proposition 1.2

Caractérisation des microparticules, des mitochondries libres et des mito-MPs produites par des cellules d’adénocarcinome rénal et de leur effet sur la polarisation des macrophages

Sous la supervision de: Luc Boudreau

Alain Simard Sandra Turcotte

Guillaume Pelletier

1. INTRODUCTION Microvésicules de RCC et immunomodulation

3

Objectifs généraux

1.  Caractériser les microparticules produites par les cellules d’adénocarcinome rénal (RCC)

2.  Caractériser l’impact des différentes populations de microparticules de RCC sur la polarisation des macrophages

4

Cancer du rein Introduction

5

RCC

Renal Cell Carcinoma ou adénocarcinome rénal •  Cancer du rein le plus fréquent (~90%) •  Causé par une mutation du gène VHL

6

RCC

25 à 30 % des patients diagnostiqués avec le RCC présentent déjà des métastases. Le taux de survie à 5 ans pour le mRCC est inférieur à 10 %.

7

n engl j med

353;23

www.nejm.org december

8

,

2005

The

new england journal

of

medicine

2478

the cases of sporadic clear-cell renal-cell carcino-ma,

8

which represents a major portion of all casesof renal-cell carcinoma.

VHL protein, the product of the

VHL

gene, func-tions as a tumor suppressor, inhibiting growthwhen reintroduced into cultures of renal-cell carci-noma.

9,10

Hypoxia-inducible genes are normally in-hibited by VHL protein,

11

including several encod-ing proteins involved in angiogenesis (e.g., vascularendothelial growth factor [VEGF]), cell growth(e.g., transforming growth factor

a

[TGF-

a

]), glu-cose uptake (e.g., the GLUT-1 glucose transporter),and acid–base balance (e.g., carbonic anhydrase IX[CA9]). When VHL protein is lost, these proteins areoverexpressed, creating a microenvironment favor-able for epithelial-cell proliferation (Fig. 4A). Thus,cells deficient in VHL protein behave as if they are hy-poxic, even in conditions of normoxia. VHL protein,with elongin proteins C and B, binds cul2 protein (amember of the cullin family of ubiquitin ligase pro-teins), indicating that some VHL protein serves asthe receptor subunit of a ubiquitin ligase complex

that promotes the ubiquitination and destructionof proteins (Fig. 4B).

12,13

VHL protein binds thetranscriptional activators hypoxia-inducible factor1

a

(HIF-1

a

) and 2

a

(HIF-2

a

) directly and destabiliz-es them.

14

Furthermore, VHL protein promotes theubiquitination and destruction of HIF-

a

.

15-17

TheseVHL-regulated pathways are being studied as po-tential targets of therapies for clear-cell renal-cellcarcinoma.

HIF is the key regulator of the hypoxic responsein multicellular organisms. Thus, VHL protein hasa central role in oxygen sensing. For HIF-

a

to bindVHL protein, a proline residue must undergo hy-droxylation, which is an unusual protein modifi-cation

18,19

(Fig. 4B). A family of proline hydroxy-lases operates on HIF-

a

in a graded fashion, so thatthe extent of hydroxylation depends on oxygen ten-sion.

20,21

Hydroxylation of an asparagine residueblocks the interaction of HIF-

a

with the transcrip-tional coactivator p300.

22

Thus, multiple hydroxyl-ation steps cooperate to inhibit HIF-

a

activity.To correlate the genotype with the disease phe-

Figure 1. Staging Overview and Five-Year Survival Rates for Renal Cancer.

Survival data

3

are based on the 1997 tumor–node–metastasis (TNM) staging guidelines.

4

More recent renal-cancer staging is described elsewhere.

5

The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF OTTAWA on October 18, 2015. For personal use only. No other uses without permission.

Copyright © 2005 Massachusetts Medical Society. All rights reserved.

