Identifying Differentially Expressed Proteins via Plasma Exosomal Proteomics

Background

Extracellular vesicle (EV) proteomics has become a powerful discipline for deciphering intercellular communication networks, identifying biomarkers of non-invasive diseases, and advancing precision medicine [1]. The protein cargo of nanoscale EVs (50–150 nm), derived from the endosome pathway, directly reflects the pathophysiological state of parent cells, making EV proteins an attractive target for liquid biopsy development. [1]. At the sample preparation stage, traditional EV separation methods, such as ultracentrifugation, size exclusion chromatography (SEC), and ultrafiltration, are often affected by the copurification of non-EV contaminants (e.g., abundant plasma proteins, lipoproteins, and protein aggregates), low recovery rates, and low batch-to-batch reproducibility [1,2].

Even after separation, low-abundance EVs are often obscured by the extreme dynamic range of biological fluids, especially in plasma, where high-abundance proteins such as albumin and immunoglobulin dominate the samples, masking disease-related low-abundance proteins [3,4]. In addition, the lack of a standardized end-to-end workflow—from vesicle separation to protein digestion, mass spectrometry collection, and data analysis—leads to poor repeatability and difficulties in cross-study verification [5,6]. At the detection level, the large number of proteins detected in proteomics measurements covers up the inherent heterogeneity among individual EVs, potentially resulting in the loss of rare but functionally important subgroups [7].

In order to address these challenges, proteomics based on mass spectrometry has been widely used to characterize the protein composition of EVs. Data-dependent acquisition (DDA) is the most commonly used mass spectrometry acquisition mode, and precursor ions are selected according to the cracking intensity [8]. Although DDA is suitable for protein identification and spectral library generation, it has some drawbacks, such as insufficient sampling of low-intensity precursors and poor cross-batch quantitative reproducibility [8,9].

In order to overcome these limitations, data-independent acquisition (DIA) has become a promising alternative. DIA systematically fragments all precursor ions within a predefined isolation window in a predefined m/z range, thus enabling unbiased and comprehensive analysis of the precursor ions and providing high quantitative reproducibility for large-scale cohort studies [10]. Unlike DDA, which selects precursor ions based on intensity, DIA enhances the identification and quantification of low-intensity peptides and significantly improves protein-group coverage [10]. For example, by combining SEC with DIA-MS, 2,896 EV-related proteins can be identified and quantified from as little as 200 μL of plasma, representing 3.5 times greater coverage than that reported in previously published studies on melanoma [11]. Although DIA provides in-depth coverage for protein omics in the discovery stage, targeted verification of candidate biomarkers requires an orthogonal strategy with higher quantitative accuracy [12].

Parallel reaction monitoring (PRM) is a targeted MS method implemented on high-resolution instruments, such as Orbitrap, which offers excellent sensitivity and quantitative accuracy [13,14]. Unlike multiple reaction monitoring (MRM), which is based on triple-quadrupole instruments, PRM scans all product ions from the target peptide (FDR:1%) to produce a complete MS/MS spectrum, minimizing ion interference and improving specificity [15]. However, PRM has its inherent limitations: because of the wider MS/MS scanning range and longer cycle time, peptide coverage is affected. In addition, high-resolution mass spectrometers require higher instrument costs and more complex operation and maintenance [16]. The throughput of PRM is also limited to 100–300 targets at a time, which makes it not suitable for target discovery [17,18]. Beyond its application in atrial fibrillation (AF) biomarker discovery, this label-free PRM EV proteomics protocol can be extended to multiple other translational contexts. In the development of liquid biopsy, this pipeline is well-suited to analyzing circulating foreign bodies from minimally invasive samples (such as plasma, urine, or saliva) to identify disease-specific protein signatures [19].

The protocol can also characterize EVs from specific cardiac cell types by targeting cell type–specific surface markers, thereby helping to dissect the contribution of cell specificity to disease pathogenesis.

In treatment monitoring, the same workflow can quantify extracellular proteins longitudinally to track the emergence of treatment response or drug resistance, especially in the context of antiarrhythmic drug therapy or catheter ablation for AF. In addition, for other cardiovascular diseases characterized by progressive structural remodeling, such as heart failure with preserved ejection fraction, hypertrophic cardiomyopathy, and cardiac amyloidosis, this approach provides a minimally invasive method that can monitor disease progression and evaluate treatment effects through a series of extracellular protein profiles. Finally, this standardized label-free PRM EV proteomics scheme provides a multifunctional platform for translational cardiovascular research and has great potential for accelerating the discovery and clinical implementation of EV-based biomarkers for AF and other diseases.