C H A P T E R

12

Gene-Environment Interactions for Metals

KARIN BROBERG, KARIN ENGSTRÖM, AND SHEGUFTA AMEER

ABSTRACT causal components, e.g. studies of human twins,

families, or populations, as well as animal studies in
It has become increasingly clear that the individual knockout mice or inbred rodent strains. However,
genetic background influences susceptibility to metal most studies within the field are based on human data
toxicity. Genetic variation in genes that regulate metal because genetic variation varies greatly between spe-
toxicokinetics and toxicodynamics influence the degree cies, and in some cases metal metabolism and metal
of metal accumulation and retention in the body, as toxicity is also different between humans and animal
well as toxic effects. Moreover, factors that regulate models such as mice and rats.
gene expression, so-called epigenetic factors, have been The type of interaction in which a genetic compo-
identified as targets for metal toxicity. This chapter nent is thought to modify the exposure-disease rela-
addresses what is currently known about such gene- tionship, is often referred to as a gene-environment
environment interactions. The picture that emerges for interaction and the genes involved as susceptibility genes.
most metals is that the genetic influence is probably The genes under study can be divided into genes that
not attributed to a single gene for each metal; rather it modify uptake and retention (so-called toxicokinetics)
is polygenic, with some genes having a stronger effect or genes that modify the toxic effects (toxicodynamics)
than others. The presence of variants of the human leu- of metals. Moreover, new genetic factors, so-called epi-
kocyte antigen system and the risk of beryllium-related genetic factors, have been identified as targets for metal
pulmonary disease was one of the first and maybe the toxicity.
strongest example of a gene-environment interaction. Metals are in several aspects ideal for studying
There are also clear gene-environment interactions for gene-environment interactions: their metabolism is
arsenic and lead. Evidence is rapidly growing for epi- often not very complex and involves only a few enzy-
genetic effects of metals, e.g. for arsenic, cadmium, and matic reactions, which means that candidate genes
lead, which may explain the association between metal for metal toxicokinetics can be more easily identified.
exposure early in life and toxic effects later in life, as Arsenic provides one of the clearest examples: genetic
well as metal carcinogenicity. variation in AS3MT, the major arsenic-metabolizing
gene, significantly contributes to arsenic metabolism
efficiency (Engström et al., 2011a; Pierce et al., 2012).
1 GENE-ENVIRONMENT INTERACTIONS The examples given here will be discussed in more
FOR METALS detail below, in the specific metal sections.
Metal toxicity has been studied for a long time and
It is generally known that there is a large interindi- consequently much is known about toxicodynamics
vidual variation within species in sensitivity to metal and the mechanism of action. There are several exam-
exposure. Different types of studies have shown that ples of specific genes involved in metal-related disease
part of this variation in susceptibility is due to genetic processes. One of the first to be identified was a human

Handbook on the Toxicology of Metals 4E

http://dx.doi.org/10.1016/B978-0-444-59453-2.00012-3 239 Copyright © 2015 Elsevier B.V. All rights reserved.
240 Karin Broberg, Karin Engström, and Shegufta Ameer

leukocyte antigen (HLA) variant that is involved in the 1979; Lalouel et al., 1985). It is caused by mutations in
presentation of beryllium to the immune system and the HFE gene that lead to increased iron uptake from
confers a high risk of beryllium-related pulmonary the gastrointestinal tract (Feder et al., 1996). The afore-
disease. In comparison, for many other types of pollut- mentioned genetic diseases will be covered elsewhere
ants, such as organic compounds, the data are much less in this handbook (apart from some aspects of hemo-
conclusive. A complex metabolism and limited knowl- chromatosis, which will be mentioned here in rela-
edge of the mechanisms of actions make it complicated tion to lead and manganese): Chapter 35 (Wilson and
to sort out how the genetic background contributes to Menkes diseases) and Chapter 41 (hemochromatosis).
differences in toxicokinetics and toxicodynamics. In contrast, genetic variants of susceptibility genes
Whereas there are clear examples of gene-metal generally have a low penetrance, but are more fre-
interactions at high levels of exposure, data are less quent. When genetic variants are present in above 1%
conclusive at low exposure levels. In the future, with of the population they are called polymorphisms. The
larger studies with enough power to tackle issues of most common type of polymorphisms in the genome
gene-environment interactions at low levels of expo- single is nucleotide polymorphisms (SNPs), which
sure, this picture will probably change. occur about every 500 base pairs (bp). Another type
is copy number variations (CNVs), in which the num-
ber of copies of DNA sequence strings > 1 kbp vary in
1.1 Genes of Concern
size between individuals: two of the first CNVs to be
Some principal functions of genes that influence described are situated in the glutathione S-transferase
metal toxicokinetics and/or toxicodynamics can be genes GSTM1 and GSTT1, where the copy number
identified. For toxicokinetics, genes that affect the range is 0-2.
uptake are important, as suggested for lead and the Two to three times higher risk or less for different
hemochromatosis HFE gene; genes involved in trans- outcomes (Engel et al., 2002; Hashibe et al., 2003) is
port, as shown for inorganic mercury and the gluta- usually found for most susceptibility genetic variants;
thione (GSH)-based ABCC2 transporter; and genes thus, there is much greater scope for interaction with
involved in metabolism as exemplified for arsenic specific exposures. However, although the individ-
and genetic variants of the methyltransferase AS3MT. ual risk for carrying a susceptibility variant might be
Genes that modify the toxic effects of different metals increased by a small amount, the presence of a com-
may affect the binding of proteins, such as for lead and mon polymorphism within the population may result
the major heme synthesis pathway-related gene ALAD; in a substantial proportion of toxic effects attributable
or be involved in defense against oxidative stress and to this type of genetic trait.
DNA repair capacity. Furthermore, since many metals
provoke hypersensitivity reactions, e.g. gold, mercury,
1.2 Epigenetics of Metals
beryllium, and nickel, the major histocompatibility
complex (also called HLA) genes involved in present- Another type of inheritance that might be important
ing metals to the immune system are influential, as has for individual susceptibility to metal toxicity is medi-
been shown for beryllium and HLA-DPB1, as well as ated by epigenetic factors. These are heritable changes
other parts of the immune system. in gene expression and regulation that are not transmit-
Metal susceptibility genes differ in several aspects ted by the DNA sequence itself, but by different modi-
from the so-called high penetrance genes that have fications of the DNA nucleotides or DNA-­associated
been identified for some rare metal-associated heredi- proteins—such as histones. So far, little is known about
tary diseases, e.g. Wilson or Menkes diseases, char- the impact of epigenetic factors on predisposition to
acterized by impaired copper homeostasis. Typically, metal toxicity, but there is increasing evidence that
high penetrance gene mutations associated with dis- metals actually affect the epigenetic machinery.
ease are rare, i.e. they have a frequency lower than 1% DNA methylation, i.e. specific methylation of the
in the population. For example, the gene frequency for cytosine directly upstream of a guanine, is essential for
Wilson disease varies between 0.3-0.7% worldwide embryogenesis and for the maintenance of cell lineage-
(Gollan and Gollan, 1998) and is the result of mutations specific gene expression throughout life. A widespread
in the ATP7B gene (Bull et al., 1993; Tanzi et al., 1993). notion is that the pre- and postnatal environments are
This type of mutation is highly penetrant, meaning that important determinants of disease susceptibility in
it confers a high absolute risk, irrespective of environ- later life (Gluckman and Hanson, 2006). This influence
mental factors. Hereditary hemochromatosis is a more is thought to be mediated mainly through alterations
common (frequency of approximately 5% in the Cau- to DNA methylation, with subsequent dysregulation
casian population) genetic disease (Cartwright et al., of gene expression that in turn alters the epigenetic
 12 Gene-Environment Interactions for Metals 241

programming of the child (Barker, 1992; Waterland and metal, or the toxicodynamics, or both. Today, many
Michels, 2007). One identified important epigenetic gene-environment studies are based on the genotype
modifier is inadequate nutrition during fetal develop- for information of the phenotype, i.e. characteristics
ment, resulting in low birth weight. However, exposure of an individual. However, the genotype-phenotype
to other environmental factors, such as metals, is also association is not established for most polymorphisms.
likely to affect the epigenome (Baccarelli and Bollati, It is probable that many genetic variants result in no
2009; Edwards and Myers, 2007; Perera and Herbstman, effect or a very limited effect on the phenotype. When
2011). In carcinogenesis, global DNA methylation is no phenotypic information is present, the selection
generally decreased, while specific regions of DNA of polymorphisms in candidate genes should be con-
containing tumor suppressor genes often show hyper- ducted based on position within the gene sequence
methylation (Deng et al., 2006; Gaudet et al., 2003; (e.g. whether or not it results in an amino acid
Robertson and Jones, 2000). exchange or might regulate expression of the protein);
Several metals are not strong mutagens and the alternatively, whether they can tag information about
mechanisms of their carcinogenicity are far from clear. variation in other polymorphisms that have a func-
Suggested mechanisms of action are interference with tional effect. An alternative approach is to explore the
the cellular response to DNA damage, dysregulation of genome without any prior hypothesis to identify genes
cell growth, and resistance to apoptosis. Recent stud- or polymorphisms associated with a phenotype. This
ies indicate that metals such as arsenic and cadmium approach is called a genome-wide association study
interfere with epigenetic processes, such as DNA meth- (GWAS), in which a very large number of SNPs and
ylation: both gene-specific DNA hypermethylation CNVs are simultaneously analyzed in each sample.
and global DNA hypomethylation are found in vitro The corresponding studies for epigenetic markers are
(Benbrahim-Tallaa et al., 2007; Chen et al., 2004), but called epigenome-wide association studies (EWAS). These
in vivo data are still limited. Gene expression is fur- studies have enormous potential to identify genetic
ther regulated by microRNAs. MicroRNAs are small, risk markers; however, interpretation of GWAS and
endogenous, single-stranded noncoding RNAs (20- EWAS data is hampered by the large number of mul-
22 nucleotides in length). They posttranscriptionally tiple comparisons that can easily lead to many false
regulate gene expression by hybridizing to messen- positive results. Thus, validation studies in other pop-
ger RNAs (mRNAs), which leads to translational ulations, animal models, or experimental systems are
repression or degradation of the target RNA (He and often required. So far, there are only a few GWAS or
Hannon, 2004). MicroRNAs are more stable than EWAS in relation to toxic metals in the literature.
mRNAs in blood; thus, they can be studied more easily. It is important to note that the geographical dis-
Another mechanisms of the regulation of gene expres- tribution of polymorphisms varies worldwide; thus,
sion is through histone modifications that contribute gene-environment interactions can differ between eth-
to either a more relaxed “open” chromatin structure nic groups. For example, the frequency of the F allele
permissive for gene expression, or to a more “closed” of the Fok1 (rs10735810, reference ID in the database for
structure (Yan and Boyd, 2006). However, apart from a SNPs) (NCBI, 2006) polymorphism of the vitamin D
few recent studies, little is known about the influence of receptor (VDR) was much higher among the African-
metals on these types of epigenetic processes. The epi- American children than the non-African-­ American
genetic effects of metals are also discussed in Chapter 9. children in a study into VDR polymorphisms and lead
(Haynes et al., 2003). The interaction between the VDR
genotype and floor-dust lead was not significant for
1.3 General Aspects of Gene-Environment African-American children, but was significant for
Interaction Studies: Example for Metals non-African-American children. In general, genetic
There are many aspects of designing a gene-­ effects are more easily observed in homogenous rather
environment interaction study, and a few are men- than mixed populations, and ethnic origin should,
tioned below. For a more comprehensive overview, the when possible, be taken into consideration.
review by Vineis and Malats (1999) is recommended.
1.3.2 Analysis and Interpretation
1.3.1 Choice of Polymorphisms and Method of
Genetic factors for metal toxicity are most often
Detection
handled as effect modifiers of exposure-effect rela-
For the selection of candidate genes, it is of course tionships. The interpretation of genotype as an effect
important to consider whether the study addresses modifier means that studies must be planned so that
genetic variants that modify the toxicokinetics of a they have sufficient power for detecting interactions
242 Karin Broberg, Karin Engström, and Shegufta Ameer

and that an interaction term should be modeled into the model that were not seen when analyzed individu-
the statistical analysis. However, it is important to note ally. Obviously, the complexity and different types of
that an interaction with one genetic variant does not mechanisms behind the interactions, as well as the lack
warrant conclusions of a causal relationship with the of a gold standard for performing statistical analysis
biochemical mechanism because the genetic variant for interaction studies makes the study and interpre-
may be linked with the causal gene. Another fact that tation of gene-environment interactions very difficult.
stresses the importance of having sufficient statistical Another point to consider is the dose-response rela-
power is that, in general, genes act in concert to exert tionship. Two modalities of dose-response relationship
their effects. Therefore, effects on one gene may induce have been suggested when taking genetic polymor-
compensatory changes in others. Moreover, many phisms into account: direct and inverse (Vineis and
metabolic enzymes have overlapping substrate speci- Malats, 1999). In the direct way, minor differences in
ficities. Thus, there are several grounds for analyzing levels of protein or function do not significantly affect
interactions between different gene loci, so-called gene- the phenotype when exposure levels are low or moder-
gene interactions. Only a limited number of studies ate, but become crucial at higher doses of exposure. In
have until now examined the combined effects of two the inverse way, the opposite occurs; this is often inter-
or more genes in relation to metal toxicity. One exam- preted as an effect of saturation of the protein, e.g. high
ple is beryllium, where the combined effect of specific dose exposure saturates the enzyme activity both in
variants of the major histocompatibility complex genes individuals with slow and in those with rapid enzyme
HLA-DQB1 and HLA-DQR1 and tumor necrosis factor activity. Accordingly, sometimes greater susceptibility
(TNF) have been analyzed in relation to clinical sever- of those with slow activity is seen only upon low dose
ity (Maier et al., 2001). Considering multiple genes exposures. The phenomenon of inverse modality is not
is more biologically relevant, but for these studies a often observed, but it has been suggested for metals:
large number of individuals are needed. Moreover, it the effect of inorganic mercury levels was associated
is necessary to formulate a strategy for handling the with a polymorphism affecting the GSH status among
problem of multiple comparisons that quickly arises low-level exposed individuals (Custodio et al., 2005). A
when analyzing the effects of a large number of poly- similar but less prominent trend was observed among
morphisms in different exposure groups, in different highly exposed subjects.
subgroups, or for different markers of effect. Simple
corrections for multiple testing, e.g. Bonferroni cor-
rection, are not very fruitful when analyzing several 2 INTERACTIONS OF SPECIFIC METALS
hundreds to millions of polymorphisms or SNPs that
are in linkage disequilibrium (LD); rather, it has been There follows a review of what presently is known
suggested that the prior credibility of the hypothesis is about gene-environment interactions and epigenetics
important for interpreting a set of observations. New for different metals. For most metals, apart from a few
statistical methods for genetic association studies are exceptions (i.e. for beryllium, lead, and arsenic), there
needed and a Bayesian approach can be used in order are none or only a few reports on gene-environment
to estimate the probability that a positive association is interactions. However, the literature in this field is rap-
false (Wacholder et al., 2004). idly growing.
There is some confusion over the appropriate
measure for assessing the absence or presence of an
2.1 Arsenic
interaction effect. The primary reason is that the sta-
tistical methods used for binary outcomes (e.g. ill ver- Gene-environment interaction studies have mainly
sus healthy) generally assume a multiplicative effect focused on genes involved in the toxicokinetics of arse-
between the two factors being analyzed. However, if nic in populations exposed to arsenic from contami-
a biological understanding of the disease process is nated drinking water. Inorganic arsenic is metabolized
put forward, the joint effect of two independent risk in the body through a series of reduction and methyla-
factors can be expected to be the sum of their indi- tion reactions, producing methylarsonic acid (MMA)
vidual effects if there is no interaction between them and dimethylarsinic acid (DMA), which are excreted in
(Rothman and Greenland, 1998). Simple comparisons the urine (Vahter, 2002). In humans, efficient methyla-
in subgroups can be very misleading or confounding. tion from inorganic arsenic to DMA is associated with
In many studies, gene-environment analysis is instead a high rate of arsenic excretion in the urine. However,
performed by multivariate regression models with a there is a marked variability in metabolism of inor-
large number of covariates; these might lead to signifi- ganic arsenic within and between population groups
cant effects for genotypes or interaction terms within (Vahter, 2002), suggesting a genetic influence.
 12 Gene-Environment Interactions for Metals 243

