37. Migliaccio E, Giorgio M, Mele S et al. The p66shc adaptor
protein controls oxidative stress response and life span in
mammals. Nature 1999; 402: 309–313
38. Menini S, Iacobini C, Ricci C et al. Ablation of the gene encoding
p66Shc protects mice against AGE-induced glomerulopathy by
preventing oxidant-dependent tissue injury and further AGE
accumulation. Diabetologia 2007; 50: 1997–2007
39. Graiani G, Lagrasta C, Migliaccio E et al. Genetic deletion of the
p66Shc adaptor protein protects from angiotensin II-induced
myocardial damage. Hypertension 2005; 46: 433–440
40. Menini S, Amadio L, Oddi G et al. Deletion of p66Shc longevity
gene protects against experimental diabetic glomerulopathy by
preventing diabetes-induced oxidative stress. Diabetes 2006; 55:
1642–1650
41. Pacini S, Pellegrini M, Migliaccio E et al. p66SHC promotes
apoptosis and antagonizes mitogenic signaling in T cells. Mol
Cell Biol 2004; 24: 1747–1757
42. Spescha RD, Shi Y, Wegener S et al.. Deletion of the ageing gene
p66Shc reduces early stroke size following ischaemia/reperfusion
brain injury. Eur Heart J 2012
43. Ejerblad E, Fored CM, Lindblad P et al. Association between
smoking and chronic renal failure in a nationwide populationbased case-control study. J Am Soc Nephrol 2004; 15: 2178–2185
29. Orsini F, Migliaccio E, Moroni M et al. The life span determinant
p66Shc localizes to mitochondria where it associates with mitochondrial heat shock protein 70 and regulates trans-membrane
potential. J Biol Chem 2004; 279: 25689–25695
30. Hall AM, Unwin RJ. The not so ‘mighty chondrion’: emergence
of renal diseases due to mitochondrial dysfunction. Nephron
Physiol 2007; 105: p1–10
31. Jassem W, Fuggle SV, Rela M et al. The role of mitochondria in
ischemia/reperfusion injury. Transplantation 2002; 73: 493–499
32. Feldkamp T, Kribben A, Weinberg JM. Assessment of mitochondrial membrane potential in proximal tubules after hypoxia-reoxygenation. Am J Physiol Renal Physiol 2005; 288: F1092–F1102
33. Zhou X, Sheng Y, Yang R et al. Nicotine promotes cardiomyocyte
apoptosis via oxidative stress and altered apoptosis-related gene
expression. Cardiology 2010; 115: 243–250
34. Wong LS, Green HM, Feugate JE et al. Effects of ‘second-hand’
smoke on structure and function of fibroblasts, cells that are critical for tissue repair and remodeling. BMC Cell Biol 2004; 5: 13
35. Berger F, Gage FH, Vijayaraghavan S. Nicotinic receptor-induced
apoptotic cell death of hippocampal progenitor cells. J Neurosci
1998; 18: 6871–6881
36. Satta R, Maloku E, Zhubi A et al. Nicotine decreases DNA
methyltransferase 1 expression and glutamic acid decarboxylase
67 promoter methylation in GABAergic interneurons. Proc Natl
Acad Sci USA 2008; 105: 16356–16361
Received for publication: 18.10.2012; Accepted in revised form: 14.11.2012
Recovery of urinary nanovesicles from ultracentrifugation
supernatants
Centre for BioAnalytical Sciences (CBAS), Dublin City University,
Luca Musante,
Dublin, Ireland
Mayank Saraswat,
Alessandra Ravidà,
Barry Byrne
and Harry Holthofer
Keywords: ammonium sulphate precipitation, exosomes, ultracentrifugation, vesicle recovery
Correspondence and offprint requests to: Harry
Holthofer; E-mail: [email protected]
represent pathological events in the kidneys and the urogenital
epithelium. The majority of currently applied isolation protocols involve cumbersome centrifugation steps to enrich vesicles from urine. To date, the efficiency of these approaches
A B S T R AC T
Background. Urinary vesicles represent a newly established
source of biological material, widely considered to faithfully
© The Author 2012. Published by Oxford University Press on
behalf of ERA-EDTA. All rights reserved.
1425
ORIGINAL ARTICLE
Nephrol Dial Transplant (2013) 28: 1425–1433
doi: 10.1093/ndt/gfs564
Advance Access publication 20 December 2012
Tamm–Horsfall glycoprotein (THP), which efficiently entraps
exosome vesicles [16]. This is followed by a high-speed ultracentrifugation (200 000 g) to pellet exosomes and other vesicular structures. Alternative methods which utilized micro- and
nano-ultrafiltration devices have been described [17, 18]
where surfactant-based recovery steps may be required. Hence,
centrifugation is preferentially selected for harvesting small
vesicle species [19]. Post-processing typically involves dithiothreitol [16, 19] and a second ultracentrifugation in the presence or absence of a sucrose gradient in the deuterated water
[6, 13] as an alternative to size exclusion chromatography retention [20].
The aim of this study was to optimize the process of vesicle
isolation and yield and, specifically, to investigate the biophysical properties of previously uncharacterized vesicle populations retained from the supernatant (SN).
ORIGINAL ARTICLE
has not been investigated with respect to performing quantitative and qualitative analyses of vesicle populations in the
pellet and supernatant (SN) fractions.
