Journal of Immunological Methods 371 (2011) 143–151

Contents lists available at ScienceDirect

Journal of Immunological Methods

j ourna l homepage: www.e lsev ie r.com/ locate / j im

Research paper

A filtration-based protocol to isolate human Plasma Membrane-derivedVesicles and exosomes from blood plasma☆

Ryan Grant a, Ephraim Ansa-Addo a,1, Dan Stratton a,1, Samuel Antwi-Baffour a, Samireh Jorfi a,Sharad Kholia a, Lizelle Krige b, Sigrun Lange c, Jameel Inal a,⁎a Cellular and Molecular Immunology Research Centre, Faculty of Life Sciences, London Metropolitan University, London, UKb Hammersmith Medicines Research, Cumberland Avenue, London, UKc University College London Institute for Women's Health, Maternal and Fetal Medicine, Perinatal Brain Repair Group, UK

a r t i c l e i n f o

☆ This work was funded by the Cellular and MResearch Centre/Faculty of Life Sciences and HamResearch. SK received a VC scholarship (London Metr⁎ Corresponding author at: Cellular and Molecular

Centre, School ofHumanSciences, Faculty of Life ScienceUniversity, London, N7 8DB, UK. Tel: +44 20 7133 214149.

E-mail address: j.inal@londonmet.ac.uk (J. Inal).1 EAA and DS contributed equally to this work.

0022-1759/$ – see front matter © 2011 Elsevier B.V.doi:10.1016/j.jim.2011.06.024

a b s t r a c t

Article history:Received 10 May 2011Received in revised form 21 June 2011Accepted 22 June 2011Available online 30 June 2011

The methods of Plasma Membrane-derived Vesicle (PMV) isolation and quantification varyconsiderably in the literature and a new standard needs to be defined. This study describes anovel filtration method to isolate PMVs in plasma, which avoids high speed centrifugation, andto quantify them using a Becton Dickinson(BD) FACS Calibur™ flow cytometer, as annexin V-positive vesicles, larger than 0.2 μm in diameter. Essentially microvesicles (which comprise amixture of PMVs and exosomes) from citrate plasma were sonicated to break up clumpedexosomes, and filtered using Millipore 0.1 μm pore size Hydrophilic Durapore membranes inSwinnex 13 mm filter holders. Phosphatidylserine-positive PMVs detected with annexin V-PEwere quantified using combined labelling and gating strategies in conjunction withPolysciences Polybead Microspheres (0.2 μm) and BDTrucount tubes. The PMV absolutecount was calculated on the analysis template using the Trucount tube lot number informationand expressed in PMV count/ml. Having estimated a normal reference range (0.51×105–

2.82×105 PMVs/ml) from a small sample of human donors, using the developed method, theeffect of certain variables was investigated. Variations such as freezing of samples and genderstatus did not significantly alter the PMV absolute count, and with age plasma PMV levels wereonly marginally reduced. Smokers appeared to have reduced PMV levels. Nicotine, as forcalpeptin was shown to dose-dependently (from 10 up to 50 μM) reduce levels of earlyapoptosis in THP-1 monocytes and to decrease the level of PMV release. Fasting individuals had2–3 fold higher PMV absolute counts compared to non-fasting subjects.

© 2011 Elsevier B.V. All rights reserved.

Keywords:Plasma Membrane-derived VesiclesExosomesPlasma PMV

1. Introduction

PlasmaMembrane-derivedVesicles (PMVs), (also knownasmicroparticles (MPs)) are small, cell-derived vesicles ~0.2–

olecular Immunologymersmith Medicinesopolitan University).Immunology Researchs, LondonMetropolitan22; fax: +44 20 7133

All rights reserved.

2 μm in diameter (Piccin et al., 2007), released from a variety ofcells. They have been identified and partially characterisedaccording to Clusters of Differentiation (CD) markers andantibody staining techniques, and this way it has been possibleto distinguish PMVs from different cell types, such as plateletsand endothelial cells.

PMVs are mainly produced through activation of the cellmembrane or apoptosis (Théry et al., 2009), but have beenobserved in both healthy and diseased human populations andanimalmodels. PMVrelease is alwaysprecededby an increase inintracellular calcium ([Ca2+]i), activation of calpain and depo-lymerisation of the actin cytoskeleton. During the process ofPMV release, summarised in Fig. 1(A), normal cell phospholipid

PMVs

early endosome

exosomes

multivesicularbody

A

B

C

D

actin calpain

phosphatidylserine

tetraspanin, eg. CD63

Fig. 1. Schematic of a typical eukaryotic cell generating microvesicles(Plasma Membrane-derived Vesicles [PMVs] and exosomes). (A) Earlyendosomes result from endocytosis and may either return to the surface orform multivesicular bodies (MVBs) by internal budding from the endosomesurface (B). The vesicles contained within MVBs are released as exosomesupon fusion of the MVB with the plasma membrane (C). Plasma Membrane-derived Vesicles (PMVs) are released (D) from the cell surface upon cellularactivation or early apoptosis, resulting in calpain-mediated cleavage of thecytoskeleton, loss of membrane asymmetry and release of PMVs. These PMVsare larger than exosomes (0.1–1 μm versus 0.1 μm in diameter, respectively)and bear more exposed phosphatidylserine. Exosomes typically carrytetraspanin family proteins (for example CD63).

