This manuscript details methods for isolation as well as sensitive MS-based protocols for molecular analyses of EMVs from HBEC. Using these methods, 1179 unique proteins were identified in HBEC-EMVs. These methods, in combination with bioinformatics tools, were used to demonstrate that the isolated HBEC-EMVs (i) are not artifacts and contain intact, potentially post-translationally modified proteins, (ii) contain a majority of known exosome-specific proteins, as well as unique ‘signature’ proteins, (iii) contain proteins implicated in receptor-mediated transcytosis across the BBB.
Are EMVs artifacts?
EMVs were originally believed to be cellular artifacts and thought of as mechanisms through which cells discard inert debris [10, 30]. Many reports have since shown that EMVs are real, released cellular sub-compartments that consist of subsets of few protein families.
The BEC EMVs isolated by the described differential centrifugation method have not been morphologically characterized and may include both small (100 nm) and larger (up to 1000 nm) EMVs. Proteomic analyses of these EMVs reported in this study confirmed that HBEC-EMVs contain specific sub-sets of intact proteins, originating from the plasma membrane, endocytic pathway(s) and the cytosol. A subgroup of higher molecular weight proteins represented in EMVs appear to be post-translationally modified, compared to same proteins in whole-cell extracts, suggesting that they may originate from compartments characterized by high glycosylation, such as BEC luminal membranes or endocytic vesicles(s).
Specificity of HBEC-EMVs
The diagnostic potential of EMVs has been aggressively investigated [12, 13], since they contain tissue and disease-specific biomarker signatures [9, 21]. The tissue-specificity of EMVs is determined by specific RNA sequences and specific cell-surface molecules. The BBB-specific EMVs in body fluids could contain biomarkers useful for diagnosis or monitoring of brain diseases, since they could be ‘shed’ into the circulation from luminal membranes of BEC and potentially shuttled across the BBB from the abluminal side. We have found that about 20% of the HBEC-EMVs MS signal originated from proteins that were absent in exosomes from other cell types, suggesting that these proteins are potentially unique to HBEC-EMVs. Some of these included cell-surface proteins, including adhesion molecules and other cell-cell interacting molecules (Figure 4B).
The molecular signatures of EMVs can change under different biological conditions (in vitro insults or diseases state) [9, 21]. For example, we have observed that HBEC-EMVs molecular profile changed significantly in response to inflammatory insults (unpublished data). Therefore, monitoring HBEC-EMV-specific and disease-modified RNAs, proteins, glycoproteins and glycans in blood-derived EMVs by targeted ‘omics’ has potential diagnostic significance for CNS disorders. However, the utility of BBB EMVs as a source of disease-specific biomarkers remains to be validated in further in vitro and in vivo studies.
HBEC-EMVs as a vehicle for cell-cell communications in CNS
The cell-cell communication mediated by EMVs occurs predominantly by two processes: surface contact of vesicles with cells triggering donor cell signaling pathways, and/or delivery of vesicle content into the recipient cell (endogenous transduction). Consistent with these roles in cell-cell communication, the surface of EMVs is typically enriched in cell-targeting/adhesion molecules (e.g., tetraspanins and integrins), membrane trafficking proteins, proteins involved in MVB formation, antigen-presenting molecules (e.g., MHC class I and class II), and membrane cytokines, whereas their luminal content mainly consists of functionally-active RNAs (e.g., mRNA, microRNA, viral RNA), RNA-binding proteins, ribosomes, functionally-active proteins including enzymes (e.g., metalloproteases, metabolic enzymes) and cytokines (Figure 5A). HBEC-EMVs molecular make-up is consistent with this ‘generic’ exosome composition.
Given tight anatomical and functional integration of the cellular elements of the neurovascular unit, including BEC, pericytes, astrocytes and neurons, we surmise that BEC exosomes could play similar roles in transducing information among the cells in the neurovascular unit. The emerging role of neuronal exosomes in neuronal-glial communication and inter-cellular transfer of signaling miRNAs contributing to neuronal development and disease mechanisms has recently been reviewed . The in silico interactomics analyses confirmed that, based on molecular profile of HBEC-EMVs, they could engage in numerous cell-surface interactions with both astrocytes and neurons. Similar EMV-mediated communication could occur among BEC and peripheral inflammatory cells during processes of immune surveillance, rolling, adhesion and transmigration.
Are HBEC-EMVs BBB ‘transcytosing’ vesicles?
The first discovery of exosomes, almost three decades ago, involved detection of anti-TFRC antibody by electron microscopy in reticulocytes (summarized by Thery et al) in the following order: (i) on the surface of the cells and clathrin-coated pits, (ii) inside early endosomes, (iii) on the surface of internal vesicles of multivesicular endosomes, and finally (iv) on the released exosomes after fusion of the multivesicular endosomes with the plasma membrane. The RMT pathway and exosome formation have notable similarities. The HBEC-EMVs contained several receptors previously shown to carry macromolecules across the BBB via RMT, including TFRC, LRPs, LDLR, INSR and TMEM30A (Table 2). A hypothetical pathway by which these receptors and their ligands are ‘sorted’ into HBEC exosomes during luminal-abluminal RMT process is shown in Figure 1. A similar process may theoretically occur in the opposite direction, resulting in RMT receptor recycling, or ‘transfer’ of parenchymal exosomes into the circulation. The presence of known BBB RMT receptors in HBEC-EMVs might suggest that, among 524 ‘unique’ proteins identified in HBEC-EMVs, there may be additional novel and more specific RMT receptors exploitable for delivery of macromolecules across the BBB.
Interestingly, after the addition of RMT-triggering antibody FC5, we observed both a 4-fold increased amount of EMVs being produced by HBEC (based on total LC-MS signal; not shown) and presence of FC5 in these EMVs. This suggests that, under specific conditions, brain endothelial cells could regulate the amount of EMVs produced and ‘shed’ into abluminal or circulatory space.
EMVs as BBB drug-delivery vehicles
The possibility of using exosomes as drug-delivery vehicles, in particular for gene therapy with siRNAs, has gained significant attention in recent literature. In the study by Alvarez-Erviti et al, autologous exosomes derived from dendritic cells engineered to express exosomal membrane protein Lamp2b fused to the neuron-specific RVG peptide, were loaded with exogenous siRNA and shown to transduce brain parenchymal cells knocking-down the therapeutic target, BACE1, after systemic injection. Exosomes were also attempted as intranasal delivery vehicle for anti-inflammatory drugs . The advantage of self-derived exosomes over other lipid-based nanocarriers is that they are immunologically inert and are thought to possess ‘intrisic ability’ to cross biological barriers. Although this assertion requires further confirmation, the possibility remains that tissue-specificity of delivery could be improved by using homologous tissue exosomes. Therefore, HBEC-EMVs could potentially be exploited as brain-selective nanocarriers for therapeutic delivery across the BBB.