Comprehensive Characterization of Nanosized Extracellular Vesicles from Central and Peripheral Organs : Implications for Preclinical and Clinical Applications

Abstract

Extracellular vesicles (EV) are nano-sized vesicles that have been garnering a lot of attention for their valuable role as potential diagnostic markers and therapeutic vehicles for a plethora of pathologies. Whilst EV markers from biofluids such as plasma, serum, urine, cerebrospinal fluid and in vitro cell culture based platforms have been extensively studied, a significant knowledge gap that remains is the characterization of specific organ derived EVs (ODE). Here, we present a standardized protocol for isolation and characterization of purified EV isolated from brain, heart, lung, kidney and liver from rat and postmortem human tissue. Next, using quantitative mass spectrometry based proteomics, we characterized the respective tissue EV proteomes that identified synaptophysin (SYP), caveolin-3 (CAV3), solute carrier family 22 member 2 (SLC22A2), surfactant protein B (SP-B), and fatty acid-binding protein 1 (FABP1) as potential markers for the brain, heart, kidney, lung, and liver-EV, respectively. These respective tissue specific markers were further validated using both immunoblotting and a nanoplasmonic platform- single EV imaging analysis in the two species. To summarize, our study for the first time using traditional biochemical and high precision technology platforms provide a valuable proof of concept approach in defining specific ODE markers which further could be developed as potential therapeutic candidates for respective end-organ associated pathologies.

Graphical Abstract

INTRODUCTION

Extracellular vesicles (EV) are phospholipid nanovesicles that include (based on their origin or size), apoptotic bodies, microvesicles (MVs) and exosomes, and are garnering significant attention as molecules involved in cell-cell communication between cells and regulating the pathophysiology of several diseases.^1-3 EV carry molecular cargo such as proteins, lipids, and RNA ^4-6, and the molecular composition of the EV-cargo generally depends on their cells/tissue of origin. Furthermore, recent works demonstrating the role of EV as next-generation biomarkers based on their feasibility as therapeutic drug-delivery nanocarriers and in immunotherapy ^{7, 8} have made them exciting candidates for diagnostics and therapeutics in an array of diseases. ^9-11 A recent study revealed the use of a quantum dot-based fluorescent immunoassay capable of distinguishing pancreatic cancer patients from healthy controls based on EV as biomarkers.¹² Similarly, assays such as nanoscale Fluorescence Analysis and Cytometric Sorting (nanoFACS) techniques have been developed to detect tumor-specific antigens on EV.¹³ Nanoengineered ExoProfile chip had enabled simultaneous detection of up to eight biomarkers from one EV sample to enhance the distinction between early and late-stage cancer diagnosis.¹⁴ Furthermore, plasma-derived exosomes have been analyzed to provide insights into the microRNA cargo profiles in the diagnosis of disease such as Type-1 diabetes mellitus.¹⁵

Recently, several studies were performed to detect neuronal EV in plasma in an effort to identify and establish biomarkers for cognitive impairment in human immunodeficiency virus (HIV) infection, and neurodegenerative diseases.^{16, 17} In addition to plasma ^{18, 19}, serum and cerebrospinal fluid-derived EV have also been studied to reveal potential diagnostic and prognostic biomarkers in glioma patients.^20-23 Since body fluids (e.g., blood plasma, urine, cerebrospinal fluid) can be serially obtained at different disease stages, isolation of EV from them lend a distinct advantage of providing real-time monitoring of the progression of the illness ²⁴ and thus making them attractive targets for diagnostic applications.^25-27

Despite the many advantages of body fluids associated EV as potential biomarkers, a significant limitation is the identification of organ specific EV markers which can offer more depth in understanding the respective end-organs and associated pathologies. Most studies have been conducted on cell-lines or bio-fluids with little understanding of how organ-specific EV are expressed in situ. Although we and others have reported the feasibility of EV isolation from brain^28-33 and liver tissues²⁹, till date there is no existing literature on comparing EV from different organs including identification of definitive organ-specific EV markers. To address these challenges, the current study performed a comprehensive analysis of the respective tissue EV proteomes from the brain, heart, lung, kidney, and liver from both preclinical (rat) and clinical (postmortem human tissues). We identified synaptophysin (SYP), caveolin-3 (CAV3), solute carrier family 22 member 2 (SLC22A2), surfactant protein B (SP-B), and fatty acid-binding protein 1 (FABP1) as potential markers for the brain, heart, kidney, lung, and liver-EV, respectively. These respective tissue specific markers were further validated using both immunoblotting and a nanoplasmonic platform- single EV imaging analysis in the two species. To summarize, our study for the first time using traditional biochemical and high precision technology platforms provide a valuable proof of concept approach in defining specific ODE markers which further could be developed as potential therapeutic candidates for respective end-organ associated pathologies.

