Defining the Genetic Landscape of Idiopathic Pulmonary Fibrosis: Role of Common and Rare Variants

Idiopathic pulmonary fibrosis (IPF) is a disease with complex pathophysiology, in which genetic and environmental factors play significant roles; however, their relative contributions remain undefined (1). Understanding the genetic basis of IPF can help define its heritability, risk, pathogenic mechanisms, and therapeutic targets. Evidence for genetic risk stems from observations that 5–20% of patients with IPF have relatives with fibrotic interstitial lung disease and that genetic variants identified in familial forms of pulmonary fibrosis (FPF) are present in nonfamilial, or “sporadic,” IPF (2). As genetic determinants could significantly contribute to disease, investigators have sought to identify the full scope of genetic variants that account for IPF risk.

Increasingly powerful methodologies have elucidated the genetic landscape of IPF, including candidate gene screens, genome-wide association studies, whole exome sequencing (WES), and most recently whole genome sequencing (WGS). These methods have identified both common (typically with minor allele frequency [MAF] ⩾ 0.05) and rare variants (RVs) (variably defined as MAF ⩽ 0.0001–0.01). Common variants are more likely to be found in healthy individuals, whereas RVs tend to cosegregate with affected families and more clearly contribute to disease pathogenicity (3). To date, at least 25 genetic loci have been associated with IPF (4, 5). The most prevalent is the MUC5B risk allele, which accounts for up to 30% of IPF risk and is also present in 10–20% of the general population (6, 7). IPF RVs have been identified in the telomere-related genes (TRGs) TERT, RTEL1, and PARN (8–12); surfactant-related genes (13–15); and the mitotic spindle gene KIF15 (4, 16). Of note, non-TRGs have not been consistently replicated across studies and methodologies (7, 10, 12, 16–18).

In this issue of the Journal (pp. 1194–1202), Peljto and colleagues (19) further define the scope of RVs in IPF by using WGS from the Trans-omics for Precision Medicine (TOPMed) program. Compared with prior methods, WGS provides a more unbiased analysis of both coding and noncoding regions. Peljto and colleagues included 2,180 IPF cases and 2,457 control subjects (without interstitial lung disease), split into discovery and validation cohorts. Cases of IPF were defined using diagnostic criteria from international society guidelines (20). The primary analysis focused on putative loss-of-function RVs (MAF ⩽ 0.01). Five RVs were significantly associated with IPF in the discovery cohort, but only RTEL1 remained significant in the validation cohort. A prespecified secondary analysis included missense variants and identified only TERT as significant. However, SPDL1, a gene involved in cell division checkpoint regulation, was modestly associated with IPF, as previously reported (4, 21, 22). Notably, the authors used the Rare Variant Filtering Tool to determine that a single variant in each of RTEL1, TERT, and SPDL1 genes accounted for most of the increased risk for IPF for that RV. Patients with FPF were more likely to carry the influential variant in TERT but not in RTEL1 and SPDL1. Two potentially novel common variants were also identified (in MCL1 and ENSG00000260803) that require further validation.

Peljto and colleagues were among the first to use WGS to examine genomic noncoding regions for RVs in IPF and have affirmed previous findings that TRGs represent the most significant RVs in IPF. The authors used well-defined and prespecified preliminary, primary, secondary, and meta-analyses, which limits potential biases. Use of the Rare Variant Filtering Tool provides potentially useful information by highlighting individual RVs in genes with the strongest influence on IPF risk. This may be important in prioritizing variants for prognostication and personalizing care for patients with IPF and their families. These results are further supported by a contemporary WGS study of exonic variants that identified TERT and RTEL1 as the only significant RVs contributing to risk (23).

Major contributions of this study are determining that the genetic heritability of IPF is 32% and providing additional evidence that most heritability is determined by common variants. Hence, we agree with the conclusion that the overall SNP contribution to IPF heritability may only modestly change as larger and more diverse population genomes are studied. However, efforts to identify additional RVs should continue, as their discovery will provide novel insights into IPF pathogenesis. For example, the discovery of KIF15 (16), a mitotic spindle–related gene, could uncover novel disease mechanisms and contribute to drug discovery. As the cost of sequencing decreases and more patients undergo genetic testing, it will be important to continue defining pathogenic RVs using larger studies in diverse populations with refinement of statistical methodologies.

There are several limitations to the study. First, the analysis of Peljto and colleagues was restricted to IPF cases of European ancestry, which potentially reduces the pool of RVs. For example, KIF15 was identified in WES and WGS studies when non-European individuals were included (4, 24). Second, MAF cutoffs have been inconsistent across studies. In a large cohort WES study, the MAF was 0.005 (10), which is lower than that used by Peljto and colleagues. Third, because RVs are, by definition, rare, splitting the sample into a discovery and validation cohort may have resulted in loss of significance of RVs in the validation cohort that were significant in the discovery cohort. Together, these limitations may explain why only two RVs (three if significance limitations are less stringent) were associated with IPF, whereas previous studies identified additional non-TRG RVs, such as SFTPC (13) and KIF15 (16). Last, in the bulk of the analysis, familial and sporadic IPF were grouped together. As FPF is more strongly associated with pathogenic RVs, it may be useful to investigate genetic risk solely in sporadic IPF.

We believe that this study represents an important contribution to further refine the genetic landscape of IPF; however, it may be difficult to estimate the risk of acquiring IPF by genetic studies alone, as myriad gene–gene and gene–environment interactions are likely to contribute to the phenotypic penetrance of individual genetic variants. To fully determine the genetic and environmental contribution to IPF risk, more studies are needed on the “exposome,” which integrates the entirety of an individual’s exposure starting from conception, including genomic, metabolic, and other -omics approaches (25). Continued efforts at understanding factors that contribute to IPF risk will ultimately allow better prognostication and therapies for patients and their families.