reviews
Annals of Oncology
Annals of Oncology 24: 901–908, 2013
doi:10.1093/annonc/mds543
Published online 7 November 2012
The contribution of molecular epidemiology
to the identification of human carcinogens: current
status and future perspectives
P. Boffetta1,2* & F. Islami1,3
1
The Tisch Cancer Institute and Institute for Translational Epidemiology, Mount Sinai School of Medicine, New York, USA; 2International Prevention Research Institute,
Lyon, France; 3Digestive Disease Research Center, Tehran University of Medical Sciences, Tehran, Iran
Received 6 August 2012; revised 14 September 2012; accepted 18 September 2012
Background: The use of biological-based markers of exposure, intermediate effect, outcome, and susceptibility has
become standard practice in cancer epidemiology, which has contributed to identification of several carcinogenic
agents. Nevertheless, with the exception of biological agents, this contribution, in terms of providing sufficiently strong
evidence as required by the International Agency for Research on Cancer (IARC) monographs, has been modest.
Materials and methods: We discuss the overall contribution of molecular epidemiology to identification of
carcinogens, with focus on IARC monographs.
Results: For many carcinogens, valid biological markers of exposure and mechanisms of actions are not available.
Molecular markers are usually assessed in single biological samples, which may not represent the actual exposure or
biological events related to carcinogens. The contribution of molecular epidemiology to identification of carcinogens has
mainly been limited to the carcinogens acting through a genotoxic mechanism, i.e. when carcinogens induce DNA
damage. A number of factors, including certain hormones and overweight/obesity, may show carcinogenic effects
through nongenotoxic pathways, for which mechanisms of carcinogenicity are not well identified and their biomarkers
are sparse.
Conclusion: Longitudinal assessment of biomarkers may provide more informative data in molecular epidemiology
studies. For many carcinogens and mechanistic pathways, in particular nongenotoxic carcinogenicity, valid biological
markers still need to be identified.
Key words: cancer, carcinogenicity, epidemiology, genotoxicity, molecular epidemiology
introduction
The field of molecular cancer epidemiology has been
established some 40 years ago, with the early applications of
molecular markers to population-based cancer studies [1].
Since then, the use of increasingly sophisticated biologicalbased markers of exposure, intermediate effect, outcome, and
susceptibility has become standard practice in cancer
epidemiology. The maturity of the field is reflected by the
increasing number of textbooks, journals, academic programs,
and professional societies [2–5].
One of the main domains of cancer epidemiology has been
etiology research, which is the study of causes of human
cancer, with the ultimate goal of identifying strategies for
cancer prevention. The causes of cancer can be classified as
*Correspondence to: Prof. P. Boffetta, Mount Sinai School of Medicine, One Gustave
L. Levy Place, Box 1079, New York, NY 10029, USA. Tel: +1-212-824-7073;
Fax: +1-212-876-6789; E-mail: [email protected]
inherited (genetic and epigenetic variants) and acquired
(sometimes defined as ‘environmental’ in a broad sense) [6].
Despite strong evidence from family-based studies of a strong
inherited component in human cancer etiology, the elucidation
of the genes and variants responsible for an increased (or
decreased) risk of cancer has been elusive: some 100 genetic
factors have been identified through family-based studies [7],
and a comparable number of low risk variants have been
identified in recent years through genome-wide association
studies [8]. Collectively, these factors explain a small
proportion of cancers in the population although some may
confer a high risk to carriers, and some special populations
comprise a relatively large proportion of carriers of risk
variants. However, our limited current knowledge of the ways
that genetic and epigenetic factors influence quantitative traits
and disease risk may be a reason for why hereditary factors
have not been recognized as strong risk factors for many
cancers at the general population level.
© The Author 2012. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: [email protected].
reviews
Compared with hereditary factors, the associations between
environmental factors and cancers are better known. Modern
cancer epidemiology was started in the early 1950s with the
conduct of carefully designed cohort and case–control studies
which rapidly lead to the identification of important,
nongenetic causes of cancer, primarily tobacco smoking,
alcohol drinking, and several occupational agents [9–13]. Since
the 1980s, however, it became apparent that the yield of cancer
epidemiology research, and the number of newly discovered
carcinogenic agents, was decreasing (Figure 1). The rising of
the new field of molecular cancer epidemiology was seen as an
opportunity to elucidate the remaining etiologic questions.
Reviews of the causes of human cancer confront the
following challenges: (i) evaluating the overall strength of
evidence for various etiologic hypotheses and (ii) identifying
the role of different components of the evidence (for example,
by study design). These problems are best exemplified by the
inconsistencies between the two reviews of the evidence of a
role of dietary factors in human cancers made by the World
Cancer Research Fund in 1997 and 2007 [14–16]. In the case of
chemical, physical, and biological agents, on the other hand, the
monographs program of the International Agency for Research
on Cancer (IARC) provides a framework for evaluation of
human, animal, and mechanistic data and for an overall
evaluation of carcinogenicity. As a result of the standard format
of the monographs, it is possible to identify the role played by
each set of data in final evaluations. This program therefore
offers the opportunity to assess the contribution of molecular
epidemiologic studies to the identification of several major
classes of human carcinogens. The list of carcinogens provided
by IARC is a widely accepted list.
molecular epidemiology and
identification of human carcinogens
the IARC monographs program
The evaluations in each of the IARC monographs on the
evaluation of carcinogenic risks to humans are collectively
made in case-by-case basis by a group of experts, who critically
review the pertinent peer-reviewed scientific literature. The
working group members have significant knowledge and
experience in research related to the carcinogenicity of the
agents being reviewed and do not have real or apparent
conflicts of interests. The first monograph was published in
1972 [17]. By 2008, 99 monographs were conducted, in each of
Figure 1. Decade of publication of key results on occupational
carcinogens.
 | Boffetta and Islami
Annals of Oncology
them, the data on carcinogenicity of one or several agents were
evaluated. In 2009, IARC conducted five sessions of the
monographs, collectively called as volume 100, to re-evaluate
the human carcinogens identified in the first 99 volumes. In
2011 and 2012 (by August), five additional groups met after
the completion of the volume 100 project. In this article, we
only consider the first 100 monographs.