Methods. After the series of differential centrifugations, the
final SN was reduced to one-twentieth of the original volume
by ammonium sulphate precipitation, with the precipitate
pellet subjected to another round of differential centrifugations. Electron microscopy, dynamic light scattering and
western blot analysis were used to characterize the vesicles
present in individual fractions of interest.
Results. Pellets obtained after the second set of centrifugations
at 200 000 g revealed the presence of vesicles which share a
common marker profile, but with distinct differences from
those seen in the initial 200 000 g pellet used as the reference.
This suggests an enrichment of previously uncharacterized
urinary vesicles still in solution after the initial centrifugation
steps. Analysis of protein yields recovered post-ultracentrifugation revealed an additional 40% of vesicles retained from the
SN. Moreover, these structures showed a formidable resistance
to harsh treatments (e.g. 95% ammonium sulphate saturation,
hypotonic dialysis, 0.3 M sodium hydroxide).
Conclusions. Methods which employ differential centrifugations of native urine are remarkably ineffective and may lose a
substantial population of biologically important vesicle species.
M AT E R I A L S A N D M E T H O D S
Urine samples
Urine samples were collected from the non-smoking
healthy volunteers among the laboratory staff (three female
and three male; aged 20–40 years) in accordance with the protocols approved by the ethical committee of Dublin City University. There was no history of renal dysfunction in any
subjects or drug administration during the sample collection.
First morning urine was processed within 3 h without protease
inhibitors (PIs). The first 50 mL was discarded to avoid artefacts due to the presence of bacteria, cells and cell debris. The
remaining urine was anonymously tested by Combur 10
Test®D dipstick (Roche Diagnostics, Germany) and pooled.
INTRODUCTION
Since the first description of exosome vesicles in the urinary
solid phase [1], recent research has aimed to establish methods
to harvest the complete urinary vesicle repertoire as a source of
valuable disease biomarkers. Accordingly, proteomic profiling
of the urinary solid phase has revealed a surprising complexity
in terms of protein identification [2, 3]. Among urinary vesicles,
exosomes comprise nanovesicles (40–100 nm) released by the
endosomal pathway [4]. During this process, cytosolic proteins
are engulfed into the invaginating membrane which maintains
the same topological orientation as the plasma membrane [5].
When fused with the cellular plasma membrane, the multi-vesicular bodies (MVBs) release vesicles into the extracellular
environment, including related, exosome-like [6] and podocyteshedding vesicles [7–9]. The latter are formed directly from the
outward protrusion of the cell membrane [10] and secreted by a
process termed ‘ectocytosis’ [11].
Urinary vesicles such as those described above are indiscriminately released from all nephron segments [1, 2] and the
urogenital tract [12, 13]. Exosomes typically share common
proteins necessary for their biogenesis, structure and trafficking in addition to the site-specific signature processes of the
cell of origin. These proteins include, e.g. aquaporin 2 (AQP2)
and podocin [1–3]. Furthermore, exosomes contain distinct
transcription factors [14] and specific nucleic acid species encoding proteins native to all nephron regions [15]. This property is of great clinical value and emphasizes the need to
precisely understand their biologic roles.
Currently, the most common procedure for harvesting vesicles from urine involves a series of centrifugations to initially
remove large cellular debris and polymers of the abundant
Vesicle purification
A schematic representation of the methodology used to
isolate vesicles is shown in Figure 1. Pooled urine samples
(≅1200 mL) were centrifuged at a relative centrifugal force
(RCF) of 1000 g for 20 min (without braking) in a swing
bucket rotor benchtop universal 320 centrifuge (Hettich Zentrifugen, Tuttingen, Germany). The resultant SN was carefully
retained and the hyaline pellet was re-suspended in 2-mL purified water per 50-mL tube before storing at −80°C. The retained SN 1000 g was split into aliquots of 300 mL each and
stored at −80°C. The cryo-precipitate, observed after thawing,
was dissolved by gentle mixing at room temperature (RT) [21].
The 1000 g SN was then centrifuged at 18 000 g for 30 min at
RT in a fixed angle rotor (Beckman JA-20; Beckman Coulter,
Fullerton, CA). The retained 18 000 g SN fraction was then ultracentrifuged at 200 000 g for 2 h (RT) using an OptimaTM
L-90K preparative ultracentrifuge (Beckman Coulter) and a
fixed-angle rotor (Beckman 70Ti; Beckman Coulter). All the
pellets were re-suspended in filtrate (0.22 µm) and deionized
water (R ≥ 18.2 MΩ cm). To determine the presence of nanovesicles in the final 200 000 g SN, this was retained and
reduced 20-fold in volume using ammonium sulphate
(NH4)2SO4 precipitation (97%, w/v saturation) [22]. The
(NH4)2SO4 precipitate was recovered by centrifugation at RCF
of 10 000 g in a fix angle rotor (Beckman JA-20) for 1 h at RT.
