
cancer patients in West Bengal (Chanda et al. 2006).The underlying
mechanisms are still unclear. Whilehypomethylation may be due to
inhibition of DNA-(cyto-sine-5) methyltransferase as in the
instance of cadmium(Takiguchi et al. 2003) or the depletion of
S-adenosylme-thionine, a common cofactor in DNA methylation
andarsenic methylation, hypermethylation is not easily under-stood
and further studies are required to resolve this ques-tion.
Antimony
Compared with arsenic, much less is known about the
car-cinogenicity and the underlying mechanisms of action
ofantimony. Epidemiological studies indicate a carcinogeniceVect of
antimony and antimony compounds on the humanlung, but co-exposure
to confounding substances does notallow a Wnal conclusion. In
animal experiments, inhalationof antimony trioxide and of antimony
trisulWde causedincreased incidences of lung tumors. Metallic
antimony hasbeen classiWed as carcinogenic, too, because after
inhala-tion exposure to metallic antimony the element has
beendetected in human body Xuids, and metallic antimonyshowed the
same toxicity proWle as antimony trioxide inrats (DFG 2007b).
Genotoxic eVects
The genotoxicity of antimony has been reviewed recently(DFG
2007b). In most bacterial mutation assays, inorganicantimony
compounds were inactive. In mammalian cells,trivalent antimony
compounds caused DNA strand breaks,enhanced chromosome aberrations
and micronuclei, butgene mutations were not detected.
In animals, administration of antimony compoundsresulted in
clastogenic eVects in some but not in all studies.Also, there are
indications of an in vivo genotoxic potentialof inorganic antimony
compounds in humans. Exposure ofworkers to antimony trioxide
resulted in oxidative DNAdamage in whole blood.
Putative mechanisms of carcinogenesis
In aqueous solution, both trivalent and pentavalent anti-mony
are stable and both forms may be mutually intercon-verted under
physiological conditions. In reductivebiological milieus,
pentavalent antimony is reduced to the123
-
Arch Toxicol (2008) 82:493512 499trivalent form, which is the
stable form in physiologicalmedia (Reglinski 1998). Trivalent
antimony reacts withsulfhydryl groups of proteins and thereby acts
as anenzyme inhibitor (Gebel et al. 1997). There exist no
con-clusive clues to the mechanism of carcinogenesis by anti-mony.
A direct eVect on DNA is unlikely, since antimonycompounds were not
mutagenic in most bacterial mutationtests. The clastogenic eVects
of antimony trichloride inmammalian cells were not induced by
DNA-protein cross-links, and the induction of micronuclei by
antimony can-not be explained by oxidative stress (SchaumlVel
andGebel 1998). As with many other carcinogenic metals,inhibition
of DNA repair might contribute to the tumori-genic activity of
antimony, but there are no published dataavailable yet.
Furthermore, at present it is not clearwhether or not antimony is
methylated to the same extentas arsenic. To sum up, it is evident
that the mechanismsunderlying the carcinogenicity of antimony are
stillobscure.
Beryllium
Occupational exposure to beryllium and beryllium com-pounds is
associated with increased lung cancer mortality,and inhalation of
beryllium metal, beryllium oxide andberyllium salts caused lung
tumors in rats and rhesus mon-keys (IARC 1993; DFG 2003). The Be2+
ion carries thesame charge as Mg2+ and competes for Mg2+ in
biochemi-cal binding sites such as the phosphate groups of
nucleo-tides and nucleic acids. Like cadmium, beryllium does
notparticipate in redox reactions under physiological
condi-tions.
Genotoxic eVects
The genotoxicity of beryllium has been reviewed (DFG2003; Gordon
and Browser 2003). In an acellular enzy-matic assay, a very high
concentration of BeCl2 (10 mM)interfered with the Wdelity of DNA
synthesis and causedincorporation of mispaired nucleotides (Zakour
and Glick-man 1984). In bacterial assays, beryllium salts were
notmutagenic, whereas in the majority of investigations
withmammalian cells, beryllium salts induced sister
chromatidexchanges, chromosomal aberrations and gene
mutations(Gordon and Browser 2003). In rats exposed to
berylliumsulfate, no increased number of micronuclei was
observed,and in a group of workers exposed to beryllium oxide
noenhanced frequencies of sister-chromatid exchanges andmicronuclei
were found. In BALB 3/T3 cells, berylliumcaused a downregulation of
genes involved in DNA repair(MCM4, MCM5, Rad23 and DNA ligase I)
(Joseph et al.2001). Possibly, beryllium impairs the DNA repair
capacityin mammalians.
Deregulation of cell proliferation
A further mechanism related to the carcinogenicity ofberyllium
may be the deregulation of cell proliferation.Like other
carcinogenic metal compounds, beryllium acti-vates mitogenic
signalling pathways. BeF2 inducedincreased phosphorylation of
mitogenic protein kinases(MEK1, ERK1, p38, MAPK and JNK) and
transcriptionfactors (NFkB and CREB) in human macrophages (Misraet
al. 2002). Furthermore, Be(II) induced the expression ofcellular
proto-oncogenes in vitro. In a study with BALB 3/T3 cells,
beryllium activated the expression of K-ras and c-jun but not of
c-myc, c-fos, c-sis or p53 genes (Keshevaet al. 2001). In another
study with the same cell line, beryl-lium upregulated c-fos, c-jun,
c-myc and c-ras, whereasseveral genes involved in DNA repair (MCM4,
MCM5,Rad23 and DNA ligase I) were downregulated (Josephet al.
2001a). Both upregulation of mitotic signals anddownregulation of
DNA repair functions are thought tocooperate to induce an
unbalanced error-prone cell prolifer-ation. There is also evidence
for epigenetic eVects in beryl-lium carcinogenesis. In lung tumors
of rats induced byexposure to beryllium metal particles, there was
hyperme-thylation of DNA in the tumor suppressor gene INK4,
asso-ciated with a reduced synthesis of the gene productp16INK4a
(Belinsky et al. 2002). Because this protein isinvolved in cell
cycle arrest at the G1-S-phase transition, itsloss may contribute
to tumor progression.
Cadmium
Human cadmium exposure is associated with cancers of thelung and
the kidney (IARC 1993; DFG 2006a). In animals,cadmium induces
carcinomas of the lung after inhalationand cancers of the prostate
after ingestion or injection(Waalkes 2003). Physicochemical
properties of the Cd2+ion may serve as clues to the interpretation
of biologicaleVects: The Cd2+ ion easily substitutes for the
calcium ionin biological systems, because it carries the same
chargeand has a similar radius. Compared to the zinc ion, theradius
of the Cd2+ ion is larger, but still Cd2+ ions can sub-stitute for
Zn2+ ions in many enzymes and transcription fac-tors.
Genotoxic eVects
In rodent experiments, cadmium salts caused increased
fre-quencies of micronuclei and chromosomal aberrations. Invitro,
in mammalian cells cadmium compounds causedDNA strand breaks, gene
mutations and chromosomal aber-rations, whereas in most bacterial
assays soluble cadmiumcompounds were not mutagenic (Waalkes 2003;
DFG2006a). Since cadmium salts do not cause DNA damage in123
-
500 Arch Toxicol (2008) 82:493512cell extracts or with isolated
DNA (Valverde and Rojas2001), the genotoxicity of cadmium has to be
explained byindirect mechanisms. Three frequently discussed
mecha-nisms are (1) the generation of reactive oxygen species,
(2)inhibition of DNA repair enzymes, and (3) deregulation ofcell
proliferation.
Induction of oxidative stress
Although cadmium(II) unlike many other carcinogenicmetal
compounds, is not able to participate in redox reac-tions under
physiological conditions, oxidative stressappears to be a relevant
mechanism of cadmium-inducedgenotoxicity. Cadmium has been shown to
induce the for-mation of reactive oxygen species, both in vitro and
in vivo.Cadmium sulWde induced H2O2 formation in human
poly-morphonuclear leukocytes, and cadmium chlorideenhanced the
production of superoxide in rat and humanphagocytes (Sugiyama
1994). Accordingly, the inductionof DNA strand breaks and
chromosomal aberrations bycadmium in mammalian cells was suppressed
by antioxi-dants and antioxidative enzymes (Ochi et al. 1987;
Stohset al. 2001; Valko et al. 2006). The induction of
oxidativestress by cadmium is interpreted by its inhibitory eVects
onantioxidant enzymes such as catalase, superoxide dismu-tase,
glutathione reductase, and glutathione peroxidase(Stohs et al.
