Diversity of extracellular vesicles addressed by immunoaffinity beads

Nina Koliha, Yvonne Wiencek, Ute Heider, Andreas Bosio, and Stefan Wild Miltenyi Biotec GmbH, Bergisch Gladbach, Germany

Analysis of surface markers on EVs Introduction 3 suggests presence of EV subpopulations Exosomes are extracellular vesicles (EVs) that are exosomes in biological fluids (e. g. plasma), specific EVs from activated B cells were incubated with the 39- released by a variety of cell types either constitutively markers are needed. plex beads and stained with a cocktail of CD9, CD63, EVs derived from activated B cells or in a stimulation-induced manner. Depending on We established a multiplex bead–based assay and CD81-APC or with selected single EVs derived from ac/vated B cells the originating cell, exosomes are loaded with a consisting of capture and detection antibodies to antibodies (fig. 5). With regard to B cell–specific markers, specific set of , , and nucleic acids. To analyze the composition of exosome surface proteins anti-CD19 beads showed stronger signals than anti- 40 35 investigate the origin, composition, and function of in a given sample by flow cytometry. CD20 beads after staining with the cocktail 30 25 (APC median signal 11.3 and 5.4, respectively). We 20 15 therefore conclude that less EVs were positive for aCD9/63/81 10 aHLA-DQ 5 aCD86 CD20 than for CD19. Conversely, on all bead types 0 aCD80 Background-corrected APC signal aCD69 CD20-APC signals were stronger than CD19-APC aCD20 Methods aCD19 signals, suggesting that the amounts of CD20 per EV Capture antibody bead were higher compared to CD19. We propose a Multiplex bead–based assay Figure 5 Color-coded polystyrene multiplex beads were 5 min. The beads were resuspended in 100 µL PAP subpopulation of B cell EVs carrying high levels of CD20. The anti-CD42a beads and anti-CD9 beads exhibited Matrix profile of B cell EVs. Background-corrected APC median signal intensity of different capture incubated with isolated EVs in 300 µL PAP (PBS, 0.1% and bound EVs were stained with 0.5 µg detection antibody bead types after incubation with 32 µg EVs from activated B cells, followed by staining with signals exclusively after staining with the CD9/63/81- CD19-APC, CD20-APC, CD69-APC, CD80-APC, CD86-APC, Anti-HLA-DQ-APC, or a cocktail of CD9-APC, Pluronic® F127, 0.09% azide) at 4 °C overnight. Beads antibody and analyzed by flow cytometry. CD63-APC, and CD81-APC antibodies. REA and mIgG1 indicate isotype control beads. were washed in PAP and centrifuged at 3,000×g for APC antibody cocktail. The signals on the anti-CD42a beads indicated the presence of during cell culture. Therefore, we performed cell stainings after Up to 39 capture antibody beads + Detection Flow cytometry antibody (cocktail) analysis B cell isolation and indeed detected 90.5% CD42a- positive events if no gates and triggers were set (data not shown), confirming the presence of platelets within the isolated B cell fraction. We conclude that the signal on the anti-CD9 beads was most likely due to Isolated exosomes or EVs and that B cell EVs most likely are CD9-negative. cell culture supernatant Incubation over 1 h incubation, night, washing washing Date:2015-03-18 Date:2015-03-18 Date:2015-03-18 Depletion of platelet EVs using superparamagnetic MicroBeads Operator: Nina Koliha Operator: Nina Koliha Operator: Nina Koliha Dye-labeled capture antibody bead (4.8 µm) Exosome APC-conjugated detection antibody 4 Version:2.6.1508.13419 Version:2.6.1508.13419 Figure 1 Version:2.6.1508.13419 Workflow of the multiplex bead platform: Isolated exosomes were incubated overnight with 39 differently labeled beads each coupled to a different capture antibody. Bound exosomes were detected with APC-conjugated antibodies. To purify the B cell EV sample, we depleted platelet 9 EVs using CD61 MicroBeads. Subsequently, the isolated A B C 8 NIKNIKNIK2015NIK201520152015----03030303----17171717____2426242624262426....0030003000300030 NIKNIKNIK2015NIK201520152015----03030303----17171717____2426...... 2015..242624262426.201520152015-...0030---03003000300030030303----17171717____2426242624262426....0030003000300030\\\\PPPP1111 ...... 2015..201520152015----03030303----17171717____2426242624262426.NIKNIKNIK2015NIK...0030003000300030\201520152015-\\\P---03PPP1030303-111 ---17171717____2426242624262426....0030003000300030 NIKNIKNIK2015NIK201520152015----03030303----17171717____2426242624262426....0030003000300030 NIKNIKNIK2015NIK201520152015----03030303----17171717____2426242624262426....00300030003000301e3 ...... 