The protocol here demonstrates that extracellular vesicles can be adequately separated from conditioned cell culture media using size exclusion chromatography.
Extracellular vesicles (EVs) are nano-sized lipid-membrane bound structures that are released from all cells, are present in all biofluids, and contain proteins, nucleic acids, and lipids that are reflective of the parent cell from which they are derived. Proper separation of EVs from other components in a sample allows for characterization of their associated cargo and lends insight into their potential as intercellular communicators and non-invasive biomarkers for numerous diseases. In the current study, oligodendrocyte derived EVs were isolated from cell culture media using a combination of state-of-the-art techniques, including ultrafiltration and size exclusion chromatography (SEC) to separate EVs from other extracellular proteins and protein complexes. Using commercially available SEC columns, EVs were separated from extracellular proteins released from human oligodendroglioma cells under both control and endoplasmic reticulum (ER) stress conditions. The canonical EV markers CD9, CD63, and CD81 were observed in fractions 1-4, but not in fractions 5-8. GM130, a protein of the Golgi apparatus, and calnexin, an integral protein of the ER, were used as negative EV markers, and were not observed in any fraction. Further, when pooling and concentrating fractions 1-4 as the EV fraction, and fractions 5-8 as the protein fraction, expression of CD63, CD81, and CD9 in the EV fraction was observed. The expression of GM130 or calnexin was not observed in either of the fraction types. The pooled fractions from both control and ER stress conditions were visualized with transmission electron microscopy and vesicles were observed in the EV fractions, but not in the protein fractions. Particles in the EV and protein fractions from both conditions were also quantified with nanoparticle tracking analysis. Together, these data demonstrate that SEC is an effective method for separating EVs from conditioned cell culture media.
The explosion of interest in studying extracellular vesicles (EVs) has been accompanied by major advancements in the technologies and techniques used to separate and study these nano-sized, heterogeneous particles. In the time since their discovery nearly four decades ago1,2, these small membranous structures have been found to contain bioactive lipids, nucleic acids, and proteins, and play major roles in intercellular communication3,4. EVs are released from all cell types and are therefore present in all biological fluids, including blood plasma and serum, saliva, and urine. EVs within these fluids hold great promise to serve as non-invasive biomarkers for various diseases, including neuroinflammatory and neurodegenerative diseases, cancers, and autoimmune disorders5,6,7. Further, in vitro mechanistic studies can be performed through cell culture techniques by separating EVs released into the culture medium3,8,9.
To understand the role of EVs in disease pathophysiology, adequate separation from the fluid in which they are found is paramount. The gold standard for EV separation has long been differential ultracentrifugation (dUC)10, however more sophisticated techniques have arisen to achieve better separation of EVs from other extracellular components. Some of these techniques include density gradients, asymmetrical-flow field-flow fraction (A4F), flow cytometry, immunocapture, polyethylene glycol precipitation, and size exclusion chromatography (SEC)11,12,13. Each technique has its own set of advantages and disadvantages; however, SEC in particular has been shown to separate EVs from both biological fluids and cell culture supernatants quite effectively8,14,15. SEC also has the added bonus of being relatively straightforward and user-friendly.
SEC is a method that separates components of a fluid based on size. With this technique, a column of resin (either made in-house or purchased commercially) is used to fractionate a sample. Small particles in the sample get trapped between the beads within the resin, while larger particles are able to pass through the resin more freely, and thus elute earlier in the process. Because EVs are larger in size than many extracellular proteins and protein aggregates, EVs pass through the column faster and elute in earlier fractions than extracellular proteins14.
