ASSOCIATION OF COMMON ALLELES OF SMALL EFFECT (POLYMORPHISMS) WITH LUNG CANCER RISK

Results of hundreds of studies using association analysis to
evaluate the effects of various polymorphisms, in metabolic
genes, growth factors, growth factor receptors, markers of DNA
damage and repair and genomic instability, and in oncogenes
and tumor suppressor loci have been published. Many of these
studies have yielded inconsistent results. The effects of risk alleles
at these loci are expected to be individually small, and they
may interact with smoking and/or other loci to increase lung
cancer risk. Two recent reviews 114,115 can help the reader obtain
an overview of these studies. On account of the relatively weak
effects of these polymorpisms, relatively few consistent replications
of effects have been provided, but more recent genomewide
association studies of very large collections of samples have
provided some highly significant and reproducible results.
Three manuscripts jointly appeared in Nature 116,117 and
Nature Genetics 118 identifying the same region of chromosome
15q as associating highly significantly with lung cancer
risk. The region that was implicated by these studies includes
a neuronal nicotinic acetylcholine receptor gene cluster comprising
the CHRNA3, CHRNA5, and CHRNB4 subunits.
Nicotinic receptors are comprised of pentamers that include
alpha and beta units, and are ubiquitously expressed, but have
higher levels in the brain. The manuscript of Thorgeirsson
initially scanned a population of 14,000 individuals who had
provided information about their smoking histories. This study
identified the 15q region as associated with smoking quantity
and then further explored the regions effects on smoking dependence
and lung cancer risk. The other two studies 116,118
started with large collections of lung cancer cases and control
samples. Results of all studies reached a surprisingly homogeneous
conclusion with increased lung cancer risks of about 1.29
among individuals carrying a heterozygous mutation (44.2% of
controls for rs8034191) in the region and about 1.80 among
individuals with homozygous variants (10.7% of controls).
Because of strong linkage disequilibrium among the markers
studied, the specific gene causing increased risks for three studies
drew conflicting conclusions about the relevance of this
region on smoking behavior and its influence on lung cancer
risk. Thorgeirsson 117 claimed that all of the risk for lung cancer
in this region appeared likely to be explained by the regions
effects on smoking behavior. Amos 118 found an association of
this region with both lung cancer risk and smoking behavior
but found stronger effects on lung cancer risk that remained
highly significant after adjusting for smoking behavior. Finally,
Hung 116 did not find any association of this region on lung
cancer risk. None of the studies had enough nonsmoking lung
cancer cases to draw strong conclusions, but a subsequent study
focusing on nonsmokers 118 fails to find any risk associated with
variants in this region and lung cancer risk in this population.
Comparing models of the risks of lung cancer among current
smokers, Hung 116 found that carriers of the common less susceptible
allele had a 14% cumulative risk of lung cancer death
compared with a 23% cumulative risk of lung cancer death
among those homozygous for the higher-risk variant.
Additional genetic factors are being identified and replicated
by using very large collections of lung cancer cases and
controls. Polymorphisms in DNA repair genes are repeatedly
associated with lung cancer risk, and these are reviewed, 119–121
but many of the individual studies reporting these associations
are of moderate size. A rare variant of CHEK2 (I157T) that
had previously been associated with increased risk for breast
cancer was found in a study of 4015 tobacco-associated cancer
cases and 3052 controls to be strongly protective for the
development of lung or head and neck cancers (RR 0.44),
but a risk factor for kidney cancer (RR 1.44). Functional
studies of biomarkers of DNA repair and mutagen sensitivity
following exposure to clastogens repeatedly show that these are
reliable predictors of lung cancer risk, 119 but applying them
for patient populations is problematic because of a lack of reference
laboratories and the need to study viable cells.
All of these lines of evidence suggest that there may be one or
several genes causing inherited increased risk to lung cancer in
the general population. Although association studies have given
evidence that alleles at various genetic loci may influence lung
cancer risk, there has frequently been disagreement between
studies. The first linkage study of lung cancer has given significant
evidence of linkage to a region on chromosome 6q.
If a susceptibility locus is identified in this region, it will be of
major public health importance because it will allow identification
of individuals at especially high risk who can then be
targeted for intensive efforts at environmental risk reduction.
In addition, identification of such a gene will lead to better
understanding of the mechanism of carcinogenesis in general,
perhaps eventually leading to better methods of prevention and
treatment. The recent identification of polymorphisms associated
with lung cancer risk provides new targets for potential
interventions for chemoprevention, but further study is needed
to evaluate these new findings and to identify particularly highrisk
subjects who might benefit most from such interventions.
Confirmation of a genetic predisposition for lung cancer
can be obtained by finding evidence for linkage of the putative
susceptibility gene(s) to genetic marker loci in a specific
chromosomal region(s). One potential problem in the search
for such a linkage is heterogeneity. There are different types
of heterogeneity of this disease and of its etiological factors:
(a) there is heterogeneity at the level of histological types of
lung cancer; (b) there is heterogeneity at the level of exposure
to various environmental risk factors; and (c) there could be
heterogeneity at the level of inherited susceptibility loci, that
is, there could be one locus involved in susceptibility for one
family and a different locus involved in susceptibility for another
family. All of these types of heterogeneity could possibly
confound the identification of a susceptibility locus (or loci)
for lung cancer. The suggestive evidence in the published linkage
study 113 for susceptibility loci at several other regions of
the genome support the possibility of locus heterogeneity in
lung cancer.
If, through linkage and positional cloning techniques, a
genetic locus or loci that contributes to inheritable risk for
lung cancer can be identified, or one of the candidate loci suggested
to modify risk by association studies can be confirmed
as a susceptibility locus, then the effects of the alleles at this
locus and its interaction with cigarette smoking and the other
well-known environmental risk factors for lung cancer can be
elucidated with much more accuracy than presently possible,
and our understanding of lung carcinogenesis in general may
be increased.