Comparative analysis of magnetically activated cell sorting and ultracentrifugation methods for exosome isolation

Mesenchymal stem cell-derived exosomes regulate cell migration, proliferation, differentiation, and synthesis of the extracellular matrix, giving great potential for the treatment of different diseases. The ultracentrifugation method is the gold standard method for exosome isolation due to the simple protocol, and high yield, but presents low purity and requires specialized equipment. Amelioration of technical optimization is required for quick and reliable confinement of exosomes to translate them to the clinic as cell therapeutics In this study, we hypothesized that magnetically activated cell sorting may provide, an effective, reliable, and rapid tool for exosome isolation when compared to ultracentrifugation. We, therefore, aimed to compare the efficiency of magnetically activated cell sorting and ultracentrifugation for human mesenchymal stem cell-derived exosome isolation from culture media by protein quantification, surface biomarker, size, number, and morphological analysis. Magnetically activated cell sorting provided a higher purity and amount of exosomes that carry visible magnetic beads when compared to ultracentrifugation. The particle number of the magnetically activated cell sorting group was higher than the ultracentrifugation. In conclusion, magnetically activated cell sorting presents a quick, and reliable method to collect and present human mesenchymal stem cell exosomes to clinics at high purity for potential cellular therapeutic approaches. The novel isolation and purification method may be extended to different clinical protocols using different autogenic or allogeneic cell sources.


Cell culture
Human bone marrow-derived MSCs were procured (PCS-500-012, ATCC, USA) and preserved in DMEM supplemented with 10% FBS, 1% penicillin/streptomycin and incubated at 5% CO 2 at 37˚C. When cells reached 70% confluency, they were washed with PBS. After the washing process, DMEM supplemented with exosome depleted FBS (10%) was added to the flask, and cells were incubated at 5% CO 2 at 37˚C for 2 days. The manufacturer already reported and warranted the characterization of hBMSCs (Lot#63208778, ATCC, USA). During the whole experiment, MSCs were used in passage 4 as recommended by the manufacturer [70].

Exosome isolation
Ultracentrifugation. Cell culture FBS exosome-depleted supernatant was filtered and transferred to sterile centrifuge tubes. Media was mixed with vortex and centrifuged at 1500xg for 10 minutes. The supernatant was transferred to an ultracentrifuge tube, completed with PBS, and centrifuged for 10 minutes at 10000xg by ultracentrifuge (XL-90 Ultracentrifuge, Beckman Coulter, USA) with SW 28 Ti Swinging-Bucket Aluminum Rotor (Beckman Coulter, USA). The supernatant was then centrifuged for 30 minutes at 30000xg. Pellet was removed and the supernatant was ultracentrifuged for 90 minutes at 100000xg. After completing the volume with PBS, the supernatant was aspirated, and the pellet was washed before being ultracentrifuged at 100000xg for 90 minutes. Finally, the supernatant was discarded, and the exosome-containing pellet was dissolved in 100 μL PBS.
Magnetic activated cell sorting (MACS). Exosome isolation with MACS Exosome Isolation Kit (#130-110-912, Miltenyi Biotec, Germany) was performed in obedience to the instruction of the manufacturer. Briefly, the cells, cell debris, and larger vesicles were removed from the cell culture media by centrifugation for 10 minutes at 300×g, 30 minutes at 2000×g, and 45 minutes at 10000×g consecutively. Then, 50 μl of CD9, CD63 and CD81 conjugated microbeads were added to 2 ml exosome-containing supernatant and vortexed. After mixing the beads and the sample, the mixture was held on incubation for 1 hour at room temperature. The μ-column was placed in the magnetic field of the μMACS separator and prepared by applying a 100 μl equilibration buffer on top of the column. Then, the column was rinsed with isolation buffer and the magnetically labeled sample was added onto the column and, the column was washed with isolation buffer. The column was removed from the magnetic separator and placed onto a 1.5 ml tube. After this step, 100 μl of isolation buffer was added to the column and the magnetically labeled vesicles were transferred to the 1.5 ml tube.

