Hidden lung-specific DPP9 protein variant linked to severe COVID-19 and lung fibrosis
核心洞察
Researchers identified a previously unrecognised lung-specific form of the DPP9 (搜索) enzyme carrying a genetic variant linked to severe COVID-19 (搜索) and lung fibrosis (搜索).
The variant rs12610495 alters the function of this hidden protein, which had remained overlooked because it is absent from standard reference genome databases.
The study combined large genetic datasets with long-read RNA sequencing and computational modelling across 22 tissues to map disease-associated variants to alternative protein forms.
A genetic difference carried by nearly one in three people may increase the risk of severe COVID-19 (搜索) and lung fibrosis (搜索) by disrupting the function of a previously unknown protein, according to a new study published in the journal Nature Communications. Researchers from the University of Edinburgh, with partners at the University of Sydney, found that the genetic variant alters a newly discovered enzyme found in lung tissue, revealing a hidden biological pathway that could help explain why some people are more susceptible to serious lung disease.
The findings add to growing evidence that the human genome contains many previously overlooked proteins that may have important roles in health and disease. Understanding where these proteins are found in the body, how they change during disease, and how genetic variation affects their function could help explain the causes of disease, improve diagnosis, and reveal new targets for treatment.
Overlooked effects in alternative protein forms
Genetic variants that alter protein function are usually interpreted according to a gene's standard, or reference, transcript. However, many human genes can produce multiple alternative forms, some of which are expressed only in particular tissues or biological contexts and are not included in standard reference genome databases. The researchers therefore investigated whether disease-associated variants could affect these alternative isoforms in ways that might otherwise be overlooked.
The study combined large genetic datasets with advanced RNA sequencing and computational modelling to examine how common and rare gene variants may affect alternative forms of proteins. The researchers found that gene sections specific to alternative isoforms contained between 1.7- and 2.4-fold more population variants per kilobase than those in standard reference forms. These alternative gene sections were also more likely to contain variants that changed the resulting protein sequence. Among common variants, 56.8% in alternative forms were missense variants, which alter a protein's amino acid sequence; by contrast, most variants in reference forms were synonymous and did not change the protein sequence.
Around 40,000 disease-associated variants listed in the Genome Wide Association Studies (GWAS) Catalog mapped to alternative gene sections across approximately 15,000 transcript isoforms identified by long-read RNA sequencing in 22 tissues, compared with around 24,000 variants mapping to reference gene sections. Many of these alternative transcripts showed marked tissue-specific expression. When the researchers looked more closely at variants found specifically in alternative forms, around 80% of the relevant transcripts had not previously been annotated. Computational analyses also identified a number of variants predicted to have structurally damaging effects on the proteins produced.
A lung-specific DPP9 isoform
As a detailed example, the researchers examined DPP9 (搜索), a gene involved in regulating inflammation, and identified the genetic variant rs12610495 within an alternative form expressed in the lung. The variant has previously been associated with both severe COVID-19 (搜索) and lung fibrosis (搜索). Long-read sequencing showed that the previously unannotated full-length DPP9 transcript had lung-specific expression in the datasets analysed. Further experiments confirmed its expression in lung epithelial cell lines and showed that expression increased during lung epithelial cell differentiation.
The variant altered the protein produced by this lung-specific form of DPP9 (搜索). Laboratory experiments indicated that this affected some aspects of DPP9 function, although its overall enzymatic activity was similar to the reference form under some experimental conditions. The findings provide an example of how a disease-associated genetic variant may have effects through a tissue-specific alternative form that could be missed when analysis focuses on conventional gene annotations.
The team also identified similar genetic changes affecting newly discovered proteins that may influence vitamin D levels and contribute to kidney or heart disease. Some of these variants are common, affecting nearly 50% of the population, but are likely to have only modest effects on disease risk. Others are extremely rare, affecting as few as one in a million people, but are thought to cause more severe forms of disease.
Implications for diagnosis and treatment
Commenting on the wider implications, study author Dr Simon Biddie, clinical lecturer at the University of Edinburgh's Institute of Genetics and Cancer, said: "For decades, we have interpreted genetic variation through the lens of the proteins we knew existed. Our findings show that many genetic changes linked to disease may actually act through previously hidden proteins that have only recently come to light. Understanding these proteins could transform how we diagnose genetic diseases and identify new targets for treatment."
Professor Mark Gorrell, of the Centenary Institute and the University of Sydney, who first discovered the enzyme and was part of the team for this study, said: "This new research both explains the association of the enzyme with lung diseases and opens up new avenues to understand this enzyme in disease and help us devise potential therapies."
The authors noted that considering full-length and tissue-specific isoforms could improve the interpretation of both common and rare genetic variants. However, they observed that current long-read transcriptomic datasets do not fully capture disease states or rare cell populations, and that GWAS associations do not necessarily identify the causal variant. Further disease-specific and cell-specific sequencing could improve the identification and interpretation of functionally relevant alternative isoforms, they suggested.