Circulating Extracellular Vesicles As the Source of Diagnostic Biomarkers for Diseases

Acute pancreatitis (AP) is an inflammatory disorder resulting from the death of the acinar cells in the pancreas. AP can be classified into three categories, mild, moderately severe, or severe according to the Atlanta classification1 and is often associated with long-term comorbidities. Severe AP (SAP) is characterized by pancreas necrosis accompanied by single or multiple organ failure with a mortality rate of 20%–40%.2 While many clinical studies have attempted to identify relevant predictive biomarkers, none of them provides a satisfactory way to predict the severity of AP.3 In the quest for new biomarkers, extracellular vesicles (EV) secreted in biological fluids represent exciting investigating opportunities. Indeed, they can be easily collected and could represent a reliable snapshot of any given tissue physiology. Furthermore, a sequential collection could be used to track disease development and monitor treatment efficiency. EV content and functions vary greatly according to the cell type of origin, the environment, or the process leading to their formation. But EV have been generally recognized as cell communication mediators and play key roles in numerous diseases including neurodegeneration or cancer.4 In their recent letter to editor (“Identification of circulating extracellular vesicle long RNAs as diagnostic biomarkers for patients with severe acute pancreatitis” by Zhu et al., in press in Clinical and Translational Medicine (CTM2-2022-06-1158)), Zhu and collaborators propose to use EV as predictive biomarkers for severe AP. First, they isolated EV from the blood of healthy controls (HC), mild (MAP) or SAP patients after their hospital admission and sequenced the total RNA content. Interestingly, they did not limit their analysis to mRNA but extended it to long non-coding RNA (lncRNA). Often neglected to the profit of mRNA, lncRNA play major roles in pathophysiology notably thanks to their regulatory role of gene expression.5 Authors therefore looked at the enriched pathways for both mRNA and lncRNA to get a better understanding of the biology of AP. Finally, authors have integrated their generated EV transcriptomic signatures of SAP to metabolomic data published by the team in a previous study.6 Such a strategy of combining transcriptomic and metabolomic signatures to identify predictive biomarkers deserves to be acknowledged and could very likely be used to predict the severity of AP, but also many other diseases. Indeed, the recent burst in transcriptomics at the single cell level has rendered the tools available to quickly and reliably determine gene expression notably in blood cells or vesicles.7 While metabolomics is considered as the “youngest of the -omics” and is still facing many technological challenges,8 ones should be optimistic considering the dynamism of this field and the potential applications once the analytical tools are available. The letter by Zhu et al. (REF) identifies TUBA1B and MIF as candidates to be used as EV predictive biomarkers for AP. These targets are in a way quite expected considering the current knowledge in the field and validate the approach used herein. TUBA1B is a well-known marker of pancreas-associated disorders from chronic pancreatitis to ductal adenocarcinoma.9 MIF is also a well-known marker for SAP10 and their presence in EV highlights relevance of such structures when considering pathophysiology. Circulating MIF-containing EV can be phagocytosed by alveolar macrophages, resulting in the activation of these lung resident macrophages leading eventually to the SAP-associated respiratory failure. Of note, other EV-transported factors have been shown to modulate macrophage activation such as S100A8/9, which bind to the macrophage receptor TLR411 and such high-throughput approach developed by Zhu et al. could assess the relative importance of these different molecules in SAP-associated co-morbidities. It should be noted that the analysis revealed a high interindividual heterogeneity in the expression of the proposed biomarkers including TUBA1B and MIF. This suggests that using a signature associated with SAP, which includes several genes and metabolites, would be more robust than a single marker, for which compensatory effects can dampen the relevance. Finally, while the letter by Zhu et al. proposes to use EV cargos as biomarkers of SAP, it is fundamental to remind that EV is a generic term covering various subtypes of secreted structures such as exosomes and microvesicles. However, specific issues arise when working with these entities, whose size and amount often make them difficult to obtain in relatively pure preparations and to characterize properly.12 In biomarker discovery, it is essential to eliminate contaminants in biofluids, like cell-free nucleic acids and lipoproteins, which interfere with EV biomarker interpretation. Interestingly, the authors employed EXODUS for small EV (sEV) isolation, a new ultrafiltration strategy, which they developed earlier. While EXODUS seems to outperform other isolation techniques in speed, purity, and yield of sEV,13 it will have to be shown whether this approach makes its path to the clinical applications in future. To conclude, this study offers a promising and exciting approach to find predictive biomarkers by using multi-omics to analyze the content of circulating EVs. While still limited by the absence of powerful analytical tools, this study paves the way for further studies highlighting the crucial role of EV and their cargos in pathophysiology.