1426
L. Musante et al.
The (NH4)2SO4 pellet was resuspended in ∼60 mL of deionized water and dialysed in membrane tubes with molecularweight cut-off (MWCO) of 3.5 kDa (Spectrum Laboratories,
Rancho Dominguez, CA) against deionized water to remove
the excess of salt. After three changes of deionized water, 10 L
each, the solution was spontaneously split into two fractions:
water soluble and pellet (water insoluble). To completely dissolve the precipitate of the water-insoluble fraction, 10-N
sodium hydroxide (NaOH) was added drop by drop until the
solution was clear (final NaOH 0.3 N). NaOH fraction was
exhaustively dialysed—MWCO of 3.5 kDa—against deionised
water to remove the excess of salt, and pH was dropped to 7.0.
Both fractions underwent a series of centrifugations at 18 000
and 200 000 g, respectively. The tube loading volume was as in
the first run. H2O and NaOH fractions were processed separately by the same series of centrifugations. These are referred
to as H2O and NaOH 18 000 and 200 000 g, respectively.
transferred to a 0.22-μm nitrocellulose membrane (Whatman,
Springfield, UK) [26]. For western blotting, membranes were
saturated with the Odyssey blocking buffer (LI-COR Biosciences, Lincoln, MA) and incubated with specific antibody:
rabbit anti-CD63 (Santa Cruz Biotechnologies, Santa Cruz,
CA); rabbit anti-tumour suppressor gene (TSG101), mouse
anti-β-actin and rabbit anti-podocin (Sigma Aldrich, Dorset,
UK); mouse anti-CD24 and rabbit anti-Wilms tumor (WT-1;
Abcam, Cambridge, UK) were used according to the manufacturer’s instructions. After six washes in phosphate-buffered
saline (PBS)–Tween (0.1%, v/v), the membranes were incubated with an infrared dye-coupled secondary antibody (LICOR Biosciences; 1:10000 dilution), with images acquired
using the Odyssey Infrared Laser Scanner (LI-COR Biosciences).
Negative transmission electron microscopy
Equivalents of 10 μg of vesicle preparations were fixed in
1% (v/v) glutaraldehyde in water. Fixed vesicle preparations
were spotted onto a Formvar/Carbon 300 mesh grid (Agar
Scientific, Stansted, UK) and dried at RT. The grids were
washed twice in 0.1 M PBS and incubated in 1% (w/v) OsO4
SDS PAGE and western blot
Five micrograms of proteins—Bradford protein assay [23]—
were separated using polyacrylamide gels as described earlier
[24]. Gels were either stained with silver nitrate [25] or
1427
Exosome optimization
ORIGINAL ARTICLE
F I G U R E 1 : Urine vesicle enrichment workflow. After the first series of centrifugations at 18 000 and 200 000 g, the final SN was saturated by
ammonium sulphate. The pellet recovered by centrifugation was dialysed and split into two fractions. The insoluble pellet was solubilized by
adding drop by drop sodium hydroxide 0.3 M final concentration and further dialysed to remove the excess of base and drop the pH to neutral.
Both solution aqueous and sodium hydroxide fractions underwent the series of centrifugations at 18 000 and 200 000 g, respectively, to recover
vesicles retained in the first SN.
in 0.1 M PBS for 30 min on ice. After five 5 min washes (three
with PBS and two with water), exosomes were stained with 5%
(w/v) uranylacetate in water for 10 min [27]. After staining,
vesicle populations were examined using the JEM-2100 transmission electron microscope (TEM; Jeol Ltd, Tokyo, Japan).
Figure 1 and involved an initial centrifugation at an RCF of
1000 g to remove cells and THP macropolymers, which start
to sediment at low centrifugation force as shown in Figure 2
( panel C, lane 7). Vesicle recovery in terms of protein quantification is reported in Table 1. Approximately 85% of the
protein was recovered during the first 18 000 g centrifugation.
For the 200 000 g fraction, 60% of the protein was harvested
while the remaining (40%) was retained in the SN. This was
recovered in the 200 000 g pellet after (NH4)2SO4 precipitation.
TEM analysis of vesicle population isolated by ultracentrifugation post-(NH4)2SO4 precipitation showed a heterogeneous population of nanoparticles whose diameters in
aqueous fractions varied between 30 and 300 nm (Figure 2A).
This size range is consistent with that of exosomes, exosomelike and shed vesicles (Figure 2A). Higher magnification revealed an intact structure in the aqueous (H2O) and sodium
hydroxide (NaOH) phase vesicles (Figure 2B and C). DLS was
selected to further define the size of the most abundant nanovesicle species in individual fractions (Figure 2D and E). Here,
homogeneous populations of vesicles with a mean size of 60,
25 and 4.5 nm were detected in crude (asterisk), H2O (spade)
and NaOH (diamond) pellets. DLS analysis expressed in terms
of intensity (Figure 2D) showed a broader curve shape
Dynamic light scattering
Dynamic light scattering was performed on a Nano-ZS
platform (Malvern Instruments Ltd, UK) equipped with a 4MW He–Ne laser (λ = 632 nm). Prior to analysis, samples
were filtered through 2-µm filters into a 1.0-mL glass cuvette
( pathlength of 1.0 cm). Each analysis comprised 16 runs (10 s
per run) and outputs were presented as a distribution of
hydrodynamic radius (RH), taking each measurement into
account.