RESULTS AND DISCUSSION

Isolation and characterization of EV from body organs

One main goal of this study was to establish a standard protocol for isolation and characterization of EV isolated from different organ tissues. Figure 1 shows the scheme for isolating organ-derived EV (ODE), with minor modifications from previously published protocols.^{31, 32} Both transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA) revealed that EV purified from non-perfused rat organs displayed sizes between ~70-250 nm (Figure 2A). Interestingly, lung derived EV displayed characteristic membrane protrusions that resemble plasma membrane budding (Figure 2A). Next, we characterized four EV protein markers (Hsp-70, Flotillin-1, CD63, and CD81), and a negative marker (calnexin) by western blot on rat organ derived EV as well as in whole tissue lysates. Figure 3A shows the expression profile of all the EV makers validated in the different ODE and tissue lysates. Expression of Hsp-70 was observed in the brain, heart, lung, and liver but not in the kidney derived EV while Flotillin-1 (Flot-1) was detected in all except the lung derived EV. Flot-1 has been shown to be ubiquitously present in all EV types, however, it is less enriched in the small EV (sEV).³⁴ Its absence from lung derived EV indicates that there may be a different biogenesis mechanism in lungs and could also explain its unique morphological characteristic. These markers were, however, detected in all the tissue lysates from the different organs. The tetraspanins CD63 and CD81 were detected in all ODE. Lastly, the negative maker calnexin was expressed only in the tissue lysates of the different organs. In summary, our western blot analysis showed positive protein bands for at least three classical EV markers for all organs (brain, heart, kidney, lung, and liver). Further, we ascertained if perfusion of animals with saline prior to harvesting tissues affects ODE isolations. Interestingly, we did not observe any differences in total EV protein yield when comparing perfused and non-perfused organs (see Table S1). TEM and western blot analysis revealed no major differences in size and EV marker expression between non-perfused and perfused ODE (Figures S1 & S2). Furthermore, western blot results also revealed similar EV marker profiles as seen with ODE isolated from non-perfused animals. This data corroborates with a recent work conducted on brain derived EV where the authors failed to see significant differences in total EV number from perfused and non-perfused brains.³⁵ Presence of heterogenous EV sizes could reveal the presence of various EV subtypes in organs. Generally, the tetraspanin markers CD63 and CD81 are predominantly present in smaller sized EV subtype, exosomes, whereas Flot-1 and Hsp-70 has been shown to be present in other larger sized EV populations. ^{34, 36} The absence of Flot-1 and Hsp-70 markers from lung and kidney, respectively, indicates the absence of a subset of larger EV population. Overall, the data demonstrates the feasibility, rigor, and reproductivity of isolating ODE from different rat organs in non-perfused and perfused conditions.

Figure 1. Schematic of isolation procedure of extracellular vesicles (EV) from different rat tissues.

~ 450 mg tissue from the respective organs was sliced in small pieces and digested at 37 °C for 15-25 min. Cold Hibernate A quenched the reaction, centrifuged sequentially at 300 g and 2000 g for 10 min at 4 °C, and 10,000 × g for 30 min at 4 °C. Further, the supernatant was passed through a 0.20 μm filter and was centrifuged at 100,000 × g for 70 min at 4 °C. EV pellet was suspended in PBS and again centrifuged at 100,000 × g for 70 min at 4 °C. Next, the EV pellet was suspended in 2 ml of 0.95 M sucrose solution, layered in between 0.60 M and 1.30 M sucrose layers, and centrifuged at 200,000 × g for 16 hr at 4 °C. The figure has been created using Adobe illustrator and Biorender.com.

Figure 2. Characterization of extracellular vesicles (EV) isolated from rat tissues.

(A) Transmission Electron Microscopy (TEM) analysis of EV from rat brain, heart, kidney, lung, and liver. Scale bars, 500 nm. (B) Nanoparticle Tracking Analysis (NTA) of EV samples isolated from non-perfused rat organs (brain, heart, kidney, lung, and liver).

Figure 3. Western blot and principal component analysis (PCA) of EV purified from rat organs.

(A) Left, Western blot of EV proteins isolated from rat tissue and Right, whole tissue lysate for positive and negative EV markers. (B) 3D- PCA plot of rat brain, heart, kidney, lung, and liver EV protein samples. Ellipses and shapes show the clustering of the samples. Replicate dots depict the proteomes of the EV from rat brain (green), heart (red), kidney (blue), lung (violet), and liver (black).

Proteomic analysis of organ derived EV

It is well known that EV protein cargo can be used as biomarkers for detection of various diseases such as cancer ³⁷, neurological ¹⁷, and cardiovascular diseases.³⁸ To determine the composition of the EV proteome from the different organs, we performed comprehensive profiling of ODE using quantitative mass spectrometry based proteomics. Since we did not find major differences in size or markers between the perfused and non-perfused organ isolations, non-perfused rat organs were used for this proteomics study. We analyzed three biological replicates of ODE isolated from our rat organs. A protein was identified as differentially expressed if the FDR (Benjamini and Hochberg’s method) corrected P-value was ≤0.05, and log2 fold change was ≥1.5. As seen in Figure 3B, principal component analysis (PCA) revealed a distinguishable organ-specific clustering of ODE proteins. Overall, PCA analysis shows that the EV protein contents of various organs are significantly different from each other.

