Microsc. Microanal. 20, 903–911, 2014 doi:10.1017/S1431927614000415
© MICROSCOPY SOCIETY OF AMERICA 2014
Ultrastructural Analysis of Healthy Synovial Fluids in Three Mammalian Species Constantin I. Matei,1,2,3 Caroline Boulocher,4,* Christelle Boulé,3 Michael Schramme,4 Eric Viguier,4 Thierry Roger,4 Yves Berthier,1 Ana-Maria Trunfio-Sfarghiu,1 and Marie-Geneviève Blanchin2 1
LaMCoS UMR5259, INSA-Lyon, CNRS, University of Lyon, 69621 Villeurbanne, France ILM, UMR5306-CNRS, University Claude Bernard Lyon 1,University of Lyon, 69622 Villeurbanne, France 3 CTmu, University Claude Bernard Lyon 1, University of Lyon, 69622 Villeurbanne, France 4 UPSP ICE 2011-03-101, VetAgro Sup, Veterinary Campus, University Claude Bernard Lyon 1, University of Lyon, 69280 Marcy lʼEtoile, France 2
Abstract: A better knowledge of synovial fluid (SF) ultrastructure is required to further understand normal joint lubrication and metabolism. The aim of the present study was to elucidate SF structural features in healthy joints from three mammalian species of different size compared with features in biomimetic SF. High-resolution structural analysis was performed using transmission electron microscopy (TEM) and scanning electron microscopy (SEM) and environmental SEM/wet scanning transmission electron microscopy mode complemented by TEM and SEM cryogenic methods. Laser-scanning confocal microscopy (LCM) was used to locate the main components of SF with respect to its ultrastructural organization. The present study showed that the ultrastructure of healthy SF is built from a network of vesicles with a size range from 100 to a few hundred nanometers. A multilayered organization of the vesicle membranes was observed with a thickness of about 5 nm. LCM study of biological SF compared with synthetic SF showed that the microvesicles consist of a lipid-based membrane enveloping a glycoprotein gel. Thus, healthy SF has a discontinuous ultrastructure based on a complex network of microvesicles. This finding offers novel perspectives for the diagnosis and treatment of synovial joint diseases. Key words: synovial fluid, ultrastructure, microvesicles, lipid membranes, lubrication
I NTRODUCTION Synovial joints, also known as diarthroses, are freely movable joints able to sustain various types of movement during an exceptionally long lifetime. These remarkable properties result from mechanisms of lubrication that have been studied since the 1950s, with particular interest in their tribological behavior and interactions between cartilage and synovial fluid (SF). The first hypothesis formulated to explain the durability of the synovial joint’s biolubrication relied on the mechanism of full-fluid film lubrication. This theory considered SF as a continuous and homogeneous full-fluid lubricant (Walker et al., 1968; Unsworth et al., 1975; Medley et al., 1984; Dowson & Jin, 1986). Nonetheless, this theory alone is insufficient to explain the lubrication in all phases of joint motion. Current physicochemical studies have demonstrated that interactions between the molecular assemblies of healthy SF generate an “elementary tribological pattern” where a nanometric layer of aqueous solution is trapped between lipid bilayers (Pawlack & Oloyede, 2008; Mirea et al., 2013). In particular, lipid vesicles filled with glycoprotein gel, which have been identified within the SF volume, seem to offset the roughness of the articular cartilage and lower the contact Received October 7, 2013; accepted February 13, 2014 *Corresponding author. [email protected]
pressure (Watanabe et al., 2003; Trunfio-Sfarghiu et al., 2007; Mirea et al., 2013). SF also supplies nutrients to the avascular articular cartilage, participates in the removal of by-products of cartilage metabolism, and delivers signaling molecules to the joint tissues (Bali & Shukla, 2001). It has been suggested that this transport occurs both by diffusion and “pumping” of nutrients and signaling molecules through the cartilage matrix. However, while impaired during joint disease, mechanisms of the pathways of transport within the SF and of SF lubrication properties are insufficiently known. Therefore, a better characterization of SF vesicular ultrastructure is required to better understand the lubrication mechanisms of the synovial joint and its contribution to normal joint metabolism. The ultrastructure of such lipid vesicular assemblies is challenging to study, as it can be dependent on sample preparation with respect to preservation of the SF structure. Knowledge of the vesicular organization is required to better understand lubrication in healthy synovial joints and especially the velocity accommodation mechanism (Mirea et al., 2013). To our knowledge no detailed study has been performed to characterize and assess the ultrastructure of such vesicular assemblies and to identify the organization of their membranes. In this context, the aim of the present study was to elucidate SF structural features of healthy joints from three mammalian species of different sizes. The structure of biological SF samples obtained from
904
Constantin I. Matei et al.
these three species was compared with biomimetic SF reconstituted ex vivo. SF samples are difficult to obtain and their volume is limited, whereas biomimetic SF reconstituted ex vivo provides an unlimited volume of fluid, offering the possibility of further validation of the structure found in biological samples. High-resolution (HR) structural investigation of the vesicular structures achieved using transmission electron microscopy (TEM), scanning electron microscopy (SEM), and environmental scanning electron microscopy (ESEM) are shown to characterize the size, morphology, and membrane structure in the vesicular assemblies. To ensure reliability of our studies with respect to the real SF structure, results from highresolution transmission electron microscopy investigations after negative staining of SF samples at room temperatures were compared with features of samples directly taken from SF and observed by ESEM/wet scanning transmission electron microscopy (Wet STEM) mode, and with those observed by SEM and TEM in samples obtained from SF by cryogenic methods. Laser-scanning confocal microscopy (LCM) was also used to determine the relative position of the vesicular structures and of the glycoprotein gel within the SF.
M ATERIALS AND M ETHODS SF Biological SF Healthy SF samples were collected in vivo from synovial joints of limbs of adult rats (males, 300 g, Wistar), dogs (males, 9–11 kg, Beagles), and horses (geldings, 490–590 kg, French trotters) in full compliance with the local ethical committee guidelines for animal protection (VetAgro Sup, Veterinary Campus of Lyon, Lyon, France) and in accordance with the legislation of the European Community. For the rat samples, SF was mixed with phosphate solution buffer during extraction because of the small size of the joints. Gel-in biomimetic SF Biomimetic healthy SF was reproduced according to the molecular composition of normal SF (Mirea et al., 2013): 0.3 mg/mL fluid phase phospholipids (DOPC 770375; Avanti Polar Lipids, Alabaster, Alabama, United States), 3 mg/mL hyaluronic acid (HA, H7630; Sigma-Aldrich, Steinheim, Germany), 18 mg/mL bovine serum albumin (A3059; SigmaAldrich), and 2 mg/mL γ-globulin (G4386, Sigma-Aldrich). Two solutions were first processed:
∙ a solution of lipids (0.3 g/L) with 1% (wt%) fluorescent
lipids β-bodipy C5-HPC (D-3803, Invitrogen; Saint Aubin, France) in a solvent composed of 90% chloroform and 10% ethanol in volume, ∙ a solution of HA (3 g/L) and bovine serum albumin (BSA) (18 g/L) with 1% (wt%) fluorescent BSA and γ-globulin (2 g/L) in phosphate buffered saline buffer pH = 7.4. This solution will further be referred to as “glycoprotein gel.” In order to obtain “included glycoprotein gel” (i.e., gel-in biomimetic SF in which the glycoprotein gel is
encapsulated by lipid layers within vesicular structures), an appropriate aliquot of the lipid/solvent solution was poured into a glass tube and the solvent was evaporated under a soft stream of nitrogen. The resulting lipid film was dried under high vacuum overnight to ensure the absence of organic solvent traces. The glycoprotein gel was then added over the dried lipid film covering the glass tube and maintained for 48 h at 37 °C on a magnetic stirrer. This solution will further be referred to as gel-in biomimetic SF.
