AI-Driven Integration of Spatial Histopathology and Single-Cell Genomics for Cardiovascular Precision Medicine
This research plan proposes a novel AI-driven framework for integrating spatial histopathology and cardiovascular genomics data at the single-cell level to advance the understanding of cardiovascular diseases (CVDs). Combining expertise in artificial intelligence, bioinformatics, and computational biology, the study aims to develop state-of-the-art methods for analyzing multimodal data and
uncovering novel insights into disease mechanisms.
The research focuses on several key areas. Advanced segmentation models, such as U-Net, Mask R-CNN, and transformer-based architectures, will be employed to extract morphological, spatial, and textural features from histopathology images, enabling a comprehensive analysis of tissue archi- tecture. For multimodal data integration, joint embeddings, graph-based models, and explainable AI (XAI) techniques will be utilized to link morphological and genomic data, facilitating a deeper understanding of cellular and molecular interactions. Predictive modeling frameworks will be de- signed to forecast disease outcomes and identify clinically relevant biomarkers, with an emphasis on interpretability and translational applications. Additionally, a scalable, user-friendly compu- tational pipeline will be developed to support researchers and clinicians in leveraging multimodal data effectively.
The research emphasizes ethical considerations, including data privacy, fairness, and bias mitigation. It leverages advanced computational resources and collaboration with experts in genomics, pathology, and AI to enhance scientific rigor and relevance. The findings will be disseminated through high-impact publications, international conferences, and open-source tools to ensure ac- cessibility and foster innovation. By integrating advanced AI methods with biomedical expertise, this research has the potential to transform CVD research, enhance diagnostic and therapeutic strategies, and contribute significantly to the broader field of precision medicine.
Key words:
Spatial histopathology, cardiovascular genomics, deep learning for image analysis, multimodal data integration, predictive modeling, biomarker discovery
Ultrasensitive detection of neurological biomarkers using upconversion nanoparticles paired with digital readout methods
Neurological diseases, including Alzheimer’s and Parkinson’s diseases, are a leading cause of mortality globally, exacerbated by increased life expectancy. Alzheimer’s disease, the most prevalent cause of dementia, often develops decades before the onset of symptoms, leading ultimately to fatal outcomes due to the progressive neurodegeneration. Early detection and diagnosis are essential for timely intervention to slow the disease progression. However, the low concentrations of neurological biomarkers, such as neurofilament light chain (NfL), phosphorylated tau proteins, and amyloid-beta peptides, in blood present significant challenges for early detection using conventional bioanalytical methods. Current diagnostic approaches primarily rely on the detection of biomarkers on cerebrospinal fluid and lack the sensitivity for early disease detection in blood.
The proposed research aims to develop ultrasensitive immunoassays leveraging photon-upconversion nanoparticles (UCNPs) to reshape neurodegenerative disease diagnostics. UCNPs, with their unique property of emitting shorter wavelength light under near-infrared excitation (anti-stokes emission) enable background free detection. Therefore, these unique nanoparticles can overcome many of the limitations of other optical labels, such as fluorophores and quantum dots, caused by autofluorescence and light scattering. The proposed project aims to develop ultrasensitive assays that integrate advanced surface chemistry and a novel digital readout mode for single-molecule detection, achieving unparalleled sensitivity.
The project is structured into five main objectives: 1) Optimization of UCNP surface chemistry to reduce nonspecific binding and enhance specific binding, 2) The development of a sandwich immunoassay for the detection of NfL using analogue and digital readout modes. 3) The construction of an upconversion microscope to perform the digital readout. 4) The implementation of an immune complex transfer assay to eliminate background due to nonspecific binding and improve sensitivity, and 5) The expansion of the platform for multiplexed detection of additional biomarkers including phosphorylated tau proteins. The real-world applicability will be demonstrated by validating the assay using clinical samples and comparing their performance to established methods like SIMOA and Elecsys.
Key words:
neurological disorders, neurological biomarkers
Julian Brandmeier
Senior Researcher
Tero Soukka
Professor
Development and Application of Novel PET Radiotracers for ImmunoPET Imaging of TREM2 in Heart and Brain
Aims:
The goal of this research is to develop novel PET radiotracers targeting Triggering Receptor Expressed on Myeloid cells 2 (TREM2) to enhance understanding of its role in cardiovascular and neuro-inflammation diseases. By developing TREM2-specific PET tracers, this project aims to create a tool to elucidate TREM2’s role in immune regulation and its dualistic function in inflammation and tissue repair. This tool will enable detailed exploration of TREM2’s involvement in the progression of conditions like atherosclerosis and Alzheimer’s disease.
Research Material and Methods:
This project employs an interdisciplinary approach, integrating immunology, radiochemistry, and molecular imaging to develop PET tracers that can quantify TREM2 expression in vivo. Advanced antibody engineering techniques are used to generate high-affinity TREM2-specific antibody constructs, such as single-chain variable fragments (scFvs), Fab fragments and small immunoproteins (SIPs). These fragments are optimized for tissue penetration, clearance rates, and binding affinity. Radiochemistry methods are tailored to each antibody format, employing mild reaction conditions for isotopes such as fluorine-18, gallium-68, and zirconium-89 to preserve antibody functionality. Pretargeting can also be employed for cardiovascular diseases such as atherosclerosis. This two-step approach utilizes bioorthogonal chemistry, enabling improved imaging contrast and reduced off-target effects, making it particularly suitable for imaging cardiovascular pathologies like atherosclerotic plaques.
In vitro assessments include autoradiography and binding specificity studies, while in vivo studies use small animal PET/CT to analyze tracer pharmacokinetics and uptake in target tissues. To validate the PET tracer’s specificity and biodistribution, atherosclerotic murine models (ApoE-/- mice) and transgenic Alzheimer’s murine models (tg-ArcSwe mice) can be utilized in the preclinical evaluation.
Ultimately, the research aims to establish TREM2 PET imaging as a translational tool, providing quantitative insights into disease progression and supporting the development of targeted diagnostic and therapeutic strategies for cardiometabolic and neurodegenerative diseases.