R E S U LT S
ORIGINAL ARTICLE
Urine samples
Combur® test analysis revealed that all urine samples were
negative for nitrite, glucose and ketones and no traces of haemoglobin or leucocytes. After pooling, the samples were processed in accordance with the methodology outlined in
F I G U R E 2 : TEM and DLS analysis. TEM showed a broad population of membrane-encapsulated structures with a diameter of 30–300 nm in
the aqueous phase (A, 4000×). Higher magnification of vesicles enriched in aqueous (B, 80 000×) and sodium hydroxide (C, 50 000×) fractions.
DLS verified a predominant vesicle size at 90 nm in the 200 000 g crude pellet. After (NH4)2SO4 precipitation, the predominant hydrodynamic
structures at 60 and 5 nm were recorded in the aqueous and base fractions, respectively (D). DLS intensity analysis revealed a more heterogeneous population of vesicles in all fractions (E).
1428
L. Musante et al.
Table 1. Protein amount recovered in the centrifugation pellets
Fraction
Sediment 1000 g
Protein amount
(mg)
Concentration
(μg/mL)
18.9
15.75
Sediment 18 000 g
2.90
2.42
Sediment 200 000 g
1.75
1.46
H2O sediment 18 000 g
0.18
0.15
H2O sediment 200 000 g
0.66
0.55
NaOH sediment 18 000 g
0.34
0.28
NaOH sediment 200 000 g
0.42
0.35
Recovery
(P18 000 g) (%)
Recovery
(P200 000 g) (%)
84.8
61.8
5.3
23.4
9.9
14.8
ORIGINAL ARTICLE
F I G U R E 3 : SDS-PAGE silver staining. Five micrograms of protein per lane: (A) 200 000 g crude, lane 1; aqueous, lane 2 and sodium hydroxide,
lane 3 SN; (B) 18 000 g crude, lane 4; aqueous, lane 5 and sodium hydroxide pellets, lane 6; (C) 1000 g, lane 7 and 200 000 g crude, lane 8;
aqueous, lane 9 and sodium hydroxide pellets, lane 10.
suggesting a more heterogeneous population of vesicles with a
mean size of ∼70 and 250 nm for the crude (asterisk) pellet;
∼130 and 650 nm for H2O (spade) and finally ∼7 and 20 nm
for NaOH (diamond) pellets. Considering the difference in the
two types of measurements taken, namely the physical size by
TEM and hydrodynamic radii measured by DLS, there was a
good agreement between these two orthogonal techniques.
Polypeptide distribution patterns for all fractions were analysed by electrophoresis, with western blotting selected for verifying the presence of exosomal markers. Figure 3 shows the
silver-stained gel of the SN ( panel A), 18 000 g pellet ( panel
B) and 1000 and 200 000 g pellets ( panel C). The large band
seen at ∼100 kDa is known to represent the THP monomer
under reducing conditions. As shown, the bulk of THP is
present in the 1000 g pellet (lane 7) and in decreasing
amounts in the first 18 000 g (lane 4) and 200 000 g (lane 8)
pellets. A small amount of THP is still visible in the 200 000 g
pellet of the aqueous phase (lane 9) and some traces appear as
a diffuse band in the 18 000 g pellet of the same step (lane 5).
No trace of THP is detectable with the sensitivity of silver
staining in the 18 000 and 200 000 g pellets of NaOH fractions
(lanes 6 and 10). For polypeptide distribution in the 200 000 g
pellets (lane 8, 9 and 10), individual fractions showed strong
similarities. The polypeptide pattern in the 18 000 g fraction
( panel B) is even more homogeneous than the 200 000 g fractions in terms of band distributions. However, the polypeptide
profiles of 18 000 and 200 000 g are not dissimilar, suggesting
a progressive recovery of vesicles at different speeds which is
due not only to the presence of THP which can entrap vesicles
in its polymer complexes [16] but also to a dilution factor.
This effect is highlighted when comparing NaOH 18 000 and
200 000 g pellets (lanes 6 and 10), which do not contain THP
when analysed by silver staining.
The bulk of the TSG101 signal (∼46 kDa) was precipitated
in the first 200 000 g ultracentrifugation (Figure 4). In
addition, two molecular-weight isoforms (∼35 and 20 kDa)
were detected as shown in lane 3. A faint 46-kDa band is detectable in the first 18 000 g pellet (lane 2). A TSG101 signal is
1429
Exosome optimization
analysed. Podocin immunodetection in the first 200 000 g
pellet revealed a ∼50-kDa band which might be a ubiquitinated isoform [32]. The intense ∼42-kDa band corresponds to
the correct molecular weight when engaged with the proteins
of the slit diaphragm. Additional bands were visible at lower
molecular weights (between 30 and 40 kDa), suggesting a proteolytic fragmentation which is not detected in the 200 000 g
pellets after (NH4)2SO4 precipitation. The high-molecularweight signal detected >250 kDa suggests incomplete solubilization and denaturation of this fraction in the Laemmli buffer
which is not strong enough to solubilize the protein complex.
Hence, not well-solubilized podocin could belong to the detergent-resistant domain of the slit diaphragm [33]. The absence
of this pattern in the other pellets suggests a different subcellular origin of these nanovesicles. The same consideration
applies to the distribution of molecular weight detected for
WT-1 antigen with a strong band at 52 kDa detected only in
the very first pellet (1000 g) plus other bands among the WT1 isoforms encoded by its gene [34]. Finally, in order to
exclude a non-specific cross reactivity due to the secondary
antibody, hybridization with only secondary antibody was performed, with no signal observed (λ = 680 and λ = 800).
