Research Article www.afm-journal.de Industrial By-Products As a Novel Circular Source of Biocompatible Extracellular Vesicles Cristina Lorca, Moisés Laparra, María Virtudes Céspedes, Laura Casaní, Sergi Florit, Mariona Jové, Natàlia Mota-Martorell, Elisabet Vilella, Xavier Gallart-Palau,* and Aida Serra* membranes remained largely and almost Extracellular vesicles (EVs) constitute an intricate system of molecular exclusively appreciated for these basic exchange that has recently gained tremendous interest. However, sustainable features, whereas their capacity to gen- sources of safe biological EVs remain scarce and elusive. This study explores erate minute exocytic structures of cel- and defines the use of food industry by-products (BP) as a circular source lular origin with outstanding properties was ignored.[1] These exocytic structures, of safe biocompatible EVs. Averaged diameter and molecular compositions currently known as extracellular vesicles indicate a large yield of exosomes and high abundancy of membrane lipids with (EVs), became once neglected as “cellular signaling capacity in these vesicles. Complex proteomes mimicking those dust” or “cell debris.”[1] Further multi- circulating in human blood plasma are also identified. Furthermore, BP-EVs disciplinary research, however, disclosed do not show relevant cytotoxicity and display excellent oral and intravenous the weightiness of these particles at the bioavailability together with specific organ targeting capacity. Collectively, it is time to protect specific cellular cargoes from extracellular hazards providing believed that the novel findings reported here will open substantial venues for them with the most appropriate signaling the use of BP as an optimal source of biocompatible nanovesicles in manifold molecules.[2] EVs, thus, constitute a highly applications of the biotechnological and biomedical fields. complex intercellular communication mechanism based on extremely resilient and efficient lipid membranes, the inclu- 1. Introduction sion of specific molecular cargoes, and the presence of preset membrane-embedded signaling molecules.[3] Additionally, Spontaneous assembling of lipid hydrophilic heads and hydro- these vesicles are shed and taken up by almost every cell type in phobic tails takes place in aqueous conditions by allowing the diverse kingdoms of life.[4] the defiant acts of definition and differentiation, basis of any Given these idiosyncratic properties, EVs have been proposed organismic existence. It is thus no wonder that bilayered lipid to perform essential functional role(s) in health and disease C. Lorca, M. Laparra, A. Serra M. V. Céspedes IMDEA-Food Research Institute Grup d’Oncologia ginecológica i peritoneal Campus of International Excellence UAM+CSIC Research Institute Old Cantoblanco Hospital Hospital de La Santa Creu I Sant Pau 8 Crta. Canto Blanco, Madrid 28049, Spain Biomedical Research Institute Sant Pau (IIB-Sant Pau) E-mail: aida.serra@imdea.org Barcelona 08041, Spain C. Lorca, X. Gallart-Palau L. Casaní, S. Florit Biomedical Research Institute of Lleida Dr. Pifarré Foundation Research Institute Hospital de La Santa Creu I Sant Pau (IRBLLEIDA) – +Pec Proteomics Research Group Biomedical Research Institute Sant Pau (IIB-Sant Pau) (+PPRG) – Neuroscience Area – University Hospital Arnau Barcelona 08041, Spain de Vilanova (HUAV) – Health Campus M. Jové, N. Mota-Martorell Psychology Department Department of Experimental Medicine University of Lleida Lleida Biomedical Research Institute (IRB Lleida) Lleida 25198, Spain Lleida University (UdL) E-mail: xgallart@irblleida.cat Lleida 25198, Spain C. Lorca, X. Gallart-Palau, A. Serra E. Vilella Proteored – Instituto de Salud Carlos III (ISCIII) Hospital Universitari Institut Pere Mata Lleida 25198, Spain Reus, Tarragona 43204, Spain The ORCID identification number(s) for the author(s) of this article E. Vilella can be found under https://doi.org/10.1002/adfm.202202700. Institut d’Investigació Sanitària Pere Virgili (IISPV) Reus, Tarragona 43204, Spain © 2022 The Authors. Advanced Functional Materials published by E. Vilella, X. Gallart-Palau Wiley-VCH GmbH. This is an open access article under the terms of the Centro de investigación Biomédica en Red en Salud Mental CIBERSAM ­Creative Commons Attribution-NonCommercial License, which permits Instituto de Salud Carlos III use, distribution and reproduction in any medium, provided the original Madrid 28029, Spain work is properly cited and is not used for commercial purposes. DOI: 10.1002/adfm.202202700 Adv. Funct. Mater. 2022, 32, 2202700 2202700 (1 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de conditions, some of which have already been uncovered.[5] Sim- BPs were supposed to exclusively contain EVs derived from ilarly, these vesicles have been identified as an optimal head- plants and microorganisms (yeast and/or bacterial strains) stream of organic biocompatible nanocarriers with a growing (coded as BP1, BP3, and BP4), while BP2 was supposed to con- interest, for their potential applications, within several areas of tain EVs derived from microorganisms and mammalian cells the bioeconomy and biomedical fields (i.e., food and supple- (further details on BP sources are provided in the Experimental ment industries, biotech industries, pharmacy, etc.).[6] Although Section). BP-EVs were initially isolated from their respective BP these vesicles display advantageous and compelling features matrices by three different isolation methods: protein organic as nanocarriers including low antigenicity, ability to trespass solvent precipitation (PROSPR), molecular weight cutoff biological barriers, stable circulatory capacity, and organ tro- (MWCO), and single-step size exclusion chromatography (SEC) pism, the applicability of EVs as optimal nanocarriers still as shown in Figure 1a. The selection of these methods was faces substantial challenges.[6] Among these safety, scalability based on industrial scalability, methodological simplicity, lower and identification of biocompatible physicochemical features execution costs, and standardization. stand out as drawbacks.[7] This is in part due to the fact that All tested isolation strategies provided substantial yields of the most researched sources of nanocarrier EVs are immortal- BP-EVs from all analyzed samples (Figure 1b), in turn dem- ized cell lines, a limiting source with significant concerns in onstrating the consistent presence of these vesicles in these terms of human safety and resource availability.[8] Similarly, the sources. However, MWCO generated significantly improved optimal progress of the EVs mimetics field, which consists of yields on solid BP sources (BP3 and BP4) (Figure 1b-I–IV), laboratory-based generation of artificial EVs and liposomes with while the difference on the quantity of EVs obtained among biocompatible nanocarrier potentialities, faces substantial chal- MWCO and the two other methods tested in liquid sources lenges based on the current available EVs sources.[9] (BP1 and BP2) was not statistically significant (Figure 1b-I–IV). Foods are known to contain a vast population of EVs, nat- In addition, the facts that MWCO presents technical advan- urally present and ingested by humans daily.[10] Presence of tages, such as lack of requirement of organic solvents in its EVs in foods has been demonstrated in multiple sources, use and that it is easier to standardize and scale from an including milk and dairy products,[11] edible plants, plant- industrial perspective, were also taken into consideration. derived products, and fermented foods.[11] Furthermore, EVs Hence, MWCO was chosen as the most optimal methodology from foods seem to tick the right boxes at the time to opti- to obtain BP-EVs in this context and thus this isolation mally perform as nanocarriers in terms of safety, biocompat- method was used in all subsequent studies performed here, as ibility, stability, and target specificity.[12] However, little is still depicted in Figure 1a. known regarding appropriate sources of food EVs, optimal Morphometric and ultrastructural characterization of the methods of obtainment, potential scalability, and advanced obtained BP-EVs indicated predominant particle sizes within physicochemical properties of these vesicles.