8

n engl j med 353;23 www.nejm.org december 8, 2005

medical progress

2483



9

n engl j med 353;23 www.nejm.org december 8, 2005

medical progress

2483



10

Lignée cellulaire 786-O

•  Cellules de RCC immortalisées •  Mutées au gène du VHL •  Vendues par ATCC

11

786-O, 10 décembre 2015

Cancer et système immunitaire Introduction

12

Rôle des macrophages

doi:1

0.10

38/n

ri307

3

13

Polarisation des macrophages

14

Rôle des macrophages dans l’inflammation

15doi:10.1038/ni.1937

Vésicules extracellulaires Introduction

16

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY VOLUME 11 | JUNE 2014 | 351

Key points

■ Microvesicles (MVs) are 0.1–1.0 μm vesicles containing lipids, proteins, RNAs and microRNAs; they are formed by budding from the cellular plasma membrane

■ Circulating levels of several subpopulations of MVs are increased in patients with liver diseases, probably due to enhanced MV production and decreased MV clearance, related to the individual liver disorder

■ MVs are now implicated at many stages of liver disease progression, including liver fibrogenesis, portal hypertension and activation of coagulation

■ Several results suggest that MVs have a role in hepatocellular carcinoma by conveying information between tumour cells and between tumour and neighbouring cells

■ High levels of circulating procoagulant MVs have been found in patients with acute liver failure and might contribute to normal or hypercoagulable global haemostasis in this setting

■ MVs have promise as diagnostic and prognostic biomarkers in patients with liver diseases

account for these changes. The potential roles of MVs in key processes of liver diseases, such as fibrosis, portal hypertension, complications of cirrhosis, thrombosis and hepatocellular carcinoma, are also presented. If specific studies in the context of liver diseases are lacking, we postulate on the potential effects of MVs on the basis of available data from other organs. Finally, we interpret available results and propose clinical situations in which MVs could be useful as biomarkers.

Increased MV levels in liver diseasesMVs have been detected in numerous human body fluids, including saliva,8 urine,9 bile,10 synovial fluid,11 vitreous fluid12 and semen,13 as well as in muscles,14 atherosclerotic plaques7 and liver tissue15 (Figure 2a). Never theless, most studies have focused on plasma MVs, because they are easily accessible. Circulating levels of MVs are increased in patients with cardiovascular disor-ders, thrombosis and cancer.16–18 Over the past ~10 years, several groups have measured the levels of circulating MVs in patients with liver diseases (Table 1).19–30 The increased levels of several subpopulations of circu-lating MVs reported in these studies are a result of increased formation and/or decreased clearance of MVs (Figure 2b; Box 2).

Enhanced MV formationSeveral causes of liver disease trigger MV production, in particular alcohol consumption,31 viral infection32 and features of the metabolic syndrome including dia-betes,33 obesity,34 dyslipidaemia35 and physical inactiv-ity.36 Liver disease itself might also induce MV release, as the main processes of MV formation (namely, apop-tosis and cell activation) are common in this context.37,38 Furthermore, many stimuli that promote MV release, including oxidative stress,39 shear stress,40,41 systemic inflammation and bacterial translocation, are present in liver diseases (Table 2).10,39,42–49

Decrease in MV clearanceThe mechanisms of MV clearance have been reviewed elsewhere.5 Under healthy conditions, spleen and liver macrophages are the primary contributors to MV clear-ance from the circulation.5,50,51 Interestingly, several lines of evidence show that cirrhosis is associated with a defect in macrophage function. Indeed, the function of the macrophage Fc gamma receptor is impaired in patients with alcoholic cirrhosis and is correlated with the degree of liver insufficiency.52 Similarly, macrophage dysfunction has been observed in animal models of cir-rhosis53 and the clearance of certain molecules (radio-labelled colloid or microaggregated human serum albumin) has been shown to be decreased in patients with cirrhosis.54,55 As such, we speculate that clear-ance of MVs might also be decreased in these patients. During pathological conditions that are associated with elevated levels of circulating MVs, such as endotoxaemia, another pathway for MV clearance is ‘turned on’ in the liver and lung endothelium.56 As cirrhosis is associated

Cells

Extracellularvesicles

Examplesof similarsize

MVB

PS

1 2

Microvesicle

0.1–1 μm

Bacteria

Exosome

40–100 nm

Virus

Apoptotic body

1–4 μm

Platelet

Figure 1 | Definitions of microvesicles. Left: Exosomes. Cells release exosomes via two mechanisms. In the classic pathway (1), intracellular vesicles appear from inward membrane budding and form MVBs before fusing with the PM and being released into the extracellular space. The direct pathway (2) involves the release of vesicles, indistinguishable from exosomes, directly from the PM without involvement of MVBs. Exosomes expose little or no PS at their surface, but harbour relatively specific markers. Middle: Microvesicles. Under normal conditions, PS is primarily located on the PM inner leaflet. After exposure to a stimulus, PS is exposed at the cell surface and cytoskeleton reorganization occurs leading to outward blebbing of the PM and release of MVs into extracellular space. However, formation of MVs might happen without externalization of PS as some MVs might be PS negative. Right: Apoptotic bodies. The early stages of apoptosis are characterized by changes in mitochondrial membrane potential and PM asymmetry leading to exposure of PS at the cell surface, but without increased cell permeability. Later stages of apoptosis are characterized by DNA fragmentation and increased PM permeability. Apoptotic cells shrink and can break up into smaller apoptotic bodies. Like MVs, apoptotic bodies harbour PS and contain various materials from mother cells (PM markers, proteins, RNA and microRNA). Unlike MVs, they contain nuclear fragments. Abbreviations: MV, microvesicle; MVB, multivesicular body; PM, plasma membrane; PS, phosphatidylserine.