Arsenic is strongly associated with cancers of the this polymorphism are rather low around the world,
lung and bladder (IARC, 2004), as well as vascular in general 10% or lower (Fujihara et al., 2007, 2008).
diseases, hepatotoxicity, diabetes, and chronic cough Drobna et al. (2004) showed that the Thr allele has
(Mazumder et al., 2005; NRC, 2001; Rahman et al., a lower methylation capacity in Hep2 cells. Wood
2007). Incomplete arsenic metabolism, with higher et al. (2006) evaluated the potential impact of poly-
urinary inorganic arsenic and MMA, but lower DMA, morphisms on arsenite methyltransferase levels
seems to be a marker of increased susceptibility to in vitro by transient expression of the common
arsenic-related diseases, including cancer (Chung AS3MT Met287 sequence and the variant Thr287
et al., 2009b; Lindberg et al., 2008b). allozyme were evaluated in COS-1 cells that have
no endogenous arsenite methyltransferase activity.
The 287Thr allozyme, associated with high MMA
2.1.1 Arsenic (+3 Oxidation State) Methyltransferase
and low DMA levels, had higher in vitro protein
The most studied gene in relation to arsenic toxicoki- levels compared to cells transfected with the AS3MT
netics is arsenic (+3 oxidation state) methyltransferase Met287 variant.
(AS3MT), which encodes arsenite methyltransferase, The link between AS3MT genotype and adverse
an enzyme that catalyzes the methylation of arsenic to health effects has been less well studied. The Thr
MMA, and further to DMA (Lin et al., 2002; Thomas allele of the Met287Thr polymorphism was associated
et al., 2007; Waters et al., 2004). The importance of with higher levels of genetic damage among Mexican
AS3MT to arsenic methylation has been shown in children (Sampayo-Reyes et al., 2010), a marginally
in vitro studies: human urothelial (UROtsa) cells that increased risk of premalignant skin lesions in a Mexi-
do not express AS3MT are unable to methylate arsenic, can population (Valenzuela et al., 2009), and an ele-
while clonal UROtsa cells that express rat AS3MT can vated risk of bladder cancer in a population from the
methylate arsenic (Drobna et al., 2005). Furthermore, USA (Beebe-Dimmer et al., 2012). There was no effect
AS3MT knockdown in human hepatocellular carci- of this polymorphism on arsenic-induced skin lesions
noma (HepG2) cells by short hairpin RNA (shRNA) among study subjects from West Bengal (De Chaud-
resulted in a large decrease in the capacity to methyl- huri et al., 2008). The association between noncoding
ate arsenic. polymorphisms and arsenic-related health effects has
Several noncoding polymorphisms in AS3MT also been evaluated in some studies. In a study on
have been associated with changes in arsenic metab- arsenic exposure from drinking water and telomere
olite pattern. The multitude of influential polymor- length, a genetic marker often altered in tumors, uri-
phisms is likely to be due to the strong LD for AS3MT nary arsenic was associated with longer telomeres (Li
polymorphisms in several populations around the et al., 2012), and individuals with noncoding AS3MT
world (Engström et al., 2011a; Fujihara et al., 2009; alleles (associated with a slower and more toxic metab-
Gomez-Rubio et al., 2010; Wood et al., 2006). An influ- olism, i.e. lower DMA and higher MMA) showed the
ence on the arsenic metabolite pattern by these non- strongest effect of arsenic on telomere length. AS3MT
coding polymorphisms and in the same direction for rs1078835 was found to be associated with heart dis-
the different alleles have been seen in several popu- ease and hyperlipidemia among individuals with low-
lations, e.g. from Argentina (Engström et al., 2011a), level arsenic exposure in Texas (Gong and O’Bryant,
Bangladesh (Engström et al., 2011a), Chile (Hernandez 2012). However, the results of other studies have been
et al., 2008), Mexico (Gomez-Rubio et al., 2010; Meza negative: noncoding AS3MT polymorphisms did not
et al., 2005), Taiwan (Chung et al., 2009a), and ­Vietnam affect the risk of skin lesions in a Mexican population
(Agusa et al., 2010). Studies indicate that some of (Valenzuela et al., 2009); nor did they affect the risk
these polymorphisms, despite being noncoding, have of cancer (all types of cancer combined) in a popula-
a functional impact by affecting gene expression tion from Taiwan (Chung et al., 2009a) or the risk of
(Engström et al., 2011a; Engström et al., 2013a). arsenic-related bladder cancer in a population from
In addition, one coding polymorphism (rs11191439, the USA (Lesseur et al., 2012).
Met287Thr), not part of the LD block, was associ- In some populations, the AS3MT LD block has
ated with altered arsenic metabolite pattern in sev- been shown to stretch into neighboring genes,
eral populations, from Chile (Hernandez et al., 2008), such as cyclin M2 (CNNM2; 3ʹ of AS3MT) and cyto-
eastern Europe (Lindberg et al., 2007), Bangladesh chrome P450, family 17, subfamily A, polypeptide 1
(Engström et al., 2011; Rodrigues et al., 2012), and (CYP17A1, 5ʹ of AS3MT; encoding steroid 17-alpha-
Mexico (Valenzuela et al., 2009), in which the Thr hydroxylase/17,20 lyase/cytochrome P45017A1)
allele was associated with higher percentage MMA (Gomez-Rubio et al., 2010; Pierce et al., 2012; Engström
and/or lower percentage DMA. The frequencies of et al., 2009). Cytochrome P45017A1 is a key enzyme in
244 Karin Broberg, Karin Engström, and Shegufta Ameer

the steroidogenic pathway, and sex steroids may play 2.1.3 Glutathione-Related Genes
a role in arsenic metabolism (Lindberg et al., 2008a).
The metabolism of arsenic involves, apart from
CNNM2 is a magnesium transporter that is widely
methylation, reduction reactions in which arsenic
expressed throughout the body. Polymorphisms in
is reduced from pentavalent to trivalent species.
these genes have been associated with an altered
This may be dependent on glutaredoxin-1 (encoded
arsenic metabolite pattern, possibly due to their LD
by GLRX) and thioredoxin (encoded by TXN) pro-
with AS3MT, or alternatively because there are sev-
tein complexes; the importance of these systems has
eral genes in the chromosomal region important for
been shown in animal models for arsenic methyla-
arsenic metabolism. A recent GWAS from Bangladesh
tion (Mukhopadhyay and Rosen, 2002; Waters et al.,
found that two polymorphisms near AS3MT modi-
2004). Peroxiredoxin-2 (encoded by PRDX2) converts
fied the risk of arsenic-related skin lesions (Pierce
thioredoxin to TXN disulfide via H2O2 (Wang et al.,
et al., 2012).
2007). Schläwicke Engström found that SNPs in GLRX
Some populations, such as indigenous people in
and PRDX2 were associated with an altered arse-
the South American Andes, have probably lived with
nic metabolite pattern in an Argentinean population
arsenic-contaminated drinking water for thousands
(Engström et al., 2009).
of years. Inhabitants of a village in the Argentinean
Functional members of the omega class glutathi-
highlands generally carry an AS3MT haplotype asso-
one S-transferases (GSTs; encoded by GSTO1 and
ciated with a lower urinary fraction of MMA. Schle-
GSTO2) have been proposed to play a role in the
busch et al. (2013) suggest that over the course of a
reduction of arsenic (Schmuck et al., 2005; Steinmaus
few thousand years, natural selection for tolerance
et al., 2005; Zakharyan et al., 2001). A first study by
to arsenic has increased the frequency of protective
Marnell et al. (2003) showed that two Mexican individ-
variants of AS3MT in the population living in this
uals heterozygous for two rare nonsynonymous amino
village because they found a significantly higher fre-
acid exchanges in GSTO1, Glu155del and Glu208Lys,
quency of the protective AS3MT haplotype in this
displayed significantly higher percentage inorganic
population compared with other indigenous groups
arsenic in the urine. Thereafter, results are conflicting
from South America, which presumably have a lower
regarding associations with arsenic metabolism: some
historical exposure to arsenic.
studies found associations between GSTO1 polymor-
phisms and the arsenic metabolite pattern (Agusa
2.1.2 Genes Involved in the One-Carbon Metabolism
et al., 2010; Chung et al., 2011; Rodrigues et al., 2012),
S-adenosyl-methionine (SAM), which is generated while most studies did not (De Chaudhuri et al., 2008;
via the one-carbon metabolism, is the main methyl Lindberg et al., 2007; Meza et al., 2005; Paiva et al.,
donor in methylation reactions, including the meth- 2008; Engström et al., 2007; Xu et al., 2009). Results are
ylation of arsenic. MTHFR (encoding methylene tet- also conflicting for GSTO2: some studies found asso-
rahydrofolatereductase) is the most studied gene in ciations between GSTO2 polymorphisms and arsenic
the one-carbon metabolism. The MTHFR 677 T-allele metabolite pattern (Chung et al., 2009a; Paiva et al.,
is associated with reduced enzyme activity and the 2010), while others did not (De Chaudhuri et al., 2008;
accumulation of plasma homocysteine (Frosst et al., Xu et al., 2009).
1995), leading to lower levels of SAM. In accordance, There are associations between GSTO genotype and
the T-allele was associated with a less efficient arse- the toxic effects of arsenic: GSTO1 has been weakly
nic methylation, i.e. higher MMA and/or lower DMA associated with metabolic syndrome (Chen et al., 2012)
(Chung et al., 2010; Lindberg et al., 2007; Engström and urothelial carcinoma (Chung et al., 2013), whereas
et al., 2007; Steinmaus et al., 2007). Another MTHFR GSTO2 has been associated with a risk of bladder
polymorphism, rs1476413, was associated with risk of cancer (Lesseur et al., 2012) and urothelial carcinoma
bladder cancer in a population from the USA exposed (Chung et al., 2011). GSTO1/GSTO2 combination hap-
to low to moderate (< 50 μg/L) levels of arsenic in their lotypes have been linked to carotid atherosclerosis
drinking water (Beebe-Dimmer et al., 2012). Two stud- (Hsieh et al., 2011) and urothelial carcinoma (Hsu et al.,
ies of Argentinean populations found associations 2011; Wang et al., 2009).
between other genes of the one-carbon metabolism, Further, other GSTs (encoded by GSTM1, GSTT1,
i.e. cystathionine-beta-synthase (CBS), 5-methyltetra- and GSTP1) have been analyzed in relation to the
hydrofolate-homocysteine methyltransferasereduc- arsenic methylation pattern. These enzymes conju-
tase (MTRR), and choline dehydrogenase (CHDH), gate a number of substrates to GSH, and differences
and the arsenic metabolite pattern (Porter et al., 2010; in metabolism and disease susceptibility between the
Engström et al., 2009). allelic forms have been seen for compounds, including
 12 Gene-Environment Interactions for Metals 245

those of arsenic (Agusa et al., 2010; Breton et al., 2007a; associated with higher 8-oxodG levels in Bangladesh.
Chiou et al., 1997; Hsu et al., 2011; Lesseur et al., 2012; The reason for this inconsistency could be differences
Marcos et al., 2006; McCarty et al., 2007; Engström in the LD pattern with an unknown functional poly-
et al., 2007). morphism. However, another explanation could be the
large difference in urinary arsenic levels between the
2.1.4 Genes Involved in DNA Repair and Defense two studies (mean urinary arsenic was 39 μg/g creati-
Against Oxidative Stress nine in Bretons’ study compared to 380 μg/g creatinine
in Engströms’ study). Breton et al. (2007b) investigated
It is generally thought that many of the toxic effects
an association between Asp148Glu and the risk of skin
associated with arsenic exposure are mediated by
lesions among arsenic-exposed individuals in Bangla-
arsenic-induced oxidative stress. Mechanisms for pro-
desh and found that the Glu allele was associated with
tection against oxidative stress in the body involve
a twofold increased risk of skin lesions.
antioxidant defense and DNA repair when the level of
oxidative stress is sufficient to damage DNA.
2.1.5 Other Genes
Among antioxidant genes, the myeloperoxidase
(MPO) 463G and catalase (CAT) 262C>T polymor- HemK methyltransferase family member 2 [encoded
phisms were associated with skin lesions in a highly by N-6-adenine-specific DNA methyltransferase 1
arsenic-exposed population in Bangladesh (Ahsan (N6AMT1)] can methylate arsenic in vitro (Ren et al.,
et al., 2003a), while polymorphisms in superoxide dis- 2011). Harari et al. (Harari et al., 2013) investigated
mutase 2 (SOD2) were associated with arsenic-related the association between N6AMT1 and the arsenic
hypertension in a Taiwanese population (Hsueh et al., metabolite pattern and found that five polymorphisms
2005). Variants in genes encoding base excision repair (rs1997605, rs2205449, rs2705671, rs16983411, and
enzymes, such as X-ray repair complementing defec- rs1048546) and two haplotypes were significantly
tive repair in Chinese hamster cells (XRCC) 1 and 3, associated with percentage MMA in urine, even after
were associated with skin lesions and different sorts taking the AS3MT haplotype into account.
of cancers in arsenic-exposed populations (Andrew Karagas et al. (2012) screened a subset of bladder
et al., 2009; Breton et al., 2007b; Hsu et al., 2008; Kundu cancer cases in a U.S. population with low to moder-
et al., 2011; Thirumaran et al., 2006). For XRCC3, all of ate drinking water levels of arsenic using a panel of
the studies mentioned above evaluated the effect of approximately 10,000 nonsynonymous SNPs. The
Thr241Met. The presence of the Met allele was associ- top hits were further analyzed in a population-based
ated with a lower cancer risk in two studies (Kundu case-control study (n = 832 cases and 1191 controls).
et al., 2011; Thirumaran et al., 2006), while Andrew Polymorphisms in fibrous sheath-interacting protein 1
et al. (2009) found a higher risk for individuals with (encoded by FSIP1) and zinc transporter ZIP2 [encoded
the Met allele with higher arsenic exposure. Polymor- by solute carrier family 39, member 2 (SLC39A2)] were
phisms in the nucleotide excision repair gene, excision associated with a risk of arsenic-related bladder cancer.
repair cross-complementing rodent repair deficiency FSIP1 is a direct target of nuclear receptor coactivator 3
complementation group 2 (ERCC2), were associated (NCoA-3/SRC-3, encoded by NCOA3/A1B1), a onco-
with nonmelanoma skin cancer (Applebaum et al., protein known to be associated with breast cancer and
2007) and skin lesions (Ahsan et al., 2003b; Banerjee a coactivator for nuclear receptors. ZIP2 is a member of
et al., 2007) among arsenic-exposed populations in the the ZIP family of metal ion transporters.
USA and Bangladesh. Genes involved in cell cycle regulatory genes have
Arsenic exposure has been related to urinary been associated with a risk of arsenic-related cancers.
8-oxodG (a marker of oxidative DNA damage) levels The Arg72Pro polymorphism in TP53 is associated with
in several studies (Chung et al., 2008b; Fujino et al., renal cell carcinoma risk (ProPro is associated with a
2005; Kubota et al., 2006; Wong et al., 2005; Xu et al., higher risk) (Huang et al., 2011), keratosis (ArgArg is
2008), and genes involved in DNA repair might influ- associated with a higher risk) (De Chaudhuri et al.,
ence the amount of 8-oxodG formed. Asp148Glu 2006), and skin cancer (ProPro is associated with a
exchange in apurinic/apyrimidinic endonuclease higher risk) (Chen et al., 2003); a polymorphism in
(APEX1) was associated with levels of 8-oxodG among CDKN1A/P21 is associated with arsenic-related uro-
arsenic-exposed individuals in Argentina (Engström thelial carcinoma risk (Chung et al., 2008a).
et al., 2010) and Bangladesh (Breton et al., 2007a). Arsenic has been linked to an increased prevalence
However, the effects were in different directions: the of cardiovascular disease. Serum paraoxonase/aryles-
AspAsp genotype was associated with lower con- terases (encoded by PON1 and PON2) hydrolyze oxi-
centrations of 8-oxodG in Argentina, while it was dized lipids and are thought to be protective against
246 Karin Broberg, Karin Engström, and Shegufta Ameer

atherosclerosis. Liao et al. (2009) examined the syner- arsenic levels in urine in nondiseased individuals.
gistic interaction of genetic factors and arsenic expo- The latter study also found hypermethylation of the
sure on electrocardiogram abnormality (in order to DNA repair gene, mutL homolog 1 (MLH1) (Hossain
evaluate long-term influence of arsenic on the cardio- et al., 2012b). The hypermethylation of specific genes,
vascular system) in a Taiwanese population (n = 216). including tumor suppressor genes, may reflect the
Subjects exposed to high levels of arsenic and carrying selection of premalignant cells during carcinogenesis,
a specific PON1/PON2 haplotype had a significantly rather than being a direct effect of arsenic DNA meth-
increased risk of electrocardiogram abnormality com- ylation. A study in individuals with arsenicosis identi-
pared to those with only one risk factor. fied a large number of genes that were differentially
Arsenic targets the immune system (Dangleben methylated by arsenic exposure (Smeester et al., 2011),
et al., 2013). Bhattacharjee et al. (2013) found that the however, the study was based on only 16 individuals.
Ala1052Glu polymorphism affecting NACHT, LRR Arsenic crosses the placenta and epigenetic
and PYD domains-containing protein 2 [encoded by changes of arsenic may alter fetal programming dur-
NLR family, pyrin domain containing 2 (NALP2)], an ing embryogenesis. Some studies have analyzed epi-
important component of the inflammasome complex, genetic markers in cord blood samples in relation to
was associated with a decreased risk of the develop- the mother’s arsenic exposure during pregnancy. Kile
ment of skin lesions (n = 219 cases and 213 controls). et al.( 2012) analyzed DNA methylation at CpG sites in
They also observed that individuals with the protec- p16 and TP53, as well as methylation of retrotranspo-
tive genotype had fewer chromosomal aberrations sons (the Alu and LINE1 families), which encompass a
and were less susceptible to arsenic-related respiratory large part of the human genome (25%) (Rollins et al.,
diseases. In another study (Wu et al., 2013), polymor- 2006) and can be used as a surrogate marker of global
phisms affecting the proinflammatory cytokines tumor methylation. Exposure to higher levels of arsenic was
necrosis factor (TNF-α; G-308A) and interleukin-8 (IL- positively associated with DNA methylation in LINE1
8; T-251A) were associated with urothelial carcinoma. elements, and to a lesser degree at CpG sites within
In summary, it is clear that a major part of the dif- the promoter region of the p16 tumor suppressor gene.
ferences in arsenic metabolism efficiency is genetically Associations were observed in both maternal and fetal
determined; there is extensive support that polymor- leukocytes. Pilsner et al. (2012) found that increasing
phisms, in particular noncoding polymorphisms, in levels of maternal urinary arsenic were associated with
AS3MT modify the efficiency of arsenic metabolism. an increase in global DNA methylation in newborns,
However, the phenotype is polygenic, and other genes as measured by the [3H]-methyl incorporation assay.
such as those involved in one-carbon metabolism or Koestler et al. (2013) investigated genome-wide dif-
reduction reactions, or other methyltransferases also ferences in DNA methylation in cord blood samples
play a role in determining the metabolism of arsenic. (n = 134) of infants with low-level arsenic exposure in
The toxicodynamics of arsenic are also polygenic. utero using the Illumina Infinium Methylation 450K
array (450,000 CpG sites). About half of the 100 CpG
loci with the lowest P-values were in CpG islands
2.1.6 Epigenetic Effects of Arsenic
(which indicates that they may be situated near a pro-
Recent studies indicate that a major mechanism of moter site and significantly affect gene expression),
arsenic toxicity is through inducing epigenetic changes, and most of these (75%) showed higher methylation
through alterations in DNA methylation, histone mod- levels in the highest exposure group compared with
ification, and noncoding RNA expression (Cheng et al., the lowest exposure group. However, no association
2012). During its metabolism, arsenic is methylated by between maternal urinary arsenic and specific CpG
SAM, the same methyl donor needed for methylation methylation in cord blood was statistically significant
of DNA (Vahter, 2002). It has therefore been proposed after adjusting for multiple comparisons. Several CpG
that arsenic metabolism may deplete intracellular loci exhibited a linear dose-dependent relationship
methyl groups, resulting in the hypomethylation of between methylation and arsenic exposure; among
genes. Indeed, treatment with arsenic results in global these, loci in the endocrine-related ESR1 (estrogen
DNA hypomethylation in both cell cultures (Zhao receptor 1) and PPARGC1A (peroxisome proliferator-
et al., 1997) and mice (Chen et al., 2004), but in arsenic- activated receptor gamma, coactivator 1 alpha) genes
related skin tumors hypermethylation of the tumor were negatively associated with arsenic. Treas et al.
suppressor genes TP53 and CDKN2A/p16 genes have (2012, 2013) revealed that exposure to arsenic, estrogen,
been reported (Banerjee et al., 2013; Chanda et al., 2006; and arsenic and estrogen combined alters the expres-
Mass and Wang, 1997). Hypermethylation of p16 and sion of epigenetic regulatory genes and changes global
lower p16 gene expression has been linked to higher DNA methylation and histone modification patterns in
 12 Gene-Environment Interactions for Metals 247