[12] Similarly, the the exosomal range for all tested BP-EVs isolation strategies use of food industry by-products (BPs) to obtain biocompat- (global averaged φ < 250 nm) as indicated in Figure 1c and ible EVs, to the best of our knowledge, has not been explored Figure S1, Supporting Information. However, collectively, BP1- yet. These BPs and residues may become an optimal source of EVs showed significantly larger averaged diameters (average φ EVs with the collateral capacity to contribute to the progress 222.4 ± 22.7 nm) than BP2-EVs (average φ 132.7 ± 21.96 nm) and of circular economy, a research priority worldwide.[13] Cir- BP3-EVs (average φ 121.1 ± 8.545 nm) (Figure 1c). It was also cular economy promotes reduction, recycling, and reusage of observed that SEC isolated larger EVs in diameter from BP4 industrial waste in front of the end-of-life concept, improving than the rest of isolation methods employed (Figure 1d). As pre- overall sustainability. Food industries generate high amounts viously described in other studies, different isolation method- and a huge variety of BPs that have a negative environmental ologies tend to generate variation in the profiles of the obtained impact and high cost of management.[14] Thus, in this study, EVs.[15] Thus, the significant variations observed here regarding we have optimized and assessed different isolation methods the use of SEC contributed to the election of the MWCO meth- to define EVs yield and scalability considering discarded food odology as preferential, which allows the obtention of EV popu- industry BPs as sources of circular EVs. Additionally, we have lations with diameters of closer compliance with the exosomal performed in-depth physicochemical characterization of the range for BP4 (Figure 1d). Finally, ultrastructural characteriza- obtained food industry by-product-derived EVs (BP-EVs) with tion of BP-EVs demonstrated that these vesicles present spher- special emphasis on lipidome and proteome compositions. ical morphology as shown in the representative micrographs in Finally, we have assessed the bioavailability, biodistribution, Figure 1e. Additionally, ultrastructural analysis confirmed low organ targeting capacity, and toxicity of these vesicles through presence of contaminant particles in isolated BP-EVs prepara- in vivo and in vitro studies. tions (Figure 1e). 2. Results 2.2. Lipidome Compositions of BP-EVs 2.1. Industry By-Products As a Novel Source of EVs Lipids are essential components of EVs predominantly acting as structural components of the lipid bilayer membrane.[16] To define whether industry BPs can become a relevant source Here, we performed next-generation untargeted lipidomics to of potentially biocompatible EVs, four different and representa- define the lipid compositions of the obtained BP-EVs. Identified tive food industry-derived BPs were analyzed. Three of these molecules were initially classified in lipid families as shown Adv. Funct. Mater. 2022, 32, 2202700 2202700 (2 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 1. Experimental flow chart and morphometric characterization of BP-EVs isolated by PROSPR, MWCO, and SEC. a) Diagram that depicts the whole experimental flow chart procedures of the study. b) Particle concentrations detected by nanoparticle tracking analysis (NTA) in I. BP1-, II. BP2-, III. BP3- and IV. BP4-EVs using the three tested EVs isolation strategies. Reported data was based on ten sequential analyses (without any field nor ­parameter restrictions) from three independent experiments (n = 3). c) Averaged particle diameter of BP-EVs. Gray external line represents the standard deviation of the mean. d) Averaged particle diameter of BP-EVs depending on the EVs enrichment strategy used. Gray external line represents the standard deviation of the mean. e) Ultrastructural transmission electron microscopy micrographs of I. BP1-EVs and II-III. BP2-EVs. * Indicates signifi- cant differences at p ≤ 0.05; ** Indicates significant differences at p ≤ 0.001. in Figure 2. Eight lipid families were commonly identified in family were found more abundant in BP-EVs of plant/yeast all BP-EVs lipidomes, which included fatty acyls (FA), glycer- origin (BP1) (Figure 2i). Significant differences on the relative ophospholipids (GP), glycerolipids (GL), sphingolipids (SP), amount of lipid families were also encountered for the GL spe- polyketides (PK), sterol lipids (ST), saccharolipids (SL), and cies, which revealed higher abundancy in BP2 compared to BP1 prenol lipids (PR) (Figure 2a–h). GP was consistently identified and BP3 (Figure 2i). as the most abundant lipid family in all BP-EVs metabolomes In a related vein, multivariate partial least squares discri- analyzed (Figure 2i); and the relative amounts of this lipid minant analysis (PLS-DA) of the molecular features identified Adv. Funct. Mater. 2022, 32, 2202700 2202700 (3 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 2. Categorization of lipid species from BP-EVs lipidomes. a–g) Heatmaps depicting the family categorization of lipid species identified from BP-EVs lipidomes, including fatty acyls (FA), glycerophospholipids (GP), glycerolipids (GL), sphingolipids (SP), polyketides (PK), sterol lipids (ST) and prenol lipids (PR). h) Heatmap that shows data from uncategorized lipid species in BP-EVs lipidomes. Species intensity in all heatmaps is expressed as relative intensity based on MS counts. i) Quantitation of lipid families identified in BP-EVs based on the sum of mass MS of the molecular species within every categorized lipid family. Multivariate partial least square discriminant analysis (PLS-DA) of the molecular features identified in BP-EVs lipidomes showing j) the BP-EVs sample separation according to lipidome compositions and k) important lipid features of component 1 based on Variable Importance in Projection (VIP) scores. k) Degree of unsaturation of lipids identified from the discriminative lipidome detected in every BP-EVs characterized. l) Quantifications are based on sum of intensities detected by LC-MS/MS. Significance was assessed by parametric ANOVA or by non-parametric Kruskal Wallis with a minimum significance level p < 0.05. Mean comparisons between BP-EVs groups were assessed by posthoc test. * Indicates significant differences at p ≤ 0.05; ** Indicates significant differences at p ≤ 0.001; # Indicates non- parametric significant differences at p ≤ 0.05. by liquid chromatography-mass spectrometry (LC-MS) dem- species could be confidently validated (Tables S2 and S3, onstrated presence of a highly specific lipidome signature Supporting Information). Of note, cholesterol ester (CE), tri- within each of the BP-EVs analyzed (Figure 2j). Of note, BP3 acylglycerols (TAG), and cardiolipin (CL), all lipid molecules was the EVs subset with higher heterogenicity in the lipid traditionally derived from contaminant particles in EVs prepa- compositions observed (Figure 2k). A total of 186 molecular rations,[17–19] did not reach the significance threshold in any of features were differentially detected by liquid cromatography the samples analyzed. with tandem mass spectrometry (LC-MS/MS) (Table S2 and Finally, the degree of unsaturation of the differentially identi- Dataset S1, Supporting Information). From these, 72 lipid fied lipids in BP-EVs was also investigated (Figure 2l). BP1-EVs Adv. Funct. Mater. 2022, 32, 2202700 2202700 (4 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de lipidomes were enriched in saturated fatty acids (SFA), as well the total proteomes (Figure 3e). Similarly, BP4-EVs showed a as in polyunsaturated fatty acids (PUFA) with low degrees of significantly higher ratio of membrane proteins compared to unsaturation. SFA were also present in BP2-EVs, together with BP3-EVs (Figure 3e). The list of identified membrane proteins monounsaturated (MUFA) and PUFA acids of up to 6 degrees in each of the BP-EVs proteomes is included in Dataset S4 of unsaturation (Figure 2l). (Supporting Information). Relative quantification of the abun- dance of these identified BP-EVs membrane proteins revealed enolases as the family of EVs membrane proteins more ubiqui- 2.3. Proteome Characterization of BP-EVs tously present across the analyzed BP-EVs and more abundant together with immunity-related membrane proteins and gly- Unbiased discovery-driven proteomics to characterize the coproteins (Figure 3f). Remarkably, although BP3 showed the obtained BP-EVs proteome compositions was also performed lowest ratio of membrane proteins compared to rest of BP-EVs by using next-generation 4D state-of-the-art technology,[20,21] proteomes, the abundance level of a subset of specific mem- BP1- and BP3-EVs were identified with richer proteomes brane protein families was significantly higher in these BP-EVs (2372 ± 23 and 2625 ± 98 proteins, respectively) than BP2 and including tetraspanins, apolipoproteins, glycoproteins, and BP4 (243 ± 4 and 402 ± 113 proteins, respectively) (Figure 3a). immunity-related proteins (Figure 3f). Total proteins compared to total lipids in BP-EVs revealed that BP1-EVs contained the highest protein-to-lipid ratio whereas BP4-EVs displayed the lowest ratio (Figure 3b). 