In this methods paper, the use of SEC for separation of EVs from cell culture media (CCM) from human oligodendrocytes under both control and endoplasmic reticulum (ER) stress conditions is outlined. Using this protocol, it is shown that EVs separated with this technique are found within specific fractions that can be pooled together and concentrated for downstream characterization, and that the separated EVs are derived from cells and not from an exogenous source such as fetal bovine serum (FBS) used to supplement the CCM. The presence of the canonical EV markers, CD63, CD81, and CD916,17,18,19 in the EV fractions, and their absence in the protein fractions is demonstrated with western blotting. Using transmission electron microscopy (TEM), EVs are visualized and display the expected morphology and are only observed in the EV fraction. Particles are also counted in the EV and protein fractions of both control and ER stress conditions, and a large number of particles within the expected size range of 50-200 nm in diameter are observed in the EV samples. Together, these data support the notion that SEC is an efficient and effective method for separating EVs from cell culture media.
1. Preparation of buffers and reagents
NOTE: Make cell culture reagents in cell culture hood to maintain sterility.
2. Culturing and treating cells
3. Collection and concentration of conditioned CCM (Figure 1)
4. Cell collection
5. EV separation using size exclusion chromatography (Figure 2)
NOTE: Perform the following steps twice, once for control retentate and once for tunicamycin treated retentate.
NOTE: The PBS utilized in this section is 0.22 μm filtered 1x PBS.
6. Ultrafiltration of EV and protein fractions
7. Media controls
8. Western blotting
9. TEM imaging
10. Nanoparticle tracking analysis (NTA)
Western blotting reveals adequate separation of EVs from CCM
To evaluate the effectiveness of SEC for separating EVs from cell culture media, a western blot was run using each individual fraction from the control samples to probe expression of the three canonical EV markers, CD9, CD63 and CD81, as well as GM130 and calnexin18, which were used as negative controls (Figure 3). Albumin expression18 was also probed to ensure that extracellular proteins in the CCM can be adequately separated from EVs. Strong expression of CD9, CD63, and CD81 was observed in fractions 1-4, with little to no expression in fractions 5-8; neither GM130 nor calnexin were observed in any fraction. Albumin was only present in fractions 6-8. Together, these data indicate that vesicles elute predominately into fractions 1-4, leaving fractions 5-8 to contain extracellular proteins. Because of this, fractions 1-4 were combined and concentrated and deemed the EV fraction, while fractions 5-8 were combined and concentrated and referred to as the protein fraction.
Next, additional western blots were run to evaluate the effectiveness of concentrating the EV and protein fractions via ultrafiltration in both control and treated samples. Expression of CD9, CD63, and CD81 was observed in the cell lysates and EV fractions of both control and tunicamycin treated samples, but not in the protein fractions (Figure 4). Tumor susceptibility gene 101 (TSG101) is associated with the endosomal sorting complex required for transport (ESCRT)24 and is commonly used as a marker for exosomes25. This protein was present in cell lysates but not in EV or protein fractions. Additionally, GM130 and calnexin were only observed in the cell lysate samples. Therefore, the data indicates that EVs were effectively separated from cell culture media.
To ensure that the positive signal observed in the EV fractions was coming from EVs released by the cells and not the media itself, the exosome-depleted media underwent the same ultrafiltration and SEC protocol as conditioned CCM and was then analyzed via western blot (Figure 5). Both control and tunicamycin containing medias were processed and cell lysates were used as a positive control on the western blots. No EV markers (CD9, CD63, CD81, or TSG101) were observed in the media samples, but were present in the cell lysates. Albumin expression was observed within the protein fractions and minimal expression in the cell lysates, likely residual from the media. As no signal is observed for any of the EV markers in the EV fractions, these data demonstrate that exosome-depleted media does not contain EVs that mask the signal of cell-derived EVs.
TEM demonstrates expected morphology of EVs
TEM images were taken of both the control and tunicamycin treated EV and protein fractions (Figure 6). Spherical structures can be observed in the EV fractions of both the control and tunicamycin treated samples, indicating the presence of EVs. These structures are not observed in the protein fractions of either sample types, which instead have significant dark staining, indicative of protein in EM imaging.
