Polygenic Risk and Rare Variants in Endotypes of Idiopathic Pulmonary Fibrosis
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Idiopathic pulmonary fibrosis (IPF) and telomere length (TL) are both strongly linked to rare and common genetic variation. Shortened TL itself may be causal for IPF. Whether rare and common variants compete or cooperate to confer genetic risk of IPF uniformly is unknown.
Methods
We used whole genome sequencing (WGS) data from a discovery case-control cohort sequenced at Columbia (777 IPF, 2905 controls) and validated findings using WGS data from Trans-Omics for Precision Medicine (TOPMed, 1148 IPF, 5202 controls) and the UK Biobank (UKBB, 2739 IPF, 395331 controls). In all cohorts, we identified rare damaging variants in disease-associated genes and computed control-normalized polygenic risk scores for IPF (IPF-PRS) and telomere length (TL-PRS). Telomere length of blood leukocytes was measured using a qPCR assay for two cohorts. We determined the association of the MUC5B rs35705950 polymorphism, an IPF-PRS excluding MUC5B (IPF-PRS-no MUC5B ), and a TL-PRS with IPF risk in the overall cohort and in subgroups stratified by genetic endotypes (rare variant carriers, non-carriers stratified by TL cutoffs). We calculated cross-validated area under the receiver operator curve (AUC) and compared the liability of IPF explained by genetic variables.
Findings
We identified independent associations between IPF risk and rare variants, the MUC5B SNP, and both polygenic scores in the discovery cohort and replicated these findings in the TOPMed and UKBB cohorts. The adjusted effect size of the TL-PRS, which includes >180 SNPs not previously associated with IPF, was comparable to the IPF-PRS-no MUC5B in the discovery (ORTL-PRS 1.63 [95% CI 1.47, 1.81] vs. ORIPF-PRS 1.60 [1.44, 1.77]) and replication cohorts (TOPMed ORTL-PRS 1.47 [1.36, 1.59] vs. ORIPF-PRS 1.37 [1.25, 1.50]; UKBB ORTL-PRS 1.24 [1.19, 1.29] vs. ORIPF-PRS 1.25 [1.21, 1.30]). The TL-PRS incrementally improved disease prediction beyond known IPF common and rare genetic predictors and clinical variables in discovery (combined AUC: 0.89, pDelong = 0.006), TOPMed (combined AUC: 0.89, pDelong = 0.01), and UKBB cohorts (combined AUC: 0.77, pDelong = 0.03). Rare and common variants jointly contributed to genetic liability of IPF. The TL-PRS increased liability of IPF explained by 13% in the discovery cohort and 8% and 13% in the TOPMed and UKBB cohorts, respectively. In IPF subjects with damaging rare variants, the TL-PRS was consistently associated with disease risk whereas the IPF-PRS-no MUC5B was not. The TL-PRS also conferred nominally greater odds of disease risk than the IPF-PRS-no MUC5B in patients with shorter TL, in the discovery and UKBB cohorts. Together, 23-43% of IPF cases have damaging rare variants or telomeres <10 th percentile, where the TL-PRS represents a major unrecognized genetic risk factor.
Interpretation
Common and rare genetic variation confer context-specific genetic risk in IPF competitively and cooperatively. In contrast to known IPF common risk variants, the TL-PRS, which includes >180 genetic loci not previously associated with IPF, increases the risk of disease specifically in certain IPF endotypes. Polygenic risk from telomere-associated common variants is a key feature of IPF genetic heterogeneity.
Funding
National Institutes of Health (NIH), Medical Research Council (MRC), National Institute for Health and Care Research (NIHR)
Research in context
Evidence before this study
We performed a search on PubMed for “common variants”, “rare variants”, “polygenic score”, “endotype”, and “IPF” on April 20 th , 2025, to identify integrated genetic studies of rare and common variants in IPF. We found no original articles that comprehensively examined common and rare IPF variants together. Nearly all articles analyzed common or rare variants in isolation; one article included both common and rare variant analyses but reported results separately without studying their combined effects. We identified numerous reviews that discuss or conceptualize the relative contributions of rare and common genetic risk to IPF risk; no studies provided or cited empiric data. Although “IPF endotypes” have been defined using transcriptomics and proteomics, we did not find any articles that utilized this terminology based on genetic factors, although several studies distinguish IPF with rare variants or telomere length below 10 th percentile.
Added value of this study
We describe the first comprehensive survey of common and rare risk variants together in IPF and identify subtype-specific genetic risk factors that substantially improve genetic explanation and disease prediction. We uncover complex relationships between polygenic factors and rare variants, including preliminary evidence of polygenic modifiers in IPF carriers of rare damaging variants that might serve as an explanation for incomplete penetrance. Genetic studies in IPF have been limited by sample size leading to “missing heritability”. We demonstrate value in leveraging large-scale genetic studies of causal molecular traits like telomere length to overcome these limitations and improve genetic understanding of IPF. These telomere-associated common variants are context-specific risk factors in certain endotypes, highlighting previously unrecognized genetic heterogeneity that will be important for future discovery of novel, reproducible genetic risk factors in IPF.
Implications of all the available evidence
A genetic basis for disease heterogeneity allows for research that focuses on relevant endotypes instead of “all-comer” disease and advances precision medicine approaches to IPF. Polygenic factors that can modify the effects of a high-risk rare variant represent a target for understanding disease-modifying pathways in IPF.
