This protocol describes methods for purifying, quantitating, and characterizing extracellular vesicles (EVs)/exosomes from non-adherent/mesenchymal mammary epithelial cells and for using them to transfer mammary gland-forming ability to luminal mammary epithelial cells. EVs/exosomes derived from stem-like mammary epithelial cells can transfer this cell property to cells that ingest the EVs/exosomes.
Cells can communicate via exosomes, ~100-nm extracellular vesicles (EVs) that contain proteins, lipids, and nucleic acids. Non-adherent/mesenchymal mammary epithelial cell (NAMEC)-derived extracellular vesicles can be isolated from NAMEC medium via differential ultracentrifugation. Based on their density, EVs can be purified via ultracentrifugation at 110,000 x g. The EV preparation from ultracentrifugation can be further separated using a continuous density gradient to prevent contamination with soluble proteins. The purified EVs can then be further evaluated using nanoparticle-tracking analysis, which measures the size and number of vesicles in the preparation. The extracellular vesicles with a size ranging from 50 to 150 nm are exosomes. The NAMEC-derived EVs/exosomes can be ingested by mammary epithelial cells, which can be measured by flow cytometry and confocal microscopy. Some mammary stem cell properties (e.g., mammary gland-forming ability) can be transferred from the stem-like NAMECs to mammary epithelial cells via the NAMEC-derived EVs/exosomes. Isolated primary EpCAMhi/CD49flo luminal mammary epithelial cells cannot form mammary glands after being transplanted into mouse fat pads, while EpCAMlo/CD49fhi basal mammary epithelial cells form mammary glands after transplantation. Uptake of NAMEC-derived EVs/exosomes by EpCAMhi/CD49flo luminal mammary epithelial cells allows them to generate mammary glands after being transplanted into fat pads. The EVs/exosomes derived from stem-like mammary epithelial cells transfer mammary gland-forming ability to EpCAMhi/CD49flo luminal mammary epithelial cells.
Exosomes can mediate cellular communication by transferring membrane and cytosolic proteins, lipids, and RNAs between cells1. Exosome-mediated communication has been demonstrated to be involved in many physiological and pathological processes (i.e., antigen presentation, development of tolerance2, and tumor progression3). Exosomes often have contents similar to those of the source cells releasing them. Thus, the exosomes can carry specific cell properties from the source cells and transfer these properties to the cells ingesting them4.
Exosomes are 50- to 150-nm double-layer membrane vesicles and present specific markers (e.g., CD9, CD81, CD63, HSP70, Alix, and TSG101). Thus, exosomes must be characterized by various methods for different aspects. Transmission electron microscopy can be used to visualize membrane vesicles such as exosomes4,5. Nanoparticle tracking analysis (NTA) and dynamic light scattering analysis (DLS) are used for measuring the size and number of purified exosomes4. The lipid membrane content of exosomes can be verified by density gradient. Exosomal markers, such as CD9, CD81, CD63, HSP70, Alix, and TSG1016,7, can be measured by Western blotting.
Mammary basal cells have the ability to generate mammary glands when implanted into fat pads, while luminal cells cannot8,9,10. Thus, mammary basal cells are also referred to as mammary repopulating units. By using the model of mammary basal and luminal cells, the ability of EVs/exosomes to transfer cell characteristics between different cell populations can be examined. This work demonstrates the method of transferring gland-forming ability from mammary basal epithelial cells to mammary luminal epithelial cells by using EVs/exosomes derived from mammary basal epithelial cells. Luminal mammary epithelial cells acquired basal cell properties following the ingestion of EVs/exosomes secreted from basal cells and can then form mammary glands4.
All research involving animals complied with protocols approved by the Institutional Committee on Animal Care.