REVIEWS

© 2014 Macmillan Publishers Limited. All rights reserved

doi:1

0.10

38/n

rgas

tro.2

014.

7

Vésicules extracellulaires

17

Taille d’une cellule 8 – 12 μm

Nomenclature Vésicules extracellulaires (EV) •  Corps apoptotiques •  Microvésicules •  Exosomes •  … etc.

18

Nomenclature

Microparticules (MP) = Microvésicules (MV) = Ectosomes •  Diamètre de 80 - 1300 nm •  Bourgeonnement externe de la membrane cellulaire Exosomes •  Diamètre de 30 - 120 nm •  Exocytose des corps multivésiculaires

19

Nomenclature freeMitos: mitochondries (fonctionnelles) retrouvées dans le milieu extracellulaire. mitoMPs: microvésicules qui contiennent des mitochondries fonctionnelles.

20

La nomenclature est pourrie, c’est une toute nouvelle science.

Introduction

21

Nombre de publications par année

0

500

1000

1500

2000

2500

3000

Ectosomes Microparticles Microvesicles Exosomes

22

Et alors? Pourquoi s’y intéresser? Introduction

23

Nature Reviews | Immunology

Donor cell Acceptor cell

mRNAs and small non-coding RNAs (miRNAs)

MVBmRNAs

Extracellular vesicle

Endosome

Fusion with limitingmembrane of endocytic vesicle?

Fusion with cell membrane?

Translation into proteins

Endocytosis

miRNAs

Post-transcriptionalregulation oftarget mRNAs

APCs78, or between APCs and other cell types (FIG. 4). Indeed, antigen-induced assembly of the immuno-logical synapse triggers the polarization of T cell MVBs to the point of APC–T cell contact, and promotes the transfer of exosomal miR-335, which is functional in miRNA reporter assays in acceptor APCs77. EBV-infected B cells transfer exosome-derived viral miRNAs to DCs, which silences mRNA transcripts that encode immune-stimulatory molecules80.

Activation of immunity by extracellular vesiclesIncreasing evidence suggests that extracellular vesi-cles transfer not only antigens to APCs but also sig-nals that may promote the activation of the acceptor cells into immunogenic APCs. Mast cell-derived extracellular vesicles, which contain a relatively high content of heat shock protein 60 (HSP60) and HSPA8 (also known as HSC70), promote DC maturation in mice58. Macrophages that have been infected with Mycobacterium tuberculosis, Salmonella enterica subsp. enterica serovar Typhimurium or T. gondii release extracellular vesicles that carry microbial antigens, as well as pathogen-associated molecular patterns that promote a Toll-like receptor-dependent inflammatory response by macrophages81. This adjuvant-intrinsic effect, together with the vesicular nature of extracel-lular vesicles, may explain why extracellular vesicles are more efficient than soluble peptides at transferring antigens between APCs.

Interleukin-1β (IL-1β), which is a cytokine that lacks the leader sequence needed for secretion by the classi-cal pathway, is released inside the extracellular vesicles secreted by DCs and macrophages82–84. How IL-1β is then released from the vesicle lumen remains unknown.

Unlike IL-1β, other cytokines are transported on the extracellular vesicle surface. Tumour necrosis factor (TNF) superfamily members, including CD95 ligand (CD95L; also known as FASL), TNF-related apopto-sis-inducing ligand (TRAIL; also known as TNFSF10) and CD40 ligand (CD40L; also known as CD154), are sorted into the exosome membrane. CTLs, natural killer (NK) cells and DCs kill target cells through the polar-ized release of CD95L-carrying extracellular vesicles85–88. Tumour cells and parenchymal cells of immune-privi-leged tissues also secrete CD95L via extracellular vesi-cles as a mechanism of immune escape89,90. In addition, mast cells release CD40L-bearing extracellular vesicles91 and platelets secrete vesicles that deliver CD40L–CD40 signalling92. The release of TNF superfamily molecules through extracellular vesicles decreases the degrada-tion of these molecules by surface proteases, augments their local concentration in the extracellular milieu and favours their aggregation into trimers, thereby increasing their biological activity87. DC-derived and macrophage-derived exosomes carry enzymes that syn-thesize leukotriene B4 (LTB4) and LTC4 (REF. 93). At sites of inflammation, unstable LTA4 that is released by neutrophils could be converted into pro-inflammatory LTB4 and LTC4 by APC-derived extracellular vesicles.