F I G U R E 4 : Western blot of exosomal marker. Mouse anti-TSG101
ORIGINAL ARTICLE
diluted 0.5 μg/mL (w/v) overnight (ON) at RT. Five micrograms of
protein was added per lane. Lane 1, 1000 g pellet; lane 2, crude pellet
18 000 g; lane 3, 200 000 g crude pellet; lane 4, aqueous 18 000 g
pellet; lane 5, aqueous 200 000 g pellet; lane 6, sodium hydroxide
18 000 g pellet; lane 7, sodium hydroxide 200 000 g pellet; lane 8,
aqueous 200 000 g SN and lane 9, sodium hydroxide 200 000 g SN.
also visible in the aqueous 200 000 g pellet (lane 5) and in the
18 000 and 200 000 g pellets of NaOH fractions (lanes 6 and
7). Finally, bands at 46 and 20 kDa are visible in these three
fractions, and the absence of a 35-kDa band suggests the enrichment of a subpopulation of exosomes positive for TSG101.
Additional western blotting, using the 1000 g pellet as a cellular positive control, was undertaken to confirm the presence of
selected antigens in the 200 000 g pellets, and to provide a better
understanding of how exosome-associated markers are distributed between these fractions (Figure 5). CD63 pattern analysis revealed a common band around 60–65 kDa and a lower
molecular mass isoform (∼55 kDa) in the 200 000 g pellets obtained after (NH4)2SO4 precipitation. CD63 is highly glycosylated and several isoforms have been ascribed to changes in the
glycan moieties [28, 29]. CD24 is a highly O-glycosylated GPI
anchor protein whose molecular weight depends on the grade of
glycosylation according to the tissue of origin [30]. The CD24
signal reveals a sharp intense band at ∼53 kDa, which is double
the molecular weight reported in the data sheet where MCF-7
cells are selected for analysis. This band is present in all 200 000
g pellets, with a faint band also displayed in the 1000 g sediment.
AQP2 shows a particular distribution of isoforms between the
200 000 g pellets. A ∼54-kDa band was present in all the fractions, with a lower molecular weight (∼29 kDa) identified primarily in the first 200 000 g pellet. The high-molecular-weight
mass bands of 36–50 kDa can be attributed to post-translational
modifications such as glycosylation, phosphorylation or ubiquitination which are known mechanisms that regulate the AQP2
activity [31]. Moreover, the detection of actin in the 1000 g
pellet and 200 000 g membrane sediment and its absence in the
200 000 g pellet after ammonium sulphate precipitation highlights the different protein compositions of these vesicle preparations.
Western blotting with two podocyte markers ( podocin and
WT-1) shows a different distribution among the fractions
DISCUSSION
One of the primary goals of modern proteomics is to enrich
the low-abundant proteins of interest within an analytical
sample, e.g. by pre-fractionation. Furthermore, it is well established that secreted vesicles in the urine (exosomes, ectosomes,
apoptotic blebs, etc.) are a precious source of biomarkers
which faithfully mirror the physio- and pathological states of
the urothelium [1, 2, 12, 35]. However, for urinary proteomic
analysis, the isolation of vesicles in the absence of interfering
soluble proteins such as THP and albumin is a key consideration. In spite of the development of numerous centrifugation
and ultrafiltration-based protocols for isolating [17–19] and
processing these entities [36, 37], an optimised method for
routine urinary exosome isolation is yet to be established.
In this study, we wanted to address the efficiency of vesicle
recovery by using differential ultracentrifugation to enrich the
previously uncharacterized vesicles retained from the SN of
the 200 000 g pellet [37] which is routinely identified as a final
analytical sample. Firstly, a centrifugation step at 1000 g is to
remove cells, cell debris, nuclei, bacteria [38, 39] and the abundant THP. Although we cannot exclude the possibility that
vesicles are retained at this point entrapped within THP polymers [16], the introduction of this low-speed centrifugation
step (1000 g) avoided a potential contamination resulting from
intracellular vesicle release.
No protease inhibitors (PIs) were added to urine at the
time of collection. This issue has been widely debated and
different reports have presented conflicting observations.
Mitchell et al. [13] demonstrated that the TSG101 and prostate-specific antigen (PSA) pattern were identical in the presence and absence of PI. In contrast, Zhou et al. [36] showed in
two different sets of urine collections that the addition of PIs
preserved the integrity of the Na–K–Cl co-transporter isoform
1430
L. Musante et al.
where populations with particles sizes of ∼5 nm were detected.
Notably, when this analysis looked at intensity distribution, two
populations (6 and ∼200 nm) were observed with correlating intensities. Both intensity distribution peaks was comparable,
suggesting that the majority of vesicles retained in this fraction
had diameters of ∼6 nm. This is because large particles scatter
much more light than small particles (the intensity of scattering
of a particle is proportional to the sixth power of its diameter
from Rayleigh approximation [40]. The number and intensity
distribution of vesicles in H2O fractions and crude pellets correlate well, and the other peaks present in the intensity distribution
in the micron range could be representative of aggregates. The
number distribution is better in understanding the predominant
particle size as the intensity distribution is biased towards detecting bigger size particles. Thus, caution should be exercised in interpreting DLS results, however, for vesicles but it is possible to
measure hydrodynamic radius, polydispersity and the presence
of aggregates. Human serum albumin (HSA) has a size of ∼6
nm at pH 7.4 measured by DLS [41] and THP has a mean diameter of 120 nm in the absence of NaCl which changes with the
increase of the salt concentration and in the presence of other
proteins [42]. In addition, high-density lipoprotein measures
around 8 nm [43]. All these factors would bias the DLS results
and therefore TEM should be used to estimate the size distribution of vesicles.