human prostate epithelial (RWPE-1) cells. The changes global levels of the H3K4 trimethylation H3K4me3
were greatest in the group treated with the combina- (Zhou et al., 2008), the latter being an open chroma-
tion of arsenic and estrogen. tin structure. Suzuki and Nohara (2013) showed that
Bailey et al. (2013) report differences in DNA meth- long-term arsenic exposure may downregulate p16
ylation profiles in peripheral blood leukocytes from in mice by targeting the recruitment of histone-lysine
patients with prediabetes mellitus or diabetes mellitus N-methyltransferase EHMT2 (G9a; an H3K9 histone
that were related to different arsenic methylation pat- methyltransferase) and increasing H3K9me2 without
terns. Hossain et al. (2012b) found there to be an effect altering DNA methylation prior to liver tumorigenesis.
of AS3MT genotype: carriers of the AS3MT haplotype Chervona et al. (2012) measured histone modifications
associated with a higher percentage MMA and lower in peripheral blood mononuclear cells and showed
percentage DMA had increased p16 methylation. The that urinary arsenic inversely correlated with H3K9ac.
AS3MT haplotype has also been associated with DNA Gender-specific patterns of association were observed
methylation of AS3MT and nearby genes (Engström between arsenic exposure and several histone markers.
et al., 2013a) in two different populations, in which the Li et al. (2012) analyzed microRNA expression pro-
specific AS3MT haplotype associated with lower per- files in human umbilical vein endothelial cells follow-
centage MMA and higher percentage DMA was also ing arsenic exposure. A total of 85 microRNAs were
associated with increased AS3MT methylation. The upregulated and 52 were downregulated by arsenic
AS3MT genotype has also been associated with LINE1 treatment, compared to the control group; phospho-
methylation (Tajuddin et al., 2013) in control subjects proteins and genes involved in alternative splicing
for bladder cancer (n = 892) with low arsenic expo- were among the top categories targeted by both up-
sure (median toenail arsenic, 0.07 μg/g). In this study, and downregulated microRNAs. Marsit et al. (2006)
associations were also seen between LINE1 methyla- evaluated whether arsenic exposure results in changes
tion and polymorphisms in genes mainly related to to microRNA (n = 385) expression in human lympho-
folate and one-carbon metabolism: DNA methyltrans- blastoid cells. Exposure to sodium arsenite led to
ferase 3A (DNMT3A; rs7581217), transcobalamin 2 global increases in microRNA expression. All of the
(TCN2; rs9606756, rs4820887, rs9621049), solute carrier microRNAs for which expression was altered by arse-
family (folate transporter) SLC19A1 (rs914238), and nic exposure were also altered, in the same direction,
5,10-methenyltetrahydrofolate synthetase (MTHFS; by folate deficiency. This data indicates that arsenic-
rs1380642). Folate status may influence the DNA meth- related effects on microRNAs involve the one-carbon
ylation; Lambrou et al. (2012) saw a trend of decreas- metabolism. Sodium arsenite upregulated the expres-
ing LINE1 and increasing Alu DNA methylation with sion of four microRNAs (hsa-miR-222, hsa-miR-22, hsa-
increasing arsenic exposure among elderly men envi- miR-34a, hsa-miR-221) and strongly downregulated
ronmentally exposed to low levels of arsenic, and these the expression of a single microRNA (hsa-miR-210).
associations varied according to folate status. Hsa-miR-222 plays an important role in human eryth-
Cantone et al. (2011) investigated whether arsenic ropoiesis, leukemic cell differentiation, and the devel-
in ambient and occupational air particulate matter opment of B-cell leukemia. Marsit et al. (2006) did
induced histone modifications in blood leukocytes not observe any significant alteration in global DNA
among steel workers (n = 63). Histone-3/lysine-4 methylation, as measured in long repetitive elements
dimethylation (H3K4me2) increased significantly (LINE1). Kong et al. (2012) investigated an association
in association with air levels of arsenic, while the between urinary arsenic and microRNA among school
increase in histone-3/lysine-9 acetylation (H3K9ac) children in Hong Kong and found that hsa-miR-21 and
was not significant. Both H3K4me2 and H3K9ac are hsa-miR-221 negatively correlated with arsenic levels.
activating histone modifications that contribute to the To summarize, an increasing number of publica-
formation of an open chromatin structure permissive tions are revealing clear effects of arsenic on different
for gene transcription. Cumulative exposures to arse- parts of the epigenetic machinery. These effects are
nic, defined as the product of years of employment probably components of arsenic toxicity but the exact
by metal air levels, positively correlated with both mechanisms controlling the interaction of arsenic and
H3K4me2 and H3K9ac. Functional studies conducted epigenetic pathways are unclear.
in human lung carcinoma A549 cells showed that
exposure to inorganic trivalent arsenic increased H3K9
2.2 Beryllium
dimethylation (H3K9me2) and decreased H3K27 tri-
methylation (H3K27me3), both of which are modifica- Genetic susceptibility to beryllium-associated
tions associated with a closed chromatin structure that sensitization and subsequent chronic beryllium
is nonpermissive for gene transcription, and increased disease (CBD) may be the clearest examples of a
248 Karin Broberg, Karin Engström, and Shegufta Ameer

gene-environment interaction in the toxicodynamics have suggested that the 308 allele in the promoter
of a toxic metal. The immunopathogenesis of CBD and region of TNF determines susceptibility to CBD (Dotti
beryllium sensitization depends on the development et al., 2004; Gaede et al., 2005; Maier et al., 2001). The
of an antigen-specific, cell-mediated immune response anti-inflammatory cytokine transforming growth fac-
(Saltini et al., 1989). Beryllium-specific CD4+ T cells tor beta (TGF-β) can inhibit the production of alveolar
probably recognize a form of beryllium as an antigen macrophages and monocytes, and it was recently dem-
and act in combination with HLA class II molecules onstrated that the frequency of a polymorphism asso-
on antigen-presenting cells. The hypothesized mecha- ciated with a low TGF-β release (labeled TGFβ1, codon
nism is that the beryllium cation may directly or indi- 25) was strongly associated with CBD (Gaede et al.,
rectly influence peptide binding to the HLA molecules 2005). However, this phenomenon was only seen in
(Amicosante et al., 2009), possibly by changing the individuals from the USA, which suggests that differ-
local electrostatic environment in the peptide-binding ent genetic factors are important in different popula-
groove of HLA class molecules. tions. Another study focused on susceptibility to CBD
In a key study, Richeldi et al. (1993) reported an caused by genetic variation in interleukins, which are
increased prevalence of HLA-DPB1 with glutamic key players in different immune processes (McCanlies
acid in position 69 (Glu69) in patients with CBD et al., 2010). They found three linked polymorphisms
(97%) compared with beryllium-exposed nondis- in interleukin 1, alpha (IL1A; rs1609682, rs3783539,
eased controls (30%) (Richeldi et al., 1993). These and rs3783543), that were significantly associated with
results were later confirmed in further studies, CBD, compared to beryllium-sensitized and control
although the frequency of the Glu69 variant was subjects.
somewhat lower (Maier et al., 2003; Rossman et al.,
2002; Saltini et al., 2001; Wang et al., 1999). These 2.3 Cadmium
studies show that specific Glu69-containing alleles
2.3.1 Genetic Influences on Cadmium Toxicokinetics
confer the greatest susceptibility to CBD in exposed
and Toxicodynamics
individuals in an allele-dosage manner (homo-
zygous or heterozygous). However, as 15-20% of Support for genetic influences on cadmium sus-
CBD patients lack Glu69, other class II markers are ceptibility comes from a number of studies performed
also likely to be involved in the beryllium-specific in different inbred strains of rodents. As an example,
response. Rosenman and coworkers did an exten- over 30 mice strains were sensitive to cadmium-
sive typing of DRB1 and DPB1 genes of the HLA induced testicular toxicity at very low cadmium levels,
class II molecules and confirmed the role of Glu69 in whereas around 10 were resistant, even at very high
conferring increased susceptibility to CBD or beryl- cadmium concentrations (Liu et al., 2001). The resis-
lium sensitization (Rosenman et al., 2011). They also tance to cadmium-induced testicular injury in certain
identified certain Glu69 alleles (DPB1*0201 negative strains is most probably due to genetic differences in
alleles) and genetic variation in the HLA-DRB1 chain, Slc39a8 (encoding zinc transporter ZIP8), located on
in particular DRBE71, that also appear to be impor- mouse chromosome 3 (Dalton et al., 2000, 2005). The
tant for disease susceptibility and sensitization. The cadmium-binding protein metallothionein gene has
total negative charge contributed by specific poly- generally been thought to be the major gene respon-
morphisms within DPB1 was, in conjunction with sible for strain susceptibility to cadmium toxicity.
Glu69, associated with CBD. The susceptibility vari- Metallothionein-null mice were indeed more sensitive
ants appear to be functional, in the sense that they to certain effects of cadmium, including nephrotoxicity
affect how peptides are presented to T cells. and testicular injuries (Liu et al., 2000).
Sato and coworkers (2007) analyzed polymorphisms Nevertheless, although cadmium is one of the
in the butyrophilin-like 2 (BTNL2) gene, situated close major toxic metals and is widespread in the environ-
to the HLA class II and III genes and previously associ- ment, little is so far known about the extent to which
ated with sarcoidosis. They found indications that this genetic factors influence its kinetics or toxicity in
gene may be a susceptibility marker for CBD. How- humans. One twin study was performed to quantify
ever, they could not distinguish whether this was due the genetic influence on variations in the blood con-
to linkage to DRB1 alleles, given the small number of centrations of cadmium and lead (Björkman et al.,
Glu69-negative CBD cases in their study. 2000). Heredity had a substantial impact on blood
Beryllium antigen stimulates TNF-α release from cadmium and lead levels, but a striking sex difference
bronchoalveolar lavage cells in CBD. Therefore, func- was also observed. Among nonsmoking women, the
tional polymorphisms in the encoding gene are sus- hereditary impact was 65%, whereas for nonsmoking
pected to modify the course of CBD. Several studies men it was only 13%.
 12 Gene-Environment Interactions for Metals 249

Two of the main metallothioneins to be widely association between these polymorphisms and the dis-
expressed in the body are MT1A and MT2A. Three ease was seen. Cadmium is associated with increased
studies in Turkish populations addressed the role of oxidative stress, as measured by oxidative DNA dam-
the rs28366003 polymorphism, located just upstream age in urine. Ketelslegers and coworkers (2008) found
of the transcription site of MT2A. G-allele carriers that GSTT1 deletion was a predictor of the effect of
showed increased blood concentrations of cadmium cadmium: significantly higher levels of DNA damage
and lead and reduced serum levels of zinc (n = 616 were found in individuals lacking the gene, compared
environmentally exposed subjects) (Kayaalti et al., to individuals that carry the gene.
2011). In addition, in kidney autopsies (n = 114), higher
cadmium levels were found in rs28366003 G-allele car- 2.3.2 Epigenetic Effects of Cadmium
riers (Kayaalti et al., 2010); however, there was only
one individual homozygous for the G-allele, so this In vitro studies have indicated that cadmium may
study needs to be cautiously interpreted. In 95 preg- interfere with epigenetic processes: both gene-spe-
nant women, the same allele was associated with cific DNA hypermethylation with gene silencing and
higher cadmium concentrations in maternal blood and global DNA hypomethylation have been reported
lower cadmium concentrations in placenta, but there (Benbrahim-Tallaa et al., 2007; Jiang et al., 2008). In
was no effect of the maternal genotype on cadmium a cross-sectional study of 202 women with environ-
concentrations in cord blood (Tekin et al., 2012). No mental cadmium exposure from the diet, urinary
effect was found on zinc status, but the genotype cor- cadmium was associated with hypomethylation of
related with higher iron levels in the tissues analyzed. LINE1 retrotransposon sequences, a marker for global
Kita et al. (2006) demonstrated a reduced expression DNA methylation, in peripheral blood (Hossain et al.,
of the rs28366003 G variant in response to Cd and Zn 2012a). DNMT1 genotypes modified this association,
exposure. which shows that genetic variation in DNA methyla-
Small iron stores and reduced erythropoiesis tion-regulating genes modifies the effects of metals on
increase cadmium uptake in the intestines (Åkesson the epigenetic machinery. Kippler et al. (2013) showed
et al., 2000). Therefore, it has been suggested that the that cadmium exposure to the fetus, measured as cad-
same proteins absorb both cadmium and iron, e.g. mium levels in maternal blood during pregnancy,
the membrane transporter, natural resistance-asso- modified genome-wide DNA methylation. Boys were
ciated macrophage protein 2 (NRAMP2; encoded particularly affected, showing increasing methylation
by SLC11A2), which has a high affinity for divalent with increasing cadmium in the mother’s blood. In a
metal ions (Picard et al., 2000), are responsible for the study (n = 81) from China, blood and urine cadmium
increased uptake of cadmium. However, a role for was related to hypermethylation of the renal fibro-
SLC11A2 in regulating cadmium concentrations could genesis genes RASAL1 and KLOTHO in blood (Zhang
not be confirmed, although the iron-related transferrin et al., 2013). In one study (n = 63 workers) into exposure
receptor (TFRC) gene was shown to be influential in to metal-rich particular matter containing cadmium
two different populations (Rentschler et al., 2013). and lead and its effects on microRNAs (Bollati et al.,
In a study on the effects of exposure to cadmium- 2010), miR-146a expression in blood was significantly
polluted rice on the kidney function, the rs11076161 reduced in subjects exposed to cadmium and lead.
G→A status of MT1A influenced the renal toxicity of
Cd: carriers of the AA genotype showed more cad-
2.4 Cobalt
mium-related low molecular weight proteins in urine
compared to those with the GG genotype (Lei et al., Hard metal disease is a pulmonary disorder associ-
2012). No strong modifying effects were found for ated with occupational exposure to inhaled cobalt ions,
MT2A rs28366003 on kidney toxicity. sintered or not to other metals. The pathogenesis of the
One study addressed genetic influences in the cad- disease is suggested to be immunological, since a strong
mium-related itai-itai disease, characterized by frac- association with hard metal disease and HLA-DP con-
tures occurring mainly in elderly multiparous women, taining Glu at position 69 has been reported (Potolicchio
with a form of osteomalacia, osteoporosis, and renal et al., 1997). This is the same HLA variant involved in
dysfunction. Since the gene encoding the estrogen beryllium toxicity (see above). The same group also
receptor (ESR1) appears to regulate bone mineral den- reported that HLA-DP molecules but not HLA-DR
sity and other determinants of osteoporotic fracture molecules bind cobalt (Potolicchio et al., 1999). On the
risk, polymorphisms in ESR1 affecting bone mineral other hand, no influence of Glu69 was observed in rela-
density have been analyzed in women with itai-itai tion to lung function changes in workers from a cobalt-­
disease (Nishio et al., 1999; Sadewa et al., 2002). No producing plant (Verougstraete et al., 2004).
250 Karin Broberg, Karin Engström, and Shegufta Ameer