In addition, extracellular vesicles seem to have a role in mediating inflammatory and autoimmune dis-eases; for example, extracellular vesicles in the serum and that are derived from the synovial fibroblasts of patients with rheumatoid arthritis have higher levels of a membrane-bound form of TNF — which is a key target in the treatment of rheumatoid arthritis — than extracellular vesicles from healthy control individuals94. Interestingly, these TNF-positive extracellular vesicles

Figure 4 | Mechanism of transfer of exosomal shuttle RNAs between cells. mRNAs and small non-coding RNAs, including microRNAs (miRNAs), are transported inside the lumen of secreted extracellular vesicles. Once released, the extracellular vesicles are trapped by acceptor cells. Release of the vesicular RNAs into the cytosol of the acceptor cell requires fusion of the vesicle membrane with the plasma membrane or probably with the limiting membrane of endocytic vesicles, after the extracellular vesicles have been internalized. MVB, multivesicular body.

REVIEWS

202 | MARCH 2014 | VOLUME 14 www.nature.com/reviews/immunol

© 2014 Macmillan Publishers Limited. All rights reserved

doi:10.1038/nri3622

Transfert horizontal d’information

24

Information: ARN non codants Lipides membranaires Protéines membranaires Récepteurs membranaires Autres métabolites Contenu du cytosol Petites organelles Pathogènes intracellulaires?

Rôle des EVs dans l’inflammation

25

doi:1

0.10

38/n

rrheu

m.2

014.

19

Transfert de PAMPs de pathogènes et de DAMPs du soi

Rôle des EVs dans l’inflammation

26

Autoantigènes de surface Cytokines phlogistiques

doi:10.1038/nrrheum.2014.19

27

doi:1

0.11

82/b

lood

-201

4-05

-573

543

Les freeMitos sont un substrat pour la sPLA2-IIA et sont très phlogistiques.

Rôle des freeMitos dans inflammation

28

Rôle des mitoMPs dans inflammation

Les neutrophiles, qui jouent aussi un rôle crucial dans l’inflammation, ont une durée de vie limitée à 6-8 h dans la circulation. L’internalisation de mitoMPs par les neutrophiles est un mécanisme proposé pour l’augmentation de leur viabilité dans des pathologies inflammatoires telles que la polyarthrite rhumatoïde.

Rôle des EVs dans le cancer

29

Les métastases les plus fréquentes sont dans les ganglions lymphatiques, les poumons, les os, le foie et le cerveau. Tumour exosome integrins determine organotropic metastasis Nature, 19 novembre 2015#(Hoshino et al.)

Bref. On s’en va où avec ça? Introduction

30

Objectifs généraux

1.  Caractériser les microparticules produites par les cellules d’adénocarcinome rénal (RCC)

2.  Caractériser l’impact des différentes populations de microparticules de RCC sur la polarisation des macrophages

31

Méthodes

•  Culture cellulaire (expériences in vitro) •  Microscopie à fluorescence (mise au point des marquages) •  Microscopie confocale (visualisation) •  FACS (caractérisation et triage) •  NanoSight (caractérisation et validation) •  TEM (visualisation) •  Analyses bioinformatiques (open source) •  ELISA (validation)

32

2. PRODUCTION DE MICROPARTICULES Microvésicules de RCC et immunomodulation

33

Production de microparticulesPas le temps de niaiser quand on veut faire la mise au point!

34

On peut utiliser du A23187 ou la thapsigargine pour augmenter le Ca2+ intracellulaire et provoquer la formation de microparticules.

Ca2+ intracellulaire et vésiculation rapide

PMID

: 902

8933

35

•  Augmenter le Ca2+ intracellulaire change complètement l’état des cellules.#

•  Les MPs produites sont probablement différentes des MPs produites normalement.#

•  On peut le faire pour faire la mise au point.