SF ultrastructural studies ESEM/Wet STEM mode An FEI XL 30 FEG ESEM microscope (FEI, Hillsboro, Oregon, USA) operated at 30 kV was used with a nominal resolution of 2 nm. A TEM copper grid coated with a holey carbon film was linked to a Peltier stage mounted on the head of a TEM sample holder, being irradiated by the incident electron beam. A circular semiconductor detector made of two semiconductor diodes was placed below the sample allowing collection of electrons scattered at energies from 5 to 30 keV (Bogner et al., 2005, 2007). About 5 μL of SF with 10% dimethyl sulfoxide (DMSO) was deposited on the backside of a holey carbon-coated TEM support grid (the carbon film within the copper squares acted as retention basin) and placed below the standard gaseous atmosphere detector in the microscope chamber. Films of SF were thinned in situ to obtain a layer thin enough to allow incident electrons to pass through and be collected to form a Wet STEM image. The film thickness was controlled by evaporation with an optimized pump-down sequence to prevent excessive evaporation from and condensation on to the SF droplet (Bogner et al., 2005, 2007). The desired film thickness was controlled on the basis of the (p, T) equilibrium water-phase diagram. Using this ESEM/Wet STEM mode, samples of biological SF could be observed directly without any preparation. With the pressure in the observation chamber kept between 6 and 3 Torr and the SF temperature maintained between 3 and 5°C, the relative humidity of the chamber atmosphere was gradually decreased from 100% (i.e., the initial liquid state of the samples) to ~10%. TEM/negative staining Biological and biomimetic SF samples were prepared for TEM using a negative staining protocol. About 10 μL of SF was placed for 5 min on a 3 mm wide electron microscopy copper grid previously covered by carbon (Agar Scientific, Stansted Essex, United Kingdom; S160 Carbon Film 200 Mesh Cu). SF samples were incubated on the grid, and then stained with 2% phosphotungstic acid (PTA) to get an optimal contrast of the lipid membranes. The excess of fluid was drained with a filter paper. After air-drying, the specimens were observed by TEM in bright-field (BF) image mode, using a Topcon 02B microscope (Topcon, Capelle, The Netherlands) operated at 120 kV. An overview of the whole grid was achieved and then a maximal number of pictures showing well-resolved vesicles
Healthy Synovial Fluid Ultrastructure
905
Figure 1. Vesicular ultrastructure in rat synovial fluid (60% dilution) observed by wet scanning transmission electron microscope : 95% hydration degree (a), 40% hydration degree (b), and 10% hydration degree (c).
were recorded. These images were analyzed using ImageJ software to obtain the statistical distribution of the vesicle size. Accounting for the nonspherical shape of the vesicles and for the bidimensional character of the images, the size for a given vesicle was taken as the longest diameter measured on the vesicle’s micrograph. SEM and TEM/cryogenic methods A high-pressure freezing procedure was used for the SEM observations. A drop of SF (biological or biomimetic) was placed in a cup and quickly introduced into a high-pressure freezing device (Leica EM PACT; Leica, Nanterre, France). After freezing the cup was stored in liquid nitrogen. For the analysis, a fracture was created within the sample in the microscope. The observations were performed at 15 KV in a Quanta FEG 250 microscope (FEI) equipped with a cryotransfer for SEM (Alto 2500; Gatan, Pleasanton, CA, USA). For TEM observations, a freeze-fracture technique was used to obtain a replica of the fracture surfaces of the frozen sample. This method involved several sequential steps: sample freezing and fracturing, fracture cold metallization, and elimination of the sample matter in the replica. Practically, a droplet of SF (biological or biomimetic) was sandwiched between two copper cups and rapidly frozen in nitrogen in vacuo (−210°C) before being transferred to liquid nitrogen. The freeze-fracture was performed in a Reichert Jung cryofract (Munich, Germany). The replicas were finally deposited on 600 Mesh copper grids (G600HHS Fine Hexagonal Mesh Fine Bar Cooper; Gilder Grids, Grantham, Lincolnshire, United Kingdom)
and observed with TEM in BF mode, using a Topcon 02B microscope at 120 kV. LCM LCM was used to locate the glycoprotein gel and the lipid components within the biomimetic SF. For this purpose 250 µL of each biomimetic SF was deposited on a 170 µm thick borosilicate Lab-Tek Chamber Slide (Z734535, Sigma-Aldrich) and fluorescence images were produced on a TCS SP5 II inverted confocal microscope (Leica Microsystems CMS Gmbh, Wetzlar, Germany). A 63 × apochromatic water immersion objective with a NA of 1.2 was used for all experiments. Two channels were used simultaneously:
∙ Fluorescent BSA dye excitation was achieved with an
Argon laser at 488 nm and the fluorescence was collected between 480 and 530 nm. Thus, the glycoprotein gel was labeled in a green fluorescent light. ∙ β-bodipy C5-HPC dye excitation was achieved with a HeNe laser at 543 nm and the emission was collected in the 530–650 nm range. Therefore, lipids were labeled in a red fluorescent light.
RESULTS Biological SF ultrastructure ESEM/Wet STEM mode At 95% relative humidity (Fig. 1a), vesicular ultrastructure exhibiting a mean diameter ranging from a few hundred
906
Constantin I. Matei et al.
Figure 2. Vesicles in healthy synovial fluid from rat (a), dog (b), and horse (c), observed by transmission electron microscope after negative staining with 2% phosphotungstic acid.
nanometers to the order of µm were observed in the SF of all three species. The average diameter of such vesicles increased to a few micrometers at a relative humidity of ~40% (Fig. 1b) and the vesicular structure merged into larger coalescing networks (Fig. 1c) when the humidity was decreased to 10%. TEM/negative staining Vesicular ultrastructure encapsulated by multilayered membranes contrasted well with the black background (negative stain PTA). Multilayered vesicles (Fig. 2) were observed in all of the biological SF samples. Distribution of the number of vesicles according to their longest diameter is reported in Figure 3 for the different species (rat, dog, horse). The size distribution varies depending on species. Values of the longest vesicle diameter determined from the distribution using ImageJ software ranged between 112 and 457 nm for rat, 100 and 336 nm for dog, 61 and 340 nm for horse with the corresponding median values being 212, 157, and 224 nm for rat, dog, and horse, respectively (Fig. 3). SEM and TEM/cryogenic methods Both techniques revealed vesicular structures exhibiting a diameter from several hundred nanometers to a few
micrometers (Fig. 4). The image features suggested that a multilayered membrane possibly surrounded the vesicles. These individual vesicles coexisted with vesicular networks.
Ultrastructure of the gel-in biomimetic SF TEM/negative staining Multilayered vesicles with two to five layers of membranes (Fig. 5), as well as broken and fused vesicles, were observed in the biomimetic SF samples with a mean longest diameter between 100 and a few hundred nanometers. Distribution of the vesicle numbers versus their longest diameter in the biomimetic SF is reported in Figure 6: the diameter range between 84 and 294 nm with a median value of 149 nm. TEM/cryogenic methods Vesicles with a mean diameter of several hundred nanometers exhibiting a possible multilayered membrane and some networks of coalescing vesicles were observed in the gel-in biomimetic SF (Fig. 7). LCM Relatively spherical structures of different sizes were observed located at the same place when imaged under either red or green fluorescent light (Fig. 8).
Healthy Synovial Fluid Ultrastructure
907
DISCUSSION
Figure 3. Histograms of size distribution of vesicles observed by transmission electron microscope after negative staining for rat, dog, and horse synovial fluid samples fitted by mathematical curves whose maximum corresponds to the median values. The values of standard deviation correspond to 87, 47, and 74 nm for rat, dog, and horse, respectively.
Extracellular vesicles (EVs) are spherical particles enclosed within a phospholipid bilayer, secreted by various cell types in most, if not all biological fluids (serum, urine, breast milk, bronchoalveolar lavage fluid) (Burger et al., 2004; György et al., 2012; van der Pol et al., 2012). Owing to their small size and heterogeneity, there are numerous challenges in the detection, classification, and characterization of EVs (van der Pol et al., 2012). The diameter of EVs ranges from 30 nm to 1 µm depending on the shedding process, cell type, and cellular compartments from which they are issued (Burger et al., 2004; György et al., 2011, 2012; van der Pol et al., 2012). Recent reviews have highlighted the huge medical potential of EVs as clinically relevant biomarkers and as conveyors of bioactive therapeutic agents (Mause & Weber, 2010; György et al., 2011; Goldberg & Klein, 2012; Gonzalez et al., 2012; van der Pol et al., 2012;). Indeed, EVs have metabolic and signaling importance owing to their pivotal role in information transfer between cells and waste management. In addition, EVs have physical and biomechanical functions,
Figure 4. Different types of structures in horse synovial fluid observed by scanning electron microscope (a, b) and transmission electron microscope (c, d).
908
Constantin I. Matei et al.
Figure 5. Multilayered vesicles in biomimetic synovial fluid observed by transmission electron microscope after negative staining with 2% phosphotungstic acid.
