A multiplex bead platform for the characterization
of cell culture–and plasma-derived exosomes
Nina Koliha, Ute Heider, Andreas Bosio, and Stefan Wild
Miltenyi Biotec GmbH, Department of Research & Development, Bergisch Gladbach, Germany
2
Introduction
Exosomes are released by a variety of cell types either constitutively
or in a stimulation-induced fashion. Depending on the originating
cell, exosomes are loaded with a specific set of proteins, lipids,
and nucleic acids. To investigate the origin, composition, and
function of exosomes in biological fluids (e. g. plasma), specific
Surface proteins of exosomes from cultured primary cells were
investigated using the multiplex bead platform. Well-established
cell markers were detected on the respective secreted exosomes,
e.g., CD2, CD3, CD4, and CD8 on T cell exosomes, CD56 on NK cell
exosomes, and CD61 and CD42a on platelet exosomes. Exosome
surface protein profiles can be used to draw conclusions on
their origin or to investigate differences between exosomes
markers are needed. We established a multiplex bead-based
platform consisting of capture and detection antibodies to analyze
the composition of exosome surface proteins in a given sample
by flow cytometry.
Methods
A
1e3
P1
1e2
PE
PE--A
1e1
1e1
1e0
1e0
1e1
1e2
1
0
-1
1e3
Date:2015-03-18
Operator: Nina Koliha
Version:2.6.1508.13419
-1 0 1
FSCFSC- A
B
NIK2015
NIK20152015 - 0303 - 17_
17 _2426.
2426 .0030
1e2
P1\1/CD9
P1\2/CD63
P1\4/CD82
1e1
P1\11/CD11c
P1\12/CD3
1e1
Date:2015-03-18
P1\15/CD13
1
P1\17/CD19
Operator: Nina Koliha
0
P1\19/CD20
Version:2.6.1508.13419
1e0
-1
P1\20/CD29
1e0
1e1
1e2
1e3
-1 0 1
1e1
1e2
1e3
P1\24/CD42a
P1\30/CD86
FSCFITCFSC- A
FITC- A
P1\33/CD209
P1\37/REA IC
P1\38/mIgG1 2015-Mar-17 13:25
File NIK2015-03-17_2426.0030.mqd
P1\39/mIgG2b
SID 33-Plex_2 + 32g platelet exos NK3 + CD9/63/81-APC Descr.
..2015
NIK2015
..20152015 - 0303 - 17_
17 _2426.
2426 .0030\
0030 \P1
NIK20152015 - 0303 - 17_
17 _2426.
2426 .0030
1e3
1e3
Path
%-#
%- #
PE
PE--A
PE
PE--A
SSC
SSC--A
1 h incubation,
washing
C
1e3
NIK2015
NIK20152015 - 0303 - 17_
17 _2426.
2426 .0030
P1
1e2
P1\1/CD9
P1\2/CD63
P1\4/CD82
P1\11/CD11c
1e1
P1\12/CD3
1e1
P1\15/CD13
1
P1\17/CD19
P1\19/CD20
0
P1\20/CD29
-1
1e0
P1\24/CD42a
1e0
1e1
1e2
1e3
-1 0 1
1e1
1e2
P1\30/CD86
FSCFITCFSC- A
FITC- A
P1\33/CD209
P1\37/REA IC
P1\38/mIgG1
File NIK2015-03-17_2426.0030.mqd
P1\39/mIgG2b 2015-Mar-17 13:25
SID 33-Plex_2 + 32g platelet exos NK3 + CD9/63/81-APC Descr.
1e2
PE
PE--A
PE
PE--A
SSC
SSC--A
1e2
Dye-labeled capture antibody bead (4.8 µm)
Exosome
APC-conjugated detection antibody
Figure 1
P1\4/CD82
P1\11/CD11c
P1\12/CD3
P1\15/CD13
P1\17/CD19
P1\19/CD20
P1\20/CD29
P1\24/CD42a
P1\30/CD86
P1\33/CD209
P1\37/REA IC
P1\38/mIgG1
P1\39/mIgG2b
1,26
1,16
0,95
1e1
1,87
1,04
0,85
1
1,59
0
6,49
-1
1,66
-1 0 1 3,101e1
1e2
5,30
CD9
CD9 /63/
63 /8181 -APCAPC- A
1,46
4,32
3,14
3,30
140
115
215
875
224
418
714
197
583
423
445
170
157
128
252
140
115
215
875
224
418
714
197
583
423
445
60,57
53,88
54,50
29,06
5,40
100,06
41,27
25,66
22,52
6,29
11,97
90,69
0,51
0,20
0,85
to the mean signal intensity of the beads coupled to the antibodies
detecting exosome markers, i.e., CD9, CD63, and CD81.