1. Extracellular Vesicle/exosome Isolation and Validation
2. Exosome Purification Using a Density Gradient
3. Extracellular Vesicle/Exosome Labeling
4. Extracellular Vesicle/Exosome Uptake Assay
5. Isolation of Primary Mouse Mammary Epithelial Cells
6. Separation of Primary Mouse Basal/Luminal Mammary Epithelial Cells
7. Extracellular Vesicle/Exosome Treatment
8. Fat Pad Injection of Mammary Epithelial Cells
9. Mammary Gland Whole Mount
Since it has been shown that blocking PGE2/EP4 signaling triggers EV/exosome release from mammary basal-like stem cells4, this work presents a method of isolating the induced EVs/exosomes from mammary epithelial basal cell (NAMEC) culture. Since NAMECs are cultured in serum-free medium, there are no pre-existing EVs/exosomes derived from serum13. For cells cultured in serum-containing medium, pre-existing exosomes in the medium must be pre-cleaned by ultracentrifugation at 110,000 x g before the medium is used to culture the source cells for the collection of EVs/exosomes5. EVs/exosomes from the 4 day induced NAMEC-conditioned medium can be isolated from the 110,000 x g pellet by differential ultracentrifugation, as illustrated in Figure 1. The number and size of the isolated vesicles in the 110,000 x g pellet can be measured using nanoparticle tracking analysis (NTA). The 110,000-g pellet fraction isolated by differential ultracentrifugation mainly contains ~100 nm vesicles (Figure 4A); this corresponds with the size of exosomes (50 – 150 nm) reported in the literature. In addition, TEM analysis showed that 110,000 g fractions of NAMEC-conditioned medium contain abundant membrane vesicles (Figure 4B). Although the differential ultracentrifugation can generate reasonably pure exosomes, the following purification step using a density gradient further eliminates contaminants (e.g., protein aggregates)5. In the density gradient, exosomes float in the gradient because of the lipid content in the vesicle, while protein aggregates, if any, stay at the bottom of the gradient. Each fraction of the gradient is collected for the detection of exosome makers (e.g., CD81, CD63, CD9, and TSG1016,7) by Western blotting. Exosome markers can be detected in the fraction with ~20% iodixanol (Figure 5). The fraction containing exosomes can be diluted with PBS and subjected to 200,000 x g ultracentrifugation to isolate the exosomes. The isolated exosome pellet is washed once in PBS and then is resuspended in PBS and stored at -20 °C for further analysis.
Before measuring the uptake of NAMEC-derived EVs/exosomes by the non-stem counterpart of mammary epithelial cells-HMLE cells-the EVs/exosomes must be labeled with a fluorescent dye (e.g., carboxyfluorescein succinimidyl ester (CFSE)). A parallel sample containing only CFSE but no EV/exosome is processed in the same labeling procedure. This sample is a negative control used in the following EV/exosome uptake assay to reflect the effect of a trace amount of residual free CFSE dye. HMLE cells are cultured with CFSE-labeled, NAMEC-derived EVs/exosomes or the negative control for 2-6 h and are then subjected to flow cytometry. Compared to the untreated HMLE cells, the HMLE cells cultured with the negative control express a slightly higher level of CFSE signal (Figure 6A, red line versus orange line), which reflects the background level of CFSE signaling caused by residual free CFSE left from the EV/exosome labeling process. Furthermore, compared to the negative control-treated HMLE cells, the HMLE cells cultured with CFSE-labeled, NAMEC-derived exosomes express a 10-fold higher CFSE signal (Figure 6A, blue line versus red line), which results from the specific uptake of CFSE-labeled EVs/exosomes. The uptake of CFSE-labeled, NAMEC-derived EVs/exosomes by HMLE cells can also be observed by confocal microscopy. While the negative control-treated HMLE cells do not exhibit a CFSE signal, the uptake of NAMEC-derived EVs/exosomes by HMLE cells can be observed with the CFSE signal under the confocal microscope in CFSE-labeled EV/exosome-treated cells (Figure 6B).
