Role of Surfactome Components of Extracellular Vesicles in Their Recognition by the Target Cell

Extracellular vesicles (EVs) are considered to be the most significant component of the mesenchymal stromal cell (MSC) secretome. Vesicles carry a set of proteins, bioactive lipids, and nucleic acids in their composition, protecting them with a lipid bilayer, and when taken up by target cells, they can demonstrate persistent regenerative effects. However, many researchers have shown that other components of the conditioned environment, besides BB, are also involved in the function of MSCs. Thus, to elucidate the mechanisms of regenerative effects of MSCs, it is important to evaluate the contribution of BBs to these processes.

BBs participate in intercellular communication, transferring proteins, bioactive lipids and nucleic acids from one cell to another. BBs produced by stem cells can deliver important information to target cells for tissue regeneration after damage [1].

The vast majority of articles studied in the course of this review reveal similar issues, such as vesicle biogenesis, their contents, size classification, and participation in intercellular interaction. Also, the fusion ligand-receptor interaction of vesicles with the recipient cell has been discussed in sufficient detail. Only a few articles mention the specificity of vesicle-cell interaction. However, when considering the given examples, one can notice that, in the end, the detailed description of the mechanism of vesicle recognition by a cell is reduced to the description of the ligand-receptor pair interaction.

Raising the question of the specificity of vesicle-cell interaction, we decided to study in detail the ligand-receptor pairs mentioned in most articles. Perhaps, the specificity of vesicle-cell/tissue interaction is not determined by their biogenesis, membrane composition, and internal composition, but merely by the specific electrostatic field of the tissue that attracts vesicles with appropriate charge characteristics [2, 3, 4].

On this basis, it can be assumed that specificity, as such, does not exist, but there is a volume distribution of differently charged vesicles in the electrostatic field of tissues [5].

To answer the question about the possibility of vesicle distribution, it is proposed to create a model predicting the biodistribution of BBs depending on their total surface charge. For this purpose it is necessary to prepare training and test (validation) samples.

Tabular data will be used for analysis, where the object for each record will be a molecular component of the extracellular vesicle membrane, and the attribute will be the location of its distribution confirmed by open data (the number of attributes may vary) (Table 1). Data collection is complicated by the lack of readymade arrays. The table is formed manually from the volume of a pre-prepared sample of literature on the given subject. The approximate size of the table is 2 (or more) * 1000.

For direct training of the model, the training (training) sample, on which the algorithm parameters are optimised, will include all cases of distribution prediction based on the knowledge of electrophoretic mobility of particles in a charged field and electrostatic effects. Seventy per cent of the data obtained will be used to create this sample (90% if insufficient data is available). To verify the accuracy of the model and to control model overfitting, the test sample will consist of experimentally validated examples of vesicle distributions. To create this sample, 30% of the obtained data will be used (10% if insufficient data).

The next step is to determine the most appropriate machine learning algorithm. After studying publications devoted to similar tasks, it was decided to use Random Forest (RF) machine learning algorithm as the most suitable for solving this type of problems. We plan to use Random Forest architecture, where each tree has Gini index as a quality criterion of tree branching, the depth of each tree will be considered as a hyperparameter, the optimum of which will be selected by early stopping, limiting the depth of the tree, setting the minimum acceptable number or cutting off branches.

Once the model is obtained, the available data on the composition of extracellular vesicles of MSCs cultured in the laboratory will be verified. The prediction result will then be validated experimentally.

1. Lopatina, T., Bruno, S., Tetta, C., Kalinina, N., Porta, M., & Camussi, G. (2014). Platelet-derived growth factor regulates the secretion of extracellular vesicles by adipose mesenchymal stem cells and enhances their angiogenic potential. Cell Communication and Signaling, 12(1), 1-12.

2. McKee T. J. et al. Extracellular matrix composition of connective tissues: a systematic review and meta-analysis //Scientific reports. — 2019. — Т. 9. — №. 1. — С. 1-15.

3. Stylianopoulos T. et al. Diffusion of particles in the extracellular matrix: the effect of repulsive electrostatic interactions //Biophysical journal. — 2010. — Т. 99. — №. 5. — С. 1342-1349.

4. Halper J., Kjaer M. Basic components of connective tissues and extracellular matrix: elastin, fibrillin, fibulins, fibrinogen, fibronectin, laminin, tenascins and thrombospondins //Progress in heritable soft connective tissue diseases. — 2014. — С. 31-47.

5. Lieleg O., Baumgärtel R. M., Bausch A. R. Selective filtering of particles by the extracellular matrix: an electrostatic bandpass // Biophysical journal. — 2009. — Т. 97. — №. 6. — С. 1569-1577.

6. Denzer, Kristin, et al. «Follicular dendritic cells carry MHC class II-expressing microvesicles at their surface.» The Journal of Immunology 165.3 (2000): 1259-1265.

7. Mallegol, Julia, et al. «T84-intestinal epithelial exosomes bear MHC class II/peptide complexes potentiating antigen presentation by dendritic cells.» Gastroenterology 132.5 (2007): 1866-1876. (177)

8. Nolte-‘t Hoen, Esther NM, et al. «Activated T cells recruit exosomes secreted by dendritic cells via LFA-1.» Blood, The Journal of the American Society of Hematology 113.9 (2009): 1977-1981. (204)

9. Chivet, Mathilde, et al. «Exosomes secreted by cortical neurons upon glutamatergic synapse activation specifically interact with neurons.» Journal of extracellular vesicles 3.1 (2014): 24722. (93)

10. Hoshino, Ayuko, et al. «Tumour exosome integrins determine organotropic metastasis.» Nature 527.7578 (2015): 329-335. (1730)

11. Purushothaman, Anurag, et al. «Fibronectin on the surface of myeloma cell-derived exosomes mediates exosome-cell interactions.» Journal of Biological Chemistry 291.4 (2016): 1652-1663. (101)

12. Zech, Daniela, et al. «Tumor-exosomes and leukocyte activation: an ambivalent crosstalk.» Cell Communication and Signaling 10.1 (2012): 37. (117)

13. Rana, Sanyukta, et al. «Toward tailored exosomes: the exosomal tetraspanin web contributes to target cell selection.» The international journal of biochemistry & cell biology 44.9 (2012): 1574-1584. (310)