NTA reveals differences in concentrations in control and treated factions
Particle concentrations of control and tunicamycin treated EV and protein fractions (n = 3) were quantified with nanoparticle tracking analysis (NTA; Figure 7). Figure 7A shows particle concentrations for the control and tunicamycin treated EV fractions, with slightly more particles present in the tunicamycin treatment relative to control. Peak particle concentrations were between 105-165 nm. Figure 7B shows concentrations for particles detected in the protein fraction of both control and tunicamycin treated protein fractions. In the protein fraction, fewer particles overall were detected relative to EVs (109 vs. 1013), and peak concentrations were also smaller, between 75-135 nm. Interestingly, the tunicamycin protein fraction had more particles than the control. The majority of EV-sized particles (approximately 50-200 nm) are observed in the EV fraction of both control and tunicamycin treated cells further supporting the use of SEC as an effective means of EV separation from conditioned cell culture media.
Figure 1: Differential centrifugation and ultrafiltration of conditioned cell culture media. Schematic outlining of the steps of differential centrifugation and ultrafiltration of media collected from cultured cells 24 h after control or 10 µg/mL tunicamycin treatment. Media is centrifuged and concentrated, leaving 500 µL of concentrated retentate suitable for SEC. Please click here to view a larger version of this figure.
Figure 2: Size exclusion chromatography. Schematic of the separation of EVs and proteins via SEC. The first 3 mL (six fractions of 500 µL each) are discarded as the void volume. Fractions 1-4 elute as EVs, while fractions 5-8 elute as proteins. Please click here to view a larger version of this figure.
Figure 3: Representative western blot analysis of individual SEC fractions. A volume of 15 µL is used from each individual SEC fraction of the control cells and probed for CD9, CD63, CD81, GM130, calnexin, and albumin expression. CD9, CD63, and CD81 were most strongly observed in fractions 1-4 while GM130 and calnexin were not observed in any fraction. Albumin was only observed in fractions 6-8. Please click here to view a larger version of this figure.
Figure 4: Representative western blot images for canonical EV makers. A final concentration of 20 µg of protein from cell lysates and 15 µg of protein from the pooled and concentrated SEC EV (1-4) and protein (5-8) fractions from control and 10 µg/mL tunicamycin treated cells were probed for CD9, CD63, CD81, TSG101, calnexin, and GM130. All markers were present in the cell lysates. CD9, CD63, and CD81 were observed in all the EV fractions, while TSG101 was not. The negative EV markers calnexin and GM130 were absent in the EV and protein fractions. Control cell lysate refers to lysates of cells that underwent control treatment, while 10 µg/mL cell lysate refers to lysates of cells that underwent tunicamycin treatment. Control EV indicates EVs from control samples. 10 µg/mL EV represents EVs from tunicamycin treated samples. Control protein implies extracellular proteins were from control samples, while 10 µg/mL protein signifies extracellular proteins from tunicamycin treated samples. Please click here to view a larger version of this figure.
Figure 5: Representative western blot images of exosome-depleted cell culture media. Control and 10 µg/mL tunicamycin exosome-depleted cell culture media underwent SEC and ultrafiltration. A total volume of 15 µL of SEC EV and protein fractions, and 20 µg of protein from cell lysates were probed for CD9, CD63, CD81, TSG101, calnexin, and albumin. No EV markers were observed in the EV or protein fractions but were observed in the cell lysates used as positive controls. Albumin was observed in the protein fractions but not the EV fraction. Control EV indicates EVs from control media samples. 10 µg/mL EV represents EVs from tunicamycin treated samples. Control protein implies extracellular proteins were from control samples. 10 µg/mL protein signifies extracellular proteins from tunicamycin treated samples. Control cell lysate stands for cells in the control treatment that were lysed, while 10 µg/mL cell lysate means lysed cells came from tunicamycin treatment. Please click here to view a larger version of this figure.
Figure 6: Representative TEM images for morphology of EVs. SEC fractions 1-4 and 5-8 were pooled and concentrated as the EV and protein fractions, respectively, from both control and 10 µg/mL tunicamycin treated cells. EVs are observed in the EV fractions of both the control and tunicamycin samples as small spherical particles less than 200 nm in diameter and are not observed in the protein samples. Scale bars = 200 nm. Please click here to view a larger version of this figure.