To evaluate whether NAMEC-derived EVs/exosomes can transfer mammary gland-forming ability from stem-like mammary basal cells to mammary luminal cells, mouse mammary luminal cells are first isolated to allow for the analysis of mammary gland formation in mice. Mouse mammary epithelial cells are isolated from 12 week-old mice. The mammary glands are cut into small pieces and are further dissociated with collagenase and trypsin. The dissociated epithelial organoids and fibroblasts can be separated by differential centrifugation, as described in steps 5.7 and 5.8. In each round of centrifugation, the pellet at the bottom of the tubes should contain mainly epithelial organoids, and the fibroblasts and single cells should float in the supernatant. Compared to the mixture containing both epithelial organoids and fibroblasts before the differential centrifugation (Figure 3, upper panels), the six rounds of centrifugation clear out most fibroblasts and single cells in the mixture (Figure 3, bottom panel).
The mammary epithelial cells (Figure 7) of the epithelial organoids are further dissociated using a natural enzyme mixture with proteolytic and collagenolytic enzyme activity and dispase to generate single cells in suspension. Sorting the single-cell suspension by the expression of cell surface CD49f and EpCAM can separate mammary luminal cells (EpCAMhi/CD49flo), mammary basal cells (EpCAMlo/CD49fhi), and non-epithelial cells (EpCAM–) (Figure 8).
The isolated EpCAMhi/CD49flo luminal mammary epithelial cells are cultured with NAMEC-derived EVs/exosomes for 10 days, and the fresh EVs/exosomes and medium are replaced every two days. After EV/exosome treatment, the mammary luminal cells are implanted into the 4th mammary fat pads (Figure 2) of mice. After 8 weeks, the fat pads are isolated and stained for the analysis of mammary gland formation (Figure 9). The treatment with induced NAMEC-derived EVs/exosomes allows luminal cells to acquire mammary gland-forming ability4. The NAMEC-derived, EV/exosome-treated mammary luminal cells form mammary glands in mouse fat pads (Figure 9).
Figure 1: Illustration of the Extracellular Vesicle/Exosome Purification Method from Cell Culture Medium by Differential Ultracentrifugation. The speed and length of each centrifugation are indicated. After each of the first three centrifugations, the supernatant is kept for the next step. After the 110,000 × g centrifugations, the pellets are kept and the supernatants discarded. Please click here to view a larger version of this figure.
Figure 2: Mouse Mammary Gland Anatomy. Mice have five pairs of mammary glands, indicated by the numbers 1-5, located in the fat pads (red) directly underneath the skin. Please click here to view a larger version of this figure.
Figure 3: Images of a mixture of epithelial organoids and cells of fat pads before and after differential centrifugation. Bright-field images of the isolated fat-pad cell mixtures before and after differential centrifugation, taken from the hemocytometer. Arrowhead: epithelial organoids. Arrow: fibroblasts and single cells. Scale bar = 0.5 mm. Please click here to view a larger version of this figure.
Figure 4: Vesicle Size and Concentration Analysis of NAMEC110,000 x g Pellet Fraction. The 110,000 g pellet fraction of the NAMEC medium is collected and subjected to (A) nano- particle tracking analysis (NTA) and (B) transmissionelectron microscopy (TEM). Scale bar = 1 µm. Please click here to view a larger version of this figure.
Figure 5: Detecting Exosomes in Fractions of the Density Gradient. Western blot analysis of exosome markers CD81, CD9, CD63, and Tsg101 and housekeeping gene GAPDH reveals the 20% iodixanol fraction containing the exosomes. Please click here to view a larger version of this figure.
Figure 6: Detecting Extracellular Vesicle/Exosome Uptake by Flow Cytometry and Confocal Microscopy. EV/exosome uptake was measured in HMLE cells. CFSE-labeled, NAMEC-derived EVs/exosomes and negative control are added to the indicated cultures for 6 h. After the incubation, the cells are subjected to (A) flow cytometry and (B) confocal microscopy. The CFSE (green; excitation/emission (nm): 492/517; microscope laser line: 488) fluorescence intensities reflect the EV/exosome uptake. Cell nuclei are stained with DAPI (blue; excitation/emission (nm): 358/461; microscope laser line: 405) and the plasma membranes are stained with plasma membrane stain (red; excitation/emission (nm): 649/666; microscope laser line: 633; see the Table of Materials). Confocal objective lens: HCX PL APO 63x/1.40-0.60 Oil. Scale bar = 20 µm. Please click here to view a larger version of this figure.