144 R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

(PL) asymmetry is lost and this leads to increased exposure ofphosphatidylserine (PS+) on the outer leaflet of the PMVmembrane. This phenomenon is detected through bindingassays with annexin V, a widely accepted method of identifyingand quantifying PMVs by flow cytometry, although, as reportedby Inal and others, not fully standardised (Hind et al., 2010). Bycontrast exosomes are formed from internal budding ofendosomes resulting in multivesicular bodies (MVB). Whenthese fuse with the plasma membrane they release theirexosomal content from the cell as summarised in Fig. 1(B–D).PMVs appear to circulate in nearly all body fluids includingserum, blood and urine, and have been documented as dwellingin lymph nodes. PMVs have also been observed by electronmicroscopy “budding” off from nearly all cell lines, (Azevedo etal., 2007; Ansa-Addo et al., 2010).

It is well known that cells communicate and exchangeinformation amongst themselves through variousmechanisms.Some of these include the use of secreted growth factors,cytokines, chemokines and small molecular mediators such asnucleotides, nitric oxide ions and bioactive lipids. Othermechanisms implemented, include the use of specialised

adhesion molecules for cell-to-cell adhesion (Majka et al.,2001; Vandendries et al., 2004). However, several investigatorsare now focusingoncell-to-cell communicationbystudying theinvolvement of PMVs. These have been reported in many casesto contain numerous proteins and lipids similar to thosepresent on the plasma membranes of parent cells. For severalyears this mechanism has been overlooked as a possiblepathway by which cells may interact. Underlying the associa-tion of PMVswith disease appears to be amajor role PMVs playin cell-to-cell communication (Morel et al., 2004), transferringorganelles, or cellular components such as proteins, receptors,bioactive molecules, micro RNA (miRNA) or mRNA (Ratajczaket al., 2006; Bajzyworzeka et al., 2006). Many reports suggestthat cells release PMVs in order to communicate with othercells through transfer of receptors, or simply to initiatesignalling or cell contact. Recent reports also suggest roles inhomeostasis (Nagai et al., 2005) and thrombosis (Matzdorff etal., 1998), angiogenesis (Mostefai et al., 2008), tumourmetastasis (Baj-Krzyworzeka et al., 2007), and stem cellengraftment (Majka et al., 2007). Other immunological rolesinclude platelet PMV-mediated IgG production dependent onplatelet CD154 (Sprague et al., 2008) or PMV-mediateddifferentiation of human myeloid blasts (THP-1) (Ansa-Addoet al., 2010).

Release of PMVs can also be a defensive mechanism forhost cells in response to stress, inflammation and tissueregeneration. It is well documented that cells shed PMVsenriched in death signals, including, Fas and caspase-1, whichare deposited on the plasma membrane. Hence, they avoidpassing the point of no return in terms of apoptosis. PMVsenriched in these signals can then deliver death messages tonearby cells (Sarkar et al., 2009; Liu et al., 2009).

As well as avoiding apoptosis, some cells release PMVs as amechanism to escape complement-mediated lysis. In thepresence of sublytic complement, release of PMVs helps cellsby removing the inserted, membrane attack complex (MAC)from the cell surface (Morgan and Campbell, 1985). Variousreviews suggest that cells release PS+-PMVs to prevent PS-induced phagocytosis of the cell as indeed we found to be thecase with the phagocytosis of apoptotic cells, which wasinhibited by PMVs (Antwi-Baffour et al., 2010).

There is a diverse range of clinical conditions in whichPMVs have been characterised and quantified. Disease statesin which they are found at elevated levels include infectiousdisease, for example Cerebral Malaria (Coltel et al., 2006) orcardiovascular disease such as stroke, aortic aneurism andvenous thrombo-embolism (Azevedo et al., 2007), SystemicLupus Erythematosus and Rheumatoid Arthritis (Ardoin et al.,2007; Antwi-Baffour et al., 2010), as well as DiabetesMellitus,and chronic renal failure (Piccin et al., 2007). PMVs have alsobeen implicated in HIV/AIDS pathogenesis (Mack et al., 2000),and cancer (Goon et al., 2006). As increased PMV levels couldbe a result of the up-regulation of cellular responses requiredby the body during these various disease states they couldthus serve as sensitive indicators of disease activity, or incytotoxic drug monitoring.

The aim of this study was to quantify all phosphatidylser-ine (PS)-positive (labelled with annexin V-PE) plasma PMVs,without the need of specialist equipment such as anultracentrifuge. This would allow standardisation of PMVanalysis in terms of apparatus and methodology.

145R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

2. Materials and methods

2.1. Sample preparation

Venous blood (1.8 ml) was drawn slowly and gently fromhealthy volunteers (n=57) into a syringe, transferred into1.8 ml vacutubes containing sodium citrate (0.109 M, finalconcentration), and processed within 15 min. Spun andseparated samples were sometimes frozen at −20 °C, untilanalysis. Samples were never repeatedly freeze/thawed. Thisstudy was approved by the relevant local ethics committees.