2015..201520152015----03030303-1e3---17171717____2426242624262426....0030003000300030\\\\P1e3PPP1111 NIKNIKNIK2015NIK201520152015-1e3---03030303----17171717____2426242624262426....00300030003000301e3 1e3 EV fractions were analyzed on the multiplex platform 1e3 1e3 1e3 7 in comparison to the original B cell EV sample (fig. 6). P1 P1 P1 1e2 1e2 1e2 1e2 6 1e2 1e2 1e2 1e2 + 1e2 A A A A A A A A In the flow-through fraction depleted of CD61 EVs, A A A A - - A A A A - - A A A A - - A A A A - - A A - - - - A A ------5 A A A A - - A A A A ------1e1 1e1 PE PE PE PE 1e1 PE PE 1e1 PE PE PE PE PE PE PE PE PE PE SSC SSC SSC SSC SSC SSC SSC SSC 1e1 1e1 PE PE PE PE PE PE PE PE the signals for each of the three investigated platelet SSC SSC SSC SSC 1e1 1e1 4 1e1 1 1 1 1 1 1 markers (CD41b, CD42a, and CD62P) were drastically 0 0 0 0 3 0 0 1e0 1e0 -1 -1 -1 -1 reduced. In contrast, the platelet EVs were detected 1e0 1e0 1e1-1 1e2 1e31e0 1e1-1 1e2 0 1 1e3 1e1-1 1e2 1e3-1 0 1 1e1-1 0 1e2 1 1e3 1e1 1e2 1e3-1 0 1 1e1 1e2 1e3 1e0 1e1 1e2 1e3 -1 0 1 1e1 1e2 1e3 -1 0 1 1e1 1e2 1e3 2 FSCFSCFSC-FSC---AAAA FSCFSCFSC-FSC---AAAA FITCFITCFITC-FITC---AAAA FITCFITCFITC-FITC---AAAA CDCDCD9CD999////63636363////81818181----APCAPCAPCAPC----AAAA CDCDCD9CD999////63636363////81818181----APCAPCAPCAPC----AAAA in the eluate of the column-based sorting and were Figure 2 FSCFSCFSC-FSC---AAAA FITCFITCFITC-FITC---AAAA CDCDCD9CD999////63636363////81818181----APCAPCAPCAPC----AAAA median signal intensity 1 found to be CD9-positive. This supports the hypothesis File NIK2015-03-17_2426.0030.mqdFile NIK2015-03-17_2426.0030.mqd2015-Mar-17 13:25 2015-Mar-17 13:25 Analysis example showingFile NIK2015-03-17_2426.0030.mqd (A) gating according to bead size, (B) discrimination2015-Mar-17 of differently 13:25 labeled bead populations, and (C) measurement of signal intensities of the single bead populations. 0 SID 33-Plex_2 + 32g platelet SIDexos NK3 33-Plex_2 + CD9/63/81-AP + 32g plateletC Descr.exos NK3 + CD9/63/81-APC Descr. SID 33-Plex_2 + 32g platelet exos NK3 + CD9/63/81-APC Descr. that B cell EVs are CD9-negative and that the signal CD62P CD41b CD19 CD20 CD9 CD63 CD81 Background corrected CD9/63/81-APC corrected CD9/63/81-APC Background PathPathPath PathPathPath %-%-%-#%-### %-%-%-#%-Count### CDCDCD9 CD999////63636363////81818181----APCAPCAPCAPC----AAAACount Median CDCDCD9 CD999////63636363////81818181----APCAPCAPCAPC----AAAA Median ImmunomagneticPathPathPath separation of EVs %-%-%-#%-### Count CDCDCD9 CD999////63636363////81818181----APCAPCAPCAPC----AAAA Median on anti-CD9 beads obtained with the original B cell Starting material (B cell exosome preparation) P1\1/CD9 P1\1/CD9 1,26 1,26170 17060,57 60,57 P1\1/CD9 1,26 170 60,57 CD61 enriched fraction (eluate) Plasma or isolated EVs fromP1\2/CD63 cell cultureP1\2/CD63 supernatant 1,16 1,16157 15753,88 53,88 preparation were indeed due to contaminating P1\2/CD63 1,16 157column, washed,53,88 and eluted after removal of the P1\4/CD82 P1\4/CD82 0,95 0,95128 12854,50 54,50 Figure 6 CD61 depleted fraction (flow through) P1\4/CD82 0,95 128 54,50 P1\11/CD11c P1\11/CD11c 1,87 1,87252 25229,06 29,06 were incubatedP1\11/CD11c with 50 µL CD9, CD61, CD63,1,87 or CD81 252column from 29,06the magnetic field. For the multiplex platelet EVs. P1\12/CD3 P1\12/CD3 1,04 1,04140 1405,40 5,40 P1\12/CD3 1,04 140 5,40 Surface marker profile of EVs from activated B cells before and after depletion of platelet EVs using P1\15/CD13 P1\15/CD13 0,85 0,85115 115100,06 100,06 MicroBeadsP1\15/CD13 (Miltenyi Biotec) for 1 h.0,85 Magnetically 115platform analysis,100,06 EVs and beads were co-eluted with CD61 MicroBeads. The initial EV sample from activated B cells as well as the EV eluate and flow-through P1\17/CD19 P1\17/CD19 1,59 1,59215 21541,27 41,27 + P1\17/CD19 1,59 215 41,27 from the CD61 EV depletion were incubated with 39 capture antibody bead types. Data indicate the P1\19/CD20 P1\19/CD20 6,49 6,49875 87525,66 25,66 P1\19/CD20 6,49 875 25,66 P1\20/CD29 P1\20/CD29 1,66 1,66224 22422,52 22,52 corrected APC median signal intensities after staining with a cocktail of CD9-, CD63-, and CD81-APC labeled EVsP1\20/CD29 were applied to a MACS® Column1,66 placed 224100 µL PAP buffer.22,52 100 µL hot (95 °C) 1× SDS Loading P1\24/CD42a P1\24/CD42a 3,10 3,10418 4186,29 6,29 antibodies. mIgG1 indicates isotype control beads. P1\24/CD42a 3,10 418 6,29 P1\30/CD86 P1\30/CD86 5,30 5,30714 71411,97 11,97 in a magneticP1\30/CD86 field. Labeled EVs were retained5,30 in the 714Buffer was used11,97 to elute EVs for Western Blot analysis. P1\33/CD209 P1\33/CD209 1,46 1,46197 19790,69 90,69 P1\33/CD209 1,46 197 90,69 P1\37/REA IC P1\37/REA IC 4,32 4,32583 5830,51 0,51 P1\37/REA IC 4,32 583 0,51 P1\38/mIgG1 P1\38/mIgG1 3,14 3,14423 4230,20 0,20 P1\38/mIgG1 3,14 423 0,20 P1\39/mIgG2b P1\39/mIgG2b 3,30 3,30445 4450,85 0,85 P1\39/mIgG2b 3,30 445 0,85 Analysis of EVs isolated from plasma Results 5 using superparamagnetic MicroBeads EVs from 2 mL of plasma were isolated using CD9, (amounts used were adjusted to a plasma volume of CD63, or CD81 MicroBeads and analyzed by the 2 mL). We conclude that the immunomagnetic multiplex bead assay (fig. 7). In comparison to EVs isolation of EVs from plasma is efficient even EVs from NK cells and platelets differ in their composition isolated by ultracentrifugation, the magnetically from plasma samples where available volumes might 1 isolated EVs showed brighter signals in most cases be small.