The evidence for carcinogenicity in epidemiological studies
is classified by the monographs as following: (i) ‘sufficient
evidence of carcinogenicity’, which indicates that a causal
association is credible, and chance, bias, and confounding
could be ruled out with reasonable confidence; (ii) ‘limited
evidence of carcinogenicity’, which shows a causal
interpretation is credible, but chance, bias, or confounding
could not be ruled out with reasonable confidence;
(iii) ‘inadequate evidence of carcinogenicity’, which indicates
no data on cancer in humans are available or the available data
do not permit a conclusion about the presence or absence of a
causal association; and (iv) ‘evidence suggesting lack of
carcinogenicity’. The evidence for carcinogenicity in animals is
classified in similar categories, but with different criteria
adapted to the nature of animal studies. Mechanistic and other
relevant evidence may include data on preneoplastic lesions,
tumor pathology, genetic and related effects, structure–activity
relationships, metabolism and toxicokinetics, physicochemical
parameters, and analogous biological agents. The strength of
the evidence that any carcinogenic effect observed is due to a
particular mechanism is evaluated and classified as weak,
moderate, or strong. Finally, the body of evidence is considered
as a whole using criteria developed by IARC, and agents are
classified as group 1, ‘carcinogenic to humans’; group 2A,
‘probably carcinogenic to humans’; group 2B, ‘possibly
carcinogenic to humans’; group 3, ‘not classifiable as to its
carcinogenicity to humans’; and group 4, ‘probably not
carcinogenic to humans’ [18].
Group 1 category is used when there is sufficient evidence of
carcinogenicity in humans. However, when evidence of
carcinogenicity in humans is less than ‘sufficient’ but there is
sufficient evidence of carcinogenicity in experimental animals
and strong evidence in exposed humans that the agent acts
through a relevant mechanism of carcinogenicity, the agent
may be classified as group 1. Therefore, strong mechanistic
data may have an important role in evaluation of
carcinogenicity of agents when the evidence from human
cancer studies is less than sufficient.
human carcinogens identified in the IARC
monographs program
In the first 100 volumes of the monographs, 107 agents were
classified as human carcinogens (Table 1). These ‘agents’ include
single compounds and infectious agents [e.g. vinyl chloride [19]
and hepatitis B virus (HBV) [20]], mixtures (e.g. wood dust
[21]), and exposures in certain occupations (e.g. coke production
[21]). For 82 of these agents, the classification was mainly based
on positive evidence from ‘traditional’ epidemiologic studies,
mainly prospective cohort studies, in which exposures, outcomes,
and covariates (e.g. confounders) were assessed using
nonmolecular tools such as environmental measures, self-reports
(exposures and confounders), and medical records and death
Volume 24 | No. 4 | April 2013
reviews
Annals of Oncology
Table 1. Agents classified in group 1 (IARC monographs volumes 1–100),
by type of evidence
Type of evidence
Direct evidence from traditional epidemiology
Direct evidence from molecular epidemiology
Mechanistic evidence from molecular epidemiology and
other relevant data
Total
Biological agent
Established mechanistic events
A
key
study
Epstein–Barr virus
Cell proliferation, inhibition of
apoptosis, genomic instability,
and cell migration
Inflammation, liver cirrhosis, and
chronic hepatitis
Inflammation, liver cirrhosis, and
liver fibrosis
Cell proliferation, inhibition of
apoptosis, genomic instability,
and cell migration
Immunosuppression
(indirect action)
Immortalization, genomic
instability, inhibition of DNA
damage response, and
antiapoptotic activity
Immortalization and
transformation of T cells
Inflammation, oxidative stress,
altered cellular turnover and
gene expression, methylation,
and mutation
–
Inflammation, oxidative stress,
and cell proliferation
Inflammation and oxidative stress
[80]
N
82
12a
13
107
a
11 biological agents.
certificates (outcomes). Tobacco smoking is a notable example of
this group of agents [22].
For an additional 12 agents, the evaluation of sufficient
evidence in humans was derived by the results of
epidemiologic studies, which used molecular markers to
measure the relevant exposures (and in some cases, potential
confounders). Eleven of these carcinogens are biological agents
(Table 2): in the most informative studies, assessment of
infection was based on laboratory methods, which justifies
their classification as molecular epidemiology studies. In some
instances, a relatively simple laboratory test had sufficient
sensitivity and specificity to allow an accurate estimate of the
association between the agent and cancer risk: a notable
example here is the use of HBV surface antigen as an indicator
of HBV infection in earlier biomarker studies of hepatocellular
carcinoma [23]. In other instances, however, early laboratory
methods were not sufficiently sensitive or specific to provide
conclusive evidence of a causal association. This was the case
of human papilloma virus infection, whose etiologic role in
cervical carcinogenesis became clear when polymerase chain
reaction (PCR)-based amplification was applied to large-scale
epidemiologic studies (Table 3) [24–26].
The only nonbiologic agent for which molecular
epidemiology played a critical role in exposure assessment and
its evaluation as an IARC group 1 carcinogen is aflatoxin. In
this case, the development of urinary markers of exposure, in
particular the aflatoxin B1-DNA adduct at N7 of guanine, and
its application in several prospective studies, in which samples
were collected at enrollment and stored during follow-up,
reduced exposure misclassification, and revealed a strong and
consistent association between aflatoxin and hepatocellular
carcinoma risk [27].
For the remaining 13 group 1 agents, the evidence of
carcinogenicity in humans was less than sufficient (Table 4).
Evidence for carcinogenicity in animals for one agent
(etoposide) was ‘inadequate’ and for the other 12 agents was
‘sufficient’. The classification of these 13 agents as established
human carcinogens was supported by strong mechanistic
evidence. This can be considered as a contribution of
molecular epidemiology: strong evidence of a relevant
mechanism of carcinogenicity has been shown in humans
exposed to 11 of these agents. For two dioxin-like compounds,
3,4,5,3⍰,4⍰-pentachlorobiphenyl and 2,3,4,7,8pentachlorodibenzofuran, evidence for similarity of their
carcinogenic action with that of 2,3,7,8-tetrachlorodibenzo-pdioxin (TCDD), an established carcinogen, was provided by
mechanistic studies in animals.