Various vesicle populations additionally have characteristic
exosomal surface markers, intra-vesicular proteins (tetraspanin, CD63, TSG101 and β-actin) and membrane nephron
2. These observed differences could be addressed by the different methodologies used herein. The discrepancy may simply
be due to the presence of more non-exosomal contaminants (i.
e. THP). Hence, the ratio between exosomal markers to the
protein load in the gel is lower. From our results the immunodetection of vesicular markers did not reveal a specific degradation pattern. However, some fragmentation can be seen in
TSG101 and podocin in the first 200 000 g pellet. Since these
two antigens reside inside the vesicle, the addition of PIs may
have been ineffective. Furthermore, other markers characterized by an ectodomain (e.g. CD63, CD24, AQP2) do not show
fragmentation.
To facilitate the detection of vesicles residing in the
200 000 g SN, the elective kosmotropic salt (NH4)SO4 was selected for protein precipitation. Notably, a small fraction of the
pellet was found to be insoluble unless the pH was increased
by drop-wise addition of 0.3 M NaOH. However, applying the
same centrifugation series on concentrated samples facilitate
the recovery of vesicles remaining from the first ultracentrifugation SN in both H2O and NaOH fractions. It is interesting
to note that 15.2 and 38.2% of the total protein yield was recovered in 18 000 and 200 000 g pellets, respectively (Table 1)
and that TEM analysis of vesicles revealed electron-dense
membranous structures (Figure 2) with an average diameter of
30–250 nm and shapes consistent with the previous reports [1,
6, 7, 12, 16–18].
DLS was used to determine the most representative size of
the vesicles in the 200 000 g crude, H2O and NaOH fractions,
1431
Exosome optimization
ORIGINAL ARTICLE
F I G U R E 5 : Western blot analysis. Five micrograms of protein per lane: lane 1, 1000 g pellet; lane 2, 200 000 g crude pellet; lane 3, aqueous
200 000 g pellet and lane 4, sodium hydroxide 200 000 g pellet. Rabbit anti-CD63 diluted 1:500; mouse anti-CD24 1 μg/mL; mouse anti-AQP2
1 μg/mL; mouse anti-β-actin 0.5 μg/mL; rabbit anti-podocin 0.5 μg/mL; rabbit anti-Wilms tumor protein antigen (WT-1) 1 μg/mL. All the
primary antibodies were incubated ON at RT. Goat anti-rabbit (λ of 680 nm) and goat anti-mouse (λ of 800 nm) were incubated for 2 h at RT.
ORIGINAL ARTICLE
segment-specific proteins, e.g. AQP2, podocin and WT which
can readily be detected by western blotting. The screening of a
selection of exosomal markers (Figures 2 and 3), particularly
CD63 [44], CD24 [45] and TSG101 [46] reveals a biomarker
distribution consistent with that of exosomes. Furthermore,
cytoskeletal β-actin and cell-type-specific markers including
podocin, WT1 [5, 14, 15] and AQP2 [1, 6, 15] showed their
presence in all the H2O and NaOH ultracentrifugation pellets.
Additionally, the presence of specific bands, e.g. CD63 and the
absence of others present in the first crude pellet, indirectly
suggests the presence of a subpopulation of vesicles which are
not harvested in the conventional set of centrifugations. The
amount of protein recovered in the first 200 000 g pellet was in
agreement with previous studies [36, 45], notwithstanding the
fact that different methodologies were used to determine the
protein concentration (Coomassie/Bicinchoninic acid). It is
also notable that protein recovered in the first crude 18 000 g
pellet was around 85% of the total yield sedimenting at this
RCF, whilst 200 000 g pellet recovery was slightly reduced to
60% (Table 1).
In conclusion, our studies show the importance of method
development for reliably recovering previously uncharacterized urinary vesicle species, as obtained from 200 000 g ultracentrifugation SNs. Achieving a standard protocol for urinary
exosomes is thus crucial to tap the full potential of urinary vesicles as new biomarkers. In fact, urine chemical–physical composition is highly variable and it greatly depends on
environmental factors, especially dietary habits [47]. This
variability influences the density of the urine specimens and
consequently introduces variable buoyancy to the solution.
