|Home | About | Journals | Submit | Contact Us | Français|
Michael Campos, M.D., Division of Pulmonary, Critical Care and Sleep Medicine (R-47), University of Miami Miller School of Medicine, 1600 N.W. 10th Ave., RMSB 7052, Miami, FL 33136, Phone: (305) 243-3045; Fax: (305) 243-6992, ude.imaim.dem@1sopmacm
The marked variability in individual susceptibility to the detrimental effects of smoking on lung function and findings suggest a significant genetic contribution to COPD, which has been demonstrated in several studies. The only known genetic risk factor for COPD, severe alpha 1 antitrypsin (AAT) deficiency, explains only 1–2% of cases of this disease. Screening for severe AAT should be conducted in all cases of COPD. Intravenous augmentation therapy should be combined with currently recommended treatment modalities for COPD when treating patients with COPD due to severe AAT deficiency. There is considerable interest in identifying susceptibility genes for COPD unrelated to severe AAT deficiency, as this could greatly enhance current efforts to prevent, diagnose and treat this disease by yielding novel insights into its pathogenesis. Genome-wide association studies (GWAS) of COPD and its intermediate phenotypes (e.g., lung function measures) have identified novel susceptibility loci for COPD. Some of these susceptibility loci may also influence lung function in the general population (e.g., HHIP and FAM13A), while others may affect not only COPD but other diseases related to smoking behavior (e.g., CHRNA3/CHRNA5). Although much work remains to be done, recent advances and the implementation of novel approaches to study COPD genetics (e.g., sequencing) and epigenetics are promising, and could have a profound impact on COPD management.
Severe alpha 1-antitrypsin (AAT) deficiency is the only well-established genetic risk factor for chronic obstructive pulmonary disease (COPD). Patients with severe AAT deficiency (most commonly, Protease Inhibitor [PI] Z) are at increased risk for developing COPD, particularly if they smoke1–2. However, severe AAT deficiency explains only a small proportion of cases of COPD (1–2%), and genes other than PI type (genetic modifiers) likely influence lung function in PI Z subjects3–4.
Burrows et al. noted that the development of chronic airflow obstruction in response to cigarette smoking is highly variable in the general population, suggesting that individuals vary in their genetic susceptibility to the detrimental effects of smoking on lung function5. Several studies have since confirmed a genetic contribution to the pathogenesis of COPD in cigarette smokers without severe AAT deficiency6–8. A search for COPD-susceptibility genes other than PI type has included genome-wide linkage analyses, candidate-gene studies and, more recently, genome-wide association studies (GWAS).
In this chapter, we review the genetics, diagnosis and treatment of severe AAT deficiency. We then assess findings from recent genetic studies of COPD and its intermediate phenotypes (e.g., lung function). Finally, we briefly discuss future directions in this field.
Severe deficiency of AAT was first recognized as a risk factor for COPD in 19639. Since then, many studies have characterized the genetics and role of this protein in the pathogenesis of COPD.
The AAT protein is encoded by a 12.2 kb gene located on chromosome 14q32.1 called SERPINA1 or PI10. The gene has seven exons and six introns, is inherited in an autosomal codominant fashion, and has more than 120 single nucleotide polymorphisms (SNPs)11. Numerous AAT protein variants can be differentiated by their speed of migration on gel electrophoresis using isoelectric focusing. The most common alleles are M, S and Z. The M variants (M1, M2, M3) result in proteins with a medium rate of migration and a normal level of AAT, the S variant is associated with mild reductions in serum AAT level, and the Z variant has the slowest rate of migration and leads to severe reduction of AAT level. Null alleles also occur, with undetectable protein levels. Combinations of the M, S and Z variants are seen in >95% of the population. Serum AAT levels in subjects with the MZ and ZZ phenotypes are 60% and 10%, respectively, of those in (normal) subjects with the MM phenotype.
The most common forms of severe AAT deficiency with a high risk of COPD involve combinations of two Z alleles (Glu342Lys) and a Z allele with a null allele (both referred as phenotype PI*Z), or a combination of two null alleles. Combinations of a Z allele and an S allele (Glu264Val) also confer a moderate risk of COPD12. The phenotype MZ has also been associated with a lower but still increased risk of COPD, while the MS phenotype has been associated with only mild reductions in AAT level and no risk of COPD13. AAT deficiency has been described in all races, although the highest frequencies of the Z allele have been described in whites. In the United States, the frequency of the ZZ phenotype is 1 in 4775, SZ phenotype 1 in 1124 and MZ phenotype 1 in 36 individuals14. Worldwide, severe AAT deficiency affects approximately 3.4 million individuals15.