Next, we assessed the differential expression of proteins expressed in ODE isolated from different organs. Proteins identified with at least two unique peptides were included in the downstream differential analysis. Figure 4A (i) shows the number of upregulated proteins in the brain when compared with heart, kidney, lung, and liver EV proteins. We identified a total of 399 EV proteins that were upregulated in brain EV when compared to the rest of the organs analyzed. Of these, 147 had significantly higher expression across all comparisons (Figure 4A and Table S2). A similar analysis of heart, kidney, lung, and liver also identified several EV proteins that were highly abundant in one organ compared the rest analyzed in this study (Figure 4A (ii-v) and Table S2). Functional annotation using the bioinformatic tool, ClueGo, revealed that among the 147 differentially upregulated brain-EV proteins, a majority of these were involved in synaptic vesicle recycling, protein-amino acid transport, and neuronal synaptic plasticity (Figure 4B (i)). The majority of heart-EV proteins were involved with proton transport and oxidative phosphorylation pathways (Figure 4B (ii)). EV proteins from the kidney (7 groups), lung (8 groups), and liver (5 groups) were enriched with transmembrane transporter activity, regulation of actin filament polymerization, and PPAR signaling pathway, respectively (Figure 4B (iii-v)). Bar charts of the data are represented in Figure S3. It is widely accepted in the EV field that most of the cargo content expressed represents proteins or RNA carried in the host cells. It is interesting to note that protein cargos expressed in ODE from organs represent the major proteins (membrane proteins, receptors and transporters) expressed in that particular organ. For example, brain derived EV express largely synaptic membrane proteins and receptors (Figure 4B (i)). Similarly, kidney derived EV express mostly transporters which are abundantly expressed in kidneys (Figure 4B (iii)). In summary, the proteomic analysis of ODE revealed distinct organ specific protein signatures.

Figure 4. Proteomic analysis of EV contents.

(A) Venn diagram showing the upregulated proteins in EV isolated from the brain (i), heart (ii), kidney (iii), lung (iv), and liver (v) each in comparison with one other tissue. (B) ClueGO analysis of up-regulated genes in EV isolated from different tissues are represented as a pie-chart showing the enrichment of functional groups in the brain (i), heart (ii), kidney (iii), lung (iv), and liver (v). The pie-charts represents the percentage of known functions of the proteins abundant in each tissue. Data presented are from three independent biological replicates.

Identification and characterization of organ specific markers

Based on the bioinformatic analysis and normalized spectral count, we characterized the top 25 most abundant ODE proteins (Figure 5 and Table S3). Since high throughout approaches such as proteomics yield a lot of data, it is imperative to validate the identified potential hits. Accordingly, we performed post-validation of the respective ODE markers by western blot. In our quest to narrow our search for a good organ specific marker, we first compared the proteins abundant in different organs with the publicly available tissue enrichment dataset (https://v15.proteinatlas.org/tissue). Using this approach, we selected the following ODE markers for the respective organs. Brain- Synaptophysin (SYP), neural cell adhesion molecule 1 (NCAM1), excitatory amino acid transporter 1 (EAA1), and myelin oligodendrocyte glycoprotein (MOG)]; Heart-Caveolin-3 (CAV3) and cytochrome c oxidase subunit 5B (COX5B); Kidney: solute carrier family 22 member 2 (SLC22A2 and solute carrier family 12 member 1 (S12A1); Lung: surfactant protein B (SP-B), Intercellular adhesion molecule 1 (ICAM1), annexin A5 (ANXA5), and acidic mammalian chitinase (CHIA); and Liver: fatty acid-binding protein 1 (FABP1), glutathione S-transferase Mu 1 (GSTM1), microsomal glutathione S-transferase 1 (MGST1), and 3 beta-hydroxysteroid dehydrogenase type 7 (HSD3B7)] (Figures 6A(i-v) & S3 (A-E)). Following immunoblotting, we confirmed the validation of SYP, CAV3, SLC22A2, SP-B, and FABP1 as a potential brain, heart, kidney, lung, and liver-EV markers respectively from non-perfused, perfused ODE and whole tissue lysates (Figure 6B). The other markers identified showed either non-specificity or significant expression in more than one primary organ of interest and, therefore, could not meet the criteria for an organ-specific marker (Figure S4).

Figure 5. Heat map of EV proteins.

(A-E) Heatmap for top 25 upregulated EV proteins of the brain (A), heart (B), kidney (C), lung (D), and liver (E), in comparison to EV proteins of other organs. Data presented is with FDR correction set at 5% and p < 0.05.

Figure 6. Characterization of organ-specific protein EV markers of different organs.

(A) Comparison of SYP (i), CAV3 (ii), SLC22A2 (iii), SP-B(iv), and FABP1 (v) for brain, heart, kidney, lung, and liver EV proteins based on log10 normalized spectral counts. (B) WB analysis of non-perfused and perfused tissue EV, and tissue lysate for brain-specific (SYP), heart-specific (CAV3), kidney-specific (SLC22A2), lung-specific (SP-B), and liver-specific (FABP1) proteins. Data are shown as ± SEM. **p < 0.01, ***p < 0.001, ****p < 0.0001.