2.2. Filtration method of PMV isolation

Within 1 h of being collected, samples were labelled with alab reference number (LRN) and spun at 160 g for 5 min andtwice at 4000 g for 30 min. All the aliquots and transfer tubesfor the same sample were labelled with the same LRN. Thisenabled correct tracking, storage and retrieval of sampleswhenrequired. The plasma samples (0.5 ml) were collected into alabelled microfuge tube for reproducibility/sample stabilitycomparison and frozen immediately at −20 °C.

The remaining 0.5 ml of plasma was then sonicated for5×1 min (Townsen and Mercer, Ltd., Croydon) before filteringusing a Swinnex 13 mm filter holder with a 0.1 μm pore sizeMillipore Hydrophilic Durapore filter placed inside. Eachsample was passed through a filter holder using a disposable2.5 ml BD Plastipack syringe inserted at the female inlet. ThePMVs (size 0.1–2 μm in diameter) were deposited on the filtermembrane and the filtrate containing exosomes was eithercollected for further analysis or discarded.

The filter membrane was then carefully removed and placedin 3 ml PBS in a 15 ml centrifuge tube. Sonication of the conicaltubes for 15 min at RT, was followed by 1 min gentle vortexing,ensuring the vibrating liquid remained in contact with the filter,to dislodge all PMVs from the filter membrane. The filter mem-brane was then removed and discarded from the tube. Cen-trifugation at 25,000 g for 60 min was used to pellet the PMVs.

2.3. Stimulation of plasma cells with calcium ionophore

Calcium Ionophore, A23187 (10 μl 50 mM Calcium Iono-phore in 250 μl Hank's Buffered Salt Solution (HBSS) and 240 μlDMSO, light protected) was freshly prepared. To stimulate cells,36 μl of calcium ionophore (A23187) was added to each 1.8 mlcitrate tube (final concentration, 20 μM). The citrate tubes wereincubated in a 37 °C incubator for 1 h and filtered as describedabove.

2.4. Annexin V/staining of plasma PMVs

For analysis of microvesicles (exosomes or PMVs) theyneeded to be centrifuged (160,000 g/1 h for exosomes or25,000 g/1 h for PMVs) and resuspended in 100 μl AnV BindingBuffer, AnVBB, (Abcam, plc) (prepared in situ from 10x BindingBuffer concentrate) resulting in a 5-fold concentration of theoriginal 0.5 ml plasma sample. The re-suspended PMV pellet(100 μl) was then transferred from the centrifuge tube to aTrucount tubewhere itwas stainedwith5 μl AnnexinV-PE. Thereactionmixturewas incubated for 15 min at RT in the dark and395 μl of 1x AnVBB (prepared in situ from 10x Binding Buffer

concentrate) added to each tube. The results were analysed byflow cytometry within 1 h. Annexin V-PE positive stainingparticles are quantified in the FL-2 channel. Mouse anti-CD63staining was similarly carried out and both stainings repeatedon isolated exosomes.

2.5. Calibration/Quality control

A 3-colour calibration was performed daily, using BDCaliBRITE 3 beads (SLE017). The calibration reports of the‘lyse-wash’ and the ‘lyse-nowash’ assayswerefiled in the FACSCalibur™ Calibration datafile. Polystyrene beads (0.2 μm)wererun as a size control before each batch of samples.

2.6. Analysis of data acquired using the BD FACS Calibur™ flowcytometer

The data was analysed after acquisition was complete. Forthis the appropriate analysis template was opened from theFacsdrive1/Analysis Templates folder. The light scatter and FL-1were set at logarithmic gains, and 10,000 total events werecollected during each sample analysis. These events wereprocessed by an Apple computer, in conjunctionwith CellquestPro Software (BD Biosciences UK). After identifying PMVs, theirabsolute counts were calculated automatically on the analysisspreadsheet after inputting the following formula provided byBD Biosciences UK:

PMV count =No: of events in region containing PMVs

No: of events in absolute count bead region

×No: of beads per test �

Test volume mLð Þ

*this value is found on the TruCount Absolute Count Tube foilpouch label, and might vary from lot to lot.

2.7. Transmission electron microscopy

Isolated PMVs or exosomeswere examined by transmissionelectronmicroscopy as described previously, by Inal and others(Ansa-Addo et al., 2010). Essentially samples were negativelystained with equal parts 2% phosphotungstic acid (pH 6.8) andbacitracin (100 g/ml). They were then placed on 400-meshcopper grids coveredwith Pioloform (Agar Scientific), renderedhydrophilic by pre-treatment for 10 min with 1% Alcian blue8GX. Sampleswere examinedusing a JEOL JEM-1200 EXMark IIelectron microscope and images recorded using an integratedAMT (Advanced Microscopy Techniques) camera.

2.8. ELISA for TGF-β

This was carried out as described previously (Ansa-Addo etal., 2010). In summary, isolated PMVs (30 μg) were lysed with1% (v/v) Triton-X-100, applied to the TGF-β ELISA kit (R&DSystems) and the concentration of TGF-β was determinedaccording to the manufacturer's instructions.