2.6. Similarities of BP-EVs with EVs in Human Plasma Subsequently, we evaluated whether the organism of origin and in Consumed Foods of the identified BP-EVs proteomes modulated the proteome compositions (Figure 3c). These analyses revealed that 99% of As expected, EVs from mammalian origin (BP2) displayed fur- BP1-EVs proteins came from yeast, specifically from Saccharo- ther complex proteome compositions in terms of biological myces cerevisiae (Figure 3c). Thus, no EVs from plant origin were processes, mirroring the proteome compositions in circulating found in BP1-EVs proteomes. Similarly, BP2-EVs proteomes human EVs (Figure 4a). On the contrary, plant and bacterial/ were identified with mammalian origin (Bos taurus) (in this yeast-derived EVs showed higher proportion of metabolic and particular case, proteins were also detected from the bacterial cellular processes-related proteins compared to mammalian- origin, though not consistently identified in at least three inde- derived EVs and scarce biological processes-related similarities pendent replicates and thus not considered) (Figure 3c). From with human circulating EVs (Figure 4a). Further analysis of EVs BP3, which originally contains a complex mixture of lactic acid proteomes from mammal origin (BP-2) showed presence of bacteria and yeast, 72% of identified EVs proteins were from the proteins involved in key biological processes, such as proteins genus Kamagataeibacter (Figure 3c); whereas in BP4-EVs, 68% of the HIF-1-alpha regulatory pathway and MHC-I class-related of the proteomes identified had plant origin (Vitis vinifera) and proteins (Figure 4b). Similarly, in EVs from this BP, the proteins the remaining 32% possessed yeast-derived origin (Figure 3c). with modulatory capacity toward lipid digestion, lipid mobiliza- Of note, although microbial EVs proteins were less rich in com- tion, and transport of lipids were more abundant (Figure 4b). plexity and diversity in BP4-EVs, these were far more abundant Finally, these analyses also revealed significant enrichment of in the total proteomes of this BP compared to plant-derived EVs specific protein domains in BP2-EVs, highlighting the 2.31-fold proteins as shown in Figure 3c. enrichment in signal peptide-containing proteins as shown in Figure 4b. In order to comparatively analyze BP-EVs with their food- 2.4. BP-EVs Proteomes Differ from Previously Identified derived counterparts, we analyzed BP2-EVs and BP4-EVs pro- EVs Proteomes teomes toward their respective previously reported bovine milk-derived EVs proteomes[24] and grape-derived EVs pro- We then investigated whether BP-EVs proteomes are alike to teomes[10] (Figure 5). BP2 and BP4-EVs were selected here as previously identified EVs proteomes curated in the EVs special- representative EVs populations to be compared with their main ized databases Vesiclepedia and Exocarta.[22,23] Of note, only food-derived EVs counterparts, given that the proteome com- EVs from the mammal BP source (BP2) obtained a relevant position of food-derived EVs isolated from the foods associated match (76%) in these comparative analyses (Figure 3d). Rest of with the generation of BP2 and BP4 are well documented and BP-EVs proteomes obtained matches lower than 3% (BP1 and easily available in front of food-derived EVs counterpart pro- BP4) or not match at all (BP3). Complete lists of uncategorized teomes from the rest of BP-EVs sources analyzed. For BP2-EVs, proteins generated in these comparative studies can be found 79% of the proteins consistently identified in our study were in Dataset S3 (Supporting Information). previously identified in bovine milk-derived EVs as shown in Figure 5a. In-depth ontology categorization of these common and uncommon proteins (Figure 5b,c) indicates no differences 2.5. Diversity and Abundance of Membrane Proteins in BP-EVs between BP2-EVs proteomes and milk EVs proteomes. Simi- larly, although grape-derived EVs proteomes have not previously Membrane proteins represent lower portions of the total pro- been in-depth characterized,[10] comparative analyses between teomes in BP1-, BP3-, and BP4-EVs (ratio membrane pro- BP4-EVs and the partial available data on grape-derived EVs pro- teins/total proteome <0.5), with the exception of BP2-EVs, in teomes also demonstrated high similarities between these two which membrane proteins represent a higher >1 portion of EVs sources (Figure 5d and Table S4, Supporting Information). Adv. Funct. Mater. 2022, 32, 2202700 2202700 (5 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 3. 4D proteomic characterization of BP-EVs by LC-MS/MS. a) Bar graph displaying the number of total proteins identified in each characterized BP-EVs proteome. b) Total intensity of lipids and proteins identified within each BP-EVs proteomes and metabolomes characterized. Total intensities were calculated based on the sum of intensities obtained from individual lipid identifications and the sum of spectral count of each protein in BP-EVs proteomes. Total intensity values were log2 transformed for a representative purpose. c) Parts of the whole graphics showing the percentage propor- tions that represent every organism in the total proteomes identified from: I. BP1-EVs, II. BP2-EVs, III. BP3-EVs and IV. BP4-EVs proteomes. Numbers in brackets indicate the percentage represented by each organism within BP-EVs proteomes. Only proteins consistently identified in at least three independent replicates are represented. d) Venn diagrams that depict overlapping between proteins identified in BP-EVs proteomes (I. BP1-EVs, II. Adv. Funct. Mater. 2022, 32, 2202700 2202700 (6 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 4. Comparative gene ontology (GO) analysis between BP-EVs proteomes and human blood plasma circulating EVs proteomes. a) Categoriza- tion of proteins according to the biological system to which they belong. Central circular graph depicts the categorization of proteins in circulating human EVs proteomes whereas external circular panels depict the categorization of proteins within each analyzed BP-EVs. Categories are distinguish- able through different colors. The color of the external lines (located at the graph corners) indicates the source of origin of EVs. Stronger gray shades denote higher similarities between EVs proteomes. b) Detailed GO categorization of the most similar EVs proteomes identified between BP-EVs (BP-2) and human circulating EVs. Panel I. depicts categorization of common proteins and fold enrichment in the total BP-EVs proteome based on biological systems and II. depicts categorization of common proteins and fold enrichment in the total BP-EVs proteome based on protein domains. 2.7. In Vitro Assessment of BP-EVs Cytotoxicity prior to their specific use in biomedical and biotechnological applications in humans. Similarly, cell viability assessed by BP-EVs were assessed on their potential ability to exert signifi- the MTT assay did not show significant cytotoxicity exerted by cant harmful features on basic cell metabolism processes and BP-EVs compared to controls. cell viability. Representative BP-EVs from animal origin (BP- EVs-A) and from vegetal/yeast origin (BP-EVs-V) were used in these experiments as detailed in the Experimental section. 