Figure 6. Histograms of size distribution of vesicles observed by transmission electron microscope after negative staining for gel-in biomimetic synovial fluid samples fitted by mathematical curve whose maximum corresponds to the median value. The values of standard deviation correspond to 58 nm.
for example, the lamellar bodies identified on the surface of lung alveoli lower surface tension and provide a hydrophobic protective lining against environmental influences (Schmitz & Muller, 1991). Normal SF is only present in small volumes, which increases pre-analytical challenges in the characterization of its ultrastructure, in particular, in small animal species. Pre-analytical factors such as storage time, temperature, buffer, and agitation must be standardized to address technical pitfalls and potential artifacts (György et al., 2012). Herein, we analyzed the ultrastructure of SF in rats, dogs, and horses combining STEM and TEM as well as cryogenic observations to validate our results. All observations showed a
similar network of multilayered EVs with a size ranging from 100 to a few hundred nanometers. In the rat only a minute amount of SF is present within healthy joints. It was therefore necessary to use buffer for joint lavage (Barton et al., 2007), which thus modified the SF ultrastructure. The use of larger species allowed us to study undiluted samples. In the dog, the average volume was 0.24 mL (0.01–1.0 mL) (Sawyer, 1963), while in the horse a volume of up to 10 mL can be collected (Van Pelt, 1967). The overall structural organization of the SF was similar in all three species studied. However, we observed that the size distribution of the EVs varied between species, possibly because of some variation artificially generated during the purification process. Alternatively, this variation might reflect real interspecies differences associated with the different forces applied in joints of animals of different weights and with different sizes of the articular surface. Because spatial resolution of the Wet STEM (around 10 nm) was insufficient to study the ultrastructure of the lipid membrane of the SF EVs, TEM with a negative staining technique was used, enabling the lipid membranes to be clearly visualized by contrast with the black background of PTA. It revealed a multilayered organization of the EV membranes with a thickness of about 5 nm. As the preparation for Wet STEM and TEM with negative staining can potentially change the sample structure (crosslinking, bursting, or fusion of the vesicles) cryogenic methods with undiluted SF were used to confirm the structures previously visualized. In comparison with STEM and TEM, the results of cryogenic observation confirmed the coexistence of a distribution of vesicles with multilayered membranes with a vesicular network in both biological and mimetic SF. These observations on the nature of EV membranes in SF are consistent with a succession of multilamellar, phospholipid-based layers (Hills, 1989; Trunfio-Sfarghiu et al., 2008). Multilamellar lipid bilayers have been proposed as playing a crucial role in joint lubrication (Mavraki et al., 2009; Mirea et al., 2013). However, other components of these lipid-based membranes such as cholesterol, proteins for transportation of hydrated ions, water or charged macromolecules also have to be considered as they have been reported to facilitate cushioning in the intra-articular space (Pawlak & Oloyede, 2008). The spheroidal shape, size, distribution, and deformable aspect of the SF vesicles match those seen in cell-derived vesicles from other sources (György et al., 2011; Van der Pol et al., 2012). There is currently no clear consensus about the nomenclature of membrane vesicles (Van der Pol et al., 2012). In particular, our findings in healthy SF are consistent with the definition of microvesicles, i.e., structures surrounded by a phospholipid bilayer with a size range from 100 to a few hundred nanometers. Recent studies have highlighted the pivotal role of EVs in information transfer between cells and waste management (Conde-Vancells et al., 2008; Van der Pol et al., 2012). EV can transfer receptors and organelles between cells, and deliver mRNA and proteins into cells (Ratajcak et al., 2006;
Healthy Synovial Fluid Ultrastructure
909
Figure 7. Multilayered vesicular ultrastructure observed by transmission electron microscope in gel-in biomimetic synovial fluid prepared by freeze-fracture in overview (a) and at higher magnification (b).
Figure 8. Confocal microscopy images of reconstituted synovial fluid (SF) gel-in biomimetic SF: lipid vesicle component (a), glycoprotein gel component (b).
Meziani et al., 2008; Al-Nedaw et al., 2009). In this way, the SF EVs might also mediate the communication between synovial tissues lining cells. The chief constituent of SF is water with addition of HA, proteins, proteoglycans, and lipids (Pasquali-Ronchetti et al., 1997; Blewis et al., 2007). Physicochemical studies have shown that the association of HA with lipids results in lipids forming multilayered vesicular and tube-like structures filled with HA (Pasquali-Ronchetti et al., 1997; Crescenzi et al., 2004). Previously, we investigated molecular interactions between HA, seric proteins (i.e., albumin and γ-globulin) on one hand and lipid multilayers on the other (Mirea et al., 2013). Force spectroscopy measurements by atomic force microscopy showed a high affinity between HA and deposited lipid layers while the recorded affinity was significantly lower for the case of the seric proteins (albumin and γ-globulin). HA is a large molecule (≈12 µm long and
70 nm of gyratory radius) and, in addition, it possesses a high miscibility property with the seric proteins. This affinity may induce reticulations, raising the viscosity and giving a gel-like aspect, which has been referred to as a “glycoprotein gel.” The results of LCM studies suggest that the glycoprotein gel (green fluorescence) is incorporated within vesicles enveloped by multiple lipid layers (red fluorescence) since both fluorescent structures were seen located at the same place (Fig. 8): this conclusion thus accounts for the chemical affinity between lipids and glycoproteins, and for the presence of multilayered vesicles within the biomimetic SF. In the present study, LCM analyses and comparison between the ultrastructure of biological and biomimetic SF supported our hypothesis. As HA was included within the SF EV, and as EVs in general carry cytosolic components specific to their cellular origin (Schmitz & Muller, 1991; Gonzalez et al., 2012; Van der Pol et al., 2012), we propose that SF EVs are
910
Constantin I. Matei et al.
mainly derived from synoviocytes (Dobbie, 1996; Schwarz & Hills, 1996). It has been shown that the inclusion of lipid vesicles in HA, or the inclusion of HA in lipid vesicles, modifies the rheological behavior of aqueous solutions of HA (Crescenzi et al., 2004). This configuration allows maintenance of a low and stable friction coefficient when SF is squeezed between a 2-hydroxyethyl methacrylate lens simulating the cartilage and a glass substrate (Mirea et al., 2013). Such a vesicular organization is known to improve the lubricating properties and resistance to high pressure of the liposomes adsorbed to the surface of phospholipid layers under biological ion concentrations (Pozo-Navas et al., 2003; Goldberg et al., 2011). However, more studies are required in order to better understand the physical properties of these vesicular systems, as they are known to display complex mechanical behavior. In particular, the interactions between the bilayers together with the effects of the bilayer undulations need to be investigated (Pozo-Navas et al., 2003).
CONCLUSION To summarize, the present study reports a complete investigation by means of several techniques showing a complex structure of SF similar in three of mammalian species. This structure consists of multilamellar vesicles or networks, which are similar to those found in the biomimetic “gel-in” SF, studied previously by us (Mirea et al., 2013), and exhibits very good lubricating properties because of the location of the slip plane inside the lipid multilayer, which forms the walls of the vesicles or networks. This study offers novel perspectives for the diagnosis and treatment of synovial joint diseases as well as a new insight into the pathophysiological mechanism with respect to the ultrastructure of SF. Further work is needed to fully characterize the composition of these vesicles, their origin, their clearance, and their participation in joint metabolism and biomechanical behavior.
ACKNOWLEDGMENTS The authors gratefully thank A. Rivoire and B. Burdin for their help with the imaging cryomethods and suggestions, P. Boulanger and S.-S. Hong for their suggestions, M.-M. Sava and B. Munteanu for the biomimetic SF and M.-E. Duclos for the rat SF samples. This work was done as the part of Region Rhone-Alpes TriboPath Project and was also partially granted by the BQR Program of INSA-Lyon and by the French National Research Agency (ANR Project Blanc-2012 SIMI 4: “Biolubrication from phospholipid membranes”).
REFERENCES AL-NEDAW, K., MEEHAN, B. & RAK, J. (2009). Microvesicles messengers and mediators of tumor progression. Cell Cycle 8, 2014–2018.