2001; Valko et al. 2006). In addition to theirprobable role in
genotoxicity, reactive oxygen species mayfunction as mitogenic
signals (see below).
Inhibition of DNA Repair
Cadmium is comutagenic and augments the mutagenicityof UV
radiaton, alkylation and oxidation in mammaliancells. These eVects
are explained by the observation thatcadmium inhibits several types
of DNA repair, that is, baseexcision, nucleotide excision, mismatch
repair and theelimination of the premutagenic DNA precursor
8-oxo-dGTP (reviewed by Hartwig and Schwerdtle 2002).Regarding base
excision repair, low concentrations of cad-mium which did not
generate oxidative damage as such,inhibited the repair of oxidative
DNA damage in mamma-lian cells (Dally and Hartwig 1997; Fatur et
al. 2003). Withrespect to nucleotide excision repair, cadmium
interferedwith the removal of thymine dimers after UV irradiation
byinhibiting the Wrst step of this repair pathway, that is,
theincision at the DNA lesion (Hartwig and Schwerdtle 2002;Fatur et
al. 2003). Furthermore, chronic exposure of yeastto very low
cadmium concentrations resulted in hypermuta-bility, and in human
cell extracts cadmium was shown toinhibit DNA mismatch repair (Jin
et al. 2003). Addition-ally, cadmium disturbed the removal of
8-oxo-dGTP fromthe nucleotide pool by inhibiting the 8-oxo-dGTPases
of
bacterial and human origin (Bialkowski and Kasprzak1998). A
molecular mechanism related to the inactivationof DNA repair
proteins involves the displacement of zincfrom zinc Wnger
structures in the DNA repair proteins suchas from the XPA protein,
which is required for nucleotideexcision repair, and from the Fpg
protein, which is involvedin base excision repair in E. coli
(Asmuss et al. 2000). Also,human OGG1 (hOGG1), a glycosylase
responsible for rec-ognition and excision of the premutagenic
8-oxodG duringbase excision repair in mammalian cells, was
inhibited bycadmium (Potts et al. 2003). Even though hOGG1
containsno zinc binding motif itself, its inhibition was shown to
bedue to diminished DNA binding of the zinc Wnger contain-ing
transcription factor SP1 (Youn et al. 2005). Finally,cadmium
induces a conformational shift in the zinc bindingdomain of the
tumor suppressor protein p53 (see below). Inaddition to inhibiting
repair proteins directly, cadmiumdownregulates genes involved in
DNA repair in vivo (Zhouet al. 2004).
The impairment of DNA repair by cadmium may beespecially
deleterious in cadmium-adapted cells. Cadmiuminduces several genes
for cadmium and ROS tolerance suchas those coding for
metallothionein, glutathione synthesisand function, catalase and
superoxide dismutase (Stohset al. 2001). Hence, a condition for
prolonged cell survivalin the presence of cadmium is established
(Chubatsu et al.1992). Taking into account the impairment of DNA
repairby cadmium, tolerance to cadmium toxicity concurrentlymay
constitute an extended chance for the induction of fur-ther
critical mutations (Achanzar et al. 2002).
Deregulation of cell proliferation
Cadmium interacts with a multitude of cellular signal
trans-duction pathways, many of them associated with
mitogenicsignaling. Submicromolar concentrations of cadmium
stim-ulated DNA synthesis and proliferation of rat myoblastcells
(Zglinicki et al. 1992) and of rat macrophages (Misraet al. 2002).
In various cell types in vitro, cadmium evokesreceptor-mediated
release of the second messengers inosi-tol-1,4,5-trisphosphate and
calcium, it activates variousmitogenic protein kinases,
transcription and translation fac-tors, and it induces the
expression of cellular proto-onco-genes c-fos, c-myc and c-jun
(reviewed by Waisberg et al.2003). However, it should be noted that
the activation ofmitogen-activated protein kinases is not a
suYcient condi-tion for enhanced cell proliferation, since
persistent low-dose exposure of cells to cadmium has been shown to
resultnot only in sustained activation of protein kinase ERK
butalso to caspase activation and apoptosis (Martin et al.2006). In
addition to directly stimulating mitogenic signals,cadmium also
inhibits negative controls of cell prolifera-tion. It inactivates
the tumor suppressor protein p53 and123
-
Arch Toxicol (2008) 82:493512 501inhibits the p53 response to
damaged DNA (Meplan et al.1999).
Recently it was reported that cadmium modulates
steroidhormone-dependent signaling in ovaries in rats, in a
breastcancer cell line and in cadmium-transformed prostate
epi-thelial cells (Benbrahim-Tallaa et al. 2007; Brama et al.2007).
Nevertheless, in in-vitro estrogenicity assays basedon estrogen
receptor activity, no eVect of cadmium wasdetected (Silva et al.
2006). It remains an open questionwhether cadmium might promote
tumor growth by anestrogen-mediated mechanism.
In addition to eVects on genes and genetic stability, cad-mium
also exerts epigenetic eVects which may contribute totumor
development. It inhibited DNA-(cytosine-5) methyl-transferase and
diminished DNA methylation during cad-mium-induced cellular
transformation (Takiguchi et al.2003). Decreased DNA methylation is
thought to have atumor promoting eVect, since it was associated
with aug-mented expression of cellular proto-oncogenes.
A unique mechanism by which cadmium deregulatescell
proliferation is the disruption of the cadherin-mediatedcellcell
adhesion system and of cellcell communication(Fig. 5). Cadmium
speciWcally displaced calcium from theprotein E-cadherin
(Prozialeck and Lamar 1999) andimpaired the cellcell adhesion in
kidney epithelial cells(Prozialeck et al. 2003). In conclusion, it
is evident thatcadmium interferes with cellular controls of
proliferation inseveral ways, which all can contribute to the
observedderegulation of cell growth by this metal. However, it is
notyet possible to assess the relative contributions of these
var-ious mechanisms.
Chromium
Chromium occurs in various oxidation states. In technol-ogy, the
prevalent materials are the chromates with hexava-lent chromium,
the chromic compounds with trivalentchromium, and metallic
chromium. Chromates are the tech-
nical intermediates in the manufacturing of Cr(III) com-pounds
and metallic chromium. Therefore, Cr(III)compounds which have not
been puriWed completely, stillcontain traces of hexavalent
chromium; a fact that hascaused erraneous Wndings of genotoxicity
with contami-nated Cr(III) compounds. Chromium traces are essential
forhuman and animal nutrition and are taken up as complexesof
Cr(III) with amino acids (Vincent 2000). Exposure tovarious
chromium(VI) compounds has been consistentlyassociated with
incidences of respiratory cancers in humansand experimental
animals. At variance, there is no evidencefor a carcinogenic action
of trivalent chromium compounds(IARC 1990; ATSDR 2000) and also the
genotoxicity ofCr(VI) is much more pronounced than that of
Cr(III).Hence, genotoxic eVects of Cr(VI) and Cr(III) are
discussedseparately below.
Genotoxic eVects of chromium(VI)
For chromium, the oxidation state is most important for
itsbiochemical activity. Chromium(VI) compounds have beenshown to
exert genotoxicity both in vivo and in vitro. Lym-phocytes of
workers exposed to dusts of chromium(VI)compounds showed elevated
frequencies of DNA strandbreaks (Gambelunghe et al 2003),
sister-chromatidexchanges (Wu et al. 2001) and micronuclei
(Vaglenovet al. 1999; Benova et al. 2002). After intratracheal
instilla-tion in rats, chromate(VI) induced DNA strand breaks
inlymphocytes (Gao et al. 1992). After intraperitoneal injec-tion
but not oral administration of chromate(VI) to mice,micronuclei
were induced in bone marrow (De Flora et al.2006). Intraperitoneal
injection of chromate(VI) induceddominant lethal mutations in rats
(Bataineh et al. 1997). Invitro, soluble chromium(VI) compounds are
mutagenic inmammalian and bacterial test systems (De Flora et
al.1990). The mutagenicity of chromium(VI) in bacteria
isexceptional because other carcinogenic metals are notgenotoxic or
produce equivocal results in bacterial tests.DiVerent from
chromate(VI), chromium(III) compoundsdid not induce genotoxic
eVects in the majority of studieswith intact cells (discussed
below).
