Xuanwei, China poses a unique opportunity to assess COPD susceptibility in a population with substantial in-home coal smoke exposure. The primary source of indoor air pollution in Xuanwei is smoke from burning coal for heating and cooking. Xuanwei’s COPD rate is more than twice the national average in China, and is strongly associated with coal use (3;5). In Xuanwei, 90% of residents are farmers with minimal industrial and automotive air pollution exposure (6). Nearly all women and few men cook, and most men and nearly no women smoke tobacco (7).

Genotyping was performed on DNA extracted from sputum samples via phenol-chloroform extraction (9). High throughput genotyping was successful for 53 cases and 111 controls with an Oligo Pool (OPA) by the Illumina GoldenGate Assay (www.illumina.com) at the National Cancer Institute’s Core Genotyping Facility (Gaithersburg, MD). There was insufficient DNA available for 10 controls. Four additional controls were excluded for missing demographic data. Duplicate samples (n = 18) of both cases and controls were randomly distributed throughout study plates to ensure quality control and determine the intra-subject concordance rate for all assays (>98%). Initially, 1,442 SNPs from the SNP500Cancer (http://snp500cancer.nci.nih.gov) database in over 400 genes were genotyped successfully. Candidate SNPs were selected based on their potentially relevance to cancer and other human diseases or they had possible functional significance, based on previously identified candidate genes. Hardy-Weinberg equilibrium (HWE) for each SNP was tested in controls with a Pearson χ² test or a Fischer’s exact test if any of the cell counts were less than five. Exclusion of 165 SNPs with low minor allele frequency (<0.01) and 16 SNPs with substantial deviation from HWE (p <0.001) left 1,261 SNPs in 380 genes for analysis.

The association between each gene and COPD was assessed by the minP test in MatLab, which assesses the significance of the minimal p-value in each gene using a permutation-based resampling procedure (20,000 permutations) that takes into account the number of SNPs genotyped in each gene region and their underlying linkage disequilibrium structure (10). False Discovery Rates (FDRs) were calculated to evaluate the significance of the minP results (11). Unconditional logistic regression was used to estimate the odds ratio (OR) and 95% confidence interval (CI) for the association between COPD risk and each SNP independently, using the homozygote of the common allele as the reference group and adjusting for age (<55 years, ≥55 years), sex, smoking (0 pack-years, 0< pack-years <25, ≥25 pack-years), and lifetime coal use (<130 tons, ≥130 tons). Dose effects for each SNP were estimated by a linear trend test by coding the genotypes based on the number of variant alleles present (0, 1, 2). Finally, interactions between the dominant genetic models for the PTEN SNPs and lifetime coal use (<130 tons, ≥130 tons), as well as smoking (0 pack-years, 0< pack-years <25, ≥25 pack-years), were tested on the multiplicative scale, adjusting for age, sex, smoking, and coal exposure.

Gene-based minP analysis identified 22 genes significantly associated with COPD risk (minP < 0.05) (Table 2). After adjusting for multiple comparisons, PTEN was the most significantly associated gene with COPD risk (FDR = 0.019). POLB (FDR = 0.076) and ARNT (FDR = 0.22) were also highly associated with COPD risk.

PTEN, or phosphatase and tensin homolog, is an important regulator of cell cycle progression and cellular survival via the AKT signaling pathway (12;13). In fact, PTEN deletion is the most common mechanism of inappropriate AKT activation in human malignancy (14). PTEN negatively regulates the AKT pathway and is essential to homeostatis (15;16). As such, AKT and PTEN expression have been observed to be inversely correlated in lung cancer tissue samples (17). While this is the first report, to the best of our knowledge, of specific PTEN SNPs associated with COPD, we previously reported the association between PTEN rs1903858 and lung cancer risk, in our parallel study in Xuanwei, which utilized the same controls (18). The inverse relationship observed in our studies between COPD and lung cancer and PTEN rs1903858 suggests that the variant may play an important role in respiratory diseases.

Smoke-induced mutations have been hypothesized to be associated with chronic inflammation of the airways and lung that accompanies COPD (19). PTEN is one such gene that is frequently mutated or deleted in the epithelium of smokers (20). In mice, PTEN expression play a crucial role in mediating airway inflammation and responsiveness for asthma (21). Further, PTEN overexpression also reduced airway hyperresponsiveness in mice (22).

Although smoking is an established risk factor for COPD, it was only slightly associated with increased risk of COPD in our analyses. Our results are similar to a cohort study in Xuanwei, which also observed comparable smoking habits between cases and the general population (23). Similarly, our lung cancer studies in Xuanwei find only a slight increased risk of lung cancer associated with smoking (24). The relatively weak effect of smoking on respiratory diseases in Xuanwei may result from the indoor air pollution levels attributed to smoky coal combustion in overwhelming the effect of tobacco smoke. Given that smoking is an established risk factor for COPD, and that smoky coal use has been associated with COPD in the cohort study in Xuanwei, it was important to explore the interactions between the significant PTEN SNPs and these important environmental risk factors, even though the statistical power was limited (<10%).

One of the major strengths of our population-based case-control study is the high participation rates. Further, our results are biologically plausible given that PTEN could contribute to COPD. Since the small sample size may lead to false positive and false negative findings (25), our findings should be viewed as hypothesis-generating. However, we adjusted the gene-based analyses for multiple comparisons to minimize the possibility of spurious findings. Due to our small sample size, we had limited power to detect modest risks at low minor allele frequencies (MAFs); however, we did have ample power (>80%) to detect risks of 2.0 and 2.5 for MAFs of ≥0.3 and ≥0.1, respectively. Specific SNP associations should be cautiously interpreted until these results are replicated, since SNP associations may be attributed to another SNP in linkage disequilibrium. Although the functionality of each SNP is not known, PTEN rs701848 is likely to alter PTEN expression and/or function provided that it is located in the 3′ translational termination codon. Further limitations of our study pertain to generalizability. As our analysis included only Asians, the generalizability of our findings to other ethnic groups is limited. The allelic frequency variation by ethnicity in HapMap found that the T allele at rs701848 is only 56% among individuals of Asian descent, compared to 81% for African descent. Similarly, the A allele at rs1903858 is only 44% among individuals of Asian descent, compared to 73% and 58% for European and African descent, respectively. Finally, our findings may not generalize to all COPD cases since disease degree of severity was not available for our analysis; however, it is likely that our study population consists of individuals with moderate to serve COPD as they sought treatment for symptoms in a hospital setting. While some controls may have had mild undiagnosed COPD, this would have underestimated our results.