Source
Isolation
Purity
Medium
Cultivation time
NK cells
Buffy coats
MACSxpress®
Technology
95–98%
T cell medium + 5% human AB serum 12 days
+ 500 IU/mL Proleukin + EBV-LCLs
Platelets
Fresh
whole blood
Serial
centrifugation
92–99%
Krebs Ringer buffer # + 50 nM calcium 30 min
ionophore A23187 + 10 mM CaCl₂¹
T cells
PBMCs
MACS® MicroBead
Technology
96–99%
T cell medium + 5 U/mL IL-2
+ 2.5 µg/mL CD28 + CD3 coating²
24 h
RPMI 1640 + 2 mM L-glutamine
72 h
2,000×g for 30 min, 10,000×g
for 45 min, 0.22 µm filtrated
supernatant. Page 2/2
Exosomes were sedimented at
100,000×g for 2 h, washed with
PBS, and then resuspended in
PBS. Protein concentration was
determined by BCA Assay and
BSA as standard.³
Page 2/2
Results
Count
CD9
CD9 /63/
63 /8181 -APCAPC- A Median
170
157
128
252
140
115
215
875
224
418
714
197
583
423
445
60,57
53,88
54,50
29,06
5,40
100,06
41,27
25,66
22,52
6,29
11,97
90,69
0,51
0,20
0,85
3
Capture antibody beads
Capture antibody beads
Plasma exosomes have mainly HLA
class II and CD61 on their surface
Plasma exosomes from four different donors showed
comparable signal intensities for the analyzed epitopes after
normalization to the mean signal intensities of anti-CD9,
anti-CD63, and anti-CD81 beads. The highest signal intensities
were detected on anti-HLA class II beads and anti-CD61 beads
(fig. 5). This suggests that HLA class II–expressing cells, such as
antigen-presenting cells, and platelets were the main sources
of exosomes in peripheral blood.
Capture antibody beads
Figure 5: Normalized signal intensities of different capture antibody beads after overnight incubation with 64 µg of plasma
exosomes from four donors (D1–D4) and staining with a cocktail of CD9-APC, CD63-APC, and CD81-APC antibodies for 1 h.
the anti-CD63 beads (fig. 3A), demonstrating that exosome binding
to the beads was specific.
Signal intensities of the capture antibody beads were comparable
in an 8-plex and 34-plex format (fig. 3B), suggesting that the
composition of the bead set or the number of beads used did not
affect the signal intensities.
B
NK exosomes, D1 (34-plex)
NK exosomes, D1 (8-plex)
NK exosomes, D2 (34-plex)
NK exosomes, D2 (8-plex)
NK exosomes, D3 (34-plex)
NK exosomes, D3 (8-plex)
NK exosomes, D4 (34-plex)
NK exosomes, D4 (8-plex)
CD9/63/81-APC median signal intensity
alone
mlgG1
aCD63
4
Exosomes from different cancer cell lines
reflect markers of the originating tumor
Two cancer-related antigens were tested for their presence on
cancer exosomes. The melanoma-associated chondroitin sulfate
proteoglycan antigen (MCSP) is expressed on the majority (>90%) of
human melanoma tissues and melanoma cell lines, but not on
–
carcinoma cells⁴ ⁶. In contrast, the epithelial cell adhesion molecule
(EpCAM) is broadly expressed on the basolateral surface of carcinomas,
but not on melanoma⁷. Consequently, MCSP could be detected on
exosomes derived from three primary melanoma cell lines, but
hardly on exosomes from the colon carcinoma cell line HT-29 (fig. 6).
Likewise, EpCAM was detected on the colon carcinoma exosomes,
but barely on the three melanoma exosome preparations (fig. 6).