Figure 7: Images of Attached Epithelial Organoids. Bright-field images of mammary epithelial cells formed by the cells migrating and growing out of the attached epithelial organoids. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 8: Sorting of Primary Mouse Mammary Epithelial Cells by Surface EpCAM and CD49f. Mouse mammary epithelial cells isolated from fat pads of 12-week-old mice are subjected to cell sorting. Mouse mammary luminal cells are enriched in the EpCAMhi/CD49flo population, marked with the blue circle; basal cells were enriched in the EpCAMlo/ CD49fhi population, marked with the red circle. Non-epithelial cells are marked with the black box in the plot. Please click here to view a larger version of this figure.
Figure 9: Mammary Gland Formation by Primary Mouse Luminal Mammary Cells. The primary mouse EpCAMhi/CD49flo luminal cells are treated with PBS or the NAMEC-derived EVs/exosomes in cell culture for 10 days and implanted into the cleared fat pads of 3-week-old mice. The mice are euthanized and necropsied after 8 weeks to analyze mammary gland formation. Scale bar = 0.75 cm. Please click here to view a larger version of this figure.
Percentage | MFI of EpCAM | MFI of CD49f | |
mammary luminal cell | 0.437 | 4.57 x 104 | 8183 |
(EpCAMhi/CD49flo) | |||
mammary basal cell | 0.09 | 5452 | 2.23 x 104 |
(EpCAMlo/CD49fhi) | |||
Non-epithelial cell (EpCAM–) | 0.309 | 53 | 4619 |
Table 1: Percentage and Mean Fluorescence Intensity (MFI) of the Populations Described in Figure 8.
Exosomes often carry characteristics of the cells that released them, and the amount of released exosomes can be induced by stimuli4. The culture medium of cells can be collected and subjected to differential ultracentrifugation for EV/exosome collection (Figure 1). There is currently no general agreement on an ideal method to isolate EVs/exosomes. The optimal method used here has been determined by the downstream application14. Ultracentrifugation is a relatively fast method for the isolation of EVs/exosomes, which can preserve the biological activity of the EVs/exosomes. However, the vesicles isolated by ultracentrifugation generally contain a mixture of EVs, which contain exosomes produced via endosomal compartments and/or vesicles produced via budding from the cell membrane.
By analyzing the sizes of vesicles with NTA (Figure 4A), the vesicles found to purify at 110,000 x g during the differential ultracentrifugation were mainly 50-150 nm, which corresponds to the size of exosomes. The NTA data suggests that the NAMEC-derived exosomes can be isolated from the culture medium at 110,000 x g during differential ultracentrifugation. However, the NTA and TEM introduced here only allow for the visualization of vesicles or the measurement of the size and number of the whole population of EVs. These techniques cannot measure the heterogeneity of the EV populations or even sort the heterogeneous EV populations. Scientists have started to develop methods of using flow cytometers to analyze and sorting EV populations15.
Although differential ultracentrifugation can be used to purify exosomes, a density gradient should be used to further remove the contamination of protein aggregates from the exosome vesicles in the 110,000 x g isolated fraction5. Exosomes contain lipid membranes. The lipid in the exosome membranes makes the exosomes float in a density gradient, while protein aggregates fall to the bottom of the density gradient. Exosomes express specific markers (e.g., CD81, CD63, CD9, and TSG1016,7). By analyzing the presence of exosome markers in the fractions of the density gradient, it is possible to identify exosomes in the fraction with ~20% iodixanol (Figure 5). Multiple exosome markers should be examined to identify exosomes in the density gradient, since each population of exosomes may express different exosome markers16.