BALI, R. & SHUKLA, A.K. (2001). Rheological effects of synovial fluid on nutritional transport. Tribol Lett 9, 233–239. BARTON, N.J., STEVENS, D.A., HUGHES, J.P., ROSSI, A.G., CHESSELL, I.P., REEVE, A.J. & MCQUEEN, D.S. (2007). Demonstration of a novel technique to quantitatively assess inflammatory mediators and cells in rat knee joints. J Inflamm (Lond) 4, 13. doi:10.1186/14769255-4-13. BLEWIS, M.E., NUGENT-DERFUS, G.E., SCHMIDT, T.A., SCHUMACHER, B.L. & SAH, R.L. (2007). A model of synovial fluid lubricant composition in normal and injured joints. Eur Cell Mater 13, 26–39. BOGNER, A., JOUNEAU, P.H., THOLLET, G., BASSET, D. & GAUTHIER, C. (2007). A history of scanning electron microscopy developments: Towards ‘Wet-STEM’ imaging. Micron 38, 390–401. BOGNER, A., THOLLET, G., BASSET, D., JOUNEAU, P.H. & GAUTHIER, C. (2005). Wet STEM: A new development in environmental SEM for imaging nano-objects included in a liquid phase. Ultramicroscopy 104, 290–301. BURGER, C., HAO, J., YING, Q., ISOBE, H., SAWAMURA, M., NAKAMURA, E. & CHU, B. (2004). Multilayer vesicles and vesicle clusters formed by the fullerene-based surfactant C60(CH3)5K. J Colloid Interface Sci 275, 632–641. CONDE-VANCELLS, J., RODRIGUEZ-SUAREZ, E., EMBADE, N., GIL, D., MATTHIESEN, R., VALLE, M., ELORTZA, F., LU, S.C., MATO, J.M. & FALCON-PEREZ, J.M. (2008). Characterization and comprehensive proteome profiling of exosomes secreted by hepatocytes. J Proteome Res 7, 5157–5166. CRESCENZI, V., TAGLIENTI, A. & PASQUALI-RONCHETTI, I. (2004). Supramolecular structures in aqueous hyaluronic acid and phospholipid vesicles mixtures: An electron microscopy and rheometric study. Colloids Surf 245, 133–135. DOBBIE, J.W. (1996). Surfactant protein A and lamellar bodies: A homologous secretory function of peritoneum, synovium, and lung. Perit Dial Int 16, 574–581. DOWSON, D. & JIN, Z.M. (1986). Micro-elastohydrodynamic lubrication of synovial joints. Eng Med 15, 63–65. GOLDBERG, R. & KLEIN, J. (2012). Liposomes as lubricants: Beyond drug delivery. Chem Phys Lipids 165, 374–381. GOLDBERG, R., SCHROEDER, A., BARENHOLZ, Y. & KLEIN, J. (2011). Interactions between adsorbed hydrogenated soy phosphatidylcholine (hspc) vesicles at physiologically high pressures and salt concentrations. Biophys J 100, 2403–2411. GONZALEZ, G., VITURET, C., DI PIETRO, A., CHANSON, M., BOULANGER, P. & HONG, S.S. (2012). Microparticle-mediated transfer of the viral receptors CAR and CD46, and the CFTR channel in a CHO cell model confers new functions to target cells. PLoS One 7(12), e52326. GYÖRGY, B., SZABÓ, T.G., PÁSZTÓI, M., PÁL, Z., MISJÁK, P., ARADI, B., LÁSZLÓ, V., PÁLLINGER, E., PAP, E., KITTEL, A., NAGY, G., FALUS, A. & BUZÁS, E.I. (2011). Membrane vesicles, current state-of-theart: Emerging role of extracellular vesicles. Cell Mol Life Sci 68, 2667–2688. GYÖRGY, B., SZABÓ, T.G., TURIÁK, L., WRIGHT, M., HERCZEG, P., LÉDECZI, Z., KITTEL, A., POLGÁR, A., TÓTH, K., DÉRFALVI, B., ZELENÁK, G., BÖRÖCZ, I., CARR, B., NAGY, G., VÉKEY, K., GAY, S., FALUS, A. & BUZÁS, E.I. (2012). Improved flow cytometric assessment reveals distinct microvesicle (cell-derived microparticle) signatures in joint diseases. PLoS One 7(11), e49726. HILLS, B.A. (1989). Oligolamellar lubrication of joints by surface active phospholipids. J Rheumatol 16, 82–91. MAUSE, F. & WEBER, C. (2010). Microparticles: Protagonists of a novel communication network for intercellular information exchange. Circ Res 107, 1047–1057.
Healthy Synovial Fluid Ultrastructure MAVRAKI, A., CANN, P.M., MISCHLER, S., BOEDO, S., BOOKER, J.F., BAYADA, G., VAN LEEUWEN, H., VAN OSTAYEN, R.A.J., BAYADA, G., MEURISSE, M.H., RENONDEAU, H., PAPKE, B.L., POZEBANCHUK, M., PARTHASARATHY, P.P., DAVIES, L., NILSSON, D., ISAKSSON, P., PRAKASH, B. & BANSAL, D. (2009). Friction and lubricant film thickness measurements on simulated synovial fluid. P I Mech Ing J-J Eng 223, 606–607. MEDLEY, J.B., DOWSON, D. & WRIGHT, V. (1984). Transient elastohydrodynamic lubrication models for the human ankle joint. Eng Med 13, 137–151. MEZIANI, F., TESSE, A. & ANDRIANTSITOHAINA, R. (2008). Microparticles are vectors of paradoxical information in vascular cells including the endothelium: Role in health and diseases. Pharmacol Rep 60, 75–84. MIREA, D., TRUNFIO-SFARGHIU, A.M., MATEI, C.I., MUNTEANU, B., PIEDNOIR, A., RIEU, J.P., BLANCHIN, M.G. & BERTHIER, Y. (2013). Role of the biomolecular interactions in the structure and tribological properties of synovial fluid. Tribol Int 59, 302–311. PASQUALI-RONCHETTI, I., QUAGLINO, D., MORI, G., BACCHELLI, B. & GHOSH, P. (1997). Hyaluronan-phospholipid interactions. J Struct Biol 120, 1–10. PAWLAK, Z. & OLOYEDE, A. (2008). Conceptualisation of articular cartilage as a giant reverse micelle: A hypothetical mechanism for joint biocushioning and lubrication. Biosystems 94, 193–201. POZO-NAVAS, B., RAGHUNATHAN, V.A., KATSARAS, J., RAPPOLT, M., LOHNER, K. & PABST, G. (2003). Discontinuous unbinding of lipid multibilayers. Phys Rev Lett 91(2), 028101. RATAJCZAK, J., WYSOCZYNSKI, M., HAYEK, F., JANOWSKA-WIECZOREK, A. & RATAJCZAK, M.Z. (2006). Membrane-derived microvesicles: Important and underappreciated mediators of cell-to-cell communication. Leukemia 20, 1487–1495.
911
SAWYER, D.C. (1963). Synovial fluid analysis of canine joints. J Am Vet Med Assoc 43, 609–612. SCHMITZ, G. & MULLER, G. (1991). Structure and function of lamellar bodies, lipid-protein complexes involved in storage and secretion of cellular lipids. J Lip Res 32, 1539–1570. SCHWARZ, I.M. & HILLS, B.A. (1996). Synovial surfactant: Lamellar bodies in type B synoviocytes and proteolipid in synovials fluid and the articular lining. J Rheumatol 35, 821–827. TRUNFIO-SFARGHIU, A.M., BERTHIER, Y., MEURISSE, M.H. & RIEU, J.P. (2007). Multiscale analysis of the tribological role of the molecular assembly of synovial fluid. Case of a healthy joint and implants. Tribol Int 40, 1500–1515. TRUNFIO-SFARGHIU, A.M., BERTHIER, Y., MEURISSE, M.H. & RIEU, J.P. (2008). Role of nanomechanical properties in the tribological performance of phospholipid biomimetic surfaces. Langmuir 24, 8765–8771. UNSWORTH, A., DOWSON, D. & WRIGHT, V. (1975). The frictional behavior of human synovial joints—Part I: Natural joints. J Lubrication Tech 97(3), 369–376. VAN DER POL, E., BÖING, A.N., HARRISON, P., STURK, A. & NIEUWLAND, R. (2012). Classification, functions, and clinical relevance of extracellular vesicles. Pharmacol Rev 64, 676–705. VAN PELT, R.W. (1967). Characteristics of normal equine tarsal synovial fluid. Can J Comp Med Vet Sci 31, 342–347. WALKER, P.S., DOWSON, D., LONGFIELD, M.D. & WRIGHT, V. (1968). Boosted lubrication of human joints by fluid enrichement and entrapment. Ann Rheum Dis 27, 512–520. WATANABE, M., LENG, C., TORIUMI, H., HAMADA, Y., AKAMATSU, N. & OHNO, S. (2003). Ultrastructural study of upper surface layer in rat articular cartilage by ‘‘in vivo cryotechnique’’ combined with various treatments. Med Electron Microsc 33, 16–24.
