Abstract
Considering the distinct etiological pathways and molecular characteristics of different lung cancer subtypes, it is crucial to develop subtype-specific prevention strategies and therapeutic targets. This study aimed to identify protein biomarkers and potential therapeutic targets for specific subtypes of lung cancer by integrating population-based observational studies and Mendelian randomisation (MR) analyses. The cohort study was conducted in the UK Biobank, including about 47,000 participants whose blood samples were measured for 2,923 unique proteins and who were followed for the development of lung cancer. Two-sample MR was performed leveraging publicly available data from genome-wide association studies (GWAS) and protein quantitative trait loci (pQTL). Proteins were prioritised based on consistent associations across logistic regression, MR, transcriptomic validation and sensitivity analyses. Tier 1 proteins passed all evaluations, including GP1BA (squamous cell carcinoma) and ACADSB (small cell carcinoma). Tier 2 proteins, supported by transcriptomic evidence but not sensitivity analyses, included AGRN, ITGB2, SEPTIN3 (adenocarcinoma) and DPP10 (squamous cell carcinoma). Tier 3 proteins, supported by logistic regression and MR only, included CD5L, GNPDA, ACAN, C7, DMP1, HEPH, CEACAM6, COX6B1, CPXM2 and IL12RB2. Druggability evaluation suggests that existing drugs targeting ITGB2, GP1BA, ACADSB and COX6B1 could potentially be repurposed for the treatment of specific lung cancer subtypes.