Plasma exosomes D1
Plasma exosomes D2
Plasma exosomes D3
Plasma exosomes D4
Figure 5
Signal intensities are specific and not impaired by the
number of beads or the composition of the bead set
A
Page 2/2
1e3
Exosome isolation
Table 1: Overview of cell isolation and cultivation for exosome production. # Krebs Ringer buffer (100 mM NaCl, 4 mM KCl, 20 mM NaHCO₃, 2 mM Na₂SO₄, 4.7 mM citric acid, 14.2 mM trisodium citrate)
In order to verify the specificity of exosome binding to the capture
antibody beads, the beads were incubated with NK cell exosomes
alone or in combination with soluble mouse IgG1 as isotype control
or soluble CD63 antibody to block the binding to the anti-CD63
beads. In contrast to the isotype control, blocking with soluble
CD63 antibody specifically inhibited binding of the exosomes to
1e3
CD9
CD9 /63/
63 /8181 -APCAPC- A Median
5,40
100,06
41,27
25,66
22,52
6,29
11,97
90,69
0,51
0,20
0,85
1e2
Figure 4: (A) Background-corrected raw data and (B) normalized signal intensities of different capture antibody beads after overnight incubation with 8 µg NK cell exosomes, 32 µg platelet exosomes, or 2.5 µg T cell exosomes and staining with a cocktail of CD9-APC,
CD63-APC, and CD81-APC antibodies for 1 h. # One outlier was excluded from the analysis.
Count
Figure 2: Analysis example
showing
gating
according
Count
CD9
-APC- A Median
CD9 /63/
63 /8181(A)
APC
to bead size, (B) discrimination of differently labeled bead
1,26
170
60,57
populations, and (C) measurement
of signal intensities
of the
1,16
157
53,88
0,95
128
54,50
single bead populations.
1,87
252
29,06
1e1
Figure 4
1e2
%-#
%- #
1,04
0,85
1,59
6,49
1,66
3,10
5,30
1,46
4,32
3,14
3,30
Originating cell type
Cell lines
1e3
4,32
3,14
3,30
Figure 2
Figure 1: Workflow of the multiplex bead platform. Isolated exosomes or cell culture supernatant were incubated overnight with up to 39 different bead types,
each coupled to a different
Path
capture antibody. The different bead types were labeled with one of two dyes or different proportions of both dyes, so that all bead types were distinguishable by flow cytometry. Exosomes
P1\1/CD9
bound to the beads were detected with CD9-APC, CD63-APC, or CD81-APC antibodies or with a cocktail of these antibodies.
P1\2/CD63
One possibility of comparing signal intensities from different
exosome populations is the normalization of the median signal
intensities from the different capture antibody bead populations
1e3
1,26
1,16
0,95
1,87
1e1
1,04
0,85
1
1,59
6,49
0
1,66
-1
-1 0 1 3,101e1
1e2
5,30
CD9
9 /63/
CD1,46
63 /8181 -APCAPC- A
-1 0 1
CD9
CD9 /63/
63 /8181 -APCAPC- A
File NIK2015-03-17_2426.0030.mqd
2015-Mar-17 13:25
SID 33-Plex_2
32g_2426.
platelet
exos
Descr.
NIK2015
- 03..2015
.0030\
\P1NK3 + CD9/63/81-APC
NIK20152015
03 - 17_
17 _2426.
2426 .0030
..20152015 - 0303+- 17_
17
2426
0030
1e3
1e3
Path
%-#
%- #
1e2
Incubation over
night, washing
1e3
FITCFITC- A
P1
Isolated exosomes or
cell culture supernatant
1e2
1
0
-1
Capture antibody beads
Figure 3
Figure 3: (A) Median signal intensities after overnight incubation of capture antibody beads with 8 µg of NK cell exosomes alone (four donors) or in combination with soluble mouse IgG1 isotype control or soluble CD63 antibody, followed by staining with a
CD81-APC antibody for 1 h. (B) Median signal intensities after overnight incubation of 8 µg of NK cell exosomes (from four donors, D1–D4) with eight (8-plex) or 34 (34-plex) different capture antibody beads and staining with a cocktail of CD9-APC, CD63-APC
and CD81-APC antibodies for 1 h. REA ctrl and mIgG1 indicate isotype control beads, i.e., beads linked to an antibody, which does not bind to exosomes.