Thus, when a comparative analysis of different samples is performed, a standardization of the sample density would be
necessary to avoid the aforementioned bias. First, it would be
recommended to introduce a dialysis step with dialysis tubing
of an appropriate MWCO in order to make the chemical–
physical conditions equal for all the urinary solutions under
investigation. As reported in this and other studies [13, 48],
urinary vesicles have shown an excellent resistance to both hypotonic and hypertonic solutions. Therefore, an exhaustive
dialysis against water or buffer at very low ionic strength
should favour and increase the harvest of vesicles following ultracentrifugation. Moreover, the introduction of this dialysis
step confers properties to the urine solution which do not
favour THP association [49] and promote precipitation of
polymers following ultracentrifugation. Secondly, it would be
necessary to normalize the protein concentration, especially in
a comparative analysis of samples characterized by different
grades of proteinuria. This can be done either by diluting or
concentrating the samples to the same final protein concentration. Finally, for all those studies which specifically and exclusively focus on the analysis of urinary vesicles, a dialysis
step utilizing high-MWCO membranes would help in eliminating the interference of soluble proteins, which in turn
would result in the avoidance of repetitive centrifugations. Alternatively, implementing immuno-affinity enrichment in a
solution free of soluble proteins (especially THP) would be a
useful alternative approach.
AC K N O W L E D G E M E N T
The skilful technical assistance of Ms Dorota Tataruch is
kindly acknowledged. This study was supported by HRB grant
number HRA/09/62 and European Union IAPP grant 286386
(UroSense).
C O N F L I C T O F I N T E R E S T S TAT E M E N T
None declared.
(See related article by Jacquillet et al. Urinary vesicles: in
splendid isolation. Nephrol Dial Transplant 2013; 28: 1332–
1335.)
REFERENCES
1. Pisitkun T, Shen RF, Knepper MA. Identification and proteomic
profiling of exosomes in human urine. Proc Natl Acad Sci USA
2004; 101: 13368–13373
2. Gonzales PA, Pisitkun T, Hoffert JD et al. Large-scale proteomics
and phosphoproteomics of urinary exosomes. J Am Soc Nephrol
2009; 20: 363–379
3. Wang Z, Hill S, Luther JM et al. Proteomic analysis of urine exosomes by multidimensional protein identification technology
(MudPIT). Proteomics 2012; 12: 329–338
4. Thery C, Zitvogel L, Amigorena S. Exosomes: composition, biogenesis and function. Nature Rev Immunol 2002; 2: 569–579
5. Babst M. A protein’s final ESCRT. Traffic 2005; 6: 2–9
6. Hogan MC, Manganelli L, Woollard JR et al. Characterization of
PKD protein-positive exosome-like vesicles. J Am Soc Nephrol
2009; 20: 278–288
7. Hara M, Yanagihara T, Hirayama Y et al. Podocyte membrane
vesicles in urine originate from tip vesiculation of podocyte microvilli. Hum Pathol 2010; 41: 1265–1275
8. Pascual M, Steiger G, Sadallah S et al. Identification of membrane
bound CR1 (CD35) in human urine: evidence for its release by
glomerular podocytes. J Exp Med 1994; 179: 889–899
9. Lescuyer P, Pernin A, Hainard A et al. Proteomics analysis of a
podocyte vesicle-enriched fraction from normal human and
pathological urines. Proteomics Clin Appl 2008; 2: 1008–1018
10. Brown DA, Rose JK. Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the
apical cell surface. Cell 1992; 68: 533–544
11. Stein JM, Luzio JP. Ectocytosis caused by sublytic autologous
complement attack on human neutrophils. The sorting of
endogenous plasma-membrane proteins and lipids into shed vesicles. Biochem J 1991; 274: 381–386
12. Welton JL, Khanna S, Giles PJ et al. Proteomics analysis of
bladder cancer exosomes. Mol Cell Proteomics 2010; 9:
1324–1338
13. Mitchell PJ, Welton J, Staffurth J et al. Can urinary exosomes act
as treatment response markers in prostate cancer? J Transl Med
2009; 7: 4–17
1432
L. Musante et al.
Received for publication: 7.9.2012; Accepted in revised form: 15.10.2012
1433
Exosome optimization
ORIGINAL ARTICLE
32. Tossidou I, Teng B, Drobot L et al. CIN85/RukL is a novel
binding partner of nephrin and podocin and mediates slit diaphragm turnover in podocytes. J Biol Chem 2010; 285:
25285–25295
33. Schwarz K, Simons M, Reiser J et al. Podocin, a raft-associated
component of the glomerular slit diaphragm, interacts with
CD2AP and nephrin. J Clin Invest 2001; 108: 1621–1629
34. Hohenstein P, Hastie ND. The many facets of the Wilms’ tumour
gene, WT1. Hum Mol Genet. 2006; 15: R196–R201
35. Moon PG, Lee JE, You S et al. Proteomic analysis of urinary exosomes from patients of early IgA nephropathy and thin basement
membrane nephropathy. Proteomics 2011; 11: 2459–2475
36. Zhou H, Yuen PS, Pisitkun T et al. Collection, storage, reservation, and normalization of human urinary exosomes for biomarker discovery. Kidney Int 2006; 69: 1471–1476
37. Gonzales PA, Zhou H, Pisitkun T et al. Isolation and purification
of exosomes in urine. Methods Mol Biol 2010; 641: 89–99
38. Thongboonkerd V, McLeish KR, Arthur JM et al. Proteomic
analysis of normal human urinary proteins isolated by acetone
precipitation or ultracentrifugation. Kidney Int 2002; 62:
1461–1469
39. Fogazzi GB, Garigali G. The clinical art and science ofurine
microscopy. Curr Opin Nephrol Hypertens 2003; 12: 625–632
40. Hallett FR, Watton J, Krygsman P. Vesicle sizing: number distributions by dynamic light scattering. Biophys J 1991; 59: 357–362
41. Luik AI, Naboka YuN, Mogilevich SE et al. Study of human
serum albumin structure by dynamic light scattering: two types
of reactions under different pH and interaction with physiologically active compounds. Spectrochim Acta A Mol Biomol Spectrosc 1998; 54: 1503–1507
42. Sanders PW, Booker BB, Bishop JB et al. Mechanisms of intranephronal proteinaceous cast formation by low molecular weight
proteins. J Clin Invest 1990; 85: 570–576
43. Sakurai T, Trirongjitmoah S, Nishibata Y et al. Measurement of
lipoprotein particle sizes using dynamic light scattering. Ann
Clin Biochem 2010; 47: 476–481
44. Escola JM, Kleijmeer MJ, Stoorvogel W et al. Selective enrichment of tetraspan proteins on the internal vesicles of multivesicular endosomes and on exosomes secreted by human Blymphocytes. J Biol Chem 1998; 273: 20121–20127.