Figure 7: Particle size and concentrations from nanoparticle tracking analysis. Particle size and quantity in (A) EV fractions and (B) protein fractions from control and tunicamycin treated cells (n = 3). Control EVs indicates EV from control samples, while treated EVs represents EVs from tunicamycin treated samples. Control protein implies extracellular proteins from control samples, and treated protein signifies extracellular proteins from tunicamycin treated samples. Error bars are ± one standard deviation. Please click here to view a larger version of this figure.
Supplementary Figure 1: Stain-free image of PVDF membranes from Figure 3. Each PVDF membrane was imaged to visualize successful protein transfer at the conclusion of step 8.3.7. The individual SEC fractions (1-8) were visualized for blots that were later probed for (A) CD9, (B) CD63, (C) CD81, (D) GM130, (E), calnexin, and (F) albumin. Please click here to download this File.
Supplementary Figure 2: Stain-free image of PBDF membranes from Figure 4. Each PVDF membrane was imaged to visualize successful protein transfer at the conclusion of step 8.3.7. Blots were later probed for (A) CD9, (B) CD63, (C) CD81, (D) TSG101, (E) GM130, and (F) calnexin. Control cell lysate refers to cell lysates from control treated cells, while 10 µg/mL cell lysate refers to cell lysates from the tunicamycin treatment. Control EV indicates EVs from control samples and 10 µg/mL EV represents EVs from tunicamycin treated samples. Control protein refers to extracellular proteins from control samples, and 10 µg/mL protein signifies extracellular proteins from tunicamycin treated samples. Please click here to download this File.
Supplementary Figure 3: Stain-free image of PVDF membranes from Figure 5. Each PVDF membrane was imaged to visualize successful protein transfer at the conclusion of step 8.3.7. Blots were later probed for (A) CD9, (B) CD63, (C) CD81, (D) TSG101, (E) calnexin, and (F) albumin. Control EV indicates EVs from control media and 10 µg/mL EV represents EVs from tunicamycin treated media. Control protein refers to extracellular proteins from control media, and 10 µg/mL protein signifies extracellular proteins from tunicamycin treated media. Control cell lysate refers to cell lysates from control treated cells, while 10 µg/mL cell lysate refers to cell lysates from the tunicamycin treatment. Please click here to download this File.
Supplementary Figure 4: Uncropped western blot images used to create Figure 3. Uncropped composite chemiluminescent and colorimetric images of western blots probed for (A) CD9, (B) CD63, (C) CD81, (D) GM130, (E) calnexin, and (F) albumin for each individual SEC fraction (fractions 1-8). Please click here to download this File.
Supplementary Figure 5: Uncropped western blot images from Figure 4. Uncropped composite chemiluminescent and colorimetric images of western blots probed for (A) CD9, (B) CD63, (C) CD81, (D) TSG101, (E) GM130, and (F) calnexin. Control cell lysate refers to cell lysates from control treated cells, while 10 µg/mL cell lysate refers to cell lysates from the tunicamycin treatment. Control EV indicates EVs from control samples and 10 µg/mL EV represents EVs from tunicamycin treated samples. Control protein implies extracellular proteins were from control samples, while 10 µg/mL protein signifies extracellular proteins from tunicamycin treated samples. Please click here to download this File.
Supplementary Figure 6: Uncropped western blot images from Figure 5. Uncropped composite chemiluminescent and colorimetric images of western blots probed for (A) CD9, (B) CD63, (C) CD81, (D) TSG101, (E) calnexin, and (F) albumin. Control EV indicates EVs from control media while 10 µg/mL EV represents EVs from tunicamycin treated media. Control protein implies extracellular proteins were from control media and 10 µg/mL protein signifies extracellular proteins from tunicamycin treated media. Control cell lysate refers to cell lysates from control treated cells, while 10 µg/mL cell lysate refers to cell lysates from the tunicamycin treatment. Please click here to download this File.