The uptake of EVs/exosomes by cells can be measured using fluorescent dye-labeled EVs/exosomes. The number of cells ingesting labeled EVs/exosomes can be measured by flow cytometry (Figure 6A). To confirm that the fluorescence from the cells result from the uptake of labeled EVs/exosomes but not from free dye, the pattern of fluorescence was examined in the cells using microscopy. Confocal microscopy shows that the pattern of fluorescence in the EV/exosome-treated cells is punctate (Figure 6B right panel). The punctate signals likely result from labeled EVs/exosomes, not from free fluorescence. The punctate signals in the cells can be further analyzed with structured illumination super-resolution microscopy, which has the highest resolution at 85 nm4,17. Super-resolution microscopy can confirm that the punctate signals are from ~100-nm hollow vesicles, which resemble exosomes4. These results suggest that NAMEC-derived exosomes can be ingested by mammary epithelial cells in culture.
NAMEC-derived EVs/exosomes often carry molecules (e.g., proteins and miRNAs) essential for the characteristics of certain cells1. This suggests that NAMEC-derived EVs/exosomes can transfer the properties of NAMECs (e.g. mammary gland-forming ability) to their epithelial cell counterparts. Since human mammary epithelial cells cannot form mammary glands xenogeneically in mouse fat pads18,19, the transfer of gland-forming ability can be examined using mouse primary mammary epithelial cells. Mouse primary mammary EpCAMhi/CD49flo luminal cells, which do not form mammary glands, can be isolated from 6-week-old mice (Figure 7 and Figure 8). The isolated cells from mouse fat pads can be divided into three groups (i.e., EpCAMhi/CD49flo luminal cells, EpCAMlo/CD49fhi basal cells, and EpCAM– non-epithelial cells) by examining the levels of surface EpCAM and CD49f (Figure 8). EpCAMlo/CD49fhi basal cells can form mammary glands in fat fads when transplanted into fat pads, but EpCAMhi/CD49flo luminal cells cannot4. Thus, the EpCAMhi/CD49flo luminal cell population can be used to examine the ability of induced NAMEC-derived EVs/exosomes to transfer mammary gland-forming ability. The isolated EpCAMhi/CD49flo luminal cells can be kept in culture for 7-10 days for the EV/exosome treatment4. It should be noted that keeping primary mammary epithelial cells in vitro for longer can attenuate the viability of the cells.
The EV/exosome-treated mammary luminal cells can be implanted into cleared fat pads to analyze the mammary gland-forming ability. The fat pads of mice used for implantation must be cleared at 3 weeks of age. At 3 weeks of age, mammary epithelial cells are confined to the region between the nipple and the lymph of a fat pad. Mammary epithelium in a fat pad can be cleared by removing the region between the nipple and the lymph at 3 weeks of age. Mammary luminal cells are implanted right after clearing the fat pad, and the gland formation by the implanted cells can be analyzed at 8 weeks after the implantation. The effect of NAMEC-derived EVs/exosomes on transferring mammary gland-forming ability to mammary luminal cells can be evaluated by the gland formation of luminal cells and EV/exosome-treated luminal cells (Figure 9). The induced EVs/exosomes from NAMECs carry the property of mammary basal epithelial cells – gland-forming ability – and the luminal cells that ingest the induced EVs/exosomes from NAMECs can acquire the property from NAMECs via EVs/exosomes. This work demonstrates evidence showing that molecules responsible for the EV/exosome-mediated transfer of mammary gland forming ability are present in lipid rafts of EVs/exosomes4.
The authors have nothing to disclose.
This work was supported by grants from the National Health Research Institutes (05A1-CSPP16-014, H.J.L.) and from the Ministry of Science and Technology (MOST 103-2320-B-400-015-MY3, H.J.L).