Although our results confirmed the presence of one predominant marker, we also found more than one significantly expressed EV protein on ODE. For example, brain given its complexity in structure and function is one of the most studied organs in EV biology. EV can readily cross the blood-brain barrier (BBB), making them attractive biomarkers for prognostic and diagnostic purposes for mental disorders. ³⁹ Recent studies have shown some success in identifying neuron-specific EV marker, L1CAM, in plasma.^40-44 In one recent study, it was shown that cleaved forms of L1CAM could be immunoprecipitated by biotin-labeled antibody from the brain and sera of HIV-1 transgenic (Tg) rats.⁴² They also show that the cleaved forms are enriched in the HIV-1 Tg rats when compared to wild type (WT). Whether the cleaved forms of L1CAM are found on the surface of EV or inside is not clear from these studies. In our proteomics screen under physiological conditions, we found that in brain L1CAM is significantly less expressed than SYP, an enriched neuronal marker (Figure S5). It is important to note that both L1CAM and SYP are expressed in other organs as well but highly enriched in the brain. SYP is an integral membrane protein localized to synaptic vesicles and the presynaptic membrane.⁴⁵ SYP has been previously shown to be differentially regulated in neuronal derived exosomes (NDE) isolated from plasma of individuals with frontotemporal dementia, Alzheimer’s disease (AD) and HIV neurological disease.^{44, 46} Interestingly, another synaptic protein, synaptosomal-associated protein, 25kDa (SNAP-25), has also been shown to increase in NDE isolated from serum of AD patients ⁴⁷. Further, we detected excitatory Amino Acid Transporter 1 (EAA1) or GLAST in brain-EV. GLAST has been recently reported as a marker for identifying astrocyte-derived EV (ADEV) in plasma ⁴⁸. Another glial specific marker, S100β was also detected in our proteomics screen.

Similarly, for heart-EV, Caveolin-3 (CAV3) was validated to be heart specific EV marker. CAV3 is a cholesterol-binding protein and is one of the main protein components of caveolae and is involved in many cardiovascular processes such as myocardial hypertrophy, myocardial ischemia, and arrhythmogenesis.⁴⁹ One recent study showed that CAV3 is present in microvesicles released from cardiomyocytes.⁵⁰ Due to its highest expression in heart-EV when compared to rest of the ODE, CAV3 could be potentially used to isolate heart-EV from peripheral biofluids to determine cardiovascular dysfunction in individuals with heart disease. Intriguingly, organ or tissue-specific EV could serve as non-invasive biomarkers to monitor immunologic rejections of transplanted tissue. ⁵¹ A recent heart transplant study in mice showed that EV released from the donor's heart could predict early acute rejection in the host. ⁵² Alterations in the expression of CAV3 could potentially serve as a biomarker to detect heart specific alterations. Kidney specific marker, SLC22A2 is an organic cation transporter 2 (OCT2), plays an important role in the renal secretion of various cationic compounds. ^53-56 Further, EV isolated from human urine could potentially serve as a non-invasive biomarker for renal dysfunction.^{57, 58} . Lung-derived EV and its associated cargo can be used as a non-invasive biomarker for diagnostic, prognostic, and therapeutic treatment of lung disorders, especially cancer.^59-62 SP-B is a lipophilic protein that is required to facilitate lung inflation and minimize the work of respiration by reducing surface tension in the lung, thus maintaining alveolar stability. Mature SP-B forms are seen in multi-vesicular bodies (MVB), indicating their likely presence on EV.⁶³ Circulating immature forms of SP-B (pro-SP-B) were identified as lung-specific biomarkers for alveolar-capillary membrane (ACM) dysfunction, increased risk of lung cancer, and chronic obstructive pulmonary disease (COPD).^64-66 Finally, for the liver, we found FABP1 to be a specific marker. FABP1 is critical for fatty acid metabolism and plays a protective role during oxidative stress, starvation, and infections.⁶⁷ Additionally, FABP1 has previously been used as molecular biomarkers in drug-induced liver injury (DILI).⁶⁸ A recent finding indicated that in patients with type 2 diabetes mellitus (T2DM), FABP1 is linked to the pathogenesis of the nonalcoholic fatty liver disease (NAFLD).⁶⁹ FABP1 plasma levels have shown a positive correlation with body mass index (BMI), creatinine levels, waist circumference, and many hepatic diseases.⁶⁹ Interestingly, one study reported the presence of FABP1 in exosomes derived from hepatocytes.⁷⁰

Intriguingly, many of the organ specific EV markers that were identified through our proteomic screening are also invariably enriched in the host tissue and subcellular organelles. For example, liver specific FABP1 and brain specific SYP are also found in organelles such as peroxisomes and synaptic vesicles respectively. However, the isolation of peroxisomes for instance uses entirely different protocol than EV which includes freezing of the density step-gradient, a prerequisite for a successful separation⁷¹. Similarly, synaptic vesicle preparations include breaking of the synaptosome membranes before the high centrifugation step to release synaptic vesicles^72-76. Hence, the variances in the isolations method largely prevent any organelle contaminations. To summarize, our results conclude that it is imperative to determine more than one organ specific EV marker in biofluids to discern the EV origin and therefore its relevance to pathogenesis.