2.8. In Vivo and Ex Vivo Investigation of BP-EVs Bioavailability Therefore, potential cell cytotoxicity of BP-EVs was investigated and Biodistribution by using immortalized human colorectal Caco-2 cells by the implementation of the cytotoxicity assays neutral red (NR) and To scrutinize whether the obtained BP-EVs show relevant bio- 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide compatible abilities (e.g., targeting of specific organ tissues, (MTT). As shown in Figure 6a, BP-EVs did not display signifi- circulation in biological fluids, ability to cross biological and cel- cant decay on the absorbance obtained by NR uptake assay of lular barriers, oral bioavailability, and postdigestion capacity) we lysosomal activity compared to negative (CN) and positive con- performed in vivo whole-body noninvasive fluorescent imaging trols (CP), although some differences on NR assay results were (FLI) experiments followed by ex vivo validation. Representa- found between the performed assays as shown in Figure 6a. Of tive BP-EVs from BP-EVs-A and BP-EVs-V were used in these note, an apparent although not significant decreasing tendency experiments as detailed in the Experimental section. Overall in NR absorbance for BP-EVs-V was also observed (Figure 6a), abdominal biodistribution of fluorescent labeled BP-EVs but this was not seconded by data obtained in the complemen- was predominantly observed at 4 h postoral administration tary assay MTT (Figure 6b). This fact indicates that further (Figure 6c,d), time point when significant differences in whole- complex metabolic assays might also be useful at the time to body fluorescent signal between BP-EVs-A and BP-EVs-V were better appreciate the cellular effects of these EVs, especially also identified (Figure 6e). Of note, the observed FLI signaling BP2-EVs, III. BP3-EVs and IV. BP4-EVs) and EVs proteins identified and curated in the EVs specialized databases Vesiclepedia and Exocarta. e) Graph bar that displays ratios of membrane proteins/other proteins within BP-EVs proteomes. These ratios were calculated based on data obtained from gene ontology analyses. f) Functional and molecular family categorizations of membrane proteins identified in BP-EVs proteomes including their relative abundance within proteomes. Significance was assessed by parametric ANOVA with a minimum significance level p < 0.05. Mean comparisons between groups were assessed by post-hoc test. * Indicates significant differences at p ≤ 0.05; ** Indicates significant differences at p ≤ 0.001. Adv. Funct. Mater. 2022, 32, 2202700 2202700 (7 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 5. Comparative study of BP-EVs proteomes with their existing previously reported counterpart food-derived EVs proteomes. a) Number of common proteins and proteome proportions identified between BP2-EV proteomes and its food-derived EVs counterpart bovine milk. Proteomics data on bovine milk EVs were obtained from ref. [24]. b) Gene ontology (GO) analysis of BP2-EVs proteins that are common with proteins identified in bovine milk-derived EVs. Data were categorized based on molecular function. c) GO analysis of BP2-EVs proteins that were not previously reported as identified in bovine milk-derived EVs. GO data were categorized based on molecular function. d) GO analysis of BP4-EVs proteins that are common with proteins identified in grape-derived EVs. Data available on grape-derived EVs proteomes were obtained from ref. [10]. GO data were categorized based on molecular identity and cellular function. differences within the thoracic region were not specifically ana- that did not show significant changes in fluorescent biodistribu- lyzed due to the difficulties encountered to confidently attribute tion are shown in Figure S2 (Supporting Information). these to any specific organ within this region (Figure 6c,d). It Finally, data from the experiments performed to compare the was also observed that biodistribution signaling peaks were biodistribution capacity of BP-EVs based on the via of adminis- respectively and differentially found at 4 h and 24 h post-oral tration (tail-intravenous versus oral), which indicated excellent administration for BP-EVs-V and BP-EVs-A (Figure 6e). Of oral bioavailability of BP-EVs-A and BP-EVs-V for both analyzed note, an apparent fluctuating behavior was observed regarding BP sources, were included in Figure S3a–c (Supporting Infor- the FLI signal at the initial time-points analyzed, which was mation). Higher hepatic accumulation of BP-EVs was observed attributed to standard pixel intensity corrections. following to tail-intravenous administration (i.v.) compared to Ex vivo organ validation at 25 h post-oral administration of oral administration (Figure S3d,e, Supporting Information), BP-EVs, indicated differential targeting of these vesicles specifi- therefore oral administration was revealed as the most optimal cally in brain, liver, and bone tissues (Figure 6f–h). Although all administration via to achieve broader whole-body biodistribu- analyzed BP-EVs sources demonstrated the ability to specifically tion of BP-EVs (Figure S3d,e, Supporting Information). target brain tissues by ≈20% of the total biodistributed vesicles (Figure 6h), only BP-EVs-V specifically targeted skeleton tissues (Figure 6f–h). Similarly, BP-EVs-A predominantly remained 3. Discussion within hepatic tissues at 25 h post-oral administration (≈90% of total biodistributed vesicles) whereas only 60% of BP-EVs-V In the present work, to the best of our knowledge, we provide were trapped in the liver (Figure 6h). The rest of tissues analyzed first vast evidence that industry BPs can become a relevant Adv. Funct. Mater. 2022, 32, 2202700 2202700 (8 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 6. Biocompatibility analysis of BP-EVs by assessing in vitro cytotoxicity, in vivo biodistribution, and organ targeting capacity. In vitro cytotoxicity was assessed in Caco-2 cells seeded in 96-well cell plate at a cell density of 15 000 cells per well treated with different concentrations of BP-EVs of animal origin (BP-EVs-A) and vegetable/yeast origin (BP-EVs-V) (Treatment 1 (T1): 108 particles per mL−1, Treatment 2 (T2): 109 particles mL−1 and Treatment 3 (T3): 1010 particles mL−1) by a) Neutral Red uptake assay of lysosomal activity and b) cell viability MTT assay. The negative control was performed treating cells with PBS (CN: 50%) and the positive control (CP) treating cells with two concentrations of DMSO (CP1: 10% and CP2: 30%). Continuous lines between bars indicate significant differences. Whole-body noninvasive monitoring of the biodistribution of fluorescent. c) BP-EVs-A and d) BP-EVs-V at 0, 2, 4, and 24 h upon oral administration. FLI corresponds to fluorescent intensity expressed in radiant efficiency. e) Total whole-body accumulation of fluorescent BP-EVs at 0, 2, 4, and 24 h quantified by measurements of FLI. f) Ex vivo tissue fluorescent imaging at 25 h post-oral administration of BP-EVs-A or BP-EVs-V. g) Total tissue accumulation was quantified by the measurement of FLI. Only tissues that showed significant differences compared to the positive control corresponding to animals treated with the fluorescent dye (CD) are displayed. Results from other investigated tissues are shown in Figure S3 (Supporting Information). h) Relative biodistribution of BP-EVs-A and BP-EVs-V at 25 h post-oral administration, considering accumulation in tissues that showed significant differences compared to CD. Significance was assessed by parametric ANOVA with a minimum significance level p < 0.05. Mean comparisons between BP-EVs groups were assessed by post-hoc test. * Indicates significant differences at p ≤ 0.05; ** Indicates significant differences at p ≤ 0.001. Adv. Funct. Mater. 2022, 32, 2202700 2202700 (9 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de source of EVs, predominantly of exosomes, with a global have previously been involved in quality control of EVs prepa- average diameter <170 nm and expressing most of the relevant rations and may act as indicators of EVs bioactive capacity.[34] molecular markers associated with this specific subfamily of Our proteome data also indicates that BP-EVs are predomi- EVs, including ESCRT machinery tetraspanins and other mul- nantly produced by probiotic microbiota strains, such as lactic tivesicular bodies-related proteins.[22,23,25] The exploitation of acid bacteria,[35] which associates our uncovered vesicles with industry BPs as a source of EVs as we investigate here, goes immune system reinforcement as opposed to proinflammatory paired with the use of an optimized, scalable, and efficient features generally associated with commensal or symbiotic EVs isolation strategy.[12] In this study, PROSPR, MWCO, and bacterial strain-generated EVs.[36,37] SEC were investigated as they share common basic features, Intact whole body FLI and ex vivo studies, as performed here, such as scalability and lowered costs of execution.