SERPINA1 codes a 52 kDa glycoprotein with 394 amino acids that contains a reactive loop with an active site at methionine 358. Although AAT is an acute phase reactant synthesized mostly by liver cells, there also is local synthesis by cells such as neutrophils, monocytes, macrophages and epithelial cells16–17. The major function of AAT is the neutralization of serine proteases, particularly neutrophil elastase (NE); other targets include cathepsin G, trypsin and proteinase-318–19. Severe deficiency of AAT results in excess protease activity in the lungs, particularly during periods of inflammation, and leads to progressive degradation of the lung parenchyma (emphysema) and accelerated decline in lung function over time20. Although protease-antiprotease imbalance is the most important cause of COPD in severe AAT deficiency, the loss of other functions of AAT likely contribute in COPD pathogenesis. These include endothelial cell protection against apoptosis by binding to caspases21, regulation of airway epithelial lining fluid balance by binding to matryptase22, and inflammatory response modulation. For example, AAT modulates endotoxin-induced inflammation23, reduces TNF-alpha-induced lung injury in rabbits24, inhibits superoxide production by neutrophils25 and regulates the response of macrophages to pro-inflammatory stimuli26.
Z-AAT has reduced capacity to inhibit NE27, as well as a conformational protein change that enables a loop-sheet polymerization process and its accumulation in the endoplasmic reticulum (ER) with loss of secretion into the circulation28. The Z-polymers are degraded by ER-associated degradation pathways via proteasomes and by autophagy as part of the ER overload response29. If these cellular mechanisms fail, gain-of-toxic function and ER stress occur, which can lead to inflammation via NFkB activation30, hepatocyte death by apoptosis31 and –ultimately- liver disease, both during childhood and later in life; a similar process may occur in the lung but has not been as well characterized.
AAT polymers can also further worsen protease-antiprotease imbalance. When instilled in the trachea of mice, Z-polymers produce a significant concentration dependent influx of neutrophils that is not mediated by chemokines32. The polymers can be detected in the circulation33 and in bronchoalveolar lavage fluid of patients with severe AAT deficiency, in whom they may also induce inflammation by a direct effect on epithelial cells34.
Compared to young subjects who have a PI*ZZ phenotype but do not smoke, those who have a PI*ZZ phenotype and do smoke have significantly lower FEV1/FVC and diffusing capacity of carbon monoxide35. At age 30 years, Pi*Z subjects who smoke have significantly more shortness of breath, sputum production and wheezing than Pi*MM smokers; among nonsmokers, Pi*ZZ subjects report only more wheezing than non-affected controls35.
Not uncommonly, clinicians stereotype the presentation of severe AAT deficiency as that of a young (<40 years old) white non-smoker who presents with COPD and pan-acinar emphysema predominantly in the lower lobes. Although some affected individuals present this way, most symptomatic individuals have features that resemble those of COPD not due to severe AAT deficiency. About a third of affected subjects have a more indolent presentation and are diagnosed after age 50 years36, about a third have emphysema predominantly located in the upper lobes37 and more than 80% have a significant smoking history38; bronchiectasis of variable severity are also common39. Current guidelines thus recommend screening for severe AAT deficiency in all patients with COPD40. Screening of relatives of affected subjects may find nonsmokers with the PI*ZZ phenotype but no respiratory symptoms41.
The significant variability observed in the development, progression and manifestations of COPD due to severe AAT deficiency strongly suggests that genetic and/or environmental factors modify disease expression. Studies of familial aggregation and heritability suggest that genetic factors other than PI type influence lung function and airflow obstruction in PI*Z individuals3, 42. Although no genetic modifiers of severe AAT deficiency have been confidently identified, candidate genes have been examined. In a case-control study, two coding polymorphisms in NOS3 were associated with severe airflow obstruction in PI*Z individuals43. In a family-based association study of 10 genes previously associated with asthma and/or COPD, variants in the genes for TNF and IL10 were associated with lung function measures44.
In subjects with severe AAT deficiency, the development of airflow obstruction is associated with age, male sex, bronchodilator responsiveness and chronic bronchitis, and -most strongly- cigarette smoke45–47.