NK cell– or platelet-derived EVs were incubated with EVs to mimic two distinct EV populations in a single Donor A the 39-plex beads and stained with APC-conjugated sample. After staining with a cocktail of CD9-, CD63-, 450 antibodies detecting the CD9, CD63, or and CD81-APC antibodies, each capture antibody 400 CD81 (Miltenyi Biotec, Germany) or with a cocktail of bead type gave a signal indicating that EVs bound 350 these antibodies. to all three bead types. Using a staining cocktail, 300 Page 2/2 Page 2/2 For the NK cell–derived EVs, no signals werePage 2/2detected CD9–CD81+ EV populations could not be distinguished 250 200 on anti-CD9 beads (fig. 3A). Additionally, no CD9-APC from CD9+CD81– populations (fig. 3C). However, 150 antibody staining was observed on any bead type. separate stainings with CD9 or CD81 antibody 100 + + We conclude that NK cell–derived EVs carry hardly any gave very low CD9 CD81 double-positive signals 50 CD9. Platelet-derived EVs appeared to carry only low demonstrating that CD9 and CD81 do not co-exist on 0 amounts of CD81, i.e., hardly any EVs were detected the same EV, i.e., the sample comprises two CD1c CD2 CD3CD4CD8CD9 CD14CD19 CD11c CD20CD24CD25CD29CD31CD40CD41bCD42aCD44CD45CD49eCD56CD62PCD63CD69CD81CD86CD105 CD142CD146CD209CD326 MCSP ROR1 mlgG1 – + CD133/1 SSEA-4 on anti-CD81 beads or on any other bead type after subpopulations of CD9 CD81 (NK cell) EVs and HLA-ABC control REA control CD81-APC antibody staining (fig. 3B). Based on this CD9+CD81– (platelet) EVs. HLA-DRDPDQ Donor B finding we prepared a mixture of NK cell and platelet 450 400 A B C Mixture of NK cell– and 350 NK cell–derived EVs Platelet-derived EVs platelet-derived EVs NK cell–derived EVs Platelet-­‐derived EVs Mixture of NK cell– and platelet-­‐derived EVs 300 250 30 30 30 200 25 25 25 150 20 20 20

