Multiomics and Multiancestry Approaches: Key Steps to Untangling the Web of Chronic Obstructive Pulmonary Disease Pathogenesis

In the past two decades, the scientific community has rapidly progressed from the sequencing of the first human genome to large-scale whole-genome sequencing of epidemiologic cohorts using affordable, direct-to-consumer sequencing. Concurrently, the use of other high-throughput “omics” technologies evaluating the molecular machinery that leads from gene to phenotype has surged. These studies have identified novel candidates for further study that, through their integration, have been used to fuel systems biology approaches to evaluate the pathogenesis of chronic diseases. The challenges of using large omics approaches to identify clinically relevant targets are visible in the history of the human genetics of chronic obstructive pulmonary disease (COPD) and informative to future systems biology approaches to the disease.

The earliest linkage and genome-wide association studies (GWASs) were focused on testing for associations between genetic loci and COPD case status or spirometric measures (1, 2). As time went on, a growing recognition of the relationship between the clinical heterogeneity of COPD and its complex genetic background led to refined phenotyping and analyses. In addition to spirometric measures of disease, shared and unique associations of radiographic, clinical, and longitudinal phenotypes of COPD with genetic loci were identified, and several large studies with deeply phenotyped populations have provided greater power to detect genetic associations, particularly in the setting of multipopulation meta-analyses (3, 4). Most recently, phenotyping and genotyping in populations of diverse genetic ancestry has expanded potential candidates even further, because the majority of early GWASs were performed in individuals of European ancestry primarily from self-identified race (5). We have learned that the power to detect and, in turn, identify clinically meaningful disease mechanisms is improved in large populations with diverse genetic ancestry that are deeply phenotyped. Next, we will need to integrate these findings with other omics approaches to untangle a web of candidates to identify those that are clinically relevant.

In this issue of AnnalsATS, Ngo and colleagues (pp. 1124–1135) present the findings of a large proteomic study designed to identify associations between circulating proteins and spirometric phenotypes of COPD in six primarily White cohorts with validation in African American participants from the Jackson Heart Study cohort (6). Using this approach, they identified 69 proteins that were significantly associated with forced expiratory volume in 1 second (FEV₁) in their discovery cohort, 56 of which were associated with only FEV₁ and not forced vital capacity. These proteins, further supported by the validation cohort, included members of the transforming growth factor-β pathway, as well as a serine protease inhibitor, kallistatin, that has not previously been associated with COPD in genetic or proteomic studies. Twenty-two of these proteins had previously been identified in genetic studies of lung function or COPD. The authors then identified 12 circulating proteins, including circulating elastase inhibitors, mediators of angiogenesis, and mucus-secreting cell-related proteins, that were associated with FEV₁ decline in a subset of four of the six discovery cohorts with longitudinal measurements of lung function. Finally, the authors investigated specific pathways enriched with identified candidate proteins, uncovering a number of previously explored pathways in COPD, including complement activation and the interleukin-1 receptor pathway.

This study has several strengths and applies several of the lessons learned from the genomics of COPD. It is the largest study to date investigating circulating protein biomarkers in obstructive lung disease. The authors used several well-phenotyped populations and accounted for spirometric measurements of disease as well as lung function decline over time when available. The authors also attempted to assess differences in disease mechanisms that may arise because of differences in genetic ancestry by including an African American validation cohort. There is an ongoing discussion in the literature on the use of race, which is a social construct, as a representative of genetic ancestry or genetic similarity (7, 8). The use of race as a classification approach oversimplifies genetic similarity and has a problematic history entangled with systemic racism (7). However, for established cohorts with defined study populations and existing data, such as those included in the present study, the use of race may be the best and only representative of genetic similarity available. The ideal solution to this problem for future studies will be to include even more diverse populations with clear measures of genetic similarity. Because race is really a surrogate for both genetic ancestry and complex environmental and social exposures, future cohort designs should also plan for the collection of comprehensive environmental and social determinants of health data to be included as covariates in analysis (9).

The results presented by Ngo and colleagues provide valuable insight into molecular mechanisms underlying COPD pathogenesis. Novel associations between lung function and circulating proteins such as kallikrein are biologically plausible and are supported by similar associations with matrix metalloproteinase 12 and members of the interleukin-1 pathway, which have previously been identified as candidate genes in GWASs. Future studies building on these findings will be strengthened by larger diverse study populations, deeper COPD phenotyping, and integrated genomic data, both for the representation of genetic ancestry or similarity and to strengthen evidence for candidate proteins. Untangling this web will improve our understanding of the underlying disease mechanisms in COPD and ultimately lead to novel interventions.