ODE Profiling Through Single EV Imaging

In additional to the traditional western blotting, we also validated the specificity of ODE markers by using a nanoplasmonic technology platform -single EV analysis (SEA). SEA technology is capable of robust, multiplexed protein biomarker measurement in individual vesicles. EV are immobilized inside a microfluidic chamber and then on-chip immuno-staining and imaging is performed.⁷⁷ Here, marker sets that showed the highest selectivity among organs were used: SYP (brain), CAV3 (heart), SLC22A (kidney), SP-B (lung), and FABP1 (liver). We first matched the concentration of ODEs harvested from different organs (1.0×10¹⁰ EV/mL). These samples were then immuno-labeled for fluorescent imaging. As a representative result, Figure 7A shows ODE images with SYP labeling. Overall, the marker expression was dominant only in brain-derived EV, corroborating the validation from our immunoblotting platform. We further used marker-positive EV numbers as a metric for quantitative comparison (Figure 7B). SYP-labeling imparted a significant contrast between brain-derived EV and other ODEs. For a given marker, we defined its selectivity as the ratio between marker-positive and negative EV counts. For SYP, these values were > 65. Similarly, we observed high selectivity for other markers (Figure 7C; see Figure S6 for all images); the average selectivity was 30.

Figure 7. Single EV imaging.

(A) Rat ODEs were labeled for the brain-specific surface marker (SYP). Strong positive signal was observed in brain-derived EV. Scale bar, 20 μm. For other organ-specific markers, see Figure S5. (B) SYP-positive EV numbers were compared. The contrast ratio between brain-derived EV and other ODEs was >50. (C) Signal contrast among ODE types and organ-specific markers was compared. The data were row-normalized.

Currently available diagnostics are less sensitive and require 10⁵–10⁶ EV per biomarker to measure protein (e.g., Western, ELISA) or 10²– 10³ EV for the more sensitive methods μNMR, nPLEX).⁷⁷ This newly developed method could be invariably used to isolate and quantitate organ specific EV changes from biofluids to follow the progression of disease pathology during therapy.

Validation of ODE from human postmortem tissues

To further extend our preclinical findings from a rat model to a clinical setting, we validated the identified ODE markers in postmortem human tissues. We isolated EV from postmortem human brain, heart, kidney, lung, and liver tissues, and performed TEM, NTA, and SEA as performed previously in our rat studies. TEM analysis confirmed that we collected nanoscale vesicles from human organs (Figure 8A and Figure S7). Intriguingly, we obtained less total EV protein when compared to EV protein isolated from rat organs even when using similar weights of the starting material (Table S5). This discrepancy could be attributed to the time elapsed between death and collection of tissues (PMI) that could invariably result in loss of protein integrity that could lead to denaturation of tissue proteins. Interestingly, EV derived from human lung showed no bud membrane protrusions as seen with rat lung derived EV (Figure S6). The size range of these vesicles ranged between 50-300 nm. Next, we performed molecular analyses on select protein markers. Immunoblotting showed that ODE samples were positive with EV-specific protein markers (Hsp-70, CD63, CD81), whereas the negative marker (calnexin) was present only in the whole tissue lysates (Figure S8). Both ODE and tissue samples also showed the presence of tissue markers identified from our rat study (Figure 8B): SYP (brain), CAV3 (heart), SLC22A2 (kidney), SP-B (lung), and FABP1 (liver). Single EV imaging further confirmed the tissue-specific nature of these markers (Figure 8C); the number of marker-positive vesicles was the highest only when tissue-type was matched. Because total EV concentrations varied among ODE sources, we normalized each column to the maximum EV numbers detected. The number of marker-positive EVs were counted and heatmap was generated as shown in Supplementary Figure S9. In summary, both rat and human ODE show similar EV marker expression profiles, thus denoting their respective use as potential markers in both pre-clinical and clinical applications.

Figure 8. Human ODE.

(A) TEM analyses of human derived ODE, Original scale bar, 500 nm (B) Immunoblotting on select tissue makers and Gapdh (*Gap) was used as a cytoplasmic marker control. (C) Human ODEs from five organs were immuno-labeled for organ-specific surface markers. Imaging confirmed high selectivity of these markers among different organ types. Scale bar, 20 μm.

CONCLUSION

Using both biochemical and single-EV imaging methods, we show a proof of concept for identifying specific EV markers for brain, lung, heart, kidney, and liver using both preclinical (rat) and clinical (human) model systems. ODE can cross tissue barriers and circulate in blood, thereby serving as a promising non-invasive biomarker to detect organ dysfunctions. Our proteomic analysis of ODE revealed that the protein signatures identified from different organs are distinct from one another. Post-validations by western blot and single-EV imaging on both rat and human ODE, we confirmed SYP, CAV3, SLC22A2, SP-B, and FABP1 as specific markers for brain, heart, kidney, lung, and liver, respectively.

