Unravelling the mechanism of action of milk-derived EV by linking their proteome to relevant signalling pathways using an unbiased comprehensive bioinformatics approach

Background: EV are multisignaling components and their functionality is likely to occur from the combined actions of their constituents, rather than single molecules. Upon deciphering their functional effects in vitro, the major challenge is to define which molecules are responsible for their mode of action. Previously, we have published the human milk-derived EV proteome and we have shown that milk-EV can enhance epithelial cell migration on the one hand and suppress T-cell activation on the other hand. While various bioanalytical approaches are available that reveal the general functions of proteins in a data set, none of these show comprehensive signalling cascades. In this study, we aimed to select and combine the relevant tools to reconstruct the mechanism of action of milk-EV. Methods: Protein-protein interaction networks were made from the common milk-EV proteome (367 proteins) using STRING. Functional enrichment analysis in STRING was used to determine which proteins were involved in any of the relevant biological processes studied in the in vitro assays. Then, the signalling pathways were constructed using Uniprot entries and their associated sources from the individual proteins, supplemented with a general literature search (including KEGG and MetaCore pathways). Finally, interacting proteins were linked to these pathways. Results: Interestingly, individual proteins and protein clusters could be linked to specific signalling events and their function (activation or inhibition) fitted the observed in vitro data. For instance, proteins were identified that can stimulate the P38 migration pathway and cytoskeleton remodelling. Additionally, the milk-EV proteome contained a great number of proteins that are known to inhibit T cells via suppression of PI3K/ AKT, RAS/RAF and MAPK pathways. Depending on the specific pathway, regulation can take place early in the signalling cascade or throughout the entire signalling pathway. Summary/Conclusion: By integrating various bioanalytical approaches we were able to identify relevant proteins and determine their action and position in distinct signalling pathways. As expected, milk-derived EV contain a cluster of proteins of which their combined actions are likely to regulate intercellular communication.