Volume 24 | No. 4 | April 2013
Table 2. Biological agents classified as human carcinogens (group 1)
Hepatitis B virus
Hepatitis C virus
Kaposi’s sarcoma herpes virus
Human immunodeficiency
virus, type 1
Human papillomavirus
Human T-cell lymphotrophic
virus, type-1
Helicobacter pylori
Clonorchis sinensis
Opisthorchis viverrini
Schistosoma haematobium
[81]
[82]
[83]
[84]
[24]
[85]
[86]
[87]
[88]
[89]
Table 3. Results of early case–control studies of cervical cancer using
different HPV assays
Assay
Odds ratio (95% CI)
Serological [25]
DNA, no PCR [26]
DNA with PCR [24]
5.7 (2.1, 15.4)
3.7 (3.0, 4.5)
46.2 (18.5, 115)
main classes of mechanisms of action of
carcinogens
Carcinogens induce their deleterious effects through various
mechanisms. Here, we only discuss the major mechanisms that
contributed to identification of carcinogens in the monograph
volume 100 series, as examples of such mechanisms of action.
We do not discuss the mechanistic evidence for the agents that
were classified as group 1 carcinogens because of sufficient
evidence from human studies. Also, our intention was not to
agree or disagree on how mechanistic data have been used in
the evaluations, but just to show them as examples of how
molecular epidemiology has been used to reach the
evaluations.
exposure and interaction with cellular targets: DNA and
protein adducts and ah receptor
A wide range of genotoxic carcinogens can covalently bind
with DNA and make DNA adducts. These adducts could
doi:10.1093/annonc/mds543 | 
reviews
Annals of Oncology
Table 4. Group 1 agents with less than sufficient evidence in humans but with strong mechanistic evidence in humans
Agents
Evidence from cancer
studies in humans
Evidence from cancer Mechanistic evidence in humansa
studies in animals
Acetaldehyde associated with
consumption of alcoholic
beverages
Areca nut
Inadequate
Sufficient
Inadequate
Sufficient
Aristolochic acid
Limited
Sufficient
Benzo[a]pyrene
Inadequate
Sufficient
Benzidine, dyes metabolized to
Inadequate
Sufficient
Ethylene oxide
Limited
Sufficient
Etoposide
Limited
Inadequate
4,4⍰-Methlenbis (2-chloroaniline)
(MOCA)
Inadequate
Sufficient
4-(N-Nitrosomethylamino)-1-(3pyridyl)1-butanone (NNK), N
⍰-Nitrosonornicotine (NNN)
Neutron radiation
Inadequate
Sufficient
Inadequate
Sufficient
3,4,5,3⍰,4⍰-Pentachlorobiphenyl
(PCB-126)
Inadequate
Sufficient
2,3,4,7,8-Pentachlorodibenzofuran
(2,3,4,7,8-PeCDF)
Inadequate
Sufficient
Ultraviolet radiation
Inadequateb
Sufficient
Very higher risk of certain upper aerodigestive tract cancers in
aldehyde dehydrogenase deficient populations in genetic
epidemiology studies
Genotoxicity, including sister chromatid exchange (SCE) and
micronuclei (MN) formation; induction of oral preneoplastic
disorders with high propensity to progress to malignancy; areca nut
is the primary ingredient in betel quid, a carcinogen in humans
Genotoxicity, including DNA adducts and A : T > T : A transversions
in TP53 in patients with severe renal nephropathy or urothelial
tumors
Genotoxicity, including SCE, MN formation, DNA damage and 8oxo-deoxyguanosine, and specific diolepoxide-induced DNA
adducts which cause TP53 and K-RAS mutations
Strong mechanistic evidence indicating that these dyes are converted
to benzidine, a carcinogen to humans, and produce DNA adducts
and other genotoxic effects similar to those of benzidine
Genotoxicity, including SCE, chromosomal aberrations, MN
formation, mutagenic and clastogenic activity in the lymphocytes of
exposed workers, and dose-related increase in ethylene oxidederived hemoglobin adducts in exposed humans
Genotoxicity, including SCE, aneuploidy, mixed lineage leukemia
(MLL) gene translocations
Genotoxicity similar to other aromatic amines that are known to
cause cancer of the urinary bladder in humans, including DNA
adducts, protein adducts, SCE, MN formation in urothelial cells of
exposed workers
Genotoxicity, including DNA and hemoglobin adducts. The uptake
and metabolic activation of NNK and NNN have been clearly
documented in smokeless tobacco users
Similar but more severe damage than from gamma rays, including
structural and numerical chromosomal aberrations (including
rings, dicentrics, and acentric fragments)
Strong evidence of action through the same Ah receptor-mediated
mechanism (based on animal studies) as 2,3,7,8tetrachlorodibenzo-p-dioxin (TCDD), a carcinogen in humans
Strong evidence of action through the same Ah receptor-mediated
mechanism (based on animal studies) as TCDD, a carcinogen in
humans
Specific C>T transition in TP53 in squamous cell carcinoma and solar
keratosis at DNA bases where known photoproducts could be
formed
a
Mechanistic evidence from animal studies is not included in this table (except for PBC-126 and 2,3,4,7,8-PeCDF).