Genotoxic eVects of chromium(III)
Cr(III) compounds have not been identiWed as carcinogenicand
their genotoxicity is questionable. Unlike the chromateanion, the
Cr3+ cation is very poorly taken up by cells.There is limited
evidence for a possible genotoxicity ofCr(III) in vitro but not in
vivo. After uptake of chro-mate(VI) and its intracellular reduction
to Cr(III), the latterforms potentially genotoxic complexes with
DNA. ThisWnding implies that Cr(III) can be genotoxic, if it
over-comes the barrier of the plasma membrane (Fig. 6). This
Fig. 5 Interference of cadmium with the cadherin cellcell
adhesionsystem (modiWed from Prozialeck 2003)
123
-
502 Arch Toxicol (2008) 82:493512can be achieved in various
ways. Insoluble particles ofCr2O3 may be taken up by phagocytosis,
and subsequentlybe solubilized in lysosomes to release Cr3+ ions.
These cat-ions bind to cellular macromolecules including
DNA.Indeed, there is some evidence for a low genotoxicity ofCr2O3
although results from diVerent laboratories are con-troversial
(reviewed by De Flora et al. 1990). In intestinalepithelial cells,
Cr(III) can be taken up via the transferrinuptake mechanism.
Another mode of Cr(III) uptakeinvolves synthetic Cr(III) complexes
with hydrophobicligands, which facilitate the permeation of
chromiumthrough plasma membranes. Thus, complexes of Cr(III)with
1,10-phenanthroline or 2,2-bipyridine (Warren et al.1981) or with
picolinic acid (Stearns et al. 2002) are takenup by cells and
induce gene mutations. It has been sug-gested that chromium(III)
might cause DNA damage if itwere reoxidised by intracellular
hydrogen peroxide to pro-duce deleterious intermediates. However,
this reaction hasbeen observed in cell-free systems only and with
high con-centrations of hydrogen peroxide (Tsou et al. 1996).
Role of chromium-DNA adducts in the genotoxicity of Cr(III) and
Cr(VI)
According to the uptake-reduction model developed byWetterhahn
(Connett and Wetterhahn 1983), Cr(VI) as thechromate anion travels
easily through anion channels of theplasma membrane and is reduced
by intracellular electrondonors in three one-electron steps via
chromium(V) andchromium(IV) to the stable form of chromium(III),
whichis accumulated in cells and bound to biochemical
macro-molecules (Fig. 7). The reduction of Cr(VI) keeps the
intra-cellular concentration of chromate low and
facilitateschromium(III) accumulation within cells. This model
hasbeen conWrmed by various investigations with mammaliancells and
in vivo. After treatment of diVerent cell lines withchromium(VI),
the intracellular formation of chromium(V)and chromium(III) has
been demonstrated by EPR spec-
troscopy (Arslan et al. 1987) and X-ray absorption spec-troscopy
(Dillon et al. 1997). Cr(V) was also detected byEPR spectroscopy in
living mice after intraveneous injec-tion of chromium(VI) (Liu et
al. 1994). Whereas theanionic chromate is unable to react with DNA
directly,chromium(III) forms stable complexes with DNA in
chro-mate-treated cells (Fornace et al. 1981; Miller and Costa1988;
Salnikow et al. 1992). Transfection of a bacterio-phage DNA treated
with Cr(III) into E. coli cells lead to adose-dependent increase in
mutation frequency (Snow1991). Also, after transfection of plasmids
with ternaryamino acid-Cr(III)-DNA adducts into human Wbroblasts
invitro, mutations were observed which predominantly weresingle
base substitutions at G:C base pairs (Voitkun et al.1998). Hence,
the formation of Cr-DNA adducts is dis-cussed as a relevant
mechanism for chromium(VI) geno-toxicity (Zhitkovich 2005).
Induction of oxidative stress
In the process of reduction of chromium(VI) to chro-mium(III) by
cellular reductants, such as ascorbate or gluta-thione, potentially
toxic intermediates such as oxygen andsulfur radicals are
generated. In a cell-free assay, chro-mate(VI) reacted with
glutathione to yield chromium(V)and thiyl radicals (Wetterhahn et
al. 1989), whereas thereduction of Cr(VI) with ascorbate resulted
in hydroxylradical formation (Shi et al. 1994). Furthermore,
chro-mium(V) was detected as an intermediate during chro-mium(VI)
reduction in experimental animals (Liu et al.1994). Pentavalent
chromium reacted with isolated DNA toproduce
8-hydroxydeoxyguanosine, whereas hexavalentchromium performed this
reaction only in the presence ofthe reductant glutathione (Faux et
al. 1992). In culturedmammalian cells, chromate(VI) induced
superoxide andnitric oxide production (Hassoun and Stohs 1995),
whereastreatment of cells with chromium(VI) in the presence of
Fig. 6 Cellular uptake and reduction of chromium compounds
No entry!
CrO42
Cr2O3
[Cr(III)Ln]
Cr3+Ions
Cr(VI)O42
Cr(IV)
Cr(V)
Hydrophobiccomplex[Cr(III)Ln]
ParticulateCr2O3
Cr(III)-Transferrin
complex
Cr(III)-Transferrin
Endocytosisand
LysosomalSolubilisation
Cr(III) bound to biochemical ligands
AnionChannel
Fig. 7 Intracellular chromium metabolism, generation of
oxidativestress and eVects on DNA
Cr(V)
X
Cr(IV)
Cr(VI) DNA
ReactiveOxygenSpecies
Cr(III)
e
Cr(III) - DNA
Oxidative Damageto Lipids, Proteins and DNA
e
e
Cellu
larR
educ
tant
s123
-
Arch Toxicol (2008) 82:493512 503glutathione reductase generated
hydroxyl radicals. How-ever, the required concentration of chromate
was in thecytotoxic range (2 mM) (Ye et al. 1995). Regarding
geno-toxicity, the relative contributions of
chromate-generatedoxidative stress on the one hand and Cr(III)-DNA
adductson the other hand are still debated.
Deregulation of cell proliferation
Besides directly causing DNA damage and mutations, chro-mium
(VI) has been shown to activate various mitogen-acti-vated protein
(MAP) kinases via formation of reactiveoxygen species. In a rat
hepatome cell line, low doses ofCr(VI) activated extracellular
signal-regulated kinases ERK-1 and ERK-2 in a persistent way (Kim
and Yurkow 1996)and in human lung carcinoma cells, Cr(VI) activated
threeMAP kinases, c-jun N-terminal kinase JNK and p38 (Chu-ang and
Yang 2001). In addition to activation of mitogenicprotein kinases,
chromium(VI) induced phosphorylation ofmitogenic transcription
factors. Nuclear factor B (NF-B)was activated in a human lymphocyte
culture (Ye et al.1995), and activating transcription factor 2
(ATF-2) and theoncogenic transcription factor c-Jun were activated
inhuman bronchial epithelial cells (Samet et al. 1998). Sincethese
protein kinases and transcription factors are known tobe involved
in both inXammation and tumor growth, theiractivation constitutes a
non-genotoxic mechanism of chro-mate(VI) carcinogenicity in
addition to direct mutagenesis.
Cobalt
Inorganic cobalt compounds, both soluble and particulateforms,
caused lung tumors in animal experiments, whereasthe
epidemiological Wndings of increased lung cancer inci-dence of
cobalt-exposed workers are regarded as not con-clusive because of
co-exposure to other carcinogenicsubstances (IARC 1991, 2006a; DFG
2007c). At variance,workers exposed to cobalt in hard metals
containing tung-sten carbide experienced a signiWcant increase in
lung can-cer (DFG 2007d).
Genotoxicity
The genotoxicity of cobalt and cobalt compounds has beenreviewed
(Beyersmann and Hartwig 1992; DFG 2007c).After intratracheal
instillation in rodents, cobalt(II) chlorideinduced aneuploidies,
micronuclei and chromosome aberra-tions in the bone marrow. In an
inhalation carcinogenicitystudy with cobalt sulfate, mutations in
the K-ras oncogenewere observed in lung tumor tissues of exposed
mice (NTP1998). Soluble cobalt(II) salts induced DNA strand
breaks,gene mutations, sister chromatid exchanges and micronu-clei
in mammalian cells in vitro but were mostly not geno-
toxic in bacterial assays. The underlying mechanismsinclude
direct mutagenicity by oxidative reactions andinterference with DNA
repair processes (see below). Withrespect to metallic cobalt,
cobalt dust caused DNA singlestrand breaks and micronuclei in
mammalian cells in vitro,and these eVects of cobalt were
considerably enhancedwhen cobalt was combined with tungsten carbide
as it isencountered in hard metals (De Boeck et al. 2003).