2.9. Flow cytometry for detecting CD63 and AnV binding onPMVs and exosomes

PMVs and exosomes were stained with annexin V-FITC(R&D Systems, UK) for PS expression and anti-CD63-PE

146 R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

antibody and analysed as described earlier using the GuavaEasyCyte flow cytometer (Ansa-Addo et al., 2010).

2.10. In vitro inhibition of PMV release from THP-1 cells

THP-1 cells were washed twice with serum-free (SF) RPMIand allowed to equilibrate by incubating at 37 °C for 30 min.Cells were pelleted by centrifuging at 200 g for 5 min andimmediately resuspended in prewarmed SF-RPMI supple-mented with 0.5 mM CaCl2. Suspended cells were seeded into24-well plates and nicotine or calpeptin was added to thewells at concentrations ranging from 10 to 100 μM, andpreincubated at 37 °C for 30 min.

NHS (5%) was immediately added to the wells withoutwashing away the chemicals and further incubated at 37 °Cfor 30 min with shaking. The reaction was stopped after30 min by placing the plates on ice for 1 min. Samples weretransferred into microfuge tubes and centrifuged at 200 g for5 min to pellet cells. The resultant supernatant was trans-ferred into fresh microfuge tubes and centrifuged at 4000 gfor 1 h. The supernatant was then collected into fresh tubesand sonicated in a water sonicator for 5 min, before beingcentrifuged at 25,000 g for 2 h. Pelleted PMVs were collected,resuspended in PBS and analysed by flow cytometry.

2.11. Statistical analyses

Method precision was determined by calculating thepercentage coefficient of variance (% CV). Box plots wereused to indicate PMV absolute count, spread amongst thehuman population, and the variables looked at. Regressionanalysis for the age variable, and t-tests were performed totest the significance of the differences in the numbers of PMVsusing SPSS Version 17.0. Statistical significance was consid-ered when P-values were less than 0.05.

3. Results

3.1. A novel filtration-based PMV isolation protocol

PMVs are typically isolated using a differential centrifugation-based protocol. The aim of the filtration protocol designed forPMV isolation (outlined in Fig. 2) was to avoid differentialcentrifugation and thereby limit contamination with smallerexosomes. Venous blood (1.8 ml)was collected in a vacutubewith 0.109 M sodium citrate. Blood plasma (0.5 ml) recov-ered after centrifugation (2125 g for 15 min) was then watersonicated for 5×1 min. This was an important stage as it hadbeen reported that exosomes tend to clump into clusters(Heijnen et al., 1999; Zakharova et al., 2007) and so co-purifywith PMVs. The ‘microvesicles,’ comprising PMVs andexosomes were then filtered through a 0.1 μm pore sizefilter. Exosomes (average diameter of ≤100 nm) wererecovered from the filtrate and set aside for later analysis.PMVs were then recovered from the filter by sonication(5 min in a sonicating water bath and vortexing for 1 min).Finally PMVs were recovered by centrifugation (25,000 g for60 min) although they could also be recovered at lowerspeeds (10,000 g for 60 min or 10,000 g for 30 min) (Fig. 3).To confirm isolation of the two microvesicle populations,filtered PMVs and filtrate (containing exosomes) were

analysed by electron microscopy and for certain markersby ELISA or flow cytometry (Fig. 4). The PMVs isolatedranged in diameter from ~0.2 μm to 1 μm (Fig. 4A), thesmaller exosomes consistently around 100 nm in diameterand showing the typical ‘cup-shaped’ morphology (Fig. 4B).PMVs had higher TGFβ concentrations (Fig. 4C) and boundmore annexin V (Fig. 4D) than exosomes, but exosomesexpressed more CD63 (Fig. 4E).

3.2. Flow cytometry output (dot plots and histograms — analysistemplates)

To ensure the flow cytometry was operational each day,and before a batch of samples was run, calcium ionophore(A23187 (20 μM) for 1 hwas used to stimulate quality control(QC) plasma, to increase the general number of plasma PMVs.This was to guarantee enough sample to gate the flowcytometer and produce quantifiable results. Calibration beads(0.2 μm) were also run every day before samples could beoverlaid with the patient sample, both gated for AnV(G2) andwith no gate; the Trucount bead population was also included(not shown). The Kolmogorov Smirnov statistics and overlaidhistograms were also used to validate that the AnV positivePMV samples being looked at on the different wavelengths(FL1 and FL2) were the same population (pb0.001). Areadout from unstimulated blood showing annexin V-positive PMVs, of diameter ≥0.2 μm is shown in the dashedbox region of the dot plot in Fig. 5A. Using the formula inSection 2.6 (where the bead count for Trucount Tubes was50595.00 [Lot no. 2731]), the absolute PMV count wascalculated in this case to be 168473.86/ml.

3.3. PMV isolation protocol to enumerate total plasma PMVlevels

The PMV isolation method we developed was then used toenumerate plasma PMVs, essentially to help establish a normalbaseline reference range in ahealthyhumanpopulation, usingatotal of 57 subjects, both male and female, ranging in age from19 to 71 years old.

3.4. The effect of freezing on PMV levels measured

Contrary to current literature, which indicates PMV levelsto become elevated after freeze/thawing (Piccin et al., 2007),we compared PMV levels from freshly prepared sampleswith those from the same samples that had been frozenat−20 °C, and observed no significant difference (p=0.226)(Fig. 6A).