© MICROSCOPY SOCIETY OF AMERICA 2014
Ultrastructural Analysis of Healthy Synovial Fluids in Three Mammalian Species Constantin I. Matei,1,2,3 Caroline Boulocher,4,* Christelle Boulé,3 Michael Schramme,4 Eric Viguier,4 Thierry Roger,4 Yves Berthier,1 Ana-Maria Trunfio-Sfarghiu,1 and Marie-Geneviève Blanchin2 1
LaMCoS UMR5259, INSA-Lyon, CNRS, University of Lyon, 69621 Villeurbanne, France ILM, UMR5306-CNRS, University Claude Bernard Lyon 1,University of Lyon, 69622 Villeurbanne, France 3 CTmu, University Claude Bernard Lyon 1, University of Lyon, 69622 Villeurbanne, France 4 UPSP ICE 2011-03-101, VetAgro Sup, Veterinary Campus, University Claude Bernard Lyon 1, University of Lyon, 69280 Marcy lʼEtoile, France 2
Abstract: A better knowledge of synovial fluid (SF) ultrastructure is required to further understand normal joint lubrication and metabolism. The aim of the present study was to elucidate SF structural features in healthy joints from three mammalian species of different size compared with features in biomimetic SF. High-resolution structural analysis was performed using transmission electron microscopy (TEM) and scanning electron microscopy (SEM) and environmental SEM/wet scanning transmission electron microscopy mode complemented by TEM and SEM cryogenic methods. Laser-scanning confocal microscopy (LCM) was used to locate the main components of SF with respect to its ultrastructural organization. The present study showed that the ultrastructure of healthy SF is built from a network of vesicles with a size range from 100 to a few hundred nanometers. A multilayered organization of the vesicle membranes was observed with a thickness of about 5 nm. LCM study of biological SF compared with synthetic SF showed that the microvesicles consist of a lipid-based membrane enveloping a glycoprotein gel. Thus, healthy SF has a discontinuous ultrastructure based on a complex network of microvesicles. This finding offers novel perspectives for the diagnosis and treatment of synovial joint diseases. Key words: synovial fluid, ultrastructure, microvesicles, lipid membranes, lubrication
I NTRODUCTION Synovial joints, also known as diarthroses, are freely movable joints able to sustain various types of movement during an exceptionally long lifetime. These remarkable properties result from mechanisms of lubrication that have been studied since the 1950s, with particular interest in their tribological behavior and interactions between cartilage and synovial fluid (SF). The first hypothesis formulated to explain the durability of the synovial joint’s biolubrication relied on the mechanism of full-fluid film lubrication. This theory considered SF as a continuous and homogeneous full-fluid lubricant (Walker et al., 1968; Unsworth et al., 1975; Medley et al., 1984; Dowson & Jin, 1986). Nonetheless, this theory alone is insufficient to explain the lubrication in all phases of joint motion. Current physicochemical studies have demonstrated that interactions between the molecular assemblies of healthy SF generate an “elementary tribological pattern” where a nanometric layer of aqueous solution is trapped between lipid bilayers (Pawlack & Oloyede, 2008; Mirea et al., 2013). In particular, lipid vesicles filled with glycoprotein gel, which have been identified within the SF volume, seem to offset the roughness of the articular cartilage and lower the contact Received October 7, 2013; accepted February 13, 2014 *Corresponding author. [email protected]
pressure (Watanabe et al., 2003; Trunfio-Sfarghiu et al., 2007; Mirea et al., 2013). SF also supplies nutrients to the avascular articular cartilage, participates in the removal of by-products of cartilage metabolism, and delivers signaling molecules to the joint tissues (Bali & Shukla, 2001). It has been suggested that this transport occurs both by diffusion and “pumping” of nutrients and signaling molecules through the cartilage matrix. However, while impaired during joint disease, mechanisms of the pathways of transport within the SF and of SF lubrication properties are insufficiently known. Therefore, a better characterization of SF vesicular ultrastructure is required to better understand the lubrication mechanisms of the synovial joint and its contribution to normal joint metabolism. The ultrastructure of such lipid vesicular assemblies is challenging to study, as it can be dependent on sample preparation with respect to preservation of the SF structure. Knowledge of the vesicular organization is required to better understand lubrication in healthy synovial joints and especially the velocity accommodation mechanism (Mirea et al., 2013). To our knowledge no detailed study has been performed to characterize and assess the ultrastructure of such vesicular assemblies and to identify the organization of their membranes. In this context, the aim of the present study was to elucidate SF structural features of healthy joints from three mammalian species of different sizes. The structure of biological SF samples obtained from
904
Constantin I. Matei et al.
these three species was compared with biomimetic SF reconstituted ex vivo. SF samples are difficult to obtain and their volume is limited, whereas biomimetic SF reconstituted ex vivo provides an unlimited volume of fluid, offering the possibility of further validation of the structure found in biological samples. High-resolution (HR) structural investigation of the vesicular structures achieved using transmission electron microscopy (TEM), scanning electron microscopy (SEM), and environmental scanning electron microscopy (ESEM) are shown to characterize the size, morphology, and membrane structure in the vesicular assemblies. To ensure reliability of our studies with respect to the real SF structure, results from highresolution transmission electron microscopy investigations after negative staining of SF samples at room temperatures were compared with features of samples directly taken from SF and observed by ESEM/wet scanning transmission electron microscopy (Wet STEM) mode, and with those observed by SEM and TEM in samples obtained from SF by cryogenic methods. Laser-scanning confocal microscopy (LCM) was also used to determine the relative position of the vesicular structures and of the glycoprotein gel within the SF.
M ATERIALS AND M ETHODS SF Biological SF Healthy SF samples were collected in vivo from synovial joints of limbs of adult rats (males, 300 g, Wistar), dogs (males, 9–11 kg, Beagles), and horses (geldings, 490–590 kg, French trotters) in full compliance with the local ethical committee guidelines for animal protection (VetAgro Sup, Veterinary Campus of Lyon, Lyon, France) and in accordance with the legislation of the European Community. For the rat samples, SF was mixed with phosphate solution buffer during extraction because of the small size of the joints. Gel-in biomimetic SF Biomimetic healthy SF was reproduced according to the molecular composition of normal SF (Mirea et al., 2013): 0.3 mg/mL fluid phase phospholipids (DOPC 770375; Avanti Polar Lipids, Alabaster, Alabama, United States), 3 mg/mL hyaluronic acid (HA, H7630; Sigma-Aldrich, Steinheim, Germany), 18 mg/mL bovine serum albumin (A3059; SigmaAldrich), and 2 mg/mL γ-globulin (G4386, Sigma-Aldrich). Two solutions were first processed:
∙ a solution of lipids (0.3 g/L) with 1% (wt%) fluorescent
lipids β-bodipy C5-HPC (D-3803, Invitrogen; Saint Aubin, France) in a solvent composed of 90% chloroform and 10% ethanol in volume, ∙ a solution of HA (3 g/L) and bovine serum albumin (BSA) (18 g/L) with 1% (wt%) fluorescent BSA and γ-globulin (2 g/L) in phosphate buffered saline buffer pH = 7.4. This solution will further be referred to as “glycoprotein gel.” In order to obtain “included glycoprotein gel” (i.e., gel-in biomimetic SF in which the glycoprotein gel is
encapsulated by lipid layers within vesicular structures), an appropriate aliquot of the lipid/solvent solution was poured into a glass tube and the solvent was evaporated under a soft stream of nitrogen. The resulting lipid film was dried under high vacuum overnight to ensure the absence of organic solvent traces. The glycoprotein gel was then added over the dried lipid film covering the glass tube and maintained for 48 h at 37 °C on a magnetic stirrer. This solution will further be referred to as gel-in biomimetic SF.