Limites expérimentales

36

3. DÉTECTION DES MICROVÉSICULES Microvésicules de RCC et immunomodulation

37

Détection des MPs: le challenge Les microvésicules et les exosomes sont très petits.

38

Le pipeline FACS

39

Pipeline d’analyses

40

Acquisitiondes données

Traitement de données à l’aide de logiciels commerciaux Publication


41


Traitement de données à l’aide de logiciels commerciaux Publication


42


Traitement de données à l’aide de logiciels open-source Publication


43




44




45



Acquisition des données FACS

46

Le MoFlo™ XDP

47

Principe de fonctionnement

48

Principe de fonctionnement

Hydrodynamic focusing Les particules en suspension sont attirées une-à-une avec le sheath fluid. Le laser ne frappe donc qu’une particule à la fois.

49

Considérations particulières pour les MPs

Problème: passage de plusieures microparticules simultanément à travers le laser Stratégie: réduire la pression de l’échantillon (réduit le diamètre du central core) Stratégie: diluer l’échantillon

50


Problème: les particules peuvent passer à travers le laser à différents endroits Stratégie: réduire la pression de l’échantillon (réduit le diamètre du central core)

51


Problème: les particules très petites donnent un signal faible Stratégie: réduire au minimum la pression du sheath fluid pour •  augmenter le temps de

transit •  augmenter la probabilité de

dispersion de la lumière •  augmenter le temps

d’émission de fluorescence

52

Traitement des données FACS

53

R est une plateforme d’analyse statistique

54

R est un language facile et un framework puissant!

CRAN7829 packages

Bioconductor1104 packages

On peut facilement: •  développer de nouveaux outils •  améliorer les outils existants

55

R est une plateforme d’analyse statistique

Analyse bioinformatique (FACS) Les données brutes obtenues du MoFlo™ XDP sont conformes au standard FCS 3.1 et sont sauvegardées en plaintext. Les analyses bioinformatiques sont faites avec R et python au besoin. Exemples de librairies utilisées •  flowCore: manipuler des données de cytométrie en flux •  flowQ: contrôle-qualité •  flowStats: analyse statistique •  flowViz et ggplot2: visualisation des données

56

Analyse bioinformatique (FACS) Où c’est approprié, les valeurs linéaires sont transformées par la fonction logicle plutôt que d’utiliser des échelles logarithmiques.

57

subset that is visible in the ‘logicle’ displays (centered near zero on both axes) seems to be missing from the logarithmic displays. However, as indicated above, this subset is present but is represented mainly by data points and contours that are ‘piled up’ on the plot axes.

The location of the median fluorescence value in each dimension (Fig. 1, dark red crosses) further demonstrates the problem with logarithmic data visualization. By defi-nition, half of the data values in each rectan-gular region are greater than the median and half are less than that value. However, whereas the locations of the median values in the ‘log-icle’ displays (Fig. 1, right) correspond to the visual centers of the subsets, the locations of the median values in the logarithmic displays (Fig. 1, left) are substantially offset from the apparent peak of the subset. In fact, each of the subsets in the logarithmic display is broken up into a ‘false peak’ above the median value, a sparsely populated region between this peak and the baseline, and a ‘pileup’ of a large frac-tion of the cells on the baseline.

This artificial subdivision is also visible when data are plotted as a one-dimensional histogram on a logarithmic scale (Fig. 1, bot-tom left). The sharp rise at the lowest value on the scale reflects the ‘pileup’ of the minimally fluorescent cells at the lowest value on the scale. This ‘peak’ tends to be overlooked because it is nearly coincident with the axis. However, it represents roughly 40% of the minimally fluorescent population and 20% of the cells in the overall population. The ‘valley’ and the ‘false peak’ are also in the histogram, creating a total of three peaks rather than the two that actually exist. This problem does not occur in histograms plotted on ‘logicle’ axes (Fig. 1, bottom right), in which the minimally fluo-rescent cells form a peak centered at or near zero and extending symmetrically above and below the peak center, thereby representing the cells whose measured fluorescence values are below zero.

In two-dimensional logarithmic displays, cells whose fluorescence measurements are at or below zero in one or both dimensions are ‘piled up’ on the axes and are represented by contours along both axes in contour displays (Fig. 1, middle left) and by colored dots, vis-ible with keen eyes, on the axes in color density displays (Fig. 1, top left). The contours (and dots) range along a large proportion of both the x and y axes, indicating that cells with fluo-rescence values at or below zero in one dimen-sion may have equivalently low fluorescence in the other dimension or may have very high fluorescence values in this second dimension. Surprisingly, given how modest the contours along the axes seem, they represent about 40%

of the cells in the sample. Notably, these ‘miss-ing’ cells are readily visible in ‘logicle’ displays, which enable visualization of fluorescent mea-surements for essentially all cells in the sample, including those whose fluorescence values are at or below zero in either dimension (Fig. 1, right).

