Here, we provide a detailed protocol to isolate small extracellular vesicles (EVs) from whole tissues, including brain and tumor specimens. This method offers a reproducible technique to extract EVs from solid tissue for further downstream analyses.
Circulating and interstitial small membrane-bound extracellular vesicles (EVs) represent promising targets for the development of novel diagnostic or prognostic biomarker assays, and likely serve as important players in the progression of a vast spectrum of diseases. Current research is focused on the characterization of vesicles secreted from multiple cell and tissue types in order to better understand the role of EVs in the pathogenesis of conditions including neurodegeneration, inflammation, and cancer. However, globally consistent and reproducible techniques to isolate and purify vesicles remain in progress. Moreover, methods for extraction of EVs from solid tissue ex vivo are scarcely described. Here, we provide a detailed protocol for extracting small EVs of interest from whole fresh or frozen tissues, including brain and tumor specimens, for further characterization. We demonstrate the adaptability of this method for multiple downstream analyses, including electron microscopy and immunophenotypic characterization of vesicles, as well as quantitative mass spectrometry of EV proteins.
Small extracellular vesicles (EVs) include endosomal-derived exosomes and small membrane-shed microvesicles that are of broad biomedical interest. Small EVs are composed of a heterogeneous population of 50-250 nm membrane-bound vesicles containing biologically active proteins, lipids, and nucleic acids that are collectively believed to play important roles in a multitude of disease processes. Advancing research has particularly implicated these vesicles in neurodegenerative and prion disorders, infectious processes, autoimmune or inflammatory conditions, and tumor growth and metastasis1,2,3,4,5,6,7,8,9,10,11,12,13. Rapidly growing biomedical research into the significance of EVs in disease pathogenesis has generated parallel interest in developing reproducible and rigorous methods for the isolation and purification of these vesicles.
A historical and current challenge in EV characterization has been the inability to fully separate small EV subpopulations. This challenge is largely due to our limited understanding of the differing molecular mechanisms governing the biogenesis of distinct vesicles. Overlapping size, density, and biologic cargo between subpopulations further convolutes these distinctions. Part of this challenge has also been the use of widely differing enrichment techniques, providing inconsistency in downstream analysis of isolated vesicles across laboratories, and undermining the global effort to illuminate categorical EV populations.
It is worth noting that the majority of EV characterization has been performed from in vitro collection of cell culture supernatant, with more recent in vivo studies describing techniques to isolate vesicles from animal or human body fluids, including plasma, urine, and saliva. While EVs are present in large quantities in circulation, it also recognized that these vesicles play important roles in cell-to-cell communication events and are present in the interstitium of cellular tissues. In the context of cancer, interstitial EVs may be particularly important in modulating the tumor microenvironment for cancer cell seeding and metastatic growth14,15. Consequently, there is value in the development and optimization of techniques to extract vesicles from solid tissue specimens. These methods will provide a means to directly study organ- or tumor- derived EVs harvested from clinical specimens, including small biopsies and partial or full organ resections.
In this study, and in a previous report published by our laboratory16, we aim to address several major current concerns in EV enrichment methodology: 1) to describe a reproducible technique to isolate and purify EVs to the highest standards currently accepted in the field; 2) to attempt to isolate small EV subpopulations highly enriched in endosomal-derived exosomes; and 3) to provide a protocol for the extraction of these vesicles from solid tissue specimens for the purpose of further characterization.
Recently, Kowal and colleagues described a relatively small-scale iodixanol density gradient to separate and purify EV subpopulations with greater efficacy than comparable sucrose density gradients17. In the cited study, dendritic cell-derived EVs captured in a relatively light density fraction, consistent with a density of 1.1 g/mL, were highly enriched in endosomal proteins believed to be most consistent with a high proportion of exosomes present in this fraction. According to the authors, these "bona fide" exosomal proteins included tumor susceptibility gene 101 (TSG101), syntenin-1, CD81, ADAM10, EHD4, and several annexin proteins17. We later adapted this technique to succeed a method of tissue dissociation described by Perez-Gonzalez et al18 and a subsequent differential centrifugation protocol to isolate whole brain-derived EVs16. We also demonstrated the utility of this method in characterizing EV proteomes by combining a sequential protocol for downstream quantitative and comparative mass spectrometry of vesicular protein, previously described by our laboratory19. This work paralleled that from the Hill laboratory, in which EVs were enriched from the frontal cortex of brains20.