Cobalt metal associated with tungsten carbide is ALAD2. On the one hand, the effects may simply be
a probable carcinogenic substance (IARC, 2006); the due to a shift in the exposure-response curve to the
mechanistic basis for carcinogenicity is that particles of right, because ALAD2 subjects have higher blood Pb
these substances induce reactive oxygen species (ROS), at the same exposure, e.g. due to a hypothetical stron-
which cause DNA breaks and other DNA lesions (De ger binding of Pb to the ALAD2 protein in the eryth-
Boeck et al., 1998; Van Goethem et al., 1997). Mateuca rocytes or, as mentioned below, through regulating the
et al. (2005) reported that variants of the OGG1 and expression of the ALAD gene, and thus the amount of
XRCC3 DNA repair genes modify the levels of DNA ALAD available to bind Pb. The same mechanism may
damage biomarkers (micronuclei and DNA comet occur in all cells, i.e. ALAD2 might sequestrate Pb bet-
tail length) in individuals exposed to cobalt-tungsten ter than ALAD1, making Pb less bioavailable and thus
carbide. less able to exert its toxic effects. Alternatively, ALAD2
No genetic determinants for susceptibility to cobalt carriers actually have less toxic effects on, for example,
allergy have been identified (Thyssen et al., 2008). heme synthesis due to differences in inhibition of the
enzyme activity.
In early studies into ALAD modification of lead
2.5 Lead effects, the individuals analyzed were from popula-
tions with relatively high levels of exposure from
2.5.1 ALAD
occupational or home environments. These studies
It is now almost 25 years since the first study by displayed an association between the ALAD2 geno-
Ziemsen et al. to describe differences in blood lead type and high blood lead levels. Wetmur and col-
levels by genotyping the ALAD gene [encoding delta- leagues demonstrated in a study based on over 1000
aminolevulinic acid dehydratase (ALAD)] (Ziemsen children from New York City that ALAD2 individu-
et al., 1986). Since then, a very large number of studies als had higher blood lead levels (Wetmur et al., 1991,
have focused on interactions among ALAD genotype, 1994). Several other studies have shown an effect of the
lead metabolism, and toxicity (reviewed in Kelada ALAD genotype on blood lead levels among individu-
et al., 2001; Scinicariello et al., 2007; Zhao et al., 2007). als with a high lead exposure (Alexander et al., 1998;
ALAD is the major binding site for lead in red blood Fleming et al., 1998), whereas others have not (Berg-
cells (Bergdahl et al., 1997b, 1998). This protein is poly- dahl et al., 1997b; Sakai et al., 2000; Schwartz et al.,
morphic, and the main variant to be analyzed is the 1995, 1997a; Sithisarankul et al., 1997; Süzen et al.,
Lys59Asn substitution (rs18000435). The two alleles 2003). Epidemiological studies of blood lead levels at
are traditionally labeled ALAD1 (the Lys variant) and background exposure levels show no clear association
ALAD2 (the Asn variant), and the resulting genotypes of effect with ALAD genotype (Bergdahl et al., 1997a;
are named ALAD1-1, ALAD1-2, and ALAD2-2. The Hsieh et al., 2000; Hu et al., 2001; Pawlas et al., 2012;
Lys59Asn amino acid position is not located near to Scinicariello et al., 2010; Smith et al., 1995; van Bemmel
the Zn-binding sites, where lead probably binds (Jaffe et al., 2011b), although in a small study Miyaki found
et al., 2000, 2001). However, since asparagine is a neu- (n = 101) higher blood lead in Japanese ALAD2 carriers
tral amino acid, whereas lysine is positively charged, at low blood lead levels (Miyaki et al., 2009). In a meta-
this amino acid substitution results in a more electro- analysis of 15 articles (Zhao et al., 2007), higher blood
negative enzyme. This fact has led to the hypothesis lead levels were found in ALAD2 carriers.
that the ALAD2 protein binds to positively charged Schwartz and colleagues found the 1-2 genotype
lead more tightly than does the ALAD1 protein, and to be associated with occupational exposures of more
that a protective effect of ALAD2 is caused by this than 6 years (Schwartz et al., 1995). This genotype
tight binding, which maintains lead within the intra- distribution may be the result of genotype selection;
vascular space and cells in a less bioavailable form the authors suggested that ALAD2 subjects were
(Wetmur et al., 1991; Wetmur, 1994). It is important to protected from the effects of lead and could tolerate
stress that there is so far limited evidence for molecu- longer exposures to lead compared to ALAD1-1 indi-
lar differences between the two isozymes. Jaffe et al. viduals. A similar finding was reported by Zheng
(2000, 2001) constructed recombinant ALAD variants et al. (2011): ALAD2 allele frequency was significantly
in Escherichia coli. They found no functional dissimi- higher in lead-exposed workers than in unexposed
larities, apart from a small difference in the half-time controls. However, if selection occurs for tolerance to
for recovery from lead inhibition between ALAD1 lead toxicity, then it must be because symptoms are
and ALAD2 in vitro (Jaffe et al., 2000, 2001). Despite easily recognized by the workers and are associated
a lack of functional data, one can speculate about the with the work. This hypothesis has not yet been prop-
causes of differences in effects between ALAD1 and erly tested.
 12 Gene-Environment Interactions for Metals 251

Kim et al. (2004) analyzed whether Korean lead Bioavailable lead has been analyzed by measur-
workers with the ALAD1-1 genotype are more sus- ing the amount of dimercaptosuccinic acid (DMSA)-
ceptible to the hematological effects of lead exposure. chelatable lead in urine after DMSA administration.
They found ALAD2 carriers to be associated with There is some evidence that ALAD-2 individuals dis-
lower blood zinc protoporphyrin (ZPP), formed as play lower levels of DMSA-chelatable lead (Schwartz
lead inhibits heme synthesis, and higher hemoglobin et al., 1997b, 2000). The ALAD genotype has also been
levels. Moreover, among individuals with normal iron analyzed in relation to the kinetics of bone lead. How-
status, those with the ALAD1-1 genotype were more ever, the results are difficult to interpret: some stud-
likely to be anemic. Alexander et al. (1998) found also ies suggest an effect of ALAD polymorphism on bone
significantly lower levels of ZPP at high blood lead lead (Bellinger et al., 1994; Hu et al., 2001; Kamel et al.,
levels among ALAD2 genotypes, although lower, non- 2003), whereas others do not (Bergdahl et al., 1997a;
significant ZPP levels among carriers of the ALAD2 Fleming et al., 1998; Schwartz et al., 2000). Moreover,
allele have been reported elsewhere (Sakai et al., 2000; a few studies have analyzed the effect modification of
Schwartz et al., 1997a; Sithisarankul et al., 1997). A ALAD on lead-induced kidney parameters (Bergdahl
similar pattern was found in Chinese battery workers et al., 1997a; Smith et al., 1995; Weaver et al., 2003;
(461 exposed and 175 unexposed workers), in which Zheng et al., 2011). The results of these studies are also
1-2/2-2 carriers had a significantly lower increase inconsistent.
in ZPP with increasing lead exposure compared to Lead is classified as a probable human carcinogen.
ALAD1-1 carriers (Zheng et al., 2011). Accumulation In a study on lead-related renal cancer risk (987 cases
of plasma aminolevulinic acid has been suggested to and 1298 controls), the most comprehensive screening
confer a greater risk of lead neurotoxicity. Reduced yet of polymorphisms (n = 19) in ALAD or in the sur-
levels of plasma aminolevulinic acid among ALAD2 rounding chromosomal region was performed (van
carriers have been reported in Korean battery workers Bemmel et al., 2011a). The authors found an increased
(Sakai et al., 2000; Schwartz et al., 1997a; Sithisarankul risk of cancer from lead exposure; moreover, there were
et al., 1997), as well as in Japanese workers (Sakai combined effects of ALAD rs2761016 and lead on can-
et al., 2000). On the other hand, the modifying effects cer risk. Lead-exposed ALAD1-1 carriers had a higher
on the cardiovascular system show another pattern: risk of cancer compared to ALAD2 carriers; however,
in a large study of environmentally exposed subjects the interaction was not statistically significant.
(n = 6016) from the USA, ALAD2 carriers in the highest Most of the results described so far show that ALAD
blood lead level tertile had a higher risk of hyperten- modifies the toxicokinetics and dynamics of lead. At
sion and higher lead-related systolic blood pressure high levels of exposure and in comparison with ALAD-
(Scinicariello et al., 2010). 1 subjects, heterozygous or homozygous ALAD-2 car-
There is a growing literature on the genetic modi- riers seem to have increased blood lead levels, and
fication of lead neurotoxicity by ALAD. A protective perhaps lower amounts of bioavailable lead. Most
effect in ALAD2 carriers on cognitive function has been studies indicate that ALAD2 carriers have some pro-
reported in children (Bellinger et al., 1994) and in adult tection against the toxic effects of lead on heme syn-
lead workers (Gao et al., 2010). No effect of ALAD2 on thesis and the nervous system. However, much is still
intelligence quotient (IQ) was found in lead-exposed unknown regarding potential genotypic effect modi-
Polish children; however, another closely located fications of lead on the kidney, of lead-related cancer,
polymorphism (rs1139488) did have an effect (Pawlas and, moreover, of whether or not the effects of ALAD
et al., 2012). In addition, Krieg et al. (2009) observed genotype depend on the level of lead exposure.
a better reaction time in adult ALAD2 carriers, but no The often inconsistent findings of the impact of
effect in children. Further, Rajan et al. (2007) recorded ALAD on lead levels and toxicity between different
a reduced effect of lead on mood in ALAD2 carriers. studies may be the result of the different methods of
Fang and coworkers (2010) found no clear interaction selecting study subjects, e.g. from highly exposed
between ALAD and lead-related risk of amyotrophic environments, or of the choice of biomarkers of effect.
lateral sclerosis, although a nonsignificantly higher If there is a dose-dependent difference in the effect
risk was reported for ALAD1-1 carriers. ALAD1-1 had of ALAD on lead and its toxicity, it will complicate
worse lead-related effects on the peripheral nervous analysis and interpretation of the data. Another pos-
system, measured as sensory and motor conduction sible reason for contradictory findings is that the effect
velocities, compared to ALAD2 carriers (Zheng et al., of ALAD2 may vary depending on the genetic back-
2011). On the other hand, there are reports of worse ground, leading to problems when different popula-
cognitive performance in elderly male ALAD2 carriers tions are studied. This could originate either from
(Rajan et al., 2008; Weuve et al., 2006). different LD patterns of ALAD polymorphisms in
252 Karin Broberg, Karin Engström, and Shegufta Ameer

different populations or from variations in other genes actually had the lowest lead levels (Garcia-Leston
influencing lead toxicokinetics and dynamics. et al., 2012).
Most studies have only analyzed the ALAD In a study of renal function in Korean lead work-
rs1800435 (ALAD1/2) polymorphism and very little is ers, the BsmI polymorphism of VDR, as well as endo-
known of other influential genetic markers, despite the thelial nitric oxide synthase (NOS3; involved in nitric
fact that ALAD rs1800435 can explain only a small pro- oxide production), were analyzed for possible effect
portion of the variation of lead in blood. Interestingly, modification (Weaver et al., 2003). However, no obvi-
there are some indications that genetic variation in ous association was present between VDR and either
ALAD can influence human health irrespective of lead lead levels or markers of renal function. For the variant
exposure. In the aforementioned study on ALAD, lead, allele of NOS3, on the other hand, a longer period of
and renal cancer, one SNP within the gene (rs8177796) working with lead was associated with higher serum
influenced the risk of cancer independently of lead. creatinine and the kidney effect markers, lower cal-
It is worth mentioning another study of 3349 subjects culated creatinine clearance, and higher N-acetyl-β-
from the USA in which the ALAD2 allele conferred a D-acetylglucosaminidase (NAG). Similarly, VDR did
reduced relative risk for overall mortality (van Bem- not influence either lead concentrations in blood or in
mel et al., 2011b). urine or lead toxicity parameters, including effects on
heme synthesis, nephrotoxicity, and peripheral ner-
vous system function (Zheng et al., 2011). In addition,
2.5.2 VDR
Pawlas et al. (2012) found no significant effects of VDR
Lead uptake increases when calcium resources are polymorphisms on blood lead levels in Polish children.
limited (Mahaffey et al., 1986). Thus, genetic differences However, VDR BsmI b carriers had larger decreases
in calcium absorption may in turn modify lead levels. in IQ with increasing lead exposure compared to BB
Lead binds to calcium proteins, which are regulated carriers. In one study there were some indications of
by the vitamin D endocrine system and the VDR. The modification by VDR TaqI (rs731236) on posture and
VDR gene displays polymorphisms and Ames et al. blood lead in 82 highly exposed children. There are
(1999) demonstrated that the VDR-fok1 (rs10735810) also indications of a modification of other aspects of
genotype FF increases bone mineral density and cal- central nervous system toxicity by VDR TaqI, in adoles-
cium absorption by 30-40% in healthy children. Haynes cents (Krieg et al., 2010), but the authors present only
and coworkers (2003) have shown that children from brief reports of their results.
New York with the Ff genotype displayed lower blood In summary, VDR polymorphisms seem to affect
lead than the FF genotype after exposure to floor-dust the toxicokinetics of lead, while information on the
lead, suggesting that this genotype is an effect modi- modification of toxic effects is scarce and inconsis-
fier. In addition, Rezende and coworkers (2008) found tent. There may be several reasons for the different
an influence of the fok1 genotype in environmentally results on VDR polymorphisms and blood lead levels.
exposed subjects: ff carriers had lower blood and One major reason is probably the fact that the gene is
plasma lead (Rezende et al., 2008). highly polymorphic and exists in several population-
Another VDR polymorphism, BsmI (rs1544410), specific haplotypes that may blur associations when
affects bone mineral density: decreased density has analyzing single polymorphisms in relation to bio-
been reported in those with the BB compared with the markers of lead.
bb genotype (Cooper and Umbach, 1996). In Korean
battery workers, individuals with the BB and Bb gen-
2.5.3 HFE and Other Iron Metabolism Genes
otypes displayed higher blood and tibia lead levels
(Schwartz et al., 2000) as well as diastolic blood pres- Lead absorption is linked to the uptake of iron in the
sure compared with participants with the bb genotype gastrointestinal tract: when iron is limited, lead absorp-
(Lee et al., 2001). The study by Rezende et al. (2008) tion is increased (Cheng et al., 1998). The autosomal
also found higher blood lead in carriers of BB and Bb; genetic disease hemochromatosis is caused by mutations
similarly, a study on low-level Pb-exposed Austrian in the HFE gene. The HFE gene is mutated at position
students (n = 324) found B carriers to have higher lev- 282 (rs1800562), which leads to cysteine-tyrosine substi-
els of lead in urine, but not in blood (Gundacker et al., tution, in 85% of affected individuals. Another variant
2009). In a study of pregnant women, the combination of this gene, reading to the nonsynonymous exchange
of VDR fok1, BsmI, and another polymorphism (ApaI) of histidine to asparagine (His63Asp; rs1799945), is
revealed lower serum lead for the haplotype consisting also associated with hemochromatosis, but with lower
of the f, b, and a alleles (Rezende et al., 2010). However, penetrance (Waheed et al., 1997). Since mutations in the
the opposite effect of VDR BsmI was found in a study HFE gene affect iron uptake, variants of this gene or oth-
on lead-exposed workers and controls: BB carriers ers affecting the iron status may also have a modifying
 12 Gene-Environment Interactions for Metals 253

effect on the levels of lead. Barton et al. (1994) showed bicarbonate cotransporter, member 7 (SLC4A7).
that subjects homozygous for the Cys282Tyr mutation Metallothioneins are low molecular weight proteins
had increased blood lead and that there was an allele- involved in the homeostasis of zinc. Their transcription
dosage effect; similar findings have been reported is induced by various heavy metals, and there is weak
by others (Hopkins et al., 2008; Wang et al., 2007). On evidence of lead-induced metallothionein synthesis
the other hand, in a Swedish study, blood lead levels (Gillis et al., 2012). Kita et al. (2006) demonstrated an
were lower among hemochromatosis-affected indi- effect of the metallothionein 2A (MT2A) polymor-
viduals than among controls (Åkesson et al., 2000). In phism, rs28366003, on expression levels in response
accordance with the Swedish study, Wright et al. (2004) to cadmium and zinc exposure in vitro. Among 616
showed among elderly men that carriers of at least one individuals, those carrying GG had significantly lower
variant HFE allele from Cys282Tyr or His63Asp had blood zinc concentrations and higher blood lead con-
lower levels of blood lead or bone lead. Adjusted analy- centrations compared to individuals with AG or AA
sis showed that one HFE variant was an independent (Kayaalti et al., 2011). In a linkage analysis between
risk factor for significantly lower levels of lead in the lead concentration in erythrocytes and chromosomal
patella. This makes the picture somewhat complicated. regions, a strong signal was found for chromosome 3,
For several of the toxic effects of lead, such as effects near SLC4A7 (Whitfield et al., 2010); however, so far,
on the cardiovascular or cognitive systems and birth specific polymorphisms in this gene have not further
weight, iron is a key player. Because iron can enhance been explored in relation to lead concentrations or
the oxidative effects of lead, differences in iron metab- toxic effects.
olism might modify the adverse effects of lead. Zhang One study focused on the role of genes related to oxi-
and colleagues (2010) reported that His63Asp, but not dative stress in relation to lead exposure and the risk of
Cys282Tyr, is a modifier of the effects of lead (measured different types of brain tumors (Bhatti et al., 2009). The
in the bone) on pulse pressure (n = 619); His63Asp vari- authors found a small increased risk of meningiomas
ant carriers showed a stronger effect of lead. This finding with increasing lead exposure, but not with other types
has been supported by experimental studies showing of tumors, and the role of lead in overall brain tumori-
that His63Asp increases, whereas Cys282Tyr decreases, genesis was weak. Nevertheless, the authors reported
oxidative stress (Lee et al., 2007; Mitchell et al., 2009). that three genes involved in handling ROS modified
In addition, HFE variants appear to play a role in the the risk association between cumulative lead exposure
association between cumulative exposure to lead and and glioblastoma. Another study focused also on oxi-
prolonged electrocardiogram QT interval (a marker for dative stress-related genes, namely GSTM1 and GSTP1,
sudden cardiac death and ventricular arrhythmia). In the and their role in lead-induced toxicity on the cognitive
same study population as Zhang et al. (2010), Park et al. function of elderly men. Eum et al. (2013) found that
(2009) showed an increased lead-related lengthening of individuals with increasing numbers of GSTP1 Val105
the QT interval with an increasing number of HFE vari- alleles had stronger associations between tibia lead
ants; however, they did not report differences in effects and low IQ scores.
for His63Asp and Cys282Tyr. A possible interaction Both lead and iron are associated with changes in
with the number of GT repeats in the heme oxygenase the dopaminergic system, which is responsible for
(HMOX1) promoter was reported, but no effects were the modulation of behavior and cognition (reviewed
found for the transferrin (TF; rs1049296) gene. The allelic in Beard and Connor, 2003; Jones and Miller, 2008).
dose of HFE was found to modify lead-related cognitive ­Froelich and colleagues (2007) found that the effect
decline (Wang et al., 2007). In a study on lead and birth of lead on executive function in children, one of the
weight, infants carrying the His63Asp variant had lower hallmark effects of lead, is modified by the number
birth weight, and the maternal 63Asp variant enhanced of repeats in one exon of the dopamine D4 receptor
the maternal lead-related negative effects (Cantonwine (DRD4) gene. Animal studies indicate that autoregula-
et al., 2010). There was no effect of the TF SNP rs1049296. tion of the dopaminergic system through the dopamine
However, this SNP played a role in lead-related effects receptor D2 (DRD2) gene plays an important role in
on IQ in Indian children (Roy et al., 2013): children carry- the neurobiological effects of lead and iron (Beard and
ing the TF variant allele had lower IQ points in relation Connor, 2003; Jones and Miller, 2008). The DRD2 Taq
to lead exposure, as well as lower hemoglobin levels. IA polymorphism is functional and linked to reduced
D2 receptor density and availability (Pohjalainen et al.,
1998; Thompson et al., 1997). Roy and colleagues (2011)
2.5.4 Other Genes
analyzed 717 children in India and found a larger nega-
A few other genes are reported to be relevant to tive effect of lead on IQ among DRD2 Taq IA (rs1800497)
the toxicokinetics of lead, i.e. genes from the family of TT carriers. Further, this polymorphism was not asso-
metallothioneins and solute carrier family 4, sodium ciated with a hemoglobin-associated increase in IQ, as
254 Karin Broberg, Karin Engström, and Shegufta Ameer

