Whole Genome Sequence Analysis of Weight Loss in 16 972 Participants With COPD Reveals Novel Risk Loci in DRAIC and...

Chronic obstructive pulmonary disease (COPD) is associated with musculoskeletal comorbidities, including cachexia. Weight loss (WL) is the major criterion for cachexia and increases risk for mortality in COPD. Risk factors for WL in COPD are incompletely understood. We performed this whole genome sequencing (WGS) analysis to identify genetic risk variants for WL in COPD.

We studied 16 972 participants from the Trans-Omics for Precision Medicine (TOPMed) Initiative and All of Us Research Program. COPD was diagnosed using spirometry in TOPMed, while diagnosis codes were used in All of Us. WL was defined as WL ≥ 5% or a final body mass index (BMI) < 20 kg/m². WGS data came from white blood cells in all cohorts. Single-variant testing was conducted on both race- and study-stratified cohorts and in a cosmopolitan, ancestry-independent manner using GENESIS in TOPMed and SAIGE in All of Us. SAIGE-GENE+ gene-based analyses were performed on race-stratified and cosmopolitan cohorts. Single variant meta-analyses were conducted using METAL within (B/AA and NHW analyses of Black/African-American and non-Hispanic white participants, respectively) and across racial groups (COSMO). Rare variant gene-based results were combined using Fisher's method. Transcriptomic effects were predicted using MetaXcan. We used the GWAS Catalogue to analyse for colocalization with other related traits.

Two single variants were associated with WL in COPD among All of Us participants: one intronic variant in HCN1 in Black/African-American participants (chr5:45271359:TACACAC:T, odds ratio with 95% confidence interval (OR (CI₉₅)) = 2.43(1.78-3.31), p = 1.95 × 10^-8) and one intergenic variant between PPP4R2 and PDZRN3 in the cosmopolitan and NHW cohorts (chr3:73345901:A:G, OR (CI₉₅) = 0.21(0.12-0.35), p = 8.84 × 10^-9 in cosmopolitan and OR (CI₉₅) = 0.18(0.10-0.33), p = 9.44 × 10^-9 in NHW). Single-variant meta-analysis identified two loci associated with WL in COPD: five variants within DRAIC in the B/AA meta-analysis (lead variant chr15:69571341:A:G, OR (CI₉₅) = 1.37(1.23-1.51), p = 1.29 × 10^-9) and two intronic variants within RFX3 in the cosmopolitan meta-analysis (lead variant chr9:3390983:T:C, OR (CI₉₅) = 1.50(1.31-1.73), p = 1.06 × 10^-8). Rare variants within RNU6-565P (NHW analysis in All of Us; p = 2.83 × 10^-7) and LOC339298 (B/AA analysis in All of Us, p = 1.85 × 10^-6) were associated with WL in COPD. The RNU-565P signal remained significant after combination with TOPMed results (p_combined = 1.23 × 10^-6). MetaXcan predicted differential expression of LNC00959 in visceral adipose tissue (p = 1.16 × 10^-6 in COSMO analysis). Colocalization analyses identified genomic associations between BMI and variants in or near DRAIC, RFX3, PDZRN3, and LINC00959.

In the first WGS analysis of WL in COPD, we have identified seven novel loci. Further characterization of these loci will validate our findings and improve our understanding of the molecular pathophysiology of this condition.