There is sufficient evidence in humans for the carcinogenicity of solar radiation, the use of ultraviolet-emitting tanning devices, and welding (another
source of ultraviolet radiation).
b
indicate the biologically effective dose of carcinogens [28]. In
addition to exposure to external factors, several endogenous
factors, including oxidative stress, can lead to DNA adduct
formation [29]. The frequency of DNA adducts is also
influenced by DNA repair ability [28]. In animal models, DNA
adduct formation has been shown to be a required step (but
might not a sufficient step) in cancer development [30]. In
humans, DNA adducts have been also suggested as biomarkers
of cancer risk [27, 31]. DNA adducts and A : T → T : A
 | Boffetta and Islami
transversions in the TP53 gene similar to those in experimental
animals exposed to aristolochic acid have been observed in
urothelial tumors from aristolochic acid nephropathy patients
and Balkan endemic nephropathy patients; these two types of
nephropathy have very similar clinical and morphological
characteristics. This evidence contributed to classification of
aristolochic acid as an IARC group 1 carcinogen [32]. The
ability for formation of DNA adducts, in combination with
other mechanistic pathways, have also contributed to the
Volume 24 | No. 4 | April 2013
reviews
Annals of Oncology
classification of benzo[a]pyrene (B[a]P), 4,4⍰-methlenbis
(2-chloroaniline) (MOCA), 4-(N-nitrosomethylamino)-1-(3pyridyl)-1-butanone (NNK), and N⍰-nitrosonornicotine
(NNN) as IARC group 1 carcinogens [19, 22]. NNK and NNN
are the most abundant strong carcinogens in smokeless
tobacco; their uptake and metabolic activation has been clearly
documented in smokeless tobacco users.
Similar to DNA adducts, various carcinogens can lead to
protein adduct formation. Protein adducts are generally stable,
and unlike most DNA adducts, they are not enzymatically
repaired [33]. Protein adducts are suitable for being used as
biomarkers of biologically effective dose of exposures. For this
purpose, globin and albumin are usually used, because they
have relatively long lifetime and are obtainable from blood
samples. However, other biological samples, such as hair and
tissue, and other proteins, including histones and collagen,
may also be used. Although protein adducts are not
mechanistically involved in carcinogenesis, they can be good
surrogates for DNA adducts and have been linked to cancer
risk [34, 35]. Therefore, they have also been used for estimating
cancer risk [34, 36]. A dose-related increase in frequency of
ethylene oxide-derived hemoglobin adducts in humans and
rodents exposed to ethylene oxide and a dose-related increase
in the frequency of ethylene oxide-derived DNA adducts in
exposed rodents have been shown [19]. Ethylene oxide also has
mutagenic and clastogenic effects and can induce sister
chromatid exchange (SCE), chromosomal aberrations, and
micronucleus formation in the lymphocytes of exposed
workers, as well as heritable translocations in the germ cells of
exposed rodents [19]. The ability of ethylene oxide in
producing protein adducts was one of the mechanisms for
which this agent was classified as an IARC group 1 carcinogen
[19]. Protein adduct formation was also one of the several
mechanisms by which B[a]P was classified in the same
category [19].
The Aryl hydrocarbon receptor (Ahr) is a ligand-dependent
cytosolic transcription factor that is normally inactive.
Following ligand binding to certain chemicals, including
polychlorinated biphenyls (PCBs), B[a]P, and dioxin-like
compounds, Ahr translocate into the nucleus and dimerize
with AhR nuclear translocator, leading to changes in gene
transcription, gene expression, cell replication, and apoptosis
[37]. The complex of Ahr/AhR nuclear translocator also
induces activity of several drug-metabolizing enzymes, notably
xenobiotic phase I metabolizing enzymes, such as cytochrome
P450 (CYP) A1, CYP1A2, and CYP1B, and phase II enzymes,
such as glutathione S-transferase [38]. Ahr is involved in
tumor initiation and plays a role in tumor promotion and
perhaps in tumor progression [39]. There is strong evidence
that TCDD has carcinogenic effects in humans through
receptor-mediated mechanisms: the primary mechanism is the
modification of cell replication and apoptosis and subsequently
promotion of tumor development; the secondary mechanism
relates to increases of oxidative stress which leads to DNA
damage [19]. Based on this mechanistic evidence, TCDD was
upgraded as an IARC Group I carcinogen in one of the earlier
monographs [40]; in the evaluation of this compound in
monograph 100, the working group also concluded that there
was sufficient evidence for carcinogenicity of TCDD in humans
Volume 24 | No. 4 | April 2013
based on epidemiological studies [19]. There is strong evidence
supporting similar Ahr-mediated mechanisms for
carcinogenicity of 3,4,5,3⍰,4⍰-pentachlorobiphenyl and
2,3,4,7,8-pentachlorodibenzofuran in humans based on their
mechanisms of carcinogenicity in experimental animals [19],
which provided strong mechanistic evidence for classifying
these two compounds as IARC group 1 carcinogens [19].
specific genetic alterations: TP53 and k-RAS mutations
TP53 is a tumor suppressor gene that encodes a transcription
factor responding to several forms of cellular stress. The p53
protein is involved in the response pathways against genotoxic
DNA damage [41] and regulation of the activity of several
genes, including some of the genes involved in the cellular
cycle control (e.g. p21 gene). The mutated TP53 found in some
neoplasias is not able to activate p21, allowing for an
uncontrolled cell growth [42–44]. Somatic TP53 gene
alterations have been reported in many cancer types. The
frequency of alterations (5%–80%) depends on the type, stage,
and etiology of tumors [45]. The majority (75%) of TP53
mutations are missense substitutions; the frequency of other
main alterations is as following: frameshift insertions and
deletions (9%), nonsense mutations (7%), and silent mutations
(5%) [46]. In addition to endogenous factors, several
environmental factors, including ionizing radiation, ultraviolet
rays, and oxidizing and cytotoxic agents, can induce TP53
mutations [47]. Exposure to ultraviolet radiation induce
mutations in several genes in human cell model systems,
and mutations have been detected in several genes in human
tumors at DNA bases where known photoproducts could be
formed, including the TP53 gene in squamous cell carcinoma
of the skin and solar keratosis [48]. This supports a major role
of these mutations in carcinogenicity of ultraviolet radiation
and has contributed to classification of ultraviolet radiation as
an IARC group 1 carcinogen. Also, the ability in inducing
TP53 mutations was one of the mechanistic pathways that
contributed to aristolochic acid and B[a]P being upgraded to
IARC group 1 carcinogens [32].