A detailed description of how to treat lung disease in severe AAT deficiency is beyond the scope of this chapter. The pharmacologic and non-pharmacologic treatment of COPD due to AATD follows the same guidelines as those for COPD in general48 Interventions showing benefit in subjects with severe AAT deficiency include disease management programs (which improve quality of life and decrease healthcare utilization49), lung volume reduction surgery (which improve 6-minute walking test and dyspnea scores50) and inhaled steroids (which reduce airflow obstruction and hyperinflation51.
Intravenous augmentation therapy with donor-derived purified AAT is the only FDA-approved specific treatment for lung disease due to severe AAT deficiency (defined as a baseline serum AAT level below 11 µM)40. The recommended dose (60 mg/kg once a week) is intended to keep trough serum AAT above 11 µM. Studies supporting this therapy are mostly observational52. Whereas results from a recent meta-analysis of observational studies suggest that augmentation therapy with AAT slows lung function decline in subjects with moderately reduced lung function53, a meta-analysis of the only two published randomized placebo-controlled trials concluded that there was not sufficient evidence to recommend this therapy54. In spite of this controversy, augmentation therapy should be used while the effectiveness, optimal dosage, therapeutic goals and target populations for this treatment are further studied. Additional potential treatments for severe AAT deficiency are being examined, including gene therapy55 and correction of the underlying genetic defect using inducible stem cells56.
Findings from studies of candidate genes for COPD susceptibility, reviewed in detail elsewhere57, have yielded additional insight into our understanding of COPD pathobiology, particularly for certain pathways (e.g., TGF-beta, matrix metalloproteinases)57–58. In contrast to studies of one or few genes selected on the basis of known biology, GWAS are hypothesis-free studies that leverage information from markers genotyped along the entire genome, and can thus yield unanticipated discoveries and insights into disease pathogenesis. Findings from recent GWAS and other new approaches to study COPD genetics are the focus of this review.
Published GWAS of COPD are summarized in Table 1. In 2009, Pillai and colleagues performed the first GWAS of COPD using a multi-stage replication design59. The discovery (primary or initial) cohort comprised 823 subjects with COPD (cases) and 810 unaffected smokers (controls) from Bergen, Norway. The 100 SNPs with the lowest P values in the GWAS of this cohort were then tested for association with COPD in 1,891 members of 606 white families from the family-based International COPD Genetics Network (ICGN). Seven of the 100 SNPs tested in ICGN showed significant evidence of association and were then tested for association with COPD in 389 cases from the National Emphysema Treatment Trial (NETT) and 472 smoking controls from the Normative Aging Study (NAS). Six of these seven SNPs were also tested for association with lung function measures in 949 members of 127 families in the Boston Early-Onset COPD Study (BEOCOPD)60. Two SNPs (rs8034191 and rs1051730) in the α-nicotinic acetylcholine receptor (CHRNA3/CHRNA5/IREB2) locus on chromosome (chr.) 15 showed significant evidence of association with COPD susceptibility in ICGN and NETT-NAS, and reached genome-wide (GW) statistical significance in the analysis of three combined cohorts (Norway, ICGN and NETT-NAS) but were not GW significant in the discovery cohort. Nominal evidence of significant associations with FEV1 was also noted in BEOCOPD (P=0.03 for each polymorphism) and ICGN (P= 1.04 × 10−4 for rs8034191 and 1.75 × 10−5 for rs1051730). Polymorphisms in the gene for hedgehog interacting protein (HHIP) on chromosome 4 were consistently replicated but did not reach GW statistical significance in the discovery cohort or the combined analysis. The hedgehog signaling pathway has received continued attention (see below) due to its involvement in branching morphogenesis of the lung61.
Family with sequence similarity 13, member A (FAM13A) was identified as a susceptibility gene for COPD in a GWAS of a cohort comprising 4,320 subjects enrolled in three case-control studies: Norway, NETT- NAS, and the multicenter Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE62)63. Replication was then attempted in 502 cases and 504 controls from COPDGene (a multicenter study of the genetics and epidemiology of COPD)64, BEOCOPD, and ICGN. A polymorphism in FAM13A (rs7671167) on chr. 4q22.1 was associated with COPD in the discovery cohort and in two of the three replication cohorts (combined P value= 1.2 × 10−11, combined odds ratio (OR) in the case-control studies=0.76, 95% confidence interval [CI]=0.69–0.83). This SNP was also associated with pre-bronchodilator FEV1 in BEOCOPD (P = 0.02) and with pre- (P=5.3 × 10−5) and post-bronchodilator FEV1 in ICGN.