It should be noted that the samples frozen at−20 °C werethawed only once, approximately 30 min prior to stainingand analysis. This suggests that frozen PMV samples, whichare therefore as reliable as those freshly prepared, can be usedwith confidence in routine laboratory analysis. It also over-comes the reproducibility obstacle of not being able tovalidate PMV counts unless freshly prepared.

3.5. Gender and plasma PMV levels

Mean plasma PMV levels from 26 males and 31 femaleswere compared (Fig. 6B). The box plot shows that the female

0.1 µm pore size filter

0.5 ml plasma or cell-free supernatant

sonicating water bath / 5 x 1 min

PMV recovery:Sonication /15 minVortexing /5 min

Exosomes (filtrate)

Annexin V staining/ Flow cytometryPMV count using TruCount beads

1.8 ml venous blood (~5x106 cells/ml) in vacutainer (0.109 M sodium citrate)or cell culture supernatant (e.g. THP-1 cells 1x106 cells/ml)

160 g / 5 min ; 4000 g / 30 min X2

0.5 ml plasma or cell-free supernatant

sonicating water bath / 5 x 1 min

25,000 g / 60 minand washed once in PBS

Annexin V staining/ Flow cytometryPMV count using TruCount beads

A B

Fig. 2. Flow chart summarising filtration-based (A) and centrifugation-based (B) PMV isolation protocols. (A) Venous blood is collected in a vacutube. After lowspeed centrifugation 0.5 ml blood plasma is sonicated (5×1 min) in a sonicating water bath and passed through a 0.1 μmpore size filter. PMVs are recovered fromthe filter by water sonication (15 min) and vortexing (5 min) and centrifugation (25,000 g/60 min). Isolated PMVs are identified by annexin V binding using a FacsCalibur™ flow cytometer and counted using TruCount beads. Using a centrifugation protocol for PMV isolation from venous blood or stimulated cultured cell lines(B), samples are and centrifuged to remove cell debris (160 g/5 min and 4000 g/30 min x2). To disrupt exosome clumps the supernatant is sonicated for 5×1 minin a sonicating water bath and then centrifuged at 25,000 g to collect PMVs. After washing, the PMVs are similarly identified as phosphatidylserine-positive (byannexin V labelling) and counted using a flow cytometer.

10,000 g 30 min 10,000 g 60 min 25,000 g 60 minA CB

0.5µm0.9µm

3.0µm 3.0µm 3.0µm

0.5µm0.9µm

0.5µm0.9µm

FSC-HLog

SS

C-H

Log

FSC-HLog

SS

C-H

Log

SS

C-H

Log

FSC-HLog

Fig. 3. Recovery of PMVs by centrifugation. PMVs are typically isolated, after removal of cell debris, by centrifugation at 25,000 g for 1 h. Analysis of microvesiclesrecovered at 25,000 g for 60 min and from centrifugation at lower speeds, 10,000 g for 60 min and 10,000 g for 30 min shows that they are annexin V positivelystaining vesicles (lower right inset in panels A, B and C) and that they show a typical forward/side scatter dot plot (upper left inset) for PMVs and that they range insize from 0.5 μm to 0.9 μm in diameter.

147R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

Exo PMVs0

10

20

30

40

50

60

70*

% C

D63

Exo PMVs0.0

0.5

1.0

1.5**

TG

F-β

1 (n

g/m

l)

Exo PMVs0

10

20

30

40

50

60 *

% A

nV

1.2 µm

PMV

100 nm

ExosomesA B

C D E

Fig. 4. Detection of PMVs and exosomes. PMVs recovered after filtration from the membrane were prepared by negative staining and visualised using atransmission electron microscope (A). Exosomes collected in the filtrate were similarly examined and as expected found to average 100 nm in diameter and tohave a typical ‘cup-shaped morphology,’ (B). PMVs were confirmed as carrying higher concentrations of TGF-β1 (C) as measured on PMVs lysed with 1% (v/v)triton-X-100, by ELISA, and to have higher levels of phosphatidylserine exposition (D), identified by annexin V-FITC binding using flow cytometry. CD63expression (E) was higher on exosomes than PMVs as measured on the Guava EasyCyte flow cytometer.

148 R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

normal baseline reference range can be defined as 0.55×105–1.8×105 PMVs/ml, whilst themale range ismarginally smaller,0.65×105–1.35×105 PMVs/ml, the lowest quartile starting ata slightly higher level with males. Females had a highermedian (1.25×105) PMVs/ml in comparison to males(0.95×105) PMVs/ml, but overall there was no significantdifference in plasma PMV levels due to gender (p=0.160).

3.6. Age and plasma PMV levels measured

We had expected that the older subjects in the samplemeasured would have higher levels of PMVs due to anincreased general amount of cell death, apoptosis and associ-ated PMV production, but this was not observed (Fig. 6C). Theslightly elevated level of plasma PMVs in younger subjectsobserved may be attributable to the higher cell turnover ormetabolism, in younger subjects. By contrast elderly individualsmay have reduced cell turnover or waste elimination. In eitherage groupelevatedplasmaPMV levelsmay thenbe indicative ofdisease.

