Novel Application of Polygenic Risk Scores Identifies Biomarkers for Early Detection of Glioma

Abstract

Abstract AIMS In the UK, brain, other CNS and intracranial tumours incidence rates are increasing with over 12,200 new cases diagnosed each year. The overall survival rate over 5 years in these tumours is approximately 12%. There is a need for predictive biomarkers to detect the presence of gliomas to improve prognosis for such patients. A novel modified Mendelian randomization (MR)-based technique has been proposed to discover predictive biomarkers for diagnostic applications. The aim of this project is to identify circulating blood biomarkers that are genetically liable to all gliomas, glioblastoma (GB), or non-GBs. METHOD We generated polygenic scores for genetic liability to glioma and a separate GWAS of plasma protein levels from the INTERVAL study. We used sensitivity analyses methods to further assess our findings. The results were then validated by the plasma pQTL GWAS data from the Fenland cohort. RESULTS We identified a potential blood biomarker called erythroblast membrane associated protein (ERMAP) which showed consistent results across all sensitivity analyses in our discovery cohort [β: -0.0778, P-value: 2.1E-04]. We identified that decreased levels of ERMAP [β: -0.078 SD of ERMAP abundance per SD increase in glioma risk, 95% confidence interval (CI) 0.037 to 0.119, P-value: 2.1E-04] in the blood are associated with increased risk of glioma. CONCLUSION ERMAP protein is found to have a potential genetic liability for all gliomas. Predictive biomarkers are novel to glioma research and could be important for diagnostic applications and population screening in early detection. However, further clinical research is required to validate these findings.