P1 melanoma exosomes
P7 melanoma exosomes
P10 melanoma exosomes
Colon carcinoma exosomes
Normalized CD81-APC signal intensity (%)
1e3
1e1
1e1
Normalized CD9/63/81-APC signal intensity (%)
+ Detection
antibody (cocktail)
PE
PE--A
Flow cytometry
analysis
NIK2015
NIK20152015 - 0303 - 17_
17 _2426.
2426 .0030
1e3
1e2
SSC
SSC--A
Up to 39 capture antibody beads
..2015
..20152015 - 0303 - 17_
17 _2426.
2426 .0030\
0030 \P1
Normalized CD9/63/81-APC signal intensity (%)
NIK2015
NIK20152015 - 0303 - 17_
17 _2426.
2426 .0030
CD9/63/81-APC median signal intensity
Version:2.6.1508.13419
1e2
CD81-APC median signal intensity
T cell exosomes (n = 4)
NK cell exosomes (n = 4)
Platelet exosomes (n = 4)
Operator: Nina Koliha
1e3
Capture antibody beads
and the membrane composition of the originating cells.
The common exosome markers CD9, CD63, and CD81 were not
equally distributed on the investigated populations. CD9 was
underrepresented on NK cell exosomes as well as CD81 on
platelet exosomes. In contrast, T cell exosomes seemed to
carry comparable amounts of CD9, CD63, and CD81 on their surface
(fig. 4).
B
Date:2015-03-18
A
1
Exosomes from primary cells carry cell-specific surface
proteins and differ in common exosome markers
Figure 6
Capture antibody beads
Figure 6: Normalized signal intensities of different capture antibody beads after incubation with 4 µg of exosomes from
melanoma or colon carcinoma cell lines and staining with a CD81-APC antibody.
Conclusion & outlook
• The multiplex bead platform allows the specific detection of
exosome surface proteins and the assessment of their relative
abundance on exosomes from different sources.
• The composition of the multiplex bead set did not affect
signal intensities.
• The common exosome markers were not equally distributed
in all exosome populations: NK cell exosomes had less CD9,
while CD81 was underrepresented on platelet exosomes.
• Well-established cell markers were detectable on the
secreted exosomes, i.e., CD3 on T cell exosomes, CD56
on NK cell exosomes, and CD61 on platelet exosomes.
• The presence of HLA class II and CD61 on plasma exosomes
suggests that HLA class II–expressing cells, such as antigenpresenting cells, and platelets are major sources of exosomes
in blood.
• Tumor-associated markers were specifically detected on
tumor cell line–derived exosomes, namely MCSP on melanoma
exosomes and EpCAM on colon carcinoma exosomes.
The relative abundance of epitopes on exosome surfaces varies
depending on the originating cell type and its status. A
comprehensive analysis of exosome surface protein composition
will enable the classification of exosome populations, e.g.,
according to their origin, and might give insight into their different
functions. In our ongoing research we investigate the impact of
cell stimulation and cancer status on the exosome protein profile.
References
1. Leong, H.S. et al. (2011) J. Thromb. Haemost. 12: 2466–2476.
2. van der Vlist, E. et al. (2012) J. Extracell. Vesicles 1: 18364.
3. Théry, C. et al. (2006) Curr. Protoc. Cell Biol. 3.22.1–3.22.29.
4. Morgan, A.C. et al. (1981) Hybridoma 1: 27–36.
5. Bumol, T.F. and Reisfeld, R.A. (1982) Proc. Natl. Acad. Sci. USA. 79: 1245–1249.
6. Pluschke, G. et al. (1996) Proc. Natl. Acad. Sci. USA 93: 9710–9715.
7. Moldenhauer, G. et al. (1987) Br. J. Cancer 56: 714–721.
Unless otherwise specifically indicated, Miltenyi Biotec products and services are for research use only and not
for therapeutic or diagnostic use. MACS and MACSxpress are registered trademarks of Miltenyi Biotec GmbH.
Copyright © 2015 Miltenyi Biotec GmbH. All rights reserved.