Identification of protein glycosylation-dependent alterations in the serum of Type 2 diabetic individuals with cognitive decline using an analytical, computational and biochemical approach

• Grant number: 25/00049-7
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: May 01, 2025
• End date: April 30, 2028
• Field of knowledge: Biological Sciences - Biochemistry - Chemistry of Macromolecules
• Principal Investigator: Giuseppe Palmisano
• Grantee: Lucas Cardoso Lázari
• Host Institution: Instituto de Ciências Biomédicas (ICB). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated research grant: 18/15549-1 - Post-translational modifications in Chagas Disease biological processes and diagnostics: novel methodological approaches and biological applications, AP.JP2

Abstract

Type 2 diabetes mellitus (T2DM), the most prevalent form of diabetes, affects millions globally and is strongly associated with cognitive decline, dementia, and neurovascular complications. Chronic hyperglycaemia in T2DM damages the blood-brain barrier and impairs microvascular function, leading to brain atrophy and cognitive deficits. While studies have linked these effects to altered protein expression and glycosylation patterns, the precise mechanisms remain unclear. This project employs an integrative approach combining glycoproteomics, proteomics, and machine learning to identify biomarkers of cognitive decline in T2DM and to gain insights into the underlying biological processes. Initially, glycoproteomics and proteomics analyses will be conducted using MALDI-TOF MS to differentiate individuals with and without cognitive decline. The resulting data will be used to narrow down the search space for potential biomarkers, which will then be further investigated through LC-MS/MS. This additional analysis will allow the identification of specific proteins and glycoproteins linked to cognitive decline and provide insights into the biological pathways affected by the condition. To enhance biomarker discovery, machine learning models will analyze protein and glycoprotein profiles generated by MALDI-TOF MS. Advanced architectures, such as autoencoders and transformers, will be trained to extract features and improve classification performance, enabling more precise differentiation of individuals. The project also includes validating identified biomarkers using orthogonal techniques, such as pseudo-SRM, lectin blotting, and biological validation, to confirm their relevance and reliability. By integrating proteomics, glycoproteomics, and machine learning, this project aims to train models for cognitive decline diagnosis, identify and validate biomarkers, and elucidate the biological processes contributing to T2DM-associated cognitive decline. These efforts will provide a deeper understanding of the mechanisms underlying this condition and support the development of novel diagnostic and therapeutic strategies. (AU)