Although we validated one marker for each organ, there is a strong possibility of more potential organ specific-markers in ODE. Therefore, more stringent strategies such as tandem or sequential immunoprecipitation might be required to pull down organ-specific EV. A more thorough analysis of the data needs to be performed to ascertain the cellular origin of EV. Therefore, we recommend using more than one EV protein markers for organ-specific screening. In summary our robust approach provides a valuable proof of concept approach in defining specific ODE markers which further could be developed as potential therapeutic candidates for respective end-organ associated pathologies.

MATERIALS AND METHODS

Animals.

Male and female Sprague Dawley rats were obtained from Charles River Laboratories Inc. (Wilmington, MA, USA) and group housed in a 12-h light-dark cycle and fed ad libitum. All procedures and protocols were approved by the Institutional Animal Care and Use Committee of the University of Nebraska Medical Center and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Isolation of Organs.

We employed both perfusion and non-perfusion methods to isolate brain, heart, kidney, lung, and liver from the animals. For isolation employing perfusion, animals were deeply anesthetized using isoflurane followed by infusion with 1X phosphate-buffered saline (PBS; 70013-032, Gibco-Thermo Fisher Scientific, Waltham, MA, USA) using a peristaltic pump (Fisherbrand Variable-Flow Peristaltic Pump, Waltham, MA, USA) to clear the blood. Animals that did not undergo perfusions were sacrificed via overdose with isoflurane followed by decapitation and retrieval of organs. The different organs were harvested and immediately flash-frozen on dry ice, followed by storage at −80 °C until further use.

Human Organs.

Human postmortem tissues (Heart, Lung, Liver and Kidney) from two donors were obtained from the National Disease Research Interchange (NDRI) (Philadelphia, PA, USA), and human postmortem Brain samples from two additional donors were obtained from the UCLA National Neurological AIDS Bank of the National NeuroAIDS Tissue Consortium (NNTC). All samples were stored at −80° C until further use. See Table S3 for additional information on donors.

Extracellular Vesicles (EV) Isolations.

EV isolations were carried out as described in our previous studies with minor modifications.^{31, 32, 78} This protocol is well verified and we have consistently got similar results in all the runs that were performed for this study. Briefly, tissue weighing ~450 mg (Table S4) each were minced and treated with 20 unit/ml papain (P4762, Sigma-Aldrich, MO, USA) in 3.5 ml of Hibernate A (NC0176976, Fisher Scientific, Hampton, NH, USA) solution and rocked for 15-25 min (brain, 15 min; kidney and liver, 20 min; heart and lung, 25 min) at 37 °C. The reaction was then quenched with 6.5 ml cold Hibernate A solution and gently homogenized using a glass pipette. Tissue homogenate was sequentially centrifuged at 300 × g for 10 min, 2000 × g for 10 min, and at 10,000 × g for 30 min respectively at 4°C to remove dead cells, cell debris, and larger vesicles, respectively. The supernatant was then filtered through a 0.20 μm syringe filter (09-754-13, Corning Fisher Scientific, Hampton, NH, USA) and centrifuged at 100,000 × g for 70 min at 4 °C to pellet down the EV. The EV pellet was then resuspended in 37 ml particle-free cold 1X PBS and centrifuged again at 100,000 × g for 70 min at 4 °C. A sucrose gradient was prepared using 2 ml steps starting from 2.0 M sucrose at the bottom with a 0.25 M sucrose solution at the top. The washed EV pellet was resuspended in 2 ml of 0.95 M sucrose solution and layered between 0.60 M and 1.30 M sucrose layers. The gradients were then centrifuged at 200,000 × g for 16 h at 4 °C. To make sure layers should not mix, we used lower acceleration (7) and lower deceleration (7) during our overnight run. After run is complete we immediately take out respective layers. For collecting the purified EV, a 2 ml fraction was discarded from the top of the gradient, and the next 8 ml of the gradient were collected, diluted to 37 ml with cold particle-free PBS, and centrifuged at 100,000 × g at 4 °C for 70 min. EV pellets were suspended in 50 - 100 μl of particle-free PBS and further used for downstream analysis.

Nanoparticle Tracking Analysis (NTA).

EV size distribution curves and concentration measurements were carried out by NTA using a Nanosight NS300TM (Malvern Instruments, UK). For NTA analysis, EV pellets were resuspended in 100 μl PBS, following which 10 μl of the sample was diluted to 1:100 – 1:1000 in PBS prior to measurements. All samples were loaded with the laser module outside the instrument. As the sample was loaded, care was taken to avoid air pockets. The machine was equipped with a 488 nm laser and a syringe pump system, with a pump infusion speed of 20. The standard measurement option was selected for the scripted workflow to capture videos. The number and duration of captures were set to 5 and 60s, respectively. The base filename and location were selected prior to starting the run. The camera level was set at 11. Background measurements were performed with filtered PBS, which revealed the absence of any kind of particles. Five video recordings were carried out for each EV preparation with a duration of the 60s with frame rates of 25 frames/sec. Once the videos were recorded, the NTA 3.1 software version was used to analyze the sample videos. For analysis, the screen gain was set to 1.0, and the detection threshold was adjusted to set the minimum brightness of pixels to be considered. At the end of the analyses, the dilution factor for the samples was updated before data export.