seen for the CT and CC carriers, indicating that this containing lead and cadmium, an inverse association
DRD2 polymorphism disrupts the protective effect of was found between lead and miR-146a expression in
hemoglobin on cognition and increases vulnerability blood (Bollati et al., 2010).
to the toxic effects of lead. However, in a study into
lead-associated executive function in younger children
2.6 Manganese
(24 and 48 months) with a somewhat lower lead expo-
sure (mean 8.1 versus 11.5 μg/L in the Indian study), Exposure to high levels of manganese is often occu-
a protective effect among DRD2 Taq IA TT carriers was pational and can have long-term consequences on the
found (Kordas et al., 2011). nervous system (Lucchini et al., 2007). Despite emerg-
ing evidence that environmental exposure to man-
ganese may also cause cognitive and motor deficits
2.5.5 Epigenetic Effects of Lead
(Bouchard et al., 2011; Riojas-Rodriguez et al., 2010;
There are so far only a few studies on the epigen- Standridge et al., 2008), there are only two studies into
etic effects of lead. Two cross-sectional studies used the genetic markers that influence the toxicokinetics of
methylation of retrotransposons (the Alu and LINE1 manganese in humans. (The lack of studies on gene-
families) as a surrogate marker of global methylation. manganese interactions is mainly due to the lack of
Pilsner and colleagues reported that maternal patella a reliable biomarker for manganese exposure.) Both
lead, representing cumulative lead exposure, was studies have investigated genes involved in iron metab-
inversely associated with LINE1 methylation in the olism because there is a well-established inverse asso-
cord blood of 103 newborns, whereas maternal tibia ciation between iron stores and manganese absorption
lead inversely correlated with Alu methylation (­Pilsner and it is likely that iron and manganese share the same
et al., 2009). Wright and colleagues (2010) found that transport and regulatory proteins. In a pilot study of
patella lead was inversely associated with LINE1 141 individuals living near a ferromanganese refin-
methylation in 517 elderly men. The consequences of ery in the USA, Haynes and coworkers (2010) did not
reduced Alu and LINE1 methylation on human health observe any significant associations between HFE or
are not clear. It may reflect hypomethylation at other TF polymorphisms with manganese levels in the hair
gene-rich areas of the genome, or hypomethylation or blood, but the association between hair manganese
may erroneously activate the retroposons or nearby and estimated ambient air manganese became signifi-
genes. Interestingly, both lead and cadmium are cant when polymorphisms in the hemochromatosis
inversely associated with hypomethylation of LINE1, gene (HFE) and the transferrin gene (TF), involved
which may indicate a shared mechanism. in iron uptake and transport, were included in linear
How lead acts on DNA methylation is not known. models. Based on this finding, the authors suggested
The mechanism may involve interference, direct or a genetic influence on manganese toxicokinetics. Fur-
indirect, with ROS generation (Pilsner et al., 2009) ther, Claus Henn and coworkers (2011) found in 332
or interaction with DNA methyltransferases, such as pregnant Mexican women environmentally exposed
inhibition of the maintenance methyltransferase, DNA to manganese that carriers of either of the HFE poly-
(cytosine-5)-methyltransferase 1 (Dnmt1), that result morphisms C282Y or H63D had 12% lower blood man-
in lower LINE1 methylation. Primates exposed to lead ganese concentrations than women with no variant
as infants showed lower levels of proteins involved alleles. They verified the influence of the HFE status
in DNA methylation as adults, consistent with hypo- on manganese levels by analyzing manganese levels in
methylating effects of lead (Bihaqi et al., 2011). More- Hfe-/- and Hfe+/+ mice, and found that the Hfe-/- mice
over, in rats, peri- or postnatally exposed to lead, effects had lower manganese and higher iron concentrations.
on DNA methyltransferase expression in hippocampi In a small case-control study on susceptibility to
were found: however, both up- and downregulation occupational chronic manganism, Zheng and col-
was observed, depending on exposure time and sex leagues analyzed genetic polymorphisms in cyto-
(Schneider et al., 2013). Notably, one study showed chrome P450 2D6 (CYP2D6) and 1a1 (CYP1A1),
experimentally and through subsequent in vivo anal- NAD(P)H:quinone oxidoreductase (NQO1), GSTM1,
ysis of lead-exposed workers and controls, increased and GSTT1 (Zheng et al., 2002). These genes were
gene-specific DNA methylation of ALAD related to chosen for their potential role in protection against
lead exposure (Li et al., 2011). These findings indicate cellular oxidant processes resulting from manganese
that lead inhibits ALAD expression by increasing gene exposure and their possible involvement in manga-
methylation, thus attenuating lead toxicity of heme nese neurotoxicity. The CYP2D6L variant was pres-
synthesis caused by enzyme inhibition. In a small ent at a lower frequency [odds ratio (OR) = 0.1; 95%
study into exposure to metal-rich particulate matter confidence interval (CI), 0.01-0.82] and there was an
 12 Gene-Environment Interactions for Metals 255

association between CYP2D6L genotype and latency of is glutamyl-cysteine ligase, which is composed of a cat-
manganese poisoning: patients with CYP2D6L devel- alytic subunit (GCLC) and a modifier subunit (GCLM).
oped symptoms 10 years later than those who were The GSTs conjugate GSH to electrophilic compounds
homozygous wild type. However, the study was based (Hayes and Strange, 2000; Strange and Fryer, 1999),
on a small number of cases and the relevance of the and may also affect the metabolism of mercury.
CYP2D6L polymorphism may be questioned because Several studies have evaluated whether polymor-
there are no convincing functional data to support a phism in GCLC, GCLM and GST genes are associated
phenotypic effect of this variant on enzyme activity with mercury biomarkers. The best-studied glutamyl-
(Johansson et al., 1993; Raimundo et al., 2000). cysteine ligase polymorphisms are -129 C/T in GCLC
In the juvenile form of Parkinsonism, one of the tar- (Koide et al., 2003) and -588 C/T in GCLM (Nakamura
get genes (ATPA13A2; encoding probable cation-trans- et al., 2002), which have been shown to affect GSH
porting ATPase 13A2) (Myhre et al., 2008) is related production (lower promoter activity for the T-alleles).
to manganese toxicity. The ATP13A2 gene product Glutathione-S-transferase pi (GSTP1) polymorphisms
belongs to a large group of lysosomal transport pro- leading to amino acid substitutions Ile105Val and
teins in the P5-type ATPase family. Dopaminergic loss Ala114Val have been associated with differences in
caused by α-synuclein overexpression in animal and enzyme activity for mercuric dichloride and meth-
neuronal cell Parkinson disease models is avoided ylmercury (Goodrich and Basu, 2012): GSTP1 Val105
by coexpression of ATP13A2, and a yeast ortholog of and Ala114 were the most sensitive to inhibition by the
ATP13A2 helps to protect cells from manganese tox- mercury species. GSTP1 105Val is more common (allele
icity (Gitler et al., 2009). Calcium-transporting ATPase frequencies range, 10-50%). In addition, influence of
type 2C member 1 (ATPase 2C1; encoded by ATP2C1) the deletion alleles of glutathione-S-transferase mu
is a probable Ca2+/Mn2+ transporter pump expressed (GSTM1) and glutathione-S-transferase theta (GSTT1)
in the membrane of the Golgi apparatus of mamma- has been analyzed.
lian cells (Murin et al., 2006) and plays an important Two studies evaluated the retention of mercury in
role in Mn2+ homeostasis. In a study of a population erythrocytes (a marker mainly of methylmercury expo-
of adolescents and elderly people from Italy, genetic sure) in Swedish populations, one with low (Custodio
variations in ATP13A2 and ATP2C1 were explored for et al., 2004) (n = 365) and one with high (n = 292) fish con-
a potential role in modifying the neurotoxic effects of sumption, the main exposure route for methylmercury
environmental exposure to manganese (Rentschler (Engström et al., 2008). In the low-exposed population,
et al., 2012). ATP13A2 polymorphisms rs4920608 and individuals with one variant allele of either GCLC-129
rs2871776 significantly modified the effects of manga- (T) or GSTP1-114 (Val) had higher erythrocyte mer-
nese exposure on impaired motor coordination in the cury, while the opposite pattern was seen in the highly
elderly, but no effects were found in adolescents or for exposed population; carriers of GSTP1-105Val or
ATP2C1. GSTP1-114Val had lower levels of erythrocyte mercury
at high fish intake, compared with individuals with
no variant alleles. In the highly exposed population,
2.7 Mercury individuals with the rare GCLM-588 TT allele (n = 12)
Mercury in its different forms, inorganic and organic had higher levels of erythrocyte mercury compared to
mercury (methyl and ethyl), shows a large interindi- individuals with CC or CT alleles.
vidual variation in both metabolism and susceptibility Goodrich et al., studied the influence of GSH-related
to toxic effects (al-Shahristani and Shihab, 1974; Clark- polymorphisms on hair mercury (a marker mainly of
son, 2002; Hursh et al., 1976), possibly due to heredi- methylmercury exposure) concentrations in a popula-
tary differences. tion (n = 515) with similar levels of hair mercury as the
general U.S. population (Goodrich et al., 2011). Val-
alleles of GSTP1-105 and GSTP1-114 were associated
2.7.1 Glutathione-Related and Other Genes and
with decreased hair mercury concentrations. In addi-
Mercury Toxicokinetics
tion, a polymorphism (rs3761144; in the 5ʹ untranslated
Glutathione (GSH) plays a central role in the region) in a gene encoding another GSH-synthetizing
metabolism of both inorganic and organic forms of enzyme, glutathione synthetase (GSS) influenced hair
mercury, since both are eliminated as GSH conjugates mercury levels. Here, an effect modification was seen:
(Ballatori and Clarkson, 1985; Dutczak and Ballatori, for each variant allele (G), more mercury was accumu-
1994); polymorphisms in GSH-synthesizing or GSH- lated in the hair with increasing fish consumption.
conjugating enzymes may thus influence the excretion Gundacker et al. (2007, 2009) evaluated GSTT1,
capacity. The rate-limiting enzyme for GSH synthesis GSTM1, GSTA1, GSTP1, and GCLC (mainly as
256 Karin Broberg, Karin Engström, and Shegufta Ameer

combinations of two polymorphisms) and hair mer- number of amalgam fillings in the subject’s mouth, the
cury concentrations among students in Austria with number of fillings removed or performed per week in
fairly low levels of hair mercury. Individuals with the the dental office, and occupation (dentist vs. nonden-
GSTP1-114Val allele had higher hair mercury com- tist). Here, the GSTT1 deletion was associated with
pared to individuals with AlaAla. Several GST/GST decreased urine mercury.
and GST/GCLC combinations affected hair mercury, The same study population was used to investigate
particularly when they included GSTP1-105 or the the influence of polymorphisms in SEPP1 (Goodrich
GSTP1-114Val-alleles. Furthermore, hair mercury et al., 2011), and metallothionein genes (Wang et al.,
concentrations were significantly increased in indi- 2012a) on urinary mercury. At higher exposure levels,
viduals with both GSTM1 and GSTT1 deletion geno- individuals with SEPP1 rs7579 CT + TT excreted more
types, compared to those carrying GSTM1 and GSTT1. urinary mercury per exposure unit (amalgam) com-
The presence of both GSTM1 and GSTT1 was also pared to those with CC. For metallothionein, subjects
associated with a higher expression of metallothio- with MT1M rs2270836 AA (notably few individuals;
nein 1X (MT1X; encoding a metal-binding enzyme) n = 10) or MT2A (rs10636) CC genotypes had lower uri-
(Gundacker et al., 2007). nary mercury levels compared to those with MT1M or
Metallothioneins bind to mercury due to their high MT2A GG genotypes.
cysteine content, and selenoprotein pi 1 (SEPP1) binds Engström et al. (2013b) investigated whether poly-
to mercury via selenocysteine residues (Chen et al., morphisms in mercury transporter genes modified
2006). Two studies examined the influence of polymor- the levels of urinary mercury (n = 1017 individuals
phisms in SEPP1 (Goodrich et al., 2011) and metallothio- from Indonesia, the Philippines, Tanzania, and Zimba-
nein genes (Wang et al., 2012a) on hair mercury levels bwe). Polymorphisms in four transporter genes had an
(n = 515). One polymorphism in SEPP1 (rs7579; in the 3ʹ effect on urinary mercury levels: solute carrier family
untranslated region) showed effect modification on the 22, members 6 and 8 (SLC22A6/OAT1 and SLC22A8/
association between estimated methylmercury intake OAT3), solute carrier family 7, member 5 (SLC7A5/
(based on fish consumption) and hair mercury, where LAT1), and ATP-binding cassette subfamily C, member
the minor allele was associated with lower hair mer- 2 (ABCC2/MRP2). ABCC2 rs1885301 (situated in the 5ʹ
cury relative to intake from fish. For metallothionein untranslated region) was associated with urinary mer-
genes, subjects with the Lys-allele of MT1A Lys51Arg cury in all populations. Three ACBB2 polymorphisms,
or the TT genotype for MT1M rs9936741 had lower rs1885301, rs2273697 (Val417Ile), and rs717620 (in the
hair mercury concentrations than subjects with MT1A promoter), showed particularly strong associations
ArgArg or MT1M CT + CC, respectively. However, with urinary mercury in the most highly exposed sub-
some genotype groups were small (n = 24 for MT1A group. LAT1 rs33916661 showed similar associations
Lys-carriers and n = 15 for MT1M TC + CC carriers). with urinary mercury in all populations, while poly-
MT1A and MT1M also showed effect modification on morphisms in OAT1 (rs4149170) and OAT3 (rs4149182)
the relationship of hair mercury level with estimated were associated with urinary mercury mainly in the
daily methylmercury intake from fish consumption. African populations.
Gundacker et al. (2009) found that MT4 Gly48Asp was Thimerosal contains ethylmercury and is a common
associated with higher hair mercury among students preservative in vaccines. Westphal et al., showed that
in Austria. individuals carrying the deletion genotypes of GSTM1
Data from a study of gold miners, gold buyers, and GSTT1 were overrepresented among patients
and referents in Ecuador (n = 309), who were highly sensitized to thimerosal (Westphal et al., 2000). The
exposed to inorganic mercury through elemental mer- mechanism for this observation could be that slow
cury vapor, showed that the presence of the less efficient elimination of ethylmercury by GSH conjugation in
GSH-­producing GCLM-588T allele was associated with some genotypes, or even an absence of this route of
increased blood, plasma, and urine mercury levels (Cus- elimination, caused a greater tendency toward ethyl-
todio et al., 2005). Another study of the same study pop- mercury accumulation and sensitization.
ulations of Ecuadorian gold miners (Harari et al., 2011)
indicated that subjects with the GCLM-588 CC genotype
2.7.2 Glutathione-Related and Other Genes and
eliminated urinary mercury only half as fast as the others.
Mercury Toxicodynamics
Goodrich et al. (2011) analyzed the influence of a
number of GSH-related polymorphisms on urinary Lee et al. (2010) examined GSTM1 and GSTT1 dele-
mercury levels (mainly a marker of inorganic expo- tions among Korean women with relatively high blood
sure) among dental professionals (n = 515). Individual mercury concentrations and their newborns (n = 417
exposures to inorganic mercury were evaluated by the mother-child pairs). Overall, increasing maternal and
 12 Gene-Environment Interactions for Metals 257