In this study, we elaborate on this technique and extend the application of the protocol recently published from our laboratory to the isolation of EVs from solid lung tumors. To our knowledge, this is the first study to describe a protocol to enrich EVs from ex vivo tumor specimens. Given the widespread interest in EVs as novel diagnostic biomarkers and their role in tumorigenesis, this method will likely prove valuable to a growing number of scientific researchers. From a clinical perspective, interstitial EVs could harbor great diagnostic value, particularly in specimens where histologic evaluation is limited. Our hope is that the method outlined here will provide a foundation for a reproducible technique to harvest EVs from fresh or frozen animal or human surgical specimens, paving the way for future work to uncover the significant roles in disease pathogenesis these small vesicles may play.
Whole brains were obtained with approval from the Institutional Animal Use and Care Committee (IACUC) of the Florida State University. A total of twelve mouse brains (3 brains from each age group: 2, 4, 6, and 8 months) from a C57BL/6 J background were used for EV extraction, as previously described16. Lung tumor specimens were generously donated by Dr. Mandip Sachdeva under approval of the Florida Agricultural and Mechanical University IACUC. Lung tumors were derived from the human adenocarcinoma cell line H1975 grown in immunodeficient Balb/c nu/nu nude mice. Data from two representative tumor replicates are highlighted in this study.
NOTE: A schematic overview of the vesicle isolation and purification method is provided in Figure 1.
1. Tissue Dissociation and Differential Centrifugation
2. Density Gradient Purification
3. Lysis and Immunoblot Confirmation of EV Proteins
NOTE: If preservation of whole vesicles for morphologic characterization or biological activity is desired, researchers can proceed directly to Step 5. In initial experiments, or before mass spectrometry analysis, all fractions recovered should be analyzed by immunoblot to confirm fractions with EV proteins.
4. EV Protein Quantification
5. Characterization of EV Morphology
6. In-gel Purification and Trypsin Digestion for Mass Spectrometry Analysis
A schematic overview of the tissue dissociation, differential centrifugation, and gradient purification of vesicles is displayed in Figure 1. The morphologic and immunophenotypic confirmation of gradient-purified EVs is highlighted in Figure 2. A diagram of reproducible densities following ultracentrifugation of the 10-30% iodixanol gradient is shown, with two distinct vesicle populations migrating upward to fraction 2 (light EVs) and fraction 5 (dense EVs), dependent on tissue type. Representative immunoblots of gradient fractions exhibit the efficient separation and purification of small tumor-derived EVs in fraction 5. Of note, lung tumor specimens appeared enriched in dense EVs compared to light EVs (in fraction 2) previously harvested from whole brain tissue, again highlighted here. As a global tissue analysis of EVs emerge, it will be interesting to determine the distinct densities of EVs produced by unique tissues. Representative nanoparticle tracking analyses and electron microscopy of tissue-derived vesicles in the predominant vesicle-containing fraction demonstrate the enrichment and preservation of whole vesicles consistent with the known size and structure of small EVs. Finally, Figure 3 highlights the utility of this method in facilitating downstream mass spectrometry characterization of gradient-purified vesicle proteins, demonstrating the detection of many proteins previously identified in small EVs. Many of these identified proteins are consistent with those enriched in small endosomal-derived EVs previously proposed by the Théry and Hill labs17,20.
Figure 1: Schematic overview of vesicle isolation and purification from whole tissue. Following tissue dissociation, pre-clearing differential centrifugation steps, filtration, and ultracentrifugation, crude EV pellets can be resuspended on the bottom of an iodixanol density gradient for floatation separation. Please click here to view a larger version of this figure.
Figure 2: Representative morphologic and immunophenotypic analysis of tissue-derived EVs. (A) Calculated densities of iodixanol gradient following ultracentrifugation, averaged over independent experiments, highlighting fractions containing light and dense EVs. (B) Representative immunoblots of tissue-derived vesicles demonstrating the isolation of predominately dense tumor EVs and light brain EVs. Vesicle isolates are depleted of non-EV protein calnexin. H, tissue homogenate. (C) Nanoparticle tracking analysis (NTA) of representative gradient-purified tissue EVs. The smallest detected particle in this sample was 78 nm, and approximately 92% of vesicles detected were 250 nm or smaller, consistent with a high enrichment of small EVs. (D) Representative electron microscopy images of brain tissue-derived EVs. Scale bars = 200 nm. Please click here to view a larger version of this figure.