Transmission Electron Microscopy (TEM).

For TEM, a 10 μl EV sample was mixed with 90 μL of TEM fix buffer (2% glutaraldehyde, 2% paraformaldehyde, and 0.1 M phosphate buffer). A 10 μl drop of EV-buffer mix sample was spotted onto formvar/silicon monoxide coated 200 mesh copper grids (Ted Pella Inc. Redding, CA) and allowed to sit for 3 min. Grids were glow discharged for 60 seconds at 20 μA with a GloQube glow discharge unit (Quorum Technologies, East Sussex, UK) prior to use. The excess solution was blotted off by using a clean piece of Whatman #50 filter paper held by forceps at 45° angle to touch only the edge of the grid, leaving behind a thin film of the sample that was allowed to dry for 1 min. Samples were negatively stained with NanoVan (Nanoprobes, New York, NY) and examined on a Tecnai G² Spirit TWIN (FEI, Hillsboro, OR) operating at an accelerating voltage of 80 kV. Images were acquired digitally with an AMT (Woburn, MA) digital imaging system.

Western Blotting.

EV pellets were homogenized in RIPA buffer containing 1% SDS, and protein estimated using the BCA protein assay kit (Thermo Fisher Scientific, Rockford, IL, USA). Western blotting was performed on the various tissue-specific EV as described previously.^{31, 78, 79} Briefly, 20-40 μg protein was loaded onto NuPAGE 4–12% Bis-Tris gels (Invitrogen) and run either under non-reducing (CD63 and CD81) or reducing (Hsp-70 and flotillin-1) conditions followed by their transfer onto nitrocellulose membranes using the iBlot 2 Gel Transfer Device (Invitrogen). Membranes were blocked in TBS SuperBlock buffer (Thermo Fisher Scientific) for 30 min, and immunoblotting carried out overnight at 4 °C with primary antibodies (see Table 1 for details). The next day, membranes were incubated with respective HRP conjugated secondary antibodies for 1.5 hr at room temperature on a rocker. Blots were developed with 1:1 solution of Radiance Chemiluminescent Substrate and Luminol/Enhancer (Azure Biosystems) and visualized using a c300 imaging system (Azure Biosystems). Images were quantified using the ImageJ software.

Table 1.

Details for primary and secondary Abs used for western blot analysis

Marker	Running Condition	Antibody (Catalog #)	Sample species	1° Ab dilution	2° Ab dilution
Hsp-70	Reducing	SAB4200714	Rat/Human	1:1000	1:2000
Flotillin-1	Reducing	abcam41927	Rat/Human	1:1000	1:2000
Calnexin	Reducing	ab22595	Rat/Human	1:1000	1:2000
Synaptophysin	Reducing	PA1-1043	Rat/Human	1:1000	1:2000
Caveolin-3	Reducing	ab2912	Rat	1:1000	1:2000
Caveolin-3	Reducing	MAB6706	Human	1:1000	1:2000
SLC22A2	Reducing	PA5-37290	Rat	1:500	1:1000
SLC22A2	Reducing	MAB6547	Human	1:500	1:1000
ANXA5	Reducing	ab14196	Rat	1:500	1:2000
FABP1	Reducing	Invitrogen 720242	Rat/Human	1:500	1:1000
COX5B	Reducing	ab180136	Rat	1:1000 0	1:2000 0
ICAM-1	Reducing	sc-107	Rat	1:500	1:1000
MOG	Reducing	MA5-24645	Rat	1:500	1:1000
SLC12A1	Reducing	ab171747	Rat	1:1000	1:2000
NCAM1	Reducing	ab232781	Rat	1:500	1:1000
MGST1	Reducing	ab129175	Rat	1:500	1:1000
GSTM1	Reducing	ab108524	Rat	1:1000	1:2000
EAAT1	Reducing	sc-515839	Rat	1:1000	1:2000
HSD3B7	Reducing	ab190223	Rat	1:500	1:1000
CHIA	Reducing	PA5-03057	Rat	1:1000	1:2000
SP-B	Non- reducing	WRAB-48604	Rat/Human	1:1000	1:2000
CD81	Non- reducing	MCA1846	Rat	1:500	1:1000
CD81	Non- reducing	MAB4615	Human	1:500	1:1000
CD63	Non- reducing	BD551458	Rat/Human	1:500	1:1000

Proteomics.

