Identifying proteomic risk factors for cancer using prospective and exome analyses: 1,463 circulating proteins and risk of 19 cancers in the UK Biobank

Preprint

Background Proteins are essential for the development and progression of cancer and for the human body’s defense against tumor onset. The availability of a large panel of protein measurements and whole exome sequence data in the UK Biobank has enabled the simultaneous examination of plasma protein associations with risk across multiple cancer sites and their potential role in cancer etiology.

 

Methods We investigated the associations of plasma proteins with incidence of 19 cancers and 9 cancer subsites in up to 44,645 middle-aged adults in the UK Biobank, who had measurements of 1,463 plasma proteins generated using Olink Explore Proximity Extension Assay in baseline blood samples (2006-2010). Using multivariable-adjusted Cox regression, we estimated the risk of each protein with each cancer overall and by time-to-diagnosis after correction for multiple-testing. Identified protein-cancer associations were further assessed in an analysis of cancer risk using cis-pQTL and exome-wide protein genetic scores (exGS) in all UK Biobank participants (n=337,543).

 

Results We identified 371 proteins associated with the risk of at least one incident cancer, represented by a total of 621 protein-cancer associations. These proteins were associated with cancers of the blood (201 proteins), liver (131), kidney (51), lung (28), esophagus (22), colorectum (15), stomach (8), breast (5), prostate (3), endometrium (3), ovary (2), bladder (1), head and neck (1), and brain (1). 100 of these 621 protein-cancer associations persisted for cases diagnosed more than seven years after blood draw. Of these 621 associations, there was further support from cis-pQTL analyses for the etiological role of TNFRSF14 in risk of non-Hodgkin lymphoma (NHL), and from whole exome protein score (exGS) analyses for 28 other protein-cancer associations, including SRP14 and risk of leukemia. Proteins with directionally concordant evidence from long time-to-diagnosis analyses and from both cis-pQTL and exGS analyses were SFTPA2 for lung cancer, TNFRSF1B and CD74 for NHL, and ADAM8 for leukemia.

 

Conclusions For the first time using an integrated multi-omics and cross-cancer approach, we have comprehensively assessed the plasma proteome in relation to cancer risk and identified multiple novel etiological candidates. Differences in the levels of many circulating proteins were detectable more than seven years before cancer diagnosis; while some of these are likely to be markers of early cancer processes that may inform risk stratification, and/or risk factors, concordant evidence from genetic analyses suggests that some may have a role in cancer development.

Authors: Papier, K; Atkins, JR; Tong, TYN; Gaitskell, K; Desai, T

• Publication status: Published
• Peer review status: Not peer reviewed
• Preprint server: medRxiv
• Publication date: 2023-08-05
• DOI: 10.1101/2023.07.28.23293330
• Language: English