MCDB 170 | USBiological | M2162 | |
DMEM/F12 | Thermo | 1250062 | |
Optima L-100K ultracentrifuge | Beckman | 393253 | |
SW28 Ti Rotor | Beckman | 342204 | |
SW41 Rotor | Beckman | 331306 | |
NANOSIGHT LM10 | Malvern | NANOSIGHT LM10 | for nanoparticle tracking analysis (NTA) |
Optiprep | Sigma-Aldrich | D1556 | 60% (w/v) solution of iodixanol in water (sterile). |
CD81 antibody | GeneTex | GTX101766 | 1:1000 in 5% w/v nonfat dry milk, 1X TBS, 0.1% Tween 20 at 4°C, overnight |
CD9 antibody | GeneTex | GTX100912 | 1:1000 in 5% w/v nonfat dry milk, 1X TBS, 0.1% Tween 20 at 4°C, overnight |
CD63 antibody | Abcam | Ab59479 | 1:1000 in 5% w/v nonfat dry milk, 1X TBS, 0.1% Tween 20 at 4°C, overnight |
TSG101 antibody | GeneTex | GTX118736 | 1:1000 in 5% w/v nonfat dry milk, 1X TBS, 0.1% Tween 20 at 4°C, overnight |
GAPDH | GeneTex | GTX100118 | 1:6000 in 5% w/v nonfat dry milk, 1X TBS, 0.1% Tween 20 at 4°C, overnight |
CFSE (carboxyfluorescein succinimidyl diacetate ester) | Thermo | V12883 | |
FACSCalibur | BD Biosciences | fluorescence cell analyzer | |
collagenase Type IV | Thermo | 17104019 | |
trypsin | Thermo | 27250018 | |
ITS | Sigma-Aldrich | I3146 | a mixture of recombinant human insulin, human transferrin, and sodium selenite |
accutase | ebioscience | 00-4555-56 | a natural enzyme mixture with proteolytic and collagenolytic enzyme activity |
dispase | STEMCELL | 7913 | 5 mg/ml = 5 U/ml |
anti-CD49f antibody | Biolegend | 313611 | 1:50 |
anti-EpCAM antibody | Biolegend | 118213 | 1:200 |
FACSAria | BD Biosciences | cell sorter | |
carmine alum | Sigma-Aldrich | C1022 | |
human mammary epithelial cells (HMLE cells, NAMECs) | gifts from Dr. Robert Weinberg | ||
permount | Thermo Fisher Scientific | SP15-500 | |
sodium bicarbonate | Zymeset | BSB101 | |
EGF | Peprotech | AF-100-015 | |
Hydrocoritisone | Sigma-Aldrich | SI-H0888 | |
Insulin | Sigma-Aldrich | SI-I9278 | |
BPE (bovine pituitary extract) | Hammod Cell Tech | 1078-NZ | |
GW627368X | Cayman | 10009162 | |
15-cm culture dish | Falcon | 353025 | |
table-top centrifuge | Eppendrof | Centrifuge 3415R | |
ultracentrifuge tube | Beckman | 344058 | |
PBS (Phosphate-buffered saline) | Corning | 46-013-CM | |
BCA Protein Assay | Thermo Fisher Scientific | 23228 | |
Transmission Electron Microscopy | Hitachi | HT7700 | |
gelatin | STEMCELL | 7903 | |
10-cm culture dish | Falcon | 353003 | |
6-well culture dish | Corning | 3516 | |
female C57BL/6 mice | NLAC (National Laboratory Animal Center | ||
FBS (Fetal Bovine Serum) | BioWest | S01520 | |
gentamycin | Thermo Fisher Scientific | 15710072 | |
Pen/Strep | Corning | 30-002-Cl | |
DNase I | 5PRIMER | 2500120 | |
isofluorane | Halocarbon | NPC12164-002-25 | |
formaldehyde | MACRON | H121-08 | |
EtOH (Ethanol) | J.T. Baker | 800605 | |
glacial acetic acid | Panreac | 131008.1611 | |
aluminum potassium sulfate | Sigma-Aldrich | 12625 | |
Xylene | Leica | 3803665 | |
0.22 μm membranes | Merck Millipore | Millex-GP | |
AUTOCLIP Wound Clips, 9 mm | BD Biosciences | 427631 | |
AUTOCLIP Wound Clip Applier | BD Biosciences | 427630 | |
CellMask™ Deep Red | Thermo Fisher Scientific | C10046 | plasma membrane stain |