[12] Opti- serve as proof of concept of the biocompatible nature of BP-EVs mized MWCO displayed the most efficient and simple way to and their ability to act as preferential organic nanocarriers. obtain BP-EVs with improved yields even in solid BP sources. Furthermore, our in vitro assays with human colorectal Caco-2 Furthermore, we believe that this isolation strategy can be cells indicated that BP-EVs did not exert relevant harmful fea- easily industrially adapted, based on our reported parameters, tures for cell metabolism nor cytotoxicity, even at higher dos- through optimization of its available industrial sibling tangen- ages. FLI experiments demonstrated excellent time-linked bio- tial flow filtration. distribution of i.v. and orally administered BP-EVs and specific Our analyses also revealed that BP-EVs contain all the innate organ targeting capacity of these vesicles. Based on the recent grandeur of food EVs with the important additions of wide systematic review performed by Chamley and co-workers,[38] availability, significantly reduced cost of obtention and higher the peak of administered BP-EVs between 2 and 24 h in intact sustainability. Morphologically, BP-EVs do not differ signifi- animals was also observed here in abdominal regions that may cantly from their food EVs counterparts. The average diameter correspond to the liver.[38] However, we did not observe any of BP2-EVs was consistent with those previously identified in attributable peak within thoracic regions matching the lungs. bovine milk,[26] whereas the average diameter of BP4-EVs was Similarly, BP-EVs signal was detected in head regions that cor- also in consonance with EVs isolated from grape,[27] the unique respond to the brain within all-time intervals measured from BPs that possess previously reported data from counterpart 2 h. These findings are consistent with previous reports on bio- food-derived EVs. In terms of biochemistry, BP-EVs lipidomes distribution of exosomes administered preferentially intrave- revealed GP, GL, and FA as the lipid families more abundantly nously,[38] though, we also demonstrate here that orally admin- present. Richness in phospholipid families, as identified here, istered BP-EVs display highly similar biodistribution patterns. is a common feature of biological eukaryotic membranes[18] Furthermore, as a novelty of this study, we confirmed by ex and a typical feature of EVs from bacteria,[28] plants, fungus,[29] vivo FLI that BP-EVs remain in brain tissues beyond the time- and human origins.[3] Moderate enrichment in ceramides point of 24 h, and that these vesicles preferentially target bone, (Cer), a family of lipids which its presence is linked to biogen- brain, and liver tissues. Consistent with previous reports,[39] esis of exosomes,[30] was also detected in BP-EVs. Similarly, distinct EVs sources display differential biodistribution patterns abundance of certain lipid families in EVs is associated with and particular tissue or cell targeting properties. These tissue their tumorigenic origin[19] and a deleterious impact on recip- targeting discrepancies are caused by specific EVs membrane ient cells.[31] Thus, it has been shown that the families of lipids proteins and lipids that require specific elucidation on the basis ST, SP, and GP are typically enriched in EVs of tumorigenic of previously performed proteome- and lipidome-wide charac- origin[19] while FA, GL, and PR abound in EVs of nontumori- terizations, as those reported in this study, that should pave the genic origin. Our obtained results were consistent with these way for further studies in the field. previous findings, as we found that BP-EVs preparations were significantly enriched in FA, GL, and PR lipids, fact that rein- forces the links between BP-EVs and their food-derived coun- 4. Conclusions terparts and their capacity to avoid potential safety concerns. Exosomes have also been associated with presence of desatu- EVs have vastly demonstrated outstanding potential as next-­ rated molecular species in their bilayer membranes,[30] finding generation delivery nanocarriers in manifold biotechnological and also consistent with our lipidome data from BP-EVs. Addi- biomedical applications. However, current sources of EVs present tionally, analysis of the degree of saturation of lipids in these substantial challenges including availability, cost, and safety as the vesicles revealed that in general terms BP-EVs are enriched in most relevant. Here, we demonstrate for the first time the pres- PUFAs with low degree of unsaturations; what confers their ence of EVs in food industry by-products. These food industry membranes with manifold and dynamic cell signaling abili- by-products, paired as detailed with an optimized and scalable ties[32] that may be specifically investigated in further studies method of EVs isolation, can become a sustainable source of based on the data generated here. In a related vein, study of EVs that surpasses any current existing source. Similarly, BP-EVs BP-EVs proteomes indicated that these vesicles are enriched show exosomal predominance, outstanding biocompatible fea- in specific coiled coil domain-containing proteins,[33] surface tures, capacity to trespass biological barriers, and excellent oral receptors, and cell adherent factors, which based on our results bioavailability. These features, extensively detailed in this work, identifies BP-EVs from mammal origin as highly similar to place BP-EVs as an optimal source of next-generation delivery EVs circulating in the human body. Enolases or GPI-anchoring nanocarriers with prospective manifold applications in biotech- proteins, among others, were the most ubiquitous proteins nology and biomedicine and extraordinary influence on the cur- present in all BP-EVs characterized here. These molecules rent and prospective translational mainstreams of the field. Adv. Funct. Mater. 2022, 32, 2202700 2202700 (10 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de 5. Experimental Section Nanoparticle Tracking Analysis: Isolated BP-EVs, as stated above, were subjected to extensive volumetric and morphometric characterization by Reagents: Caco-2 cells were obtained from DSMZ (Braunschweig, NTA. NTA was done as previously described with minor modifications.[42] Germany). All reagents were purchased from Sigma-Aldrich (St. Louis, Here, NTA analysis of BP-EVs was performed using a ZetaView BASIC MO) unless otherwise specified. Sequencing-grade modified trypsin was NTA—Nanoparticle Tracking Video Microscope PMX-120 operated obtained from Promega (Madison, WI). with ZetaView software (version 8.05.11) (Particle-Metrix, Ammersee, Food Industry By Products: Four different fermented food industry Germany). For each BP-EVs sample analyzed, 10 independent readings BP (FFIBP) were used in this study as potential source of circular EVs: covering the entire sample run were achieved. BP-EVs samples were i) brewer spend yeast derived from the beer fermentation production, diluted to 1/1000 prior to NTA analysis in PBS to reduce the number of generously provided by Mahou San Miguel brewery (Alovera, Spain); ii) particles within the observational field at <200 per image. whey from unbranded plain yogurt purchased in a local supermarket; Ultrastructural Analysis of BP-EVs: Isolated BP-EVs preparations from iii) symbiotic culture of bacteria and yeast (SCOBY) derived from the the representative sources (BP1 and BP2) were mounted on Cu-Formvar- production of tea fermented drink (Kombucha) on-line purchased Carbon grids, let for 20 min at RT, washed in distilled water, and fixed by from Kirandia—Ingenio Corporate Group, S.L. (Madrid, Spain) and 1% glutaraldehyde in PBS for 5 min. Subsequently, samples were stained iv) fermented wine pomace from red grapes derived from the wine with uranyl-oxalate solution for 5 min. Finally, BP-EVs samples were fermentation production, generously provided by Castell del Remei embedded by methyl-cellulose-uranyl-oxalate and dried for permanent winery (Penelles, Spain). preservation. Electron micrographs were collected using a Jeol Jem1010 Preparation of Food Industry By-Products Prior EVs Isolation: Prior electron microscopy. at a voltage of 80 kV. to EVs isolation, BP were pre-processed as follows: brewer spend Characterization of BP-EVs by Lipidomics—BP-EVs Lipid extraction: yeast (≈15 mL) was centrifuged at 10 000 × g, 30 min at 5 °C to obtain For lipid extraction, 5 µL (≈2.5 × 105 particles) of concentrated BP-EVs by-product 1 (BP1); Whey from plain yogurt (BP2) was obtained by were mixed with 5 µL of milliQ water and 20 µL of ice-cold methanol. centrifugation of commercial yogurt (≈25 mL) at 10 000 × g, 30 min at Samples were vigorously shaken by vortexing for 2 min and then, 5 °C of plain; SCOBY derived from the production of kombucha (BP3) 250 µL of methyl tert-butyl ether (MTBE), were added. Samples were and fermented wine pomace (BP4) were obtained by respective manual immersed in a water bath (ATU Ultrasonidos, Valencia, Spain) with an homogenization of 3 g each of SCOBY and wine pomace in 15 mL of ultrasound frequency and power of 40 KHz and 100 W, respectively, at PBS by using a glass Dounce homogenizer. All pre-processed BPs were 10 °C for 30 min. Then, 25 µL of milliQ water was added to the mixture, stored -80 °C until further use. and organic phase was separated by centrifugation (1400 g) at 10 °C for EVs Isolation Methods: Three different industrially scalable methods 10 min.