Figure 3: Summary of highlighted vesicle proteins detected by mass spectrometry of a representative tissue-derived EV sample. Following EV extraction and gradient purification, 10 µg of EV protein in enriched fraction was loaded into a polyacrylamide gel for electrophoresis and in-gel trypsin digestion. In this representative tissue-derived EV sample, a total of 918 proteins were identified by LC-MS/MS analysis. Peptides were identified, analyzed, and found to be enriched in exosomal, lysosomal, and plasma membrane proteins as previously described16. Here, we demonstrate the presence of small EV or exosomal proteins in our preparations that have been described by the Théry and Hill labs17,20. Please click here to view a larger version of this figure.
Much scientific interest has been generated with regards to the roles small EVs play in the tumor microenvironment, as well as in organ development, maturation, and function. Overall, this study provides an optimized workflow for the extraction of intact EVs from whole brain or tumor specimens. While here we simply demonstrate the applicability of this technique to lung tumor-derived EVs, this method could easily be adapted for work on other solid tissues, providing opportunities for further ex vivo characterization of small secreted vesicles that are of broad biomedical interest. While beyond the scope of this study, future comparison of EV content to that of originating tissue expression will also be important to establish the utility of isolated vesicles as tissue and disease biomarkers.
In the study previously published by our laboratory, we clearly demonstrate the advantages of the floatation-based density gradient over sedimentation gradients for purification of enriched small vesicles. In essence, a buoyancy-based separation improves purification of vesicles by avoiding contaminant migration through the fractions of interest. This principle appears particularly important when starting from complex samples like whole tissue. Regardless, we stress the importance of the pre-analytical confirmation studies of isolated vesicles described, including examination of EV morphology through complementary techniques such as electron microscopy and nanoparticle tracking analysis, as well as biochemical analyses including immunoblotting. EV proteins in enriched fractions should be confirmed with at least three recognized EV markers, including one tetraspanin protein, as recommended by the Minimal Information for Studies of EVs (MISEV) report published in the Journal of Extracellular Vesicles27. A negative EV marker should be demonstrated to be absent or highly reduced in EV fractions compared to tissue homogenates. All steps stated above are crucial for quality assurance of small vesicle purification. In addition, early demonstration of laboratory- and user- specific consistency in EV-containing density fractions can save valuable time and cost associated with processing and analyzing all fractions following the density gradient ultracentrifugation step.
It is clear that vesicles from diverse tissues may float to slightly different density fractions, or may separate into predominately light or dense subpopulations. These differences have been previously seen in whole brain-derived EVs compared to EVs harvested from dendritic cells grown in culture16,17, and are further demonstrated in this study. We show that lung tumor specimens may harbor predominately dense small EVs compared to the lighter EVs isolated from brain tissue. These differences highlighted by the method proposed in this study may promote valuable inquiries into the differential composition of EVs, including tumor EVs that may be packed with dense DNA strands28,29,30. Because of these differences in EV composition, we strongly underscore the importance of these initial technical experiments. Future application of this technique to the isolation of EVs from diverse tumor specimens will be important to determine if dense EVs are predominately secreted in the interstitium of other tumor types as well.
While the complete purification of small EVs and isolation of distinct vesicle subpopulations remains a current limitation across laboratories, this method provides a basis for further studies to characterize small vesicles from a diversity of tissue or tumor types, and may therefore facilitate a greater understanding of EV diversity. Recent separation of EV particles by asymmetric-flow field-flow fractionation has illuminated the possibility of distinct cargo and functions of small (60-80 nm) versus larger (90-120 nm) exosomes, and further identified a subset of non-membrane bound <50 nm particles denominated as "exomeres"31,32. However, cargo packaged in these unique vesicle populations varied across cell lines, highlighting the substantial heterogeneity of secreted vesicles, and supporting the importance of comprehensive in vitro or ex vivo vesicle analysis. In this study, we most likely enrich for both classes of exosomes identified. A proportion of small microvesicles with similar densities to exosomes may be present in isolates as well. Although we cannot exclude the possibility of exomeres in our isolates, membrane-bound vesicles were certainly most abundant in the gradient fractions enriched in vesicle protein. In addition to small EVs, large oncosomes (>1,000 nm) have been of recent interest to cancer biologists33. While the protocol described in this study filters out vesicles greater than 450 nm, exclusion of this step may facilitate examination of these larger EVs described. Future examination of densities associated with various vesicle subpopulations will be important to distinguish these entities in addition to their described sizes.