Antibody | Host Species | Dilution |
CD9 | Mouse | 1 : 500 |
CD63 | Mouse | 1 : 1000 |
CD81 | Mouse | 1 : 500 |
GM130 | Rabbit | 1 : 500 |
Albumin | Rabbit | 1 : 1000 |
TSG101* | Rabbit | 1 : 1000 |
Calnexin* | Rabbit | 1 : 1000 |
Anti-mouse^ | Horse | 1 : 1000 |
Anti-rabbit^ | Goat | 1 : 1000 |
Table 1: Antibody dilutions. Dilutions used for antibodies. Stock antibodies were diluted in 5% milk in TBS-Tween. *Represents antibodies that require samples to be run under reducing conditions; ^ represents secondary antibodies.
SEC is a user-friendly method for adequately separating EVs from conditioned CCM. In order to specifically isolate cell derived EVs, careful consideration of the type of CCM and its supplements must be taken into account. Many cell culture medias need to be supplemented with FBS, which contains EVs derived from the animal in which the serum was harvested. These serum EVs may saturate and mask any signal produced by EVs derived from cells in culture26. Therefore, when performing experiments, EV-depleted FBS should be used whenever possible to ensure that the EVs found can be attributed to cell derived, and not bovine derived, EVs This can either be purchased commercially or made in-house through various techniques reviewed elsewhere26,27.
When running samples through a SEC column, it is imperative that filtered PBS is used. Otherwise, each fraction will be contaminated by particles from the PBS instead of solely cell derived EVs and extracellular proteins. By filtering PBS through a 0.22 µm filter prior to its utilization for SEC, one can ensure that any particle in the fractions is from the cells themselves and not contaminated PBS.
The protocol outlined here can also be modified to alter the number and size of the collected fractions. For example, fractions could be made larger (1 mL vs. 500 µL), or more fractions could be collected as there are likely extracellular proteins that elute past what is deemed fraction 8 in this protocol. This is especially true for other sample types like blood plasma, or other column compositions28,29. These small extracellular proteins that potentially elute in the later fractions may contain important signaling molecules that are not collected or assayed with the current methodology. In addition, the use of the AFC to collect fractions alleviates any potential user errors that may be introduced through manual fraction collection. As the fractions elute droplet by droplet, great care must be taken to ensure that each droplet is collected into the appropriate fraction, which can be challenging when performing manually.
One major limitation of SEC is the limited volume of sample that can be run through the column. With the commercially available columns used in this protocol, 500 µL of sample is the maximum volume that can be used. Other commercially available columns or in-house made columns may accommodate other volumes of sample, however these are still relatively small volumes29,30. Other methods, like dUC for example, can accommodate larger sample volumes, however this technique cannot easily separate EVs from other extracellular components. Gradients of sucrose or iodixanol can be useful for separating particles based on density, but this technique is time consuming and requires steady hands and very strong pipetting skills31,32.
Overall, SEC is an important method for EV research that allows for sufficient separation of EVs from conditioned CCM. This is especially true for commercial columns as they allow for greater reproducibility between biological replicates, and fractionation can be automated, reducing the chances of user error. The protocol outlined here demonstrates that EVs elute predominately in early fractions (1-4), while later fractions contain extracellular proteins. These fractions can then be combined and undergo ultrafiltration to effectively concentrate the sample for downstream analyses including western blotting, TEM imaging, and NTA particle sizing and quantification.
The authors have nothing to disclose.
The authors would like to thank Penn State Behrend and the Hamot Health Foundation for funding, as well as the Penn State Microscopy Facility in University Park, PA.