RAS genes encode Ras proteins, which are among
Guanosine-5⍰-triphosphate-binding proteins [49]. Normal
Ras proteins are involved in the control of several cellular
functions, including proliferation, differentiation, and survival
[50]. In ∼30% of human cancers, activating mutations in RAS
have been reported [51]. This activating transformation is
implicated in various cancers. The major categories of RAS
genes, including K-, H-, and N-RAS, seem to be primarily
associated with different types of cancers, among which K-RAS
are mainly associated with lung, colon, pancreas, and biliary
track cancers [52]. K-RAS mutation can be induced by various
environmental exposures, including polycyclic aromatic
hydrocarbons (including B[a]P) and aristolochic acid [51, 53].
K-Ras mutations, as well as some other mechanistic pathways,
contributed to classification of B[a]P and aristolochic acid as
IARC group 1 carcinogens [19, 32].
structural genomic alterations
Several structural alternations in the genome have been used in
cancer studies. Some of these alternations, including SCE and
micronuclei (MN) formation, are unspecific, whereas some
doi:10.1093/annonc/mds543 | 
reviews
others, including mixed lineage leukemia (MLL) translocations,
are specific to certain cancer types.
SCE may occur during DNA replication. Two sister
chromatids break and rejoin with one another, which results in
symmetrical exchange of DNA segments between the sister
chromatids [54]. As the exchanges are symmetrical, SCEs
themselves do not necessarily cause adverse health outcomes,
and they are not considered mutations [55]. Although
mechanisms of SCE formation are not clear, SCEs can be
induced by UV light and a large number of genotoxic chemicals,
including tobacco use [56]. Genotoxic factors can induce high
SCE levels in vitro and in animal models, and quantification of
these levels is generally easy [54]. Therefore, SCE test has been
used to investigate the DNA damage in blood lymphocytes as an
indicator of exposure to genotoxic factors. However, high SCE
frequencies have not been shown as a predictor of cancer [57].
This may be attributable to high interindividual variability in
exposure to genotoxic factors and DNA repair capacity, the two
main factors that determine SCE levels [57].
MNs are small cytoplasmic bodies having a portion of
acentric chromosome or the whole chromosome, which were
not carried during the anaphase to the opposite poles, due to
defect in the repair of DNA breaks [58]. This results in missing
of part or whole chromosomes in the daughter cell. MN
formation is influenced by age, gender, diet, several lifestyle
factors, including exercise, alcohol drinking, and smoking, and
some genetic defects and hereditary disorders [59]. MN test
have shown a high reliability in identifying genome damage in
genotoxicity studies [60]. Association between MN frequency
and risk of several types of cancer has been reported in
epidemiological studies, including prospective studies [60, 61].
Induction of SCE and MN by an agent per se is not generally
considered as strong evidence for carcinogenicity of the agent,
but it may support carcinogenicity of agents in the presence of
other mechanistic evidence. SCE and MN formation have been
reported with exposure to many carcinogenic agents; here, we
present two examples. Areca nut chewing, alone or in
combination with other substances, is a practice in Southern
Asia [22]. Areca nut is the primary ingredient in betel quid,
which is a known carcinogen to humans [22]. Continuous local
irritation of buccal epithelial cells by areca nut can generate
chronic inflammation, oxidative stress, and cytokine production,
and subsequently genotoxic effects, including SCE and MN, in
exposed oral keratinocytes with uncontrolled proliferation and
hyperplastic/dysplastic lesions [22]. With long-term use, these
preneoplastic lesions may progress to malignancy. Based on this
evidence, areca nut chewing was upgraded from IARC
carcinogenicity group 2A to group 1 [22]. The mechanisms for
which MOCA was upgraded to a group 1 carcinogen were
genotoxic activities similar to those of other aromatic amines
that are known to cause bladder cancer in humans, including
DNA and protein adducts, SCE, and MN formation [19].
MLL protein is a histone methyltransferase that seems to be
involved in the epigenetic maintenance of transcriptional
memory [62]. Translocations in MLL gene (in 11q23) are
involved in human leukemia [63]. Overall, MLL translocations
are found in ∼10% of human acute leukemia cases, but there is
a large variation in this proportion depending on the type of
leukemia and the age of patients (reported in >70% of infant
 | Boffetta and Islami
Annals of Oncology
leukemia and in <10% of leukemia in older children) [64, 65].
There is sufficient evidence from cancer studies in human that
etoposide, a topoisomerase-II-inhibiting chemotherapy agent,
in combination with cisplatin and bleomycin can cause acute
myeloid leukemia [32]. Each of these chemotherapy agents are
genotoxic, with evidence of induction of DNA damage,
chromosomal aberrations, and aneuploidy. However,
etoposide alone has been classified as IARC group 1 human
carcinogens because, unlike the other two drugs, it can induce
chromosomal translocations affecting the MLL gene in
cultured mouse and human cells [32]. Among treatmentrelated 11q23 translocations in leukemia patients, ∼85% have
been treated with topoisomerase-II-inhibiting drugs, primarily
etoposide [66].
challenges and future perspectives
Molecular epidemiology has contributed to identification of
several carcinogens. However, with the exception of biological
agents, its contribution in terms of providing sufficiently strong
evidence as required by the IARC monographs has been
modest. Molecular epidemiology may provide evidence for
exposure to and biological effects of carcinogens. This can be
particularly helpful with mixed compounds, as identification of
the effect of individual exposures in traditional epidemiologic
studies could be challenging. As an example, molecular
epidemiology has provided information on exposures or has
been used for validation of exposure assessment methods in
occupational studies, in which individuals are usually exposed
to several compounds simultaneously, and these exposures
differ by job type, location, and time [6]. Also, as discussed
above, the carcinogenic effects of several such compounds have
been shown in molecular epidemiological studies.
Nevertheless, biological markers of exposure and
mechanisms of actions for many other carcinogens need to be
identified. Some cases of cancer may be associated with longterm, low-dose exposures to carcinogens [67, 68]. Owing to
temporal variability of exposures, assessment of biomarkers in
a single biological sample may not provide accurate
information about lifelong exposure to the respective
carcinogens. Even in cases of cancers associated with high
exposures in a relatively short time, samples may not be
collected in the most appropriate time for assessment of the
exposure. Therefore, attempts should be made to assess
exposures in repeated samples in different periods to have a
more comprehensive picture about patterns of exposure. The
concept of ‘exposome’, assessment of exposure from perinatal
period until adulthood [69, 70], has been introduced to address
the variability of exposure in epidemiological studies, which
also applies to molecular epidemiology.
