Exsosome Purification

HEK-293T cells (maintained within 10 passages) were trans - fected either with pEGFP-N1 or pCD-UPRT-EGFP using Lipofectamine 2000 (Invitrogen), according to manufacturer’s protocol. Five hours later, transfection media was replaced with media containing 5% MV-free FBS. 22 Three days after transfection, MVs were harvested, as described. 22 Briefly, culture medium from 2 × 150 mm plates containing 2 × 10 7 cells per plate was first centrifuged at 300 g for 15 minutes to separate cells from medium, then at 16,000 g for 30 minutes to precipitate cellular debris. The supernatants were filtered through 0.22 μm filters (Millex, Billerica, MA) and then a final ultracentrifugation was performed at 110,000 g for 80 minutes using Beckman Quick seal tubes and a 70Ti rotor (Beckman Coulter, Fullerton, CA). Pelleted MVs were eluted either in 50 μl (for in vitro experiments) or 20 μl (for in vivo studies) in a mixture containing 1× phosphate-buffered saline, RNase inhibitor (2 μl, 10U-RNAse-OUT—Invitrogen) and rDNase I (1 μl, 2U-DNA-free—Ambion, Grand Island, NY). One μl MVs were deliv - ered into each tumor once a week for 2 months, with fresh MVs prepared for each injection. Every 2 weeks, we generated new, early passage cultures of the cells to isolate MVs to make sure that there were no extended passage differences between the cells from which the MVs were isolated. MV yields were determined by measuring total RNA content using a NanoDrop 1000 (NanoDrop, Wilmington, DE) which correlated directly with MV number as assessed using a Nanosight, NS500 ( Supplementary Figure S8 ). For in vivo studies, MVs were prepared in the same manner and similar amounts of MVs were used in both groups as assessed by total RNA content: 189 ± 21 ngin 20 μl for CD-UPRT-EGFP MVs and 201 ± 38 ng in 20 μl for the control EGFP MVs. In some cases MVs were treated with 1 μl DNaseI (2U; Ambion) in 50 μl of total reaction for 30 minutes at 37 °C to remove DNA bound to the surface. In order to make sure that our DNase treatment protocol was functional, we performed qPCR reactions in the absence of RT directly from the MVs content and found no PCR amplification suggesting that the treatment protocol completely removed any plasmid DNA purified during MVs isolation. Sucrose gradient ultracentrifuga - tion was performed, as described previously. 34 Briefly, MVs were layered onto a sucrose density gradient (8, 30, 45, 60% layers) and centrifuged for 38 minutes at 50,000 rpm in SW40Ti swinging bucket rotor (Beckman Coulter) in a Beckman Optima ultracentrifuge with deceleration set to slow. Fractions 3–7 (density: fraction 3 contains: half 8%, half 30%, frac - tions 4 and 5 contains 30% only, fractions 6 and 7 contains 45% only) were collected, diluted in phosphate-buffered saline and MVs pelleted at 100,000 g for 75 minutes in the S50A rotor using a Sorvall MX-120 micro - centrifuge (Thermo Fisher Scientific, Agawam, MA) and used for western blot analysis for MV-associated proteins

Briefly, the culture supernatant was cleared from
cell debris and large vesicles by centrifugation at 2,000 rpm
for 10 minutes followed by 0.8 and 0.2 μm filtrations.
Supernatant concentration and buffer-exchange to PBS was
performed by ultrafiltration/diafiltration using a fiber car-
tridge with a 500 kD cut-off. The concentrated fraction was
subjected to ultracentrifugation on a 30%-sucrose cushion
in PBS/D2O and exosomes were recovered by collecting the
sucrose-containing layer. The exosome fraction was further
concentrated and diafiltered to yield a 500X-concentrated
solution of purified exosomes. Exosomes were aliquoted
and stored at –80 °C until use. Protein content of purified
exosomes was determined using a Bradford assay (Pierce,
Rockford, IL). Exosomes prepared following this method
appear as a homogeneous population of vesicles 50–100 nm in
diameter when evaluated by electron microscopy (Lamparski
et al 2002).