Finally, though we appreciate the limitations including cost and time associated with this proposed technique, we also acknowledge the need for rigorous foundational methods to characterize tissue EVs and to serve as an end-product comparison for constantly evolving and more sophisticated methods in the future. These advancing technologies will hopefully illuminate more rapid and clinically-compatible techniques to extract EVs from medical or surgical specimens.
The authors have nothing to disclose.
The authors thank Dr. Richard Nowakowski and the Florida State University Laboratory Animal Resources for providing and caring for the animals used to develop this protocol, respectfully. We thank Xia Liu, Dr. Rakesh Singh, and the Florida State University Translational Science Laboratory for assistance with the mass spectrometry work used for downstream vesicle characterization, as well as the FSU Biological Science Imaging Resource facility for the usage of the transmission electron microscope in this study. Finally, we thank Dr. Mandip Sachdeva (Florida A&M University) for donation of the tumor specimens used in the development of this method. This study was supported by grants from the Florida Department of Health Ed and Ethel Moore Alzheimer's Disease Research Program awarded to D.G.M. and J.M.O (6AZ11) and the National Cancer Institute of the National Institutes of Health under Award Number R01CA204621 awarded to D.G.M.
0.45 µm filter | VWR | 28145-505 | |
12 mL ultracentrifuge tubes | Beckman Coulter | 331372 | |
5.5 mL ultracentrifuge tubes | Beckman Coulter | 344057 | |
anti-Alix antibody | Santa Cruz | sc-7129 | |
anti-Calnexin antibody | Santa Cruz | sc-11397 | |
anti-CD63 antibody | Abcam | ab59479 | |
anti-CD81 antibody | Santa Cruz | sc-9158 | |
anti-Flotillin 2 antibody | Santa Cruz | sc-25507 | |
anti-HSC70 antibody | Santa Cruz | sc-7298 | |
anti-Syntenin-1 antibody | Santa Cruz | sc-100336 | |
anti-TSG101 antibody | Santa Cruz | sc-7964 | |
Dounce homogenizer | DWK Life Sciences | 885300-0015 | Loose-fit pestle (clearance of 0.889–0.165 mm) used. |
EZQ protein quantification kit | ThermoFisher Scientific | R33200 | |
FA-45-6-30 rotor | Eppendorf | 5820715006 | |
FEI CM120 Electron Microscope | TSS Microscopy | ||
goat anti-rabbit IgG (Fab fragment) | Genetex | 27171 | |
HALT phosphatase inhibitor (100x solution) | ThermoFisher Scientific | 78420 | |
HALT protease inhibitor (100x solution) | ThermoFisher Scientific | 78438 | |
Hibernate E medium | ThermoFisher Scientific | A1247601 | |
MLS-50 swinging-bucket rotor | Beckman Coulter | 367280 | |
NanoSight LM10 | Malvern | ||
Optima MAX-XP Benchtop Ultracentrifuge | Beckman Coulter | 393315 | |
Optima XE-100 ultracentrifuge | Beckman Coulter | A94516 | |
Optiprep | Sigma | D1556 | 60% iodixanol in sterile water solution |
Q Exactive HF Mass Spectrometer | ThermoFisher Scientific | ||
rabbit anti-goat IgG | Genetex | 26741 | |
rabbit anti-mouse IgG | Genetex | 26728 | |
Refracto 30PX (refractometer) | Mettler Toledo | 51324650 | |
S-4-104 rotor | Eppendorf | 5820759003 | |
SW 41 Ti swinging-bucket Rotor | Beckman Coulter | 333790 | |
Tabletop 5804R centrifuge | Eppendorf | 22623508 |