Induction of oxidative stress
Cobalt ions are able to induce the formation of reactive oxy-gen
species (ROS) in vivo and in vitro. Cobalt(II) catalyzesthe
generation of hydroxyl radicals from hydrogen peroxidein a Fenton
type reaction. After intraperitoneal injection inrats, cobalt(II)
evoked the formation of oxidative DNA basedamage in kidney, liver
and lung (Kasprzak et al. 1994). Theanalysis of mutations in tumor
tissues in a carcinogenicitystudy with cobalt sulfate in mice
revealed that Wve of ninemutations were G-T transversions in codon
12 of the K-rasoncogene (Asmuss et al. 2000; National Toxicology
Program1998). The authors interpret this eVect as evidence
thatcobalt(II) causes oxidative DNA damage. A special mecha-nism
was delineated for the increased genotoxicity of cobaltin
combination with tungsten carbide as it is used in hardmetal dust.
When metallic cobalt was combined with tungstencarbide in an
acellular system, reactive oxygen species weregenerated (Lison et
al. 1995). These authors conclude thattungsten carbide (WC)
catalyzes the transfer of electrons fromcobalt to oxygen to yield
superoxide as depicted in Fig. 8.
Inhibition of DNA repair
The genotoxicity of other mutagenic agents was augmentedby
soluble cobalt salts (Beyersmann and Hartwig 1992) andcobalt metal
dust (De Boeck et al. 1998). In human Wbro-
Fig. 8 Proposed mechanism of reactive oxygen species (ROS)
forma-tion by transfer of electrons from cobalt metal to molecular
oxygen ascatalyzed by tungsten carbide (modiWed from Lison et al.
1995)
O2
ROS
WC
eCo
Co2+
Co
Co Co2+ + 2 e O2 + e ROS123
-
504 Arch Toxicol (2008) 82:493512blasts, cobalt(II) inhibited
the nucleotide excision repair ofDNA damage caused by UV-C
radiation. Both the incisionand polymerisation steps were inhibited
(Kasten et al.1997). In particular, cobalt inhibited the Xeroderma
pig-mentosum group A (XPA) protein, a zinc Wnger proteininvolved in
nucleotide excision repair (Asmuss et al. 2000)where it substituted
for the zinc ion (Kopera et al. 2004).The comutagenicity of cobalt
observed in vitro correspondsto its cocarcinogenic eVect in an
animal study, wherecobalt(II) oxide enhanced the carcinogenicity
ofbenzo[a]pyrene (SteinhoV and Mohr 1991).
Upregulation of hypoxia-inducible factor HIF-1
Cobalt(II) is known to evoke a hypoxia-like state in vivoand in
vitro even in the presence of normal molecular oxy-gen pressure.
The underlying mechanism involves the sta-bilization of
hypoxia-inducible factor HIF-1, whichnormally is degraded when
suYcient oxygen is present. Inthe hypoxic state, HIF-1 acts as a
subunit of a transcrip-tion factor inducing the expression of genes
controllingerythropoietin synthesis, glucose uptake, glycolytic
enzymeactivities, blood vessel formation (angiogenesis) and
otherprocesses allowing cell survival at low oxygen
pressure.Hypoxia is a common feature of tumor tissues, and
thegrowth of tumors beneWts from HIF-1 activation, whichleads to
enhanced glycolytic and angiogenetic activities.For a more detailed
discussion of this area, readers arereferred to the review of
Maxwell and Salnikow (2004)who also discuss similar eVects of
nickel(II).
Lead
The toxicity of lead and its compounds is well known formany
centuries, with anemia and developmental distur-bances being most
prominent. Nevertheless, during the lastyears potential
carcinogenic eVects came into focus, leadingto the classiWcation of
inorganic lead compounds as Proba-bly carcinogenic to humans (Group
2A) by IARC andGroup 2 by the German MAK Commission (considered
tobe carcinogenic to man based on long-term animal studies).These
classiWcations were mainly based on animal experi-ments, where
increased tumor incidences were observed inmultiple organs,
including kidney and brain. Nevertheless,the exact mechanisms are
still unclear, but as with mostother metals and there compounds,
indirect mechanisms likethe induction of oxidative stress and the
interaction withDNA repair processes appear to be relevant.
Genotoxic eVects
Genotoxic eVects of lead compounds are well documentedin
in-vitro systems, experimental animals and in lead-
exposed humans (summarized in IARC 2006b). Equivocalresults have
been published with respect to the mutagenic-ity of water soluble
lead compounds in mammalian cells inculture; in most classical test
systems, eVects were ratherweak and/or restricted to toxic doses.
Nevertheless, whenapplying mammalian AS52 cells carrying a single
copy ofan E. coli gpt gene, which are suited for the detection
ofsmall and large deletions, lead chloride induced mutationsin a
dose-dependent manner, starting at the non-cytotoxicconcentration
of 0.1 M (Ariza and Williams 1996; Arizaet al. 1998; Ariza and
Williams 1999). High mutant fre-quencies and mutation spectra
similar to those induced byreactive oxygen species were also
observed in a diVerentstudy in CHO K1 cells (Yang et al. 1996).
Furthermore,two studies revealed an increase in mutation frequency
incombination with UVC irradiation and MNNG, indicativeof the
disturbance of DNA repair processes (see below). Incontrast to the
equivocal results of gene mutation studies,chromosomal damage and
micronuclei have been observedconsistently in mammalian cells in
culture, in experimentalanimals and in several cases also in
lead-exposed humans;however, with respect to population-based
studies, con-founding exposures cannot be ruled out (reviewed in
IARC2006b). At low concentrations realistic for human expo-sure,
two mechanisms may underlie lead-induced genotoxi-city, namely a
disruption of the pro-oxidant/anti-oxidantbalance and an
interference with DNA repair systems.
Induction of oxidative stress
At diVerent experimental levels, there are strong indicationsfor
the involvement of ROS in lead-induced genotoxicity.Proposed
molecular mechanisms include enhanced lipidperoxidation, inhibition
of antioxidant defense systems,catalysis of Fenton-type reactions
and, interestingly, alsothe long-known inhibition of aminolevulinic
acid dehydra-tase. The latter reaction leads to the accumulation of
theheme precursor aminolevulinic acid, with the
subsequentgeneration or ROS and oxidative DNA damage (reviewedin
IARC 2006b).
DNA repair inhibition
A further mechanism, which has been quite well docu-mented
during the last years, is the interaction of lead withtwo major DNA
repair systems, that is nucleotide excisionrepair and base excision
repair, and comutagenic eVectshave been observed in combination
with UVC radiationand MNNG (reviewed in IARC 2006b). As one
moleculartarget with respect to base excision repair, lead has
beenshown to inhibit the apurinic/apyrimidinic endonuclease(APE1)
in the low micromolar concentration range both inan isolated
enzymic test system and in cultured AA8 cells,123
-
Arch Toxicol (2008) 82:493512 505leading to an accumulation of
apurinic sites in DNA and anincrease in MMS-induced mutagenicity
(McNeill et al.2007). Furthermore, lead interferes with the repair
of DNAdouble strand breaks via interaction with the stress
responsepathway induced by ATM (a phosphoinositol-3-kinaserelated
kinase) (Gastaldo et al. 2007). Due to its high aYn-ity for
sulfhydryl groups, one mechanism for lead interac-tion with
proteins could be the displacement of zinc fromzinc binding
structures. In support of this assumption, incell-free systems lead
has been shown to reduce DNA bind-ing of transcription factors
TFIIIA and Sp1 (Hanas et al.1999; Huang et al. 2004). However, no
impact was seen onthe zinc-containing DNA repair proteins Fpg or
XPA(Asmuss et al. 2000). Thus, zinc binding proteins cannot
beconsidered as general target, but interactions depend on
theactual protein.
Deregulation of cell proliferation
Low concentrations of lead have been shown to stimulatecell
growth (reviewed in IARC 2006b). A probable mecha-nism consists of
the mobilization of free intracellular Ca2+and the activation of
protein kinase C (PKC) by lead, whichtriggers a signal transduction
cascade Wnally leading to thestimulation of DNA synthesis. In
animals, lead signiWcantlyincreases proliferative lesions in the
kidney below cyto-toxic concentrations, indicating that
genotoxicity and accel-erated growth stimuli may act in concert in
lead-inducedcarcinogenicity.
