Exosome Structures Supported by Machine Learning Can Be Used as a Promising Diagnostic Tool.

Area of Science:

• Biophysics
• Machine Learning
• Cell Biology

Background:

• Exosomes, extracellular vesicles (EVs), play roles in intercellular communication and disease.
• Characterizing exosome morphology is crucial for understanding their function and potential in diagnostics.
• Immature dendritic cells (IDCs) are a source of EVs, but their exosome morphology requires specialized analysis.

Purpose of the Study:

• To investigate the utility of Principal Component Analysis (PCA) for evaluating the dynamic morphological features of exosomes.
• To assess the potential of PCA combined with Cryo-Transmission Electron Microscopy (Cryo-TEM) imaging for disease diagnosis and prognosis.
• To determine if crude isolation of exosomes from immature dendritic cells (IDCs) is sufficient for PCA-based morphological analysis.

Main Methods:

• Crude isolation of exosomes from JAWSII immature dendritic cells (IDCs).
• Cryo-Transmission Electron Microscopy (Cryo-TEM) for high-resolution imaging of exosome morphology.
• Principal Component Analysis (PCA) applied to Cryo-TEM images, analyzing parameters like membrane projections, Gaussians, Hessian, hue, and class.
• Comparison of Brownian motion data from Nanoparticle Tracking Analysis (NTA) with SEM and confocal microscopy images.
• Sodium-Dodecyl-Sulphate-Polyacrylamide-Gel-Electrophoresis (SDS-PAGE) to assess protein contamination in crude isolates.

Main Results:

• PCA effectively analyzed morphological features (3D orientation, shape, size, brightness) of IDC-derived exosomes using Cryo-TEM data.
• FTIR identified functional molecular groups, but Cryo-TEM revealed unique physical and morphological characteristics.
• SDS-PAGE confirmed no significant protein contamination in crude exosome isolates, simplifying the analysis workflow.
• Brownian motion analysis provided complementary data to imaging techniques.

Conclusions:

• PCA is a valuable and novel machine-learning tool for analyzing exosome morphology from Cryo-TEM images.
• Crude isolation of exosomes from IDCs is sufficient for PCA-based morphological evaluation, eliminating the need for further purification.
• This approach holds promise for disease diagnosis and prognosis by leveraging dynamic exosome features.