SF ultrastructural studies ESEM/Wet STEM mode An FEI XL 30 FEG ESEM microscope (FEI, Hillsboro, Oregon, USA) operated at 30 kV was used with a nominal resolution of 2 nm. A TEM copper grid coated with a holey carbon film was linked to a Peltier stage mounted on the head of a TEM sample holder, being irradiated by the incident electron beam. A circular semiconductor detector made of two semiconductor diodes was placed below the sample allowing collection of electrons scattered at energies from 5 to 30 keV (Bogner et al., 2005, 2007). About 5 μL of SF with 10% dimethyl sulfoxide (DMSO) was deposited on the backside of a holey carbon-coated TEM support grid (the carbon film within the copper squares acted as retention basin) and placed below the standard gaseous atmosphere detector in the microscope chamber. Films of SF were thinned in situ to obtain a layer thin enough to allow incident electrons to pass through and be collected to form a Wet STEM image. The film thickness was controlled by evaporation with an optimized pump-down sequence to prevent excessive evaporation from and condensation on to the SF droplet (Bogner et al., 2005, 2007). The desired film thickness was controlled on the basis of the (p, T) equilibrium water-phase diagram. Using this ESEM/Wet STEM mode, samples of biological SF could be observed directly without any preparation. With the pressure in the observation chamber kept between 6 and 3 Torr and the SF temperature maintained between 3 and 5°C, the relative humidity of the chamber atmosphere was gradually decreased from 100% (i.e., the initial liquid state of the samples) to ~10%. TEM/negative staining Biological and biomimetic SF samples were prepared for TEM using a negative staining protocol. About 10 μL of SF was placed for 5 min on a 3 mm wide electron microscopy copper grid previously covered by carbon (Agar Scientific, Stansted Essex, United Kingdom; S160 Carbon Film 200 Mesh Cu). SF samples were incubated on the grid, and then stained with 2% phosphotungstic acid (PTA) to get an optimal contrast of the lipid membranes. The excess of fluid was drained with a filter paper. After air-drying, the specimens were observed by TEM in bright-field (BF) image mode, using a Topcon 02B microscope (Topcon, Capelle, The Netherlands) operated at 120 kV. An overview of the whole grid was achieved and then a maximal number of pictures showing well-resolved vesicles
Healthy Synovial Fluid Ultrastructure
905
Figure 1. Vesicular ultrastructure in rat synovial fluid (60% dilution) observed by wet scanning transmission electron microscope : 95% hydration degree (a), 40% hydration degree (b), and 10% hydration degree (c).
were recorded. These images were analyzed using ImageJ software to obtain the statistical distribution of the vesicle size. Accounting for the nonspherical shape of the vesicles and for the bidimensional character of the images, the size for a given vesicle was taken as the longest diameter measured on the vesicle’s micrograph. SEM and TEM/cryogenic methods A high-pressure freezing procedure was used for the SEM observations. A drop of SF (biological or biomimetic) was placed in a cup and quickly introduced into a high-pressure freezing device (Leica EM PACT; Leica, Nanterre, France). After freezing the cup was stored in liquid nitrogen. For the analysis, a fracture was created within the sample in the microscope. The observations were performed at 15 KV in a Quanta FEG 250 microscope (FEI) equipped with a cryotransfer for SEM (Alto 2500; Gatan, Pleasanton, CA, USA). For TEM observations, a freeze-fracture technique was used to obtain a replica of the fracture surfaces of the frozen sample. This method involved several sequential steps: sample freezing and fracturing, fracture cold metallization, and elimination of the sample matter in the replica. Practically, a droplet of SF (biological or biomimetic) was sandwiched between two copper cups and rapidly frozen in nitrogen in vacuo (−210°C) before being transferred to liquid nitrogen. The freeze-fracture was performed in a Reichert Jung cryofract (Munich, Germany). The replicas were finally deposited on 600 Mesh copper grids (G600HHS Fine Hexagonal Mesh Fine Bar Cooper; Gilder Grids, Grantham, Lincolnshire, United Kingdom)
and observed with TEM in BF mode, using a Topcon 02B microscope at 120 kV. LCM LCM was used to locate the glycoprotein gel and the lipid components within the biomimetic SF. For this purpose 250 µL of each biomimetic SF was deposited on a 170 µm thick borosilicate Lab-Tek Chamber Slide (Z734535, Sigma-Aldrich) and fluorescence images were produced on a TCS SP5 II inverted confocal microscope (Leica Microsystems CMS Gmbh, Wetzlar, Germany). A 63 × apochromatic water immersion objective with a NA of 1.2 was used for all experiments. Two channels were used simultaneously:
∙ Fluorescent BSA dye excitation was achieved with an
Argon laser at 488 nm and the fluorescence was collected between 480 and 530 nm. Thus, the glycoprotein gel was labeled in a green fluorescent light. ∙ β-bodipy C5-HPC dye excitation was achieved with a HeNe laser at 543 nm and the emission was collected in the 530–650 nm range. Therefore, lipids were labeled in a red fluorescent light.
RESULTS Biological SF ultrastructure ESEM/Wet STEM mode At 95% relative humidity (Fig. 1a), vesicular ultrastructure exhibiting a mean diameter ranging from a few hundred
906
Constantin I. Matei et al.
Figure 2. Vesicles in healthy synovial fluid from rat (a), dog (b), and horse (c), observed by transmission electron microscope after negative staining with 2% phosphotungstic acid.
nanometers to the order of µm were observed in the SF of all three species. The average diameter of such vesicles increased to a few micrometers at a relative humidity of ~40% (Fig. 1b) and the vesicular structure merged into larger coalescing networks (Fig. 1c) when the humidity was decreased to 10%. TEM/negative staining Vesicular ultrastructure encapsulated by multilayered membranes contrasted well with the black background (negative stain PTA). Multilayered vesicles (Fig. 2) were observed in all of the biological SF samples. Distribution of the number of vesicles according to their longest diameter is reported in Figure 3 for the different species (rat, dog, horse). The size distribution varies depending on species. Values of the longest vesicle diameter determined from the distribution using ImageJ software ranged between 112 and 457 nm for rat, 100 and 336 nm for dog, 61 and 340 nm for horse with the corresponding median values being 212, 157, and 224 nm for rat, dog, and horse, respectively (Fig. 3). SEM and TEM/cryogenic methods Both techniques revealed vesicular structures exhibiting a diameter from several hundred nanometers to a few
micrometers (Fig. 4). The image features suggested that a multilayered membrane possibly surrounded the vesicles. These individual vesicles coexisted with vesicular networks.
Ultrastructure of the gel-in biomimetic SF TEM/negative staining Multilayered vesicles with two to five layers of membranes (Fig. 5), as well as broken and fused vesicles, were observed in the biomimetic SF samples with a mean longest diameter between 100 and a few hundred nanometers. Distribution of the vesicle numbers versus their longest diameter in the biomimetic SF is reported in Figure 6: the diameter range between 84 and 294 nm with a median value of 149 nm. TEM/cryogenic methods Vesicles with a mean diameter of several hundred nanometers exhibiting a possible multilayered membrane and some networks of coalescing vesicles were observed in the gel-in biomimetic SF (Fig. 7). LCM Relatively spherical structures of different sizes were observed located at the same place when imaged under either red or green fluorescent light (Fig. 8).
Healthy Synovial Fluid Ultrastructure
907
DISCUSSION
Figure 3. Histograms of size distribution of vesicles observed by transmission electron microscope after negative staining for rat, dog, and horse synovial fluid samples fitted by mathematical curves whose maximum corresponds to the median values. The values of standard deviation correspond to 87, 47, and 74 nm for rat, dog, and horse, respectively.
Extracellular vesicles (EVs) are spherical particles enclosed within a phospholipid bilayer, secreted by various cell types in most, if not all biological fluids (serum, urine, breast milk, bronchoalveolar lavage fluid) (Burger et al., 2004; György et al., 2012; van der Pol et al., 2012). Owing to their small size and heterogeneity, there are numerous challenges in the detection, classification, and characterization of EVs (van der Pol et al., 2012). The diameter of EVs ranges from 30 nm to 1 µm depending on the shedding process, cell type, and cellular compartments from which they are issued (Burger et al., 2004; György et al., 2011, 2012; van der Pol et al., 2012). Recent reviews have highlighted the huge medical potential of EVs as clinically relevant biomarkers and as conveyors of bioactive therapeutic agents (Mause & Weber, 2010; György et al., 2011; Goldberg & Klein, 2012; Gonzalez et al., 2012; van der Pol et al., 2012;). Indeed, EVs have metabolic and signaling importance owing to their pivotal role in information transfer between cells and waste management. In addition, EVs have physical and biomechanical functions,
Figure 4. Different types of structures in horse synovial fluid observed by scanning electron microscope (a, b) and transmission electron microscope (c, d).
908
Constantin I. Matei et al.
Figure 5. Multilayered vesicles in biomimetic synovial fluid observed by transmission electron microscope after negative staining with 2% phosphotungstic acid.