The importance of seeing less than nothingAs FACS instruments measure cell-associated fluorescence that can be essentially zero but not negative, how can a FACS data set contain values below zero and why are some cell popu-lations found to be distributed symmetrically around zero? The answer lies in the way the FACS instrument collects data and routinely corrects (compensates for) the data for spec-

tral overlap before visualization in histograms, dot plots and contour maps. In essence, back-ground light and electronic noise make a small but important contribution to the overall sig-nals the FACS instrument obtains for each cell. The FACS apparatus measures this background when no cells are present and subtracts it from the signals the detectors record for each cell. This instrument background correction intro-duces statistical variation, which may cause the measurements that the detectors record for cells that have no cell-associated fluores-cence to be a little above or a little below the measurement set as the instrument zero. As a result, the corrected measurements distribute symmetrically around zero.

This distribution is often broadened sub-

0 103 104 1050

1000

2000

3000

400050%

30%

101 102 103 104 1050

50

1000

1500

2000

30%

50%

10 1

10 2

10 3

10 4

10 5

101 102 103 104 105 0 103 104 105

0

103

104

105

+

++

+

++

10 1

10 2

10 3

10 4

10 5

0

103

104

105

+ +

++++

Logarithmic ‘Logicle’

CD5

lgD

CD5

Figure 1 ‘Logicle’ displays provide improved representation of cells with minimal fluorescence. Cells with minimal fluorescence can be visualized with ‘logicle’ displays (right) but are ‘piled up’ on the axis with logarithmic displays (left). The true center of each gated population (median fluorescence value in each dimension; dark red crosses) matches the visual peak for that population in ‘logicle’ displays (right, top and middle) but does not match the visual peak in the logarithmic displays (left, top and middle). Because logarithmic scales cannot display cells with zero or negative values, these cells are ‘piled up’ on the axis in the logarithmic displays. However, they are properly visualized in the ‘logicle’ display (bottom right, red shaded region). Data provided by E. Ghosn (Stanford University, Stanford, California).

682 VOLUME 7 NUMBER 7 JULY 2006 NATURE IMMUNOLOGY

COMMENTARY

©20

06 N

atur

e Pu

blis

hing

Gro

up h

ttp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

subset that is visible in the ‘logicle’ displays (centered near zero on both axes) seems to be missing from the logarithmic displays. However, as indicated above, this subset is present but is represented mainly by data points and contours that are ‘piled up’ on the plot axes.

The location of the median fluorescence value in each dimension (Fig. 1, dark red crosses) further demonstrates the problem with logarithmic data visualization. By defi-nition, half of the data values in each rectan-gular region are greater than the median and half are less than that value. However, whereas the locations of the median values in the ‘log-icle’ displays (Fig. 1, right) correspond to the visual centers of the subsets, the locations of the median values in the logarithmic displays (Fig. 1, left) are substantially offset from the apparent peak of the subset. In fact, each of the subsets in the logarithmic display is broken up into a ‘false peak’ above the median value, a sparsely populated region between this peak and the baseline, and a ‘pileup’ of a large frac-tion of the cells on the baseline.

This artificial subdivision is also visible when data are plotted as a one-dimensional histogram on a logarithmic scale (Fig. 1, bot-tom left). The sharp rise at the lowest value on the scale reflects the ‘pileup’ of the minimally fluorescent cells at the lowest value on the scale. This ‘peak’ tends to be overlooked because it is nearly coincident with the axis. However, it represents roughly 40% of the minimally fluorescent population and 20% of the cells in the overall population. The ‘valley’ and the ‘false peak’ are also in the histogram, creating a total of three peaks rather than the two that actually exist. This problem does not occur in histograms plotted on ‘logicle’ axes (Fig. 1, bottom right), in which the minimally fluo-rescent cells form a peak centered at or near zero and extending symmetrically above and below the peak center, thereby representing the cells whose measured fluorescence values are below zero.