2-Mercaptoethanol | VWR | 97064-588 | |
4X Laemmli Sample Buffer | BioRad | 1610747 | |
Amicon Ultra-15 Centrifugal Filter Unit, Ultracel, 3 KDa, 15mL | Sigma-Aldrich | UFC900308 | 3 kDa cutoff |
Amicon Ultra-2 Centrifugal Filter Unit with Ultracel-3 membrane | Sigma-Aldrich | UFC200324 | 3 kDa cutoff |
Ammonium Persulfate | Sigma-Aldrich | A3678-100G | |
Anti rabbit IgG, HRP linked Antibody | Cell Signaling Technology | 7074V | 1:1000 Dilution |
Anti-Calnexin antibody | Abcam | ab22595 | 1:500 Dilution |
Anti-CD9 Mouse Monoclonal Antibody | BioLegend | 312102 | 1:500 Dilution |
Anti-GM130 antibody [EP892Y] – cis-Golgi Marker | Abcam | ab52649 | 1:500 Dilution |
Anti-mouse IgG, HRP-linked Antibody | Cell Signaling Technology | 7076V | 1:1000 Dilution |
Automatic Fraction Collector | Izon Science | ||
BCA assay Kit | Bio-Rad | ||
CCD camera | Gatan Orius SC200 | ||
Cd63 Mouse anti Human | BD | 556019 | 1:1000 Dilution |
CD81 Antibody | Santa Cruz Biotechnology | sc-23962 | 1:1000 Dilution |
Cellstar Filter Cap Cell Culture Flasks | Greiner Bio-One | 660175 | |
ChemiDoc MP Imager | BioRad | ||
Clarity Western ECL Substrate | BioRad | 1705061 | |
deoxycholate | Sigma-Aldrich | D6750-10G | |
dithiothreitol | Sigma | 3483-12-3 | |
DMEM/High glucose with L-glutamine; without sodium | Cytiva | SH300022.FS | |
Fetal Bovine Serum Premium grade | VWR | 97068-085 | |
Fetal Bovine Serum, exosome-depleted | Thermo Scientific | A2720801 | |
Glycine | BioRad | 1610718 | |
Great Value Nonfat Dry Milk | Amazon | B076NRD2TZ | |
HOG Human Oligodendroglioma Cell Line | Sigma-Aldrich | SCC163 | |
Izon Science Usa Ltd qev Size Exclusion Columns 5pk | Izon Science | ||
Methanol >99.8% ACS | VWR | BDH1135-4LP | |
Mini-PROTEAN Glass plates | BioRad | 1653310 | with 0.75mm spacers |
Mini-PROTEAN Short plates | BioRad | 1653308 | |
NP-40 | Sigma-Aldrich | 492016 | |
Penicillin-Streptomycin,Solution | Sigma-Aldrich | P4458-100mL | |
Phosphate Buffered Saline PBS | Fisher Scientific | BP66150 | |
Pierce BCA Protein Assay Kits and Reagents | Thermo Fisher Scientific | 23227 | |
Pierce PVDF Transfer Membranes | Thermo Scientific | 88518 | |
Pierce Western Blotting Filter Paper | Thermo Scientific | 84783 | |
Polyoxyethylene-20 (TWEEN 20), 500mL | Bio Basic | TB0560 | |
Protease/phosphatase Inhibitor Cocktail (100X) | Cell Signaling Technology | 5872S | |
Recombinant Anti-TSG101 antibody [EPR7130(B)] | ABCam | ab125011 | 1:1000 dilution |
Slodium hydroxide | Sigma-Aldrich | SX0603 | |
Sodium azide | Fisher Scientific | BP922I-500 | |
Sodium Chloride | Sigma-Aldrich | S9888-500G | |
Sodium dodecyl sulfate,≥99.0% (GC), dust-free pellets | Sigma-Aldrich | 75746-1KG | |
Tetramethylethylenediamine | Sigma-Aldrich | T9281-25ML | |
TGX Stain-Free FastCast Acrylamide Kit, 10% | BioRad | 1610183 | |
Transmission Electron Microscope | FEI Tecnai 12 Biotwin | ||
Tris | BioRad | 1610716 | |
Trypsin 0.25% protease with porcine trypsin, HBSS, EDTA; without calcium, magnesium | Cytiva | SH30042.01 | |
Tunicamycin | Tocris | 3516 | |
Zeta View software | Analytik | NTA software |