Exosome characterization
Determination of protein quantification. Bicinchoninic acid (BCA) protein quantification (#23225, Thermo Fisher Scientific, USA) was performed [71]. Bovine serum albumin standards were prepared by serial dilutions (in between 2000 and 2 μg/ ml). Then, 25 μL serial standard solutions and samples were put into the 96-well plate and the working solution was added into the wells. After the incubation period for 30 minutes at 37˚C, absorbance was measured at 562 nm by a microplate reader (Synergy HT, Biotek, Winooski, VT, USA). This experiment was performed with 6 replicates.
Exosome characterization by nanoparticle tracking analysis. The particle numbers and particle size dispersity were identified by nanoparticle tracking analysis (NTA) (qNano Gold, Izon Science Ltd, New Zealand) [43]. Exosome samples isolated by ultracentrifuge and MACS were diluted to 1:20 and 1:2 respectively. NP80 nanopore which targets particles with 40-255nm diameter was used for analysis. The voltage was set to 0.56V and the particle counts began to be recorded when the current was stabilized. Two sets and 3 repeats of the experiment have been performed.
Purity ratio measurement of exosome rich samples. The ratio of particle number, obtained by NTA (n = 6), to protein concentration, determined by BCA (n = 6), presented the purity of exosome-rich samples [72][73][74]. Purity ratio calculated with the equation as follows; Exosome characterization by flow cytometry (FCM). The carboxyl latex beads (#C37282, Thermo Fisher Scientific) were combined with antibodies of exosome surface markers [75]. The unconjugated anti-CD9 (#312102, Biolegend), and PE-labelled anti-CD81 (#349502, Biolegend) antibodies were used to capture exosomes. For the immune labeling process, 10 μL latex bead was completed to 500 μL with PBS and precipitated at 12000xg for 10 minutes. Pellet was dissolved in an anti-CD9 antibody so that the mixture was at 1μL bead for 1μg antibody, and the volume was completed to 100 μL with PBS. It was incubated while rotating at room temperature for 30 minutes. Incubated samples volume was completed to 500 μL with PBS and it was incubated overnight at room temperature. Bead and the antibody-containing solution was centrifuged at 12000xg for 10 minutes and dissolved in 1 ml 5% BSA solution. After incubation for 4 hours at room temperature for blocking, the mixture was centrifuged again at 12000xg for 10 minutes and the pellet was resuspended in 200 μL 1% BSA in PBS. Five micrograms of exosome were combined with 1 μL of the antibody-bead solution for staining. The final volume reached 50 μL with PBS and the mixture was held on 30 minutes incubation at room temperature. Upon incubation, the volume reached to 500 μL with PBS and the mixture was held on overnight incubation at room temperature while slowly rotating. The solution was precipitated at 12000xg for 10 minutes and the supernatant was discarded. Antibodies and their isotype control; mouse IgG1, κ PE (#400112, Biolegend, USA) were hold on 2 hours incubation with 1 μg/ml bead-exosome solution in 100 μL final volume at room temperature and avoid from the light. Incubated samples were washed with PBS. Samples were centrifuged at 12000xg for 10 minutes. The pellet was suspended in an appropriate volume of PBS and analyzed by flow cytometer (Novocyte, ACEA Biosciences, USA).

Statistical analysis
Shapiro Wilk test was used for evaluation of the normality of the distribution of the data obtained from analyses. Bonferroni and one-way ANOVA methods were used for the comparison of parametric data. Kruskal Wallis and Mann Whitney U tests for comparisons of nonparametric outputs. The SPSS statistics software (v23, IBM, USA) was used for these analyses. The degree of significance was p <0.05.

Protein quantification is determined by BCA analysis
The protein concentrations of MACS and ultracentrifuged isolated exosome sample were detected as 2.2 mg/ml and 1.8 mg/ml respectively when the results were normalized to 16 ml of starting sample. The particle number of the MACS group was higher when compared the that of the ultracentrifugation group, but the difference was not statistically significant ( Fig 1A).

