Validation of soluble amyloid-β precursor protein assays as diagnostic CSF biomarkers for neurodegenerative diseases

Abstract

Analytical validation of a biomarker assay is essential before implementation in clinical practice can occur. In this study, we analytically validated the performance of assays detecting soluble amyloid-β precursor protein (sAPP) α and β in CSF in two laboratories according to previously standard operating procedures serving this goal. sAPPα and sAPPβ ELISA assays from two vendors (IBL-international, Meso Scale Diagnostics) were validated. The performance parameters included precision, sensitivity, dilutional linearity, recovery, and parallelism. Inter-laboratory variation, biomarker comparison (sAPPα vs. sAPPβ) and clinical performance was determined in three laboratories using 60 samples of patients with subjective memory complaints, Alzheimer's disease, or frontotemporal dementia. All performance parameters of the assays were similar between labs and within predefined acceptance criteria. The only exceptions were minor out-of-range results for recovery at low concentrations and, despite being within predefined acceptance criteria, non-comparability of the results for evaluation of the dilutional linearity and hook-effect. Based on the inter-laboratory correlation between Lab #1 and Lab #2, the IBL-international assays were more robust (sAPPα: r2 = 0.92, sAPPβ: r2 = 0.94) than the Meso Scale Diagnostics (MSD) assay (sAPPα: r2 = 0.70, sAPPβ: r2 = 0.80). Specificity of assays was confirmed using assay-specific peptide competitors. Clinical validation showed consistent results across the clinical groups in the different laboratories for all assays. The validated sAPP assays appear to be of sufficient technical quality and perform well. Moreover, the study shows that the newly developed standard operating procedures provide highly useful tools for the validation of new biomarker assays. A recommendation was made for renewed instructions to evaluate the dilutional linearity and hook-effect.