15 median signal intensity Background corrected CD9/63/81-APC 15 15 100 10 10 10 50 CD9, 63, 81 5 CD9, 63, 81 CD9, 63, 81 5 CD81 0 CD81 5 CD63 0

Background-corrected APC signal 0 CD63 CD9 CD63 Background-corrected APC signal 0 CD81

Background-corrected APC signal CD2 CD3CD4CD8CD9 CD9 CD9 CD63 CD9 CD9 CD9 CD1c CD14CD19CD20CD24CD25CD29CD31CD40 CD44CD45 CD56 CD63CD69CD81CD86 ROR1 CD63 CD81 CD63 CD11c CD41bCD42a CD49e CD62P CD105 CD142CD146CD209CD326 MCSP mlgG1 CD81 CD81 CD133/1 SSEA-4 mIgG1 mIgG1 mIgG1 HLA-ABC control Capture antibody bead Capture antibody bead REA control Capture antibody bead HLA-DRDPDQ Ultracentrifuge CD9 MicroBeads CD63 MicroBeads CD81 MicroBeads Figure 3 Figure 7

(A–C) Background-corrected APC median signal intensities of anti-CD9-, anti-CD63-, anti-CD81 beads and isotype control beads after incubation with (A) 16 µg NK cell–derived EVs, (B) 16 µg platelet-derived EVs, Surface marker profile of EVs isolated from plasma of donor A (top) or B (bottom) by ultracentrifugation or immunomagnetic isolation using CD9, CD63, or CD81 MicroBeads. Data indicate APC median signal and (C) a mixture of both NK cell–derived and platelet-derived EVs (8 µg each) followed by staining with CD9-APC, CD63-APC, or CD81-APC antibodies or with a cocktail of these antibodies. mIgG1 indicates isotype intensities of isolated EVs incubated with the 39 capture antibody bead types and stained with a cocktail of CD9-, CD63-, and CD81-APC antibodies. REA and mIgG1 indicate isotype control beads. control beads.

EV subpopulations can be separated 2 using superparamagnetic MicroBeads Conclusion and outlook We used the same mixture of NK cell and platelet EVs • The multiplex bead platform allows the specific The relative abundance of epitopes on the surface of and separated the respective subpopulation using detection of exosome surface proteins and the exosomes varies depending on the originating cell CD9 MicroBeads or CD81 MicroBeads. Western blot determination of their relative abundance on type and its status. Comprehensive analysis of exosome analysis of the mixture of NK cell and platelet EVs gave exosomes from different sources. surface compositions will allow the no indication of the two subpopulations as CD29, • The common exosome markers were not classification of exosome populations according to CD81, and CD9 were detected (fig. 4 middle). However, equally distributed in all exosome populations: their origin, for example. Isolation of specific exosome using CD9 MicroBeads and CD81 MicroBeads, the NK cell exosomes carried less CD9, while CD81 populations with immunoaffinity beads might give CD9+ EVs (platelet origin) and CD81+ NK cell EVs could further insight into different EV functions. was underrepresented on platelet exosomes. be separated from each other. CD63 MicroBeads were used as control. • Immunoaffinity-based isolation allows physical References Figure 4 1. Théry, C. et al. (2006) Curr. Protoc. Cell Biol. 3.22.1–3.22.29. separation of EV subpopulations from a sample 2. Koliha, N. et al. (2016) J. Extracell. Vesicles 5: 29975. doi: 10.3402/jev.v5.29975. Western blot analysis of NK cell EVs (left) a mixture of NK cell and platelet EVs (middle) and platelet EVs (right). Ctrl.: Loading control of 1.7 µg NK cell EVs (left), a mixture of 0.8 µg NK cell EVs and 1.7 µg for further analysis. platelet EVs (middle), and 3.3 µg platelet EVs (right). CD63-MB, CD9-MB, CD81-MB: EVs isolated with Unless otherwise specifically indicated, Miltenyi Biotec products and services are for research use only CD63, CD9, or CD81 MicroBeads (25 µg NK cell EVs, a mixture of 12.5 µg NK cell EVs and 25 µg platelet • Superparamagnetic MicroBeads can be used and not for therapeutic or diagnostic use. MACS is a registered trademark of Miltenyi Biotec GmbH. EVs, and 50 µg platelet EVs were used as starting material for magnetic sorting). CD29, CD81, and CD9: All other trademarks mentioned in this document are the property of their respective owners and are detection antibodies. for direct isolation of exosomes from plasma. used for identification purposes only. Copyright © 2016 Miltenyi Biotec GmbH. All rights reserved.