This meta-analysis of 30 case-control studies assessed the association of polymorphisms in DNA damage response genes with oral, pharyngeal and laryngeal cancer risk. A previous review by Vineis et al.(Vineis et al., 2009
), evaluated the association of variants in DNA repair genes and cancer susceptibility in general, without in-depth analysis of head and neck cancer, given the broad scope of their paper. Here, we provide an updated systematic revision of the literature analyzing a larger number of studies and genetic polymorphisms. We have also reported results according to race and head and neck subsite, when possible.
There are three major pathways involved in DNA repair, depending on the type and magnitude of the damage. First, the base excision repair (BER) pathway repairs small base modifications, including oxidatively-induced lesions and single-strand breaks (SSBs), through exposure of the cells to reactive oxygen species (ROS), an endogenous toxic agent. For this pathway, we report results for three polymorphisms in the XRCC1
gene. The nucleotide excision repair (NER) pathway removes a broader spectrum of genomic damage, including bulky adducts induced by large polycyclic aromatic hydrocarbons, such as those present in benzo[α]pyrene in cigarette smoke, and crosslinks caused by UV-light photoproducts and chemotherapeutic agents. We have evaluated eleven polymorphisms in nucleotide excision repair genes XPA, XPC, XPD, XPF
. Finally, single (SSBs) and double strand breaks (DSBs), endogenously produced by reactive oxygen species among other factors, can undergo either an error-prone (by non-homologous DNA end joining) or an error-free (by homologous recombination) repair process(Hakem, 2008
). For this pathway, we have evaluated one polymorphism in the XRCC3
Although there is little evidence about the direct influence of genetic polymorphisms on the functionality of the BER pathway, recent publications with conflicting results have addressed the association between various polymorphisms in BER genes, such as XRCC1
, and the risk of oral, pharyngeal and laryngeal cancer. Similar to Vineis et al.(Vineis et al., 2009
), our meta-analysis revealed an almost two-fold statistically significant increased association between the XRCC1
codon 194 homozygous T/T variant and oral, pharyngeal and laryngeal cancer. We also report that this statistically significant two-fold increased risk was observed for Asian populations and for Asian oral cavity cancer cases. Comparison of the meta-ORs between Asians and Caucasians was not possible, since the Caucasian studies included more than one head and neck subsite, while the Asian studies were more homogeneous and included oral cavity cancer cases only. Therefore, studies that investigate this association between XRCC1
and cancer according to race and head and neck subsite are warranted.
is an important component in the BER, because it has the ability to interact with and serves as a scaffold for other key proteins that are responsible for strand incision at the DNA damage site, as well as DNA polymerase β and DNA ligase III, responsible for synthesis and re-joining of the DNA strand break, respectively(Altieri et al., 2008
). Although the functional impact of the XRCC1
codon 194 polymorphisms remains unknown since it was first reported(Shen et al., 1998
), it is plausible that changes in amino acid sequence at conserved sites may alter the functionality of the protein. This eventually could lead to a defective BER pathway, increased genomic instability and cancer risk.
XPC, XPA and XPD play important roles the nucleotide excision repair pathway. We observed marginal significant increased associations between XPD Asp312Asn heterozygotes and combined variants, as well as the XPC-PAT homozygous variant and the risk of oral, pharyngeal and laryngeal cancer. Our findings contrast with those reported by Vineis et al.(Vineis et al., 2009
) and Manuguerra et al.(Manuguerra et al., 2006
), who found no association. However, our meta-analysis included twice as many studies for each of these polymorphisms. Although no associations were seen between the XPA
5′-UTR homozygous and XPD
codon 156 variants, for Caucasians, a marginally increased association and a significantly inverse association were observed, respectively. XPC is responsible for the detection of the DNA damage lesion, while XPA and XPD, along with other proteins are responsible for the local unwinding of the DNA helix and the demarcation of the lesion. The formation of the open complex enables incorporation of endonucleases to excise the damaged site and further gap filling and sealing by DNA polymerase δ and ligase I, respectively(Altieri et al., 2008
). It has been reported that the XPD Asp312Asn variant in smokers is significantly correlated with increased aromatic DNA adduct levels(Hou et al., 2002
), while another study found decreased DNA damage-induced apoptosis in lymphoblastoid cells(Seker et al., 2001
). Although the effect of the XPD
codon 156 variant on this pathway is unknown, based on our findings, it would be interesting to determine whether this polymorphism provides a gain of function on the XPD protein activity. The functional implication of the XPA
5′-UTR (A23G) and XPC-PAT
variants are unknown. The NER pathway is responsible for removing bulky adducts generated from cigarette smoke, among other environmental carcinogens(Altieri et al., 2008
). Cigarette smoke is one of the primary risk factors for head and neck cancer, leading to chromosomal instability(Reshmi and Gollin, 2005
). Thus, further investigation of these polymorphisms in the context of tobacco dose is needed.