Another challenge in molecular epidemiology is that its
contribution has mainly been limited to the identification of
carcinogens acting through a genotoxic mechanism, i.e. when
carcinogens induce DNA damage. There is increasing
understanding of the role of nongenotoxic carcinogens. Some
examples of such carcinogens are certain hormones (such as
17β-estradiol and androstenedione) [71, 72],
immunosuppressants (such as ciclosporine) [73, 74], metals
(such as arsenic) [75, 76], and overweight/obesity [77, 78].
Volume 24 | No. 4 | April 2013
Annals of Oncology
These carcinogens may act through different mechanisms,
including receptor modification, effects on immune system,
and epigenetic changes, but these mechanisms are not clear in
the majority of cases. The potential role of nongenotoxic
mechanisms in carcinogenicity of IARC group 1 carcinogens is
reviewed elsewhere [79]. Overall, as nongenotoxic mechanisms
of carcinogenicity are not well identified and their biomarkers
are sparse, the contribution of these mechanisms to
identification of human carcinogens has been limited. This
area of molecular epidemiology deserves priority through the
development of biological markers of nongenotoxic
mechanisms and alterations that are specific to carcinogens.
disclosure
The authors have declared no conflicts of interest.
references
1. Perera FP, Weinstein IB. Molecular epidemiology and carcinogen-DNA adduct
detection: new approaches to studies of human cancer causation. J Chronic Dis
1982; 35: 581–600.
2. Porta M, Malats N, Vioque J et al Incomplete overlapping of biological, clinical,
and environmental information in molecular epidemiological studies: a variety of
causes and a cascade of consequences. J Epidemiol Community Health 2002;
56: 734–738.
3. Lin BK, Clyne M, Walsh M et al Tracking the epidemiology of human genes in the
literature: the HuGE Published Literature database. Am J Epidemiol 2006; 164:
1–4.
4. Ugolini D, Puntoni R, Perera FP et al A bibliometric analysis of scientific production
in cancer molecular epidemiology. Carcinogenesis 2007; 28: 1774–1779.
5. Ambrosone CB, Harris CC. The development of molecular epidemiology to
elucidate cancer risk and prognosis: a historical perspective. Int J Mol Epidemiol
Genet 2010; 1: 84–91.
6. Boffetta P. Molecular epidemiology. J Intern Med 2000; 248: 447–454.
7. Fletcher O, Houlston RS. Architecture of inherited susceptibility to common
cancer. Nat Rev Cancer 2010; 10: 353–361.
8. Stadler ZK, Gallagher DJ, Thom P et al Genome-wide association studies of
cancer: principles and potential utility. Oncology 2010; 24: 629–637.
9. Doll R, HILL AB. Smoking and carcinoma of the lung; preliminary report. Br Med J
1950; 2: 739–748.
10. Wynder EL, Graham EA. Tobacco smoking as a possible etiologic factor in
bronchiogenic carcinoma; a study of 684 proved cases. J Am Med Assoc 1950;
143: 329–336.
11. Levin ML, Goldstein H, Gerhardt PR. Cancer and tobacco smoking; a preliminary
report. J Am Med Assoc 1950; 143: 336–338.
12. Breslow L, Hoaglin L, Rasmussen G et al Occupations and cigarette smoking as
factors in lung cancer. Am J Public Health Nations Health 1954; 44: 171–181.
13. Wynder EL, Bross IJ, Day E. A study of environmental factors in cancer of the
larynx. Cancer 1956; 9: 86–110.
14. World Cancer Research Fund/American Institute for Cancer Research. Food,
nutrition, and the prevention of cancer: a global perspective. Washington, DC:
American Institute for Cancer Research 1999.
15. World Cancer Research Fund/American Institute for Cancer Research. Food,
nutrition, physical activity, and the prevention of cancer: a global perspective.
Washington, DC: American Institute for Cancer Research 2007.
16. Boyle P, Boffetta P, Autier P. Diet, nutrition and cancer: public, media and
scientific confusion. Ann Oncol 2008; 19: 1665–1667.
17. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 1: Some inorganic substances, chlorinated hydrocarbons,
aromatic amines, N-nitroso compounds, and natural products. Lyon: IARC Press
1972.
Volume 24 | No. 4 | April 2013
reviews
18. IARC. Preamble. In: IARC Monographs on the Evaluation of Carcinogenic Risks to
Humans, Vol 99: Some aromatic amines, organic dyes, and related exposures.
Lyon: IARC Press 2010; 9–38.
19. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 100F: Chemical agents and related occupations. Lyon: IARC
Press 2012.
20. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 100B: Biological agents. Lyon: IARC Press 2012.
21. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 100C: Arsenic, metals, fibres, and dusts. Lyon: IARC Press
2012.
22. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 100E: Personal habits and indoor combustions. Lyon: IARC Press
2012.
23. Vogel CL, Anthony PP, Mody N et al Hepatitis-associated antigen in Ugandan
patients with hepatocellular carcinoma. Lancet 1970; 2: 621–624.
24. Munoz N, Bosch FX, de Sanjose S et al The causal link between human
papillomavirus and invasive cervical cancer: a population-based case-control
study in Colombia and Spain. Int J Cancer 1992; 52: 743–749.
25. Wideroff L, Schiffman MH, Nonnenmacher B et al Evaluation of seroreactivity to
human papillomavirus type 16 virus-like particles in an incident case-control
study of cervical neoplasia. J Infect Dis 1995; 172: 1425–1430.
26. Reeves WC, Brinton LA, Garcia M et al Human papillomavirus infection
and cervical cancer in Latin America. N Engl J Med 1989;
320: 1437–1441.
27. Ross RK, Yuan JM, Yu MC et al Urinary aflatoxin biomarkers and risk of
hepatocellular carcinoma. Lancet 1992; 339: 943–946.