45. Keller S, Rupp C, Stoeck A et al. CD24 is a marker of exosomes
secreted into urine and amniotic fluid. Kidney Int 2007; 72:
1095–1102
46. Babst M, Odorizzi G, Estepa EJ et al. Mammalian tumor susceptibility gene 101 (TSG101) and the yeast homologue, Vps23p, both
function in late endosomal trafficking. Traffic 2000; 1: 248–258
47. Pradella M, Dorizzi RM, Rigolin F. Relative density of urine:
methods and clinical significance. Crit Rev Clin Lab Sci 1988; 26:
195–242
48. Musante L, Saraswat M, Duriez E et al. Biochemical and physical
characterisation of urinary nanovesicles following CHAPS treatment. PLoS One 2012; 7: e37279
49. McQueen EG, Engel EG. Factors determining the aggregation of
urinary mucoprotein. J Clin Pathol 1966; 19: 392–396
14. Zhou H, Cheruvanky A, Hu X et al. Urinary exosomal transcription factors, a new class of biomarkers for renal disease. Kidney
Int 2008; 74: 613–621
15. Miranda KC, Bond DT, McKee M et al. Nucleic acids within
urinary exosomes/microvesicles are potential biomarkers for
renal disease. Kidney Int 2010; 78: 191–199
16. Fernández-Llama P, Khositseth S, Gonzales PA et al. Tamm
Horsfal protein and urinary exosomes isolation. Kidney Int 2010;
77: 736–742
17. Merchant ML, Powell DW, Wilkey DW et al. Microfiltration isolation of human urinary exosomes for characterization by MS.
Proteomics Clin Appl 2010; 4: 84–96
18. Cheruvanky A, Zhou H, Pisitkun T et al. Rapid isolation of urinary
exosomal biomarkers using a nanomembrane ultrafiltration concentrator. Am J Physiol Renal Physiol 2007; 292: 1657–1661
19. Alvarez ML, Khosroheidari M, Kanchi Ravi R et al. Comparison
of protein, microRNA, and mRNA yields using different
methods of urinary exosome isolation for the discovery of kidney
disease biomarkers. Kidney Int 2012; 82: 1024–1032
20. Rood IL, Deegens JK, Merchant ML et al. Comparison of three
methods for isolation of urinary microvesicles to identify biomarkers of nephrotic syndrome. Kidney Int 2010; 78: 810–816
21. Mataija-Botelho D, Murphy P, Pinto DM et al. A qualitative proteome investigation of the sediment portion of human urine:
implications in the biomarker discovery process. Proteomics Clin
Appl 2009; 3: 95–105
22. Wood WI. Tables for the preparation of ammonium sulfate solutions. Anal Biochem 1976; 73: 250–257
23. Bradford MM. A rapid and sensitive method for the quantitation
of microgram quantities of protein utilizing the principle of
protein-dye binding. Analyt Biochem 1976; 72: 248–254
24. Laemmli UK. Cleavage of structural proteins during the assembly
of the head of bacteriophage T4. Nature 1970; 227: 680–685
25. Shevchenko A, Wilm M, Vorm O et al. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal
Chem 1996; 68: 850–858
26. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure
and some applications. Proc Natl Acad Sci USA 1979; 76:
4350–4354
27. Mathias RA, Lim JW, Ji H. Proteomics Analysis of Membrane
Proteins: Methods and Protocols. Totowa, NJ: Humana Press,
2009, pp. 227–242
28. Agerberg M, Lindmark A. Characterisation of the biosynthesis
and processing of the neutrophil granule membrane protein
CD63 in myeloid cells. Clin Lab Haem 2003; 25: 297–306
29. Engering A, Kuhn L, Fluitsma D et al. Differential post-translational modification of CD63 molecules during maturation of
human dendritic cells. Eur J Biochem 2003; 270: 2412–2420
30. Poncet C, Frances V, Gristina R et al. CD24 a glycosylphosphatidylinositol-anchored molecules is transiently expressed during
the development of human central nervous system and is a
marker of human neural cell lineage tumors. Acta Neuropathol
1996; 9: 400–408
31. Moeller HB, Olesen ET, Fenton RA. Regulation of the water
channel aquaporin-2 by posttranslational modification. Am J
Physiol Renal Physiol 2011; 300: F1062–F1073