Nickel
Inorganic, both soluble and particulate nickel compoundswere
associated with lung tumors in exposed workers (Doll1990). In
experimental animals, inhalation of particulatenickel(II) compounds
but not nickel(II) sulfate caused lungtumors in rats and mice
(Dunnick et al. 1995). The absenceof carcinogenic eVects of nickel
sulfate in experimental ani-mals may be attributed to the
relatively low maximum tol-erated dose when compared with human
exposure. Atvariance, insoluble particulate nickel oxides and
sulWdesenter cells by phagocytosis, accumulate within cells to
highconcentrations and release nickel ions after gradual
dissolu-tion in lysosomes (Fig. 1) (Costa and Mollenhauer
1980).
Genotoxic eVects
The genotoxicity of nickel and its compounds has beenreviewed
recently (DFG 2006b). Workers exposed to solu-ble nickel compounds
or to poorly soluble sulWdic andoxidic nickel exhibited an elevated
incidence of metaphaseswith gaps, but no signiWcant increase in
sister-chromatidexchanges in lymphocytes. In animal experiments,
intra-
peritoneal injection of soluble nickel salts caused chromo-some
aberrations or micronuclei in some but not in allstudies. In
mammalian cells, nickel(II) ions evoked chro-mosome aberrations,
sister chromatid exchange, DNAbreaks and DNA-protein cross links,
but only at millimolarcytotoxic concentrations. Furthermore,
soluble nickel(II)salts were only weakly mutagenic in mammalian
cells andinactive in almost all bacterial mutagenicity tests.
Threemajor mechanisms are discussed for the genotoxic eVectsof
nickel: generation of reactive oxgen species, interferencewith DNA
repair processes, and epigenetic mechanismsinducing enhanced cell
proliferation. In all mechanistic pro-posals, nickel ions are
regarded as the ultimately genotoxicform of nickel and inorganic
nickel compounds.
Induction of oxidative stress
Like many other carcinogenic metals, nickel compoundsare able to
induce the formation of reactive oxygen species.Nickel ions can
catalyze the generation of hydroxyl radicalsfrom hydrogen peroxide
in a Fenton type reaction. Accord-ingly, in the presence of
hydrogen peroxide, nickel(II) ionsproduce oxidative DNA damage in
isolated DNA and chro-matin (Kasprzak and Hernandez 1989; Lloyd and
Phillips1999). In living cells, the contribution of oxidative
mecha-nisms to the genotoxicity of nickel seems to depend onnickel
speciation and cell type. While NiCl2 in HeLa cellscaused oxidative
DNA damage only at elevated cytotoxicdoses (Dally and Hartwig
1997), soluble nickel carbonatehydroxide induced sister-chromatid
exchanges involvingthe production of reactive oxygen species in
human lym-phocytes at lower concentrations (MBemba-Meka et
al.2007). Furthermore, the redox activity of nickel(II) maychange
considerably if it is complexed to certain aminoacid sequences as
demonstrated in subcellular systems forhistone binding (Bal et al.
2000).
Inhibition of DNA repair
Nickel is a distinct comutagen and it interferes with variousDNA
repair pathways. Nickel ions enhanced the the muta-genicity of
methyl methanesulfonate in E. coli and theinduction of mutations
and sister chromatid exchanges byUV radiation in hamster cells.
These comutagenic eVectsare explained by the inhibition of all
major types of DNArepair processes. DNA excision repair, repair of
O6-alkyl-guanine and repair of oxidative DNA damage were inhib-ited
at subtoxic concentrations of nickel(II) chloride, whichwere not
yet mutagenic themselves (Dally and Hartwig1997; Hartwig et al.
1994; Iwitzki et al. 1998; Kruegeret al. 1999). Recently, in human
bronchial epithelial cellstransformed by nickel sulWde, silencing
of the O6-methyl-guanine-DNA methyltransferase gene was observed
(Ji123
-
506 Arch Toxicol (2008) 82:493512et al. 2008). Furthermore, the
degradation of the promuta-genic DNA precursor 8-oxo-dGTP by a
speciWc GTPase isalso inhibited by nickel(II) (Porter et al. 1997).
The comu-tagenic properties of nickel ions are also reXected by
epide-miological results. Occupational exposure to readilysoluble
nickel salts led to lung tumors only at relativelyhigh exposure
levels, but it increased the tumor incidenceafter simultaneous
exposure to either poorly soluble nickelcompounds (Doll 1990) or
tobacco smoke (Andersen et al.1996).
Deregulation of cell proliferation
In addition to its genotoxic activity, nickel deregulates
nor-mal growth control by several epigenetic mechanisms(reviewed by
Salnikow and Zhitkovich 2008). In culturedmammalian cells, nickel
chloride caused increased methyl-ation of cytosine bases and
decreased expression of tumorsuppressor genes resulting in
accelerated cell proliferation.Also in nickel-induced tumors, DNA
hypermethylation wasobserved together with reduced expression of
tumor sup-pressor genes p16 and Fhit. As a second epigenetic
mecha-nism, nickel chloride inhibits acetylation of several
histonesfollowed by chromatin condensation in eukaryotic
cells,probably by binding of nickel ions to histone proteins.Since
histone acetylation aids the access of transcriptionfactors to DNA,
inhibition of histone acetylation is believedto contribute to the
observed silencing of telomeric markergenes. As a third mechanism,
the activation of hypoxic sig-naling is suggested. Nickel ions are
strong inducers of thehypoxia-inducible factor HIF-1 and
HIF-dependent tran-scription. Mimicking of the hypoxic state may
provide themetabolic condition for the selection of transformed
cellsthat have altered energy metabolism, changed growth con-trol
and resistance to apoptosis (reviewed by Maxwell andSalnikow
2004).
Vanadium
Vanadum occurs in the oxidation states 0, +2, +3, +4, and+5. In
the presence of oxygen, pentavalent vanadium is thestable state,
whereas in biological media both the vanadateanion H2VO4 with
pentavalent vanadium and the vanadylcation VO2+ with tetravalent
vandium are stable and mutu-ally interconverted easily.
Divanadium(V) pentoxideinduced lung tumors in mice and rats (NTP
2002).
Genotoxicity
The genotoxicity of vanadium compounds has beenreviewed recently
(IARC 2006a; DFG 2006c). In animalexperiments, vanadium(V) and
vanadium(IV) compoundsinduced micronuclei, vanadium(V) compounds
caused
chromosomal aberrations and aneuploidy in bone marrowcells. Both
vanadium(IV) and vanadium(V) compoundswere positive in dominant
lethal tests. In human cells invitro, vanadium(V) compounds induced
DNA strandbreaks. In mammalian cells, vanadium(III),
vanadium(IV)and vanadium(V) compounds caused the formation of
chro-mosome aberrations and vanadium(IV) and vanadium(V)induced
aneuploidies in mammalian cells. Similar to mostcarcinogenic metal
compounds, vanadate(V) exhibited noconsistent results in bacterial
mutagenicity assays. Thegenotoxicity of vanadium compounds is
interpreted bymechanisms of induction of oxidative stress,
inhibition ofDNA repair and interference with the activity of
proteinphosphatases and kinases. The observed induction of
aneu-ploidy by vanadate(V) is interpreted by the inhibition
ofspindle formation and the disruption of microtubule assem-bly
(Ramirez et al. 1997; Mailhes et al. 2003). Vanadate(V)inhibits
relative speciWcally the activity of protein-tyrosinephosphatases
(Stankiewicz et al. 1995). Because theseenzymes regulate the
aggregation of the meiotic spindleand the spindle checkpoint during
meiosis, the inhibition ofprotein-tyrosine phophatases may
contribute to the geno-toxicity of vanadium(V).
Induction of oxidative stress
The genotoxicity of vanadium(V) can be attributed to oxi-dative
mechanisms. In a cell-free system containing ratliver microsomes,
vanadate(V) was reduced by NADH tovanadium(IV) and generated
hydroxyl radicals as detectedby ESR spectroscopy (Shi and Dalal
1992). Vanadate(V)reacted with thiols to produce vanadium(IV) and
thiyl radi-cals (Shi et al. 1990). Vanadyl(IV) sulfate catalyzed
thereaction of 2-deoxyguanosin with molecular oxygen toform
8-hydroxydeoxyguanosin, and it caused strand breaksin isolated
plasmid DNA (Shi et al. 1996).