Figure 6. Histograms of size distribution of vesicles observed by transmission electron microscope after negative staining for gel-in biomimetic synovial fluid samples fitted by mathematical curve whose maximum corresponds to the median value. The values of standard deviation correspond to 58 nm.
for example, the lamellar bodies identified on the surface of lung alveoli lower surface tension and provide a hydrophobic protective lining against environmental influences (Schmitz & Muller, 1991). Normal SF is only present in small volumes, which increases pre-analytical challenges in the characterization of its ultrastructure, in particular, in small animal species. Pre-analytical factors such as storage time, temperature, buffer, and agitation must be standardized to address technical pitfalls and potential artifacts (György et al., 2012). Herein, we analyzed the ultrastructure of SF in rats, dogs, and horses combining STEM and TEM as well as cryogenic observations to validate our results. All observations showed a
similar network of multilayered EVs with a size ranging from 100 to a few hundred nanometers. In the rat only a minute amount of SF is present within healthy joints. It was therefore necessary to use buffer for joint lavage (Barton et al., 2007), which thus modified the SF ultrastructure. The use of larger species allowed us to study undiluted samples. In the dog, the average volume was 0.24 mL (0.01–1.0 mL) (Sawyer, 1963), while in the horse a volume of up to 10 mL can be collected (Van Pelt, 1967). The overall structural organization of the SF was similar in all three species studied. However, we observed that the size distribution of the EVs varied between species, possibly because of some variation artificially generated during the purification process. Alternatively, this variation might reflect real interspecies differences associated with the different forces applied in joints of animals of different weights and with different sizes of the articular surface. Because spatial resolution of the Wet STEM (around 10 nm) was insufficient to study the ultrastructure of the lipid membrane of the SF EVs, TEM with a negative staining technique was used, enabling the lipid membranes to be clearly visualized by contrast with the black background of PTA. It revealed a multilayered organization of the EV membranes with a thickness of about 5 nm. As the preparation for Wet STEM and TEM with negative staining can potentially change the sample structure (crosslinking, bursting, or fusion of the vesicles) cryogenic methods with undiluted SF were used to confirm the structures previously visualized. In comparison with STEM and TEM, the results of cryogenic observation confirmed the coexistence of a distribution of vesicles with multilayered membranes with a vesicular network in both biological and mimetic SF. These observations on the nature of EV membranes in SF are consistent with a succession of multilamellar, phospholipid-based layers (Hills, 1989; Trunfio-Sfarghiu et al., 2008). Multilamellar lipid bilayers have been proposed as playing a crucial role in joint lubrication (Mavraki et al., 2009; Mirea et al., 2013). However, other components of these lipid-based membranes such as cholesterol, proteins for transportation of hydrated ions, water or charged macromolecules also have to be considered as they have been reported to facilitate cushioning in the intra-articular space (Pawlak & Oloyede, 2008). The spheroidal shape, size, distribution, and deformable aspect of the SF vesicles match those seen in cell-derived vesicles from other sources (György et al., 2011; Van der Pol et al., 2012). There is currently no clear consensus about the nomenclature of membrane vesicles (Van der Pol et al., 2012). In particular, our findings in healthy SF are consistent with the definition of microvesicles, i.e., structures surrounded by a phospholipid bilayer with a size range from 100 to a few hundred nanometers. Recent studies have highlighted the pivotal role of EVs in information transfer between cells and waste management (Conde-Vancells et al., 2008; Van der Pol et al., 2012). EV can transfer receptors and organelles between cells, and deliver mRNA and proteins into cells (Ratajcak et al., 2006;
Healthy Synovial Fluid Ultrastructure
909
Figure 7. Multilayered vesicular ultrastructure observed by transmission electron microscope in gel-in biomimetic synovial fluid prepared by freeze-fracture in overview (a) and at higher magnification (b).
Figure 8. Confocal microscopy images of reconstituted synovial fluid (SF) gel-in biomimetic SF: lipid vesicle component (a), glycoprotein gel component (b).
Meziani et al., 2008; Al-Nedaw et al., 2009). In this way, the SF EVs might also mediate the communication between synovial tissues lining cells. The chief constituent of SF is water with addition of HA, proteins, proteoglycans, and lipids (Pasquali-Ronchetti et al., 1997; Blewis et al., 2007). Physicochemical studies have shown that the association of HA with lipids results in lipids forming multilayered vesicular and tube-like structures filled with HA (Pasquali-Ronchetti et al., 1997; Crescenzi et al., 2004). Previously, we investigated molecular interactions between HA, seric proteins (i.e., albumin and γ-globulin) on one hand and lipid multilayers on the other (Mirea et al., 2013). Force spectroscopy measurements by atomic force microscopy showed a high affinity between HA and deposited lipid layers while the recorded affinity was significantly lower for the case of the seric proteins (albumin and γ-globulin). HA is a large molecule (≈12 µm long and
70 nm of gyratory radius) and, in addition, it possesses a high miscibility property with the seric proteins. This affinity may induce reticulations, raising the viscosity and giving a gel-like aspect, which has been referred to as a “glycoprotein gel.” The results of LCM studies suggest that the glycoprotein gel (green fluorescence) is incorporated within vesicles enveloped by multiple lipid layers (red fluorescence) since both fluorescent structures were seen located at the same place (Fig. 8): this conclusion thus accounts for the chemical affinity between lipids and glycoproteins, and for the presence of multilayered vesicles within the biomimetic SF. In the present study, LCM analyses and comparison between the ultrastructure of biological and biomimetic SF supported our hypothesis. As HA was included within the SF EV, and as EVs in general carry cytosolic components specific to their cellular origin (Schmitz & Muller, 1991; Gonzalez et al., 2012; Van der Pol et al., 2012), we propose that SF EVs are
910
Constantin I. Matei et al.
mainly derived from synoviocytes (Dobbie, 1996; Schwarz & Hills, 1996). It has been shown that the inclusion of lipid vesicles in HA, or the inclusion of HA in lipid vesicles, modifies the rheological behavior of aqueous solutions of HA (Crescenzi et al., 2004). This configuration allows maintenance of a low and stable friction coefficient when SF is squeezed between a 2-hydroxyethyl methacrylate lens simulating the cartilage and a glass substrate (Mirea et al., 2013). Such a vesicular organization is known to improve the lubricating properties and resistance to high pressure of the liposomes adsorbed to the surface of phospholipid layers under biological ion concentrations (Pozo-Navas et al., 2003; Goldberg et al., 2011). However, more studies are required in order to better understand the physical properties of these vesicular systems, as they are known to display complex mechanical behavior. In particular, the interactions between the bilayers together with the effects of the bilayer undulations need to be investigated (Pozo-Navas et al., 2003).
CONCLUSION To summarize, the present study reports a complete investigation by means of several techniques showing a complex structure of SF similar in three of mammalian species. This structure consists of multilamellar vesicles or networks, which are similar to those found in the biomimetic “gel-in” SF, studied previously by us (Mirea et al., 2013), and exhibits very good lubricating properties because of the location of the slip plane inside the lipid multilayer, which forms the walls of the vesicles or networks. This study offers novel perspectives for the diagnosis and treatment of synovial joint diseases as well as a new insight into the pathophysiological mechanism with respect to the ultrastructure of SF. Further work is needed to fully characterize the composition of these vesicles, their origin, their clearance, and their participation in joint metabolism and biomechanical behavior.
ACKNOWLEDGMENTS The authors gratefully thank A. Rivoire and B. Burdin for their help with the imaging cryomethods and suggestions, P. Boulanger and S.-S. Hong for their suggestions, M.-M. Sava and B. Munteanu for the biomimetic SF and M.-E. Duclos for the rat SF samples. This work was done as the part of Region Rhone-Alpes TriboPath Project and was also partially granted by the BQR Program of INSA-Lyon and by the French National Research Agency (ANR Project Blanc-2012 SIMI 4: “Biolubrication from phospholipid membranes”).
REFERENCES AL-NEDAW, K., MEEHAN, B. & RAK, J. (2009). Microvesicles messengers and mediators of tumor progression. Cell Cycle 8, 2014–2018.