[43] BP-EVs-derived lipid extracts, contained in the upper phase, were selected based on scientific literature and tested for enrichment were collected and subjected to liquid chromatography electrospray of BP-EVs. All experiments were performed in independent batches per ionization quadrupole time-of-flight mass spectrometry (LC-ESI-Q-TOF triplicate. MS/MS). Internal isotopically labeled lipid standards for each class Protein Organic Solvent Precipitation: PROSPR was applied to obtain were used for signal normalization.[44] Stock solutions were prepared by EVs from 4 mL of BP1-4 as previously described.[3,40] Briefly, BP was mixed dissolving lipid standards in MTBE at a concentration of 1 mg mL−1, and with four volumes of chilled acetone (-20 °C), the mixture was vortexed working solutions were diluted to 2.5 µg mL−1 in MTBE. and centrifuged for <1 min at 5000 × g. The supernatant containing Nontargeted Lipidomics LC-MS: BP-EVs-derived lipid extracts were EVs was concentrated in a vacuum concentrator till near dryness analyzed following a previously published method.[45] For LC-Q-TOF- (Concentrator Plus, Eppendorf AG, Hamburg, Germany). Concentrated based lipid molecular species analysis, BP-EVs-derived lipid extracts BPs-EVs samples were stored at -80 °C until further analysis. were subjected to LC-MS using a UPLC 1290 series coupled to ESI- Single-Step Size Exclusion chromatography: Isolation of BP-EVs Q-TOF MS/MS 6545 (Agilent Technologies, Barcelona, Spain). Sample by SEC was performed as previously reported by Böing et al.,[41] with compartment of the UHPLC was refrigerated at 4 °C and for each sample, minor modifications. Briefly, Sepharose CL-2B was washed with 0.32% 10 µL of FFIBP EVs-derived lipid extract were applied onto 1.8 µm particle trisodium citrate prepared in PBS (0.22 µm filtered) by spinning 10 mL 100 × 2.1 mm id Waters Acquity HSS T3 column (Waters, Milford, MA) of Sepharose at 4500 × g, 5 °C for 5 min. Supernatant was removed, heated at 55 °C. The flow rate was 400 µL min−1 with solvent A composed 6 mL of PBS was added, and the mixtures were vortexed for 1 min of 10 × 10−3 m ammonium acetate in acetonitrile–water (40:60, v/v) and followed by centrifugation 4500 × g at 4 °C for 5 min. Sepharose was solvent B composed of 10 × 10−3 m ammonium acetate in acetonitrile– then three times washed. The tip of a 10 mL plastic medical HENKE- isopropanol (10:90, v/v). The gradient started at 40% of mobile phase JECT syringe (Henke Sass Wolf GmbH, Tuttlingen, Germany) was B and reached 100% B in 10 min and held for 2 min. Finally, the system stuffed with commercial nylon stocking (20 denier HM, Västerås, was switched back to 60% of mobile phase B and was equilibrated for Sweden). Subsequently, the suspension of Sepharose in PBS was filled 3 min. Duplicate runs of the samples were performed to collect positive and stacked in the syringe by creating an in-house made column of and negative electrospray ionized lipid species in TOF mode, operated approximately 6 cm height. For EVs enrichment by single-step SEC, in full-scan at 100 to 3000 m/z in an extended dynamic range (2 GHz), 4 mL of preprocessed BPs sample was loaded into the column. Elution using N2 as nebulizer gas (5 L min−1, 350 °C). The capillary voltage was was performed with a PBS solution including 0.32% citrate (pH 7.4, set at 3500 V with a scan rate of 1 scan s−1. Continuous infusion using 0.22 µm filtered). Eluates were then collected in three sequential steps a double spray with masses 121.050873, 922.009798 (positive ion mode) by using 4 mL in each step. Based on the strategy designed by Böing, and 119.036320, 966.000725 (negative ion mode) was used for in-run et al.,[41] BP-EVs were eluted in the second fraction. Presence of BP-EVs calibration of the mass spectrometer instrument.[46] was experimentally confirmed by Nanoparticle Tracking Analysis (NTA). Nontargeted Lipidomics Bioinformatics and Data Analysis: MassHunter SEC isolated BP-EVs samples were finally stored at -80 °C till further Qualitative Analysis Software (Agilent Technologies, Barcelona, Spain) use. was used to obtain the molecular features of the samples, representing Molecular Weight Cut-Off: Isolation of BP-EVs by MWCO was different comigrating ionic species of a given molecular entity using the performed from 3 mL of pre-processed BP by using Vivaspin PES Molecular Feature Extractor (MFE) algorithm (Agilent Technologies, 300 kDa MWCO spin-filters. Spin-filters were centrifuged at 6000 × g Barcelona, Spain). MassHunter Mass Profiler Professional Software for 35 min at 4 °C. After initial centrifugation of 3 mL of BPs sample, (Agilent Technologies, Barcelona, Spain) and Metabolanalyst Software the sample remaining in the upper side of the MWCO membrane was were used to perform a nontargeted metabolomics and lipidomic washed twice with 10 mL of PBS and spin-filter was centrifuged again analysis over the extracted features. Only those features with a minimum using the same conditions. Concentrated BP-EVs sample remaining in of 2 ions were selected. After that, the molecular characteristics in the the upper side of the filter was collected and stored at -80 °C for further samples were aligned using a retention time window of 0.1% ± 0.25 min downstream analysis. and 30.0 ppm ± 2.0 mDa. Only features found in at least 70% of the Adv. Funct. Mater. 2022, 32, 2202700 2202700 (11 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de QC samples were considered to correct for individual bias and signal MTT and Neutral Red Assays: Caco-2 cells were cultured in P96 well was corrected using LOESS approach. Multivariate statistics (partial plates at 15.000 cells per well with a volume of 100 µL. 24 h after culturing, least-squares discriminant analysis (PLS-DA)) were performed using the medium was removed and BP-EVs were added at three different Metaboanalyst software. Features representing significant differences concentrations: Treatment 1 (T1): 108 particles mL−1, Treatment 2 (T2) by univariate statistics ANOVA (p < 0.05), defined by exact mass and 109 particles mL−1 and Treatment 3 (T3): 1010 particles mL−1 prepared in retention time, were reported. An initial tentative classification into 50 µL PBS and 50 µL of complete medium without serum. The negative lipid families of the molecular features present consistently in three control was performed treating cells with PBS (CN: 50%) and the replicates was performed using CEU Mass Mediator v.3.0 (accuracy positive control was performed treating cells with two concentrations of < 10 ppm). Identities of interest were also searched against the HMDB DMSO (CP1: 10% and CP2: 30% diluted in PBS). Cells were incubated (accuracy < 30 ppm) and LIPID MAPS (accuracy < 20 ppm) databases. for 24 h and after the incubation period, 3-(4,5-dimethylthiazol-2-yl)- The identities obtained were compared with the retention time of the 2,5-diphenyltetrazolium bromide (MTT) and Neutral Red (NR) assays authentic standards added. Finally, identities were confirmed by MS/ were performed. Briefly, the medium was retired and MTT/NR solutions MS by checking the MS/MS spectrums using LipidBlast software and (1:10 dilution in PBS) were added and incubated for 40 min at 37 °C LipidMatch, a R-based tool for lipid identification. with 5% CO2. Cells were then washed with PBS and distaining solutions Proteomics Characterization of EVs—In Solution Digestion of BP-Derived (10% sodium dodecyl sulfate in 0.01 m HC1 for MTT and 50% ethanol EVs Proteomes: BP-EVs (≈4.2 × 107 particles) were processed for vesicle 96%, 49% deionized water, 1% glacial acetic acid for NR) were added. lysis and denaturalization of proteins by dissolution in 16 m urea Absorbance was measured at 590 and 540 nm respectively using a prepared in 100 × 10−3 m ammonium bicarbonate. After 20 min of mixture spectrophotometer VICTOR Nivo Multimode Microplate Reader. incubation at room temperature, samples were diluted with HPLC- Fluorescent Imaging Monitoring of BP-EVs Bioavailability and grade water to 8 m urea. Protein amount in each analyzed sample was Biodistribution In Vivo—Animals: All in vivo experiments were carried quantified by bicinchoninic acid (BCA) assay. BP-EVs constituent proteins