Interference with DNA repair
In addition to its own genotoxicity, vanadium(V) mayenhance the
eVects of other genotoxic agents. In humanWbroblasts, a low
concentration of vanadate(V) (1 M)impaired the repair of DNA damage
caused by UV irradia-tion or by bleomycin (Ivancsits et al.
2002).
Deregulation of cell proliferation
Inhibition of protein tyrosine phosphatases by vanadate(V)is
thought to enhance mitogenic signalling, because inhibi-tion of
dephosphorylation stabilizes active phosphorylatedproteins. In
mammalian cells, vanadyl(IV)sulfate activatedphosphatidylinositol-3
kinase, and vanadyl(IV)sulfate andvanadate(V) stimulated
mitogen-activated protein kinases123
-
Arch Toxicol (2008) 82:493512 507ERK-1 and ERK-2 (Pandey et al.
1999; Wang and Bonner2000). In mouse epidermis cells, vanadate(V)
activatedprotein kinase B (Akt kinase) and stimulated the entry
ofcells into the S-phase (Zhang et al. 2004). A further mecha-nism
stimulating proliferation is the activation of severaltranscription
factors by oxidative mechanisms. In a murinemacrophage cell line,
vanadate(V) induced the activation ofTNF (tumor necrosis factor )
(Ye et al. 1999), and inmurine epidermis cells vanadate(V)
activated the transcrip-tion factor AP-1 (activator protein 1)
(Ding et al. 1999). Onthe level of gene exporession, vanadate(V)
activated theproliferin gene and induced morphological cell
transforma-tion of murine Wbroblasts (Parfett and Pilon 1995).
Thesestimulatory eVects of vanadium compounds on
mitogenicsignalling enzymes, transcription factors and gene
expres-sion are thought to promote cell transformation and
malig-nant growth by carcinogenic vanadium compounds.
Conclusions
Carcinogenic metals are widely distributed over the peri-odic
table of the elements. They occur in eight diVerentgroups from
clear-cut metals to metalloids, from hard met-als like beryllium,
which form ionic compounds only, tosoft metals like lead which are
able to form covalent bonds.In spite of the wide range of
physicochemical properties,some common mechanisms of carcinogenesis
emergewhich can be regarded as typical for metal carcinogens
ingeneral. They include the induction of oxidative
stress,inhibition of DNA repair, activation of mitogenic
signal-ling, and epigenetic modiWcation of gene expression,
whichmay even be based on the same or similar molecular inter-
actions. Figure 9 gives an overview over these genetic
andepigenetic mechanisms ultimately concurring in the dereg-ulation
of cell growth and development of tumors. Never-theless, each metal
and also each metal species exertcharacteristic interactions, and
even though similar cellularpathways are aVected, the underlying
mechanisms are quitediVerent.
One decisive factor in metal carcinogenesis is the
bio-availability of diVerent metal species, and an important
bar-rier is the cell membrane. Depending largely on the
actualspecies present in physiological environments, metals
canenter the cell via anion channels or cation transporters.
Poorwater soluble particulate metal species may be endocytosedand
are gradually dissolved in the acidic environment of thelysosomes,
where respective metal ions are deliberated anddistributed within
the cytoplasm and also the nucleus. Thepotential impact of the
membrane passage is most evidentin case of chromium compounds:
while the human and ani-mal carcinogen chromium(VI) is readily
taken up via theanion transporter due to its similarity to sulfate,
the cellmembrane is nearly impermeable for chromium(III), forwhich
no carcinogenicity has been observed so far.
Once inside the cell, in most cases, the DNA appears notto be
the primary binding site for carcinogenic metal ions.Even though
due to their cationic character, in principlethey can form adducts
with DNA bases as shown in isolatedsystems, interactions with
proteins appear to be preferred inintact cells. One important
exception is chromium(VI):after its intracellular reduction to
chromium(III), it bindsreadily to DNA forming DNA-protein and
DNADNAcrosslinks.
Nevertheless, in spite of the missing DNA binding, theinduction
of oxidative DNA damage has been observed for
Fig. 9 Major mechanisms in metal carcinogenicity. Not shown are
unique mechanisms found with speciWc metal com-pounds such as
chromium-DNA adduct formation, cadmium interference with cellcell
adhe-sion or vanadate inhibition of protein phosphatases
Inhibition of DNA Repair
Inhibition of AntioxidantDefences
Activation of Mitotic Signalling
Modulation ofGene Expression
DecreasedGenomic Stabilty
Oxidative Stress
Induction of Protooncogenes
Inactivation of TumorSuppressorGenes
Accumulationof CriticalMutations
Deregulation of Cell Proliferation
MetalCompound
TumorDevelopment123
-
508 Arch Toxicol (2008) 82:493512most metals, and common
mechanisms include the interfer-ence with the cellular defense
system against reactive oxy-gen species, including DNA repair
systems, and/or thecatalysis of Fenton-type reactions where
endogenouslyformed ROS like hydrogen peroxide are converted into
thefar more reactive hydroxyl radicals. Furthermore, ROS maybe
generated also in the course of intracellular reduction ofmetals,
as is the case of chromium(VI) reduction to chro-mium(III) with
instable chromium(V) and chromium(IV)intermediates, and also redox
reactions occurring for exam-ple in the course of methylation of
arsenite within meta-bolic competent cells. In many cases, the
relevance ofoxidative DNA damage for metal carcinogenesis
remainsquestionable, since in experimental systems, frequently
butnot in all cases, comparatively high concentrations arerequired
to yield signiWcant increases of the endogenouslevel of oxidative
DNA damage. However, oxidative modi-Wcations may also play a role
in the interaction with pro-teins, as outlined below.
For most metal compounds, interactions with proteinsappear to be
more relevant for carcinogenicity as comparedto direct DNA damage,
and several targets have been iden-tiWed, such as DNA repair, tumor
suppressor and signaltransduction proteins. Even though diVerent
metal com-pounds exert diVerent eVects on protein functions,
commonmechanisms include the displacement of essential metalions
and/or the oxidation of critical amino acids leadingalso to altered
redox regulation, perhaps best investigatedfor DNA repair proteins.
Since metal ions can bind in prin-ciple to many electron rich
centers in proteins, this raisesthe question whether there are
particularly metal-sensitiveprotein structures. During the last
years, so-called zincWnger proteins have been identiWed as
potential moleculartargets for toxic metal compounds. They
represent a familyof proteins where zinc is complexed through four
invariantcysteine and/or histidine residues forming a zinc
Wngerdomain, which is mostly involved not only in DNA bindingbut
also in proteinprotein interactions (Mackay and Cross-ley, 1998).
Besides transcription factors, several proteinsinvolved in DNA
damage signaling and repair belong tothis family, and also the
tumor suppressor protein p53 has azinc binding structure in its DNA
binding domain, essentialfor its transcriptional activity. For
several zinc Wnger pro-teins, molecular interactions with toxic
metal ions havebeen elucidated in detail. Thus, cadmium can
substitute forzinc in the zinc Wnger domain of the nucleotide
excisionrepair protein XPA, leading to structural distortions
whichdisturb its correct function within the nucleotide
excisionrepair complex. In contrast, nickel can substitute for zinc
inthe XPA protein and increase its sensitivity towards oxidiz-ing
agents. Perhaps most relevant are the results in the caseof
arsenite and its trivalent methylated metabolites. Whileall of them
inhibit the poly(ADP-ribosyl)ation, mediated
predominantly by the zinc Wnger protein PARP-1 atextremely low
concentrations, detailed molecular studieswith the zinc Wnger
structure of XPA revealed an oxidationof the zinc complexing thiol
groups for arsenite, while incase of MMA(III) the binding to the
thiol group andthereby the replacement of zinc was observed.
Finally, withrespect to cadmium, an unfolding of the zinc
bindingdomain of p53 was observed, leading to a complete loss
oftumor suppressor functions. Thus, there is accumulatingevidence
for zinc binding structures being very sensitivetargets for toxic
metal compounds, but whether or not therespective proteins are
indeed inhibited and if so by whichmechanism depends on the speciWc
interaction of the metalion with the respective protein. Decisive
factors appear tobe not only physicochemical properties but also
assessibil-ity and the microenvironment within the protein
underinvestigation.