cord blood mercury levels were associated with lower work histories. Genetic polymorphisms evaluated
birth weights. The birth weight of newborns whose were in coproporphyrinogen oxidase (CPOX) (Ech-
mothers had the GSTT1 deletion genotype decreased everria et al., 2006), brain-derived neurotropic factor
with increasing maternal blood mercury levels dur- (BDNF) (Echeverria et al., 2005; Heyer et al., 2004),
ing late pregnancy. For mothers with both GSTM1 and catechol O-methyltransferase (COMT) (Heyer et al.,
GSTT1 deletion genotypes, both maternal and cord 2009), and the serotonin transporter gene promoter
blood mercury levels were associated with lower birth region (5-HTTLPR) (Echeverria et al., 2006; Heyer
weight. et al., 2008). CPOX, along with mercury, impacts the
Engström et al. (2011b) investigated the modify- biosynthesis of heme, a factor likely to be involved in
ing effects of GSH-related genes on the association neurological signaling and neuronal functions. Cat-
between plasma polyunsaturated fatty acids or eryth- echol O-methyltransferase (COMT) is an important
rocyte mercury and the risk of myocardial infarction (a enzyme in the methylation of endogenous and exog-
potential effect of mercury) in a case-control study in enous catechol compounds, including catecholamine
a Swedish population (n = 1027). There were no statis- neurotransmitters and catechol estrogens. BDNF and
tically significant associations, but the few carriers of 5-HTTLPR directly influence mood and behavior. The
the relatively rare GCLM-588 TT genotype (frequency studies showed independent and additive effects of
3%) had a lower risk relative to the CC genotype. The inorganic mercury and each of these polymorphisms
GCLM-588 TT genotype had the highest levels of eryth- on outcomes, but no effect modification was found in
rocyte mercury among controls in this study. any of the studies.
Wang et al. (2012b) evaluated whether polymor- Mercury affects porphyrin metabolism. The CPOX
phisms in GSH-related genes modify the relationship Asn272His polymorphism modified the effect of mer-
between mercury in hair and peripheral nerve func- cury on porphyrin excretion in adults (Heyer et al.,
tion. Individuals with GCLC-129 T had higher sural 2006; Woods et al., 2005). This polymorphism causes
nerve amplitude, compared to those carrying GG, for an atypical response to mercury exposure, resulting in
the same increment of hair mercury levels. However, a highly specific urinary porphyrin profile, and might
there was a problem with false positives: only three out serve as a biomarker of susceptibility to mercury toxic-
of 504 multivariate linear regression models (polymor- ity. Woods et al. (2012) examined whether CPOX Asn-
phisms in selenoproteins and metallothioneins were 272His also modified the neurotoxic effects of mercury
also evaluated) achieved statistical significance. in children exposed to dental amalgam tooth fillings
De Marco et al. (2011, 2012) analyzed endothe- (n = 330). All dose-response associations between uri-
lial nitric oxide synthase (eNOS/NOS3) in relation nary mercury and test performance were restricted
to nitrite concentrations following methylmercury to boys with the CPOX4 variant, and all of these
exposure in a highly exposed study population from associations showed that increased urinary mercury
Brazil (n = 202). Nitrite is a source of the potent vaso- decreased test performance (Woods et al., 2013). In the
dilator nitric oxide (NO), which has protective effects same study population there were numerous signifi-
on cardiovascular functions; animal studies suggest cant interaction effects between variants of the metallo-
that methylmercury exposure inhibits NO synthesis thionein genes MT1M (rs2270837; G-allele) and MT2A
(Grotto et al., 2009). Homozygotes for a 27 nucleotide (rs10636; C-allele), alone and combined, and mercury
repeat polymorphism of eNOS intron 4 had lower exposure on multiple domains of neurobehavioral
nitrite levels than individuals with zero or one cop- function. In this study, all dose-response associations
ies of the polymorphism (all genotype groups had were restricted to boys and showed that increased uri-
similar levels of mercury in blood), and the presence nary mercury decreased test performance.
of the polymorphism contributed to the reduction of In summary, several studies indicate modifica-
nitrite plasma concentration with increasing blood tion of mercury toxicokinetics by polymorphisms in
mercury levels (de Marco et al., 2012). However, the glutathione-related genes, based on the role of GSH-
number of individuals with the homozygous geno- related enzymes in mercury metabolism. The GSTP1
type was low (n = 10). 105Val variant has been, at high exposure levels, asso-
The genetic impact on inorganic mercury-related ciated with lower levels of mercury, possibly through
neurological outcomes or mood has been analyzed direct inhibition of the encoded enzyme. ABCC2 and
in a cohort of occupationally exposed subjects with metallothionein genes are also promising markers of a
rather low inorganic mercury exposure (male dentists genetic influence on mercury retention. Several genes
and female dental assistants) from Washington State, have been described for the toxic effects, in particular
USA. Exposure variables were urinary mercury or a of inorganic mercury, but they need to be validated in
chronic mercury exposure index, based upon subjects’ more populations. Nevertheless, there is still limited
258 Karin Broberg, Karin Engström, and Shegufta Ameer

knowledge about genes that influence mercury toxico- expression of cell cycle regulators (p16 and p21) and
kinetics and toxicity. Partial inconsistent in the results senescence-associated markers, as well as global DNA
may be due to the fact that GSH-related genes have methylation in exposed cells. These changes were
very broad substrate specificities, and their interac- observed in cells directly exposed to methylmercury
tions with mercury may be influenced by the presence (parent cells), as well as in daughter cells cultured
of other substrates or by mercury exposure levels, such under methylmercury-free conditions. Goodrich et al.
as in many of the studies described above (especially (2013) investigated the association between methyl-
for methylmercury) where mercury exposures were mercury (mercury in hair) or inorganic mercury (mer-
rather low. At these low levels and narrow ranges of cury in urine) and altered DNA methylation at LINE1,
exposure, genetic effects, especially effect modifica- DNMT1, SEPW1, and SEPP1 in 131 dental profession-
tions, may be difficult to discern. In addition, self- als. There was a trend toward SEPP1 hypomethylation
reported fish consumption is often used as a proxy for with increasing hair mercury levels in males.
methylmercury exposure. This may be a poor marker,
partly due to the varying content of methylmercury
2.8 Nickel
in fish. Underlying genetic differences, such as differ-
ences in LD patterns between populations, may also Nickel allergy was suggested to have a genetic
affect the outcomes since the genetic polymorphisms causal component in earlier twin and familial studies
studied can differ in associations with functional poly- (Fleming et al., 1999; Menne and Holm, 1983). Based
morphisms. Engström et al. (2013b) found large differ- on the involvement of the immune system, the first
ences in the LD patterns of mercury transporter genes studies analyzed variations in HLA genes and found
between African and Asian populations. an association between HLA-DQA and nickel allergy
(Olerup and Emtestam, 1988), but associations with
this, or other HLA-types, have not been confirmed
2.7.3 Epigenetic Effects
(Emtestam et al., 1993; Ikaheimo et al., 1993). Later,
In utero and postnatal exposure to methylmercury HLA-DR was considered a susceptibility locus for
results in adverse health effects, in particular on the nickel allergy (Gamerdinger et al., 2003).
nervous system; DNA methylation and histone modi- Recently, the filaggrin gene was in several studies
fications have been suggested to mediate these effects identified as a candidate gene for nickel allergy (Novak
(Cheng et al., 2012). Onishchenko et al. (2008) exposed et al., 2008; Thyssen et al., 2008). Filaggrin is an epider-
pregnant mice to methylmercury both during gestation mal protein in the horned layer of the skin. It has been
and postnatally, resulting in depression-like behavior suggested that null mutations of this gene increases
in the offspring. Because reduced mRNA expression of nickel penetration through the epidermis (Thyssen
BDNF is implicated in the pathophysiology of depres- et al., 2008). Ross-Hansen and coworkers (Ross-­Hansen
sion and anxiety, the BDNF promoter region was exam- et al., 2011) found in a study on 3471 Danes that filag-
ined and multiple transcription-repressive epigenetic grin (FLG) null mutations lowered the age of onset for
signatures were identified, such as increased histone nickel dermatitis with about 3 years and was associ-
H3K27me3, decreased histone H3 acetylation, and ated with increased sensitivity to nickel in patch test-
increased CpG methylation. Gadhia et al. (2012) evalu- ing. In the same population associations between the
ated whether acute mercury exposure epigenetically gene claudin-1, important for epidermal barrier func-
altered mouse embryonic stem cells. Mercury signifi- tion, and contact sensitization was found. However,
cantly decreased the total histone protein production the study results were hampered by multiple compari-
per cell and also decreased the total H3K27me1 resi- sons (Ross-Hansen et al., 2013). Probably other genes,
dues. Desaulniers et al. (2009) exposed pregnant rats such as from the p38/MKK6 pathway, may also be
to methylmercury, resulting in a significant decrease in important, as indicated from mouse models of nickel
mRNA levels for the DNA methyltransferases Dnmt1 allergy (Watanabe et al., 2011).
and Dnmt3b in the offspring. They also found a sig- Epigenetic effects of nickel are covered elsewhere
nificant decrease in promoter methylation of cyclin- (Chapter 9).
dependent kinase inhibitor 2A (Cdkn2a), encoding the
p16INK4a tumor suppressor protein, compared with
2.9 Platinum
controls. Bose et al. (2012) investigated short-term direct
and long-term inherited effects of exposure to methyl- Gene-exposure interactions have been reported for
mercury using primary cultures of rat embryonic corti- platinum-containing substances such as cisplatin and
cal neural stem cells. They found that methylmercury oxaliplatin when used in high doses in chemotherapy,
reduced neural stem cell proliferation and altered the and genetic variants have an impact on treatment
 12 Gene-Environment Interactions for Metals 259

outcome and survival in different types of cancer. The Newly discovered mechanisms for gene expression
literature has rapidly grown in recent years. Examples regulation—referred to as epigenetics—have emerged
of genes that influence the response to platinum-based and evidence is rapidly growing for metal-related
chemotherapy are those that prevent platinum-DNA epigenetic effects. So far, these are most clear for arse-
adduct formation by regulating the intake and excre- nic and cadmium, and may explain the association
tion of platinum, e.g. glutathione-related genes GSTT1 between metal exposure early in life and toxic effects
and GSTM4 (Moyer et al., 2010) or transporters ABCC2, later in life, as well as metal carcinogenicity.
SLC22A5, SLCO4C1 (Moyer et al., 2010; Ziliak et al.,
2011); genes involved in the removal of DNA damage
(mismatch-repair gene, MLH1) (Rossi et al., 2010, 2011); References
genes involved in epigenetic regulation (Wagner et al.,
Agusa, T., Iwata, H., Fujihara, J., et al., 2010. Toxicol. Appl. Pharma-
2012); and genes with an unknown functional relation- col. 242, 352–362.
ship to platinum (DAPK3, METTL6) (Tan et al., 2011). Ahsan, H., Chen, Y., Kibriya, M.G., et al., 2003a. Cancer Lett. 201,
These variants are candidates for personalized cancer 57–65.
treatment based on genetic information. Although Ahsan, H., Chen, Y., Wang, Q., et al., 2003b. Toxicol. Lett. 143,
123–131.
some of the aforementioned polymorphisms might
Åkesson, A., Stål, P., Vahter, M., 2000. Environ. Health Perspect. 108,
have an impact on susceptibility to platinum-related 289–291.
health effects from low-level occupational or environ- al-Shahristani, H., Shihab, K.M., 1974. Arch. Environ. Health 28,
mental exposures, nothing is currently known. 342–344.
Alexander, B.H., Checkoway, H., Costa-Mallen, P., et al., 1998. Envi-
ron. Health Perspect. 106, 213–216.
Ames, S.K., Ellis, K.J., Gunn, S.K., et al., 1999. J. Bone Miner. Res. 14,
3 CONCLUSIONS 740–746.
Amicosante, M., Berretta, F., Dweik, R., et al., 2009. Immunology 128,
There is increasing evidence for a major role of e462–e470.
genetic factors in the toxicokinetics and dynamics of Andrew, A.S., Mason, R.A., Kelsey, K.T., et al., 2009. Toxicol. Lett.
187, 10–14.
metals. The picture that emerges for most metals is that
Applebaum, K.M., Karagas, M.R., Hunter, D.J., et al., 2007. Environ.
the genetic influence is probably not attributed to a sin- Health Perspect. 115, 1231–1236.
gle gene for each metal; rather, it is polygenic, with some Baccarelli, A., Bollati, V., 2009. Curr. Opin. Pediatr. 21, 243–251.
genes having a stronger effect than others. The rela- Bailey, K.A., Wu, M.C., Ward, W.O., et al., 2013. J. Biochem. Mol.
tionship between HLA-DPB1 and toxic effects of expo- Toxicol. 27, 106–115.
Ballatori, N., Clarkson, T.W., 1985. Fundam. Appl. Toxicol. 5,
sure to beryllium was one of the first and may be the
816–831.
strongest example of a gene-environment interaction. Banerjee, M., Sarkar, J., Das, J.K., et al., 2007. Carcinogenesis 28,
Genetic variation in the major arsenic-­ metabolizing 672–676.
gene AS3MT predicts arsenic metabolism efficiency Banerjee, N., Paul, S., Sau, T.J., et al., 2013. Toxicol. Sci. (in press).
and probably arsenic-related toxic effects. The filaggrin Barker, D.J., 1992. Br. J. Obstet. Gynaecol. 99, 275–276.
Barton, J.C., Patton, M.A., Edwards, C.Q., et al., 1994. J. Lab. Clin.
gene (FLG) appears to be a promising novel genetic
Med. 124, 193–198.
marker for susceptibility to nickel allergy. The associa- Beard, J.L., Connor, J.R., 2003. Annu. Rev. Nutr. 23, 41–58.
tion between ALAD genotype and lead exposure is the Beebe-Dimmer, J.L., Iyer, P.T., Nriagu, J.O., et al., 2012. Environ.
best-studied interaction; there are definite modifica- Health 11, 43.
tions by ALAD polymorphisms on lead metabolism, Bellinger, D., Hu, H., Titlebaum, L., 1994. Arch. Environ. Health 49,
98–105.
and most studies also indicate an effect on lead toxicity.
Benbrahim-Tallaa, L., Waterland, R.A., Dill, A.L., et al., 2007. Envi-
The partly heterogeneous results of studies on ron. Health Perspect. 115, 1454–1459.
ALAD-lead highlights problems with gene-environ- Bergdahl, I.A., Gerhardsson, L., Schutz, A., et al., 1997a. Arch. Envi-
ment interaction studies, such as uncharacterized bio- ron. Health 52, 91–96.
logical mechanisms; the usage of different biomarkers Bergdahl, I.A., Grubb, A., Schutz, A., et al., 1997b. Pharmacol. Toxi-
of exposure and markers of effect; the low prevalence col. 81, 153–158.
Bergdahl, I.A., Sheveleva, M., Schutz, A., et al., 1998. Toxicol. Sci. 46,
of some alleles; small studies with low power; and the 247–253.
lack of a gold standard for statistical analysis for this Bhattacharjee, P., Das, N., Chatterjee, D., et al., 2013. Mutat. Res.
study type. However, it is important to solve these 755, 1–5.
problems in order to identify of subpopulations that are Bhatti, P., Stewart, P.A., Hutchinson, A., et al., 2009. Cancer Epide-
more vulnerable to metal-induced toxic effects. More- miol. Biomarkers Prev. 18, 1841–1848.
Bihaqi, S.W., Huang, H., Wu, J., et al., 2011. J. Alzheimers Dis. JAD
over, genetic variants may be significant with respect 27, 819–833.
to low-level exposures by influencing risk assessment Björkman, L., Vahter, M., Pedersen, N.L., 2000. Environ. Health Per-
and the setting of tolerable limits of exposure. spect. 108, 719–722.
260 Karin Broberg, Karin Engström, and Shegufta Ameer