100 μg of EV protein prepared using RIPA buffer with 1% SDS per sample from three biological replicates and detergent removed by chloroform/methanol extraction. Samples were then subjected to quantitative mass spectrometry based proteomics. Briefly, the protein pellet was re-suspended in 100 mM ammonium bicarbonate and digested with MS-grade trypsin (Pierce) overnight at 37 °C. Peptides cleaned with PepClean C18 spin columns (Thermo Fisher Scientific) were resuspended in 2% acetonitrile (ACN) and 0.1% formic acid (FA). 500 ng from each sample was loaded onto trap column Acclaim PepMap 100 75μm x 2 cm C18 LC Columns (Thermo Fisher Scientific) at a flow rate of 4 μl/min. Subsequently, samples were separated with a Thermo RSLC Ultimate 3000 (Thermo Fisher Scientific) on a Thermo Easy-Spray PepMap RSLC C18 75μm x 50cm C-18 2 μm column (Thermo Fisher Scientific, Waltham, MA, USA) with a step gradient of 4–25% solvent B (0.1% FA in 80 % ACN) from 10-130 min and 25–45% solvent B for 130–145 min at 300 nL/min and 50 °C with a 180 min total run time. Thermo Orbitrap Fusion Lumos Tribrid (Thermo Fisher Scientific) mass spectrometer (data-dependent acquisition mode) was used to analyze the eluted peptides. By Orbitrap, with a resolution of 120,000, the full survey scan MS (from m/z 350–1800) was acquired. The ion filling time was set as 100 ms, while the AGC target for MS1 was set as 4 × 10⁵. The precursor ions with charge state 2-6 were isolated in 3 s cycle and fragmented using HCD fragmentation with 35 % normalized collision energy and were detected at a mass resolution of 30,000 at 200 m/z. For MS/MS, the AGC target was set as 5 × 10⁴, and ion filling time was set 60 ms while dynamic separation was set for 30 s with a 10 ppm mass window. The protein identification was performed by searching MS/MS data against the swiss-prot Rattus norvegicus protein database downloaded on Feb 13, 2019, using the in house mascot 2.6.2 (Matrix Science) search engine. With a maximum of two unexploited cleavage sites, the search was set up for full tryptic peptides. N-terminus acetylation and oxidized methionine were included as variable modifications, whereas cysteine carbamidomethylation was set as a fixed modification. Forerunner mass tolerance threshold was set at 10 ppm, and the maximum fragment mass error was 0.02 Da. The significance threshold score of the ion was calculated based on a false discovery rate (FDR, calculated using false positive matches/true positive matches+false positive matches) of ≤ 1% (http://www.matrixscience.com/help/decoy_help.html). Qualitative analysis was performed using progenesis QI proteomics 4.1 (Nonlinear Dynamics). Sample to sample experimental variation during sample runs was normalized using default settings (Normalize to all proteins) in progenesis QI proteomics.

Data Analysis and Bioinformatics.

Proteins identified with at least two unique peptides were included in the downstream differential analysis. Proteins were further filtered from downstream analysis if the spectral counts were missing in 80% of the samples, irrespective of the sample type. Spectral counts were analyzed using limma ⁸⁰ to identify differentially expressed proteins between any two EV types. A total of ten pairwise differential expression analyses were performed across all sample groups based on the tissue of origin. A protein was identified as differentially expressed if the FDR (Benjamini and Hochberg’s method) corrected P-value was ≤0.05, and log2 fold change was ≥1.5. A heatmap of the top 25 proteins in each organ that exhibited significantly higher or lower expression with respect to other comparisons was plotted using the function heatmap.2 in the R (version 3.6.0) package gplots. Principal Component Analysis (PCA) was performed using ClustVis.⁸¹ Gene Ontology (GO) analysis of differentially expressed proteins was performed using the Cytoscape plugin ClueGO.⁸² Only proteins exhibiting significantly higher expression across all comparisons were included in this analysis. The biological process, molecular function, and KEGG pathways were included for GO enrichment analysis.

Single EV analysis (SEA) for organ specific protein markers.

EV from different organ tissues were prepared as described above. EV pellets were suspended in PBS and then filtered with a 0.22 μm syringe filter (SLGVR04NL, Millipore Sigma). Filtered EV were captured on a glass slide. Following 30 min incubation (20 °C), the slide was washed with PBS buffer containing 0.001% Tween 20 (PBST). After incubation with fixation buffer (4% formaldehyde), permiabilization buffer (BD perm/wash buffer, 554723, BD Biosciences) and blocking buffer (Superblock, 37515, Thermo Scientific), EV samples were incubated with primary antibodies (see Table 2 for antibody details). After washing with PBST, the fluorescence-conjugated secondary antibody was incubated for 30 min (20 °C) and washed again. Fluorescence images were taken with an upright fluorescent microscope (BX63, Olympus) using a 40× objective lense. Acquisition settings (i.e., magnification, exposure time, camera gains, illumination) were kept constant for all images.

Table 2.

Details for primary and secondary Abs used for single EV analysis

Marker	Antibody (Catalog #)	Sample species	1° Ab dilution	2° Ab dilution
Synaptophysin	PA1-1043	Rat/Human	1:100/1:50	1:500/1:200
Caveolin-3	ab2912	Rat	1:100	1:500
Caveolin-3	MAB6706	Human	1:50	1:400
SLC22A2	PA5-37290	Rat	1:100	1:500
SLC22A2	MAB6547	Human	1:50	1:400
ANXA5	ab14196	Rat	1:100	1:500
SP-B	WRAB-48604	Human	1:50	1:100
FABP1	CST 13368	Rat/Human	1:50/1:50	1:200/1:100