28. Vineis P, Perera F. DNA adducts as markers of exposure to carcinogens and risk
of cancer. Int J Cancer 2000; 88: 325–328.
29. Swenberg JA, Lu K, Moeller BC et al Endogenous versus exogenous DNA
adducts: their role in carcinogenesis, epidemiology, and risk assessment. Toxicol
Sci 2011; 120(Suppl 1): S130–S145.
30. Poirier MC. DNA adducts as exposure biomarkers and indicators of cancer risk.
Environ Health Perspect 1997; 105(Suppl 4): 907–912.
31. Veglia F, Loft S, Matullo G et al DNA adducts and cancer risk in prospective
studies: a pooled analysis and a meta-analysis. Carcinogenesis 2008; 29:
932–936.
32. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 100A: Pharmaceuticals. Lyon: IARC Press 2012.
33. Farmer PB. Exposure biomarkers for the study of toxicological impact on
carcinogenic processes. IARC Sci Publ 2004; 157: 71–90.
34. Bryant MS, Vineis P, Skipper PL et al Hemoglobin adducts of aromatic amines:
associations with smoking status and type of tobacco. Proc Natl Acad Sci USA
1988; 85: 9788–9791.
35. Phillips DH. Smoking-related DNA and protein adducts in human tissues.
Carcinogenesis 2002; 23: 1979–2004.
36. Tornqvist M, Ehrenberg L. Estimation of cancer risk caused by environmental
chemicals based on in vivo dose measurement. J Environ Pathol Toxicol Oncol
2001; 20: 263–271.
37. Chiba T, Uchi H, Yasukawa F et al Role of the arylhydrocarbon receptor in lung
disease. Int Arch Allergy Immunol 2011; 155(Suppl 1): 129–134.
38. Stockinger B, Hirota K, Duarte J et al External influences on the immune system
via activation of the aryl hydrocarbon receptor. Semin Immunol 2011; 23:
99–105.
39. Dietrich C, Kaina B. The aryl hydrocarbon receptor (AhR) in the regulation
of cell-cell contact and tumor growth. Carcinogenesis 2010; 31:
1319–1328.
40. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 69: Polychlorinated dibenzo-para-dioxins and polychlorinated
dibenzofurans. Lyon: IARC Press 1997.
41. Evans HJ, Prosser J. Tumor-suppressor genes: cardinal factors in inherited
predisposition to human cancers. Environ Health Perspect 1992; 98: 25–37.
42. Harris CC. p53 tumor suppressor gene: from the basic research laboratory to
the clinic–an abridged historical perspective. Carcinogenesis 1996; 17:
1187–1198.
doi:10.1093/annonc/mds543 | 
reviews
43. Semenza JC, Weasel LH. Molecular epidemiology in environmental health: the
potential of tumor suppressor gene p53 as a biomarker. Environ Health Perspect
1997; 105(Suppl 1): 155–163.
44. Szymanska K, Hainaut P. TP53 and mutations in human cancer. Acta Biochim
Pol 2003; 50: 231–238.
45. Petitjean A, Mathe E, Kato S et al Impact of mutant p53 functional properties on
TP53 mutation patterns and tumor phenotype: lessons from recent developments
in the IARC TP53 database. Hum Mutat 2007; 28: 622–629.
46. Olivier M, Eeles R, Hollstein M et al The IARC TP53 database: new online
mutation analysis and recommendations to users. Hum Mutat 2002; 19:
607–614.
47. Hainaut P, Hollstein M. p53 and human cancer: the first ten thousand mutations.
Adv Cancer Res 2000; 77: 81–137.
48. IARC Working Group. IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, Vol 100D: Radiation. Lyon: IARC Press 2012.
49. Arozarena I, Calvo F, Crespo P. Ras, an actor on many stages: posttranslational
modifications, localization, and site-specified events. Genes Cancer 2011; 2:
182–194.
50. Malumbres M, Barbacid M. RAS oncogenes: the first 30 years. Nat Rev Cancer
2003; 3: 459–465.
51. Alguacil J, Porta M, Kauppinen T et al Occupational exposure to dyes, metals,
polycyclic aromatic hydrocarbons and other agents and K-ras activation in human
exocrine pancreatic cancer. Int J Cancer 2003; 107: 635–641.
52. Karnoub AE, Weinberg RA. Ras oncogenes: split personalities. Nat Rev Mol Cell
Biol 2008; 9: 517–531.
53. Schmeiser HH, Janssen JW, Lyons J et al Aristolochic acid activates ras genes in
rat tumors at deoxyadenosine residues. Cancer Res 1990; 50: 5464–5469.
54. Garcia-Sagredo JM. Fifty years of cytogenetics: a parallel view of the evolution of
cytogenetics and genotoxicology. Biochim Biophys Acta 2008; 1779: 363–375.
55. Wilson DM, III, Thompson LH. Molecular mechanisms of sister-chromatid
exchange. Mutat Res 2007; 616: 11–23.
56. van Poppel G, de Vogel N, van Balderen PJ et al Increased cytogenetic damage
in smokers deficient in glutathione S-transferase isozyme mu. Carcinogenesis
1992; 13: 303–305.
57. Norppa H, Bonassi S, Hansteen IL et al Chromosomal aberrations and SCEs as
biomarkers of cancer risk. Mutat Res 2006; 600: 37–45.
58. Fenech M, Kirsch-Volders M, Natarajan AT et al Molecular mechanisms of
micronucleus, nucleoplasmic bridge and nuclear bud formation in mammalian
and human cells. Mutagenesis 2011; 26: 125–132.
59. Fenech M, Bonassi S. The effect of age, gender, diet and lifestyle on DNA
damage measured using micronucleus frequency in human peripheral blood
lymphocytes. Mutagenesis 2011; 26: 43–49.
60. Bonassi S, El Zein R, Bolognesi C et al Micronuclei frequency in peripheral blood
lymphocytes and cancer risk: evidence from human studies. Mutagenesis 2011;
26: 93–100.
61. Bonassi S, Znaor A, Ceppi M et al An increased micronucleus frequency in
peripheral blood lymphocytes predicts the risk of cancer in humans.