Particle size and concentration is determined by NTA
The mean particle diameter of exosomes was recorded as 125 nm and 121 nm for MACS isolated and ultracentrifuged samples respectively. The maximum particle size range was found to be 88-426 nm with a peak of 110 nm in MACS isolated exosomes whereas it was noted as 80-383 nm with a peak of 110nm in ultracentrifuged exosomes. The mode of the particle diameter was detected as averagely 110 nm and 96 nm for MACS isolated and ultracentrifuged MSCs-exosomes respectively. The span value (d90-d10)/d50 reflecting the dispersity of particles was calculated as 0.52 and 0.60 for MACS isolated and ultracentrifuged particles respectively. The particle concentration of MACS isolated, and ultracentrifuged exosomes was 9.31 ±4.4x10 9 particles/ml and 3.34±1.7x10 9 particle/ml respectively. The particle number of the MACS group was higher (p = 0.011) when compared to the ultracentrifugation ( Fig 1B). All results were normalized to a 16 ml starting sample.

Morphological characterizations of isolated exosomes are occurred by TEM
Both the MACS and ultracentrifugation isolated exosomes exhibited negatively stained, spherical nano-sized particles. The samples that isolated MACS were in a range of sizes between 90-170 nm, and the ultracentrifuged samples were between 80-150 nm (Fig 3A-3C). The exosomes of both groups made round-shaped, intact particles or aggregated groups at the ultrastructural level. They have been distinct from the presence of the magnetic beads at MACS isolated samples (Fig 3B).