3.7. The effect of fasting status on plasma PMV levels

The number of subjects that were comparable for fasting(defined ashavinghadnothing to eat or drink, exceptwater, for12 h) was limited. Only 4 subjects from the whole populationwere analysed both fasting and non fasting, that didn't smoke.This was mainly due to the practical inconvenience caused tovolunteers, and the other variables used in this investigation.However within a paid clinical trial setting this should beobtainable. Despite the small population size (n=4), there is a

statistical significance as shown by the independent sample t-test p=0.005 (equal variances not assumed) (Fig. 6D).

The mechanism as to why absolute PMVs/ml appearelevated when the volunteers fasted may be attributed toother compounds present in the plasma, for example in the nonfasting state, e.g. lipaemic samples (that frequently interferewith common laboratory tests) which were not noted, butshould for future purposes. Food or other related contaminantsmay increase sample turbidity, falsely reducing or maskingPMVs, or potentially partially blocking or inhibiting AnVlabelling of the PMVs, due to increased viscosity of the plasma.Our findings of increased plasma PMV levels in the fasting stateprompt further investigation and could potentially set animportant new standard as to how PMVs are measured, just asother fasting analytes, such as glucose, are measured in aroutine laboratory.

3.8. Smoking and plasma PMV levels

Plasma PMV levels were substantially diminished insmokers compared to non-smokers (Fig. 6E). Although theremay bemore thanone component contributing to an inhibitionof PMV release, nicotine, like calpeptin is known to inhibitapoptosis and therefore seemed a likely candidate (Wright etal., 1993 and Lee et al., 2008) Experiments were thereforecarried out to show whether increasing concentrations ofnicotine would result in decreasing levels of early apoptosis inTHP-1 monocytes, measured by annexin V binding as well asdecreasing levels of PMV release, as would be expected usingcalpeptin. Indeed, as for calpeptin (Fig. 6F), nicotine (Fig. 6G)showed a dose-dependent decrease of early (AnV+, 7AAD−)

Absolute PMV count (before activation) = 1913/1149 x 50595/0.5 = 168474 PMVs/ml

A

No. of events in region containing PMVs X No. of beads per test * = PMV count

No. of events in absolute count bead region Test volume (mL)

B

Region

>0.2um CalibrationAnnexin V PE +Trucount Beads

Events

863419131149

100.0022.1613.31

% Gated

Fig. 5. Flow cytometric analysis of isolated PMVs. Peripheral blood PMVs isolated from non stimulated blood, using the filtrationmethod, were analysed on the FacsCalibur™ flow cytometer. Using Trucount beads, those PMVs of diameter greater than 0.2 μm and staining positive for annexin V-PE binding, shown in the dashedbox on the scatter plot (A) were enumerated and an absolute PMV count calculated (B) using the formula in materials and methods (Section 2.6).

149R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

but not late (AnV+, 7AAD+) apoptosis as well as a decrease ofPMV release, as the cells were treated with increasingconcentrations of nicotine, up to ~50 μM.

4. Discussion

The filtration method developed has been used to quantifyphosphatidylserine (PS-positive) plasma PMVs, ≥0.2 μm indiameter, by staining with Annexin V-PE. In the population ofhealthydonors studied (n=57)plasmaPMV levelswere foundto range between 0.51 and 2.82×105 PMVs/ml. These figures(expressed as 1 PMV released from 32 blood cells, consideringthere to be ~5×106 cells/ml in venous blood) are comparablebut understandably slightly lower than those released fromcells in culture (1 PMV released from 8 carcinoma cells, on thebasis of 1×106 cells having been stimulated (Ansa-Addo et al.,2010). It should be noted that this study also confirmed thatcarcinoma cell lines, THP-1 monocytes in particular, releasetwice as many PMVs as primary, peripheral blood monocytes.Most of the variables looked at in this study including freezingat−20 °C, gender and subject age, did not alter PMV absolutecount significantly. Fasting individuals had a wider spread andappeared to have higher PMV levels (2.8×105–5.8×105)PMVs/ml. This could be defined as the normal baseline fastingreference range, which is almost 3-fold higher compared to thenon-fasting group reference range (0.9–1.5×105) PMVs/ml.These results suggest that when analysing total PMVs, further

investigations should use fasting subjects, as for quantificationof other fasting analytes. Plasma PMV levels were alsocompared using the described method between non-smokersand smokers and found to be significantly lower in the latter. Inexperiments on THP-1 monocytes, positive for the nicotinereceptor (Morgan et al., 2001) it was found that increasingdoses of nicotine up to 50 μM, just as for calpeptin, anotherknown inhibitor of apoptosis, resulted in decreasing levels ofcells stainedwith annexinV (early apoptosis) and indecreasinglevels of PMVs released.

