GENE–ENVIRONMENT INTERACTIONS

Both genetic and environmental factors affect lung cancer risks,
but the molecular pathophysiology of gene–environment interactions
is complex. The genes influencing cancer susceptibility
may consist of heterogeneous alleles at one locus or a combination
of alleles at multiple loci. In a study of familial aggregation of
lung cancer, Tokuhata and Lilienfeld 151 reported a significantly
increased risk of lung cancer mortality among nonsmoking
relatives of lung cancer cases when compared with nonsmoking
relatives of age-, race-, and sex-matched controls. Kreuzer et al. 5
concluded that lung cancer in a first-degree relative was associated
with a 2.6-fold increase in risk of lung cancer in cases diagnosed
in patients younger than 50 years of age. A similar pattern of
familial aggregation limited to probands with lung cancer diagnosed
at a younger age than the median age in the general population
was reported by Bromen et al. 152 In a segregation analysis
involving 337 high-risk families with lung cancer, Sellers et al. 153
described a pattern of autosomal codominant inheritance, and
hypothesized that segregation at the putative gene locus would
account for 69% of the lung cancer cases diagnosed in persons
up to age 50 years. Samet et al. 154 concluded that the personal
risk of lung cancer was increased more than fivefold if at least
one parent had lung cancer. In a study of families of women with
lung cancer, an odds ratio gradient was noted: never-smoker with
a positive family history (5.7), smoker with a negative family history
(15.1), and smoker with a positive family history (30.0). 155
Familial aggregation of lung cancer may be attributed to shared
exposures to tobacco smoking or other environmental and/or
heritable determinants. On the assumption that a lung cancer
susceptibility gene with a frequency of 0.3 to 0.5 would increase
the risk of lung cancer in carriers, then an autosomal recessive
model of inheritance would predict that siblings of cases would
manifest a twofold to fourfold increased RR of lung cancer. 156
Multiple inherited and acquired mechanisms of susceptibility
to lung cancer have been proposed. Individual susceptibility
to tobacco-induced lung cancer may be dependent on
competitive gene–enzyme interactions that affect activation
or detoxification of procarcinogens and levels of DNA adduct
formation, or on the integrity of endogenous mechanisms for
repairing lesions in DNA. 157 Nicotine is converted to cotinine
in a two-step enzymatic process for which the rate- limiting
step is the drug-metabolizing enzyme, cytochrome P450, a
genetically polymorphic enzyme. Glucuronyl transferase enzymes
conjugate and inactivate carcinogenic compounds,
including 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone
(NNK), the tobacco-specific methylnitrosamino metabolite,
a potent procarcinogen in tobacco smoke. In a case-control
study by Nakachi et al. 158 in which they assessed DNA polymorphisms
in the cytochrome P4501A1 gene in relationship
to squamous cell lung carcinoma, persons with the susceptible
genotype had a RR of 7.31 after adjusting for cigarette smoking
history. DNA polymorphisms in the cytochrome P450
gene (CYP1A1), or aromatic hydrocarbon hydroxylase, which
is responsible for the metabolic activation of benzo(a)pyrene
and other polyaromatic hydrocarbons, may represent a locus
of a susceptibility gene for lung cancer. Increased activity of
CYP1A1 has been demonstrated in lung cancer cells when
compared with normal tissue in the same patient, suggesting
that dysregulation of the gene may occur in carcinogenesis.
However, no association between lung cancer and CYP1A1
polymorphisms was reported in studies in Finland conducted
by Hirvonen and coworkers. 159,160
The genetically controlled ability to metabolize the antihypertensive
agent debrisoquine has been linked to the risk of lung
cancer. The P450 gene (CYP2D6) that regulates debrisoquine
metabolism appears to influence the metabolism of nicotine to
cotinine and metabolic activation of NNK, which is a potent carcinogen
in experimental animals. Those who are extensive metabolizers
in the hydroxylation of a 10-mg test dose of debrisoquine,
a dominant trait affecting up to 90% of the U.S. white population,
have been characterized as being at increased risk of lung
cancer. Caporaso et al. 161 had initially estimated the smokingadjusted
RR to be increased sixfold, but more recently, studies
have suggested more modest increases that are twofold or have
failed to demonstrate an association using either the debrisoquine
metabolic phenotype or polymerase chain reaction assays for
detecting the genotype. 162 Various phase II detoxification systems
serve to modulate risk in relation to cumulative levels of exposure
to chemical metabolites. GST alleles encode a family of enzymes
that catalyze the conjugation of electrophilic substrates. 163,164
Inherited genetic traits can influence an individual’s smoking
addictive behavior. The candidate genes affecting smoking
behavior include the dopamine receptors, dopamine and serotonin
transporter alleles, and the cytochrome P450 alleles (e.g.,
CYP2A6). These genetic factors collectively influence binding
and metabolism of nicotine and other neurotransmitters. 165
Several case-control studies have suggested that subjects with deficiency
of the GST- isoenzyme or the GSTM1 null genotype,
may have a 10% to 60% increase in lung cancer risk. Metabolites
of constituents of cigarette smoke, including polycyclic aromatic
hydrocarbons, aryl amines, and nitrosamines are potential substrates
for GSTM1. Some studies have also evaluated potential interactions
between CYP1A1 and GSTM1 genotypes. 166 Studies
in Japan have reported that subjects with the combined GSTM1
null genotype and CYP1A1 polymorphisms were at increased
risk. 167 The risk appears to be greater than additive in cigarette
smokers with homozygous deletions of GSTM1 and CYP1A1
polymorphisms. Alexandrov et al. 168 noted that with both variant
genes, the concentration levels of benzo(a)pyrene diol epoxide
adducts of DNA were increased in lung parenchyma.
It is now clear that human tumors result from a complex
sequence of mutational events. The bronchial epithelium of sustained
smokers progresses from squamous metaplasia, to dysplasia,
to invasive carcinoma, which is accompanied by progressive
genomic instability. Many of the genetic defects that have been
described in somatic cells of lung neoplasms are acquired during
adult life and are related to exposures to environmental carcinogens.
Some genetic events, however, are inherited and are present
in all somatic cells. Mechanistic interactions of genes and exogenous
agents may result from environmental agents altering the
expression of genes involved in the regulation of the cell cycle, intercellular
signaling, cell cycle arrest, and apoptosis. Susceptibility
genes, in addition, include those concerned with the fidelity of
DNA repair, DNA replication, and genomic stability. Individuals
with combinations of alleles that dysfunctionally controlled enzyme
systems regulating activation or detoxification pathways
may be at increased risk of lung cancer when exposed to even
low dose levels of tobacco smoke or other mutagens. However,
the validity and efficiency of screening for carcinogenic metabolites
in predicting human lung cancer risk is questionable in the
context of a population. Strategic targeting of phenotypic or genotypic
testing as a cancer control measure in high-risk families,
in conjunction with behavioral counseling, may be more costeffective.
In a cohort study of monozygotic and dizygotic twin
pairs followed in the National Academy of Sciences/National
Research Council Twin Registry, it was concluded that inherited
predisposition was not demonstrable in relationship to smokinginduced
lung cancer diagnosed in men older than 50 years. 169 If
one were to assume that 50% of lung cancer deaths before the
age of 50 result from genetic predispositions, this would represent
only 5% of lung cancer deaths.