Plasma biomarkers of neuroinflammation and vascular injury predict cognitive decline in patients with mild cognitive impairment: Biomarkers (non‐neuroimaging)/Plasma/Serum/Urine biomarkers. (7th December 2020)

• Record Type: Journal Article
• Authors: Trombetta, Bianca A.; Kivisäkk, Pia; Carlyle, Becky C.; Magdamo, Colin; Merrill, Emily; Chibnik, Lori B.; Serrano‐Pozo, Alberto; Hyman, Bradley T.; Das, Sudeshna; Arnold, Steven E.
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 5
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.046134
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: The development of robust blood‐based biomarkers for pathological and clinical characterization of Alzheimer's disease (AD) is necessary, particularly as therapeutic approaches shift towards early intervention. Novel, ultra‐sensitive Olink TM technology allows for high‐throughput quantification of protein signatures in biofluids. In this study, we examined the utility of this technology to identify plasma biomarkers associated with longitudinal cognitive performance in individuals presenting with mild cognitive impairment (MCI). Methods: 60 participants from the Massachusetts ADRC Longitudinal Cohort with a clinical diagnosis of MCI, baseline global CDR of 0.5, and at least five annual follow‐up evaluations were classified into two groups based on their global CDR trajectory: MCI‐decline (an increase in global CDR from 0.5 to 1) and MCI‐stable (no change in global CDR). Five Olink TM Proximity Extension Assay (PEA) panels were assessed for technical reliability and used to quantify 460 plasma proteins at baseline visit. We built statistical models to identify proteins whose plasma levels were predictive of future cognitive decline, and used machine‐learning to evaluate the utility of these proteins for predicting conversion from MCI to dementia, relative to clinical and demographic predictors. Result: Olink TM panels exhibited excellent technical precision and plate‐to‐plate replication. Of the 460 proteins measured, 60 did not pass quality control thresholds. A final model adjusting for age and sex identified 14 proteins with a significant fold‐change in MCI‐decline compared to MCI‐stable (p<0.05), including NfL, CX3CL1, M‐CSF, VEGF‐A, IL‐8, NOS3, TNFRSF12A, and PHOSPHO1. An additional 10 proteins approached significance (p<0.1), primarily representative of inflammatory and vascular processes. Addition of these proteins to a supervised classification model informed by baseline cognitive measured substantially improved the overall accuracy from 63% to 83.3% (AUC: 0.88, sensitivity: 0.76, specificity: 0.87). Conclusion: Our findings suggest that plasma biomarkers of neuroinflammation and vascular injury have value as predictors of MCI‐to‐dementia progression. These circulating biomarkers may reflect CNS processes, blood‐brain barrier dysfunction, and/or systemic inflammation that drive AD associated pathophysiologies, or represent comorbid vascular pathologies. Developing a plasma biomarker panel could aid in prognostic deliberations by identifying MCI individuals at higher risk of progressing to dementia in clinical practice.