out at the Animal Experimental Service (SEA) of the Institut de Recerca were then digested for shotgun proteomics as previously described,[47] de l’Hospital de la. Santa Creu i Sant Pau. All the in vivo procedures with little modifications. Briefly, solubilized BP-EVs proteomes were were approved by the Hospital de Sant Pau and Hospital Vall d’Hebron reduced with 20 × 10−3 m dithiothreitol at 30 °C for 3 h and alkylated with Animal Ethics Committees (Approval Ref #10234; date of approval 20th 40 × 10−3 m iodoacetamide for 1 h at room temperature prevented from March 2019; Project “Evaluation of nanoconjugates for directed therapy in the light. Trypsin digestion was subsequently performed overnight at murine models”) and performed according to and in strict accordance 37 °C by adding a 20 µg of sequencing-grade trypsin to the samples and with the International Guiding Principles for Animal Research. subsequently quenched by addition of a final concentration of 0.5% formic Fluorescent Labeling of BP-EVs: Fluorescent labeled BP-EVs were acid (FA). Tryptic digested peptidomes were then desalted through 100 mg prepared fresh prior to the in vivo study. BP-EVs resuspended in PBS C18 Sep-pack cartridge (Waters, Milford, MA). Elution was performed were labeled with 1 × 10−6 m of Vybrant DiD Cell-Labeling Solution using 1 mL of 75% acetonitrile, 0.1% FA. Eluates were dried using a (Invitrogen, Waltham, USA) at 37 °C for 20 min. Subsequently, excess vacuum concentrator and stored at -20 °C till further proteomics analysis. of dye was removed by filtration using a 10 kDa MWCO filters following Liquid Chromatography Tandem-Mass Spectrometry of BP-EV manufacturer’s instructions. Passed-through liquid containing the Proteomes: Desalted BP-EVs peptidomes were resuspended in 0.1% FA unbound dye and fluorescent-labeled BP-EVs were collected. prior to LC-MS/MS analysis by using a nanoElute liquid chromatograph In Vivo Administration of Fluorescent-Labeled BP-EVs: 10-weeks old (Bruker Daltonics, MA, USA) at 300 nL min−1 and ran using a 60 min C57BL/6NT mice (n = 16) were housed in cages on a 12 h dark/light cycle linear gradient of 3–45% ACN. The nanoElute liquid chromatograph was at stable temperature (21 °C) with water provided ad libitum and fed online coupled to a state-of-the-art timsTOF Pro mass spectrometer with standard commercial chow for a minimum of 2 weeks (adaptation (Bruker Daltonics, MA, USA) using parallel accumulation—serial period) before starting the study. Mice were maintained in fasting fragmentation (PASEF) data acquisition. conditions 16 h before the experimental procedure. Mice were treated Proteomics Bioinformatics and Data Analysis: Analysis of BP-EVs with ≈3 × 109 particles suspended in 200 µL of PBS administered orally obtained proteomics raw data was carried out in the specialized (oral) or by tail-intravenous injection (i.v.). BP1 and BP2 were chosen in bioinformatics suite software PEAKS Studio X, allowing a precursor ion this experiment as representative BP-EVs from animal (BP-EVs-A) and tolerance of 10 ppm and a fragment ion tolerance of 0.05 Da. In-house yeast/vegetal (BP-EVs-V) sources, respectively. Two different controls databases were created by combining the existent proteome available in were also carried out; a control of the dye where the same amount of National Center for Biotechnology Information (NCBI) from all possible dye was mixed with PBS (CD) and a negative control where PBS were organisms present in every BP sample based on existent literature. administered (CN). Detailed information about organisms considered in each database is In Vivo and Ex Vivo Whole-Body Biodistribution of Fluorescent-Labeled included in Table S1 (Supporting Information). Carbamidomethyl of BP-EVs: The oral bioavailability and whole-body biodistribution of Cys was set as fixed modification during the proteomics bioinformatics fluorescent-labeled BP-EVs were monitored noninvasively by FLI using database search. PTM algorithm available in PEAKS Studio software was the IVIS Spectrum imaging system (PerkinElmer Life Science, Waltham, used for the identification of protein post-translational modifications. MA, USA). Images acquisition and measurements were performed using FDR < 1% was established for protein identification in all samples and the Living Image 4.3 software (PerkinElmer). Mice were anesthetized trypsin was set as a proteolytic enzyme. Data were exported to Microsoft using 1–3% isoflurane (Abbott Laboratories, Abbott Park, IL) and Excel CSV files and in-house generated macros were used for protein imaged at 0, 2, 4, and 24 h postadministration. Mice were maintained quantification analyses. Gene ontology (GO) analysis for comparison of with water ad libitum and hydrogel during the experimental procedure. BP-EVs proteomes with human circulating EVs proteomes was performed Mice were then euthanized at 25 h postadministration and major with PANTHER version 16.0. List of proteins identified from human organs (lung, brain, liver, heart, bone from ribs, muscle from leg, kidney, circulating EVs isolated by ultracentrifugation with sucrose cushion was digestive system, pancreas, and spleen) were excised and imaged by obtained from the previously published work of the authors.[40] Further ex vivo FLI. The signal emitted by the fluorescent-labeled BP-EVs was categorization of membrane proteins was performed by using in-house detected and digitalized using the software Living Image 4.7.3 as a macro software or by performing functional enrichment analysis in the pseudo-color overlay onto a grey-scale image of the animal or excised open-source specialized software FunRich version 3.1.4. tissue. Fluorescent signal of regions of interest was quantified as radiant Cell Toxicity Assessment by In Vitro Assays—Cell Culture: Caco-2 cells efficiency. were cultured in Dulbecco’s modified Eagle’s medium (Biowest) Statistical Analysis: Data were reported as mean ± standard error of complemented with 1% fetal bovine serum (Biowest), non-essential mean (SEM), otherwise specified. Data are derived from at least three amino acids, sodium pyruvate, and HEPES buffer solution (Biowest), at biological replicates in all the experiments. Log2 transformation of 37 °C with 5% CO2. data was performed by the software GraphPrism version 8.4.3. Gene Adv. Funct. Mater. 2022, 32, 2202700 2202700 (12 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Ontology (GO) analysis for comparison of BP-EVs proteomes with processing, lipidomics, and data analysis. C.L., E.V., X.G.-P., and A.S. human circulating EVs proteomes was performed with PANTHER wrote the manuscript. C.L., X.G.-P., and A.S planned the experiments. version 16.0. List of proteins identified from human circulating EVs X.G.-P. and A.S. conceived the study, obtained samples, obtained isolated by ultracentrifugation with sucrose cushion was obtained from competitive funds, and supervised the project. All authors have read and a previously published work of the authors.[40] Further categorization of approved the final version of the manuscript for submission. membrane proteins was performed by using in-house macro software or by performing functional enrichment analysis with the open- source specialized software FunRich version 3.1.4. Significance was Data Availability Statement assessed by parametric ANOVA or by nonparametric Kruskal Wallis with a minimum significance level p < 0.05. Statistical significance was The proteomics data that support the findings of this study are openly indicated as: *p < 0.05 (parametric), **p < 0.001 (parametric), #p < 0.05 available through the specialized repository PRIDE (https://www.ebi. (nonparametric), otherwise specified in the specific figure captions. ac.uk/pride/) with the following identifier: PXD025780. The Lipidomics GraphPad Prism version 8.4.3 was used for statistical analyses and for data that support the findings of this study are openly available through creating and rendering of data plots. Figures were assembled for the the specialized repository MetaboLights (https://www.ebi.ac.uk/ final version by using Illustrator 2020 software (Adobe, CA, USA). metabolights/index) with the following identifier: MTBLS2801. The rest of data generated remain at full disposition upon reasonable request. Supporting Information Keywords Supporting Information is available from the Wiley Online Library or from the author. circular economy, extracellular vesicles, food industry by-products, lipidomics, mass spectrometry, nanocarriers, proteomics Acknowledgements Received: March 8, 2022 Revised: March 29, 2022 The authors sincerely thank Gemma Plaza, oenologist at the Castell Published online: April 21, 2022 del Remei winery in Penelles, Lleida, Spain and Juan Carlos Blanco, production manager at Mahou San Miguel in Alovera, Madrid, Spain for their kind and altruistic help on the obtention of their respective industry by-product samples. The authors also thank Dr. Hector Peinado and his research group at the National Center for Oncology Research (CNIO) in [1] L. Margolis, Y. Sadovsky, PLoS Biol. 2019, 17, e3000363. Madrid (Spain) for their support on the morphometric characterization [2] G. Raposo, W. Stoorvogel, J. Cell Biol. 2013, 200, 373. of BP-EVs. Support for this work was provided by the Research and [3] X. Gallart-Palau, A. Serra, S. K. Sze, Mol. Neurodegener. 