In two-dimensional logarithmic displays, cells whose fluorescence measurements are at or below zero in one or both dimensions are ‘piled up’ on the axes and are represented by contours along both axes in contour displays (Fig. 1, middle left) and by colored dots, vis-ible with keen eyes, on the axes in color density displays (Fig. 1, top left). The contours (and dots) range along a large proportion of both the x and y axes, indicating that cells with fluo-rescence values at or below zero in one dimen-sion may have equivalently low fluorescence in the other dimension or may have very high fluorescence values in this second dimension. Surprisingly, given how modest the contours along the axes seem, they represent about 40%

of the cells in the sample. Notably, these ‘miss-ing’ cells are readily visible in ‘logicle’ displays, which enable visualization of fluorescent mea-surements for essentially all cells in the sample, including those whose fluorescence values are at or below zero in either dimension (Fig. 1, right).

The importance of seeing less than nothingAs FACS instruments measure cell-associated fluorescence that can be essentially zero but not negative, how can a FACS data set contain values below zero and why are some cell popu-lations found to be distributed symmetrically around zero? The answer lies in the way the FACS instrument collects data and routinely corrects (compensates for) the data for spec-

tral overlap before visualization in histograms, dot plots and contour maps. In essence, back-ground light and electronic noise make a small but important contribution to the overall sig-nals the FACS instrument obtains for each cell. The FACS apparatus measures this background when no cells are present and subtracts it from the signals the detectors record for each cell. This instrument background correction intro-duces statistical variation, which may cause the measurements that the detectors record for cells that have no cell-associated fluores-cence to be a little above or a little below the measurement set as the instrument zero. As a result, the corrected measurements distribute symmetrically around zero.

This distribution is often broadened sub-

0 103 104 1050

1000

2000

3000

400050%

30%

101 102 103 104 1050

50

1000

1500

2000

30%

50%

10 1

10 2

10 3

10 4

10 5

101 102 103 104 105 0 103 104 105

0

103

104

105

+

++

+

++

10 1

10 2

10 3

10 4

10 5

0

103

104

105

+ +

++++

Logarithmic ‘Logicle’

CD5

lgD

CD5

Figure 1 ‘Logicle’ displays provide improved representation of cells with minimal fluorescence. Cells with minimal fluorescence can be visualized with ‘logicle’ displays (right) but are ‘piled up’ on the axis with logarithmic displays (left). The true center of each gated population (median fluorescence value in each dimension; dark red crosses) matches the visual peak for that population in ‘logicle’ displays (right, top and middle) but does not match the visual peak in the logarithmic displays (left, top and middle). Because logarithmic scales cannot display cells with zero or negative values, these cells are ‘piled up’ on the axis in the logarithmic displays. However, they are properly visualized in the ‘logicle’ display (bottom right, red shaded region). Data provided by E. Ghosn (Stanford University, Stanford, California).

682 VOLUME 7 NUMBER 7 JULY 2006 NATURE IMMUNOLOGY

COMMENTARY

©20

06 N

atur

e Pu

blis

hing

Gro

up h

ttp://

ww

w.n

atur

e.co

m/n

atur

eim

mun

olog

y

doi:10.1038/ni0706-681

Résultats préliminaires FACS

58

Détection par FACS avec le MoFlo™ XDP

59Signal-to-noise ratio: 17

424 événements 7098 événements

Limite de détection Les mesures de fluorescence sont en général plus sensibles que les mesures de transmittance: •  Le FSC est une mesure d’une petite variation sur un signal de forte

intensité. •  Les détecteurs de fluorescence mesurent des signaux relativement

forts par rapport à un signal de fond excessivement faible. Un marquage des membranes au PKH67 a été validé pour la détection de microparticules au FACS. Des essais sont en cours présentement.

60

TEM Détection des MPs

61

Microscopie électronique à transmission (UPEI)

doi:10.3402/jev.v1i0.19179

62

4. CARACTÉRISATION DES MICROVÉSICULES Microvésicules de RCC et immunomodulation

63

•  Taille des microparticules (FACS vs NanoSight) •  Quantification absolue et études cinétiques •  Visualisation des mitochondries par fluorescence et

microscopie confocale •  Quantification relative des freeMitos et des mitoMPs •  Marqueurs de surface et triage par FACS

Caractérisation des microparticules

64

Distribution de taille des MPs FACS et NanoSight

65

Taille des microparticules (MoFlo™ XDP)

66


67

250 nm

580 nm

790 nm

1 340 nm

Bruit de fond


68

250 nm

580 nm

790 nm

1 340 nm

Bruit de fond


69

250 nm

580 nm

790 nm

1 340 nm

Bruit de fond

600

800

1300

Taille approximative (nm)

250 200

Résolution de taille du MoFlo™ XDP: •  < 200 nm (en théorie) •  400-500 nm (en réalité) Puisque ~80 % des microvésicules sont de taille inférieure à 130 nm, l’information obtenue sur la distribution de taille est peu utile. En revanche, le NanoSight résolve la taille des particules entre 10 et 2 000 nm.