Altogether, the inhibition of DNA repair processes andthe
interference with cell growth, cell cycle control andtumor
suppressor functions appears to be more evident forcarcinogenic
metal compounds as opposed to direct muta-genicity. Nevertheless,
the outcome, that is the decrease ingenomic stability, is very
similar or even more severe.Since DNA repair systems not only
provide pronouncedprotection towards environmental mutagens, but
alsotowards endogenous DNA damage occurring permanently,for example
due to oxygen metabolism their disturbanceresults in an increase in
mutations and carcinogenesis. Thisis evident, for example, in the
high tumor frequency inpatients with the DNA repair disorder
Xeroderma pigmen-tosum. Nevertheless, other mechanisms contribute
as well,such as epigenetic alterations of gene expression and
altera-tions in signal transduction pathways leading, for
example,to growth stimulation or deregulated apoptosis; the
applica-tion of new techniques like genomics and proteomics
willprovide much more information in the near future. Also,these
common mechanisms do not exclude the existence ofunique
interactions of speciWc metal species, such as thebinding of
vanadate to phosphate binding sites. Consideringrisk assessment,
future research will have to focus on therelevance of the
respective mechanisms in experimentalanimals and exposed humans,
especially with respect toeVective concentrations.
References
Achanzar WE, Webber MM, Waalkes MP (2002) Altered apoptoticgene
expression and acquired apoptotic resistance in cadmium-transformed
human prostate epithelial cells. Prostate 52:236244
Andersen A, Berge SR, Engeland A, Norseth T (1996) Exposure
tonickel compounds and smoking in relation to incidence of lungand
nasal cancer among nickel reWnery workers. Occup EnvironMed
53:708713123
-
Arch Toxicol (2008) 82:493512 509Aposhian HV, Aposhian MM (2006)
Arsenic toxicology: Wve ques-tions. Chem Res Toxicol 19:115
Ariza ME, Williams MV (1996) Mutagenesis of AS52 cells by
lowconcentrations of lead(II) and mercury(II). Environ Mol
Mutagen27:3033
Ariza ME, Williams MV (1999) Lead and mercury mutagenesis:
typeof mutation dependent upon metal concentration. J Biochem
MolToxicol 13:107112
Ariza ME, Bijur GN, Williams MV (1998) Lead and mercury
muta-genesis: role of H2O2, superoxide dismutase, and xanthine
oxi-dase. Environ Mol Mutagen 31:352361
Arslan P, Beltrame M, Tomasi A (1987) Intracellular chromium
reduc-tion. Biochim Biophys Acta. 931:1015
Asmuss M, Mullenders LH, Eker A, Hartwig A (2000)
DiVerentialeVects of toxic metal compounds on the activities of Fpg
andXPA, two zinc Wnger proteins involved in DNA repair.
Carcino-genesis 21:20972104
ATSDR (Agency for Toxic Substances Disease Registry) (2000)
Tox-icological proWle of chromium. US Department of Health and
Hu-man Services, Public Health Services Atlanta, USA
Bal W, Liang R, Lukszo J, Lee SH, Dizdaroglu M, Kasprzak
KS(2000) Ni(II) speciWcally cleaves the C-terminal tail of the
majorvariant of histone H2A and forms an oxidative
damage-mediatingcomplex with the cleaved-oV octapeptide. Chem Res
Toxicol13:616624
Bataineh H, al-Hamood MH, Elbetieha A, Bani Hani I (1997) EVect
oflong-term ingestion of chromium compounds on aggression,
sexbehavior and fertility in adult male rat. Drug Chem
Toxicol20:133149
Belinsky SA, Snow SS, Nikula KJ, Finch GL, Tellez CS,
PalmisanoWA (2002) Aberrant CpG island methylation of the
p16(INK4a)and estrogen receptor genes in rat lung tumors induced by
partic-ulate carcinogens. Carcinogenesis 23:335339
Benbrahim-Tallaa L, Liu J, Webber MM, Waalkes MP (2007)
Estro-gen signaling and disruption of androgen metabolism in
acquiredandrogen-independence during cadmium carcinogenesis in
hu-man prostate epithelial cells. Prostate 67:135145
Benova D, Hadjidekova V, Hristova R, Nikolova T, Boulanova
M,Georgieva I, Grigorova M, Popov T, Panev T, Georgieva R,
Nat-arajan AT, Darroudi F, Nilsson R (2002) Cytogenetic eVects
ofhexavalent chromium in Bulgarian chromium platers. Mutat
Res514:2938
Beyersmann D (1995) Physicochemical aspects of the interference
ofdetrimental metal ions with normal metal metabolism. In: Ber-thon
G (ed) Handbook on metal-ligand interactions in biologicalXuids.
Marcel Dekker, New York, pp 813826
Beyersmann D, Hartwig A (1992) The genetic toxicology of
cobalt.Toxicol Appl Pharmacol 115:137145
Bialkowski K, Kasprzak KS (1998) A novel assay of
8-oxo-2-deoxy-guanosine 5-triphosphate pyrophosphohydrolase
(8-oxo-dGT-Pase) activity in cultured cells and its use for
evaluation ofcadmium (II) inhibition of this activity. Nucleic
Acids Res26:31943201
Brama M, Gnessi L, Basciani S, Cerulli N, Politi L, Spera G,
MarianiS, Cherubini S, Scotto dAbusco A, Scandurra R, Migliaccio
S(2007) Cadmium induces mitogenic signaling in breast cancercell by
an ER-dependent mechanism. Mol Cell Endocrinol264:102108
Chanda S, Dasgupta UB, Guhamazumder D, Gupta M, Chaudhuri
U,Lahiri S, Das S, Ghosh N, Chatterjee D (2006) DNA
hypermethy-lation of promoter of gene p53 and p16 in
arsenic-exposed peoplewith and without malignancy. Toxicol Sci
89:431437
Cheung WY (1984) Calmodulin: its potential role in cell
proliferationand heavy metal toxicity. Fed Proc 43:29952999
Chuang SM, Yang JL (2001) Comparison of roles of three
mitogen-activated protein kinases induced by chromium(VI) and
cadmium
in non-small-cell lung carcinoma cells. Mol Cell
Biochem222:8595
Chubatsu LS, Gennari M, Meneghini R (1992) Glutathione is the
anti-oxidant responsible for resistance to oxidative stress in V79
Chi-nese hamster Wbroblasts rendered resistant to cadmium. ChemBiol
Interact 82:99110
Connett PH, Wetterhahn KE (1983) In vitro reaction of the
carcinogenchromate with cellular thiols and carboxylic acids. J Am
ChemSoc 107:42824288
Costa M, Mollenhauer HH (1980) Carcinogenic activity of
particulatenickel compounds is proportional to their cellular
uptake. Science209:515517
Cui X, Wakai T, Shirai Y, Hatakeyama K, Hirano S. (2006)
Chronicoral exposure to inorganic arsenate interferes with
methylationstatus of p16INK4a and RASSF1A and induces lung cancer
in A/J mice. Toxicol Sci 91:372381
Dally H, Hartwig A (1997) Induction and repair inhibition of
oxidativeDNA damage by nickel(II) and cadmium(II) in mammalian
cells.Carcinogenesis 18:10211026
De Boeck M, Lison D, Kirsch-Volders M (1998) Evaluation of the
invitro direct and indirect genotoxic eVects of cobalt compounds
us-ing the alkaline comet assay. InXuence of interdonor and
interex-perimental variability. Carcinogenesis 19:20212029
De Boeck M, Lombaert N, de Backer S, Finsy R, Lison D,
Kirsch-Vol-ders M (2003) In vitro eVects of diVerent combinations
of cobaltand metallic carbide particles. Mutagenesis 18:177186
De Flora S, Bagnasco M, Serra D, Zanacchi P (1990) Genotoxicity
ofchromium compounds. A review. Mutat Res 238:99172
De Flora S, Iltcheva M, Balansky RM (2006) Oral chromium(VI)
doesnot aVect the frequency of micronuclei in hematopoietic cells
ofadult mice and of transplacentally exposed fetuses. Mutat
Res610:3847
DFG (2003) Beryllium und seine anorganischen Verbindungen.
In:Greim H (ed) Gesundheitsschdliche ArbeitsstoVe,
toxi-kologisch-medizinische Begrndungen von MAK-Werten.Wiley-VCH,
Weinheim
DFG (2006a) Cadmium and its compounds (in the form of
inhalabledusts/aerosols). In: Deutsche Forschungsgemeinschaft (ed)
TheMAK collection for occupational health and safety. Part I:
MAKvalue documentations, vol 22. Wiley-VCH, Weinheim, pp 119146
DFG (2006b) Nickel and its inorganic compounds. In: Deutsche
Fors-chungsgemeinschaft (ed) The MAK collection for
occupationalhealth and safety. Part I: MAK value documentations,
vol 22.Wiley-VCH, Weinheim, pp141
DFG (2006c) Vanadium und seine anorganischen Verbindungen.