2016, 11, 41. Education Council of the Community of Madrid, Spain (2018-T1/ [4] L. M. Doyle, M. Z. Wang, Cells 2019, 8, 727. BIO-10633), Ministry of Science and Innovation, Spain (PID2020- [5] M. Malloci, L. Perdomo, M. Veerasamy, R. Andriantsitohaina, 114885RB-C21) and a FIS project by Carlos III Institute of Health (ISCIII), G. Simard, M. C. Martínez, Antioxid. Redox Signaling 2019, 30, 813. Spain (PI20/00623). A.S. acknowledges a grant from the Talento [6] W. Meng, C. He, Y. Hao, L. Wang, L. Li, G. Zhu, Drug Delivery 2020, Program 2018 of the Community of Madrid. X.G.-P. acknowledges grants 27, 585. from Sara Borrell postdoctoral program (CD19/00243) and Miguel [7] P. Gangadaran, B.-C. Ahn, Pharmaceutics 2020, 12, 442. Servet tenure track program (CP21/00096) of the Instituto de Salud [8] K. Adlerz, D. Patel, J. Rowley, K. Ng, T. Ahsan, Stem Cell Res. 2020, Carlos III (ISCIII, Spain), respectively awarded on the 2019 and 2021 calls 48, 101978. under ISCIII-Health Strategy Actions [These grants are cofunded with [9] P. Gangadaran, C. M. Hong, J. M. Oh, R. L. Rajendran, European Union Funds (ISCIII Miguel Servet Program 2021 is cofunded S. Kalimuthu, S. H. Son, A. Gopal, L. Zhu, S. H. Baek, S. Y. Jeong, by Fondo Social Europeo Plus, FSE+)]. M.V.C. acknowledges a Miguel S.-W. Lee, J. Lee, B.-C. Ahn, Front. Pharmacol. 2018, 9, 817. Servet program contract (CPII20/00007). C.L.’s Ph.D. was funded by [10] P. Pérez-Bermúdez, J. Blesa, J. M. Soriano, A. Marcilla, the Regional Ministry of Science, Universities and Innovation of the Community of Madrid and the European Social Fund for the recruitment Eur. J. Pharm. Sci. 2017, 98, 40. of predoctoral researchers (PEJD-2019-PRE/BIO-16475). IRBLLEIDA and [11] E. A. Kang, H. I. Choi, S. W. Hong, S. Kang, H. Y. Jegal, E. W. Choi, X.G.-P. are co-funded by CERCA Program/Generalitat de Catalunya. B. S. Park, J. S. Kim, Biomedicines 2020, 8, 522. [12] P. Akuma, O. D. Okagu, C. C. Udenigwe, Front. Sustainable Food Syst. 2019, 3, 23. Conflict of Interest [13] G. Moraga, S. Huysveld, F. Mathieux, G. A. Blengini, L. Alaerts, K. Van Acker, S. de Meester, J. Dewulf, Resour. Conserv. Recycl. 2019, CL, EV, XG-P. & AS are inventors in a European patent application filed 146, 452. by the non-profit public research institutions: Institute of Research [14] F. Girotto, L. Alibardi, R. Cossu, Waste Manag. 2015, 45, 32. in Health Pere Virgili (IISPV), IMDEA Food & Health Sciences and [15] R. E. Veerman, L. Teeuwen, P. Czarnewski, G. Güclüler Akpinar, CIBERSAM-ISCIII. The referred patent covers the obtention and any A. Sandberg, X. Cao, M. Pernemalm, L. M. Orre, S. Gabrielsson, potential application(s) of BP-EVs obtained from industrial residues of M. Eldh, J. Extracell. Vesicles 2021, 10, e12128. fermented beverages, fermented dairy products and fermented plant- [16] H. Sunshine, M. L. Iruela-Arispe, Curr. Opin. Lipidol. 2017, 28, 408. based products. [17] T. Skotland, K. Sandvig, A. Llorente, Prog. Lipid Res. 2017, 66, 30. [18] G. Van Meer, D. R. Voelker, G. W. Feigenson, Nat. Rev. Mol. Cell Biol. 2008, 9, 112. Author Contributions [19] J. S. Brzozowski, H. Jankowski, D. R. Bond, S. B. McCague, C.L., X.G.-P., and A.S. performed isolation of BP-EVs, morphometric & B. R. Munro, M. J. Predebon, C. J. Scarlett, K. A. Skelding, ultrastructural characterization, sample processing for systems biology J. Weidenhofer, Lipids Health Dis. 2018, 17, 211. studies, proteomics experiments, and analysis of data. M.L. and C.L. [20] F. Meier, A. D. Brunner, S. Koch, H. Koch, M. Lubeck, M. Krause, performed in vitro studies. C.L., M.V.C., L.C., S.F., X.G.-P., and A.S. N. Goedecke, J. Decker, T. Kosinski, M. A. Park, N. Bache, O. Hoerning, performed in vivo studies. C.L., M.J., and N.M.-M. performed sample J. Cox, O. Räther, M. Mann, Mol. Cell Proteomics 2018, 17, 2534. Adv. Funct. Mater. 2022, 32, 2202700 2202700 (13 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2022, 29, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202202700 by Readcube (Labtiva Inc.), Wiley Online Library on [03/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de [21] C. G. Vasilopoulou, K. Sulek, A.-D. Brunner, N. S. Meitei, [33] P. R. Odgren, L. W. HarvieJr, E. G. Fey, Proteins 1996, 24, 467. U. Schweiger-Hufnagel, S. W. Meyer, A. Barsch, M. Mann, F. Meier, [34] H. Shao, H. Im, C. M. Castro, X. Breakefield, R. Weissleder, H. Lee, Nat. Commun. 2020, 11, 331. Chem. Rev. 2018, 118, 1917. [22] S. Mathivanan, C. J. Fahner, G. E. Reid, R. J. Simpson, Nucleic Acids [35] M. L. Marco, D. Heeney, S. Binda, C. J. Cifelli, P. D. Cotter, Res. 2012, 40, D1241. B. Foligné, M. Gänzle, R. Kort, G. Pasin, A. Pihlanto, E. J. Smid, [23] H. Kalra, R. J. Simpson, H. Ji, E. Aikawa, P. Altevogt, P. Askenase, R. Hutkins, Curr. Opin. Biotechnol. 2017, 44, 94. V. C. Bond, F. E. Borras, X. Breakefield, V. Budnik, E. Buzas, [36] J. A. Tsatsaronis, S. Franch-Arroyo, U. Resch, E. Charpentier, G. Camussi, A. Clayton, E. Cocucci, J. M. Falcon-Perez, Trends Microbiol. 2018, 26, 401. S. Gabrielsson, Y. S. Gho, D. Gupta, H. C. Harsha, A. Hendrix, [37] L. Macia, R. Nanan, E. Hosseini-Beheshti, G. E. Grau, Int. J. Mol. A. F. Hill, J. M. Inal, G. Jenster, E. M. Kramer-Albers, Sci. 2019, 21, 107. S. K. Lim, A. Llorente, J. Lotvall, A. Marcilla, L. Mincheva-Nilsson, [38] M. Kang, V. Jordan, C. Blenkiron, L. W. Chamley, J. Extracell. Vesicles I. Nazarenko, et al., PLoS Biol. 2012, 10, e1001450. 2021, 10, e12085. [24] T. A. Reinhardt, J. D. Lippolis, B. J. Nonnecke, R. E. Sacco, J. Prot- [39] D. E. Murphy, O. G. de Jong, M. Brouwer, M. J. Wood, G. Lavieu, eomics 2012, 75, 1486. R. M. Schiffelers, P. Vader, Exp. Mol. Med. 2019, 51, 32. [25] S. Keerthikumar, D. Chisanga, D. Ariyaratne, H. Al Saffar, [40] X. Gallart-Palau, A. Serra, A. S. W. Wong, S. Sandin, M. K. P. Lai, S. Anand, K. Zhao, M. Samuel, M. Pathan, M. Jois, N. Chilamkurti, C. P. Chen, O. L. Kon, S. K. Sze, Sci. Rep. 2015, 5, 14664. L. Gangoda, S. Mathivanan, J. Mol. Biol. 2016, 428, 688. [41] A. N. Böing, E. van der Pol, A. E. Grootemaat, F. A. W. Coumans, [26] B. Pieters, O. Arntz, M. Bennink, M. Broeren, A. Caam, A. Sturk, R. Nieuwland, J. Extracell. Vesicles 2014, 3, 23430. M. Koenders, P. Lent, W. Berg, M. de Vries, P. Kraan, F. A. J. Loo, [42] X. Gallart-Palau, A. Serra, Y. Hase, C. F. Tan, C. P. Chen, PLoS One 2015, 10, e0121123. R. N. Kalaria, S. K. Sze, Brain Pathol. 2019, 29, 593. [27] A. Stensballe, T. Bennike, J. Extracell. Vesicles 2014, 3, 27. [43] C. Pizarro, I. Arenzana-Rámila, N. Pérez-del-Notario, P. Pérez-Matute, [28] H. Kim, M. Kim, K. Myoung, W. Kim, J. Ko, K. P. Kim, E. G. Cho, J.-M. González-Sáiz, Anal. Chem. 2013, 85, 12085. Int. J. Mol. Sci. 2020, 21, 8076. [44] I. Pradas, K. Huynh, R. Cabré, V. Ayala, P. J. Meikle, M. Jové, [29] E. Woith, G. Fuhrmann, M. F. Melzig, Int. J. Mol. Sci. 2019, 20, 5695. R. Pamplona, Front. Physiol. 2018, 9, 1165. [30] K. Trajkovic, C. Hsu, S. Chiantia, L. Rajendran, D. Wenzel, [45] J. M. Castro-Perez, J. Kamphorst, J. DeGroot, F. Lafeber, J. Goshawk, F. Wieland, P. Schwille, B. Brügger, M. Simons, Science 2008, 319, K. Yu, J. P. Shockcor, R. J. Vreeken, T. Hankemeier, J. Proteome Res. 1244. 2010, 9, 2377. [31] H. S. Lee, I. C. Boulton, K. Reddin, H. Wong, D. Halliwell, [46] I. Pradas, S. Rovira-Llopis, A. Naudí, C. Bañuls, M. Rocha, O. Mandelboim, A. R. Gorringe, S. D. Gray-Owen, Infect. Immun. A. Hernandez-Mijares, R. Pamplona, V. M. Victor, M. Jové, Sci. Rep. 2007, 75, 4449. 2019, 9, 16033. [32] M. J. Bruno, R. Rusinova, N. J. Gleason, R. E. KoeppeII, [47] X. Gallart-Palau, X. Guo, A. Serra, S. K. Sze, Alzheimer’s Res. Ther. O. S. Andersen, Faraday Discuss. 2013, 161, 461. 2020, 12, 54. Adv. Funct. Mater. 2022, 32, 2202700 2202700 (14 of 14) © 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH