Genetics and racial, ethnic, and gender characteristics of COPD


Chronic obstructive pulmonary disease (COPD) may be influenced by race, ethnicity, gender, and genetic factors. Limited data exist that compare COPD in different racial/ethnic groups; however, the available data suggest that differences in COPD may exist. Potential differences in COPD between racial/ ethnic groups include genetic and biological differences; disparities in diagnosis and treatment; increasing exposure to cigarette in nonwhite populations world wide; and a lack of enrollment of minorities in epidemiological and clinical trials. Gender appears to also influence COPD. Historically, men have had higher prevalence rates of COPD than women, but recent data suggests that women may actually be more susceptible to COPD. COPD in omen may have different characteristics than in men, and it may be more severe. Importantly, in the United States, more women now die of COPD than men.

The marked variability in lung function and risk for COPD in people with similar cigarette smoking histories, together with studies of familial aggregation, support an important role for genetic risk factors in COPD. A small but important fraction of COPD cases harbor a major genetic determinant, α1-antitrypsin deficiency (AATD). This condition is most common in populations of Northern European ancestry, although affected individuals in other populations can be found. Despite significant advances in diagnosis and treatment, AATD remains highly under diagnosed. Manifestations other than classic lower-lobe predominant emphysema can include bronchiectasis, liver disease, panniculitis, and vasculitis. Intravenous augmentation with ATT protein is a commonly used treatment for severe AATD; it may result in improved pulmonary outcomes, although randomized clinical trials that support the efficacy of this treatment are not available.

Other genetic determinants of COPD have been more difficult to establish, as has been the case with many other common complex diseases. A number of genetic methods have been used. The most common studies to date have been association analysis of pathophysiological candidate genes. These studies have had mixed results; the vast majority have not been well replicated, though a few genes, such as mircosomal epoxide hydrolase, glutathione S-transferase PI, and transforming growth factor-β1, have had more consistently positive results. Studies of specific COPD-related phenotypes using these candidate genes have also been performed, with varying success. Animal models have been instrumental in analyzing COPD-related molecular pathways and they may also provide another source of candidate genes for association analysis. Several human monogenic diseases include emphysema or emphysema-like characteristics; while accounting for an extremely small percentage of obstructive lung disease or emphysema in the general population, these conditions may provide insights into mechanisms that are relevant for more common types of COPD.

Finally, genome-wide approaches represent an unbiased approach to identify novel COPD susceptibility genes. One method is to identify quantitative trait loci in animals. Traditional linkage studies in humans, of which one has been performed in COPD, have led to the discovery of several genomic regions of interest and a novel COPD-associated gene, SERPINE2. Genome-wide association studies have greater power to identify smaller effects; while they have yet to be published in COPD, they have led to successes in other complex diseases.


Figure 1
Racial/ ethnic differences in COPD. Overall, limited data exist on the racial/ ethnic differences in COPD susceptibility. Smoking prevalence does vary by ethnic group; some evidence suggests that susceptibility to the impact of cigarette smoking may differ as well, which may be due to genetic or other environmental factors. Available data on treatment and outcomes suggest that disparities exist between racial/ ethnic groups, but such data is sparse.



Figure 2
Gender differences in COPD. Though the prevalence of smoking among women has historically been less than in men, female smoking rates have been on the rise, especially in developing counties. In developing countries, exposure to biomass can be an important risk factor for COPD, especially in women. Women dying of COPD in the United States now outnumber men. Data suggest that there are differences between men and women in susceptibility to cigarette smoke. Gender may affect clinical presentation and management, although further studies are needed. GOLD -Global Initiative on Obstructive Lung Disease; BOLD - Burden of Obstructive Lung Disease; ED- emergency department; FEV1 - Forced Expiratory Volume in 1 second.




Figure 3
Cigarette smoking prevalence by race and gender in the United States, 1992-2004. Among black men, white men, black women, and white women, smoking rates have historically been highest in black men and lowest in black women, although rates of smoking are declining among all four groups.




Figure 4
Self-reported United States chronic obstructive pulmonary disease (COPD) prevalence by race and gender, 1980-2000. Data are age adjusted to the 2000 United States population. The prevalence of COPD based on self-reports of emphysema or chronic bronchitis has generally been higher among whites and women.




Figure 5
United States COPD death rates by gender, 1960-2000. COPD is the leading cause of death in the United States for which the absolute death rate is rising. The age-adjusted death rate between 1980 and 2000 increased by 67% overall; during this period, the increase in the COPD death rate was higher among blacks (87%) than whites and tripled among women. By 2000, the absolute number of women dying from COPD was higher than for men. Increases observed in 1999 may be the result of a different International Classification of Disease system.




Figure 6
Worldwide prevalence of COPD. The Prevalence of COPD in 2007 in different geographic regions based on the Burden of Obstructive Lung Disease data is presented based on results for specific cities within the countries listed. The reason for these regional differences are not entirely clear and may be due to different environmental exposures (smoking, biomass, occupation), as well as possible genetic factors.









Figure 7
Mechanism of protease inhibition and polymerization by α1- antitrypsin (ATT). A. The mechanism of protease inhibition by ATT, which is encoded by the SERPINA1 gene, resembles a mouse trap; the protease Typically neutrophil elastase) docks to the center loop at ATT, which swings to a more stable conformation, resulting in distortion of the structure of the protease and destruction of both molecules. B. In the Z variant (Glu342Lys) of ATT, a gap in the β-sheet A can either accept its own loop to form a latent conformation of process to irreversible polyimerization in the liver. Polymerization of the Z ATT within the liver leads to reduced circulating ATT levels; individuals with the genotype ZZ have approximately 15% of the normal plasma ATT levels.



Figure 8
Contributors to reduced alpha1-antitrypsin (AAT) activity among PI ZZ individuals. The most important determinant of reduced AAT activity in PI ZZ subjects is the reduced circulating level of AAT within the blood stream. However, other mechanisms have also been proposed, including reduced activity of each Z AAT molecule, inactivation of circulating AAT, and polymerization of circulating AAT.




Figure 9
Barriers to the diagnosis of alpha1-antitrypsin (AAT) deficiency (AATD). Despite the potentially devastating consequences, the majority of AATD subjects have not been identified; in one study, only 4% of the estimated PI ZZ individuals could be identified through local physicians. The major problem is underrecognition, though it is likely that the variable expression of disease phenotypes also plays a significant role. The role of deferral of genetic testing in AATD has not been accurately quantified.




Figure 10
Selection alpha1-antitrypsin alleles. The allele names are based historically on migration through an electrophoretic gel; for example, M (medium), S (small), Z (very low), and F (fast). The null alleles result in undetectable alpha1-antitrypsin and, therefore cannot be detected using serum electrophorisis. The most common alleles are the M alleles; phylogenic analysis indicate that M1A is the common ancestral form. S alleles result in a mild deficiency, while Z alleles result in both deficiency and dysfunction, with intracellular accumulation, increasing the risk for liver disease. The null alleles and other alleles (of which there are many) are quire rare.



Figure 11
Selected alpha1-antitrypsin (ATT) phenotype-genotype correlations. The combination of alleles that an individual inherits determines their ATT phenotype. The inheritance pattern for lung and liver disease manifestations is autosomal recessive, though the mode of inheritance for AAT levels is codominant. Levels exceeding 11uM (or 80 mg/dL) are generally felt to be protective against the development of chronic obstructive pulmonary disease (COPD), although epidemiologic studies indicate a possible increased risk of COPD in individuals with the PI MZ type.




Figure 12
Heterogeneity among PI ZZ subjects in the development of airflow obstruction. Despite inheriting identical variants in SERPINA1, PI ZZ subjects demonstrate significant variability in the development of COPD. In a study of 378 individuals, all with confirmed ZZ genotypes, forced expiratory volume I none second (FEV1) percentage predicted stratified by smoking status shows a wide variability, with some nonsmokers having low lung function and some smokers having preserved lung function. Other factors likely influence the development of COPD among PI ZZ subjects, which may include other diseases such as asthma or pneumonia, environmental factors, and genetic modifiers, such as interleukin-10.




Figure 13
Clinical presentation of alpha1-antitrypsin deficient (AATD) subjects. While early-onset of severe, lower-lobe predominant emphysema along with family history of COPD may be relatively easy to diagnose, underrecognitions of the disease still is a significant problem. Patients may present without classic pulmonary findings such as asthma or bronchiectasis, or with extrapulmonary manifestations such as liver disease, skin disease (panniculitis), or vasculitis. C-ANCA-cytoplasmic antineutrophilic cytoplasmic antibody.



Figure 14
Efficacy of alpha1-antitrypsin (AAT) augmentation therapy. Several investigations have assessed the efficacy of this therapy. Although there is no randomized controlled clinical trial that confirms efficacy, several observational studies provide supportive evidence. FEV1- Forced expiratory volume in 1 second.

Genetic Determinants of COPD Other Than alpha1-Antitrypsin Deficiency




Figure 15
A general theoretical model of the growth and decline of lung function with time. This diagram may also serve as a model of how different genetic and environmental factors could affect the development of COPD. The onset of symptoms increases with decreasing lung function but varies between individuals. Peak lung function is reached at approximately 20 years, which is typically followed by a plateau phase. Genetic variants or environmental factors (eg childhood respiratory infections) may result in reduced lung growth; with a similar rate of decline, this could result in an early onset of COPD. Genetic or environmental factors may also cat later in life, resulting in an early or more rapid decline in lung function. These mechanisms may also act in combination.




Figure 16
Steps to identify genetic determinants for a complex disease are as follows. 1) the phenotype(S) studied should be well defined; consider quantitative, objective, intermediate phenotypes (eg, serum markers, lung function). 2) A genetic component should be demonstrated. Studies of familial aggregation can asses the risk to relatives of affected individuals compared to unaffected individuals. Similarly, twin studies that demonstrate a higher correlation between monozygotic (identical) twins and dizygotic (fraternal) twins support a genetic etiology. The relative genetic component of the disease can be estimated by its heritability, which can be calculated by twin studies (potentially resulting in an overestimate) or family based studies. 3) Linkage analysis followed by gene localization (positional cloning) has been the standard approach for mendelian disorders and has been used with some success in complex trait genetics. Genome-wide association analyses are a more powerful method to search for novel susceptibility loci. Alternatively, candidate genes can b chosen for biological plausibility from human and animal models, gene expression, monogenic syndromes, etc. 4) Once an association is demonstrate, it is important to replicate findings in other populations. 5) an association with a genetic variant such as a single-nucleotide polymorphism is usually due to linkage disequilibrium; in other words, the variant is not in its self causative, but co-segregates in the genome with the true variant(s). Thus, further studies are usually needed to identify the functional variant(s). Gene-gene and gene-environment interactions can be studied. Functional work is important in determining pathophysiology. 6) Once the key functional variant(s) are found, the impact of the variant on the overall disease burden in the population can be determined, estimation of risk and prognosis can be provided, and, ultimately, improved understanding of the pathophysiology can new treatments can result.




Figure 17
Methods to identify the genetic determinants of COPD. Linkage analysis has been successful in mendelian traits, though there has been only modest success in finding novel genes associated with complex traits. The vast majority of studies have employed candidate gene-association analysis. Results of these association studies have been mixed, with lack of consistent replication a significant problem. Genome-wide association analyses have been promising in many other complex diseases. SNP - single nucleotide polymorphism; GWAS - genome-wide association studies; QTL - quantitative trait locus; CNV - copy number variation.




Figure 18
Population-based, case-control, and family-based genetic- association analysis. In all three types of studies, there is an association of disease status (bold outline) with the risk allele (purple). In case-control or population-based study designs, a higher frequency of AB and BB genotypes is seem among cases, which can be tested using chi-square, Armitage trend, or other tests, depending on the genetic model. In family-based analysis, parent-child trios or larger pedigrees are collected. Based on random assortment, each offspring has a 50% change of inheriting either A or B from a heterozygous AB parent. In this example, the cases are much more likely to have received the B allele, indicating deviation from random assortment. The statistical significance of the associations can be tested using the transmission disequilibrium test or its variations.




Figure 19
Evidence for genetic susceptibility to COPD and COPD-related lung-functions phenotypes. Familial aggregation may be demonstrated in several ways. In twin studies, higher concordance among (or smaller differences between) monozygotic versus dizigotic twins is suggestive of genetic influences. Studies of lung function have ranged from showing the same correlation in monozygotic and dizygotic twins to show that monozygotic twins have twice as much correlation; however, evidence for greater monozsygotic concordance has been demonstrated in most studies. In many family studies, a higher correlation has been demonstrated among more closely related family members (eg, siblings versus spouses), which also suggests a genetic influence. Heritability is an estimate of the proportion of genetic, as opposed to environmental, contributions to phenotypic variation. Several factors need to be considered, such as: 1) twin studies, 2) inclusion of covariates such as height and smoking can significantly affect heritability estimates, 3) different models and different populations (eg, general population versus COPD families) make direct comparisons difficult. Nevertheless, most heritability estimates suggest a significant impact of genes on lung-function levels. Finally, several studies in COPS using various phenotypes of airflow obstruction, chronic bronchitis, or a combination, have demonstrated a relative risk greater than 1 to siblings, indicate the involvement of familial/genetic components in COPD. FEF -forced expiratory flow; FEV1 - forced expiratory volume in 1 second; FVC - forced vital capacity.



Figure 20
Linkage analysis of lung function in the general population and in COPD families. Linkage analysis uses relatively widely spaced genetic markers throughout the genome and relies on larger regions of chromosomal sharing within families to find markers that co-segregate with disease loci. While highly successful for mendelian diseases, linkage analyses have been more challenging for complex diseases. Several studies have performed linkage analysis of lung function in families that were not selected for lung disease. To date, only one linkage analysis in COPD families has been reported, in the Boston Early-Onset COPD Study. Few regions of interest have been replicated between the general population studies and the COPD study; reasons for this include phenotypic and genetic heterogeneity and small sample sizes. One candidate gene foe lung function (SMOC2) and two candidate genes for COPD (TGFbeta1, SERPINE2) were identified within these linkage regions. Genomic regions with suggestive evidence of linkage in extended pedigrees (logarithm of the odds [LOD] greater than or equal to 1.9) were included; regions linked to more than one phenotype or in more than one study are shown in bold. FEF - forced expiratory flow; FEV1 - forced expiratory volume in 1 second; FVC - forced vital capacity.



Figure 21
Summary of the genetic-association studies of COPD and COPD related lung-function phenotypes (eg, forced expiratory volume in 1 second, lung-function decline). Candidate genes with significant associations in at least two studies are reported, excluding studies of asthma and SERPINA1. Studies were considered positive or negative based on the author's report and primary analysis (suggestive findings from gene-gene interactions and subgroup analyses were generally not included). Most candidate genes have had both positive and negative studies.



Figure 22
Reasons for a lack of replication in COPD genetic association studies. Of the numerous COPD genetic association studies published, few have been consistently replicated. Some have replicated an association, but in the opposite direction. Small sample size, genetic heterogeneity in different study populations, failure to adjust for multiple statistical testing, lack of assessment/ adjustment for population stratification, and genotyping error all are potential contributors to these inconsistent results. With the rapidly decreasing cost of single-nucleotide polymorphism genotyping and the availability of the HapMap, a lack of appropriate coverage of a gene or the genome) via genome-wide association studies) may become less of a problem, except for rare variants. TDT-transmission-disequilibrium test; FBAT- family-based association test; SNP- single-nucleotide polymorphism.




Figure 23
Selected replication genetic associations in COPD. Several genes have been associated in at least two studies for COPD or lung function phenotypes such as forced expiratory volume in 1 second (FEV1), FEV1/forced vital capacity (FVC), and FEV1 decline. These genes have generally come from three groups: protease-antiprotease, oxidant/antioxidant and xenobiotic metabolizing enzymes, and immune/inflammatory mediators. However, these pathways have been the most heavily studied source for candidate genes, so they may not be the most biologically important pathways. SNP- single-nucleotide polymorphism.



Figure 24
Selected Genetic Association Studies for COPD-Related Phenotypes. One of the problems in the genetics of COPD (and in fact, for most complex diseases) is the inherent heterogeneity of COPD cases. COPD-related subtypes, such as emphysema- versus airway-predominant disease may have different genetic determinants. In addition, COPD-related phenotypes, such as pulmonary hypertension may share genetic determinants with non-COPD related pulmonary hypertension, or have determinants unique to COPD.



Figure 25
Other Rare / Genetic Diseases Associated with COPD and/or Emphysema-like Phenotypes. While alpha 1-antitrypsin deficiency is the most widely studied single gene disorder leading to COPD, other genetic disorders also have COPD or emphysema-like conditions as part of their syndrome constellations. While these disorders can account for only a small proportion of COPD or COPD-like conditions, they may be helpful in finding the genetic determinants of more common varieties of COPD, and (in some cases) should also be considered in cases that appear to be typical COPD but have other unusual features. These disorders can be: 1) single gene disorders with the causal gene identified, such as Marfan syndrome, 2) disorders where the specific causal gene responsible for COPD has not been found (e.g. Down syndrome), or 3) the genetic mechanisms are unclear (Hypocomplementemic urticarial vasculitis). In addition, the strength of the association with COPD or emphysema varies from clear disease association (cutis laxa, Marfan syndrome) to case reports of an association (Sialuria, Niemann-Pick disease).




Mouse Models of COPD



Figure 26
Gene Knockout / Loss of Function and Transgenic Animal Models of COPD and/or Emphysema. Animal models have played a key role in our understanding of COPD, beginning in the 1960's with providing key support for the protease-antiprotease hypothesis. Mouse strains with differential susceptibilities to emphysema and COPD have been described. In some cases, the genetic determinants have been found; in others, quantitative trait locus (QTL) mapping to determine the loci responsible for these findings holds great promise. The advent of mouse knockouts and transgenic mice that overexpress a protein of interest (often limited to a specific tissue) has led to significant improvement in our understanding of disease pathways, especially in models involving cigarette smoke exposure. The applicability of these models to human disease is not always clear; however, as 1) fundamental differences exist between human and animal lung structure and function, 2) human disease genetic variants generally do not lead to complete loss of protein function; 3) loss of protein function at a blastocyst stage may lead to compensation, and overexpression may interfere with expression of other genes; 4) the lung may be involved nonspecifically in systemic disease, and 5) models may be incomplete or have poor overlap with human disease.




Figure 27
Mouse Strains Associated with COPD and/or Emphysema. In addition to genetically engineered knock-out or transgenic mouse models of emphysema, there are a variety of naturally occurring murine strains that have increased susceptibility to emphysema. Adapted from Mahadeva.