BALI, R. & SHUKLA, A.K. (2001). Rheological effects of synovial fluid on nutritional transport. Tribol Lett 9, 233–239. BARTON, N.J., STEVENS, D.A., HUGHES, J.P., ROSSI, A.G., CHESSELL, I.P., REEVE, A.J. & MCQUEEN, D.S. (2007). Demonstration of a novel technique to quantitatively assess inflammatory mediators and cells in rat knee joints. J Inflamm (Lond) 4, 13. doi:10.1186/14769255-4-13. BLEWIS, M.E., NUGENT-DERFUS, G.E., SCHMIDT, T.A., SCHUMACHER, B.L. & SAH, R.L. (2007). A model of synovial fluid lubricant composition in normal and injured joints. Eur Cell Mater 13, 26–39. BOGNER, A., JOUNEAU, P.H., THOLLET, G., BASSET, D. & GAUTHIER, C. (2007). A history of scanning electron microscopy developments: Towards ‘Wet-STEM’ imaging. Micron 38, 390–401. BOGNER, A., THOLLET, G., BASSET, D., JOUNEAU, P.H. & GAUTHIER, C. (2005). Wet STEM: A new development in environmental SEM for imaging nano-objects included in a liquid phase. Ultramicroscopy 104, 290–301. BURGER, C., HAO, J., YING, Q., ISOBE, H., SAWAMURA, M., NAKAMURA, E. & CHU, B. (2004). Multilayer vesicles and vesicle clusters formed by the fullerene-based surfactant C60(CH3)5K. J Colloid Interface Sci 275, 632–641. CONDE-VANCELLS, J., RODRIGUEZ-SUAREZ, E., EMBADE, N., GIL, D., MATTHIESEN, R., VALLE, M., ELORTZA, F., LU, S.C., MATO, J.M. & FALCON-PEREZ, J.M. (2008). Characterization and comprehensive proteome profiling of exosomes secreted by hepatocytes. J Proteome Res 7, 5157–5166. CRESCENZI, V., TAGLIENTI, A. & PASQUALI-RONCHETTI, I. (2004). Supramolecular structures in aqueous hyaluronic acid and phospholipid vesicles mixtures: An electron microscopy and rheometric study. Colloids Surf 245, 133–135. DOBBIE, J.W. (1996). Surfactant protein A and lamellar bodies: A homologous secretory function of peritoneum, synovium, and lung. Perit Dial Int 16, 574–581. DOWSON, D. & JIN, Z.M. (1986). Micro-elastohydrodynamic lubrication of synovial joints. Eng Med 15, 63–65. GOLDBERG, R. & KLEIN, J. (2012). Liposomes as lubricants: Beyond drug delivery. Chem Phys Lipids 165, 374–381. GOLDBERG, R., SCHROEDER, A., BARENHOLZ, Y. & KLEIN, J. (2011). Interactions between adsorbed hydrogenated soy phosphatidylcholine (hspc) vesicles at physiologically high pressures and salt concentrations. Biophys J 100, 2403–2411. GONZALEZ, G., VITURET, C., DI PIETRO, A., CHANSON, M., BOULANGER, P. & HONG, S.S. (2012). Microparticle-mediated transfer of the viral receptors CAR and CD46, and the CFTR channel in a CHO cell model confers new functions to target cells. PLoS One 7(12), e52326. GYÖRGY, B., SZABÓ, T.G., PÁSZTÓI, M., PÁL, Z., MISJÁK, P., ARADI, B., LÁSZLÓ, V., PÁLLINGER, E., PAP, E., KITTEL, A., NAGY, G., FALUS, A. & BUZÁS, E.I. (2011). Membrane vesicles, current state-of-theart: Emerging role of extracellular vesicles. Cell Mol Life Sci 68, 2667–2688. GYÖRGY, B., SZABÓ, T.G., TURIÁK, L., WRIGHT, M., HERCZEG, P., LÉDECZI, Z., KITTEL, A., POLGÁR, A., TÓTH, K., DÉRFALVI, B., ZELENÁK, G., BÖRÖCZ, I., CARR, B., NAGY, G., VÉKEY, K., GAY, S., FALUS, A. & BUZÁS, E.I. (2012). Improved flow cytometric assessment reveals distinct microvesicle (cell-derived microparticle) signatures in joint diseases. PLoS One 7(11), e49726. HILLS, B.A. (1989). Oligolamellar lubrication of joints by surface active phospholipids. J Rheumatol 16, 82–91. MAUSE, F. & WEBER, C. (2010). Microparticles: Protagonists of a novel communication network for intercellular information exchange. Circ Res 107, 1047–1057.
Healthy Synovial Fluid Ultrastructure MAVRAKI, A., CANN, P.M., MISCHLER, S., BOEDO, S., BOOKER, J.F., BAYADA, G., VAN LEEUWEN, H., VAN OSTAYEN, R.A.J., BAYADA, G., MEURISSE, M.H., RENONDEAU, H., PAPKE, B.L., POZEBANCHUK, M., PARTHASARATHY, P.P., DAVIES, L., NILSSON, D., ISAKSSON, P., PRAKASH, B. & BANSAL, D. (2009). Friction and lubricant film thickness measurements on simulated synovial fluid. P I Mech Ing J-J Eng 223, 606–607. MEDLEY, J.B., DOWSON, D. & WRIGHT, V. (1984). Transient elastohydrodynamic lubrication models for the human ankle joint. Eng Med 13, 137–151. MEZIANI, F., TESSE, A. & ANDRIANTSITOHAINA, R. (2008). Microparticles are vectors of paradoxical information in vascular cells including the endothelium: Role in health and diseases. Pharmacol Rep 60, 75–84. MIREA, D., TRUNFIO-SFARGHIU, A.M., MATEI, C.I., MUNTEANU, B., PIEDNOIR, A., RIEU, J.P., BLANCHIN, M.G. & BERTHIER, Y. (2013). Role of the biomolecular interactions in the structure and tribological properties of synovial fluid. Tribol Int 59, 302–311. PASQUALI-RONCHETTI, I., QUAGLINO, D., MORI, G., BACCHELLI, B. & GHOSH, P. (1997). Hyaluronan-phospholipid interactions. J Struct Biol 120, 1–10. PAWLAK, Z. & OLOYEDE, A. (2008). Conceptualisation of articular cartilage as a giant reverse micelle: A hypothetical mechanism for joint biocushioning and lubrication. Biosystems 94, 193–201. POZO-NAVAS, B., RAGHUNATHAN, V.A., KATSARAS, J., RAPPOLT, M., LOHNER, K. & PABST, G. (2003). Discontinuous unbinding of lipid multibilayers. Phys Rev Lett 91(2), 028101. RATAJCZAK, J., WYSOCZYNSKI, M., HAYEK, F., JANOWSKA-WIECZOREK, A. & RATAJCZAK, M.Z. (2006). Membrane-derived microvesicles: Important and underappreciated mediators of cell-to-cell communication. Leukemia 20, 1487–1495.
911
SAWYER, D.C. (1963). Synovial fluid analysis of canine joints. J Am Vet Med Assoc 43, 609–612. SCHMITZ, G. & MULLER, G. (1991). Structure and function of lamellar bodies, lipid-protein complexes involved in storage and secretion of cellular lipids. J Lip Res 32, 1539–1570. SCHWARZ, I.M. & HILLS, B.A. (1996). Synovial surfactant: Lamellar bodies in type B synoviocytes and proteolipid in synovials fluid and the articular lining. J Rheumatol 35, 821–827. TRUNFIO-SFARGHIU, A.M., BERTHIER, Y., MEURISSE, M.H. & RIEU, J.P. (2007). Multiscale analysis of the tribological role of the molecular assembly of synovial fluid. Case of a healthy joint and implants. Tribol Int 40, 1500–1515. TRUNFIO-SFARGHIU, A.M., BERTHIER, Y., MEURISSE, M.H. & RIEU, J.P. (2008). Role of nanomechanical properties in the tribological performance of phospholipid biomimetic surfaces. Langmuir 24, 8765–8771. UNSWORTH, A., DOWSON, D. & WRIGHT, V. (1975). The frictional behavior of human synovial joints—Part I: Natural joints. J Lubrication Tech 97(3), 369–376. VAN DER POL, E., BÖING, A.N., HARRISON, P., STURK, A. & NIEUWLAND, R. (2012). Classification, functions, and clinical relevance of extracellular vesicles. Pharmacol Rev 64, 676–705. VAN PELT, R.W. (1967). Characteristics of normal equine tarsal synovial fluid. Can J Comp Med Vet Sci 31, 342–347. WALKER, P.S., DOWSON, D., LONGFIELD, M.D. & WRIGHT, V. (1968). Boosted lubrication of human joints by fluid enrichement and entrapment. Ann Rheum Dis 27, 512–520. WATANABE, M., LENG, C., TORIUMI, H., HAMADA, Y., AKAMATSU, N. & OHNO, S. (2003). Ultrastructural study of upper surface layer in rat articular cartilage by ‘‘in vivo cryotechnique’’ combined with various treatments. Med Electron Microsc 33, 16–24.