In this study we developed a filtration method for theisolation ofmicrovesicles,with the advantage over thepreviousdifferential centrifugation-based method of being able toprocess larger volumes, especially of cell culture supernatant.The instrument settings and methodology have been validatedon a BD Facs Calibur™ Flow Cytometer, showing that PMVanalysis can be implemented routinely and as a potentially highthroughput test (up to 10 batch runs per day). We are nowcontinuing to validate PMV counting on different flowcytometers. It would also be desirable to compare identicalsamples for plasma PMV levels on 2 ormore of the samemodelof flow cytometer for standardising PMV analysis. Additionalresearch could also look at quantifying lymph node dwellingPMVs, as this study only focuses on circulating plasma PMVs.Future improvements under investigation include the use of CDmarkers to differentiate PMVs derived from different cell typessuch as CD61 to identify/isolate platelet-derived PMVs. These

Fresh Frozen0

100000

200000P

MV

s/m

l

Male Female0

100000

200000

300000

PM

Vs/

ml

fasting non fasting0

100000

200000

300000

400000

500000

600000

PM

Vs/

ml

non-smoker smoker0

50000

100000

150000

200000

250000

300000

350000

PM

Vs/

ml

10 20 30 40 50 60 70 800

100000

200000

300000

Age (years)

PM

Vs/

ml

r2=0.012

A B

C D

E***

0

25

50

75

100

125

150

0

20000

40000

60000

80000

100000

1200007-AADPMVs/ml

0 50 1002510

Annexin V+

Nicotine (µM)

Cel

ls (

%) P

MV

s/ml

F

G

0

25

50

75

100

125

150

0

25000

50000

75000

100000

125000

1500007-AAD+

PMVs/ml

Annexin V+

0 25 50 10010

Calpeptin (µM)

Cel

ls (

%) P

MV

s/ml

**

ns ns

Fig. 6. Effect of freezing, gender, age, fasting and smoking status on plasma PMV levels. PMVs isolated from plasma were enumerated as described. Comparing thesame plasma PMV samples, freeze/thawing had no effect on the measureable PMV levels (A) (n=38) and there was a negligible difference between males andfemales (B) (n=26 male; n=31 female). Although most samples were obtained from donors in the 20–30 age range, there was no effect of age on PMV levels(C) (n=57). Although only four samples could be compared between the same donors after fasting or non-fasting, PMVs levels were higher after fasting (D).Smokers had reduced PMV levels (E) and increasing doses of nicotine (F) or calpeptin (G), reaching saturation at 50 μM, dose-dependently decreased annexin Vbinding, as a marker of early apoptosis, as well as the number of PMVs released. These latter experiments with calpeptin and nicotine were repeated twice intriplicate.

150 R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

151R. Grant et al. / Journal of Immunological Methods 371 (2011) 143–151

studies could also include an assessment of sample stability byassaying the same samples 1, 2, and 3 h post draw.

Acknowledgements

We are grateful to Maria McCrossan for assisting with thetransmission electron microscopy. We are also indebted tomembers of CMIRC for critically reading the manuscript andfor providing constructive comments.

References

Ansa-Addo, E.A., Lange, S., Stratton, D., Antwi-Baffour, S., Cestari, I., Ramirez,M.I., McCrossan, M.V., Inal, J.M., 2010. Human plasma membrane-derived vesicles halt proliferation and induce differentiation of THP-1acute monocytic leukemia cells. J. Immunol. 185, 5236.

Antwi-Baffour, S.S., Kholia, S., Aryee, Y.K.-D., Ansa-Addo, E.A., Stratton, D.,Lange, S., Inal, J.M., 2010. Human plasma membrane-derived vesiclesinhibit the phagocytosis of apoptotic cells — possible role in SLE.Biochem. Biophys. Res. Commun. 398, 278.

Baj-Krzyworzeka, M., Szatanek, R., Weglarczyk, K., Baran, J., Urbanowicz, B.,Brański, P., Ratajczak, M.Z., Zembala, M., 2006. Tumour-derived micro-vesicles carry several surface determinants and mRNA of tumour cellsand transfer some of these determinants to monocytes. Cancer Immunol.Immunother. 55, 808.

Baj-Krzyworzeka, M., Szatanek, R., Weglarczyk, K., Baran, J., Zembala, M.,2007. Tumour-derived microvesicles modulate biological activity ofhuman monocytes. Immunol. Lett. 113, 76.

Ardoin, S.P., Shanahan, J.C., Pisetsky, D.S., 2007. The role of microparticles ininflammation and thrombosis. Scand. J. Immunol. 66, 159.

Azevedo, L.C., Pedro, M.A., Laurindo, F.R., 2007. Circulating microparticles astherapeutic targets in cardiovascular diseases. Recent Pat. Cardiovasc.Drug Discov. 2, 41.

Coltel, N., Combes, V., Wassmer, S.C., Chimini, G., Grau, G.E., 2006. Cellvesiculation and immunopathology: implications in cerebral malaria.Microbes Infect. 8, 2305.

Goon, P.K.Y., Lip, G.Y.H., Boos, C.J., Stonelake, P.S., Blann, A.D., 2006. Circulatingendothelial cells, endothelial progenitor cells, and endothelial microparti-cles in cancer. Neoplasia 8, 79.

Heijnen, H.F., Schiel, A.E., Fijnheer, R., Geuze, H.J., Sixma, J.J., 1999. Activatedplatelets release two types of membrane vesicles: microvesicles bysurface shedding and exosomes derived from exocytosis of multi-vesicular bodies and a-granules. Blood 94, 3791.