Carcinogenesis 2007; 28: 625–631.
62. Guenther MG, Jenner RG, Chevalier B et al Global and Hox-specific roles for the
MLL1 methyltransferase. Proc Natl Acad Sci USA 2005; 102: 8603–8608.
63. Liu H, Cheng EH, Hsieh JJ. MLL fusions: pathways to leukemia. Cancer Biol Ther
2009; 8: 1204–1211.
64. Huret JL, Dessen P, Bernheim A. An atlas of chromosomes in hematological
malignancies. Example: 11q23 and MLL partners. Leukemia 2001; 15:
987–989.
65. Krivtsov AV, Armstrong SA. MLL translocations, histone modifications and
leukaemia stem-cell development. Nat Rev Cancer 2007; 7: 823–833.
66. Bloomfield CD, Archer KJ, Mrozek K et al 11q23 balanced chromosome
aberrations in treatment-related myelodysplastic syndromes and acute leukemia:
report from an international workshop. Genes Chromosomes Cancer 2002; 33:
362–378.
67. Taylor R, Najafi F, Dobson A. Meta-analysis of studies of passive smoking and
lung cancer: effects of study type and continent. Int J Epidemiol 2007; 36:
1048–1059.
 | Boffetta and Islami
Annals of Oncology
68. Darby S, Hill D, Deo H et al Residential radon and lung cancer-detailed results of
a collaborative analysis of individual data on 7148 persons with lung cancer and
14,208 persons without lung cancer from 13 epidemiologic studies in Europe.
Scand J Work Environ Health 2006; 32: 1–83.
69. Wild CP. Complementing the genome with an ‘exposome’: the outstanding
challenge of environmental exposure measurement in molecular epidemiology.
Cancer Epidemiol Biomarkers Prev 2005; 14: 1847–1850.
70. Rappaport SM. Implications of the exposome for exposure science. J Expo Sci
Environ Epidemiol 2011; 21: 5–9.
71. Spink BC, Bennett JA, Pentecost BT et al Long-term estrogen exposure promotes
carcinogen bioactivation, induces persistent changes in gene expression, and
enhances the tumorigenicity of MCF-7 human breast cancer cells. Toxicol Appl
Pharmacol 2009; 240: 355–366.
72. National Toxicology Program Technical Report. Toxicology and carcinogenesis
studies of androstenedione (CAS No. 63-05-8) in F344/N rats and
B6C3F1 mice (gavage studies). Natl Toxicol Program Tech Rep Ser 2010; 560:
1–171.
73. van Kesteren PC, Beems RB, Luijten M et al DNA repair-deficient Xpa/p53
knockout mice are sensitive to the non-genotoxic carcinogen cyclosporine A:
escape of initiated cells from immunosurveillance? Carcinogenesis 2009; 30:
538–543.
74. Bugelski PJ, Volk A, Walker MR et al Critical review of preclinical approaches to
evaluate the potential of immunosuppressive drugs to influence human
neoplasia. Int J Toxicol 2010; 29: 435–466.
75. Ren X, McHale CM, Skibola CF et al An emerging role for epigenetic
dysregulation in arsenic toxicity and carcinogenesis. Environ Health Perspect
2011; 119: 11–19.
76. Singh AP, Goel RK, Kaur T. Mechanisms pertaining to arsenic toxicity. Toxicol Int
2011; 18: 87–93.
77. Moon HS, Chamberland JP, Aronis K et al Direct role of adiponectin and
adiponectin receptors in endometrial cancer: in vitro and ex vivo studies in
humans. Mol Cancer Ther 2011; 10: 2234–2243.
78. Kant P, Hull MA. Excess body weight and obesity–the link with gastrointestinal
and hepatobiliary cancer. Nat Rev Gastroenterol Hepatol 2011; 8: 224–238.
79. Hernandez LG, van Steeg H, Luijten M et al Mechanisms of non-genotoxic
carcinogens and importance of a weight of evidence approach. Mutat Res 2009;
682: 94–109.
80. Henle G, Henle W, Klein G et al Antibodies to early Epstein-Barr virus-induced
antigens in Burkitt’s lymphoma. J Natl Cancer Inst 1971; 46: 861–871.
81. Vogel CL, Anthony PP, Sadikali F et al Hepatitis-associated antigen and antibody
in hepatocellular carcinoma: results of a continuing study. J Natl Cancer Inst
1972; 48: 1583–1588.
82. Seeff LB, Buskell-Bales Z, Wright EC et al Long-term mortality after transfusionassociated non-A, non-B hepatitis. The National Heart, Lung, and Blood Institute
Study Group. N Engl J Med 1992; 327: 1906–1911.
83. Chang Y, Cesarman E, Pessin MS et al Identification of herpesvirus-like DNA
sequences in AIDS-associated Kaposi’s sarcoma. Science 1994; 266:
1865–1869.
84. Bayley AC, Downing RG, Cheingsong-Popov R et al HTLV-III serology
distinguishes atypical and endemic Kaposi’s sarcoma in Africa. Lancet 1985; 1:
359–361.
85. The T- and B-Cell Malignancy Study Group. Statistical analyses of clinicopathological, virological and epidemiological data on lymphoid malignancies with
special reference to adult T-cell leukemia/lymphoma: a report of the second
nationwide study of Japan. Jpn J Clin Oncol 1985; 15: 517–535.
86. Nomura A, Stemmermann GN, Chyou PH et al Helicobacter pylori infection and
gastric carcinoma among Japanese Americans in Hawaii. N Engl J Med 1991;
325: 1132–1136.
87. Kim Y, Yang DH, Chang KR. Relationship between clonorchis sinensis infestation
and cholangiocarcinoma of the liver in Korea. Seoul J Med 1974; 15: 247–255.
88. Parkin DM, Srivatanakul P, Khlat M et al Liver cancer in Thailand. I. A casecontrol study of cholangiocarcinoma. Int J Cancer 1991; 48: 323–328.
89. Makhyoun NA. Smoking and bladder cancer in Egypt. Br J Cancer 1974; 30:
577–581.
Volume 24 | No. 4 | April 2013