Discussion
We isolated hMSCs-exosomes by MACS and ultracentrifugation and comparatively characterized by protein quantification, NTA, and TEM in this study. hMSCs-exosomes that isolated by MACS and ultracentrifuge, revealed similar protein quantity (2.2±0.17 mg/ml and 1.8±0.25 mg/ml, respectively), particle size (in a range between 80-170nm), and round-shaped morphology. The particle number of exosomes that were isolated by MACS (9.31±4.4x10 9 particles/ml) was higher than that of ultracentrifugation (3.34±1.7x10 9 ). MACS isolation technique provided 2.3-fold high purity to exosomes when compared to ultracentrifugation. MACS isolated exosomes exhibited a 2-fold lower MFI value that was 331 when compared to ultracentrifuged (632).
MACS and ultracentrifugation provided 2.2±0.17 mg/ml and 1.8±0.25 mg/ml protein quantity by BCA in this study. Previous studies on ultracentrifuged BMSCs-Exo samples revealed a protein quantity within a wide range between 0.27 mg/ml-1.18x10 3 mg/ml [14,25,77]. In a recent study, protein quantity was reported as 0.27 mg/ml in rat BMSCs-Exo samples [14]. Our protein quantity was about six to eight times higher than that study. Protein concentration was 1.18x10 3 mg/ml in another study that reported on ultracentrifuged hBMSCs-exosomes samples [77]. Our findings were in line with that study. Sajeesh et al reported hBMSCsexosomes samples content as 1 mg/ml protein [25]. Different than our study, none of these studies reported their starting volume. This might be the reason of a wide range of protein quantity results in between these studies. BCA assessment should be supported by NTA analysis for particle number as it may not be solely related to the exosome purity [72,73,78]. MACS revealed a higher particle number when compared to ultracentrifugation by NTA in this study. Tan et al. [79] isolated hBMSCs-exosomes by phosphatidylserine labeled magnetic beads but did not compare their technique with ultracentrifugation. They reported the magnetically isolated hBMSCs-exosomes particle content as 4.1x10 9 particles per ml which contributes to the lower quantity when compared to our MACS samples (9.31±4.4x10 9 ) [79]. Our exosome-specific triple surface marker selection (CD9, CD63, and CD81) on the other hand presented a more reliable method than the phospholipid capturing technique since it is based on the presence of three exosome specific tetraspanins [80,81]. Apoptotic bodies with similar sizes with exosomes may be captured with magnetic beads when they are selected by phosphatidylserine which is a molecule that is shared by all extracellular vesicles [82]. Our ultracentrifuged samples revealed approximately 5-times higher particle numbers when compared to recent studies [25, 33, 83] reporting a range between 6 to 7.31x10 8 particle/ml and similar particle numbers when compared to Mead et al (1.17x10 9 ±1.42x10 8 particle per ml) [84] for hBMSCs-exosomes.
The ratio of protein to particle concentration determines the purity of exosome-rich samples [72,73,78]. Our results revealed the particle concentration to protein amount of exosomes as 4.23x10 9 by MACS and 1.85x10 9 by ultracentrifugation. The purity of human plasma exosomes was reported as > 3x10 10 particle/mg protein by precipitation technique [72]. The human urinary exosomes exhibited a particle to protein ratio of 1x10 9 particle/mg by purity ratio technique [73]. The particle to protein ratio of sucrose cushion-isolated prostate, urinary bladder and breast cancer cell lines [73] was reported as 5x10 9 -4x10 10 particle/mg by ultracentrifugation. There was no information on human BMSCs-Exo purity by MACS and ultracentrifuge methods in the literature. Weber et al stated the purity ratio range of cancer cell line exosome samples as highly pure when > 3x10 10 particle/mg protein, low purity when between 2x10 9 and 2x10 10 particle/mg and un-pure when less than 1.5x10 9 particle/mg. In this study, purity of MACS isolated, and ultracentrifuged hBMSC-exosomes were categorized at low purity according to that classification. The source of exosomes may influence the ratio of purity. Healthy hBMSCs may release lower exosomes when compared to cancer cell lines and other body fluids. Nevertheless, MACS purity was higher than that of the ultracentrifugation as presented in this study.
In our study, exosomes isolated by MACS and ultracentrifuge had both low spans; 0.52 and 0.60 respectively. The span value provides the particle size distribution (PSD) and a smaller value means a homogenous sample with a similar particle size in the NTA system [72,85]. Both samples were monodispersed whereas ultracentrifuge exosomes had more polydispersity when compared to MACS exosomes. Since MACS uses magnetic beads and lacks error-prone steps such as centrifugation, a little less polydispersity of MACS exosomes is understandable.
Here we additionally report the transmission electron microscopic appearance of exosomes that were isolated by MACS and ultracentrifugation as characteristic homogeneous, spherically shaped, and double membrane-layered structures. The MACS and ultracentrifugation isolated exosomes diameter range were noted in between 80 to 200 nm with a peak at 110 nm by NTA, and TEM revealed the size range between 90-170 nm, and 80-150 nm for MACS and ultracentrifugation respectively. Recent studies reported MACS isolated MSCs-exosomes having an average size in between 80.0±1.9 nm [68] and 170nm [79], with no information on protein quantity. Our particle size data by MACS  MACS and flow cytometric (FCM) characterization and selection are based on the same triple tetraspanin markers CD 9, CD 63, and CD 81 that compete when binding. So, MACS and FCM characterization were not applicable together. This could limit the comparison of MACS performance with FCM characterization when applied to the same sample. Those tetraspanins are stated as general exosome markers and present the most reliable tool when used three together by MACS [67,93] and FCM [80,81].
Our data may have several limitations. The in vitro and in vivo efficiency tests might be made to compare the performance of MACS and ultracentrifugation techniques in the isolation and characterization of exosomes. This limitation, however, does not exclude future in vitro, in vivo, and clinical investigations because the statistical correctness of our study was confirmed. Furthermore, MACS and ultracentrifugation are reliable in vitro methods that should be evaluated for potential future customized therapies before the clinic. The study does not comprise any technical assessment of body fluids or other cell types since the objective is to isolate, characterize and expand autogenic or allogenic MSC exosomes from the bone marrow to use as potential cell therapeutics. Candidate exosomes from each cell type should be separately tested and optimized before clinical translation.

Conclusion
The exosome isolation by MACS approach has many advantages such as providing isolation and characterization of exosomes on a single step, contributing high purity and rational yield; and disadvantages such as low working volume and biomarker screening problems when compared to ultracentrifugation. Ultracentrifugation is regarded as a gold standard and expensive technique for isolation of exosomes allowing working with large volumes. In this study, we proposed MACS as a practical and reliable alternative isolation technique that improves purity and speed to ultracentrifugation for exosomes. MACS technique is a promising isolation method for potential clinical applications of autogenic or allogeneic mesenchymal stem cells in which time, reliability, repeatability, and practicality have vital importance.