Intermediate phenotypes offer several advantages for genetic studies of complex diseases such as COPD because they are objectively defined and may be influenced by fewer genes than the disease per se. Kong et al. conducted a meta-analysis of GWAS results from the Norway, ECLIPSE and NETT cohorts on percent emphysema from computed tomography (defined by densitometry as the percentage of lung voxels at −950 Hounsfield Units, and also qualitatively by visual scoring)65. Though no SNPs in any of the three cohorts were significantly associated with emphysema, a SNP (rs10844154) in the gene for bicaudal homolog 1 (BICD1) on chr. 12 reached GW significance for association in the meta-analysis of emphysema (OR for at least mild emphysema=1.46, P= 5.2 × 10−7 and OR for at least moderate emphysema=1.56, P=4.8 × 10−8) in computed tomography (CT). Interestingly, the strongest signal in this study came from radiologist scoring rather than density mask analysis though the authors stressed that this did not imply superiority of one method over the other. The authors noted that variants in BICD1 are associated with length of telomeres, suggesting that a mechanism linked to accelerated aging may be involved in the pathogenesis of emphysema.
Five SNPs in loci previously associated with COPD (HHIP, CHRNA3/CHRNA5/IREB2, and FAM13A), were tested for association with COPD-related phenotypes (smoking behavior, lung function, body mass index [BMI], fat-free body mass, BODE index66, emphysema and airway wall thickness determined by CT) in ECLIPSE and then validated in ICGN67. A SNP in the CHRNA3/CHRNA5 locus (rs8034191) was associated with increased smoking intensity (expressed as pack-years), radiologist’s assessment of emphysema on high-resolution CT of the chest, and airflow obstruction in the ECLIPSE and ICGN cohorts. In ECLIPSE, subjects with COPD who were current or former smokers and homozygous for the rs8034191 risk allele had 7.5 more cumulative pack-years of smoking (P=0.002) than those who were heterozygous or homozygous for the non-risk allele; this association was confirmed in the ICGN. SNPs in HHIP or FAM13A loci were not associated with smoking intensity. A SNP in HHIP was associated with FEV1/FVC in the ECLIPSE and ICGN cohorts; this SNP was also significantly associated with fat-free body mass and COPD exacerbations in ECLIPSE. FAM13A SNPs were associated with lung function but this association was not consistently significant across cohorts. These findings, taken together with others, suggest that: 1) the CHRNA3/CHRNA5 locus, which has also been associated lung cancer68, influences COPD-related phenotypes –at least partly-through its effects on smoking behavior69; 2) the HHIP locus, which has also been associated with lung cancer70 and lung function in subjects with asthma71, has effects on lung function and the systemic component(s) of COPD; and 3) FAM13A influences lung function and airflow obstruction.
Cachexia, which is common in subjects with advanced COPD, is associated with increased severity of airflow obstruction and increased mortality. Wan et al. conducted a GWAS of BMI in 2,950 subjects with COPD in three cohorts (ECLIPSE, Norway and NETT), with replication attempted in 502 subjects from COPDGene72. A GWAS of fat-free mass index (FFMI) was also conducted in the ECLIPSE and Norway cohorts. SNP rs8050136, located in the intron of the fat mass and obesity–associated gene (FTO) was significantly associated with BMI (P=4.97 × 10−7) and FFMI (p = 1.19 × 10−7) in the discovery cohort. Findings for BMI were replicated in COPDGene (P=6 × 10−3). These findings suggest that FMO influences anthropometric measures in subjects with COPD.
Spirometric measures of lung function such as FEV1 and FEV1/FVC are key intermediate phenotypes of COPD. Thus, some genes that influence lung function in the general population may also be relevant to the pathogenesis of COPD. Wilk and colleagues conducted a GWAS of lung function measures in 7,691 participants in the Framingham Heart Study (FHS) with validation in an independent cohort of 835 subjects in the Family Heart Study that was enriched for airflow obstruction73. Four SNPs in tight linkage disequilibrium (e.g., highly correlated) on chr. 4q31were significantly associated with FEV1/FVC percent predicted in the discovery (FHS) cohort. The association between one of the four SNPs (rs13147758) and FEV1/FVC was replicated in the Family Heart Study, in which significant associations were also shown with FEV1 and binary airflow obstruction phenotypes (particularly in smokers). The associated SNPs were not in a gene transcript but were near HHIP, a candidate gene for COPD susceptibility (see above).
In a meta-analysis of GWAS results for lung function in 20,890 participants from four CHARGE consortium studies (Atherosclerosis Risk in Communities, Cardiovascular Heath Study, FHS, and the Rotterdam Study), HHIP and FAM13A (see above) were among the eight loci associated with FEV1/FVC (the other six were GPR126, ADAM19, AGER-PPT2, PTCH1, PID1, and HTR4) at or near the threshold for GW statistical significance74. In a separate study, a GWAS of lung function measures was conducted in 20,288 subjects of European ancestry in the Spirometa Consortium (discovery cohort); this GWAS was then followed by a meta-analysis of data for the top signals from the discovery cohort and 32,184 subjects in the CHARGE consortium, as well as in silico summary association data from 21,209 individuals from the CHARGE consortium and 883 individuals in the Health 2000 Survey75. This study confirmed the previously reported locus on chromosome 4q31 near HHIP (see above) and identified five novel loci for FEV1 or FEV1/FVC: 2q35 in TNS1, 4q24 in GSTCD, 5q33 in HTR4, 6p21 in AGER and 15q23 in THSD4.
Two recent studies have assessed whether loci identified by GWAS on lung function also influence COPD. Soler-Artigas and colleagues studied a large sample of subjects of European ancestry (including individuals with and without COPD) and constructed a risk score including six SNPs in HTR4, GSTCD, TNS1, AGER, THSD4, and near HHIP76. Compared to subjects in a common baseline group, those in the highest risk category (estimated as approximately 5% or Europeans) had a 1.6-fold increased risk of developing COPD76. In another study, Castaldi et al tested whether 32 SNPs in or near 11 loci associated with lung function in prior GWAS were associated with COPD in 5,362 subjects in four cohorts (NETT-NAS, ECLIPSE, Norway, and the first 1000 subjects in COPDGene)77. Of the previously identified susceptibility loci for lung function, three genomic regions harbored polymorphisms associated with susceptibility to COPD at a 5% false discovery rate: the FLJ20184/INTS12/GSTCD/NPNT locus on chromosome 4q24, the chromosome 6p21 locus including AGER and PPT2, and the chromosome 5q33 locus that includes ADAM19.
As with other complex diseases, GWAS have identified susceptibility loci that explain a modest fraction of the heritability (the proportion of variation in a phenotype due to genetic factors) of COPD, with nominal explanation of disease risk78. Additional work is needed for functional characterization and full assessment of variation in these susceptibility loci, as well as for understanding their individual and combined (e.g., gene-by-gene and gene-by-environment [i.e., smoking]) effects on COPD-related phenotypes.
Beyond GWAS, additional approaches to the study of COPD genetics include sequencing exomes and/or the whole genome to identify rare susceptibility variants, and conducting studies of integrative genomics (where genetic variants are evaluated for their contribution to gene expression) and epigenetics (heritable changes in gene expression that occur without changes in DNA sequence).
Genetic variants that are uncommon (minor allele frequency [MAF]=1%–4%) or rare (MAF <1%) may confer greater susceptibility to COPD in certain individuals and/or ethnic groups. Because the genotyping platforms used for GWAS of COPD predominantly include variants with MAF ≥5%, they would not be able to detect these uncommon/rare variants with moderate to strong genetic effects. Next-generation sequencing technologies provide this capability79. Exome sequencing allows investigators to detect rare variants in protein-coding regions of the genome. Compared to whole-genome sequencing, exome sequencing is less expensive and thus allows studying a larger number of subjects80. Exome sequencing has been successful in identifying the genetic etiology of a rare Mendelian disorder81 and is currently being applied to the investigation of COPD. As prices of whole-genome sequencing drop, this will become a feasible and attractive method to detect rare variants with strong effects on COPD, particularly regulatory (non-coding) variants. Integration of data from GWAS and sequencing studies with those from studies of gene expression and proteomics should allow researchers to focus on the most promising candidate genes for COPD susceptibility.
Studying epigenetic mechanisms (including DNA methylation, histone modification, and micro-RNA) provides a unique opportunity to examine the potential impact of demographic, environmental and lifestyle factors (e.g., diet, aging, and cigarette smoking)82) on gene expression in the lung and COPD. Whole-genome studies of DNA methylation, the best characterized epigenetic mechanism, and COPD are now in progress.
Although much remains to be done, recent advances and the advent of new methodologies are promising and should yield increased understanding of the genetic and epigenetic mechanisms influencing the pathogenesis of COPD, both related and unrelated to severe AAT deficiency. Such understanding should ultimately be translated into novel approaches to prevent, diagnose and treat COPD.
Sources of support: Grant HL073373 from the U.S. National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.