Bollati, V., Marinelli, B., Apostoli, P., et al., 2010. Environ Health. Per- De Chaudhuri, S., Mahata, J., Das, J.K., et al., 2006. Mutat. Res. 601,
spect 118, 763–768. 102–112.
Bose, R., Onishchenko, N., Edoff, K., et al., 2012. Toxicol. Sci. 130, De Chaudhuri, S., Ghosh, P., Sarma, N., et al., 2008. Environ. Health
383–390. Perspect. 116, 501–505.
Bouchard, M.F., Sauve, S., Barbeau, B., et al., 2011. Environ. Health de Marco, K.C., Braga, G.U., Barbosa Jr, F., 2011. J. Toxicol. Environ.
Perspect. 119, 138–143. Health A 74, 1323–1333.
Breton, C.V., Kile, M.L., Catalano, P.J., et al., 2007a. Environ. de Marco, K.C., Antunes, L.M., Tanus-Santos, J.E., et al., 2012. Sci.
Health 6, 39. Total Environ. 414, 708–712.
Breton, C.V., Zhou, W., Kile, M.L., et al., 2007b. Carcinogenesis 28, Deng, G., Nguyen, A., Tanaka, H., et al., 2006. Int. J. Cancer 118,
1520–1525. 2999–3005.
Bull, P.C., Thomas, G.R., Rommens, J.M., et al., 1993. Nat. Genet. 5, Desaulniers, D., Xiao, G.H., Lian, H., et al., 2009. Int. J. Toxicol. 28,
327–337. 294–307.
Cantone, L., Nordio, F., Hou, L., et al., 2011. Environ. Health Per- Dotti, C., D’Apice, M.R., Rogliani, P., et al., 2004. Sarcoidosis Vasc.
spect. 119, 964–969. Diffuse Lung Dis. 21, 29–34.
Cantonwine, D., Hu, H., Tellez-Rojo, M.M., et al., 2010. Environ. Drobna, Z., Waters, S.B., Walton, F.S., et al., 2004. Toxicol. Appl. Phar-
Health 9, 43. macol. 201, 166–177.
Cartwright, G.E., Edwards, C.Q., Kravitz, K., et al., 1979. N. Engl. J. Drobna, Z., Waters, S.B., Devesa, V., et al., 2005. Toxicol. Appl. Phar-
Med. 301, 175–179. macol. 207, 147–159.
Chervona, Y., Hall, M.N., Arita, A., et al., 2012. Cancer Epidemiol. Dutczak, W.J., Ballatori, N., 1994. J. Biol. Chem. 269, 9746–9751.
Biomarkers Prev 21, 2252–2260. Echeverria, D., Woods, J.S., Heyer, N.J., et al., 2005. Neurotoxicol.
Chanda, S., Dasgupta, U.B., Guhamazumder, D., et al., 2006. Toxicol. Teratol. 27, 781–796.
Sci. 89, 431–437. Echeverria, D., Woods, J.S., Heyer, N.J., et al., 2006. Neurotoxicol.
Chen, C., Yu, H., Zhao, J., et al., 2006. Environ. Health Perspect. 114, Teratol. 28, 39–48.
297–301. Edwards, T.M., Myers, J.P., 2007. Environ. Health Perspect. 115,
Chen, H., Li, S., Liu, J., et al., 2004. Carcinogenesis 25, 1779–1786. 1264–1270.
Chen, J.W., Wang, S.L., Wang, Y.H., et al., 2012. Chemosphere 88, Emtestam, L., Zetterquist, H., Olerup, O., 1993. J. Invest. Dermatol.
432–438. 100, 271–274.
Chen, Y.C., Xu, L., Guo, Y.L., et al., 2003. J. Environ. Sci. Health A Tox. Engel, L.S., Taioli, E., Pfeiffer, R., et al., 2002. Am. J. Epidemiol. 156,
Hazard Subst. Environ. Eng. 38, 201–211. 95–109.
Cheng, T.F., Choudhuri, S., Muldoon-Jacobs, K., 2012. J. Appl. Toxi- Engström, K., Broberg, K., Concha, G., et al., 2007. Environ. Health
col. 32, 643–653. Perspect. 115, 599–605.
Cheng, Y., Willett, W.C., Schwartz, J., et al., 1998. Am. J. Epidemiol. Engström, K., Stromberg, U., Lundh, T., et al., 2008. Environ. Health
147, 1162–1174. Perspect. 116, 734–739.
Chiou, H.Y., Hsueh, Y.M., Hsieh, L.L., et al., 1997. Mutat. Res. 386, Engström, K., Nermell, B., Concha, G., et al., 2009. Mutat. Res.
197–207. 667, 4–14.
Chung, C.J., Huang, C.J., Pu, Y.S., et al., 2008a. Toxicol. Appl. Phar- Engström, K.S., Vahter, M., Lindh, C., et al., 2010. Mutat. Res. 683,
macol. 232, 203–209. 98–105.
Chung, C.J., Huang, C.J., Pu, Y.S., et al., 2008b. Toxicol. Appl. Phar- Engström, K., Vahter, M., Mlakar, S.J., et al., 2011a. Environ. Health
macol. 226, 14–21. Perspect. 119, 182–188.
Chung, C.J., Hsueh, Y.M., Bai, C.H., et al., 2009a. Cancer Causes Con- Engström, K.S., Wennberg, M., Stromberg, U., et al., 2011b. Environ.
trol 20, 1653–1661. Health 10, 33.
Chung, C.J., Pu, Y.S., Su, C.T., et al., 2010. Cancer Causes Control 21, Engström, K.S., Hossain, M.B., Lauss, M., et al., 2013a. PloS One 8,
1605–1613. e53732.
Chung, C.J., Pu, Y.S., Su, C.T., et al., 2011. Sci. Total Environ. 409, Engström, K., Ameer, S., Bernaudat, L., et al., 2013b. Environ. Health
465–470. Perspect. 121, 85–91.
Chung, C.J., Huang, C.Y., Pu, Y.S., et al., 2013. Toxicol. Appl. Phar- Eum, K.D., Wang, F.T., Schwartz, J., et al., 2013. Neurotoxicology
macol. 266, 254–259. 39C, 65–71.
Chung, W.H., Sung, B.H., Kim, S.S., et al., 2009b. Int. J. Oncol. 34, Fang, F., Kwee, L.C., Allen, K.D., et al., 2010. Am. J. Epidemiol. 171,
1669–1679. 1126–1133.
Clarkson, T.W., 2002. Environ. Health Perspect 110 (Suppl. 1), 11–23. Feder, J.N., Gnirke, A., Thomas, W., et al., 1996. Nat. Genet. 13, 399–408.
Claus Henn, B., Kim, J., Wessling-Resnick, M., et al., 2011. Environ. Fleming, D.E., Chettle, D.R., Wetmur, J.G., et al., 1998. Environ. Res.
Health 10, 97. 77, 49–61.
Cooper, G.S., Umbach, D.M., 1996. J. Bone Miner. Res. 11, 1841–1849. Fleming, C.J., Burden, A.D., Forsyth, A., 1999. Contact Dermatitis 41,
Custodio, H.M., Broberg, K., Wennberg, M., et al., 2004. Arch. Envi- 251–253.
ron. Health 59, 588–595. Froehlich, T.E., Lanphear, B.P., Dietrich, K.N., et al., 2007. Biol. Psy-
Custodio, H.M., Harari, R., Gerhardsson, L., et al., 2005. Arch. Envi- chiatry 62, 243–249.
ron. Occup. Health 60, 17–23. Frosst, P., Blom, H.J., Milos, R., et al., 1995. Nat. Genet. 10, 111–113.
Dalton, T.P., Miller, M.L., Wu, X., et al., 2000. Pharmacogenetics 10, Fujihara, J., Kunito, T., Agusa, T., et al., 2007. Toxicol. Appl. Pharma-
141–151. col 225, 251–254.
Dalton, T.P., He, L., Wang, B., et al., 2005. Proc. Natl. Acad. Sci. U.S.A. Fujihara, J., Soejima, M., Koda, Y., et al., 2008. Mutat. Res. 654, 158–
102, 3401–3406. 161.
Dangleben, N.L., Skibola, C.F., Smith, M.T., 2013. Environ. Health Fujihara, J., Soejima, M., Yasuda, T., et al., 2009. Toxicol. Appl. Phar-
12, 73. macol. 243, 292–299.
De Boeck, M., Lison, D., Kirsch-Volders, M., 1998. Carcinogenesis 19, Fujino, Y., Guo, X., Liu, J., et al., 2005. J. Expo. Anal. Environ. Epide-
2021–2029. miol. 15, 147–152.
 12 Gene-Environment Interactions for Metals 261

Gadhia, S.R., Calabro, A.R., Barile, F.A., 2012. Toxicol. Lett. 212, Hsu, L.I., Chiu, A.W., Huan, S.K., et al., 2008. Toxicol. Appl. Pharma-
169–179. col. 228, 144–155.
Gaede, K.I., Amicosante, M., Schurmann, M., et al., 2005. J. Mol. Hsu, L.I., Chen, W.P., Yang, T.Y., et al., 2011. J. Biomed. Sci. 18, 51.
Med. 83, 397–405. Hsueh, Y.M., Lin, P., Chen, H.W., et al., 2005. J. Toxicol. Environ.
Gamerdinger, K., Moulon, C., Karp, D.R., et al., 2003. J. Exp. Med. Health A 68, 1471–1484.
197, 1345–1353. Hu, H., Wu, M.T., Cheng, Y., et al., 2001. Environ. Health Perspect.
Gao, A., Lu, X.T., Li, Q.Y., et al., 2010. Sci. Total Environ. 408, 4052– 109, 827–832.
4055. Huang, C.Y., Su, C.T., Chu, J.S., et al., 2011. Toxicol. Appl. Pharmacol.
Garcia-Leston, J., Roma-Torres, J., Vilares, M., et al., 2012. Environ. 257, 349–355.
Int. 43, 29–36. Hursh, J.B., Cherian, M.G., Clarkson, T.W., et al., 1976. Arch. Envi-
Gaudet, F., Hodgson, J.G., Eden, A., et al., 2003. Science 300, ron. Health 31, 302–309.
489–492. IARC (International Agency for Research on Cancer), 2004. IARC
Gillis, B.S., Arbieva, Z., Gavin, I.M., 2012. BMC Genomics 13, 344. Monographs. Some drinking-water disinfectants and contami-
Gitler, A.D., Chesi, A., Geddie, M.L., et al., 2009. Nat. Genet. 41, nants, including arsenic, vol. 84, International Agency for Re-
308–315. search on Cancer, Lyon.
Gluckman, P.D., Hanson, M.A., 2006. Developmental origins of IARC, 2006. IARC Monographs. Evaluation of Carcinogenic Risks
health and disease. Cambridge University Press, Cambridge; to Humans. Cobalt in Hard-metals and Cobalt Sulfate, Gallium
New York. Arsenide, Indium Phosphide and Vanadium Pentoxide, vol. 86,
Gollan, J.L., Gollan, T.J., 1998. J. Hepatol. 28 (Suppl. 1), 28–36. International Agency for Research on Cancer, Lyon.
Gomez-Rubio, P., Meza-Montenegro, M.M., Cantu-Soto, E., et al., Ikaheimo, I., Tiilikainen, A., Karvonen, J., et al., 1993. Int. Arch. Al-
2010. J. Appl. Toxicol. 30, 260–270. lergy Immunol. 100, 248–250.
Gong, G., O’Bryant, S.E., 2012. Environ. Res. 113, 52–57. Jaffe, E.K., Volin, M., Bronson-Mullins, C.R., et al., 2000. J. Biol.
Goodrich, J.M., Basu, N., 2012. Toxicol. In Vitro 26, 630–635. Chem. 275, 2619–2626.
Goodrich, J.M., Wang, Y., Gillespie, B., et al., 2011. Toxicol. Appl. Jaffe, E.K., Martins, J., Li, J., et al., 2001. J. Biol. Chem. 276, 1531–1537.
Pharmacol. 257, 301–308. Jiang, G., Xu, L., Song, S., et al., 2008. Toxicology 244, 49–55.
Goodrich, J.M., Basu, N., Franzblau, A., et al., 2013. Environ. Mol. Johansson, I., Lundqvist, E., Bertilsson, L., et al., 1993. Proc. Natl.
Mutagen. 54, 195–203. Acad. Sci. U.S.A. 90, 11825–11829.
Grotto, D., de Castro, M.M., Barcelos, G.R., et al., 2009. Arch. Toxicol. Jones, D.C., Miller, G.W., 2008. Biochem. Pharmacol. 76, 569–581.
83, 653–662. Kamel, F., Umbach, D.M., Lehman, T.A., et al., 2003. Environ. Health
Gundacker, C., Komarnicki, G., Jagiello, P., et al., 2007. Sci. Total En- Perspect. 111, 1335–1339.
viron. 385, 37–47. Karagas, M.R., Andrew, A.S., Nelson, H.H., et al., 2012. Hum. Genet.
Gundacker, C., Wittmann, K.J., Kukuckova, M., et al., 2009. Environ. 131, 453–461.
Res. 109, 786–796. Kayaalti, Z., Mergen, G., Soylemezoglu, T., 2010. Toxicol. Appl. Phar-
Harari, R., Harari, F., Gerhardsson, L., et al., 2011. Toxicol. Lett. 213, macol. 245, 252–255.
75–82. Kayaalti, Z., Aliyev, V., Soylemezoglu, T., 2011. Toxicol. Appl. Phar-
Harari, F., Engström, K., Concha, G., et al., 2013. Environ. Health Per- macol. 256, 1–7.
spect. 121, 797–803. Kelada, S.N., Shelton, E., Kaufmann, R.B., et al., 2001. Am. J. Epide-
Hashibe, M., Brennan, P., Strange, R.C., et al., 2003. Cancer Epide- miol. 154, 1–13.
miol. Biomarkers Prev 12, 1509–1517. Ketelslegers, H.B., Gottschalk, R.W., Koppen, G., et al., 2008. Cancer
Hayes, J.D., Strange, R.C., 2000. Pharmacology 61, 154–166. Epidemiol. Biomarkers Prev. 17, 1902–1912.
Haynes, E.N., Kalkwarf, H.J., Hornung, R., et al., 2003. Environ. Kile, M.L., Baccarelli, A., Hoffman, E., et al., 2012. Environ. Health
Health Perspect. 111, 1665–1669. Perspect. 120, 1061–1066.
Haynes, E.N., Heckel, P., Ryan, P., et al., 2010. Neurotoxicology 31, Kim, H.S., Lee, S.S., Lee, G.S., et al., 2004. Environ. Health Perspect.
468–474. 112, 538–541.
He, L., Hannon, G.J., 2004. Nat. Rev. Genet. 5, 522–531. Kippler, M., Engström, K., Mlakar, S.J., et al., 2013. Epigenetics 8,
Hernandez, A., Xamena, N., Surralles, J., et al., 2008. Mutat. Res. 637, 494–503.
80–92. Kita, K., Miura, N., Yoshida, M., et al., 2006. Hum. Genet. 120,
Heyer, N.J., Echeverria, D., Bittner Jr., A.C., et al., 2004. Toxicol. Sci. 553–560.
81, 354–363. Koestler, D.C., Avissar-Whiting, M., Houseman, E.A., et al., 2013. En-
Heyer, N.J., Bittner Jr., A.C., Echeverria, D., et al., 2006. Toxicol. Lett. viron. Health Perspect. 121, 971–977.
161, 159–166. Koide, S., Kugiyama, K., Sugiyama, S., et al., 2003. J. Am. Coll. Car-
Heyer, N.J., Echeverria, D., Farin, F.M., et al., 2008. J. Toxicol. Envi- diol. 41, 539–545.
ron. Health A 71, 1318–1326. Kong, A.P., Xiao, K., Choi, K.C., et al., 2012. Clin. Chim. Acta 413,
Heyer, N.J., Echeverria, D., Martin, M.D., et al., 2009. J. Toxicol. Envi- 1053–1057.
ron. Health A 72, 599–609. Kordas, K., Ettinger, A.S., Bellinger, D.C., et al., 2011. J. Pediatr. 159,
Hopkins, M.R., Ettinger, A.S., Hernandez-Avila, M., et al., 2008. En- 638–643.
viron. Health Perspect. 116, 1261–1266. Krieg Jr., E.F., Butler, M.A., Chang, M.H., et al., 2009. Neurotoxicol.
Hossain, M.B., Vahter, M., Concha, G., et al., 2012a. Environ. Health Teratol. 31, 364–371.
Perspect. 120, 879–884. Krieg Jr., E.F., Butler, M.A., Chang, M.H., et al., 2010. Neurotoxicol.
Hossain, M.B., Vahter, M., Concha, G., et al., 2012b. Metallomics 4, Teratol. 32, 262–272.
1167–1175. Kubota, R., Kunito, T., Agusa, T., et al., 2006. J. Environ. Monit. 8,
Hsieh, L.L., Liou, S.H., Chen, Y.H., et al., 2000. J. Occup. Environ. 293–299.
Med. 42, 151–155. Kundu, M., Ghosh, P., Mitra, S., et al., 2011. Mutat Res. 706, 7–12.
Hsieh, Y.C., Lien, L.M., Chung, W.T., et al., 2011. Environ. Res. 111, Lalouel, J.M., Le Mignon, L., Simon, M., et al., 1985. Am. J. Hum.
804–810. Genet. 37, 700–718.
262 Karin Broberg, Karin Engström, and Shegufta Ameer