Résolution de taille

70

NanoSight (Nanoparticle Tracking Analysis) La dispersion de la lumière est mesurée avec une caméra vidéo ou un microscope. Le principe est similaire au FACS, mais l’échantillon est en suspension et est plus ou moins immobile. Mesure: •  Concentration •  Taille des particules •  Fluorescence spécifique On obtient un graphe de la distribution#de tailles des microparticules.

71

NanoSight (Nanoparticle Tracking Analysis) La dispersion de la lumière est mesurée avec une caméra vidéo ou un microscope. Le principe est similaire au FACS, mais l’échantillon est en suspension et est plus ou moins immobile. Mesure: •  Concentration •  Taille des particules •  Fluorescence spécifique On obtient un graphe de la distribution#de tailles des microparticules.

72

doi:10.3402/jev.v1i0.19179

Cinétique de production des MPs Quantification absolue

73

Simulation d’un time-course et quantification absolue à l’aide d’une concentration connue de billes de polystyrène (rouge), corrigée pour la taille de la population de cellules.

Cinétique de production des MPs

74

freeMitos et mitoMPs FACS, TEM et microscopie confocale

75

Est-ce qu’il y a beaucoup de mitochondries dans les cellules de RCC?

76

Tem

ps d

’incu

batio

n: 4

5 m

inut

es

77

Excellent! Prochaines étapes: FACS, TEM et microscopie confocale

78

freeMitos et mitoMPs

doi:10.1182/blood-2014-05-573543 (L. Boudreau et al.) 79

Marqueurs de surface et triage FACS

80

FACS, marqueurs de surface et triage

81

Les marqueurs de surface permettent d’identifier des sous-populations de EVs d’intérêt.

L’option triage du FACS a déjà été testée avec succès dans notre labo avec des microparticules de plaquettes.

Le rendement est relativement faible, mais cette technique permet néanmoins l’enrichissement de sous-populations de EVs.

FACS, marqueurs de surface et triage

82

5. EFFET DES MPs SUR LE SYSTÈME IMMUNITAIRE Microvésicules de RCC et immunomodulation

83

•  Macrophages et HL-60 •  Analyse du phénotype (M1 vs M2) par FACS •  Analyse fonctionnelle (cytokines) par ELISA

Effet des MPs sur le système immunitaire

84

Modèle cellulaire Polarisation des macrophages

85

La lignée cellulaire HL-60 est dérivée d’un patient atteint de leucémie aiguë promyélocytaire. C’est le modèle qu’on utilise actuellement afin de tester les agonistes silencieux de la nicotine.

HL-60 et macrophages

86

L’hypothèse nulle c’est que les microparticules n’ont aucun effet sur la polarisation des macrophages.

HL-60 et macrophages – H0

87

L’hypothèse alternative c’est que les microparticules ont un effet sur la polarisation des macrophages.

HL-60 et macrophages – H1

88

Analyse de phénotype (M1 vs M2) Polarisation des macrophages

89

Le concept de macrophages M1 et M2 est très joli, mais en pratique ils représentent des extrêmes d’un continuum de phénotypes (et il existe d’autres classifications). Il existe peu de marqueurs de surface utiles pour les M1, mais on peut utiliser une combinaison de marqueurs générals de macrophages et de marqueurs de M2 pour identifier nos populations de cellules. CD11b: macrophages M1 et M2 CD80 / CD86 high: M1; CD80 / CD86 low: M2 CX3CR1 low: M1; CX3CR1 high: M2 CCR2 / Ly-6c high: M1; CCR2 / Ly-6c low: M2 CD163: M2

FACS et marqueurs de surface

90

Analyse fonctionnelle (cytokines) Polarisation des macrophages

91

Cytokines et ELISA

92

Timeline approximatif

95

Perspectives futures

•  Quel rôle jouent les mitoMPs dans le RCC?

•  Quel est le rôle des MPs de RCC dans l’angiogénèse?

•  Quel est l’effet des MPs de RCC sur la polarisation des autres cellules du système immunitaire (eg.: cellules lymphoïdes)?

•  Opportunité pour études –omiques?

96

proposition 1.2

Documents