Hind, E., Heugh, S., Ansa-Addo, E.A., Antwi-Baffour, S., Lange, S., Inal, J.M.,2010. Red cell PMVs, plasma membrane-derived vesicles calling out forstandards. Biochem. Biophys. Res. Commun. 399, 465.

Lee, S.-Y., Lee, M.-S., Cherla, R.P., Tesh, V.L., 2008. Shiga toxin 1 inducesapoptosis through the endoplasmic reticulum stress response in humanmonocytic cells. Cell. Microbiol. 10, 770.

Liu, R., Klich, I., Ratajczak, J., Ratajczak, M.Z., Zuba-Surma, E.K., 2009.Erythrocyte-derived microvesicles may transfer phosphatidylserine tothe surface of nucleated cells and falsely 'mark' them as apoptotic. Eur.J. Haematol. 83, 220.

Mack, M., Kleinschmidt, A., Bruhl, H., Klier, C., Nelson, P.J., Cihak, J., Plachy, J.,Stangassinger, M., Erfle, V., Schlondorff, D., 2000. Transfer of thechemokine receptor CCR5 between cells by membrane derived micro-particles: a mechanism for cellular human immunodeficiency virus 1infection. Nat. Med. 6, 769.

Majka, M., Kijowski, J., Lesko, E., Gózdizk, J., Zupanska, B., Ratajczak, M.Z.,2007. Evidence that platelet-derivedmicrovesicles may transfer platelet-specific immunoreactive antigens to the surface of endothelial cells andCD34+ hematopoietic stem/progenitor cells—implication for the path-ogenesis of immune thrombocytopenias. Folia Histochem. Cytobiol. 45,27.

Majka, M., Janowska-Wieczorek, A., Ratajczak, J., Ehrenman, K., Pietrzkowski,Z., Kowalska, M.A., Gewirtz, A.M., Emerson, S.G., Ratajczak, M.Z., 2001.Numerous growth factors, cytokines, and chemokines are secreted byhuman CD34(+) cells, myeloblasts, erythroblasts, and megakaryoblastsand regulate normal hematopoiesis in an autocrine/paracrine manner.Blood 97, 3075.

Matzdorff, A.C., Berchner, D., Kuhnel, G., Kemkes-Matthes, B., Pralle, H., Voss,R., 1998. Relative and absolute changes of activated platelets, micropar-ticles and platelet aggregates after activation in vitro. Haemostasis 28,277.

Morel, O., Toti, F., Hugel, B., Freyssinet, J.-M., 2004. Cellular microparticles: adisseminated storage pool of bioactive vascular effectors. Curr. Opin.Hematol. 11, 156.

Morgan, B.P., Campbell, A.K., 1985. The recovery of human polymorphonu-clear leucocytes from sublytic complement attack is mediated bychanges in intracellular free calcium. Biochem. J. 231, 205.

Morgan, D., Parsons, M.E., Whelan, C.J., 2001. Investigation of nicotinebinding to THP-1 cells: evidence for a non-cholinergic binding site.Biochem Pharmacology 61, 733.

Mostefai, H.A., Andriantsitohaina, R., Martinez, M.C., 2008. Plasmamembranemicroparticles in angiogenesis: role in ischemic diseases and in cancer.Physiol. Res. 57, 311.

Nagai, A., Sato, T., Akimoto, N., Ito, A., Sumida, M., 2005. Isolation andidentification of histone H3 protein enriched in microvesicles secretedfrom cultured sebocytes. Endocrinology 146, 2593.

Piccin, A., Murphy, W.G., Smith, O.P., 2007. Circulating microparticles:pathophysiology and clinical implications. Blood Rev. 21, 157.

Ratajczak, J., Wysoczynski, M., Hayek, F., Janowska-Wieczorek, A., Ratajczak,M.Z., 2006. Membrane-derived microvesicles: important and underap-preciated mediators of cell-to-cell communication. Leukemia 20, 1487.

Sarkar, A., Mitra, S., Mehta, S., Raices, R., Wewers, M.D., 2009. Monocytederived microvesicles deliver a cell death message via encapsulatedcaspase-1. PLoS One 4, e7140.

Sprague, D.L., Bennett, D.E., Crist, S.C., Waldschmidt, T.J., Jensen, R.J., Ratliff, T.L.,2008. Platelet-mediated modulation of adaptive immunity: uniquedelivery of CD154 signal by platelet-derived membrane vesicles. Blood111, 5028.

Théry, C., Ostrowski, M., Segura, E., 2009. Membrane vesicles as conveyors ofimmune responses. Nat. Rev. Immunol. 9, 581.

Vandendries, E.R., Furie, B.C., Furie, B., 2004. Role of P-selectin and PSGL-1 incoagulation and thrombosis. Thromb. Haemost. 92, 459.

Wright, S.C., Zhong, J., Zheng, H., Larrick, J.W., 1993. Nicotine inhibition ofapoptosis suggests a role in tumor promotion. FASEB J. 7, 1045.

Zakharova, L., Svetlova, M., Fomina, A.F., 2007. T cell exosomes inducecholesterol accumulation in human monocytes via phosphatidylserinereceptor. J. Cell. Physiol. 212, 174.