It is also important to consider our findings in the context of Human Papillomavirus (HPV), an additional independent risk factor for head and neck cancer. HPV distribution in head and neck cancer seems to be subsite-specific and associated with improved outcome. It has been reported that HPV is mainly distributed in the oropharynx, with the highest distribution in the tonsils(Ragin and Taioli, 2007
). Patients with HPV-positive tumors are less likely to have subsequent tumors, recurrences, metastases and new primary tumors, which contrast with what is observed in patients with HPV-negative tumors(Ragin et al., 2004
), and distinct molecular profiles are observed between HPV-positive and HPV-negative tumors(Ragin et al., 2006
). Therefore, we were interested in exploring whether there was an association between DNA repair gene polymorphisms by anatomic sub-site and HPV status. This analysis could not be performed due to the lack of the studies reporting on gene polymorphisms by anatomic sub-site or HPV status. Further studies are needed to address this interesting question.
A recent report describes an unequal burden of head and neck cancer in the US, in which the disparities were greater in African-American males, who showed a higher incidence and mortality rate for head and neck cancer compared to Caucasians(Goodwin et al., 2008
). Also, African-Americans have been reported to have a younger age of onset compared to Caucasians(Gourin and Podolsky, 2006
), and a greater likelihood to be smokers(Arbes et al., 1999
). Therefore, it is important to have a better understanding of health disparities in minority populations by knowing whether genetic polymorphisms can identify high-risk individuals in the population who could be targeted with chemoprevention strategies. Surprisingly, in our meta-analysis, just one study by Shen et al.(Shen et al., 2001
) reported genetic polymorphisms by race (non-Hispanic Whites, African-Americans, Hispanic-Americans, and native Chinese) without finding significant ethnic differences among the four groups. We have also observed a lack of publications concerning African-Americans or individuals of African descent while evaluating other gene polymorphisms(Ragin et al., 2010
). Future assessments of genetic polymorphisms in the DNA repair pathway in minority populations are needed.
There are some limitations to this meta-analysis. First, the majority of the studies did not report gene polymorphism by sub-site and smoking status. Therefore, we were unable to perform stratification by those variables, which may explain some negative results. Second, heterogeneity due to ethnic ancestry (mostly Caucasians and Asians) and the small number of studies per ethnic group for the majority of the gene polymorphisms may have limited the ability of this meta-analysis to find true associations. Nevertheless, while performing a summary estimate, an average of each OR is weighted for the precision of each study, thus reducing the possibility of a biased estimate. Furthermore, despite performing stratification by race, when possible, to further assess heterogeneity, at times heterogeneity could not be resolved possibly due to the variation in the PCR methodology (PCR-RFLP, sequencing, melting curve analysis, 5′-exonuclease assay, MALDITOF-MS) employed in some of the studies in the same subgroups. In addition, subsite-specific analyses could only be performed for oral cavity cases, but these studies were only found in Asian populations. Therefore, the level of heterogeneity according to head and neck subsite in each racial group was not comparable. Third, despite conducting a meta-analysis with an almost overall absence of publication bias, it was only observed for the XRCC1 exon 9 codon 280 homozygous variant. We have not included any unpublished data, which may lead to false-positive results and/or bias. The source of publication bias for that particular variant remains unknown. Despite these limitations, the current meta-analysis has also some advantages. First the overall number of studies and genetic polymorphisms included were consistently large, compared to previously conducted meta-analysis, which significantly increased the statistical power of the analysis. Second, each of the studies included in the meta-analysis met our inclusion criteria. Third, we did not detect publication bias in the overall estimate that yielded statistically significant associations; which indicates that the pooled results should be unbiased.