| Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous in the environment and several of the compounds in its class are known carcinogens. However, the cellular response to PAH exposure is complex. Polymorphisms in metabolizing and conjugating genes may modify the level of DNA damage resulting from the direct (DNA adducts) and indirect (oxidized DNA bases) action by PAHs. In addition, polymorphisms in DNA repair genes may alter an individual's ability to repair this damage. Evaluation of the association between prostate cancer and factors involved in PAH metabolism, conjugation and DNA repair using conventional "one-at-a-time" approaches has produced equivocal findings. Furthermore, although quantitative kinetic models have been attempted, the pathways involved in the cellular response to PAHs and their interrelations remain incompletely characterized in humans, restricting accurate construction of gene regulatory networks. In an attempt to overcome these current methodological limitations, we present an approach that allows for the simultaneous evaluation of multiple sources of PAHs and key candidate genes (single nucleotide polymorphisms, SNPs) involved in PAH metabolism, conjugation and DNA repair pathways. This approach uses hierarchical latent variable path analysis (HLVPA) to generate DNA Damage from PAHs and DNA Repair Capacity Indices (composite scores); that account for the relevant pathway interrelations, followed by a regression framework that enables statistical inference about their role in prostate cancer. We compare this approach to conventional "one-at-a-time" and joint regression approaches in two empirical evaluations. Although we observed significant associations between individual factors involved in PAH metabolism (e.g., mEH A9970G SNP), PAH conjugation (e.g., GSTM1 +/- SNP), base excision repair (e.g., OGG1 C6803G SNP) and prostate cancer risk, the HLVPA approach was also able to reveal novel pathway and source level relationships such as the strong inverse association between the interaction of the Total PAH Body Burden Index and the Base Excision Repair Capacity Index and prostate cancer risk, lending new insights into the mechanisms involved in prostate cancer. Although interpretation of these interactions beyond a one-unit change is not straightforward, this approach provides a more holistic, structured evaluation of the numerous factors involved in complex diseases and can serve to supplement current methods. |