In:Greim H (ed) Gesundheitsschdliche ArbeitsstoVe,
toxi-kologisch-medizinische Begrndungen von MAK-Werten.Wiley-VCH,
Weinheim
DFG (2007a). Deutsche Forschungsgemeinschaft: List of MAK andBAT
values 2007. Wiley-VCH, Weinheim.
DFG (2007b) Antimony and its inorganic compounds (inhalable
frac-tion). In: Deutsche Forschungsgemeinschaft (ed) The MAK
col-lection for occupational health and safety. Part I: MAK
valuedocumentations, vol 23. Wiley-VCH, Weinheim, pp 173
DFG (2007c) Cobalt and its compounds (inhalable dusts or
aerosols).In: Deutsche Forschungsgemeinschaft (ed) The MAK
collectionfor occupational health and safety. Part I: MAK value
documen-tations, vol 23. Wiley-VCH, Weinheim, pp 75113
DFG (2007d) Hard metal containing tungsten carbide and cobalt
(inha-lable fraction). In: Deutsche Forschungsgemeinschaft (ed)
TheMAK collection for occupational health and safety. Part I:
MAKvalue documentations, vol 23. Wiley-VCH, Weinheim, pp 217234
Dillon CT, Lay PA, Cholewa M, Legge GJ, Bonin AM, Collins
TJ,Kostka KL, Shea-McCarthy G (1997) Microprobe X-ray
absorp-123
-
510 Arch Toxicol (2008) 82:493512tion spectroscopic
determination of the oxidation state of intracel-lular chromium
following exposure of V79 Chinese hamster lungcells to genotoxic
chromium complexes. Chem Res Toxicol10:533535
Ding M, Li JJ, Leonard SS, Ye JP, Shi X, Colburn NH, Castranova
V,Vallyathan V (1999) Vanadate-induced activation of
activatorprotein-1: role of reactive oxygen species.
Carcinogenesis20:663668
Doll R (1990) Report of the international committee on nickel
carcino-genesis in man. Scand J Work Environ Health 16:182
Dunnick JK, Elwell MR, Radovsky AE, Benson JM, Hahn FF,
NikulaKJ, Barr EB, Hobbs CH. (1995) Comparative carcinogenic
eVectsof nickel subsulWde, nickel oxide, or nickel sulfate
hexahydratechronic exposures in the lung. Cancer Res
55:52515256
Fatur T, Tusek M, Falnoga I, Scancar J, Lah TT, Filipic M (2003)
Cad-mium inhibits repair of UV-, methyl methanesulfonate- and
N-methyl-N-nitrosourea-induced DNA damage in Chinese hamsterovary
cells. Mutat Res 529:109116
Faux SP, Gao M, Chipman JK, Levy LS (1992) Production of
8-hy-droxydeoxyguanosine in isolated DNA by chromium(VI)
andchromium(V). Carcinogenesis 13:16671669
Fornace AJ Jr, Seres DS, Lechner JF, Harris CC (1981)
DNA-proteincross-linking by chromium salts. Chem Biol Interact
36:345354
Gao M, Binks SP, Chipman JK, Levy LS, Braithwaite RA, Brown
SS(1992) Induction of DNA strand breaks in peripheral lymphocytesby
soluble chromium compounds. Hum Exp Toxicol 11:7782
Gambelunghe A, Piccinini R, Ambrogi M, Villarini M, Moretti
M,Marchetti C, Abbritti G, Muzi G (2003) Primary DNA damage
inchrome-plating workers. Toxicology 188:187195
Gastaldo J, Viau M, Bencokova Z, Joubert A, Charvet AM, Balosso
J,Foray N (2007) Lead contamination results in late and
slowlyrepairable DNA double-strand breaks and impacts upon
theATM-dependent signaling pathways. Toxicol Lett 173:201214
Gebel T, Christensen S, Dunkelberg H (1997) Comparative and
envi-ronmental genotoxicity of antimony and arsenic. Anticancer
Res17:26032608
Genestra M (2007) Oxyl radicals, redox-sensitive signalling
cascadesand antioxidants. Cell Signal 19:18071819
Gordon T, Browser D (2003) Beryllium: genotoxicity and
carcinoge-nicity. Mutat Res 533:99105
Hanas JS, Rodgers JS, Bantle JA, Cheng YG (1999) Lead inhibition
ofDNA-binding mechanism of Cys(2) His(2) zinc Wnger proteins.Mol
Pharmacol 56:982988
Hartwig A (2001) Zinc Wnger proteins as potential targets for
toxicmetal ions: diVerential eVects on structure and function.
AntioxidRedox Signal 3:625634
Hartwig A (2007) Kanzerogene Metallverbindungen. Aktuelle
As-pekte zu Wirkungsmechanismen und Risikobewertung.
Oester-reichisches Forum Arbeitsmedizin 01/07:510
Hartwig A, Schwerdtle T (2002) Interactions of carcinogenic
metalcomounds with DNA repair processes: toxicological
implica-tions. Toxicol Lett 127:4754
Hartwig A, Mullenders LH, Schlepegrell R, Kasten U, Beyersmann
D(1994) Nickel(II) interferes with the incision step in
nucleotideexcision repair in mammalian cells. Cancer Res
54:40454051
Hartwig A, Pelzer A, Asmuss M, Burkle A (2003) Very low
concen-trations of arsenite suppress poly(ADP-ribosyl) ation in
mamma-lian cells. Int J Cancer 104:16
Hassoun EA, Stohs SJ (1995) Chromium-induced production of
reac-tive oxygen species, DNA single-strand breaks, nitric oxide
pro-duction, and lactate dehydrogenase leakage in J774A.1
cellcultures. J Biochem Toxicol 10:315321
Huang M, Krepkiy D, Hu W, Petering DH (2004) Zn-, Cd-, and
Pb-transcription factor IIIA: properties, DNA binding, and
compari-son with TFIIIA-Wnger 3 metal complexes. J Inorg
Biochem98:775785
International Agency for Research on Cancer (IARC) (1990)
Chro-mium, nickel and welding. IARC monographs on the evaluationof
carcinogenic risks to humans 49. IARC, Lyon, pp 49256
International Agency for Research on Cancer (IARC) (1991)
Chlori-nated drinking water; chlorination byproducts; some other
halo-genated compounds; cobalt and cobalt compounds. IARCmonographs
on the evaluation of carcinogenic risks to humans 52.IARC, Lyon, pp
363472
International Agency for Research on Cancer (IARC) (1993)
Beryl-lium, cadmium, mercury, and exposures in the glass
manufactur-ing Industry. IARC monographs on the evaluation
ofcarcinogenic risks to humans, vol 58. Lyon, pp 119237
International Agency for Research on Cancer (IARC) (2006a)
Cobaltin hard metals and cobalt sulfate, gallium arsenide, indium
phos-phide and vanadium pentoxide. IARC monographs on the
evalu-ation of carcinogenic risks to humans, vol 86. Lyon, pp
119237
International Agency for Research on Cancer IARC (2006)
Inorganicand organic lead compounds. IARC Monogr Eval Carcinog
RisksHum 87:1471
Ivancsits S, Pilger A, Diem E, SchaVer A, Rdiger HW (2002)
Vana-date induces DNA strand breaks in cultured human Wbroblasts
atdoses relevant to occupational exposure. Mutat Res 519:2535
Iwitzki F, Schlepegrell R, Eichhorn U, Kaina B, Beyersmann D,
Har-twig A (1998) Nickel(II) inhibits the repair of
O6-methylguaninein mammalian cells. Arch Toxicol. 72:681689
Ji W, Yang L, Yu L, Yuan J, Hu D, Zhang W, Yang J, Pang Y, Li
W,Lu J, Fu J, Chen J, Lin Z, Chen W, Zhuang Z. (2008)
Epigeneticsilencing of O6-methylguanine-DNA methyltransferase gene
inNiS-transformed cells. Carcinogenesis, 19 January 2008 (Epubahead
of print)
Jin YH, Clark AB, Slebos RJ, Al-Refal H, Taylor JA, Kunkel TA,
Re-snick MA, Gordenin DA (2003) Cadmium as a mutagen that actsby
inhibiting mismatch repair. Nat Genet 34:326329
Joseph P, Muchnok T, Ong TM (2001) Gene expression proWle
inBAL