Lambrou, A., Baccarelli, A., Wright, R.O., et al., 2012. Epidemiology NCBI (National Center for Biotechnology Information), 2006. Snp
23, 668–676. database. http://www.ncbi.nlm.nih.gov/snp (accessed 9.01.12).
Lee, B.E., Hong, Y.C., Park, H., et al., 2010. Environ. Health Perspect. Nishio, H., Hayashi, C., Lee, M.J., et al., 1999. Arch. Toxicol. 73,
118, 437–443. 496–498.
Lee, B.K., Lee, G.S., Stewart, W.F., et al., 2001. Environ. Health Per- Novak, N., Baurecht, H., Schafer, T., et al., 2008. J. Invest. Dermatol.
spect. 109, 383–389. 128, 1430–1435.
Lee, S.Y., Patton, S.M., Henderson, R.J., et al., 2007. FASEB J. 21, NRC (National Research Council), 2001. Arsenic in drinking water:
564–576. 2001 update. National Academy Press, Washington DC.
Lei, L., Chang, X., Rentschler, G., et al., 2012. Toxicol. Appl. Pharma- Olerup, O., Emtestam, L., 1988. Immunogenetics 28, 310–313.
col. 265, 373–379. Onishchenko, N., Karpova, N., Sabri, F., et al., 2008. J. Neurochem.
Lesseur, C., Gilbert-Diamond, D., Andrew, A.S., et al., 2012. Toxicol. 106, 1378–1387.
Lett. 210, 100–106. Paiva, L., Marcos, R., Creus, A., et al., 2008. Pharmacogenet. Genom-
Li, C., Xu, M., Wang, S., et al., 2011. Toxicol. Lett. 203, 48–53. ics 18, 1–10.
Li, X., Shi, Y., Wei, Y., et al., 2012. Environ. Toxicol. Pharmacol. 34, Paiva, L., Hernandez, A., Martinez, V., et al., 2010. Environ. Res. 110,
381–387. 463–468.
Liao, Y.T., Li, W.F., Chen, C.J., et al., 2009. Toxicol. Appl. Pharmacol. Park, S.K., Hu, H., Wright, R.O., et al., 2009. Environ. Health Perspect
239, 178–183. 117, 80–85.
Lin, S., Shi, Q., Nix, F.B., et al., 2002. J. Biol. Chem. 277, 10795–10803. Pawlas, N., Broberg, K., Olewinska, E., et al., 2012. Neurotoxicology
Lindberg, A.L., Kumar, R., Goessler, W., et al., 2007. Environ. Health 33, 37–43.
Perspect. 115, 1081–1086. Perera, F., Herbstman, J., 2011. Reprod. Toxicol. 31, 363–373.
Lindberg, A.L., Ekstrom, E.C., Nermell, B., et al., 2008a. Environ. Res. Picard, V., Govoni, G., Jabado, N., et al., 2000. J. Biol. Chem. 275,
106, 110–120. 35738–35745.
Lindberg, A.L., Rahman, M., Persson, L.A., et al., 2008b. Toxicol. Pierce, B.L., Kibriya, M.G., Tong, L., et al., 2012. PLoS Genet 8,
Appl. Pharmacol. 230, 9–16. e1002522.
Liu, J., Corton, C., Dix, D.J., et al., 2001. Toxicol. Appl. Pharmacol. Pilsner, J.R., Hu, H., Ettinger, A., et al., 2009. Environ. Health Per-
176, 1–9. spect. 117, 1466–1471.
Liu, Y., Liu, J., Habeebu, S.M., et al., 2000. Toxicol. Sci. 57, 167–176. Pilsner, J.R., Hall, M.N., Liu, X., et al., 2012. PloS One 7, e37147.
Lucchini, R.G., Albini, E., Benedetti, L., et al., 2007. Am. J. Ind. Med. Pohjalainen, T., Rinne, J.O., Nagren, K., et al., 1998. Mol. Psychiatry
50, 788–800. 3, 256–260.
Mahaffey, K.R., Gartside, P.S., Glueck, C.J., 1986. Pediatrics 78, Porter, K.E., Basu, A., Hubbard, A.E., et al., 2010. Environ. Res. 110,
257–262. 580–587.
Maier, L.A., Sawyer, R.T., Bauer, R.A., et al., 2001. Am. J. Respir. Crit. Potolicchio, I., Mosconi, G., Forni, A., et al., 1997. Eur. J. Immunol.
Care Med. 164, 1192–1199. 27, 2741–2743.
Maier, L.A., McGrath, D.S., Sato, H., et al., 2003. J. Immunol. 171, Potolicchio, I., Festucci, A., Hausler, P., et al., 1999. Eur. J. Immunol.
6910–6918. 29, 2140–2147.
Marcos, R., Martinez, V., Hernandez, A., et al., 2006. J. Occup. Envi- Rahman, A., Vahter, M., Ekstrom, E.C., et al., 2007. Am. J. Epidemiol.
ron. Med. 48, 334–341. 165, 1389–1396.
Marnell, L.L., Garcia-Vargas, G.G., Chowdhury, U.K., et al., 2003. Raimundo, S., Fischer, J., Eichelbaum, M., et al., 2000. Pharmacoge-
Chem. Res. Toxicol. 16, 1507–1513. netics 10, 577–581.
Marsit, C.J., Eddy, K., Kelsey, K.T., 2006. Cancer Res. 66, 10843–10848. Rajan, P., Kelsey, K.T., Schwartz, J.D., et al., 2007. Am. J. Epidemiol.
Mass, M.J., Wang, L., 1997. Mutat. Res. 386, 263–277. 166, 1400–1408.
Mateuca, R., Aka, P.V., De Boeck, M., et al., 2005. Toxicol. Lett. 156, Rajan, P., Kelsey, K.T., Schwartz, J.D., et al., 2008. J. Occup. Environ.
277–288. Med. 50, 1053–1061.
Mazumder, D.N., Steinmaus, C., Bhattacharya, P., et al., 2005. Epide- Ren, X., Aleshin, M., Jo, W.J., et al., 2011. Environ. Health Perspect.
miology 16, 760–765. 119, 771–777.
McCanlies, E.C., Yucesoy, B., Mnatsakanova, A., et al., 2010. J. Occup. Rentschler, G., Covolo, L., Ahmadi Haddad, A., et al., 2012. Neuro-
Environ. Med. 52, 680–684. toxicology 33, 697–702.
McCarty, K.M., Chen, Y.C., Quamruzzaman, Q., et al., 2007. Environ. Rentschler, G., Kippler, M., Axmon, A., et al., 2013. Environ. Health
Health Perspect. 115, 341–345. Perspect. 121, 467–472.
Menne, T., Holm, N.V., 1983. Int. J. Dermatol. 22, 22–28. Rezende, V.B., Barbosa Jr., F., Montenegro, M.F., et al., 2008. Arch.
Meza, M.M., Yu, L., Rodriguez, Y.Y., et al., 2005. Environ. Health Per- Toxicol. 82, 29–36.
spect. 113, 775–781. Rezende, V.B., Amaral, J.H., Quintana, S.M., et al., 2010. Sci. Total
Mitchell, R.M., Lee, S.Y., Randazzo, W.T., et al., 2009. J. Neuroinflam- Environ. 408, 4955–4960.
mation 6, 6. Richeldi, L., Sorrentino, R., Saltini, C., 1993. Science 262, 242–244.
Miyaki, K., Lwin, H., Masaki, K., et al., 2009. Int. J. Environ. Res. Riojas-Rodriguez, H., Solis-Vivanco, R., Schilmann, A., et al., 2010.
Public Health 6, 999–1009. Environ. Health Perspect. 118, 1465–1470.
Moyer, A.M., Sun, Z., Batzler, A.J., et al., 2010. Cancer Epidemiol. Robertson, K.D., Jones, P.A., 2000. Carcinogenesis 21, 461–467.
Biomarkers Prev. 19, 811–821. Rodrigues, E.G., Kile, M., Hoffman, E., et al., 2012. Biomarkers 17,
Mukhopadhyay, R., Rosen, B.P., 2002. Environ. Health Perspect 110 240–247.
(Suppl. 5), 745–748. Rollins, R.A., Haghighi, F., Edwards, J.R., et al., 2006. Genome Res.
Murin, R., Verleysdonk, S., Raeymaekers, L., et al., 2006. Cell Mol. 16, 157–163.
Neurobiol. 26, 1355–1365. Rosenman, K.D., Rossman, M., Hertzberg, V., et al., 2011. Occup. En-
Myhre, R., Steinkjer, S., Stormyr, A., et al., 2008. BMC Neurol. 8, 47. viron. Med. 68, 487–493.
Nakamura, S., Kugiyama, K., Sugiyama, S., et al., 2002. Circulation Ross-Hansen, K., Menne, T., Johansen, J.D., et al., 2011. Contact Der-
105, 2968–2973. matitis 64, 24–31.
 12 Gene-Environment Interactions for Metals 263

Ross-Hansen, K., Linneberg, A., Johansen, J.D., et al., 2013. Br. J. Der- Tekin, D., Kayaalti, Z., Aliyev, V., et al., 2012. J. Appl. Toxicol. 32,
matol. 168, 762–770. 270–275.
Rossi, D., Fangazio, M., De Paoli, L., et al., 2010. Cancer 116, Thirumaran, R.K., Bermejo, J.L., Rudnai, P., et al., 2006. Carcinogen-
2188–2200. esis 27, 1676–1681.
Rossi, D., Rasi, S., Di Rocco, A., et al., 2011. Blood 117, 2405–2413. Thomas, D.J., Li, J., Waters, S.B., et al., 2007. Exp. Biol. Med. (May-
Rossman, M.D., Stubbs, J., Lee, C.W., et al., 2002. Am. J. Respir. Crit. wood) 232, 3–13.
Care Med. 165, 788–794. Thompson, J., Thomas, N., Singleton, A., et al., 1997. Pharmacoge-
Rothman, K., Greenland, S., 1998. Modern epidemiology, Second ed. netics 7, 479–484.
Lippincott-Raven, Philadelphia. Thyssen, J.P., Carlsen, B.C., Menne, T., 2008. Dermatitis 19,
Roy, A., Hu, H., Bellinger, D.C., et al., 2011. Environ. Health Perspect. 303–307.
119, 144–149. Treas, J.N., Tyagi, T., Singh, K.P., 2012. PloS One 7, e43880.
Roy, A., Ettinger, A.S., Hu, H., et al., 2013. Neurotoxicology Treas, J., Tyagi, T., Singh, K.P., 2013. Prostate 73, 1660–1672.
38, 17–22. Vahter, M., 2002. Toxicology 181-182, 211–217.
Sadewa, H.A., Miyabe, Y., Nishio, H., et al., 2002. Arch. Toxicol. 76, Valenzuela, O.L., Drobna, Z., Hernandez-Castellanos, E., et al., 2009.
467–469. Toxicol. Appl. Pharmacol. 239, 200–207.
Sakai, T., Morita, Y., Araki, T., et al., 2000. Am. J. Ind. Med. 38, van Bemmel, D.M., Boffetta, P., Liao, L.M., et al., 2011a. PLoS One
355–360. 6, e20432.
Saltini, C., Winestock, K., Kirby, M., et al., 1989. N. Engl. J. Med. 320, van Bemmel, D.M., Li, Y., McLean, J., et al., 2011b. Epidemiology 22,
1103–1109. 273–278.
Saltini, C., Richeldi, L., Losi, M., et al., 2001. Eur. Respir. J. 18, Van Goethem, F., Lison, D., Kirsch-Volders, M., 1997. Mutat. Res.
677–684. 392, 31–43.
Sampayo-Reyes, A., Hernandez, A., El-Yamani, N., et al., 2010. Toxi- Verougstraete, V., Mallants, A., Buchet, J.P., et al., 2004. Am. J. Respir.
col. Sci. 117, 63–71. Crit. Care Med. 170, 162–166.
Sato, H., Spagnolo, P., Silveira, L., et al., 2007. Tissue Antigens 70, Vineis, P., Malats, N., 1999. IARC Sci Publ 148, 51–61.
480–486. Wacholder, S., Chanock, S., Garcia-Closas, M., et al., 2004. J. Natl.
Schlebusch, C.M., Lewis Jr., C.M., Vahter, M., et al., 2013. Environ. Cancer Inst. 96, 434–442.
Health Perspect. 121, 53–58. Wagner, K.W., Ye, Y., Lin, J., et al., 2012. Clin. Cancer Res. 18, 585–592.
Schmuck, E.M., Board, P.G., Whitbread, A.K., et al., 2005. Pharmaco- Waheed, A., Parkkila, S., Zhou, X.Y., et al., 1997. Proc. Natl. Acad. Sci.
genet. Genomics 15, 493–501. U.S.A 94, 12384–12389.
Schneider, J.S., Kidd, S.K., Anderson, D.W., 2013. Toxicol. Lett. 217, Wang, F.T., Hu, H., Schwartz, J., 2007. Environ. Health Perspect. 115,
75–81. 1210–1215.
Schwartz, B.S., Lee, B.K., Stewart, W., et al., 1995. Am. J. Epidemiol. Wang, Y., Goodrich, J.M., Gillespie, B., et al., 2012a. Environ. Health
142 (7), 38–745. Perspect. 120, 530–534.
Schwartz, B.S., Lee, B.K., Stewart, W., et al., 1997a. Arch. Environ. Wang, Y., Goodrich, J.M., Werner, R., et al., 2012b. Sci. Total Environ.
Health 52, 97–103. 417-418, 32–38.
Schwartz, B.S., Lee, B.K., Stewart, W., et al., 1997b. Occup. Environ. Wang, Y.H., Yeh, S.D., Shen, K.H., et al., 2009. Toxicol. Appl. Pharma-
Med. 54, 241–246. col. 241, 111–118.
Schwartz, B.S., Lee, B.K., Lee, G.S., et al., 2000. Environ. Health Per- Wang, Z., White, P.S., Petrovic, M., et al., 1999. J. Immunol. 163,
spect. 108, 949–954. 1647–1653.
Scinicariello, F., Murray, H.E., Moffett, D.B., et al., 2007. Environ. Watanabe, M., Ishimaru, N., Ashrin, M.N., et al., 2011. PLoS One 6,
Health Perspect. 115, 35–41. e19017.
Scinicariello, F., Yesupriya, A., Chang, M.H., et al., 2010. Environ. Waterland, R.A., Michels, K.B., 2007. Annu. Rev. Nutr. 27, 363–388.
Health Perspect. 118, 259–264. Waters, S.B., Devesa, V., Del Razo, L.M., et al., 2004. Chem. Res. Toxi-
Sithisarankul, P., Schwartz, B.S., Lee, B.K., et al., 1997. Am. J. Ind. col. 17, 404–409.
Med. 32, 15–20. Weaver, V.M., Schwartz, B.S., Ahn, K.D., et al., 2003. Environ. Health
Smeester, L., Rager, J.E., Bailey, K.A., et al., 2011. Chem. Res. Toxicol. Perspect. 111, 1613–1619.
24, 165–167. Westphal, G.A., Schnuch, A., Schulz, T.G., et al., 2000. Int. Arch. Oc-
Smith, C.M., Wang, X., Hu, H., et al., 1995. Environ. Health Perspect. cup. Environ. Health 73, 384–388.
103, 248–253. Wetmur, J.G., 1994. Environ. Health Perspect. 102 (Suppl. 3), 215–219.
Standridge, J.S., Bhattacharya, A., Succop, P., et al., 2008. J. Occup. Wetmur, J.G., Kaya, A.H., Plewinska, M., et al., 1991. Am. J. Hum.
Environ. Med. 50, 1421–1429. Genet. 49, 757–763.
Steinmaus, C., Yuan, Y., Kalman, D., et al., 2005. Cancer Epidemiol. Weuve, J., Kelsey, K.T., Schwartz, J., et al., 2006. Occup. Environ.
Biomarkers Prev. 14, 919–924. Med. 63, 746–753.
Steinmaus, C., Moore, L.E., Shipp, M., et al., 2007. J. Toxicol. Environ. Whitfield, J.B., Dy, V., McQuilty, R., et al., 2010. Environ. Health Per-
Health A 70, 159–170. spect. 118, 776–782.
Strange, R.C., Fryer, A.A., 1999. IARC Sci Publ 148, 231–249. Wong, R.H., Kuo, C.Y., Hsu, M.L., et al., 2005. Environ. Health Per-
Süzen, H.S., Duydu, Y., Aydin, A., et al., 2003. Am. J. Ind. Med. 43, spect. 113, 1386–1390.
165–171. Wood, T.C., Salavagionne, O.E., Mukherjee, B., et al., 2006. J. Biol.
Suzuki, T., Nohara, K., 2013. J. Appl. Toxicol. 33, 951–958. Chem. 281, 7364–7373.
Tajuddin, S.M., Amaral, A.F., Fernandez, A.F., et al., 2013. Environ. Woods, J.S., Echeverria, D., Heyer, N.J., et al., 2005. Toxicol. Appl.
Health Perspect. 121, 650–656. Pharmacol. 206, 113–120.
Tan, X.L., Moyer, A.M., Fridley, B.L., et al., 2011. Clin. Cancer Res. Woods, J.S., Heyer, N.J., Echeverria, D., et al., 2012. Neurotoxicol.
17, 5801–5811. Teratol. 34, 513–521.
Tanzi, R.E., Petrukhin, K., Chernov, I., et al., 1993. Nat. Genet. 5, Woods, J.S., Heyer, N.J., Russo, J.E., et al., 2013. Neurotoxicol. Tera-
344–350. tol. 39C, 36–44.
264 Karin Broberg, Karin Engström, and Shegufta Ameer

Wright, R.O., Silverman, E.K., Schwartz, J., et al., 2004. Environ. Zhang, C., Liang, Y., Lei, L., et al., 2013. Toxicol. Appl. Pharmacol.
Health Perspect. 112, 746–750. 271, 78–85.
Wright, R.O., Schwartz, J., Wright, R.J., et al., 2010. Environ. Health Zhao, C.Q., Young, M.R., Diwan, B.A., et al., 1997. Proc. Natl. Acad.
Perspect. 118, 790–795. Sci. U.S.A. 94, 10907–10912.
Wu, C.C., Huang, Y.K., Chung, C.J., et al., 2013. Toxicol. Appl. Phar- Zhao, Y., Wang, L., Shen, H.B., et al., 2007. J. Toxicol. Environ. Health.
macol. 272, 30–36. A 70, 1986–1994.
Xu, Y., Wang, Y., Zheng, Q., et al., 2008. Toxicol. Appl. Pharmacol. Zheng, G., Tian, L., Liang, Y., et al., 2011. Neurotoxicology 32, 374–382.
232, 142–149. Zheng, Y.X., Chan, P., Pan, Z.F., et al., 2002. Biomarkers 7, 337–346.
Xu, Y., Li, X., Zheng, Q., et al., 2009. Mutat. Res. 679, 44–49. Zhou, X., Sun, H., Ellen, T.P., et al., 2008. Carcinogenesis 29,
Yan, C., Boyd, D.D., 2006. Mol. Cell. Biol. 26, 6357–6371. 1831–1836.
Zakharyan, R.A., Sampayo-Reyes, A., Healy, S.M., et al., 2001. Chem. Ziemsen, B., Angerer, J., Lehnert, G., et al., 1986. Int. Arch. Occup.
Res. Toxicol. 14, 1051–1057. Environ. Health 58, 245–247.
Zhang, A., Park, S.K., Wright, R.O., et al., 2010. Environ. Health Per- Ziliak, D., O’Donnell, P.H., Im, H.K., et al., 2011. Transl. Res. 157,
spect. 118, 1261–1266. 265–272.